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 1.0.2m and 1.0.2n [7 Dec 2017]
12 *) Read/write after SSL object in error state
14 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
15 mechanism. The intent was that if a fatal error occurred during a handshake
16 then OpenSSL would move into the error state and would immediately fail if
17 you attempted to continue the handshake. This works as designed for the
18 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
19 SSL_connect()), however due to a bug it does not work correctly if
20 SSL_read() or SSL_write() is called directly. In that scenario, if the
21 handshake fails then a fatal error will be returned in the initial function
22 call. If SSL_read()/SSL_write() is subsequently called by the application
23 for the same SSL object then it will succeed and the data is passed without
24 being decrypted/encrypted directly from the SSL/TLS record layer.
26 In order to exploit this issue an application bug would have to be present
27 that resulted in a call to SSL_read()/SSL_write() being issued after having
28 already received a fatal error.
30 This issue was reported to OpenSSL by David Benjamin (Google).
34 *) rsaz_1024_mul_avx2 overflow bug on x86_64
36 There is an overflow bug in the AVX2 Montgomery multiplication procedure
37 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
38 Analysis suggests that attacks against RSA and DSA as a result of this
39 defect would be very difficult to perform and are not believed likely.
40 Attacks against DH1024 are considered just feasible, because most of the
41 work necessary to deduce information about a private key may be performed
42 offline. The amount of resources required for such an attack would be
43 significant. However, for an attack on TLS to be meaningful, the server
44 would have to share the DH1024 private key among multiple clients, which is
45 no longer an option since CVE-2016-0701.
47 This only affects processors that support the AVX2 but not ADX extensions
48 like Intel Haswell (4th generation).
50 This issue was reported to OpenSSL by David Benjamin (Google). The issue
51 was originally found via the OSS-Fuzz project.
55 Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
57 *) bn_sqrx8x_internal carry bug on x86_64
59 There is a carry propagating bug in the x86_64 Montgomery squaring
60 procedure. No EC algorithms are affected. Analysis suggests that attacks
61 against RSA and DSA as a result of this defect would be very difficult to
62 perform and are not believed likely. Attacks against DH are considered just
63 feasible (although very difficult) because most of the work necessary to
64 deduce information about a private key may be performed offline. The amount
65 of resources required for such an attack would be very significant and
66 likely only accessible to a limited number of attackers. An attacker would
67 additionally need online access to an unpatched system using the target
68 private key in a scenario with persistent DH parameters and a private
69 key that is shared between multiple clients.
71 This only affects processors that support the BMI1, BMI2 and ADX extensions
72 like Intel Broadwell (5th generation) and later or AMD Ryzen.
74 This issue was reported to OpenSSL by the OSS-Fuzz project.
78 *) Malformed X.509 IPAddressFamily could cause OOB read
80 If an X.509 certificate has a malformed IPAddressFamily extension,
81 OpenSSL could do a one-byte buffer overread. The most likely result
82 would be an erroneous display of the certificate in text format.
84 This issue was reported to OpenSSL by the OSS-Fuzz project.
88 Changes between 1.0.2k and 1.0.2l [25 May 2017]
90 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
91 platform rather than 'mingw'.
94 Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
96 *) Truncated packet could crash via OOB read
98 If one side of an SSL/TLS path is running on a 32-bit host and a specific
99 cipher is being used, then a truncated packet can cause that host to
100 perform an out-of-bounds read, usually resulting in a crash.
102 This issue was reported to OpenSSL by Robert Święcki of Google.
106 *) BN_mod_exp may produce incorrect results on x86_64
108 There is a carry propagating bug in the x86_64 Montgomery squaring
109 procedure. No EC algorithms are affected. Analysis suggests that attacks
110 against RSA and DSA as a result of this defect would be very difficult to
111 perform and are not believed likely. Attacks against DH are considered just
112 feasible (although very difficult) because most of the work necessary to
113 deduce information about a private key may be performed offline. The amount
114 of resources required for such an attack would be very significant and
115 likely only accessible to a limited number of attackers. An attacker would
116 additionally need online access to an unpatched system using the target
117 private key in a scenario with persistent DH parameters and a private
118 key that is shared between multiple clients. For example this can occur by
119 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
120 similar to CVE-2015-3193 but must be treated as a separate problem.
122 This issue was reported to OpenSSL by the OSS-Fuzz project.
126 *) Montgomery multiplication may produce incorrect results
128 There is a carry propagating bug in the Broadwell-specific Montgomery
129 multiplication procedure that handles input lengths divisible by, but
130 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
131 and DH private keys are impossible. This is because the subroutine in
132 question is not used in operations with the private key itself and an input
133 of the attacker's direct choice. Otherwise the bug can manifest itself as
134 transient authentication and key negotiation failures or reproducible
135 erroneous outcome of public-key operations with specially crafted input.
136 Among EC algorithms only Brainpool P-512 curves are affected and one
137 presumably can attack ECDH key negotiation. Impact was not analyzed in
138 detail, because pre-requisites for attack are considered unlikely. Namely
139 multiple clients have to choose the curve in question and the server has to
140 share the private key among them, neither of which is default behaviour.
141 Even then only clients that chose the curve will be affected.
143 This issue was publicly reported as transient failures and was not
144 initially recognized as a security issue. Thanks to Richard Morgan for
145 providing reproducible case.
149 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
150 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
151 prevent issues where no progress is being made and the peer continually
152 sends unrecognised record types, using up resources processing them.
155 Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
157 *) Missing CRL sanity check
159 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
160 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
161 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
163 This issue only affects the OpenSSL 1.0.2i
167 Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
169 *) OCSP Status Request extension unbounded memory growth
171 A malicious client can send an excessively large OCSP Status Request
172 extension. If that client continually requests renegotiation, sending a
173 large OCSP Status Request extension each time, then there will be unbounded
174 memory growth on the server. This will eventually lead to a Denial Of
175 Service attack through memory exhaustion. Servers with a default
176 configuration are vulnerable even if they do not support OCSP. Builds using
177 the "no-ocsp" build time option are not affected.
179 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
183 *) In order to mitigate the SWEET32 attack, the DES ciphers were moved from
186 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
191 *) OOB write in MDC2_Update()
193 An overflow can occur in MDC2_Update() either if called directly or
194 through the EVP_DigestUpdate() function using MDC2. If an attacker
195 is able to supply very large amounts of input data after a previous
196 call to EVP_EncryptUpdate() with a partial block then a length check
197 can overflow resulting in a heap corruption.
199 The amount of data needed is comparable to SIZE_MAX which is impractical
202 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
206 *) Malformed SHA512 ticket DoS
208 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
209 DoS attack where a malformed ticket will result in an OOB read which will
212 The use of SHA512 in TLS session tickets is comparatively rare as it requires
213 a custom server callback and ticket lookup mechanism.
215 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
219 *) OOB write in BN_bn2dec()
221 The function BN_bn2dec() does not check the return value of BN_div_word().
222 This can cause an OOB write if an application uses this function with an
223 overly large BIGNUM. This could be a problem if an overly large certificate
224 or CRL is printed out from an untrusted source. TLS is not affected because
225 record limits will reject an oversized certificate before it is parsed.
227 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
231 *) OOB read in TS_OBJ_print_bio()
233 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
234 the total length the OID text representation would use and not the amount
235 of data written. This will result in OOB reads when large OIDs are
238 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
242 *) Pointer arithmetic undefined behaviour
244 Avoid some undefined pointer arithmetic
246 A common idiom in the codebase is to check limits in the following manner:
249 Where "p" points to some malloc'd data of SIZE bytes and
252 "len" here could be from some externally supplied data (e.g. from a TLS
255 The rules of C pointer arithmetic are such that "p + len" is only well
256 defined where len <= SIZE. Therefore the above idiom is actually
259 For example this could cause problems if some malloc implementation
260 provides an address for "p" such that "p + len" actually overflows for
261 values of len that are too big and therefore p + len < limit.
263 This issue was reported to OpenSSL by Guido Vranken
267 *) Constant time flag not preserved in DSA signing
269 Operations in the DSA signing algorithm should run in constant time in
270 order to avoid side channel attacks. A flaw in the OpenSSL DSA
271 implementation means that a non-constant time codepath is followed for
272 certain operations. This has been demonstrated through a cache-timing
273 attack to be sufficient for an attacker to recover the private DSA key.
275 This issue was reported by César Pereida (Aalto University), Billy Brumley
276 (Tampere University of Technology), and Yuval Yarom (The University of
281 *) DTLS buffered message DoS
283 In a DTLS connection where handshake messages are delivered out-of-order
284 those messages that OpenSSL is not yet ready to process will be buffered
285 for later use. Under certain circumstances, a flaw in the logic means that
286 those messages do not get removed from the buffer even though the handshake
287 has been completed. An attacker could force up to approx. 15 messages to
288 remain in the buffer when they are no longer required. These messages will
289 be cleared when the DTLS connection is closed. The default maximum size for
290 a message is 100k. Therefore the attacker could force an additional 1500k
291 to be consumed per connection. By opening many simulataneous connections an
292 attacker could cause a DoS attack through memory exhaustion.
294 This issue was reported to OpenSSL by Quan Luo.
298 *) DTLS replay protection DoS
300 A flaw in the DTLS replay attack protection mechanism means that records
301 that arrive for future epochs update the replay protection "window" before
302 the MAC for the record has been validated. This could be exploited by an
303 attacker by sending a record for the next epoch (which does not have to
304 decrypt or have a valid MAC), with a very large sequence number. This means
305 that all subsequent legitimate packets are dropped causing a denial of
306 service for a specific DTLS connection.
308 This issue was reported to OpenSSL by the OCAP audit team.
312 *) Certificate message OOB reads
314 In OpenSSL 1.0.2 and earlier some missing message length checks can result
315 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
316 theoretical DoS risk but this has not been observed in practice on common
319 The messages affected are client certificate, client certificate request
320 and server certificate. As a result the attack can only be performed
321 against a client or a server which enables client authentication.
323 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
327 Changes between 1.0.2g and 1.0.2h [3 May 2016]
329 *) Prevent padding oracle in AES-NI CBC MAC check
331 A MITM attacker can use a padding oracle attack to decrypt traffic
332 when the connection uses an AES CBC cipher and the server support
335 This issue was introduced as part of the fix for Lucky 13 padding
336 attack (CVE-2013-0169). The padding check was rewritten to be in
337 constant time by making sure that always the same bytes are read and
338 compared against either the MAC or padding bytes. But it no longer
339 checked that there was enough data to have both the MAC and padding
342 This issue was reported by Juraj Somorovsky using TLS-Attacker.
346 *) Fix EVP_EncodeUpdate overflow
348 An overflow can occur in the EVP_EncodeUpdate() function which is used for
349 Base64 encoding of binary data. If an attacker is able to supply very large
350 amounts of input data then a length check can overflow resulting in a heap
353 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
354 the PEM_write_bio* family of functions. These are mainly used within the
355 OpenSSL command line applications, so any application which processes data
356 from an untrusted source and outputs it as a PEM file should be considered
357 vulnerable to this issue. User applications that call these APIs directly
358 with large amounts of untrusted data may also be vulnerable.
360 This issue was reported by Guido Vranken.
364 *) Fix EVP_EncryptUpdate overflow
366 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
367 is able to supply very large amounts of input data after a previous call to
368 EVP_EncryptUpdate() with a partial block then a length check can overflow
369 resulting in a heap corruption. Following an analysis of all OpenSSL
370 internal usage of the EVP_EncryptUpdate() function all usage is one of two
371 forms. The first form is where the EVP_EncryptUpdate() call is known to be
372 the first called function after an EVP_EncryptInit(), and therefore that
373 specific call must be safe. The second form is where the length passed to
374 EVP_EncryptUpdate() can be seen from the code to be some small value and
375 therefore there is no possibility of an overflow. Since all instances are
376 one of these two forms, it is believed that there can be no overflows in
377 internal code due to this problem. It should be noted that
378 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
379 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
380 of these calls have also been analysed too and it is believed there are no
381 instances in internal usage where an overflow could occur.
383 This issue was reported by Guido Vranken.
387 *) Prevent ASN.1 BIO excessive memory allocation
389 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
390 a short invalid encoding can casuse allocation of large amounts of memory
391 potentially consuming excessive resources or exhausting memory.
393 Any application parsing untrusted data through d2i BIO functions is
394 affected. The memory based functions such as d2i_X509() are *not* affected.
395 Since the memory based functions are used by the TLS library, TLS
396 applications are not affected.
398 This issue was reported by Brian Carpenter.
404 ASN1 Strings that are over 1024 bytes can cause an overread in applications
405 using the X509_NAME_oneline() function on EBCDIC systems. This could result
406 in arbitrary stack data being returned in the buffer.
408 This issue was reported by Guido Vranken.
412 *) Modify behavior of ALPN to invoke callback after SNI/servername
413 callback, such that updates to the SSL_CTX affect ALPN.
416 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
420 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
421 methods are enabled and ssl2 is disabled the methods return NULL.
424 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
426 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
427 Builds that are not configured with "enable-weak-ssl-ciphers" will not
428 provide any "EXPORT" or "LOW" strength ciphers.
431 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
432 is by default disabled at build-time. Builds that are not configured with
433 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
434 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
435 will need to explicitly call either of:
437 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
439 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
441 as appropriate. Even if either of those is used, or the application
442 explicitly uses the version-specific SSLv2_method() or its client and
443 server variants, SSLv2 ciphers vulnerable to exhaustive search key
444 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
445 ciphers, and SSLv2 56-bit DES are no longer available.
449 *) Fix a double-free in DSA code
451 A double free bug was discovered when OpenSSL parses malformed DSA private
452 keys and could lead to a DoS attack or memory corruption for applications
453 that receive DSA private keys from untrusted sources. This scenario is
456 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
461 *) Disable SRP fake user seed to address a server memory leak.
463 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
465 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
466 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
467 was changed to ignore the "fake user" SRP seed, even if the seed
470 Users should use SRP_VBASE_get1_by_user instead. Note that in
471 SRP_VBASE_get1_by_user, caller must free the returned value. Note
472 also that even though configuring the SRP seed attempts to hide
473 invalid usernames by continuing the handshake with fake
474 credentials, this behaviour is not constant time and no strong
475 guarantees are made that the handshake is indistinguishable from
476 that of a valid user.
480 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
482 In the BN_hex2bn function the number of hex digits is calculated using an
483 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
484 large values of |i| this can result in |bn_expand| not allocating any
485 memory because |i * 4| is negative. This can leave the internal BIGNUM data
486 field as NULL leading to a subsequent NULL ptr deref. For very large values
487 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
488 In this case memory is allocated to the internal BIGNUM data field, but it
489 is insufficiently sized leading to heap corruption. A similar issue exists
490 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
491 is ever called by user applications with very large untrusted hex/dec data.
492 This is anticipated to be a rare occurrence.
494 All OpenSSL internal usage of these functions use data that is not expected
495 to be untrusted, e.g. config file data or application command line
496 arguments. If user developed applications generate config file data based
497 on untrusted data then it is possible that this could also lead to security
498 consequences. This is also anticipated to be rare.
500 This issue was reported to OpenSSL by Guido Vranken.
504 *) Fix memory issues in BIO_*printf functions
506 The internal |fmtstr| function used in processing a "%s" format string in
507 the BIO_*printf functions could overflow while calculating the length of a
508 string and cause an OOB read when printing very long strings.
510 Additionally the internal |doapr_outch| function can attempt to write to an
511 OOB memory location (at an offset from the NULL pointer) in the event of a
512 memory allocation failure. In 1.0.2 and below this could be caused where
513 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
514 could be in processing a very long "%s" format string. Memory leaks can
517 The first issue may mask the second issue dependent on compiler behaviour.
518 These problems could enable attacks where large amounts of untrusted data
519 is passed to the BIO_*printf functions. If applications use these functions
520 in this way then they could be vulnerable. OpenSSL itself uses these
521 functions when printing out human-readable dumps of ASN.1 data. Therefore
522 applications that print this data could be vulnerable if the data is from
523 untrusted sources. OpenSSL command line applications could also be
524 vulnerable where they print out ASN.1 data, or if untrusted data is passed
525 as command line arguments.
527 Libssl is not considered directly vulnerable. Additionally certificates etc
528 received via remote connections via libssl are also unlikely to be able to
529 trigger these issues because of message size limits enforced within libssl.
531 This issue was reported to OpenSSL Guido Vranken.
535 *) Side channel attack on modular exponentiation
537 A side-channel attack was found which makes use of cache-bank conflicts on
538 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
539 of RSA keys. The ability to exploit this issue is limited as it relies on
540 an attacker who has control of code in a thread running on the same
541 hyper-threaded core as the victim thread which is performing decryptions.
543 This issue was reported to OpenSSL by Yuval Yarom, The University of
544 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
545 Nadia Heninger, University of Pennsylvania with more information at
546 http://cachebleed.info.
550 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
551 if no keysize is specified with default_bits. This fixes an
552 omission in an earlier change that changed all RSA/DSA key generation
553 apps to use 2048 bits by default.
556 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
558 *) DH small subgroups
560 Historically OpenSSL only ever generated DH parameters based on "safe"
561 primes. More recently (in version 1.0.2) support was provided for
562 generating X9.42 style parameter files such as those required for RFC 5114
563 support. The primes used in such files may not be "safe". Where an
564 application is using DH configured with parameters based on primes that are
565 not "safe" then an attacker could use this fact to find a peer's private
566 DH exponent. This attack requires that the attacker complete multiple
567 handshakes in which the peer uses the same private DH exponent. For example
568 this could be used to discover a TLS server's private DH exponent if it's
569 reusing the private DH exponent or it's using a static DH ciphersuite.
571 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
572 TLS. It is not on by default. If the option is not set then the server
573 reuses the same private DH exponent for the life of the server process and
574 would be vulnerable to this attack. It is believed that many popular
575 applications do set this option and would therefore not be at risk.
577 The fix for this issue adds an additional check where a "q" parameter is
578 available (as is the case in X9.42 based parameters). This detects the
579 only known attack, and is the only possible defense for static DH
580 ciphersuites. This could have some performance impact.
582 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
583 default and cannot be disabled. This could have some performance impact.
585 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
589 *) SSLv2 doesn't block disabled ciphers
591 A malicious client can negotiate SSLv2 ciphers that have been disabled on
592 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
593 been disabled, provided that the SSLv2 protocol was not also disabled via
596 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
597 and Sebastian Schinzel.
601 *) Reject DH handshakes with parameters shorter than 1024 bits.
604 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
606 *) BN_mod_exp may produce incorrect results on x86_64
608 There is a carry propagating bug in the x86_64 Montgomery squaring
609 procedure. No EC algorithms are affected. Analysis suggests that attacks
610 against RSA and DSA as a result of this defect would be very difficult to
611 perform and are not believed likely. Attacks against DH are considered just
612 feasible (although very difficult) because most of the work necessary to
613 deduce information about a private key may be performed offline. The amount
614 of resources required for such an attack would be very significant and
615 likely only accessible to a limited number of attackers. An attacker would
616 additionally need online access to an unpatched system using the target
617 private key in a scenario with persistent DH parameters and a private
618 key that is shared between multiple clients. For example this can occur by
619 default in OpenSSL DHE based SSL/TLS ciphersuites.
621 This issue was reported to OpenSSL by Hanno Böck.
625 *) Certificate verify crash with missing PSS parameter
627 The signature verification routines will crash with a NULL pointer
628 dereference if presented with an ASN.1 signature using the RSA PSS
629 algorithm and absent mask generation function parameter. Since these
630 routines are used to verify certificate signature algorithms this can be
631 used to crash any certificate verification operation and exploited in a
632 DoS attack. Any application which performs certificate verification is
633 vulnerable including OpenSSL clients and servers which enable client
636 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
640 *) X509_ATTRIBUTE memory leak
642 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
643 memory. This structure is used by the PKCS#7 and CMS routines so any
644 application which reads PKCS#7 or CMS data from untrusted sources is
645 affected. SSL/TLS is not affected.
647 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
652 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
653 This changes the decoding behaviour for some invalid messages,
654 though the change is mostly in the more lenient direction, and
655 legacy behaviour is preserved as much as possible.
658 *) In DSA_generate_parameters_ex, if the provided seed is too short,
659 use a random seed, as already documented.
660 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
662 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
664 *) Alternate chains certificate forgery
666 During certificate verfification, OpenSSL will attempt to find an
667 alternative certificate chain if the first attempt to build such a chain
668 fails. An error in the implementation of this logic can mean that an
669 attacker could cause certain checks on untrusted certificates to be
670 bypassed, such as the CA flag, enabling them to use a valid leaf
671 certificate to act as a CA and "issue" an invalid certificate.
673 This issue was reported to OpenSSL by Adam Langley/David Benjamin
678 *) Race condition handling PSK identify hint
680 If PSK identity hints are received by a multi-threaded client then
681 the values are wrongly updated in the parent SSL_CTX structure. This can
682 result in a race condition potentially leading to a double free of the
687 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
689 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
690 incompatibility in the handling of HMAC. The previous ABI has now been
693 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
695 *) Malformed ECParameters causes infinite loop
697 When processing an ECParameters structure OpenSSL enters an infinite loop
698 if the curve specified is over a specially malformed binary polynomial
701 This can be used to perform denial of service against any
702 system which processes public keys, certificate requests or
703 certificates. This includes TLS clients and TLS servers with
704 client authentication enabled.
706 This issue was reported to OpenSSL by Joseph Barr-Pixton.
710 *) Exploitable out-of-bounds read in X509_cmp_time
712 X509_cmp_time does not properly check the length of the ASN1_TIME
713 string and can read a few bytes out of bounds. In addition,
714 X509_cmp_time accepts an arbitrary number of fractional seconds in the
717 An attacker can use this to craft malformed certificates and CRLs of
718 various sizes and potentially cause a segmentation fault, resulting in
719 a DoS on applications that verify certificates or CRLs. TLS clients
720 that verify CRLs are affected. TLS clients and servers with client
721 authentication enabled may be affected if they use custom verification
724 This issue was reported to OpenSSL by Robert Swiecki (Google), and
725 independently by Hanno Böck.
729 *) PKCS7 crash with missing EnvelopedContent
731 The PKCS#7 parsing code does not handle missing inner EncryptedContent
732 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
733 with missing content and trigger a NULL pointer dereference on parsing.
735 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
736 structures from untrusted sources are affected. OpenSSL clients and
737 servers are not affected.
739 This issue was reported to OpenSSL by Michal Zalewski (Google).
743 *) CMS verify infinite loop with unknown hash function
745 When verifying a signedData message the CMS code can enter an infinite loop
746 if presented with an unknown hash function OID. This can be used to perform
747 denial of service against any system which verifies signedData messages using
749 This issue was reported to OpenSSL by Johannes Bauer.
753 *) Race condition handling NewSessionTicket
755 If a NewSessionTicket is received by a multi-threaded client when attempting to
756 reuse a previous ticket then a race condition can occur potentially leading to
757 a double free of the ticket data.
761 *) Removed support for the two export grade static DH ciphersuites
762 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
763 were newly added (along with a number of other static DH ciphersuites) to
764 1.0.2. However the two export ones have *never* worked since they were
765 introduced. It seems strange in any case to be adding new export
766 ciphersuites, and given "logjam" it also does not seem correct to fix them.
769 *) Only support 256-bit or stronger elliptic curves with the
770 'ecdh_auto' setting (server) or by default (client). Of supported
771 curves, prefer P-256 (both).
774 *) Reject DH handshakes with parameters shorter than 768 bits.
775 [Kurt Roeckx and Emilia Kasper]
777 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
779 *) ClientHello sigalgs DoS fix
781 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
782 invalid signature algorithms extension a NULL pointer dereference will
783 occur. This can be exploited in a DoS attack against the server.
785 This issue was was reported to OpenSSL by David Ramos of Stanford
788 [Stephen Henson and Matt Caswell]
790 *) Multiblock corrupted pointer fix
792 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
793 feature only applies on 64 bit x86 architecture platforms that support AES
794 NI instructions. A defect in the implementation of "multiblock" can cause
795 OpenSSL's internal write buffer to become incorrectly set to NULL when
796 using non-blocking IO. Typically, when the user application is using a
797 socket BIO for writing, this will only result in a failed connection.
798 However if some other BIO is used then it is likely that a segmentation
799 fault will be triggered, thus enabling a potential DoS attack.
801 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
805 *) Segmentation fault in DTLSv1_listen fix
807 The DTLSv1_listen function is intended to be stateless and processes the
808 initial ClientHello from many peers. It is common for user code to loop
809 over the call to DTLSv1_listen until a valid ClientHello is received with
810 an associated cookie. A defect in the implementation of DTLSv1_listen means
811 that state is preserved in the SSL object from one invocation to the next
812 that can lead to a segmentation fault. Errors processing the initial
813 ClientHello can trigger this scenario. An example of such an error could be
814 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
817 This issue was reported to OpenSSL by Per Allansson.
821 *) Segmentation fault in ASN1_TYPE_cmp fix
823 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
824 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
825 certificate signature algorithm consistency this can be used to crash any
826 certificate verification operation and exploited in a DoS attack. Any
827 application which performs certificate verification is vulnerable including
828 OpenSSL clients and servers which enable client authentication.
832 *) Segmentation fault for invalid PSS parameters fix
834 The signature verification routines will crash with a NULL pointer
835 dereference if presented with an ASN.1 signature using the RSA PSS
836 algorithm and invalid parameters. Since these routines are used to verify
837 certificate signature algorithms this can be used to crash any
838 certificate verification operation and exploited in a DoS attack. Any
839 application which performs certificate verification is vulnerable including
840 OpenSSL clients and servers which enable client authentication.
842 This issue was was reported to OpenSSL by Brian Carpenter.
846 *) ASN.1 structure reuse memory corruption fix
848 Reusing a structure in ASN.1 parsing may allow an attacker to cause
849 memory corruption via an invalid write. Such reuse is and has been
850 strongly discouraged and is believed to be rare.
852 Applications that parse structures containing CHOICE or ANY DEFINED BY
853 components may be affected. Certificate parsing (d2i_X509 and related
854 functions) are however not affected. OpenSSL clients and servers are
859 *) PKCS7 NULL pointer dereferences fix
861 The PKCS#7 parsing code does not handle missing outer ContentInfo
862 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
863 missing content and trigger a NULL pointer dereference on parsing.
865 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
866 otherwise parse PKCS#7 structures from untrusted sources are
867 affected. OpenSSL clients and servers are not affected.
869 This issue was reported to OpenSSL by Michal Zalewski (Google).
873 *) DoS via reachable assert in SSLv2 servers fix
875 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
876 servers that both support SSLv2 and enable export cipher suites by sending
877 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
879 This issue was discovered by Sean Burford (Google) and Emilia Käsper
880 (OpenSSL development team).
884 *) Empty CKE with client auth and DHE fix
886 If client auth is used then a server can seg fault in the event of a DHE
887 ciphersuite being selected and a zero length ClientKeyExchange message
888 being sent by the client. This could be exploited in a DoS attack.
892 *) Handshake with unseeded PRNG fix
894 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
895 with an unseeded PRNG. The conditions are:
896 - The client is on a platform where the PRNG has not been seeded
897 automatically, and the user has not seeded manually
898 - A protocol specific client method version has been used (i.e. not
899 SSL_client_methodv23)
900 - A ciphersuite is used that does not require additional random data from
901 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
903 If the handshake succeeds then the client random that has been used will
904 have been generated from a PRNG with insufficient entropy and therefore the
905 output may be predictable.
907 For example using the following command with an unseeded openssl will
908 succeed on an unpatched platform:
910 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
914 *) Use After Free following d2i_ECPrivatekey error fix
916 A malformed EC private key file consumed via the d2i_ECPrivateKey function
917 could cause a use after free condition. This, in turn, could cause a double
918 free in several private key parsing functions (such as d2i_PrivateKey
919 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
920 for applications that receive EC private keys from untrusted
921 sources. This scenario is considered rare.
923 This issue was discovered by the BoringSSL project and fixed in their
928 *) X509_to_X509_REQ NULL pointer deref fix
930 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
931 the certificate key is invalid. This function is rarely used in practice.
933 This issue was discovered by Brian Carpenter.
937 *) Removed the export ciphers from the DEFAULT ciphers
940 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
942 *) Change RSA and DH/DSA key generation apps to generate 2048-bit
946 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
947 ARMv5 through ARMv8, as opposite to "locking" it to single one.
948 So far those who have to target multiple plaforms would compromise
949 and argue that binary targeting say ARMv5 would still execute on
950 ARMv8. "Universal" build resolves this compromise by providing
951 near-optimal performance even on newer platforms.
954 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
955 (other platforms pending).
956 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
958 *) Add support for the SignedCertificateTimestampList certificate and
959 OCSP response extensions from RFC6962.
962 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
963 for corner cases. (Certain input points at infinity could lead to
964 bogus results, with non-infinity inputs mapped to infinity too.)
967 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
968 This covers AES, SHA256/512 and GHASH. "Initial" means that most
969 common cases are optimized and there still is room for further
970 improvements. Vector Permutation AES for Altivec is also added.
973 *) Add support for little-endian ppc64 Linux target.
974 [Marcelo Cerri (IBM)]
976 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
977 SHA1, SHA256 and GHASH. "Initial" means that most common cases
978 are optimized and there still is room for further improvements.
979 Both 32- and 64-bit modes are supported.
980 [Andy Polyakov, Ard Biesheuvel (Linaro)]
982 *) Improved ARMv7 NEON support.
985 *) Support for SPARC Architecture 2011 crypto extensions, first
986 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
987 SHA256/512, MD5, GHASH and modular exponentiation.
988 [Andy Polyakov, David Miller]
990 *) Accelerated modular exponentiation for Intel processors, a.k.a.
992 [Shay Gueron & Vlad Krasnov (Intel Corp)]
994 *) Support for new and upcoming Intel processors, including AVX2,
995 BMI and SHA ISA extensions. This includes additional "stitched"
996 implementations, AESNI-SHA256 and GCM, and multi-buffer support
999 This work was sponsored by Intel Corp.
1002 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1003 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1004 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1007 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1008 this fixes a limiation in previous versions of OpenSSL.
1011 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1012 MGF1 digest and OAEP label.
1015 *) Add EVP support for key wrapping algorithms, to avoid problems with
1016 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1017 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1018 algorithms and include tests cases.
1021 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1023 [Douglas E. Engert, Steve Henson]
1025 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1026 difference in days and seconds between two tm or ASN1_TIME structures.
1029 *) Add -rev test option to s_server to just reverse order of characters
1030 received by client and send back to server. Also prints an abbreviated
1031 summary of the connection parameters.
1034 *) New option -brief for s_client and s_server to print out a brief summary
1035 of connection parameters.
1038 *) Add callbacks for arbitrary TLS extensions.
1039 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1041 *) New option -crl_download in several openssl utilities to download CRLs
1042 from CRLDP extension in certificates.
1045 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1048 *) New function X509_CRL_diff to generate a delta CRL from the difference
1049 of two full CRLs. Add support to "crl" utility.
1052 *) New functions to set lookup_crls function and to retrieve
1053 X509_STORE from X509_STORE_CTX.
1056 *) Print out deprecated issuer and subject unique ID fields in
1060 *) Extend OCSP I/O functions so they can be used for simple general purpose
1061 HTTP as well as OCSP. New wrapper function which can be used to download
1062 CRLs using the OCSP API.
1065 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1068 *) SSL_CONF* functions. These provide a common framework for application
1069 configuration using configuration files or command lines.
1072 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1073 message callback and prints the results. Needs compile time option
1074 "enable-ssl-trace". New options to s_client and s_server to enable
1078 *) New ctrl and macro to retrieve supported points extensions.
1079 Print out extension in s_server and s_client.
1082 *) New functions to retrieve certificate signature and signature
1086 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1090 *) New Suite B modes for TLS code. These use and enforce the requirements
1091 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1092 only use Suite B curves. The Suite B modes can be set by using the
1093 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1096 *) New chain verification flags for Suite B levels of security. Check
1097 algorithms are acceptable when flags are set in X509_verify_cert.
1100 *) Make tls1_check_chain return a set of flags indicating checks passed
1101 by a certificate chain. Add additional tests to handle client
1102 certificates: checks for matching certificate type and issuer name
1106 *) If an attempt is made to use a signature algorithm not in the peer
1107 preference list abort the handshake. If client has no suitable
1108 signature algorithms in response to a certificate request do not
1109 use the certificate.
1112 *) If server EC tmp key is not in client preference list abort handshake.
1115 *) Add support for certificate stores in CERT structure. This makes it
1116 possible to have different stores per SSL structure or one store in
1117 the parent SSL_CTX. Include distint stores for certificate chain
1118 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1119 to build and store a certificate chain in CERT structure: returing
1120 an error if the chain cannot be built: this will allow applications
1121 to test if a chain is correctly configured.
1123 Note: if the CERT based stores are not set then the parent SSL_CTX
1124 store is used to retain compatibility with existing behaviour.
1128 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1129 mask based on the current session, check mask when sending client
1130 hello and checking the requested ciphersuite.
1133 *) New ctrls to retrieve and set certificate types in a certificate
1134 request message. Print out received values in s_client. If certificate
1135 types is not set with custom values set sensible values based on
1136 supported signature algorithms.
1139 *) Support for distinct client and server supported signature algorithms.
1142 *) Add certificate callback. If set this is called whenever a certificate
1143 is required by client or server. An application can decide which
1144 certificate chain to present based on arbitrary criteria: for example
1145 supported signature algorithms. Add very simple example to s_server.
1146 This fixes many of the problems and restrictions of the existing client
1147 certificate callback: for example you can now clear an existing
1148 certificate and specify the whole chain.
1151 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1152 the certificate can be used for (if anything). Set valid_flags field
1153 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1154 to have similar checks in it.
1156 Add new "cert_flags" field to CERT structure and include a "strict mode".
1157 This enforces some TLS certificate requirements (such as only permitting
1158 certificate signature algorithms contained in the supported algorithms
1159 extension) which some implementations ignore: this option should be used
1160 with caution as it could cause interoperability issues.
1163 *) Update and tidy signature algorithm extension processing. Work out
1164 shared signature algorithms based on preferences and peer algorithms
1165 and print them out in s_client and s_server. Abort handshake if no
1166 shared signature algorithms.
1169 *) Add new functions to allow customised supported signature algorithms
1170 for SSL and SSL_CTX structures. Add options to s_client and s_server
1174 *) New function SSL_certs_clear() to delete all references to certificates
1175 from an SSL structure. Before this once a certificate had been added
1176 it couldn't be removed.
1179 *) Integrate hostname, email address and IP address checking with certificate
1180 verification. New verify options supporting checking in opensl utility.
1183 *) Fixes and wildcard matching support to hostname and email checking
1184 functions. Add manual page.
1185 [Florian Weimer (Red Hat Product Security Team)]
1187 *) New functions to check a hostname email or IP address against a
1188 certificate. Add options x509 utility to print results of checks against
1192 *) Fix OCSP checking.
1193 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1195 *) Initial experimental support for explicitly trusted non-root CAs.
1196 OpenSSL still tries to build a complete chain to a root but if an
1197 intermediate CA has a trust setting included that is used. The first
1198 setting is used: whether to trust (e.g., -addtrust option to the x509
1202 *) Add -trusted_first option which attempts to find certificates in the
1203 trusted store even if an untrusted chain is also supplied.
1206 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1207 platform support for Linux and Android.
1210 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1213 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1214 When in FIPS mode the approved implementations are used as normal,
1215 when not in FIPS mode the internal unapproved versions are used instead.
1216 This means that the FIPS capable OpenSSL isn't forced to use the
1217 (often lower perfomance) FIPS implementations outside FIPS mode.
1220 *) Transparently support X9.42 DH parameters when calling
1221 PEM_read_bio_DHparameters. This means existing applications can handle
1222 the new parameter format automatically.
1225 *) Initial experimental support for X9.42 DH parameter format: mainly
1226 to support use of 'q' parameter for RFC5114 parameters.
1229 *) Add DH parameters from RFC5114 including test data to dhtest.
1232 *) Support for automatic EC temporary key parameter selection. If enabled
1233 the most preferred EC parameters are automatically used instead of
1234 hardcoded fixed parameters. Now a server just has to call:
1235 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1236 support ECDH and use the most appropriate parameters.
1239 *) Enhance and tidy EC curve and point format TLS extension code. Use
1240 static structures instead of allocation if default values are used.
1241 New ctrls to set curves we wish to support and to retrieve shared curves.
1242 Print out shared curves in s_server. New options to s_server and s_client
1243 to set list of supported curves.
1246 *) New ctrls to retrieve supported signature algorithms and
1247 supported curve values as an array of NIDs. Extend openssl utility
1248 to print out received values.
1251 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1252 between NIDs and the more common NIST names such as "P-256". Enhance
1253 ecparam utility and ECC method to recognise the NIST names for curves.
1256 *) Enhance SSL/TLS certificate chain handling to support different
1257 chains for each certificate instead of one chain in the parent SSL_CTX.
1260 *) Support for fixed DH ciphersuite client authentication: where both
1261 server and client use DH certificates with common parameters.
1264 *) Support for fixed DH ciphersuites: those requiring DH server
1268 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1270 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1271 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1272 X509_CINF_get_signature were reverted post internal team review.
1274 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1276 *) Build fixes for the Windows and OpenVMS platforms
1277 [Matt Caswell and Richard Levitte]
1279 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1281 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1282 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1283 dereference. This could lead to a Denial Of Service attack. Thanks to
1284 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1288 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1289 dtls1_buffer_record function under certain conditions. In particular this
1290 could occur if an attacker sent repeated DTLS records with the same
1291 sequence number but for the next epoch. The memory leak could be exploited
1292 by an attacker in a Denial of Service attack through memory exhaustion.
1293 Thanks to Chris Mueller for reporting this issue.
1297 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1298 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1299 method would be set to NULL which could later result in a NULL pointer
1300 dereference. Thanks to Frank Schmirler for reporting this issue.
1304 *) Abort handshake if server key exchange message is omitted for ephemeral
1307 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1308 reporting this issue.
1312 *) Remove non-export ephemeral RSA code on client and server. This code
1313 violated the TLS standard by allowing the use of temporary RSA keys in
1314 non-export ciphersuites and could be used by a server to effectively
1315 downgrade the RSA key length used to a value smaller than the server
1316 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1317 INRIA or reporting this issue.
1321 *) Fixed issue where DH client certificates are accepted without verification.
1322 An OpenSSL server will accept a DH certificate for client authentication
1323 without the certificate verify message. This effectively allows a client to
1324 authenticate without the use of a private key. This only affects servers
1325 which trust a client certificate authority which issues certificates
1326 containing DH keys: these are extremely rare and hardly ever encountered.
1327 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1332 *) Ensure that the session ID context of an SSL is updated when its
1333 SSL_CTX is updated via SSL_set_SSL_CTX.
1335 The session ID context is typically set from the parent SSL_CTX,
1336 and can vary with the CTX.
1339 *) Fix various certificate fingerprint issues.
1341 By using non-DER or invalid encodings outside the signed portion of a
1342 certificate the fingerprint can be changed without breaking the signature.
1343 Although no details of the signed portion of the certificate can be changed
1344 this can cause problems with some applications: e.g. those using the
1345 certificate fingerprint for blacklists.
1347 1. Reject signatures with non zero unused bits.
1349 If the BIT STRING containing the signature has non zero unused bits reject
1350 the signature. All current signature algorithms require zero unused bits.
1352 2. Check certificate algorithm consistency.
1354 Check the AlgorithmIdentifier inside TBS matches the one in the
1355 certificate signature. NB: this will result in signature failure
1356 errors for some broken certificates.
1358 Thanks to Konrad Kraszewski from Google for reporting this issue.
1360 3. Check DSA/ECDSA signatures use DER.
1362 Reencode DSA/ECDSA signatures and compare with the original received
1363 signature. Return an error if there is a mismatch.
1365 This will reject various cases including garbage after signature
1366 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1367 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1368 (negative or with leading zeroes).
1370 Further analysis was conducted and fixes were developed by Stephen Henson
1371 of the OpenSSL core team.
1376 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1377 results on some platforms, including x86_64. This bug occurs at random
1378 with a very low probability, and is not known to be exploitable in any
1379 way, though its exact impact is difficult to determine. Thanks to Pieter
1380 Wuille (Blockstream) who reported this issue and also suggested an initial
1381 fix. Further analysis was conducted by the OpenSSL development team and
1382 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1383 the OpenSSL core team.
1387 *) Do not resume sessions on the server if the negotiated protocol
1388 version does not match the session's version. Resuming with a different
1389 version, while not strictly forbidden by the RFC, is of questionable
1390 sanity and breaks all known clients.
1391 [David Benjamin, Emilia Käsper]
1393 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1394 early CCS messages during renegotiation. (Note that because
1395 renegotiation is encrypted, this early CCS was not exploitable.)
1398 *) Tighten client-side session ticket handling during renegotiation:
1399 ensure that the client only accepts a session ticket if the server sends
1400 the extension anew in the ServerHello. Previously, a TLS client would
1401 reuse the old extension state and thus accept a session ticket if one was
1402 announced in the initial ServerHello.
1404 Similarly, ensure that the client requires a session ticket if one
1405 was advertised in the ServerHello. Previously, a TLS client would
1406 ignore a missing NewSessionTicket message.
1409 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1411 *) SRTP Memory Leak.
1413 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1414 sends a carefully crafted handshake message, to cause OpenSSL to fail
1415 to free up to 64k of memory causing a memory leak. This could be
1416 exploited in a Denial Of Service attack. This issue affects OpenSSL
1417 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1418 whether SRTP is used or configured. Implementations of OpenSSL that
1419 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1421 The fix was developed by the OpenSSL team.
1425 *) Session Ticket Memory Leak.
1427 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1428 integrity of that ticket is first verified. In the event of a session
1429 ticket integrity check failing, OpenSSL will fail to free memory
1430 causing a memory leak. By sending a large number of invalid session
1431 tickets an attacker could exploit this issue in a Denial Of Service
1436 *) Build option no-ssl3 is incomplete.
1438 When OpenSSL is configured with "no-ssl3" as a build option, servers
1439 could accept and complete a SSL 3.0 handshake, and clients could be
1440 configured to send them.
1442 [Akamai and the OpenSSL team]
1444 *) Add support for TLS_FALLBACK_SCSV.
1445 Client applications doing fallback retries should call
1446 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1448 [Adam Langley, Bodo Moeller]
1450 *) Add additional DigestInfo checks.
1452 Reencode DigestInto in DER and check against the original when
1453 verifying RSA signature: this will reject any improperly encoded
1454 DigestInfo structures.
1456 Note: this is a precautionary measure and no attacks are currently known.
1460 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1462 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1463 SRP code can be overrun an internal buffer. Add sanity check that
1464 g, A, B < N to SRP code.
1466 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1467 Group for discovering this issue.
1471 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1472 TLS 1.0 instead of higher protocol versions when the ClientHello message
1473 is badly fragmented. This allows a man-in-the-middle attacker to force a
1474 downgrade to TLS 1.0 even if both the server and the client support a
1475 higher protocol version, by modifying the client's TLS records.
1477 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1478 researching this issue.
1482 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1483 to a denial of service attack. A malicious server can crash the client
1484 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1485 ciphersuite and sending carefully crafted handshake messages.
1487 Thanks to Felix Gröbert (Google) for discovering and researching this
1492 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1493 to leak memory. This can be exploited through a Denial of Service attack.
1494 Thanks to Adam Langley for discovering and researching this issue.
1498 *) An attacker can force openssl to consume large amounts of memory whilst
1499 processing DTLS handshake messages. This can be exploited through a
1500 Denial of Service attack.
1501 Thanks to Adam Langley for discovering and researching this issue.
1505 *) An attacker can force an error condition which causes openssl to crash
1506 whilst processing DTLS packets due to memory being freed twice. This
1507 can be exploited through a Denial of Service attack.
1508 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1513 *) If a multithreaded client connects to a malicious server using a resumed
1514 session and the server sends an ec point format extension it could write
1515 up to 255 bytes to freed memory.
1517 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1522 *) A malicious server can crash an OpenSSL client with a null pointer
1523 dereference (read) by specifying an SRP ciphersuite even though it was not
1524 properly negotiated with the client. This can be exploited through a
1525 Denial of Service attack.
1527 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1528 discovering and researching this issue.
1532 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1533 X509_name_oneline, X509_name_print_ex et al. to leak some information
1534 from the stack. Applications may be affected if they echo pretty printing
1535 output to the attacker.
1537 Thanks to Ivan Fratric (Google) for discovering this issue.
1539 [Emilia Käsper, and Steve Henson]
1541 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1542 for corner cases. (Certain input points at infinity could lead to
1543 bogus results, with non-infinity inputs mapped to infinity too.)
1546 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1548 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1549 handshake can force the use of weak keying material in OpenSSL
1550 SSL/TLS clients and servers.
1552 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1553 researching this issue. (CVE-2014-0224)
1554 [KIKUCHI Masashi, Steve Henson]
1556 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1557 OpenSSL DTLS client the code can be made to recurse eventually crashing
1560 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1562 [Imre Rad, Steve Henson]
1564 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1565 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1566 client or server. This is potentially exploitable to run arbitrary
1567 code on a vulnerable client or server.
1569 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1570 [Jüri Aedla, Steve Henson]
1572 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1573 are subject to a denial of service attack.
1575 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1576 this issue. (CVE-2014-3470)
1577 [Felix Gröbert, Ivan Fratric, Steve Henson]
1579 *) Harmonize version and its documentation. -f flag is used to display
1581 [mancha <mancha1@zoho.com>]
1583 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1584 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1585 [mancha <mancha1@zoho.com>]
1587 *) Fix some double frees. These are not thought to be exploitable.
1588 [mancha <mancha1@zoho.com>]
1590 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1592 *) A missing bounds check in the handling of the TLS heartbeat extension
1593 can be used to reveal up to 64k of memory to a connected client or
1596 Thanks for Neel Mehta of Google Security for discovering this bug and to
1597 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1598 preparing the fix (CVE-2014-0160)
1599 [Adam Langley, Bodo Moeller]
1601 *) Fix for the attack described in the paper "Recovering OpenSSL
1602 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1603 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1604 http://eprint.iacr.org/2014/140
1606 Thanks to Yuval Yarom and Naomi Benger for discovering this
1607 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1608 [Yuval Yarom and Naomi Benger]
1610 *) TLS pad extension: draft-agl-tls-padding-03
1612 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1613 TLS client Hello record length value would otherwise be > 255 and
1614 less that 512 pad with a dummy extension containing zeroes so it
1615 is at least 512 bytes long.
1617 [Adam Langley, Steve Henson]
1619 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1621 *) Fix for TLS record tampering bug. A carefully crafted invalid
1622 handshake could crash OpenSSL with a NULL pointer exception.
1623 Thanks to Anton Johansson for reporting this issues.
1626 *) Keep original DTLS digest and encryption contexts in retransmission
1627 structures so we can use the previous session parameters if they need
1628 to be resent. (CVE-2013-6450)
1631 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1632 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1633 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1634 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1635 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1636 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1637 [Rob Stradling, Adam Langley]
1639 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1641 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1642 supporting platforms or when small records were transferred.
1643 [Andy Polyakov, Steve Henson]
1645 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1647 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1649 This addresses the flaw in CBC record processing discovered by
1650 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1651 at: http://www.isg.rhul.ac.uk/tls/
1653 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1654 Security Group at Royal Holloway, University of London
1655 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1656 Emilia Käsper for the initial patch.
1658 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1660 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1661 ciphersuites which can be exploited in a denial of service attack.
1662 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1663 and detecting this bug and to Wolfgang Ettlinger
1664 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1668 *) Return an error when checking OCSP signatures when key is NULL.
1669 This fixes a DoS attack. (CVE-2013-0166)
1672 *) Make openssl verify return errors.
1673 [Chris Palmer <palmer@google.com> and Ben Laurie]
1675 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1676 the right response is stapled. Also change SSL_get_certificate()
1677 so it returns the certificate actually sent.
1678 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1679 [Rob Stradling <rob.stradling@comodo.com>]
1681 *) Fix possible deadlock when decoding public keys.
1684 *) Don't use TLS 1.0 record version number in initial client hello
1688 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1690 *) Sanity check record length before skipping explicit IV in TLS
1691 1.2, 1.1 and DTLS to fix DoS attack.
1693 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1694 fuzzing as a service testing platform.
1698 *) Initialise tkeylen properly when encrypting CMS messages.
1699 Thanks to Solar Designer of Openwall for reporting this issue.
1702 *) In FIPS mode don't try to use composite ciphers as they are not
1706 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1708 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1709 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1710 mean any application compiled against OpenSSL 1.0.0 headers setting
1711 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1712 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1713 0x10000000L Any application which was previously compiled against
1714 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1715 will need to be recompiled as a result. Letting be results in
1716 inability to disable specifically TLS 1.1 and in client context,
1717 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1720 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1721 disable just protocol X, but all protocols above X *if* there are
1722 protocols *below* X still enabled. In more practical terms it means
1723 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1724 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1725 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1729 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1731 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1732 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1733 in CRYPTO_realloc_clean.
1735 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1736 issue and to Adam Langley <agl@chromium.org> for fixing it.
1738 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1740 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1743 *) Workarounds for some broken servers that "hang" if a client hello
1744 record length exceeds 255 bytes.
1746 1. Do not use record version number > TLS 1.0 in initial client
1747 hello: some (but not all) hanging servers will now work.
1748 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1749 the number of ciphers sent in the client hello. This should be
1750 set to an even number, such as 50, for example by passing:
1751 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1752 Most broken servers should now work.
1753 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1754 TLS 1.2 client support entirely.
1757 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1760 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1762 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1763 STRING form instead of a DigestInfo.
1766 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1767 and the RSA_sign/RSA_verify functions. This was made more apparent when
1768 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1769 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1770 the correct format in RSA_verify so both forms transparently work.
1773 *) Some servers which support TLS 1.0 can choke if we initially indicate
1774 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1775 encrypted premaster secret. As a workaround use the maximum pemitted
1776 client version in client hello, this should keep such servers happy
1777 and still work with previous versions of OpenSSL.
1780 *) Add support for TLS/DTLS heartbeats.
1781 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1783 *) Add support for SCTP.
1784 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1786 *) Improved PRNG seeding for VOS.
1787 [Paul Green <Paul.Green@stratus.com>]
1789 *) Extensive assembler packs updates, most notably:
1791 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1792 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1793 - x86_64: bit-sliced AES implementation;
1794 - ARM: NEON support, contemporary platforms optimizations;
1795 - s390x: z196 support;
1796 - *: GHASH and GF(2^m) multiplication implementations;
1800 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1801 (removal of unnecessary code)
1802 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1804 *) Add TLS key material exporter from RFC 5705.
1807 *) Add DTLS-SRTP negotiation from RFC 5764.
1810 *) Add Next Protocol Negotiation,
1811 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1812 disabled with a no-npn flag to config or Configure. Code donated
1814 [Adam Langley <agl@google.com> and Ben Laurie]
1816 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1817 NIST-P256, NIST-P521, with constant-time single point multiplication on
1818 typical inputs. Compiler support for the nonstandard type __uint128_t is
1819 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1820 Code made available under Apache License version 2.0.
1822 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1823 line to include this in your build of OpenSSL, and run "make depend" (or
1824 "make update"). This enables the following EC_METHODs:
1826 EC_GFp_nistp224_method()
1827 EC_GFp_nistp256_method()
1828 EC_GFp_nistp521_method()
1830 EC_GROUP_new_by_curve_name() will automatically use these (while
1831 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1833 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1835 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1836 all platforms. Move ssize_t definition from e_os.h to the public
1837 header file e_os2.h as it now appears in public header file cms.h
1840 *) New -sigopt option to the ca, req and x509 utilities. Additional
1841 signature parameters can be passed using this option and in
1845 *) Add RSA PSS signing function. This will generate and set the
1846 appropriate AlgorithmIdentifiers for PSS based on those in the
1847 corresponding EVP_MD_CTX structure. No application support yet.
1850 *) Support for companion algorithm specific ASN1 signing routines.
1851 New function ASN1_item_sign_ctx() signs a pre-initialised
1852 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1853 the appropriate parameters.
1856 *) Add new algorithm specific ASN1 verification initialisation function
1857 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1858 handling will be the same no matter what EVP_PKEY_METHOD is used.
1859 Add a PSS handler to support verification of PSS signatures: checked
1860 against a number of sample certificates.
1863 *) Add signature printing for PSS. Add PSS OIDs.
1864 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1866 *) Add algorithm specific signature printing. An individual ASN1 method
1867 can now print out signatures instead of the standard hex dump.
1869 More complex signatures (e.g. PSS) can print out more meaningful
1870 information. Include DSA version that prints out the signature
1874 *) Password based recipient info support for CMS library: implementing
1878 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1879 neatly separates the code into cipher and PBE sections and is required
1880 for some algorithms that split PBES2 into separate pieces (such as
1881 password based CMS).
1884 *) Session-handling fixes:
1885 - Fix handling of connections that are resuming with a session ID,
1886 but also support Session Tickets.
1887 - Fix a bug that suppressed issuing of a new ticket if the client
1888 presented a ticket with an expired session.
1889 - Try to set the ticket lifetime hint to something reasonable.
1890 - Make tickets shorter by excluding irrelevant information.
1891 - On the client side, don't ignore renewed tickets.
1892 [Adam Langley, Bodo Moeller (Google)]
1894 *) Fix PSK session representation.
1897 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1899 This work was sponsored by Intel.
1902 *) Add GCM support to TLS library. Some custom code is needed to split
1903 the IV between the fixed (from PRF) and explicit (from TLS record)
1904 portions. This adds all GCM ciphersuites supported by RFC5288 and
1905 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
1906 add a special AESGCM string for GCM only.
1909 *) Expand range of ctrls for AES GCM. Permit setting invocation
1910 field on decrypt and retrieval of invocation field only on encrypt.
1913 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1914 As required by RFC5289 these ciphersuites cannot be used if for
1915 versions of TLS earlier than 1.2.
1918 *) For FIPS capable OpenSSL interpret a NULL default public key method
1919 as unset and return the appopriate default but do *not* set the default.
1920 This means we can return the appopriate method in applications that
1921 swicth between FIPS and non-FIPS modes.
1924 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1925 ENGINE is used then we cannot handle that in the FIPS module so we
1926 keep original code iff non-FIPS operations are allowed.
1929 *) Add -attime option to openssl utilities.
1930 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1932 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1935 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1936 FIPS EC methods unconditionally for now.
1939 *) New build option no-ec2m to disable characteristic 2 code.
1942 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1943 all cases can be covered as some introduce binary incompatibilities.
1946 *) Redirect RSA operations to FIPS module including keygen,
1947 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1950 *) Add similar low level API blocking to ciphers.
1953 *) Low level digest APIs are not approved in FIPS mode: any attempt
1954 to use these will cause a fatal error. Applications that *really* want
1955 to use them can use the private_* version instead.
1958 *) Redirect cipher operations to FIPS module for FIPS builds.
1961 *) Redirect digest operations to FIPS module for FIPS builds.
1964 *) Update build system to add "fips" flag which will link in fipscanister.o
1965 for static and shared library builds embedding a signature if needed.
1968 *) Output TLS supported curves in preference order instead of numerical
1969 order. This is currently hardcoded for the highest order curves first.
1970 This should be configurable so applications can judge speed vs strength.
1973 *) Add TLS v1.2 server support for client authentication.
1976 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1980 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1981 FIPS modules versions.
1984 *) Add TLS v1.2 client side support for client authentication. Keep cache
1985 of handshake records longer as we don't know the hash algorithm to use
1986 until after the certificate request message is received.
1989 *) Initial TLS v1.2 client support. Add a default signature algorithms
1990 extension including all the algorithms we support. Parse new signature
1991 format in client key exchange. Relax some ECC signing restrictions for
1992 TLS v1.2 as indicated in RFC5246.
1995 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1996 to new signature format when needed using client digest preference.
1997 All server ciphersuites should now work correctly in TLS v1.2. No client
1998 support yet and no support for client certificates.
2001 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2002 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2003 ciphersuites. At present only RSA key exchange ciphersuites work with
2004 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2005 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2006 and version checking.
2009 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2010 with this defined it will not be affected by any changes to ssl internal
2011 structures. Add several utility functions to allow openssl application
2012 to work with OPENSSL_NO_SSL_INTERN defined.
2016 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
2018 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2021 *) Permit abbreviated handshakes when renegotiating using the function
2022 SSL_renegotiate_abbreviated().
2023 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2025 *) Add call to ENGINE_register_all_complete() to
2026 ENGINE_load_builtin_engines(), so some implementations get used
2027 automatically instead of needing explicit application support.
2030 *) Add support for TLS key exporter as described in RFC5705.
2031 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2033 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2034 a few changes are required:
2036 Add SSL_OP_NO_TLSv1_1 flag.
2037 Add TLSv1_1 methods.
2038 Update version checking logic to handle version 1.1.
2039 Add explicit IV handling (ported from DTLS code).
2040 Add command line options to s_client/s_server.
2043 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2045 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2046 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2047 content decryption and always return the same error. Note: this attack
2048 needs on average 2^20 messages so it only affects automated senders. The
2049 old behaviour can be reenabled in the CMS code by setting the
2050 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2051 an MMA defence is not necessary.
2052 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2053 this issue. (CVE-2012-0884)
2056 *) Fix CVE-2011-4619: make sure we really are receiving a
2057 client hello before rejecting multiple SGC restarts. Thanks to
2058 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2061 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2063 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2064 Thanks to Antonio Martin, Enterprise Secure Access Research and
2065 Development, Cisco Systems, Inc. for discovering this bug and
2066 preparing a fix. (CVE-2012-0050)
2069 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2071 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2072 of the Vaudenay padding oracle attack on CBC mode encryption
2073 which enables an efficient plaintext recovery attack against
2074 the OpenSSL implementation of DTLS. Their attack exploits timing
2075 differences arising during decryption processing. A research
2076 paper describing this attack can be found at:
2077 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2078 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2079 Security Group at Royal Holloway, University of London
2080 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2081 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2082 for preparing the fix. (CVE-2011-4108)
2083 [Robin Seggelmann, Michael Tuexen]
2085 *) Clear bytes used for block padding of SSL 3.0 records.
2087 [Adam Langley (Google)]
2089 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2090 Kadianakis <desnacked@gmail.com> for discovering this issue and
2091 Adam Langley for preparing the fix. (CVE-2011-4619)
2092 [Adam Langley (Google)]
2094 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2095 [Andrey Kulikov <amdeich@gmail.com>]
2097 *) Prevent malformed RFC3779 data triggering an assertion failure.
2098 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2099 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2100 [Rob Austein <sra@hactrn.net>]
2102 *) Improved PRNG seeding for VOS.
2103 [Paul Green <Paul.Green@stratus.com>]
2105 *) Fix ssl_ciph.c set-up race.
2106 [Adam Langley (Google)]
2108 *) Fix spurious failures in ecdsatest.c.
2109 [Emilia Käsper (Google)]
2111 *) Fix the BIO_f_buffer() implementation (which was mixing different
2112 interpretations of the '..._len' fields).
2113 [Adam Langley (Google)]
2115 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2116 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2117 threads won't reuse the same blinding coefficients.
2119 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2120 lock to call BN_BLINDING_invert_ex, and avoids one use of
2121 BN_BLINDING_update for each BN_BLINDING structure (previously,
2122 the last update always remained unused).
2123 [Emilia Käsper (Google)]
2125 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2126 [Bob Buckholz (Google)]
2128 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2130 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2131 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2132 [Kaspar Brand <ossl@velox.ch>]
2134 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2135 for multi-threaded use of ECDH. (CVE-2011-3210)
2136 [Adam Langley (Google)]
2138 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2141 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2142 signature public key algorithm by using OID xref utilities instead.
2143 Before this you could only use some ECC ciphersuites with SHA1 only.
2146 *) Add protection against ECDSA timing attacks as mentioned in the paper
2147 by Billy Bob Brumley and Nicola Tuveri, see:
2149 http://eprint.iacr.org/2011/232.pdf
2151 [Billy Bob Brumley and Nicola Tuveri]
2153 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2155 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2156 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2158 *) Fix bug in string printing code: if *any* escaping is enabled we must
2159 escape the escape character (backslash) or the resulting string is
2163 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2165 *) Disable code workaround for ancient and obsolete Netscape browsers
2166 and servers: an attacker can use it in a ciphersuite downgrade attack.
2167 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2170 *) Fixed J-PAKE implementation error, originally discovered by
2171 Sebastien Martini, further info and confirmation from Stefan
2172 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2175 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2177 *) Fix extension code to avoid race conditions which can result in a buffer
2178 overrun vulnerability: resumed sessions must not be modified as they can
2179 be shared by multiple threads. CVE-2010-3864
2182 *) Fix WIN32 build system to correctly link an ENGINE directory into
2186 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2188 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2190 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2192 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2194 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2195 context. The operation can be customised via the ctrl mechanism in
2196 case ENGINEs want to include additional functionality.
2199 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2202 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2203 output hashes compatible with older versions of OpenSSL.
2204 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2206 *) Fix compression algorithm handling: if resuming a session use the
2207 compression algorithm of the resumed session instead of determining
2208 it from client hello again. Don't allow server to change algorithm.
2211 *) Add load_crls() function to apps tidying load_certs() too. Add option
2212 to verify utility to allow additional CRLs to be included.
2215 *) Update OCSP request code to permit adding custom headers to the request:
2216 some responders need this.
2219 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2221 [Julia Lawall <julia@diku.dk>]
2223 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2224 needlessly dereferenced structures, used obsolete functions and
2225 didn't handle all updated verify codes correctly.
2228 *) Disable MD2 in the default configuration.
2231 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2232 indicate the initial BIO being pushed or popped. This makes it possible
2233 to determine whether the BIO is the one explicitly called or as a result
2234 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2235 it handles reference counts correctly and doesn't zero out the I/O bio
2236 when it is not being explicitly popped. WARNING: applications which
2237 included workarounds for the old buggy behaviour will need to be modified
2238 or they could free up already freed BIOs.
2241 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2242 renaming to all platforms (within the 0.9.8 branch, this was
2243 done conditionally on Netware platforms to avoid a name clash).
2244 [Guenter <lists@gknw.net>]
2246 *) Add ECDHE and PSK support to DTLS.
2247 [Michael Tuexen <tuexen@fh-muenster.de>]
2249 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2253 *) Add "missing" function EVP_MD_flags() (without this the only way to
2254 retrieve a digest flags is by accessing the structure directly. Update
2255 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2256 or cipher is registered as in the "from" argument. Print out all
2257 registered digests in the dgst usage message instead of manually
2258 attempting to work them out.
2261 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2262 this allows the use of compression and extensions. Change default cipher
2263 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2264 by default unless an application cipher string requests it.
2267 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2268 key ids to find matching certificates and keys but some PKCS#12 files
2269 don't follow the (somewhat unwritten) rules and this strategy fails.
2270 Now just gather all certificates together and the first private key
2271 then look for the first certificate that matches the key.
2274 *) Support use of registered digest and cipher names for dgst and cipher
2275 commands instead of having to add each one as a special case. So now
2282 openssl dgst -sha256 foo
2284 and this works for ENGINE based algorithms too.
2288 *) Update Gost ENGINE to support parameter files.
2289 [Victor B. Wagner <vitus@cryptocom.ru>]
2291 *) Support GeneralizedTime in ca utility.
2292 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2294 *) Enhance the hash format used for certificate directory links. The new
2295 form uses the canonical encoding (meaning equivalent names will work
2296 even if they aren't identical) and uses SHA1 instead of MD5. This form
2297 is incompatible with the older format and as a result c_rehash should
2298 be used to rebuild symbolic links.
2301 *) Make PKCS#8 the default write format for private keys, replacing the
2302 traditional format. This form is standardised, more secure and doesn't
2303 include an implicit MD5 dependency.
2306 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2307 committed to OpenSSL should pass this lot as a minimum.
2310 *) Add session ticket override functionality for use by EAP-FAST.
2311 [Jouni Malinen <j@w1.fi>]
2313 *) Modify HMAC functions to return a value. Since these can be implemented
2314 in an ENGINE errors can occur.
2317 *) Type-checked OBJ_bsearch_ex.
2320 *) Type-checked OBJ_bsearch. Also some constification necessitated
2321 by type-checking. Still to come: TXT_DB, bsearch(?),
2322 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2326 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2327 seconds to a tm structure directly, instead of going through OS
2328 specific date routines. This avoids any issues with OS routines such
2329 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2330 and X509_time_adj_ex() to cover the extended range. The existing
2331 X509_time_adj() is still usable and will no longer have any date issues.
2334 *) Delta CRL support. New use deltas option which will attempt to locate
2335 and search any appropriate delta CRLs available.
2337 This work was sponsored by Google.
2340 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2341 code and add additional score elements. Validate alternate CRL paths
2342 as part of the CRL checking and indicate a new error "CRL path validation
2343 error" in this case. Applications wanting additional details can use
2344 the verify callback and check the new "parent" field. If this is not
2345 NULL CRL path validation is taking place. Existing applications wont
2346 see this because it requires extended CRL support which is off by
2349 This work was sponsored by Google.
2352 *) Support for freshest CRL extension.
2354 This work was sponsored by Google.
2357 *) Initial indirect CRL support. Currently only supported in the CRLs
2358 passed directly and not via lookup. Process certificate issuer
2359 CRL entry extension and lookup CRL entries by bother issuer name
2360 and serial number. Check and process CRL issuer entry in IDP extension.
2362 This work was sponsored by Google.
2365 *) Add support for distinct certificate and CRL paths. The CRL issuer
2366 certificate is validated separately in this case. Only enabled if
2367 an extended CRL support flag is set: this flag will enable additional
2368 CRL functionality in future.
2370 This work was sponsored by Google.
2373 *) Add support for policy mappings extension.
2375 This work was sponsored by Google.
2378 *) Fixes to pathlength constraint, self issued certificate handling,
2379 policy processing to align with RFC3280 and PKITS tests.
2381 This work was sponsored by Google.
2384 *) Support for name constraints certificate extension. DN, email, DNS
2385 and URI types are currently supported.
2387 This work was sponsored by Google.
2390 *) To cater for systems that provide a pointer-based thread ID rather
2391 than numeric, deprecate the current numeric thread ID mechanism and
2392 replace it with a structure and associated callback type. This
2393 mechanism allows a numeric "hash" to be extracted from a thread ID in
2394 either case, and on platforms where pointers are larger than 'long',
2395 mixing is done to help ensure the numeric 'hash' is usable even if it
2396 can't be guaranteed unique. The default mechanism is to use "&errno"
2397 as a pointer-based thread ID to distinguish between threads.
2399 Applications that want to provide their own thread IDs should now use
2400 CRYPTO_THREADID_set_callback() to register a callback that will call
2401 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2403 Note that ERR_remove_state() is now deprecated, because it is tied
2404 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2405 to free the current thread's error state should be replaced by
2406 ERR_remove_thread_state(NULL).
2408 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2409 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2410 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2411 application was previously providing a numeric thread callback that
2412 was inappropriate for distinguishing threads, then uniqueness might
2413 have been obtained with &errno that happened immediately in the
2414 intermediate development versions of OpenSSL; this is no longer the
2415 case, the numeric thread callback will now override the automatic use
2417 [Geoff Thorpe, with help from Bodo Moeller]
2419 *) Initial support for different CRL issuing certificates. This covers a
2420 simple case where the self issued certificates in the chain exist and
2421 the real CRL issuer is higher in the existing chain.
2423 This work was sponsored by Google.
2426 *) Removed effectively defunct crypto/store from the build.
2429 *) Revamp of STACK to provide stronger type-checking. Still to come:
2430 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2431 ASN1_STRING, CONF_VALUE.
2434 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2435 RAM on SSL connections. This option can save about 34k per idle SSL.
2438 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2439 STACK, TXT_DB, bsearch, qsort.
2442 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2443 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2444 support for data, signedData, compressedData, digestedData and
2445 encryptedData, envelopedData types included. Scripts to check against
2446 RFC4134 examples draft and interop and consistency checks of many
2447 content types and variants.
2450 *) Add options to enc utility to support use of zlib compression BIO.
2453 *) Extend mk1mf to support importing of options and assembly language
2454 files from Configure script, currently only included in VC-WIN32.
2455 The assembly language rules can now optionally generate the source
2456 files from the associated perl scripts.
2459 *) Implement remaining functionality needed to support GOST ciphersuites.
2460 Interop testing has been performed using CryptoPro implementations.
2461 [Victor B. Wagner <vitus@cryptocom.ru>]
2463 *) s390x assembler pack.
2466 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2470 *) Implement Opaque PRF Input TLS extension as specified in
2471 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2472 official specification yet and no extension type assignment by
2473 IANA exists, this extension (for now) will have to be explicitly
2474 enabled when building OpenSSL by providing the extension number
2475 to use. For example, specify an option
2477 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2479 to the "config" or "Configure" script to enable the extension,
2480 assuming extension number 0x9527 (which is a completely arbitrary
2481 and unofficial assignment based on the MD5 hash of the Internet
2482 Draft). Note that by doing so, you potentially lose
2483 interoperability with other TLS implementations since these might
2484 be using the same extension number for other purposes.
2486 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2487 opaque PRF input value to use in the handshake. This will create
2488 an interal copy of the length-'len' string at 'src', and will
2489 return non-zero for success.
2491 To get more control and flexibility, provide a callback function
2494 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2495 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2499 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2502 Callback function 'cb' will be called in handshakes, and is
2503 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2504 Argument 'arg' is for application purposes (the value as given to
2505 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2506 be provided to the callback function). The callback function
2507 has to return non-zero to report success: usually 1 to use opaque
2508 PRF input just if possible, or 2 to enforce use of the opaque PRF
2509 input. In the latter case, the library will abort the handshake
2510 if opaque PRF input is not successfully negotiated.
2512 Arguments 'peerinput' and 'len' given to the callback function
2513 will always be NULL and 0 in the case of a client. A server will
2514 see the client's opaque PRF input through these variables if
2515 available (NULL and 0 otherwise). Note that if the server
2516 provides an opaque PRF input, the length must be the same as the
2517 length of the client's opaque PRF input.
2519 Note that the callback function will only be called when creating
2520 a new session (session resumption can resume whatever was
2521 previously negotiated), and will not be called in SSL 2.0
2522 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2523 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2524 for applications that need to enforce opaque PRF input.
2528 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2531 [Victor B. Wagner <vitus@cryptocom.ru>]
2533 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2534 RFC4507bis. The encrypted ticket format is an encrypted encoded
2535 SSL_SESSION structure, that way new session features are automatically
2538 If a client application caches session in an SSL_SESSION structure
2539 support is transparent because tickets are now stored in the encoded
2542 The SSL_CTX structure automatically generates keys for ticket
2543 protection in servers so again support should be possible
2544 with no application modification.
2546 If a client or server wishes to disable RFC4507 support then the option
2547 SSL_OP_NO_TICKET can be set.
2549 Add a TLS extension debugging callback to allow the contents of any client
2550 or server extensions to be examined.
2552 This work was sponsored by Google.
2555 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2556 OpenSSL should now compile cleanly on gcc 4.2
2557 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2559 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2560 support including streaming MAC support: this is required for GOST
2561 ciphersuite support.
2562 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2564 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2565 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2566 to output in BER and PEM format.
2569 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2570 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2571 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2572 ENGINE support for HMAC keys which are unextractable. New -mac and
2573 -macopt options to dgst utility.
2576 *) New option -sigopt to dgst utility. Update dgst to use
2577 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2578 alternative signing paramaters such as X9.31 or PSS in the dgst
2582 *) Change ssl_cipher_apply_rule(), the internal function that does
2583 the work each time a ciphersuite string requests enabling
2584 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2585 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2586 the order of disabled ciphersuites such that those ciphersuites
2587 that most recently went from enabled to disabled not only stay
2588 in order with respect to each other, but also have higher priority
2589 than other disabled ciphersuites the next time ciphersuites are
2592 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2593 the same ciphersuites as with "HIGH" alone, but in a specific
2594 order where the PSK ciphersuites come first (since they are the
2595 most recently disabled ciphersuites when "HIGH" is parsed).
2597 Also, change ssl_create_cipher_list() (using this new
2598 funcionality) such that between otherwise identical
2599 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2603 *) Change ssl_create_cipher_list() so that it automatically
2604 arranges the ciphersuites in reasonable order before starting
2605 to process the rule string. Thus, the definition for "DEFAULT"
2606 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2607 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2608 This makes it much easier to arrive at a reasonable default order
2609 in applications for which anonymous ciphers are OK (meaning
2610 that you can't actually use DEFAULT).
2611 [Bodo Moeller; suggested by Victor Duchovni]
2613 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2614 processing) into multiple integers instead of setting
2615 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2616 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2617 (These masks as well as the individual bit definitions are hidden
2618 away into the non-exported interface ssl/ssl_locl.h, so this
2619 change to the definition of the SSL_CIPHER structure shouldn't
2620 affect applications.) This give us more bits for each of these
2621 categories, so there is no longer a need to coagulate AES128 and
2622 AES256 into a single algorithm bit, and to coagulate Camellia128
2623 and Camellia256 into a single algorithm bit, which has led to all
2626 Thus, among other things, the kludge introduced in 0.9.7m and
2627 0.9.8e for masking out AES256 independently of AES128 or masking
2628 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2630 With the change, we also introduce new ciphersuite aliases that
2631 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2635 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2636 Use the leftmost N bytes of the signature input if the input is
2637 larger than the prime q (with N being the size in bytes of q).
2640 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2641 it yet and it is largely untested.
2644 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2647 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2648 some compilers (gcc 4.2 and later) reject their use. Safestack is
2649 reimplemented. Update ASN1 to avoid use of legacy functions.
2652 *) Win32/64 targets are linked with Winsock2.
2655 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2656 to external functions. This can be used to increase CRL handling
2657 efficiency especially when CRLs are very large by (for example) storing
2658 the CRL revoked certificates in a database.
2661 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2662 new CRLs added to a directory can be used. New command line option
2663 -verify_return_error to s_client and s_server. This causes real errors
2664 to be returned by the verify callback instead of carrying on no matter
2665 what. This reflects the way a "real world" verify callback would behave.
2668 *) GOST engine, supporting several GOST algorithms and public key formats.
2669 Kindly donated by Cryptocom.
2672 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2673 partitioned by DP are handled but no indirect CRL or reason partitioning
2674 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2675 selected via a scoring technique which handles IDP and AKID in CRLs.
2678 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2679 will ultimately be used for all verify operations: this will remove the
2680 X509_STORE dependency on certificate verification and allow alternative
2681 lookup methods. X509_STORE based implementations of these two callbacks.
2684 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2685 Modify get_crl() to find a valid (unexpired) CRL if possible.
2688 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2689 this would be called X509_CRL_cmp() but that name is already used by
2690 a function that just compares CRL issuer names. Cache several CRL
2691 extensions in X509_CRL structure and cache CRLDP in X509.
2694 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2695 this maps equivalent X509_NAME structures into a consistent structure.
2696 Name comparison can then be performed rapidly using memcmp().
2699 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2703 *) Allow digests to supply their own micalg string for S/MIME type using
2704 the ctrl EVP_MD_CTRL_MICALG.
2707 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2708 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2709 ctrl. It can then customise the structure before and/or after signing
2713 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2714 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2715 to free up any added signature OIDs.
2718 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2719 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2720 digest and cipher tables. New options added to openssl utility:
2721 list-message-digest-algorithms and list-cipher-algorithms.
2724 *) Change the array representation of binary polynomials: the list
2725 of degrees of non-zero coefficients is now terminated with -1.
2726 Previously it was terminated with 0, which was also part of the
2727 value; thus, the array representation was not applicable to
2728 polynomials where t^0 has coefficient zero. This change makes
2729 the array representation useful in a more general context.
2732 *) Various modifications and fixes to SSL/TLS cipher string
2733 handling. For ECC, the code now distinguishes between fixed ECDH
2734 with RSA certificates on the one hand and with ECDSA certificates
2735 on the other hand, since these are separate ciphersuites. The
2736 unused code for Fortezza ciphersuites has been removed.
2738 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2739 (not "ECDHE"). For consistency with the code for DH
2740 certificates, use of ECDH certificates is now considered ECDH
2741 authentication, not RSA or ECDSA authentication (the latter is
2742 merely the CA's signing algorithm and not actively used in the
2745 The temporary ciphersuite alias "ECCdraft" is no longer
2746 available, and ECC ciphersuites are no longer excluded from "ALL"
2747 and "DEFAULT". The following aliases now exist for RFC 4492
2748 ciphersuites, most of these by analogy with the DH case:
2750 kECDHr - ECDH cert, signed with RSA
2751 kECDHe - ECDH cert, signed with ECDSA
2752 kECDH - ECDH cert (signed with either RSA or ECDSA)
2753 kEECDH - ephemeral ECDH
2754 ECDH - ECDH cert or ephemeral ECDH
2760 AECDH - anonymous ECDH
2761 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2765 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2766 Use correct micalg parameters depending on digest(s) in signed message.
2769 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2770 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2773 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2774 an engine to register a method. Add ENGINE lookups for methods and
2775 functional reference processing.
2778 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2779 EVP_{Sign,Verify}* which allow an application to customise the signature
2783 *) New -resign option to smime utility. This adds one or more signers
2784 to an existing PKCS#7 signedData structure. Also -md option to use an
2785 alternative message digest algorithm for signing.
2788 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2789 create PKCS7 structures containing multiple signers. Update smime
2790 application to support multiple signers.
2793 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2797 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2798 Reorganize PBE internals to lookup from a static table using NIDs,
2799 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2800 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2801 PRF which will be automatically used with PBES2.
2804 *) Replace the algorithm specific calls to generate keys in "req" with the
2808 *) Update PKCS#7 enveloped data routines to use new API. This is now
2809 supported by any public key method supporting the encrypt operation. A
2810 ctrl is added to allow the public key algorithm to examine or modify
2811 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2815 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2816 a default digest type to use. In most cases this will be SHA1 but some
2817 algorithms (such as GOST) need to specify an alternative digest. The
2818 return value indicates how strong the prefernce is 1 means optional and
2819 2 is mandatory (that is it is the only supported type). Modify
2820 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2821 use the default md. Update openssl utilities to use the default digest
2822 type for signing if it is not explicitly indicated.
2825 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2826 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2827 signing method from the key type. This effectively removes the link
2828 between digests and public key types.
2831 *) Add an OID cross reference table and utility functions. Its purpose is to
2832 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2833 rsaEncryption. This will allow some of the algorithm specific hackery
2834 needed to use the correct OID to be removed.
2837 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2838 structures for PKCS7_sign(). They are now set up by the relevant public
2842 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2845 *) Add support for key derivation (agreement) in the API, DH method and
2849 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2850 public and private key formats. As a side effect these add additional
2851 command line functionality not previously available: DSA signatures can be
2852 generated and verified using pkeyutl and DH key support and generation in
2857 [Oliver Tappe <zooey@hirschkaefer.de>]
2859 *) New make target "install_html_docs" installs HTML renditions of the
2861 [Oliver Tappe <zooey@hirschkaefer.de>]
2863 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
2864 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2865 support key and parameter generation and add initial key generation
2866 functionality for RSA.
2869 *) Add functions for main EVP_PKEY_method operations. The undocumented
2870 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2871 EVP_PKEY_{encrypt,decrypt}_old.
2874 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2875 key API, doesn't do much yet.
2878 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2879 public key algorithms. New option to openssl utility:
2880 "list-public-key-algorithms" to print out info.
2883 *) Implement the Supported Elliptic Curves Extension for
2884 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2887 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2888 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2891 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2892 utilities such as rsa, dsa, dsaparam etc except they process any key
2896 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2897 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2898 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2902 *) Initial support for pluggable public key ASN1.
2903 De-spaghettify the public key ASN1 handling. Move public and private
2904 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2905 algorithm specific handling to a single module within the relevant
2906 algorithm directory. Add functions to allow (near) opaque processing
2907 of public and private key structures.
2910 *) Implement the Supported Point Formats Extension for
2911 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2914 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2915 for the psk identity [hint] and the psk callback functions to the
2916 SSL_SESSION, SSL and SSL_CTX structure.
2919 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2923 SSL_CTX_use_psk_identity_hint
2924 SSL_get_psk_identity_hint
2925 SSL_get_psk_identity
2926 SSL_use_psk_identity_hint
2928 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2930 *) Add RFC 3161 compliant time stamp request creation, response generation
2931 and response verification functionality.
2932 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2934 *) Add initial support for TLS extensions, specifically for the server_name
2935 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2936 have new members for a host name. The SSL data structure has an
2937 additional member SSL_CTX *initial_ctx so that new sessions can be
2938 stored in that context to allow for session resumption, even after the
2939 SSL has been switched to a new SSL_CTX in reaction to a client's
2940 server_name extension.
2942 New functions (subject to change):
2944 SSL_get_servername()
2945 SSL_get_servername_type()
2948 New CTRL codes and macros (subject to change):
2950 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2951 - SSL_CTX_set_tlsext_servername_callback()
2952 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2953 - SSL_CTX_set_tlsext_servername_arg()
2954 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2956 openssl s_client has a new '-servername ...' option.
2958 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2959 '-key2 ...', '-servername_fatal' (subject to change). This allows
2960 testing the HostName extension for a specific single host name ('-cert'
2961 and '-key' remain fallbacks for handshakes without HostName
2962 negotiation). If the unrecogninzed_name alert has to be sent, this by
2963 default is a warning; it becomes fatal with the '-servername_fatal'
2966 [Peter Sylvester, Remy Allais, Christophe Renou]
2968 *) Whirlpool hash implementation is added.
2971 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2972 bn(64,32). Because of instruction set limitations it doesn't have
2973 any negative impact on performance. This was done mostly in order
2974 to make it possible to share assembler modules, such as bn_mul_mont
2975 implementations, between 32- and 64-bit builds without hassle.
2978 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2979 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2983 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2984 dedicated Montgomery multiplication procedure, is introduced.
2985 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2986 "64-bit" performance on certain 32-bit targets.
2989 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2990 in SSL structures. New SSL ctrl to set maximum send fragment size.
2991 Save memory by seeting the I/O buffer sizes dynamically instead of
2992 using the maximum available value.
2995 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2996 in addition to the text details.
2999 *) Very, very preliminary EXPERIMENTAL support for printing of general
3000 ASN1 structures. This currently produces rather ugly output and doesn't
3001 handle several customised structures at all.
3004 *) Integrated support for PVK file format and some related formats such
3005 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3006 these in the 'rsa' and 'dsa' utilities.
3009 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3012 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3013 place for the (very old) "NETSCAPE" format certificates which are now
3014 handled using new ASN1 code equivalents.
3017 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3018 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3019 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3022 *) Modify CRL distribution points extension code to print out previously
3023 unsupported fields. Enhance extension setting code to allow setting of
3027 *) Add print and set support for Issuing Distribution Point CRL extension.
3030 *) Change 'Configure' script to enable Camellia by default.
3033 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3035 *) When rejecting SSL/TLS records due to an incorrect version number, never
3036 update s->server with a new major version number. As of
3037 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3038 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3039 the previous behavior could result in a read attempt at NULL when
3040 receiving specific incorrect SSL/TLS records once record payload
3041 protection is active. (CVE-2010-0740)
3042 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3044 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3045 could be crashed if the relevant tables were not present (e.g. chrooted).
3046 [Tomas Hoger <thoger@redhat.com>]
3048 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3050 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3051 [Martin Olsson, Neel Mehta]
3053 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3054 accommodate for stack sorting, always a write lock!).
3057 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3058 excessive delays in the RAND_poll(): over a minute. As a workaround
3059 include a time check in the inner Heap32Next loop too.
3062 *) The code that handled flushing of data in SSL/TLS originally used the
3063 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3064 the problem outlined in PR#1949. The fix suggested there however can
3065 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3066 of Apache). So instead simplify the code to flush unconditionally.
3067 This should be fine since flushing with no data to flush is a no op.
3070 *) Handle TLS versions 2.0 and later properly and correctly use the
3071 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
3072 off ancient servers have a habit of sticking around for a while...
3075 *) Modify compression code so it frees up structures without using the
3076 ex_data callbacks. This works around a problem where some applications
3077 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
3078 restarting) then use compression (e.g. SSL with compression) later.
3079 This results in significant per-connection memory leaks and
3080 has caused some security issues including CVE-2008-1678 and
3084 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3085 change when encrypting or decrypting.
3088 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3089 connect and renegotiate with servers which do not support RI.
3090 Until RI is more widely deployed this option is enabled by default.
3093 *) Add "missing" ssl ctrls to clear options and mode.
3096 *) If client attempts to renegotiate and doesn't support RI respond with
3097 a no_renegotiation alert as required by RFC5746. Some renegotiating
3098 TLS clients will continue a connection gracefully when they receive
3099 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3100 waiting for a server hello which it will never receive. Now we treat a
3101 received no_renegotiation alert as a fatal error. This is because
3102 applications requesting a renegotiation might well expect it to succeed
3103 and would have no code in place to handle the server denying it so the
3104 only safe thing to do is to terminate the connection.
3107 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3108 peer supports secure renegotiation and 0 otherwise. Print out peer
3109 renegotiation support in s_client/s_server.
3112 *) Replace the highly broken and deprecated SPKAC certification method with
3113 the updated NID creation version. This should correctly handle UTF8.
3116 *) Implement RFC5746. Re-enable renegotiation but require the extension
3117 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3118 turns out to be a bad idea. It has been replaced by
3119 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3120 SSL_CTX_set_options(). This is really not recommended unless you
3121 know what you are doing.
3122 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3124 *) Fixes to stateless session resumption handling. Use initial_ctx when
3125 issuing and attempting to decrypt tickets in case it has changed during
3126 servername handling. Use a non-zero length session ID when attempting
3127 stateless session resumption: this makes it possible to determine if
3128 a resumption has occurred immediately after receiving server hello
3129 (several places in OpenSSL subtly assume this) instead of later in
3133 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3134 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3135 fixes for a few places where the return code is not checked
3137 [Julia Lawall <julia@diku.dk>]
3139 *) Add --strict-warnings option to Configure script to include devteam
3140 warnings in other configurations.
3143 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3144 makes it possible to install openssl libraries in locations which
3145 have names other than "lib", for example "/usr/lib64" which some
3147 [Steve Henson, based on patch from Jeremy Utley]
3149 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3150 X690 8.9.12 and can produce some misleading textual output of OIDs.
3151 [Steve Henson, reported by Dan Kaminsky]
3153 *) Delete MD2 from algorithm tables. This follows the recommendation in
3154 several standards that it is not used in new applications due to
3155 several cryptographic weaknesses. For binary compatibility reasons
3156 the MD2 API is still compiled in by default.
3159 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3163 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3164 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3166 [Guenter <lists@gknw.net>]
3168 *) Fix the server certificate chain building code to use X509_verify_cert(),
3169 it used to have an ad-hoc builder which was unable to cope with anything
3170 other than a simple chain.
3171 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3173 *) Don't check self signed certificate signatures in X509_verify_cert()
3174 by default (a flag can override this): it just wastes time without
3175 adding any security. As a useful side effect self signed root CAs
3176 with non-FIPS digests are now usable in FIPS mode.
3179 *) In dtls1_process_out_of_seq_message() the check if the current message
3180 is already buffered was missing. For every new message was memory
3181 allocated, allowing an attacker to perform an denial of service attack
3182 with sending out of seq handshake messages until there is no memory
3183 left. Additionally every future messege was buffered, even if the
3184 sequence number made no sense and would be part of another handshake.
3185 So only messages with sequence numbers less than 10 in advance will be
3186 buffered. (CVE-2009-1378)
3187 [Robin Seggelmann, discovered by Daniel Mentz]
3189 *) Records are buffered if they arrive with a future epoch to be
3190 processed after finishing the corresponding handshake. There is
3191 currently no limitation to this buffer allowing an attacker to perform
3192 a DOS attack with sending records with future epochs until there is no
3193 memory left. This patch adds the pqueue_size() function to detemine
3194 the size of a buffer and limits the record buffer to 100 entries.
3196 [Robin Seggelmann, discovered by Daniel Mentz]
3198 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3199 parent structure is freed. (CVE-2009-1379)
3202 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3203 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3206 [Ilya O. <vrghost@gmail.com>]
3208 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3210 *) Disable renegotiation completely - this fixes a severe security
3211 problem (CVE-2009-3555) at the cost of breaking all
3212 renegotiation. Renegotiation can be re-enabled by setting
3213 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3214 run-time. This is really not recommended unless you know what
3218 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3220 *) Don't set val to NULL when freeing up structures, it is freed up by
3221 underlying code. If sizeof(void *) > sizeof(long) this can result in
3222 zeroing past the valid field. (CVE-2009-0789)
3223 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3225 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3226 checked correctly. This would allow some invalid signed attributes to
3227 appear to verify correctly. (CVE-2009-0591)
3228 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3230 *) Reject UniversalString and BMPString types with invalid lengths. This
3231 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3232 a legal length. (CVE-2009-0590)
3235 *) Set S/MIME signing as the default purpose rather than setting it
3236 unconditionally. This allows applications to override it at the store
3240 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3241 to handle some structures.
3244 *) Improve efficiency of mem_gets: don't search whole buffer each time
3246 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3248 *) New -hex option for openssl rand.
3251 *) Print out UTF8String and NumericString when parsing ASN1.
3254 *) Support NumericString type for name components.
3257 *) Allow CC in the environment to override the automatically chosen
3258 compiler. Note that nothing is done to ensure flags work with the
3262 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3264 *) Properly check EVP_VerifyFinal() and similar return values
3266 [Ben Laurie, Bodo Moeller, Google Security Team]
3268 *) Enable TLS extensions by default.
3271 *) Allow the CHIL engine to be loaded, whether the application is
3272 multithreaded or not. (This does not release the developer from the
3273 obligation to set up the dynamic locking callbacks.)
3274 [Sander Temme <sander@temme.net>]
3276 *) Use correct exit code if there is an error in dgst command.
3277 [Steve Henson; problem pointed out by Roland Dirlewanger]
3279 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3280 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3283 *) Add experimental JPAKE support, including demo authentication in
3284 s_client and s_server.
3287 *) Set the comparison function in v3_addr_canonize().
3288 [Rob Austein <sra@hactrn.net>]
3290 *) Add support for XMPP STARTTLS in s_client.
3291 [Philip Paeps <philip@freebsd.org>]
3293 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3294 to ensure that even with this option, only ciphersuites in the
3295 server's preference list will be accepted. (Note that the option
3296 applies only when resuming a session, so the earlier behavior was
3297 just about the algorithm choice for symmetric cryptography.)
3300 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3302 *) Fix NULL pointer dereference if a DTLS server received
3303 ChangeCipherSpec as first record (CVE-2009-1386).
3306 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
3307 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3310 *) The fix in 0.9.8c that supposedly got rid of unsafe
3311 double-checked locking was incomplete for RSA blinding,
3312 addressing just one layer of what turns out to have been
3313 doubly unsafe triple-checked locking.
3315 So now fix this for real by retiring the MONT_HELPER macro
3316 in crypto/rsa/rsa_eay.c.
3318 [Bodo Moeller; problem pointed out by Marius Schilder]
3320 *) Various precautionary measures:
3322 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3324 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3325 (NB: This would require knowledge of the secret session ticket key
3326 to exploit, in which case you'd be SOL either way.)
3328 - Change bn_nist.c so that it will properly handle input BIGNUMs
3329 outside the expected range.
3331 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3334 [Neel Mehta, Bodo Moeller]
3336 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3337 the load fails. Useful for distros.
3338 [Ben Laurie and the FreeBSD team]
3340 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3343 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3346 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3348 This work was sponsored by Logica.
3351 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3352 keystores. Support for SSL/TLS client authentication too.
3353 Not compiled unless enable-capieng specified to Configure.
3355 This work was sponsored by Logica.
3358 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3359 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3360 attribute creation routines such as certifcate requests and PKCS#12
3364 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3366 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3367 handshake which could lead to a cilent crash as found using the
3368 Codenomicon TLS test suite (CVE-2008-1672)
3369 [Steve Henson, Mark Cox]
3371 *) Fix double free in TLS server name extensions which could lead to
3372 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3375 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3377 Clear the error queue to ensure that error entries left from
3378 older function calls do not interfere with the correct operation.
3379 [Lutz Jaenicke, Erik de Castro Lopo]
3381 *) Remove root CA certificates of commercial CAs:
3383 The OpenSSL project does not recommend any specific CA and does not
3384 have any policy with respect to including or excluding any CA.
3385 Therefore it does not make any sense to ship an arbitrary selection
3386 of root CA certificates with the OpenSSL software.
3389 *) RSA OAEP patches to fix two separate invalid memory reads.
3390 The first one involves inputs when 'lzero' is greater than
3391 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3392 before the beginning of from). The second one involves inputs where
3393 the 'db' section contains nothing but zeroes (there is a one-byte
3394 invalid read after the end of 'db').
3395 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3397 *) Partial backport from 0.9.9-dev:
3399 Introduce bn_mul_mont (dedicated Montgomery multiplication
3400 procedure) as a candidate for BIGNUM assembler implementation.
3401 While 0.9.9-dev uses assembler for various architectures, only
3402 x86_64 is available by default here in the 0.9.8 branch, and
3403 32-bit x86 is available through a compile-time setting.
3405 To try the 32-bit x86 assembler implementation, use Configure
3406 option "enable-montasm" (which exists only for this backport).
3408 As "enable-montasm" for 32-bit x86 disclaims code stability
3409 anyway, in this constellation we activate additional code
3410 backported from 0.9.9-dev for further performance improvements,
3411 namely BN_from_montgomery_word. (To enable this otherwise,
3412 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3414 [Andy Polyakov (backport partially by Bodo Moeller)]
3416 *) Add TLS session ticket callback. This allows an application to set
3417 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3418 values. This is useful for key rollover for example where several key
3419 sets may exist with different names.
3422 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3423 This was broken until now in 0.9.8 releases, such that the only way
3424 a registered ENGINE could be used (assuming it initialises
3425 successfully on the host) was to explicitly set it as the default
3426 for the relevant algorithms. This is in contradiction with 0.9.7
3427 behaviour and the documentation. With this fix, when an ENGINE is
3428 registered into a given algorithm's table of implementations, the
3429 'uptodate' flag is reset so that auto-discovery will be used next
3430 time a new context for that algorithm attempts to select an
3432 [Ian Lister (tweaked by Geoff Thorpe)]
3434 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3435 implemention in the following ways:
3437 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3440 Lack of BER streaming support means one pass streaming processing is
3441 only supported if data is detached: setting the streaming flag is
3442 ignored for embedded content.
3444 CMS support is disabled by default and must be explicitly enabled
3445 with the enable-cms configuration option.
3448 *) Update the GMP engine glue to do direct copies between BIGNUM and
3449 mpz_t when openssl and GMP use the same limb size. Otherwise the
3450 existing "conversion via a text string export" trick is still used.
3451 [Paul Sheer <paulsheer@gmail.com>]
3453 *) Zlib compression BIO. This is a filter BIO which compressed and
3454 uncompresses any data passed through it.
3457 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3458 RFC3394 compatible AES key wrapping.
3461 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3462 sets string data without copying. X509_ALGOR_set0() and
3463 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3464 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3465 from an X509_ATTRIBUTE structure optionally checking it occurs only
3466 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3470 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3471 to get the expected BN_FLG_CONSTTIME behavior.
3472 [Bodo Moeller (Google)]
3476 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3477 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3478 - added some more tests to do_tests.pl
3479 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3480 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3481 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3482 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3483 - various changes to netware.pl to enable gcc-cross builds on Win32
3485 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3486 - various changes to fix missing prototype warnings
3487 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3488 - added AES, WHIRLPOOL and CPUID assembler code to build files
3489 - added missing AES assembler make rules to mk1mf.pl
3490 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3491 [Guenter Knauf <eflash@gmx.net>]
3493 *) Implement certificate status request TLS extension defined in RFC3546.
3494 A client can set the appropriate parameters and receive the encoded
3495 OCSP response via a callback. A server can query the supplied parameters
3496 and set the encoded OCSP response in the callback. Add simplified examples
3497 to s_client and s_server.
3500 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3502 *) Fix various bugs:
3503 + Binary incompatibility of ssl_ctx_st structure
3504 + DTLS interoperation with non-compliant servers
3505 + Don't call get_session_cb() without proposed session
3506 + Fix ia64 assembler code
3507 [Andy Polyakov, Steve Henson]
3509 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3511 *) DTLS Handshake overhaul. There were longstanding issues with
3512 OpenSSL DTLS implementation, which were making it impossible for
3513 RFC 4347 compliant client to communicate with OpenSSL server.
3514 Unfortunately just fixing these incompatibilities would "cut off"
3515 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3516 server keeps tolerating non RFC compliant syntax. The opposite is
3517 not true, 0.9.8f client can not communicate with earlier server.
3518 This update even addresses CVE-2007-4995.
3521 *) Changes to avoid need for function casts in OpenSSL: some compilers
3522 (gcc 4.2 and later) reject their use.
3523 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3526 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3527 RFC4507bis. The encrypted ticket format is an encrypted encoded
3528 SSL_SESSION structure, that way new session features are automatically
3531 If a client application caches session in an SSL_SESSION structure
3532 support is transparent because tickets are now stored in the encoded
3535 The SSL_CTX structure automatically generates keys for ticket
3536 protection in servers so again support should be possible
3537 with no application modification.
3539 If a client or server wishes to disable RFC4507 support then the option
3540 SSL_OP_NO_TICKET can be set.
3542 Add a TLS extension debugging callback to allow the contents of any client
3543 or server extensions to be examined.
3545 This work was sponsored by Google.
3548 *) Add initial support for TLS extensions, specifically for the server_name
3549 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3550 have new members for a host name. The SSL data structure has an
3551 additional member SSL_CTX *initial_ctx so that new sessions can be
3552 stored in that context to allow for session resumption, even after the
3553 SSL has been switched to a new SSL_CTX in reaction to a client's
3554 server_name extension.
3556 New functions (subject to change):
3558 SSL_get_servername()
3559 SSL_get_servername_type()
3562 New CTRL codes and macros (subject to change):
3564 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3565 - SSL_CTX_set_tlsext_servername_callback()
3566 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3567 - SSL_CTX_set_tlsext_servername_arg()
3568 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3570 openssl s_client has a new '-servername ...' option.
3572 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3573 '-key2 ...', '-servername_fatal' (subject to change). This allows
3574 testing the HostName extension for a specific single host name ('-cert'
3575 and '-key' remain fallbacks for handshakes without HostName
3576 negotiation). If the unrecogninzed_name alert has to be sent, this by
3577 default is a warning; it becomes fatal with the '-servername_fatal'
3580 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3582 *) Add AES and SSE2 assembly language support to VC++ build.
3585 *) Mitigate attack on final subtraction in Montgomery reduction.
3588 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3589 (which previously caused an internal error).
3592 *) Squeeze another 10% out of IGE mode when in != out.
3595 *) AES IGE mode speedup.
3596 [Dean Gaudet (Google)]
3598 *) Add the Korean symmetric 128-bit cipher SEED (see
3599 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3600 add SEED ciphersuites from RFC 4162:
3602 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3603 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3604 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3605 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3607 To minimize changes between patchlevels in the OpenSSL 0.9.8
3608 series, SEED remains excluded from compilation unless OpenSSL
3609 is configured with 'enable-seed'.
3610 [KISA, Bodo Moeller]
3612 *) Mitigate branch prediction attacks, which can be practical if a
3613 single processor is shared, allowing a spy process to extract
3614 information. For detailed background information, see
3615 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3616 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3617 and Necessary Software Countermeasures"). The core of the change
3618 are new versions BN_div_no_branch() and
3619 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3620 respectively, which are slower, but avoid the security-relevant
3621 conditional branches. These are automatically called by BN_div()
3622 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3623 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3624 remove a conditional branch.
3626 BN_FLG_CONSTTIME is the new name for the previous
3627 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3628 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3629 in the exponent causes BN_mod_exp_mont() to use the alternative
3630 implementation in BN_mod_exp_mont_consttime().) The old name
3631 remains as a deprecated alias.
3633 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3634 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3635 constant-time implementations for more than just exponentiation.
3636 Here too the old name is kept as a deprecated alias.
3638 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3639 the BN_BLINDING structure gets an independent copy of the
3640 modulus. This means that the previous "BIGNUM *m" argument to
3641 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3642 essentially becomes "const BIGNUM *m", although we can't actually
3643 change this in the header file before 0.9.9. It allows
3644 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3645 enable BN_FLG_CONSTTIME.
3647 [Matthew D Wood (Intel Corp)]
3649 *) In the SSL/TLS server implementation, be strict about session ID
3650 context matching (which matters if an application uses a single
3651 external cache for different purposes). Previously,
3652 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3653 set. This did ensure strict client verification, but meant that,
3654 with applications using a single external cache for quite
3655 different requirements, clients could circumvent ciphersuite
3656 restrictions for a given session ID context by starting a session
3657 in a different context.
3660 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3661 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3662 authentication-only ciphersuites.
3665 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3666 not complete and could lead to a possible single byte overflow
3667 (CVE-2007-5135) [Ben Laurie]
3669 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3671 *) Since AES128 and AES256 (and similarly Camellia128 and
3672 Camellia256) share a single mask bit in the logic of
3673 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3674 kludge to work properly if AES128 is available and AES256 isn't
3675 (or if Camellia128 is available and Camellia256 isn't).
3678 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3679 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3680 When a point or a seed is encoded in a BIT STRING, we need to
3681 prevent the removal of trailing zero bits to get the proper DER
3682 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3683 of a NamedBitList, for which trailing 0 bits need to be removed.)
3686 *) Have SSL/TLS server implementation tolerate "mismatched" record
3687 protocol version while receiving ClientHello even if the
3688 ClientHello is fragmented. (The server can't insist on the
3689 particular protocol version it has chosen before the ServerHello
3690 message has informed the client about his choice.)
3693 *) Add RFC 3779 support.
3694 [Rob Austein for ARIN, Ben Laurie]
3696 *) Load error codes if they are not already present instead of using a
3697 static variable. This allows them to be cleanly unloaded and reloaded.
3698 Improve header file function name parsing.
3701 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3702 or CAPABILITY handshake as required by RFCs.
3705 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3707 *) Introduce limits to prevent malicious keys being able to
3708 cause a denial of service. (CVE-2006-2940)
3709 [Steve Henson, Bodo Moeller]
3711 *) Fix ASN.1 parsing of certain invalid structures that can result
3712 in a denial of service. (CVE-2006-2937) [Steve Henson]
3714 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3715 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3717 *) Fix SSL client code which could crash if connecting to a
3718 malicious SSLv2 server. (CVE-2006-4343)
3719 [Tavis Ormandy and Will Drewry, Google Security Team]
3721 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3722 match only those. Before that, "AES256-SHA" would be interpreted
3723 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3724 the same strength classification in 0.9.7h) as we currently only
3725 have a single AES bit in the ciphersuite description bitmap.
3726 That change, however, also applied to ciphersuite strings such as
3727 "RC4-MD5" that intentionally matched multiple ciphersuites --
3728 namely, SSL 2.0 ciphersuites in addition to the more common ones
3729 from SSL 3.0/TLS 1.0.
3731 So we change the selection algorithm again: Naming an explicit
3732 ciphersuite selects this one ciphersuite, and any other similar
3733 ciphersuite (same bitmap) from *other* protocol versions.
3734 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3735 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3737 Since SSL 2.0 does not have any ciphersuites for which the
3738 128/256 bit distinction would be relevant, this works for now.
3739 The proper fix will be to use different bits for AES128 and
3740 AES256, which would have avoided the problems from the beginning;
3741 however, bits are scarce, so we can only do this in a new release
3742 (not just a patchlevel) when we can change the SSL_CIPHER
3743 definition to split the single 'unsigned long mask' bitmap into
3744 multiple values to extend the available space.
3748 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3750 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3751 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3753 *) Add AES IGE and biIGE modes.
3756 *) Change the Unix randomness entropy gathering to use poll() when
3757 possible instead of select(), since the latter has some
3758 undesirable limitations.
3759 [Darryl Miles via Richard Levitte and Bodo Moeller]
3761 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3762 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3763 cannot be implicitly activated as part of, e.g., the "AES" alias.
3764 However, please upgrade to OpenSSL 0.9.9[-dev] for
3765 non-experimental use of the ECC ciphersuites to get TLS extension
3766 support, which is required for curve and point format negotiation
3767 to avoid potential handshake problems.
3770 *) Disable rogue ciphersuites:
3772 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3773 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3774 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3776 The latter two were purportedly from
3777 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3780 Also deactivate the remaining ciphersuites from
3781 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3782 unofficial, and the ID has long expired.
3785 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3786 dual-core machines) and other potential thread-safety issues.
3789 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3790 versions), which is now available for royalty-free use
3791 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3792 Also, add Camellia TLS ciphersuites from RFC 4132.
3794 To minimize changes between patchlevels in the OpenSSL 0.9.8
3795 series, Camellia remains excluded from compilation unless OpenSSL
3796 is configured with 'enable-camellia'.
3799 *) Disable the padding bug check when compression is in use. The padding
3800 bug check assumes the first packet is of even length, this is not
3801 necessarily true if compresssion is enabled and can result in false
3802 positives causing handshake failure. The actual bug test is ancient
3803 code so it is hoped that implementations will either have fixed it by
3804 now or any which still have the bug do not support compression.
3807 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3809 *) When applying a cipher rule check to see if string match is an explicit
3810 cipher suite and only match that one cipher suite if it is.
3813 *) Link in manifests for VC++ if needed.
3814 [Austin Ziegler <halostatue@gmail.com>]
3816 *) Update support for ECC-based TLS ciphersuites according to
3817 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3818 TLS extensions, which are supported starting with the 0.9.9
3819 branch, not in the OpenSSL 0.9.8 branch).
3822 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3823 opaque EVP_CIPHER_CTX handling.
3826 *) Fixes and enhancements to zlib compression code. We now only use
3827 "zlib1.dll" and use the default __cdecl calling convention on Win32
3828 to conform with the standards mentioned here:
3829 http://www.zlib.net/DLL_FAQ.txt
3830 Static zlib linking now works on Windows and the new --with-zlib-include
3831 --with-zlib-lib options to Configure can be used to supply the location
3832 of the headers and library. Gracefully handle case where zlib library
3836 *) Several fixes and enhancements to the OID generation code. The old code
3837 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3838 handle numbers larger than ULONG_MAX, truncated printing and had a
3839 non standard OBJ_obj2txt() behaviour.
3842 *) Add support for building of engines under engine/ as shared libraries
3843 under VC++ build system.
3846 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3847 Hopefully, we will not see any false combination of paths any more.
3850 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3852 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3853 (part of SSL_OP_ALL). This option used to disable the
3854 countermeasure against man-in-the-middle protocol-version
3855 rollback in the SSL 2.0 server implementation, which is a bad
3856 idea. (CVE-2005-2969)
3858 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3859 for Information Security, National Institute of Advanced Industrial
3860 Science and Technology [AIST], Japan)]
3862 *) Add two function to clear and return the verify parameter flags.
3865 *) Keep cipherlists sorted in the source instead of sorting them at
3866 runtime, thus removing the need for a lock.
3869 *) Avoid some small subgroup attacks in Diffie-Hellman.
3870 [Nick Mathewson and Ben Laurie]
3872 *) Add functions for well-known primes.
3875 *) Extended Windows CE support.
3876 [Satoshi Nakamura and Andy Polyakov]
3878 *) Initialize SSL_METHOD structures at compile time instead of during
3879 runtime, thus removing the need for a lock.
3882 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3883 attempting to decrypt each encrypted key in turn. Add support to
3887 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3889 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3892 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3895 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3896 key into the same file any more.
3899 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3902 *) Add -utf8 command line and config file option to 'ca'.
3903 [Stefan <stf@udoma.org]
3905 *) Removed the macro des_crypt(), as it seems to conflict with some
3906 libraries. Use DES_crypt().
3909 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3910 involves renaming the source and generated shared-libs for
3911 both. The engines will accept the corrected or legacy ids
3912 ('ncipher' and '4758_cca' respectively) when binding. NB,
3913 this only applies when building 'shared'.
3914 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3916 *) Add attribute functions to EVP_PKEY structure. Modify
3917 PKCS12_create() to recognize a CSP name attribute and
3918 use it. Make -CSP option work again in pkcs12 utility.
3921 *) Add new functionality to the bn blinding code:
3922 - automatic re-creation of the BN_BLINDING parameters after
3923 a fixed number of uses (currently 32)
3924 - add new function for parameter creation
3925 - introduce flags to control the update behaviour of the
3926 BN_BLINDING parameters
3927 - hide BN_BLINDING structure
3928 Add a second BN_BLINDING slot to the RSA structure to improve
3929 performance when a single RSA object is shared among several
3933 *) Add support for DTLS.
3934 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3936 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3937 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3940 *) Remove buggy and incompletet DH cert support from
3941 ssl/ssl_rsa.c and ssl/s3_both.c
3944 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3945 the apps/openssl applications.
3948 *) Compile clean with "-Wall -Wmissing-prototypes
3949 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3950 DEBUG_SAFESTACK must also be set.
3953 *) Change ./Configure so that certain algorithms can be disabled by default.
3954 The new counterpiece to "no-xxx" is "enable-xxx".
3956 The patented RC5 and MDC2 algorithms will now be disabled unless
3957 "enable-rc5" and "enable-mdc2", respectively, are specified.
3959 (IDEA remains enabled despite being patented. This is because IDEA
3960 is frequently required for interoperability, and there is no license
3961 fee for non-commercial use. As before, "no-idea" can be used to
3962 avoid this algorithm.)
3966 *) Add processing of proxy certificates (see RFC 3820). This work was
3967 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3968 EGEE (Enabling Grids for E-science in Europe).
3971 *) RC4 performance overhaul on modern architectures/implementations, such
3972 as Intel P4, IA-64 and AMD64.
3975 *) New utility extract-section.pl. This can be used specify an alternative
3976 section number in a pod file instead of having to treat each file as
3977 a separate case in Makefile. This can be done by adding two lines to the
3980 =for comment openssl_section:XXX
3982 The blank line is mandatory.
3986 *) New arguments -certform, -keyform and -pass for s_client and s_server
3987 to allow alternative format key and certificate files and passphrase
3991 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3992 update associated structures and add various utility functions.
3994 Add new policy related verify parameters, include policy checking in
3995 standard verify code. Enhance 'smime' application with extra parameters
3996 to support policy checking and print out.
3999 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
4000 Nehemiah processors. These extensions support AES encryption in hardware
4001 as well as RNG (though RNG support is currently disabled).
4002 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
4004 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
4007 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
4008 [Andy Polyakov and a number of other people]
4010 *) Improved PowerPC platform support. Most notably BIGNUM assembler
4011 implementation contributed by IBM.
4012 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
4014 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
4015 exponent rather than 'unsigned long'. There is a corresponding change to
4016 the new 'rsa_keygen' element of the RSA_METHOD structure.
4017 [Jelte Jansen, Geoff Thorpe]
4019 *) Functionality for creating the initial serial number file is now
4020 moved from CA.pl to the 'ca' utility with a new option -create_serial.
4022 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
4023 number file to 1, which is bound to cause problems. To avoid
4024 the problems while respecting compatibility between different 0.9.7
4025 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
4026 CA.pl for serial number initialization. With the new release 0.9.8,
4027 we can fix the problem directly in the 'ca' utility.)
4030 *) Reduced header interdepencies by declaring more opaque objects in
4031 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
4032 give fewer recursive includes, which could break lazy source code - so
4033 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
4034 developers should define this symbol when building and using openssl to
4035 ensure they track the recommended behaviour, interfaces, [etc], but
4036 backwards-compatible behaviour prevails when this isn't defined.
4039 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
4042 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
4043 This will generate a random key of the appropriate length based on the
4044 cipher context. The EVP_CIPHER can provide its own random key generation
4045 routine to support keys of a specific form. This is used in the des and
4046 3des routines to generate a key of the correct parity. Update S/MIME
4047 code to use new functions and hence generate correct parity DES keys.
4048 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
4049 valid (weak or incorrect parity).
4052 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
4053 as looking them up. This is useful when the verified structure may contain
4054 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
4055 present unless the new PKCS7_NO_CRL flag is asserted.
4058 *) Extend ASN1 oid configuration module. It now additionally accepts the
4061 shortName = some long name, 1.2.3.4
4064 *) Reimplemented the BN_CTX implementation. There is now no more static
4065 limitation on the number of variables it can handle nor the depth of the
4066 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
4067 information can now expand as required, and rather than having a single
4068 static array of bignums, BN_CTX now uses a linked-list of such arrays
4069 allowing it to expand on demand whilst maintaining the usefulness of
4070 BN_CTX's "bundling".
4073 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
4074 to allow all RSA operations to function using a single BN_CTX.
4077 *) Preliminary support for certificate policy evaluation and checking. This
4078 is initially intended to pass the tests outlined in "Conformance Testing
4079 of Relying Party Client Certificate Path Processing Logic" v1.07.
4082 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
4083 remained unused and not that useful. A variety of other little bignum
4084 tweaks and fixes have also been made continuing on from the audit (see
4088 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4089 associated ASN1, EVP and SSL functions and old ASN1 macros.
4092 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4093 and this should never fail. So the return value from the use of
4094 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4095 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4098 *) BN_CTX_get() should return zero-valued bignums, providing the same
4099 initialised value as BN_new().
4100 [Geoff Thorpe, suggested by Ulf Möller]
4102 *) Support for inhibitAnyPolicy certificate extension.
4105 *) An audit of the BIGNUM code is underway, for which debugging code is
4106 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4107 is considered valid when processing BIGNUMs, and causes execution to
4108 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4109 further steps are taken to deliberately pollute unused data in BIGNUM
4110 structures to try and expose faulty code further on. For now, openssl will
4111 (in its default mode of operation) continue to tolerate the inconsistent
4112 forms that it has tolerated in the past, but authors and packagers should
4113 consider trying openssl and their own applications when compiled with
4114 these debugging symbols defined. It will help highlight potential bugs in
4115 their own code, and will improve the test coverage for OpenSSL itself. At
4116 some point, these tighter rules will become openssl's default to improve
4117 maintainability, though the assert()s and other overheads will remain only
4118 in debugging configurations. See bn.h for more details.
4119 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4121 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4122 that can only be obtained through BN_CTX_new() (which implicitly
4123 initialises it). The presence of this function only made it possible
4124 to overwrite an existing structure (and cause memory leaks).
4127 *) Because of the callback-based approach for implementing LHASH as a
4128 template type, lh_insert() adds opaque objects to hash-tables and
4129 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4130 to clean up those corresponding objects before destroying the hash table
4131 (and losing the object pointers). So some over-zealous constifications in
4132 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4133 objects as "const" and the lh_doall[_arg] callback wrappers are not
4134 prototyped to have "const" restrictions on the object pointers they are
4135 given (and so aren't required to cast them away any more).
4138 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4139 (speed) prefers to use its own implementation. The two implementations
4140 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4141 its object type properly exposed (MS_TM) instead of casting to/from "char
4142 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4143 aren't necessarily the greatest nomenclatures - but this is what was used
4144 internally to the implementation so I've used that for now.
4147 *) Ensure that deprecated functions do not get compiled when
4148 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4149 the self-tests were still using deprecated key-generation functions so
4150 these have been updated also.
4153 *) Reorganise PKCS#7 code to separate the digest location functionality
4154 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
4155 New function PKCS7_set_digest() to set the digest type for PKCS#7
4156 digestedData type. Add additional code to correctly generate the
4157 digestedData type and add support for this type in PKCS7 initialization
4161 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4162 structure of type "other".
4165 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4166 sure the loop does correctly stop and breaking ("division by zero")
4167 modulus operations are not performed. The (pre-generated) prime
4168 table crypto/bn/bn_prime.h was already correct, but it could not be
4169 re-generated on some platforms because of the "division by zero"
4170 situation in the script.
4171 [Ralf S. Engelschall]
4173 *) Update support for ECC-based TLS ciphersuites according to
4174 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4175 SHA-1 now is only used for "small" curves (where the
4176 representation of a field element takes up to 24 bytes); for
4177 larger curves, the field element resulting from ECDH is directly
4178 used as premaster secret.
4179 [Douglas Stebila (Sun Microsystems Laboratories)]
4181 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4182 curve secp160r1 to the tests.
4183 [Douglas Stebila (Sun Microsystems Laboratories)]
4185 *) Add the possibility to load symbols globally with DSO.
4186 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4188 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4189 control of the error stack.
4192 *) Add support for STORE in ENGINE.
4195 *) Add the STORE type. The intention is to provide a common interface
4196 to certificate and key stores, be they simple file-based stores, or
4197 HSM-type store, or LDAP stores, or...
4198 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4201 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4202 pass a list of arguments to any function as well as provide a way
4203 for a function to pass data back to the caller.
4206 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4207 works like BUF_strdup() but can be used to duplicate a portion of
4208 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4212 *) Add the function sk_find_ex() which works like sk_find(), but will
4213 return an index to an element even if an exact match couldn't be
4214 found. The index is guaranteed to point at the element where the
4215 searched-for key would be inserted to preserve sorting order.
4218 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4219 takes an extra flags argument for optional functionality. Currently,
4220 the following flags are defined:
4222 OBJ_BSEARCH_VALUE_ON_NOMATCH
4223 This one gets OBJ_bsearch_ex() to return a pointer to the first
4224 element where the comparing function returns a negative or zero
4227 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4228 This one gets OBJ_bsearch_ex() to return a pointer to the first
4229 element where the comparing function returns zero. This is useful
4230 if there are more than one element where the comparing function
4234 *) Make it possible to create self-signed certificates with 'openssl ca'
4235 in such a way that the self-signed certificate becomes part of the
4236 CA database and uses the same mechanisms for serial number generation
4237 as all other certificate signing. The new flag '-selfsign' enables
4238 this functionality. Adapt CA.sh and CA.pl.in.
4241 *) Add functionality to check the public key of a certificate request
4242 against a given private. This is useful to check that a certificate
4243 request can be signed by that key (self-signing).
4246 *) Make it possible to have multiple active certificates with the same
4247 subject in the CA index file. This is done only if the keyword
4248 'unique_subject' is set to 'no' in the main CA section (default
4249 if 'CA_default') of the configuration file. The value is saved
4250 with the database itself in a separate index attribute file,
4251 named like the index file with '.attr' appended to the name.
4254 *) Generate muti valued AVAs using '+' notation in config files for
4258 *) Support for nameConstraints certificate extension.
4261 *) Support for policyConstraints certificate extension.
4264 *) Support for policyMappings certificate extension.
4267 *) Make sure the default DSA_METHOD implementation only uses its
4268 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4269 and change its own handlers to be NULL so as to remove unnecessary
4270 indirection. This lets alternative implementations fallback to the
4271 default implementation more easily.
4274 *) Support for directoryName in GeneralName related extensions
4278 *) Make it possible to link applications using Makefile.shared.
4279 Make that possible even when linking against static libraries!
4282 *) Support for single pass processing for S/MIME signing. This now
4283 means that S/MIME signing can be done from a pipe, in addition
4284 cleartext signing (multipart/signed type) is effectively streaming
4285 and the signed data does not need to be all held in memory.
4287 This is done with a new flag PKCS7_STREAM. When this flag is set
4288 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4289 is done after the data is output (and digests calculated) in
4290 SMIME_write_PKCS7().
4293 *) Add full support for -rpath/-R, both in shared libraries and
4294 applications, at least on the platforms where it's known how
4298 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4299 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4300 will now compute a table of multiples of the generator that
4301 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4302 faster (notably in the case of a single point multiplication,
4303 scalar * generator).
4304 [Nils Larsch, Bodo Moeller]
4306 *) IPv6 support for certificate extensions. The various extensions
4307 which use the IP:a.b.c.d can now take IPv6 addresses using the
4308 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4312 *) Added an ENGINE that implements RSA by performing private key
4313 exponentiations with the GMP library. The conversions to and from
4314 GMP's mpz_t format aren't optimised nor are any montgomery forms
4315 cached, and on x86 it appears OpenSSL's own performance has caught up.
4316 However there are likely to be other architectures where GMP could
4317 provide a boost. This ENGINE is not built in by default, but it can be
4318 specified at Configure time and should be accompanied by the necessary
4319 linker additions, eg;
4320 ./config -DOPENSSL_USE_GMP -lgmp
4323 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4324 testing availability of engines with "-t" - the old behaviour is
4325 produced by increasing the feature's verbosity with "-tt".
4328 *) ECDSA routines: under certain error conditions uninitialized BN objects
4329 could be freed. Solution: make sure initialization is performed early
4330 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4334 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4335 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4336 software implementations. For DSA and DH, parameter generation can
4337 also be overriden by providing the appropriate method callbacks.
4340 *) Change the "progress" mechanism used in key-generation and
4341 primality testing to functions that take a new BN_GENCB pointer in
4342 place of callback/argument pairs. The new API functions have "_ex"
4343 postfixes and the older functions are reimplemented as wrappers for
4344 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4345 declarations of the old functions to help (graceful) attempts to
4346 migrate to the new functions. Also, the new key-generation API
4347 functions operate on a caller-supplied key-structure and return
4348 success/failure rather than returning a key or NULL - this is to
4349 help make "keygen" another member function of RSA_METHOD etc.
4351 Example for using the new callback interface:
4353 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4357 BN_GENCB_set(&my_cb, my_callback, my_arg);
4359 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4360 /* For the meaning of a, b in calls to my_callback(), see the
4361 * documentation of the function that calls the callback.
4362 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4363 * my_callback should return 1 if it wants BN_is_prime_ex()
4364 * to continue, or 0 to stop.
4369 *) Change the ZLIB compression method to be stateful, and make it
4370 available to TLS with the number defined in
4371 draft-ietf-tls-compression-04.txt.
4374 *) Add the ASN.1 structures and functions for CertificatePair, which
4375 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4377 CertificatePair ::= SEQUENCE {
4378 forward [0] Certificate OPTIONAL,
4379 reverse [1] Certificate OPTIONAL,
4380 -- at least one of the pair shall be present -- }
4382 Also implement the PEM functions to read and write certificate
4383 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4385 This needed to be defined, mostly for the sake of the LDAP
4386 attribute crossCertificatePair, but may prove useful elsewhere as
4390 *) Make it possible to inhibit symlinking of shared libraries in
4391 Makefile.shared, for Cygwin's sake.
4394 *) Extend the BIGNUM API by creating a function
4395 void BN_set_negative(BIGNUM *a, int neg);
4396 and a macro that behave like
4397 int BN_is_negative(const BIGNUM *a);
4399 to avoid the need to access 'a->neg' directly in applications.
4402 *) Implement fast modular reduction for pseudo-Mersenne primes
4403 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4404 EC_GROUP_new_curve_GFp() will now automatically use this
4406 [Nils Larsch <nla@trustcenter.de>]
4408 *) Add new lock type (CRYPTO_LOCK_BN).
4411 *) Change the ENGINE framework to automatically load engines
4412 dynamically from specific directories unless they could be
4413 found to already be built in or loaded. Move all the
4414 current engines except for the cryptodev one to a new
4416 The engines in engines/ are built as shared libraries if
4417 the "shared" options was given to ./Configure or ./config.
4418 Otherwise, they are inserted in libcrypto.a.
4419 /usr/local/ssl/engines is the default directory for dynamic
4420 engines, but that can be overriden at configure time through
4421 the usual use of --prefix and/or --openssldir, and at run
4422 time with the environment variable OPENSSL_ENGINES.
4423 [Geoff Thorpe and Richard Levitte]
4425 *) Add Makefile.shared, a helper makefile to build shared
4426 libraries. Addapt Makefile.org.
4429 *) Add version info to Win32 DLLs.
4430 [Peter 'Luna' Runestig" <peter@runestig.com>]
4432 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4433 can be added using this API to created arbitrary PKCS#12
4434 files while avoiding the low level API.
4436 New options to PKCS12_create(), key or cert can be NULL and
4437 will then be omitted from the output file. The encryption
4438 algorithm NIDs can be set to -1 for no encryption, the mac
4439 iteration count can be set to 0 to omit the mac.
4441 Enhance pkcs12 utility by making the -nokeys and -nocerts
4442 options work when creating a PKCS#12 file. New option -nomac
4443 to omit the mac, NONE can be set for an encryption algorithm.
4444 New code is modified to use the enhanced PKCS12_create()
4445 instead of the low level API.
4448 *) Extend ASN1 encoder to support indefinite length constructed
4449 encoding. This can output sequences tags and octet strings in
4450 this form. Modify pk7_asn1.c to support indefinite length
4451 encoding. This is experimental and needs additional code to
4452 be useful, such as an ASN1 bio and some enhanced streaming
4455 Extend template encode functionality so that tagging is passed
4456 down to the template encoder.
4459 *) Let 'openssl req' fail if an argument to '-newkey' is not
4460 recognized instead of using RSA as a default.
4463 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4464 As these are not official, they are not included in "ALL";
4465 the "ECCdraft" ciphersuite group alias can be used to select them.
4466 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4468 *) Add ECDH engine support.
4469 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4471 *) Add ECDH in new directory crypto/ecdh/.
4472 [Douglas Stebila (Sun Microsystems Laboratories)]
4474 *) Let BN_rand_range() abort with an error after 100 iterations
4475 without success (which indicates a broken PRNG).
4478 *) Change BN_mod_sqrt() so that it verifies that the input value
4479 is really the square of the return value. (Previously,
4480 BN_mod_sqrt would show GIGO behaviour.)
4483 *) Add named elliptic curves over binary fields from X9.62, SECG,
4484 and WAP/WTLS; add OIDs that were still missing.
4486 [Sheueling Chang Shantz and Douglas Stebila
4487 (Sun Microsystems Laboratories)]
4489 *) Extend the EC library for elliptic curves over binary fields
4490 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4493 EC_GF2m_simple_method
4497 EC_GROUP_new_curve_GF2m
4498 EC_GROUP_set_curve_GF2m
4499 EC_GROUP_get_curve_GF2m
4500 EC_POINT_set_affine_coordinates_GF2m
4501 EC_POINT_get_affine_coordinates_GF2m
4502 EC_POINT_set_compressed_coordinates_GF2m
4504 Point compression for binary fields is disabled by default for
4505 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4508 As binary polynomials are represented as BIGNUMs, various members
4509 of the EC_GROUP and EC_POINT data structures can be shared
4510 between the implementations for prime fields and binary fields;
4511 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4512 are essentially identical to their ..._GFp counterparts.
4513 (For simplicity, the '..._GFp' prefix has been dropped from
4514 various internal method names.)
4516 An internal 'field_div' method (similar to 'field_mul' and
4517 'field_sqr') has been added; this is used only for binary fields.
4519 [Sheueling Chang Shantz and Douglas Stebila
4520 (Sun Microsystems Laboratories)]
4522 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4523 through methods ('mul', 'precompute_mult').
4525 The generic implementations (now internally called 'ec_wNAF_mul'
4526 and 'ec_wNAF_precomputed_mult') remain the default if these
4527 methods are undefined.
4529 [Sheueling Chang Shantz and Douglas Stebila
4530 (Sun Microsystems Laboratories)]
4532 *) New function EC_GROUP_get_degree, which is defined through
4533 EC_METHOD. For curves over prime fields, this returns the bit
4534 length of the modulus.
4536 [Sheueling Chang Shantz and Douglas Stebila
4537 (Sun Microsystems Laboratories)]
4539 *) New functions EC_GROUP_dup, EC_POINT_dup.
4540 (These simply call ..._new and ..._copy).
4542 [Sheueling Chang Shantz and Douglas Stebila
4543 (Sun Microsystems Laboratories)]
4545 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4546 Polynomials are represented as BIGNUMs (where the sign bit is not
4547 used) in the following functions [macros]:
4550 BN_GF2m_sub [= BN_GF2m_add]
4551 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4552 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4553 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4555 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4556 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4557 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4558 BN_GF2m_cmp [= BN_ucmp]
4560 (Note that only the 'mod' functions are actually for fields GF(2^m).
4561 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4563 For some functions, an the irreducible polynomial defining a
4564 field can be given as an 'unsigned int[]' with strictly
4565 decreasing elements giving the indices of those bits that are set;
4566 i.e., p[] represents the polynomial
4567 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4569 p[0] > p[1] > ... > p[k] = 0.
4570 This applies to the following functions:
4575 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4576 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4578 BN_GF2m_mod_sqrt_arr
4579 BN_GF2m_mod_solve_quad_arr
4583 Conversion can be performed by the following functions:
4588 bntest.c has additional tests for binary polynomial arithmetic.
4590 Two implementations for BN_GF2m_mod_div() are available.
4591 The default algorithm simply uses BN_GF2m_mod_inv() and
4592 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4593 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4594 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4596 [Sheueling Chang Shantz and Douglas Stebila
4597 (Sun Microsystems Laboratories)]
4599 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4600 functionality is disabled at compile-time.
4601 [Douglas Stebila <douglas.stebila@sun.com>]
4603 *) Change default behaviour of 'openssl asn1parse' so that more
4604 information is visible when viewing, e.g., a certificate:
4606 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4607 mode the content of non-printable OCTET STRINGs is output in a
4608 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4609 avoid the appearance of a printable string.
4610 [Nils Larsch <nla@trustcenter.de>]
4612 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4614 EC_GROUP_set_asn1_flag()
4615 EC_GROUP_get_asn1_flag()
4616 EC_GROUP_set_point_conversion_form()
4617 EC_GROUP_get_point_conversion_form()
4618 These control ASN1 encoding details:
4619 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4620 has been set to OPENSSL_EC_NAMED_CURVE.
4621 - Points are encoded in uncompressed form by default; options for
4622 asn1_for are as for point2oct, namely
4623 POINT_CONVERSION_COMPRESSED
4624 POINT_CONVERSION_UNCOMPRESSED
4625 POINT_CONVERSION_HYBRID
4627 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4630 EC_GROUP_get0_seed()
4631 EC_GROUP_get_seed_len()
4632 This is used only for ASN1 purposes (so far).
4633 [Nils Larsch <nla@trustcenter.de>]
4635 *) Add 'field_type' member to EC_METHOD, which holds the NID
4636 of the appropriate field type OID. The new function
4637 EC_METHOD_get_field_type() returns this value.
4638 [Nils Larsch <nla@trustcenter.de>]
4643 EC_POINT_point2hex()
4644 EC_POINT_hex2point()
4645 providing useful interfaces to EC_POINT_point2oct() and
4646 EC_POINT_oct2point().
4647 [Nils Larsch <nla@trustcenter.de>]
4649 *) Change internals of the EC library so that the functions
4650 EC_GROUP_set_generator()
4651 EC_GROUP_get_generator()
4652 EC_GROUP_get_order()
4653 EC_GROUP_get_cofactor()
4654 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4655 to methods, which would lead to unnecessary code duplication when
4656 adding different types of curves.
4657 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4659 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4660 arithmetic, and such that modified wNAFs are generated
4661 (which avoid length expansion in many cases).
4664 *) Add a function EC_GROUP_check_discriminant() (defined via
4665 EC_METHOD) that verifies that the curve discriminant is non-zero.
4667 Add a function EC_GROUP_check() that makes some sanity tests
4668 on a EC_GROUP, its generator and order. This includes
4669 EC_GROUP_check_discriminant().
4670 [Nils Larsch <nla@trustcenter.de>]
4672 *) Add ECDSA in new directory crypto/ecdsa/.
4674 Add applications 'openssl ecparam' and 'openssl ecdsa'
4675 (these are based on 'openssl dsaparam' and 'openssl dsa').
4677 ECDSA support is also included in various other files across the
4678 library. Most notably,
4679 - 'openssl req' now has a '-newkey ecdsa:file' option;
4680 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4681 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4682 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4683 them suitable for ECDSA where domain parameters must be
4684 extracted before the specific public key;
4685 - ECDSA engine support has been added.
4686 [Nils Larsch <nla@trustcenter.de>]
4688 *) Include some named elliptic curves, and add OIDs from X9.62,
4689 SECG, and WAP/WTLS. Each curve can be obtained from the new
4691 EC_GROUP_new_by_curve_name(),
4692 and the list of available named curves can be obtained with
4693 EC_get_builtin_curves().
4694 Also add a 'curve_name' member to EC_GROUP objects, which can be
4696 EC_GROUP_set_curve_name()
4697 EC_GROUP_get_curve_name()
4698 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4700 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4701 was actually never needed) and in BN_mul(). The removal in BN_mul()
4702 required a small change in bn_mul_part_recursive() and the addition
4703 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4704 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4705 bn_sub_words() and bn_add_words() except they take arrays with
4709 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4711 *) Cleanse PEM buffers before freeing them since they may contain
4713 [Benjamin Bennett <ben@psc.edu>]
4715 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4716 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4717 authentication-only ciphersuites.
4720 *) Since AES128 and AES256 share a single mask bit in the logic of
4721 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4722 kludge to work properly if AES128 is available and AES256 isn't.
4725 *) Expand security boundary to match 1.1.1 module.
4728 *) Remove redundant features: hash file source, editing of test vectors
4729 modify fipsld to use external fips_premain.c signature.
4732 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4733 run algorithm test programs.
4736 *) Make algorithm test programs more tolerant of whitespace.
4739 *) Have SSL/TLS server implementation tolerate "mismatched" record
4740 protocol version while receiving ClientHello even if the
4741 ClientHello is fragmented. (The server can't insist on the
4742 particular protocol version it has chosen before the ServerHello
4743 message has informed the client about his choice.)
4746 *) Load error codes if they are not already present instead of using a
4747 static variable. This allows them to be cleanly unloaded and reloaded.
4750 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4752 *) Introduce limits to prevent malicious keys being able to
4753 cause a denial of service. (CVE-2006-2940)
4754 [Steve Henson, Bodo Moeller]
4756 *) Fix ASN.1 parsing of certain invalid structures that can result
4757 in a denial of service. (CVE-2006-2937) [Steve Henson]
4759 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4760 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4762 *) Fix SSL client code which could crash if connecting to a
4763 malicious SSLv2 server. (CVE-2006-4343)
4764 [Tavis Ormandy and Will Drewry, Google Security Team]
4766 *) Change ciphersuite string processing so that an explicit
4767 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4768 will no longer include "AES128-SHA"), and any other similar
4769 ciphersuite (same bitmap) from *other* protocol versions (so that
4770 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4771 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4772 changes from 0.9.8b and 0.9.8d.
4775 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4777 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4778 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4780 *) Change the Unix randomness entropy gathering to use poll() when
4781 possible instead of select(), since the latter has some
4782 undesirable limitations.
4783 [Darryl Miles via Richard Levitte and Bodo Moeller]
4785 *) Disable rogue ciphersuites:
4787 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4788 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4789 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4791 The latter two were purportedly from
4792 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4795 Also deactive the remaining ciphersuites from
4796 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4797 unofficial, and the ID has long expired.
4800 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4801 dual-core machines) and other potential thread-safety issues.
4804 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4806 *) Adapt fipsld and the build system to link against the validated FIPS
4807 module in FIPS mode.
4810 *) Fixes for VC++ 2005 build under Windows.
4813 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4814 from a Windows bash shell such as MSYS. It is autodetected from the
4815 "config" script when run from a VC++ environment. Modify standard VC++
4816 build to use fipscanister.o from the GNU make build.
4819 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4821 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4822 The value now differs depending on if you build for FIPS or not.
4823 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4824 safely run with a non-FIPSed libcrypto, as it may crash because of
4825 the difference induced by this change.
4828 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4830 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4831 (part of SSL_OP_ALL). This option used to disable the
4832 countermeasure against man-in-the-middle protocol-version
4833 rollback in the SSL 2.0 server implementation, which is a bad
4834 idea. (CVE-2005-2969)
4836 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4837 for Information Security, National Institute of Advanced Industrial
4838 Science and Technology [AIST], Japan)]
4840 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4841 mainly for FIPS compliance and not fully integrated at this stage.
4844 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4845 the exponentiation using a fixed-length exponent. (Otherwise,
4846 the information leaked through timing could expose the secret key
4847 after many signatures; cf. Bleichenbacher's attack on DSA with
4851 *) Make a new fixed-window mod_exp implementation the default for
4852 RSA, DSA, and DH private-key operations so that the sequence of
4853 squares and multiplies and the memory access pattern are
4854 independent of the particular secret key. This will mitigate
4855 cache-timing and potential related attacks.
4857 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4858 and this is automatically used by BN_mod_exp_mont() if the new flag
4859 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4860 will use this BN flag for private exponents unless the flag
4861 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4862 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4864 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4866 *) Change the client implementation for SSLv23_method() and
4867 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4868 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4869 (Previously, the SSL 2.0 backwards compatible Client Hello
4870 message format would be used even with SSL_OP_NO_SSLv2.)
4873 *) Add support for smime-type MIME parameter in S/MIME messages which some
4877 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4878 a threadsafe manner. Modify rsa code to use new function and add calls
4879 to dsa and dh code (which had race conditions before).
4882 *) Include the fixed error library code in the C error file definitions
4883 instead of fixing them up at runtime. This keeps the error code
4884 structures constant.
4887 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4889 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4892 *) Fixes for newer kerberos headers. NB: the casts are needed because
4893 the 'length' field is signed on one version and unsigned on another
4894 with no (?) obvious way to tell the difference, without these VC++
4895 complains. Also the "definition" of FAR (blank) is no longer included
4896 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4897 some needed definitions.
4900 *) Undo Cygwin change.
4903 *) Added support for proxy certificates according to RFC 3820.
4904 Because they may be a security thread to unaware applications,
4905 they must be explicitely allowed in run-time. See
4906 docs/HOWTO/proxy_certificates.txt for further information.
4909 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4911 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4912 server and client random values. Previously
4913 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4914 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4916 This change has negligible security impact because:
4918 1. Server and client random values still have 24 bytes of pseudo random
4921 2. Server and client random values are sent in the clear in the initial
4924 3. The master secret is derived using the premaster secret (48 bytes in
4925 size for static RSA ciphersuites) as well as client server and random
4928 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4931 [Stephen Henson, reported by UK NISCC]
4933 *) Use Windows randomness collection on Cygwin.
4936 *) Fix hang in EGD/PRNGD query when communication socket is closed
4937 prematurely by EGD/PRNGD.
4938 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4940 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4943 *) Back-port of selected performance improvements from development
4944 branch, as well as improved support for PowerPC platforms.
4947 *) Add lots of checks for memory allocation failure, error codes to indicate
4948 failure and freeing up memory if a failure occurs.
4949 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4951 *) Add new -passin argument to dgst.
4954 *) Perform some character comparisons of different types in X509_NAME_cmp:
4955 this is needed for some certificates that reencode DNs into UTF8Strings
4956 (in violation of RFC3280) and can't or wont issue name rollover
4960 *) Make an explicit check during certificate validation to see that
4961 the CA setting in each certificate on the chain is correct. As a
4962 side effect always do the following basic checks on extensions,
4963 not just when there's an associated purpose to the check:
4965 - if there is an unhandled critical extension (unless the user
4966 has chosen to ignore this fault)
4967 - if the path length has been exceeded (if one is set at all)
4968 - that certain extensions fit the associated purpose (if one has
4972 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4974 *) Avoid a race condition when CRLs are checked in a multi threaded
4975 environment. This would happen due to the reordering of the revoked
4976 entries during signature checking and serial number lookup. Now the
4977 encoding is cached and the serial number sort performed under a lock.
4978 Add new STACK function sk_is_sorted().
4981 *) Add Delta CRL to the extension code.
4984 *) Various fixes to s3_pkt.c so alerts are sent properly.
4985 [David Holmes <d.holmes@f5.com>]
4987 *) Reduce the chances of duplicate issuer name and serial numbers (in
4988 violation of RFC3280) using the OpenSSL certificate creation utilities.
4989 This is done by creating a random 64 bit value for the initial serial
4990 number when a serial number file is created or when a self signed
4991 certificate is created using 'openssl req -x509'. The initial serial
4992 number file is created using 'openssl x509 -next_serial' in CA.pl
4993 rather than being initialized to 1.
4996 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4998 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4999 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5000 [Joe Orton, Steve Henson]
5002 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
5004 [Joe Orton, Steve Henson]
5006 *) Make it possible to have multiple active certificates with the same
5007 subject in the CA index file. This is done only if the keyword
5008 'unique_subject' is set to 'no' in the main CA section (default
5009 if 'CA_default') of the configuration file. The value is saved
5010 with the database itself in a separate index attribute file,
5011 named like the index file with '.attr' appended to the name.
5014 *) X509 verify fixes. Disable broken certificate workarounds when
5015 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
5016 keyUsage extension present. Don't accept CRLs with unhandled critical
5017 extensions: since verify currently doesn't process CRL extensions this
5018 rejects a CRL with *any* critical extensions. Add new verify error codes
5022 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
5023 A clarification of RFC2560 will require the use of OCTET STRINGs and
5024 some implementations cannot handle the current raw format. Since OpenSSL
5025 copies and compares OCSP nonces as opaque blobs without any attempt at
5026 parsing them this should not create any compatibility issues.
5029 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
5030 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
5031 this HMAC (and other) operations are several times slower than OpenSSL
5035 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
5036 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5038 *) Use the correct content when signing type "other".
5041 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5043 *) Fix various bugs revealed by running the NISCC test suite:
5045 Stop out of bounds reads in the ASN1 code when presented with
5046 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5048 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
5050 If verify callback ignores invalid public key errors don't try to check
5051 certificate signature with the NULL public key.
5055 *) New -ignore_err option in ocsp application to stop the server
5056 exiting on the first error in a request.
5059 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5060 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5064 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5065 extra data after the compression methods not only for TLS 1.0
5066 but also for SSL 3.0 (as required by the specification).
5067 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5069 *) Change X509_certificate_type() to mark the key as exported/exportable
5070 when it's 512 *bits* long, not 512 bytes.
5073 *) Change AES_cbc_encrypt() so it outputs exact multiple of
5074 blocks during encryption.
5077 *) Various fixes to base64 BIO and non blocking I/O. On write
5078 flushes were not handled properly if the BIO retried. On read
5079 data was not being buffered properly and had various logic bugs.
5080 This also affects blocking I/O when the data being decoded is a
5084 *) Various S/MIME bugfixes and compatibility changes:
5085 output correct application/pkcs7 MIME type if
5086 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5087 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5088 of files as .eml work). Correctly handle very long lines in MIME
5092 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5094 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5095 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5096 a protocol version number mismatch like a decryption error
5097 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5100 *) Turn on RSA blinding by default in the default implementation
5101 to avoid a timing attack. Applications that don't want it can call
5102 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5103 They would be ill-advised to do so in most cases.
5104 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5106 *) Change RSA blinding code so that it works when the PRNG is not
5107 seeded (in this case, the secret RSA exponent is abused as
5108 an unpredictable seed -- if it is not unpredictable, there
5109 is no point in blinding anyway). Make RSA blinding thread-safe
5110 by remembering the creator's thread ID in rsa->blinding and
5111 having all other threads use local one-time blinding factors
5112 (this requires more computation than sharing rsa->blinding, but
5113 avoids excessive locking; and if an RSA object is not shared
5114 between threads, blinding will still be very fast).
5117 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5118 ENGINE as defaults for all supported algorithms irrespective of
5119 the 'flags' parameter. 'flags' is now honoured, so applications
5120 should make sure they are passing it correctly.
5123 *) Target "mingw" now allows native Windows code to be generated in
5124 the Cygwin environment as well as with the MinGW compiler.
5127 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5129 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5130 via timing by performing a MAC computation even if incorrrect
5131 block cipher padding has been found. This is a countermeasure
5132 against active attacks where the attacker has to distinguish
5133 between bad padding and a MAC verification error. (CVE-2003-0078)
5135 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5136 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5137 Martin Vuagnoux (EPFL, Ilion)]
5139 *) Make the no-err option work as intended. The intention with no-err
5140 is not to have the whole error stack handling routines removed from
5141 libcrypto, it's only intended to remove all the function name and
5142 reason texts, thereby removing some of the footprint that may not
5143 be interesting if those errors aren't displayed anyway.
5145 NOTE: it's still possible for any application or module to have it's
5146 own set of error texts inserted. The routines are there, just not
5147 used by default when no-err is given.
5150 *) Add support for FreeBSD on IA64.
5151 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5153 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5154 Kerberos function mit_des_cbc_cksum(). Before this change,
5155 the value returned by DES_cbc_cksum() was like the one from
5156 mit_des_cbc_cksum(), except the bytes were swapped.
5157 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5159 *) Allow an application to disable the automatic SSL chain building.
5160 Before this a rather primitive chain build was always performed in
5161 ssl3_output_cert_chain(): an application had no way to send the
5162 correct chain if the automatic operation produced an incorrect result.
5164 Now the chain builder is disabled if either:
5166 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5168 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5170 The reasoning behind this is that an application would not want the
5171 auto chain building to take place if extra chain certificates are
5172 present and it might also want a means of sending no additional
5173 certificates (for example the chain has two certificates and the
5177 *) Add the possibility to build without the ENGINE framework.
5178 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5180 *) Under Win32 gmtime() can return NULL: check return value in
5181 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5184 *) DSA routines: under certain error conditions uninitialized BN objects
5185 could be freed. Solution: make sure initialization is performed early
5186 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5187 Nils Larsch <nla@trustcenter.de> via PR#459)
5190 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5191 checked on reconnect on the client side, therefore session resumption
5192 could still fail with a "ssl session id is different" error. This
5193 behaviour is masked when SSL_OP_ALL is used due to
5194 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5195 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5196 followup to PR #377.
5199 *) IA-32 assembler support enhancements: unified ELF targets, support
5200 for SCO/Caldera platforms, fix for Cygwin shared build.
5203 *) Add support for FreeBSD on sparc64. As a consequence, support for
5204 FreeBSD on non-x86 processors is separate from x86 processors on
5205 the config script, much like the NetBSD support.
5206 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5208 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5210 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5213 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5214 code (06) was taken as the first octet of the session ID and the last
5215 octet was ignored consequently. As a result SSLv2 client side session
5216 caching could not have worked due to the session ID mismatch between
5218 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5222 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5223 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5227 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5228 seems that in spite of existing for more than a year, many application
5229 author have done nothing to provide the necessary callbacks, which
5230 means that this particular engine will not work properly anywhere.
5231 This is a very unfortunate situation which forces us, in the name
5232 of usability, to give the hw_ncipher.c a static lock, which is part
5234 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5235 appear in 0.9.8 or later. We EXPECT application authors to have
5236 dealt properly with this when 0.9.8 is released (unless we actually
5237 make such changes in the libcrypto locking code that changes will
5238 have to be made anyway).
5241 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5242 octets have been read, EOF or an error occurs. Without this change
5243 some truncated ASN1 structures will not produce an error.
5246 *) Disable Heimdal support, since it hasn't been fully implemented.
5247 Still give the possibility to force the use of Heimdal, but with
5248 warnings and a request that patches get sent to openssl-dev.
5251 *) Add the VC-CE target, introduce the WINCE sysname, and add
5252 INSTALL.WCE and appropriate conditionals to make it build.
5253 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5255 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5256 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5257 edit numbers of the version.
5258 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5260 *) Introduce safe string copy and catenation functions
5261 (BUF_strlcpy() and BUF_strlcat()).
5262 [Ben Laurie (CHATS) and Richard Levitte]
5264 *) Avoid using fixed-size buffers for one-line DNs.
5265 [Ben Laurie (CHATS)]
5267 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5268 resizing buffers containing secrets, and use where appropriate.
5269 [Ben Laurie (CHATS)]
5271 *) Avoid using fixed size buffers for configuration file location.
5272 [Ben Laurie (CHATS)]
5274 *) Avoid filename truncation for various CA files.
5275 [Ben Laurie (CHATS)]
5277 *) Use sizeof in preference to magic numbers.
5278 [Ben Laurie (CHATS)]
5280 *) Avoid filename truncation in cert requests.
5281 [Ben Laurie (CHATS)]
5283 *) Add assertions to check for (supposedly impossible) buffer
5285 [Ben Laurie (CHATS)]
5287 *) Don't cache truncated DNS entries in the local cache (this could
5288 potentially lead to a spoofing attack).
5289 [Ben Laurie (CHATS)]
5291 *) Fix various buffers to be large enough for hex/decimal
5292 representations in a platform independent manner.
5293 [Ben Laurie (CHATS)]
5295 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5296 resizing buffers containing secrets, and use where appropriate.
5297 [Ben Laurie (CHATS)]
5299 *) Add BIO_indent() to avoid much slightly worrying code to do
5301 [Ben Laurie (CHATS)]
5303 *) Convert sprintf()/BIO_puts() to BIO_printf().
5304 [Ben Laurie (CHATS)]
5306 *) buffer_gets() could terminate with the buffer only half
5308 [Ben Laurie (CHATS)]
5310 *) Add assertions to prevent user-supplied crypto functions from
5311 overflowing internal buffers by having large block sizes, etc.
5312 [Ben Laurie (CHATS)]
5314 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5316 [Ben Laurie (CHATS)]
5318 *) Eliminate unused copy of key in RC4.
5319 [Ben Laurie (CHATS)]
5321 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5322 [Ben Laurie (CHATS)]
5324 *) Fix off-by-one error in EGD path.
5325 [Ben Laurie (CHATS)]
5327 *) If RANDFILE path is too long, ignore instead of truncating.
5328 [Ben Laurie (CHATS)]
5330 *) Eliminate unused and incorrectly sized X.509 structure
5332 [Ben Laurie (CHATS)]
5334 *) Eliminate unused and dangerous function knumber().
5335 [Ben Laurie (CHATS)]
5337 *) Eliminate unused and dangerous structure, KSSL_ERR.
5338 [Ben Laurie (CHATS)]
5340 *) Protect against overlong session ID context length in an encoded
5341 session object. Since these are local, this does not appear to be
5343 [Ben Laurie (CHATS)]
5345 *) Change from security patch (see 0.9.6e below) that did not affect
5346 the 0.9.6 release series:
5348 Remote buffer overflow in SSL3 protocol - an attacker could
5349 supply an oversized master key in Kerberos-enabled versions.
5351 [Ben Laurie (CHATS)]
5353 *) Change the SSL kerb5 codes to match RFC 2712.
5356 *) Make -nameopt work fully for req and add -reqopt switch.
5357 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5359 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5360 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5362 *) Make sure tests can be performed even if the corresponding algorithms
5363 have been removed entirely. This was also the last step to make
5364 OpenSSL compilable with DJGPP under all reasonable conditions.
5365 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5367 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5368 to allow version independent disabling of normally unselected ciphers,
5369 which may be activated as a side-effect of selecting a single cipher.
5371 (E.g., cipher list string "RSA" enables ciphersuites that are left
5372 out of "ALL" because they do not provide symmetric encryption.
5373 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5374 [Lutz Jaenicke, Bodo Moeller]
5376 *) Add appropriate support for separate platform-dependent build
5377 directories. The recommended way to make a platform-dependent
5378 build directory is the following (tested on Linux), maybe with
5381 # Place yourself outside of the OpenSSL source tree. In
5382 # this example, the environment variable OPENSSL_SOURCE
5383 # is assumed to contain the absolute OpenSSL source directory.
5384 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5385 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5386 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5387 mkdir -p `dirname $F`
5388 ln -s $OPENSSL_SOURCE/$F $F
5391 To be absolutely sure not to disturb the source tree, a "make clean"
5392 is a good thing. If it isn't successfull, don't worry about it,
5393 it probably means the source directory is very clean.
5396 *) Make sure any ENGINE control commands make local copies of string
5397 pointers passed to them whenever necessary. Otherwise it is possible
5398 the caller may have overwritten (or deallocated) the original string
5399 data when a later ENGINE operation tries to use the stored values.
5400 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5402 *) Improve diagnostics in file reading and command-line digests.
5403 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5405 *) Add AES modes CFB and OFB to the object database. Correct an
5406 error in AES-CFB decryption.
5409 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5410 allows existing EVP_CIPHER_CTX structures to be reused after
5411 calling EVP_*Final(). This behaviour is used by encryption
5412 BIOs and some applications. This has the side effect that
5413 applications must explicitly clean up cipher contexts with
5414 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5417 *) Check the values of dna and dnb in bn_mul_recursive before calling
5418 bn_mul_comba (a non zero value means the a or b arrays do not contain
5419 n2 elements) and fallback to bn_mul_normal if either is not zero.
5422 *) Fix escaping of non-ASCII characters when using the -subj option
5423 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5426 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5427 form for "surname", serialNumber has no short form.
5428 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5429 therefore remove "mail" short name for "internet 7".
5430 The OID for unique identifiers in X509 certificates is
5431 x500UniqueIdentifier, not uniqueIdentifier.
5432 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5435 *) Add an "init" command to the ENGINE config module and auto initialize
5436 ENGINEs. Without any "init" command the ENGINE will be initialized
5437 after all ctrl commands have been executed on it. If init=1 the
5438 ENGINE is initailized at that point (ctrls before that point are run
5439 on the uninitialized ENGINE and after on the initialized one). If
5440 init=0 then the ENGINE will not be iniatialized at all.
5443 *) Fix the 'app_verify_callback' interface so that the user-defined
5444 argument is actually passed to the callback: In the
5445 SSL_CTX_set_cert_verify_callback() prototype, the callback
5446 declaration has been changed from
5449 int (*cb)(X509_STORE_CTX *,void *);
5450 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5451 i=s->ctx->app_verify_callback(&ctx)
5452 has been changed into
5453 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5455 To update applications using SSL_CTX_set_cert_verify_callback(),
5456 a dummy argument can be added to their callback functions.
5457 [D. K. Smetters <smetters@parc.xerox.com>]
5459 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5460 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5462 *) Add and OPENSSL_LOAD_CONF define which will cause
5463 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5464 This allows older applications to transparently support certain
5465 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5466 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5467 load the config file and OPENSSL_add_all_algorithms_conf() which will
5468 always load it have also been added.
5471 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5472 Adjust NIDs and EVP layer.
5473 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5475 *) Config modules support in openssl utility.
5477 Most commands now load modules from the config file,
5478 though in a few (such as version) this isn't done
5479 because it couldn't be used for anything.
5481 In the case of ca and req the config file used is
5482 the same as the utility itself: that is the -config
5483 command line option can be used to specify an
5487 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5488 use "openssl_conf" if filename is NULL use default openssl config file.
5491 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5492 config section name. Add a new flag to tolerate a missing config file
5493 and move code to CONF_modules_load_file().
5496 *) Support for crypto accelerator cards from Accelerated Encryption
5497 Processing, www.aep.ie. (Use engine 'aep')
5498 The support was copied from 0.9.6c [engine] and adapted/corrected
5499 to work with the new engine framework.
5500 [AEP Inc. and Richard Levitte]
5502 *) Support for SureWare crypto accelerator cards from Baltimore
5503 Technologies. (Use engine 'sureware')
5504 The support was copied from 0.9.6c [engine] and adapted
5505 to work with the new engine framework.
5508 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5509 make the newer ENGINE framework commands for the CHIL engine work.
5510 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5512 *) Make it possible to produce shared libraries on ReliantUNIX.
5513 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5515 *) Add the configuration target debug-linux-ppro.
5516 Make 'openssl rsa' use the general key loading routines
5517 implemented in apps.c, and make those routines able to
5518 handle the key format FORMAT_NETSCAPE and the variant
5520 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5522 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5523 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5525 *) Add -keyform to rsautl, and document -engine.
5526 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5528 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5529 BIO_R_NO_SUCH_FILE error code rather than the generic
5530 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5533 *) Add new functions
5535 ERR_peek_last_error_line
5536 ERR_peek_last_error_line_data.
5537 These are similar to
5540 ERR_peek_error_line_data,
5541 but report on the latest error recorded rather than the first one
5542 still in the error queue.
5543 [Ben Laurie, Bodo Moeller]
5545 *) default_algorithms option in ENGINE config module. This allows things
5547 default_algorithms = ALL
5548 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5551 *) Prelminary ENGINE config module.
5554 *) New experimental application configuration code.
5557 *) Change the AES code to follow the same name structure as all other
5558 symmetric ciphers, and behave the same way. Move everything to
5559 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5560 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5562 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5563 [Ben Laurie and Theo de Raadt]
5565 *) Add option to output public keys in req command.
5566 [Massimiliano Pala madwolf@openca.org]
5568 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5569 (up to about 10% better than before for P-192 and P-224).
5572 *) New functions/macros
5574 SSL_CTX_set_msg_callback(ctx, cb)
5575 SSL_CTX_set_msg_callback_arg(ctx, arg)
5576 SSL_set_msg_callback(ssl, cb)
5577 SSL_set_msg_callback_arg(ssl, arg)
5579 to request calling a callback function
5581 void cb(int write_p, int version, int content_type,
5582 const void *buf, size_t len, SSL *ssl, void *arg)
5584 whenever a protocol message has been completely received
5585 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5586 protocol version according to which the SSL library interprets
5587 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5588 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5589 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5590 specification (change_cipher_spec(20), alert(21), handshake(22)).
5591 'buf' and 'len' point to the actual message, 'ssl' to the
5592 SSL object, and 'arg' is the application-defined value set by
5593 SSL[_CTX]_set_msg_callback_arg().
5595 'openssl s_client' and 'openssl s_server' have new '-msg' options
5596 to enable a callback that displays all protocol messages.
5599 *) Change the shared library support so shared libraries are built as
5600 soon as the corresponding static library is finished, and thereby get
5601 openssl and the test programs linked against the shared library.
5602 This still only happens when the keyword "shard" has been given to
5603 the configuration scripts.
5605 NOTE: shared library support is still an experimental thing, and
5606 backward binary compatibility is still not guaranteed.
5607 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5609 *) Add support for Subject Information Access extension.
5610 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5612 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5613 additional bytes when new memory had to be allocated, not just
5614 when reusing an existing buffer.
5617 *) New command line and configuration option 'utf8' for the req command.
5618 This allows field values to be specified as UTF8 strings.
5621 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5622 runs for the former and machine-readable output for the latter.
5625 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5626 of the e-mail address in the DN (i.e., it will go into a certificate
5627 extension only). The new configuration file option 'email_in_dn = no'
5628 has the same effect.
5629 [Massimiliano Pala madwolf@openca.org]
5631 *) Change all functions with names starting with des_ to be starting
5632 with DES_ instead. Add wrappers that are compatible with libdes,
5633 but are named _ossl_old_des_*. Finally, add macros that map the
5634 des_* symbols to the corresponding _ossl_old_des_* if libdes
5635 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5636 desired, the des_* symbols will be mapped to DES_*, with one
5639 Since we provide two compatibility mappings, the user needs to
5640 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5641 compatibility is desired. The default (i.e., when that macro
5642 isn't defined) is OpenSSL 0.9.6c compatibility.
5644 There are also macros that enable and disable the support of old
5645 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5646 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5647 are defined, the default will apply: to support the old des routines.
5649 In either case, one must include openssl/des.h to get the correct
5650 definitions. Do not try to just include openssl/des_old.h, that
5653 NOTE: This is a major break of an old API into a new one. Software
5654 authors are encouraged to switch to the DES_ style functions. Some
5655 time in the future, des_old.h and the libdes compatibility functions
5656 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5657 default), and then completely removed.
5660 *) Test for certificates which contain unsupported critical extensions.
5661 If such a certificate is found during a verify operation it is
5662 rejected by default: this behaviour can be overridden by either
5663 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5664 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5665 X509_supported_extension() has also been added which returns 1 if a
5666 particular extension is supported.
5669 *) Modify the behaviour of EVP cipher functions in similar way to digests
5670 to retain compatibility with existing code.
5673 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5674 compatibility with existing code. In particular the 'ctx' parameter does
5675 not have to be to be initialized before the call to EVP_DigestInit() and
5676 it is tidied up after a call to EVP_DigestFinal(). New function
5677 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5678 EVP_MD_CTX_copy() changed to not require the destination to be
5679 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5680 requires the destination to be valid.
5682 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5683 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5686 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5687 so that complete 'Handshake' protocol structures are kept in memory
5688 instead of overwriting 'msg_type' and 'length' with 'body' data.
5691 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5692 [Massimo Santin via Richard Levitte]
5694 *) Major restructuring to the underlying ENGINE code. This includes
5695 reduction of linker bloat, separation of pure "ENGINE" manipulation
5696 (initialisation, etc) from functionality dealing with implementations
5697 of specific crypto iterfaces. This change also introduces integrated
5698 support for symmetric ciphers and digest implementations - so ENGINEs
5699 can now accelerate these by providing EVP_CIPHER and EVP_MD
5700 implementations of their own. This is detailed in crypto/engine/README
5701 as it couldn't be adequately described here. However, there are a few
5702 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5703 were changed in the original introduction of ENGINE code have now
5704 reverted back - the hooking from this code to ENGINE is now a good
5705 deal more passive and at run-time, operations deal directly with
5706 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5707 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5708 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5709 they were not being used by the framework as there is no concept of a
5710 BIGNUM_METHOD and they could not be generalised to the new
5711 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5712 ENGINE_cpy() has been removed as it cannot be consistently defined in
5716 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5719 *) Change mkdef.pl to sort symbols that get the same entry number,
5720 and make sure the automatically generated functions ERR_load_*
5721 become part of libeay.num as well.
5724 *) New function SSL_renegotiate_pending(). This returns true once
5725 renegotiation has been requested (either SSL_renegotiate() call
5726 or HelloRequest/ClientHello receveived from the peer) and becomes
5727 false once a handshake has been completed.
5728 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5729 sends a HelloRequest, but does not ensure that a handshake takes
5730 place. SSL_renegotiate_pending() is useful for checking if the
5731 client has followed the request.)
5734 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5735 By default, clients may request session resumption even during
5736 renegotiation (if session ID contexts permit); with this option,
5737 session resumption is possible only in the first handshake.
5739 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5740 more bits available for options that should not be part of
5741 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5744 *) Add some demos for certificate and certificate request creation.
5747 *) Make maximum certificate chain size accepted from the peer application
5748 settable (SSL*_get/set_max_cert_list()), as proposed by
5749 "Douglas E. Engert" <deengert@anl.gov>.
5752 *) Add support for shared libraries for Unixware-7
5753 (Boyd Lynn Gerber <gerberb@zenez.com>).
5756 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5757 be done prior to destruction. Use this to unload error strings from
5758 ENGINEs that load their own error strings. NB: This adds two new API
5759 functions to "get" and "set" this destroy handler in an ENGINE.
5762 *) Alter all existing ENGINE implementations (except "openssl" and
5763 "openbsd") to dynamically instantiate their own error strings. This
5764 makes them more flexible to be built both as statically-linked ENGINEs
5765 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5766 Also, add stub code to each that makes building them as self-contained
5767 shared-libraries easier (see README.ENGINE).
5770 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5771 implementations into applications that are completely implemented in
5772 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5773 commands that can be used to configure what shared-library to load and
5774 to control aspects of the way it is handled. Also, made an update to
5775 the README.ENGINE file that brings its information up-to-date and
5776 provides some information and instructions on the "dynamic" ENGINE
5777 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5780 *) Make it possible to unload ranges of ERR strings with a new
5781 "ERR_unload_strings" function.
5784 *) Add a copy() function to EVP_MD.
5787 *) Make EVP_MD routines take a context pointer instead of just the
5788 md_data void pointer.
5791 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5792 that the digest can only process a single chunk of data
5793 (typically because it is provided by a piece of
5794 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5795 is only going to provide a single chunk of data, and hence the
5796 framework needn't accumulate the data for oneshot drivers.
5799 *) As with "ERR", make it possible to replace the underlying "ex_data"
5800 functions. This change also alters the storage and management of global
5801 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5802 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5803 index counters. The API functions that use this state have been changed
5804 to take a "class_index" rather than pointers to the class's local STACK
5805 and counter, and there is now an API function to dynamically create new
5806 classes. This centralisation allows us to (a) plug a lot of the
5807 thread-safety problems that existed, and (b) makes it possible to clean
5808 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5809 such data would previously have always leaked in application code and
5810 workarounds were in place to make the memory debugging turn a blind eye
5811 to it. Application code that doesn't use this new function will still
5812 leak as before, but their memory debugging output will announce it now
5813 rather than letting it slide.
5815 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5816 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5817 has a return value to indicate success or failure.
5820 *) Make it possible to replace the underlying "ERR" functions such that the
5821 global state (2 LHASH tables and 2 locks) is only used by the "default"
5822 implementation. This change also adds two functions to "get" and "set"
5823 the implementation prior to it being automatically set the first time
5824 any other ERR function takes place. Ie. an application can call "get",
5825 pass the return value to a module it has just loaded, and that module
5826 can call its own "set" function using that value. This means the
5827 module's "ERR" operations will use (and modify) the error state in the
5828 application and not in its own statically linked copy of OpenSSL code.
5831 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5832 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5833 the operation, and provides a more encapsulated way for external code
5834 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5835 to use these functions rather than manually incrementing the counts.
5837 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5840 *) Add EVP test program.
5843 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5846 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5847 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5848 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5849 These allow a CRL to be built without having to access X509_CRL fields
5850 directly. Modify 'ca' application to use new functions.
5853 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5854 bug workarounds. Rollback attack detection is a security feature.
5855 The problem will only arise on OpenSSL servers when TLSv1 is not
5856 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5857 Software authors not wanting to support TLSv1 will have special reasons
5858 for their choice and can explicitly enable this option.
5859 [Bodo Moeller, Lutz Jaenicke]
5861 *) Rationalise EVP so it can be extended: don't include a union of
5862 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5863 (similar to those existing for EVP_CIPHER_CTX).
5868 EVP_MD_CTX_init(&md); /* new function call */
5869 EVP_DigestInit(&md, EVP_sha1());
5870 EVP_DigestUpdate(&md, in, len);
5871 EVP_DigestFinal(&md, out, NULL);
5872 EVP_MD_CTX_cleanup(&md); /* new function call */
5876 *) Make DES key schedule conform to the usual scheme, as well as
5877 correcting its structure. This means that calls to DES functions
5878 now have to pass a pointer to a des_key_schedule instead of a
5879 plain des_key_schedule (which was actually always a pointer
5882 des_key_schedule ks;
5884 des_set_key_checked(..., &ks);
5885 des_ncbc_encrypt(..., &ks, ...);
5887 (Note that a later change renames 'des_...' into 'DES_...'.)
5890 *) Initial reduction of linker bloat: the use of some functions, such as
5891 PEM causes large amounts of unused functions to be linked in due to
5892 poor organisation. For example pem_all.c contains every PEM function
5893 which has a knock on effect of linking in large amounts of (unused)
5894 ASN1 code. Grouping together similar functions and splitting unrelated
5895 functions prevents this.
5898 *) Cleanup of EVP macros.
5901 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5902 correct _ecb suffix.
5905 *) Add initial OCSP responder support to ocsp application. The
5906 revocation information is handled using the text based index
5907 use by the ca application. The responder can either handle
5908 requests generated internally, supplied in files (for example
5909 via a CGI script) or using an internal minimal server.
5912 *) Add configuration choices to get zlib compression for TLS.
5915 *) Changes to Kerberos SSL for RFC 2712 compliance:
5916 1. Implemented real KerberosWrapper, instead of just using
5917 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5918 2. Implemented optional authenticator field of KerberosWrapper.
5920 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5921 and authenticator structs; see crypto/krb5/.
5923 Generalized Kerberos calls to support multiple Kerberos libraries.
5924 [Vern Staats <staatsvr@asc.hpc.mil>,
5925 Jeffrey Altman <jaltman@columbia.edu>
5926 via Richard Levitte]
5928 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5929 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5930 values for each of the key sizes rather than having just
5931 parameters (and 'speed' generating keys each time).
5934 *) Speed up EVP routines.
5937 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5938 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5939 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5940 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5942 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5943 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5944 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5947 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5949 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5952 *) Added the OS2-EMX target.
5953 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5955 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5956 to support NCONF routines in extension code. New function CONF_set_nconf()
5957 to allow functions which take an NCONF to also handle the old LHASH
5958 structure: this means that the old CONF compatible routines can be
5959 retained (in particular wrt extensions) without having to duplicate the
5960 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5963 *) Enhance the general user interface with mechanisms for inner control
5964 and with possibilities to have yes/no kind of prompts.
5967 *) Change all calls to low level digest routines in the library and
5968 applications to use EVP. Add missing calls to HMAC_cleanup() and
5969 don't assume HMAC_CTX can be copied using memcpy().
5970 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5972 *) Add the possibility to control engines through control names but with
5973 arbitrary arguments instead of just a string.
5974 Change the key loaders to take a UI_METHOD instead of a callback
5975 function pointer. NOTE: this breaks binary compatibility with earlier
5976 versions of OpenSSL [engine].
5977 Adapt the nCipher code for these new conditions and add a card insertion
5981 *) Enhance the general user interface with mechanisms to better support
5982 dialog box interfaces, application-defined prompts, the possibility
5983 to use defaults (for example default passwords from somewhere else)
5984 and interrupts/cancellations.
5987 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5988 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5991 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5992 tidy up some unnecessarily weird code in 'sk_new()').
5993 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5995 *) Change the key loading routines for ENGINEs to use the same kind
5996 callback (pem_password_cb) as all other routines that need this
6000 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
6001 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
6002 than this minimum value is recommended.
6005 *) New random seeder for OpenVMS, using the system process statistics
6006 that are easily reachable.
6009 *) Windows apparently can't transparently handle global
6010 variables defined in DLLs. Initialisations such as:
6012 const ASN1_ITEM *it = &ASN1_INTEGER_it;
6014 wont compile. This is used by the any applications that need to
6015 declare their own ASN1 modules. This was fixed by adding the option
6016 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
6017 needed for static libraries under Win32.
6020 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
6021 setting of purpose and trust fields. New X509_STORE trust and
6022 purpose functions and tidy up setting in other SSL functions.
6025 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
6026 structure. These are inherited by X509_STORE_CTX when it is
6027 initialised. This allows various defaults to be set in the
6028 X509_STORE structure (such as flags for CRL checking and custom
6029 purpose or trust settings) for functions which only use X509_STORE_CTX
6030 internally such as S/MIME.
6032 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
6033 trust settings if they are not set in X509_STORE. This allows X509_STORE
6034 purposes and trust (in S/MIME for example) to override any set by default.
6036 Add command line options for CRL checking to smime, s_client and s_server
6040 *) Initial CRL based revocation checking. If the CRL checking flag(s)
6041 are set then the CRL is looked up in the X509_STORE structure and
6042 its validity and signature checked, then if the certificate is found
6043 in the CRL the verify fails with a revoked error.
6045 Various new CRL related callbacks added to X509_STORE_CTX structure.
6047 Command line options added to 'verify' application to support this.
6049 This needs some additional work, such as being able to handle multiple
6050 CRLs with different times, extension based lookup (rather than just
6051 by subject name) and ultimately more complete V2 CRL extension
6055 *) Add a general user interface API (crypto/ui/). This is designed
6056 to replace things like des_read_password and friends (backward
6057 compatibility functions using this new API are provided).
6058 The purpose is to remove prompting functions from the DES code
6059 section as well as provide for prompting through dialog boxes in
6060 a window system and the like.
6063 *) Add "ex_data" support to ENGINE so implementations can add state at a
6064 per-structure level rather than having to store it globally.
6067 *) Make it possible for ENGINE structures to be copied when retrieved by
6068 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
6069 This causes the "original" ENGINE structure to act like a template,
6070 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
6071 operational state can be localised to each ENGINE structure, despite the
6072 fact they all share the same "methods". New ENGINE structures returned in
6073 this case have no functional references and the return value is the single
6074 structural reference. This matches the single structural reference returned
6075 by ENGINE_by_id() normally, when it is incremented on the pre-existing
6079 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
6080 needs to match any other type at all we need to manually clear the
6084 *) Changes to the "openssl engine" utility to include;
6085 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6086 about an ENGINE's available control commands.
6087 - executing control commands from command line arguments using the
6088 '-pre' and '-post' switches. '-post' is only used if '-t' is
6089 specified and the ENGINE is successfully initialised. The syntax for
6090 the individual commands are colon-separated, for example;
6091 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6094 *) New dynamic control command support for ENGINEs. ENGINEs can now
6095 declare their own commands (numbers), names (strings), descriptions,
6096 and input types for run-time discovery by calling applications. A
6097 subset of these commands are implicitly classed as "executable"
6098 depending on their input type, and only these can be invoked through
6099 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6100 can be based on user input, config files, etc). The distinction is
6101 that "executable" commands cannot return anything other than a boolean
6102 result and can only support numeric or string input, whereas some
6103 discoverable commands may only be for direct use through
6104 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6105 pointers, or other custom uses. The "executable" commands are to
6106 support parameterisations of ENGINE behaviour that can be
6107 unambiguously defined by ENGINEs and used consistently across any
6108 OpenSSL-based application. Commands have been added to all the
6109 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6110 control over shared-library paths without source code alterations.
6113 *) Changed all ENGINE implementations to dynamically allocate their
6114 ENGINEs rather than declaring them statically. Apart from this being
6115 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6116 this also allows the implementations to compile without using the
6117 internal engine_int.h header.
6120 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6121 'const' value. Any code that should be able to modify a RAND_METHOD
6122 should already have non-const pointers to it (ie. they should only
6123 modify their own ones).
6126 *) Made a variety of little tweaks to the ENGINE code.
6127 - "atalla" and "ubsec" string definitions were moved from header files
6128 to C code. "nuron" string definitions were placed in variables
6129 rather than hard-coded - allowing parameterisation of these values
6130 later on via ctrl() commands.
6131 - Removed unused "#if 0"'d code.
6132 - Fixed engine list iteration code so it uses ENGINE_free() to release
6133 structural references.
6134 - Constified the RAND_METHOD element of ENGINE structures.
6135 - Constified various get/set functions as appropriate and added
6136 missing functions (including a catch-all ENGINE_cpy that duplicates
6137 all ENGINE values onto a new ENGINE except reference counts/state).
6138 - Removed NULL parameter checks in get/set functions. Setting a method
6139 or function to NULL is a way of cancelling out a previously set
6140 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6141 and doesn't justify the extra error symbols and code.
6142 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6143 flags from engine_int.h to engine.h.
6144 - Changed prototypes for ENGINE handler functions (init(), finish(),
6145 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6148 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6149 to the algorithm using long division. The binary algorithm can be
6150 used only if the modulus is odd. On 32-bit systems, it is faster
6151 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6152 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6153 up to 450 bits. In 64-bit environments, the binary algorithm
6154 appears to be advantageous for much longer moduli; here we use it
6155 for moduli up to 2048 bits.
6158 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6159 could not support the combine flag in choice fields.
6162 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6163 extensions from a certificate request to the certificate.
6166 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6167 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6168 file: this allows the display of the certificate about to be
6169 signed to be customised, to allow certain fields to be included
6170 or excluded and extension details. The old system didn't display
6171 multicharacter strings properly, omitted fields not in the policy
6172 and couldn't display additional details such as extensions.
6175 *) Function EC_POINTs_mul for multiple scalar multiplication
6176 of an arbitrary number of elliptic curve points
6177 \sum scalars[i]*points[i],
6178 optionally including the generator defined for the EC_GROUP:
6179 scalar*generator + \sum scalars[i]*points[i].
6181 EC_POINT_mul is a simple wrapper function for the typical case
6182 that the point list has just one item (besides the optional
6186 *) First EC_METHODs for curves over GF(p):
6188 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6189 operations and provides various method functions that can also
6190 operate with faster implementations of modular arithmetic.
6192 EC_GFp_mont_method() reuses most functions that are part of
6193 EC_GFp_simple_method, but uses Montgomery arithmetic.
6195 [Bodo Moeller; point addition and point doubling
6196 implementation directly derived from source code provided by
6197 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6199 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6200 crypto/ec/ec_lib.c):
6202 Curves are EC_GROUP objects (with an optional group generator)
6203 based on EC_METHODs that are built into the library.
6205 Points are EC_POINT objects based on EC_GROUP objects.
6207 Most of the framework would be able to handle curves over arbitrary
6208 finite fields, but as there are no obvious types for fields other
6209 than GF(p), some functions are limited to that for now.
6212 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6213 that the file contains a complete HTTP response.
6216 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6217 change the def and num file printf format specifier from "%-40sXXX"
6218 to "%-39s XXX". The latter will always guarantee a space after the
6219 field while the former will cause them to run together if the field
6220 is 40 of more characters long.
6223 *) Constify the cipher and digest 'method' functions and structures
6224 and modify related functions to take constant EVP_MD and EVP_CIPHER
6228 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6229 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6232 *) Modify EVP_Digest*() routines so they now return values. Although the
6233 internal software routines can never fail additional hardware versions
6237 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6239 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6240 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6243 ERR_R_NESTED_ASN1_ERROR
6245 ERR_R_MISSING_ASN1_EOS
6246 were 4 .. 9, conflicting with
6247 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6249 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6250 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6252 Add new error code 'ERR_R_INTERNAL_ERROR'.
6255 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6259 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6260 sets the subject name for a new request or supersedes the
6261 subject name in a given request. Formats that can be parsed are
6262 'CN=Some Name, OU=myOU, C=IT'
6264 'CN=Some Name/OU=myOU/C=IT'.
6266 Add options '-batch' and '-verbose' to 'openssl req'.
6267 [Massimiliano Pala <madwolf@hackmasters.net>]
6269 *) Introduce the possibility to access global variables through
6270 functions on platform were that's the best way to handle exporting
6271 global variables in shared libraries. To enable this functionality,
6272 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6273 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6274 is normally done by Configure or something similar).
6276 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6277 in the source file (foo.c) like this:
6279 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6280 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6282 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6283 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6285 OPENSSL_DECLARE_GLOBAL(int,foo);
6286 #define foo OPENSSL_GLOBAL_REF(foo)
6287 OPENSSL_DECLARE_GLOBAL(double,bar);
6288 #define bar OPENSSL_GLOBAL_REF(bar)
6290 The #defines are very important, and therefore so is including the
6291 header file everywhere where the defined globals are used.
6293 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6294 of ASN.1 items, but that structure is a bit different.
6296 The largest change is in util/mkdef.pl which has been enhanced with
6297 better and easier to understand logic to choose which symbols should
6298 go into the Windows .def files as well as a number of fixes and code
6299 cleanup (among others, algorithm keywords are now sorted
6300 lexicographically to avoid constant rewrites).
6303 *) In BN_div() keep a copy of the sign of 'num' before writing the
6304 result to 'rm' because if rm==num the value will be overwritten
6305 and produce the wrong result if 'num' is negative: this caused
6306 problems with BN_mod() and BN_nnmod().
6309 *) Function OCSP_request_verify(). This checks the signature on an
6310 OCSP request and verifies the signer certificate. The signer
6311 certificate is just checked for a generic purpose and OCSP request
6315 *) Add OCSP_check_validity() function to check the validity of OCSP
6316 responses. OCSP responses are prepared in real time and may only
6317 be a few seconds old. Simply checking that the current time lies
6318 between thisUpdate and nextUpdate max reject otherwise valid responses
6319 caused by either OCSP responder or client clock inaccuracy. Instead
6320 we allow thisUpdate and nextUpdate to fall within a certain period of
6321 the current time. The age of the response can also optionally be
6322 checked. Two new options -validity_period and -status_age added to
6326 *) If signature or public key algorithm is unrecognized print out its
6327 OID rather that just UNKNOWN.
6330 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6331 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6332 ID to be generated from the issuer certificate alone which can then be
6333 passed to OCSP_id_issuer_cmp().
6336 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6337 ASN1 modules to export functions returning ASN1_ITEM pointers
6338 instead of the ASN1_ITEM structures themselves. This adds several
6339 new macros which allow the underlying ASN1 function/structure to
6340 be accessed transparently. As a result code should not use ASN1_ITEM
6341 references directly (such as &X509_it) but instead use the relevant
6342 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6343 use of the new ASN1 code on platforms where exporting structures
6344 is problematical (for example in shared libraries) but exporting
6345 functions returning pointers to structures is not.
6348 *) Add support for overriding the generation of SSL/TLS session IDs.
6349 These callbacks can be registered either in an SSL_CTX or per SSL.
6350 The purpose of this is to allow applications to control, if they wish,
6351 the arbitrary values chosen for use as session IDs, particularly as it
6352 can be useful for session caching in multiple-server environments. A
6353 command-line switch for testing this (and any client code that wishes
6354 to use such a feature) has been added to "s_server".
6355 [Geoff Thorpe, Lutz Jaenicke]
6357 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6358 of the form '#if defined(...) || defined(...) || ...' and
6359 '#if !defined(...) && !defined(...) && ...'. This also avoids
6360 the growing number of special cases it was previously handling.
6363 *) Make all configuration macros available for application by making
6364 sure they are available in opensslconf.h, by giving them names starting
6365 with "OPENSSL_" to avoid conflicts with other packages and by making
6366 sure e_os2.h will cover all platform-specific cases together with
6368 Additionally, it is now possible to define configuration/platform-
6369 specific names (called "system identities"). In the C code, these
6370 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6371 macro with the name beginning with "OPENSSL_SYS_", which is determined
6372 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6376 *) New option -set_serial to 'req' and 'x509' this allows the serial
6377 number to use to be specified on the command line. Previously self
6378 signed certificates were hard coded with serial number 0 and the
6379 CA options of 'x509' had to use a serial number in a file which was
6383 *) New options to 'ca' utility to support V2 CRL entry extensions.
6384 Currently CRL reason, invalidity date and hold instruction are
6385 supported. Add new CRL extensions to V3 code and some new objects.
6388 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6389 disable standard block padding (aka PKCS#5 padding) in the EVP
6390 API, which was previously mandatory. This means that the data is
6391 not padded in any way and so the total length much be a multiple
6392 of the block size, otherwise an error occurs.
6395 *) Initial (incomplete) OCSP SSL support.
6398 *) New function OCSP_parse_url(). This splits up a URL into its host,
6399 port and path components: primarily to parse OCSP URLs. New -url
6400 option to ocsp utility.
6403 *) New nonce behavior. The return value of OCSP_check_nonce() now
6404 reflects the various checks performed. Applications can decide
6405 whether to tolerate certain situations such as an absent nonce
6406 in a response when one was present in a request: the ocsp application
6407 just prints out a warning. New function OCSP_add1_basic_nonce()
6408 this is to allow responders to include a nonce in a response even if
6409 the request is nonce-less.
6412 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6413 skipped when using openssl x509 multiple times on a single input file,
6414 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6417 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6418 set string type: to handle setting ASN1_TIME structures. Fix ca
6419 utility to correctly initialize revocation date of CRLs.
6422 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6423 the clients preferred ciphersuites and rather use its own preferences.
6424 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6425 Internet Explorer by ensuring unchanged hash method during stepup.
6426 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6429 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6430 to aes and add a new 'exist' option to print out symbols that don't
6434 *) Additional options to ocsp utility to allow flags to be set and
6435 additional certificates supplied.
6438 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6439 OCSP client a number of certificate to only verify the response
6443 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6444 handle the new API. Currently only ECB, CBC modes supported. Add new
6447 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6448 Encryption Standard (AES) Ciphersuites for Transport Layer
6449 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6450 not enabled by default and were not part of the "ALL" ciphersuite
6451 alias because they were not yet official; they could be
6452 explicitly requested by specifying the "AESdraft" ciphersuite
6453 group alias. In the final release of OpenSSL 0.9.7, the group
6454 alias is called "AES" and is part of "ALL".)
6455 [Ben Laurie, Steve Henson, Bodo Moeller]
6457 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6458 request to response.
6461 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6462 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6463 extract information from a certificate request. OCSP_response_create()
6464 creates a response and optionally adds a basic response structure.
6465 OCSP_basic_add1_status() adds a complete single response to a basic
6466 response and returns the OCSP_SINGLERESP structure just added (to allow
6467 extensions to be included for example). OCSP_basic_add1_cert() adds a
6468 certificate to a basic response and OCSP_basic_sign() signs a basic
6469 response with various flags. New helper functions ASN1_TIME_check()
6470 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6471 (converts ASN1_TIME to GeneralizedTime).
6474 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6475 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6476 structure from a certificate. X509_pubkey_digest() digests the public_key
6477 contents: this is used in various key identifiers.
6480 *) Make sk_sort() tolerate a NULL argument.
6481 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6483 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6484 passed by the function are trusted implicitly. If any of them signed the
6485 response then it is assumed to be valid and is not verified.
6488 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6489 to data. This was previously part of the PKCS7 ASN1 code. This
6490 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6491 [Steve Henson, reported by Kenneth R. Robinette
6492 <support@securenetterm.com>]
6494 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6495 routines: without these tracing memory leaks is very painful.
6496 Fix leaks in PKCS12 and PKCS7 routines.
6499 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6500 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6501 effectively meant GeneralizedTime would never be used. Now it
6502 is initialised to -1 but X509_time_adj() now has to check the value
6503 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6504 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6505 [Steve Henson, reported by Kenneth R. Robinette
6506 <support@securenetterm.com>]
6508 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6509 result in a zero length in the ASN1_INTEGER structure which was
6510 not consistent with the structure when d2i_ASN1_INTEGER() was used
6511 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6512 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6513 where it did not print out a minus for negative ASN1_INTEGER.
6516 *) Add summary printout to ocsp utility. The various functions which
6517 convert status values to strings have been renamed to:
6518 OCSP_response_status_str(), OCSP_cert_status_str() and
6519 OCSP_crl_reason_str() and are no longer static. New options
6520 to verify nonce values and to disable verification. OCSP response
6521 printout format cleaned up.
6524 *) Add additional OCSP certificate checks. These are those specified
6525 in RFC2560. This consists of two separate checks: the CA of the
6526 certificate being checked must either be the OCSP signer certificate
6527 or the issuer of the OCSP signer certificate. In the latter case the
6528 OCSP signer certificate must contain the OCSP signing extended key
6529 usage. This check is performed by attempting to match the OCSP
6530 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6531 in the OCSP_CERTID structures of the response.
6534 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6535 and related routines. This uses the standard OpenSSL certificate
6536 verify routines to perform initial checks (just CA validity) and
6537 to obtain the certificate chain. Then additional checks will be
6538 performed on the chain. Currently the root CA is checked to see
6539 if it is explicitly trusted for OCSP signing. This is used to set
6540 a root CA as a global signing root: that is any certificate that
6541 chains to that CA is an acceptable OCSP signing certificate.
6544 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6545 extensions from a separate configuration file.
6546 As when reading extensions from the main configuration file,
6547 the '-extensions ...' option may be used for specifying the
6549 [Massimiliano Pala <madwolf@comune.modena.it>]
6551 *) New OCSP utility. Allows OCSP requests to be generated or
6552 read. The request can be sent to a responder and the output
6553 parsed, outputed or printed in text form. Not complete yet:
6554 still needs to check the OCSP response validity.
6557 *) New subcommands for 'openssl ca':
6558 'openssl ca -status <serial>' prints the status of the cert with
6559 the given serial number (according to the index file).
6560 'openssl ca -updatedb' updates the expiry status of certificates
6562 [Massimiliano Pala <madwolf@comune.modena.it>]
6564 *) New '-newreq-nodes' command option to CA.pl. This is like
6565 '-newreq', but calls 'openssl req' with the '-nodes' option
6566 so that the resulting key is not encrypted.
6567 [Damien Miller <djm@mindrot.org>]
6569 *) New configuration for the GNU Hurd.
6570 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6572 *) Initial code to implement OCSP basic response verify. This
6573 is currently incomplete. Currently just finds the signer's
6574 certificate and verifies the signature on the response.
6577 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6578 value of OPENSSLDIR. This is available via the new '-d' option
6579 to 'openssl version', and is also included in 'openssl version -a'.
6582 *) Allowing defining memory allocation callbacks that will be given
6583 file name and line number information in additional arguments
6584 (a const char* and an int). The basic functionality remains, as
6585 well as the original possibility to just replace malloc(),
6586 realloc() and free() by functions that do not know about these
6587 additional arguments. To register and find out the current
6588 settings for extended allocation functions, the following
6589 functions are provided:
6591 CRYPTO_set_mem_ex_functions
6592 CRYPTO_set_locked_mem_ex_functions
6593 CRYPTO_get_mem_ex_functions
6594 CRYPTO_get_locked_mem_ex_functions
6596 These work the same way as CRYPTO_set_mem_functions and friends.
6597 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6598 extended allocation function is enabled.
6599 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6600 a conventional allocation function is enabled.
6601 [Richard Levitte, Bodo Moeller]
6603 *) Finish off removing the remaining LHASH function pointer casts.
6604 There should no longer be any prototype-casting required when using
6605 the LHASH abstraction, and any casts that remain are "bugs". See
6606 the callback types and macros at the head of lhash.h for details
6607 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6610 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6611 If /dev/[u]random devices are not available or do not return enough
6612 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6614 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6615 /etc/entropy will be queried once each in this sequence, quering stops
6616 when enough entropy was collected without querying more sockets.
6619 *) Change the Unix RAND_poll() variant to be able to poll several
6620 random devices, as specified by DEVRANDOM, until a sufficient amount
6621 of data has been collected. We spend at most 10 ms on each file
6622 (select timeout) and read in non-blocking mode. DEVRANDOM now
6623 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6624 (previously it was just the string "/dev/urandom"), so on typical
6625 platforms the 10 ms delay will never occur.
6626 Also separate out the Unix variant to its own file, rand_unix.c.
6627 For VMS, there's a currently-empty rand_vms.c.
6630 *) Move OCSP client related routines to ocsp_cl.c. These
6631 provide utility functions which an application needing
6632 to issue a request to an OCSP responder and analyse the
6633 response will typically need: as opposed to those which an
6634 OCSP responder itself would need which will be added later.
6636 OCSP_request_sign() signs an OCSP request with an API similar
6637 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6638 response. OCSP_response_get1_basic() extracts basic response
6639 from response. OCSP_resp_find_status(): finds and extracts status
6640 information from an OCSP_CERTID structure (which will be created
6641 when the request structure is built). These are built from lower
6642 level functions which work on OCSP_SINGLERESP structures but
6643 wont normally be used unless the application wishes to examine
6644 extensions in the OCSP response for example.
6646 Replace nonce routines with a pair of functions.
6647 OCSP_request_add1_nonce() adds a nonce value and optionally
6648 generates a random value. OCSP_check_nonce() checks the
6649 validity of the nonce in an OCSP response.
6652 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6653 This doesn't copy the supplied OCSP_CERTID and avoids the
6654 need to free up the newly created id. Change return type
6655 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6656 This can then be used to add extensions to the request.
6657 Deleted OCSP_request_new(), since most of its functionality
6658 is now in OCSP_REQUEST_new() (and the case insensitive name
6659 clash) apart from the ability to set the request name which
6660 will be added elsewhere.
6663 *) Update OCSP API. Remove obsolete extensions argument from
6664 various functions. Extensions are now handled using the new
6665 OCSP extension code. New simple OCSP HTTP function which
6666 can be used to send requests and parse the response.
6669 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6670 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6671 uses the special reorder version of SET OF to sort the attributes
6672 and reorder them to match the encoded order. This resolves a long
6673 standing problem: a verify on a PKCS7 structure just after signing
6674 it used to fail because the attribute order did not match the
6675 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6676 it uses the received order. This is necessary to tolerate some broken
6677 software that does not order SET OF. This is handled by encoding
6678 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6679 to produce the required SET OF.
6682 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6683 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6684 files to get correct declarations of the ASN.1 item variables.
6687 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6688 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6689 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6690 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6691 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6692 ASN1_ITEM and no wrapper functions.
6695 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6696 replace the old function pointer based I/O routines. Change most of
6697 the *_d2i_bio() and *_d2i_fp() functions to use these.
6700 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6701 lines, recognice more "algorithms" that can be deselected, and make
6702 it complain about algorithm deselection that isn't recognised.
6705 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6706 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6707 to use new functions. Add NO_ASN1_OLD which can be set to remove
6708 some old style ASN1 functions: this can be used to determine if old
6709 code will still work when these eventually go away.
6712 *) New extension functions for OCSP structures, these follow the
6713 same conventions as certificates and CRLs.
6716 *) New function X509V3_add1_i2d(). This automatically encodes and
6717 adds an extension. Its behaviour can be customised with various
6718 flags to append, replace or delete. Various wrappers added for
6719 certifcates and CRLs.
6722 *) Fix to avoid calling the underlying ASN1 print routine when
6723 an extension cannot be parsed. Correct a typo in the
6724 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6727 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
6728 entries for variables.
6731 *) Add functionality to apps/openssl.c for detecting locking
6732 problems: As the program is single-threaded, all we have
6733 to do is register a locking callback using an array for
6734 storing which locks are currently held by the program.
6737 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6738 SSL_get_ex_data_X509_STORE_idx(), which is used in
6739 ssl_verify_cert_chain() and thus can be called at any time
6740 during TLS/SSL handshakes so that thread-safety is essential.
6741 Unfortunately, the ex_data design is not at all suited
6742 for multi-threaded use, so it probably should be abolished.
6745 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6746 [Broadcom, tweaked and integrated by Geoff Thorpe]
6748 *) Move common extension printing code to new function
6749 X509V3_print_extensions(). Reorganise OCSP print routines and
6750 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6753 *) New function X509_signature_print() to remove duplication in some
6757 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6758 set (this was treated exactly the same as SET OF previously). This
6759 is used to reorder the STACK representing the structure to match the
6760 encoding. This will be used to get round a problem where a PKCS7
6761 structure which was signed could not be verified because the STACK
6762 order did not reflect the encoded order.
6765 *) Reimplement the OCSP ASN1 module using the new code.
6768 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6769 for its ASN1 operations. The old style function pointers still exist
6770 for now but they will eventually go away.
6773 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6774 completely replaces the old ASN1 functionality with a table driven
6775 encoder and decoder which interprets an ASN1_ITEM structure describing
6776 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6777 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6778 has also been converted to the new form.
6781 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6782 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6783 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6784 for negative moduli.
6787 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6788 of not touching the result's sign bit.
6791 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6795 *) Changed the LHASH code to use prototypes for callbacks, and created
6796 macros to declare and implement thin (optionally static) functions
6797 that provide type-safety and avoid function pointer casting for the
6798 type-specific callbacks.
6801 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6803 [Veers Staats <staatsvr@asc.hpc.mil>,
6804 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6806 *) Reformat the FAQ so the different questions and answers can be divided
6807 in sections depending on the subject.
6810 *) Have the zlib compression code load ZLIB.DLL dynamically under
6814 *) New function BN_mod_sqrt for computing square roots modulo a prime
6815 (using the probabilistic Tonelli-Shanks algorithm unless
6816 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6817 be handled deterministically).
6818 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6820 *) Make BN_mod_inverse faster by explicitly handling small quotients
6821 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6822 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6825 *) New function BN_kronecker.
6828 *) Fix BN_gcd so that it works on negative inputs; the result is
6829 positive unless both parameters are zero.
6830 Previously something reasonably close to an infinite loop was
6831 possible because numbers could be growing instead of shrinking
6832 in the implementation of Euclid's algorithm.
6835 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6836 sign of the number in question.
6838 Fix BN_is_word(a,w) to work correctly for w == 0.
6840 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6841 because its test if the absolute value of 'a' equals 'w'.
6842 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6843 it exists mostly for use in the implementations of BN_is_zero(),
6844 BN_is_one(), and BN_is_word().
6847 *) New function BN_swap.
6850 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6851 the exponentiation functions are more likely to produce reasonable
6852 results on negative inputs.
6855 *) Change BN_mod_mul so that the result is always non-negative.
6856 Previously, it could be negative if one of the factors was negative;
6857 I don't think anyone really wanted that behaviour.
6860 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6861 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6862 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6863 and add new functions:
6872 BN_mod_lshift1_quick
6876 These functions always generate non-negative results.
6878 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6879 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6881 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6882 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6883 be reduced modulo m.
6884 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6887 The following entry accidentily appeared in the CHANGES file
6888 distributed with OpenSSL 0.9.7. The modifications described in
6889 it do *not* apply to OpenSSL 0.9.7.
6891 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6892 was actually never needed) and in BN_mul(). The removal in BN_mul()
6893 required a small change in bn_mul_part_recursive() and the addition
6894 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6895 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6896 bn_sub_words() and bn_add_words() except they take arrays with
6901 *) In 'openssl passwd', verify passwords read from the terminal
6902 unless the '-salt' option is used (which usually means that
6903 verification would just waste user's time since the resulting
6904 hash is going to be compared with some given password hash)
6905 or the new '-noverify' option is used.
6907 This is an incompatible change, but it does not affect
6908 non-interactive use of 'openssl passwd' (passwords on the command
6909 line, '-stdin' option, '-in ...' option) and thus should not
6913 *) Remove all references to RSAref, since there's no more need for it.
6916 *) Make DSO load along a path given through an environment variable
6917 (SHLIB_PATH) with shl_load().
6920 *) Constify the ENGINE code as a result of BIGNUM constification.
6921 Also constify the RSA code and most things related to it. In a
6922 few places, most notable in the depth of the ASN.1 code, ugly
6923 casts back to non-const were required (to be solved at a later
6927 *) Make it so the openssl application has all engines loaded by default.
6930 *) Constify the BIGNUM routines a little more.
6933 *) Add the following functions:
6935 ENGINE_load_cswift()
6937 ENGINE_load_atalla()
6939 ENGINE_load_builtin_engines()
6941 That way, an application can itself choose if external engines that
6942 are built-in in OpenSSL shall ever be used or not. The benefit is
6943 that applications won't have to be linked with libdl or other dso
6944 libraries unless it's really needed.
6946 Changed 'openssl engine' to load all engines on demand.
6947 Changed the engine header files to avoid the duplication of some
6948 declarations (they differed!).
6951 *) 'openssl engine' can now list capabilities.
6954 *) Better error reporting in 'openssl engine'.
6957 *) Never call load_dh_param(NULL) in s_server.
6960 *) Add engine application. It can currently list engines by name and
6961 identity, and test if they are actually available.
6964 *) Improve RPM specification file by forcing symbolic linking and making
6965 sure the installed documentation is also owned by root.root.
6966 [Damien Miller <djm@mindrot.org>]
6968 *) Give the OpenSSL applications more possibilities to make use of
6969 keys (public as well as private) handled by engines.
6972 *) Add OCSP code that comes from CertCo.
6975 *) Add VMS support for the Rijndael code.
6978 *) Added untested support for Nuron crypto accelerator.
6981 *) Add support for external cryptographic devices. This code was
6982 previously distributed separately as the "engine" branch.
6983 [Geoff Thorpe, Richard Levitte]
6985 *) Rework the filename-translation in the DSO code. It is now possible to
6986 have far greater control over how a "name" is turned into a filename
6987 depending on the operating environment and any oddities about the
6988 different shared library filenames on each system.
6991 *) Support threads on FreeBSD-elf in Configure.
6994 *) Fix for SHA1 assembly problem with MASM: it produces
6995 warnings about corrupt line number information when assembling
6996 with debugging information. This is caused by the overlapping
6998 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
7001 NCONF_get_number() has no error checking at all. As a replacement,
7002 NCONF_get_number_e() is defined (_e for "error checking") and is
7003 promoted strongly. The old NCONF_get_number is kept around for
7004 binary backward compatibility.
7005 Make it possible for methods to load from something other than a BIO,
7006 by providing a function pointer that is given a name instead of a BIO.
7007 For example, this could be used to load configuration data from an
7011 *) Fix for non blocking accept BIOs. Added new I/O special reason
7012 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
7013 with non blocking I/O was not possible because no retry code was
7014 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
7018 *) Added the beginnings of Rijndael support.
7021 *) Fix for bug in DirectoryString mask setting. Add support for
7022 X509_NAME_print_ex() in 'req' and X509_print_ex() function
7023 to allow certificate printing to more controllable, additional
7024 'certopt' option to 'x509' to allow new printing options to be
7028 *) Clean old EAY MD5 hack from e_os.h.
7031 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
7033 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7034 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7035 [Joe Orton, Steve Henson]
7037 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
7039 *) Fix additional bug revealed by the NISCC test suite:
7041 Stop bug triggering large recursion when presented with
7042 certain ASN.1 tags (CVE-2003-0851)
7045 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
7047 *) Fix various bugs revealed by running the NISCC test suite:
7049 Stop out of bounds reads in the ASN1 code when presented with
7050 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7052 If verify callback ignores invalid public key errors don't try to check
7053 certificate signature with the NULL public key.
7057 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7058 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7062 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7063 extra data after the compression methods not only for TLS 1.0
7064 but also for SSL 3.0 (as required by the specification).
7065 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7067 *) Change X509_certificate_type() to mark the key as exported/exportable
7068 when it's 512 *bits* long, not 512 bytes.
7071 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
7073 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7074 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7075 a protocol version number mismatch like a decryption error
7076 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7079 *) Turn on RSA blinding by default in the default implementation
7080 to avoid a timing attack. Applications that don't want it can call
7081 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7082 They would be ill-advised to do so in most cases.
7083 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7085 *) Change RSA blinding code so that it works when the PRNG is not
7086 seeded (in this case, the secret RSA exponent is abused as
7087 an unpredictable seed -- if it is not unpredictable, there
7088 is no point in blinding anyway). Make RSA blinding thread-safe
7089 by remembering the creator's thread ID in rsa->blinding and
7090 having all other threads use local one-time blinding factors
7091 (this requires more computation than sharing rsa->blinding, but
7092 avoids excessive locking; and if an RSA object is not shared
7093 between threads, blinding will still be very fast).
7096 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7098 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7099 via timing by performing a MAC computation even if incorrrect
7100 block cipher padding has been found. This is a countermeasure
7101 against active attacks where the attacker has to distinguish
7102 between bad padding and a MAC verification error. (CVE-2003-0078)
7104 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7105 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7106 Martin Vuagnoux (EPFL, Ilion)]
7108 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7110 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7111 memory from it's contents. This is done with a counter that will
7112 place alternating values in each byte. This can be used to solve
7113 two issues: 1) the removal of calls to memset() by highly optimizing
7114 compilers, and 2) cleansing with other values than 0, since those can
7115 be read through on certain media, for example a swap space on disk.
7118 *) Bugfix: client side session caching did not work with external caching,
7119 because the session->cipher setting was not restored when reloading
7120 from the external cache. This problem was masked, when
7121 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7122 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7125 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7126 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7127 [Zeev Lieber <zeev-l@yahoo.com>]
7129 *) Undo an undocumented change introduced in 0.9.6e which caused
7130 repeated calls to OpenSSL_add_all_ciphers() and
7131 OpenSSL_add_all_digests() to be ignored, even after calling
7135 *) Change the default configuration reader to deal with last line not
7136 being properly terminated.
7139 *) Change X509_NAME_cmp() so it applies the special rules on handling
7140 DN values that are of type PrintableString, as well as RDNs of type
7141 emailAddress where the value has the type ia5String.
7142 [stefank@valicert.com via Richard Levitte]
7144 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7145 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7146 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7147 the bitwise-OR of the two for use by the majority of applications
7148 wanting this behaviour, and update the docs. The documented
7149 behaviour and actual behaviour were inconsistent and had been
7150 changing anyway, so this is more a bug-fix than a behavioural
7152 [Geoff Thorpe, diagnosed by Nadav Har'El]
7154 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7155 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7158 *) Fix initialization code race conditions in
7159 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7160 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7161 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7162 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7163 ssl2_get_cipher_by_char(),
7164 ssl3_get_cipher_by_char().
7165 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7167 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7168 the cached sessions are flushed, as the remove_cb() might use ex_data
7169 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7170 (see [openssl.org #212]).
7171 [Geoff Thorpe, Lutz Jaenicke]
7173 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7174 length, instead of the encoding length to d2i_ASN1_OBJECT.
7177 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7179 *) [In 0.9.6g-engine release:]
7180 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7181 [Lynn Gazis <lgazis@rainbow.com>]
7183 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7185 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7186 and get fix the header length calculation.
7187 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7188 Alon Kantor <alonk@checkpoint.com> (and others),
7191 *) Use proper error handling instead of 'assertions' in buffer
7192 overflow checks added in 0.9.6e. This prevents DoS (the
7193 assertions could call abort()).
7194 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7196 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7198 *) Add various sanity checks to asn1_get_length() to reject
7199 the ASN1 length bytes if they exceed sizeof(long), will appear
7200 negative or the content length exceeds the length of the
7202 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7204 *) Fix cipher selection routines: ciphers without encryption had no flags
7205 for the cipher strength set and where therefore not handled correctly
7206 by the selection routines (PR #130).
7209 *) Fix EVP_dsa_sha macro.
7213 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7214 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7215 that was added in OpenSSL 0.9.6d.
7217 As the countermeasure turned out to be incompatible with some
7218 broken SSL implementations, the new option is part of SSL_OP_ALL.
7219 SSL_OP_ALL is usually employed when compatibility with weird SSL
7220 implementations is desired (e.g. '-bugs' option to 's_client' and
7221 's_server'), so the new option is automatically set in many
7225 *) Changes in security patch:
7227 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7228 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7229 Air Force Materiel Command, USAF, under agreement number
7232 *) Add various sanity checks to asn1_get_length() to reject
7233 the ASN1 length bytes if they exceed sizeof(long), will appear
7234 negative or the content length exceeds the length of the
7235 supplied buffer. (CVE-2002-0659)
7236 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7238 *) Assertions for various potential buffer overflows, not known to
7240 [Ben Laurie (CHATS)]
7242 *) Various temporary buffers to hold ASCII versions of integers were
7243 too small for 64 bit platforms. (CVE-2002-0655)
7244 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7246 *) Remote buffer overflow in SSL3 protocol - an attacker could
7247 supply an oversized session ID to a client. (CVE-2002-0656)
7248 [Ben Laurie (CHATS)]
7250 *) Remote buffer overflow in SSL2 protocol - an attacker could
7251 supply an oversized client master key. (CVE-2002-0656)
7252 [Ben Laurie (CHATS)]
7254 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7256 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7257 encoded as NULL) with id-dsa-with-sha1.
7258 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7260 *) Check various X509_...() return values in apps/req.c.
7261 [Nils Larsch <nla@trustcenter.de>]
7263 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7264 an end-of-file condition would erronously be flagged, when the CRLF
7265 was just at the end of a processed block. The bug was discovered when
7266 processing data through a buffering memory BIO handing the data to a
7267 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7268 <ptsekov@syntrex.com> and Nedelcho Stanev.
7271 *) Implement a countermeasure against a vulnerability recently found
7272 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7273 before application data chunks to avoid the use of known IVs
7274 with data potentially chosen by the attacker.
7277 *) Fix length checks in ssl3_get_client_hello().
7280 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7281 to prevent ssl3_read_internal() from incorrectly assuming that
7282 ssl3_read_bytes() found application data while handshake
7283 processing was enabled when in fact s->s3->in_read_app_data was
7284 merely automatically cleared during the initial handshake.
7285 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7287 *) Fix object definitions for Private and Enterprise: they were not
7288 recognized in their shortname (=lowercase) representation. Extend
7289 obj_dat.pl to issue an error when using undefined keywords instead
7290 of silently ignoring the problem (Svenning Sorensen
7291 <sss@sss.dnsalias.net>).
7294 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7295 generators, i.e. generators other than 2 and 5. (Previously, the
7296 code did not properly initialise the 'add' and 'rem' values to
7297 BN_generate_prime().)
7299 In the new general case, we do not insist that 'generator' is
7300 actually a primitive root: This requirement is rather pointless;
7301 a generator of the order-q subgroup is just as good, if not
7305 *) Map new X509 verification errors to alerts. Discovered and submitted by
7306 Tom Wu <tom@arcot.com>.
7309 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7310 returning non-zero before the data has been completely received
7311 when using non-blocking I/O.
7312 [Bodo Moeller; problem pointed out by John Hughes]
7314 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7315 [Ben Laurie, Lutz Jaenicke]
7317 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7318 Yoram Zahavi <YoramZ@gilian.com>).
7321 *) Add information about CygWin 1.3 and on, and preserve proper
7322 configuration for the versions before that.
7323 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7325 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7326 check whether we deal with a copy of a session and do not delete from
7327 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7328 <izhar@checkpoint.com>.
7331 *) Do not store session data into the internal session cache, if it
7332 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7333 flag is set). Proposed by Aslam <aslam@funk.com>.
7336 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7340 *) [In 0.9.6d-engine release:]
7341 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7342 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7344 *) Add the configuration target linux-s390x.
7345 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7347 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7348 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7349 variable as an indication that a ClientHello message has been
7350 received. As the flag value will be lost between multiple
7351 invocations of ssl3_accept when using non-blocking I/O, the
7352 function may not be aware that a handshake has actually taken
7353 place, thus preventing a new session from being added to the
7356 To avoid this problem, we now set s->new_session to 2 instead of
7357 using a local variable.
7358 [Lutz Jaenicke, Bodo Moeller]
7360 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7361 if the SSL_R_LENGTH_MISMATCH error is detected.
7362 [Geoff Thorpe, Bodo Moeller]
7364 *) New 'shared_ldflag' column in Configure platform table.
7367 *) Fix EVP_CIPHER_mode macro.
7368 ["Dan S. Camper" <dan@bti.net>]
7370 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7371 type, we must throw them away by setting rr->length to 0.
7372 [D P Chang <dpc@qualys.com>]
7374 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7376 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7377 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7378 worked incorrectly for those cases where range = 10..._2 and
7379 3*range is two bits longer than range.)
7382 *) Only add signing time to PKCS7 structures if it is not already
7386 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7387 OBJ_ld_ce should be OBJ_id_ce.
7388 Also some ip-pda OIDs in crypto/objects/objects.txt were
7389 incorrect (cf. RFC 3039).
7390 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7392 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7393 returns early because it has nothing to do.
7394 [Andy Schneider <andy.schneider@bjss.co.uk>]
7396 *) [In 0.9.6c-engine release:]
7397 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7398 [Andy Schneider <andy.schneider@bjss.co.uk>]
7400 *) [In 0.9.6c-engine release:]
7401 Add support for Cryptographic Appliance's keyserver technology.
7402 (Use engine 'keyclient')
7403 [Cryptographic Appliances and Geoff Thorpe]
7405 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7406 is called via tools/c89.sh because arguments have to be
7407 rearranged (all '-L' options must appear before the first object
7409 [Richard Shapiro <rshapiro@abinitio.com>]
7411 *) [In 0.9.6c-engine release:]
7412 Add support for Broadcom crypto accelerator cards, backported
7414 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7416 *) [In 0.9.6c-engine release:]
7417 Add support for SureWare crypto accelerator cards from
7418 Baltimore Technologies. (Use engine 'sureware')
7419 [Baltimore Technologies and Mark Cox]
7421 *) [In 0.9.6c-engine release:]
7422 Add support for crypto accelerator cards from Accelerated
7423 Encryption Processing, www.aep.ie. (Use engine 'aep')
7424 [AEP Inc. and Mark Cox]
7426 *) Add a configuration entry for gcc on UnixWare.
7427 [Gary Benson <gbenson@redhat.com>]
7429 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7430 messages are stored in a single piece (fixed-length part and
7431 variable-length part combined) and fix various bugs found on the way.
7434 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7435 instead. BIO_gethostbyname() does not know what timeouts are
7436 appropriate, so entries would stay in cache even when they have
7438 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7440 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7441 faced with a pathologically small ClientHello fragment that does
7442 not contain client_version: Instead of aborting with an error,
7443 simply choose the highest available protocol version (i.e.,
7444 TLS 1.0 unless it is disabled). In practice, ClientHello
7445 messages are never sent like this, but this change gives us
7446 strictly correct behaviour at least for TLS.
7449 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7450 never resets s->method to s->ctx->method when called from within
7451 one of the SSL handshake functions.
7452 [Bodo Moeller; problem pointed out by Niko Baric]
7454 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7455 (sent using the client's version number) if client_version is
7456 smaller than the protocol version in use. Also change
7457 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7458 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7459 the client will at least see that alert.
7462 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7466 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7467 client receives HelloRequest while in a handshake.
7468 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7470 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7471 should end in 'break', not 'goto end' which circuments various
7472 cleanups done in state SSL_ST_OK. But session related stuff
7473 must be disabled for SSL_ST_OK in the case that we just sent a
7476 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7477 before just sending a HelloRequest.
7478 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7480 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7481 reveal whether illegal block cipher padding was found or a MAC
7482 verification error occured. (Neither SSLerr() codes nor alerts
7483 are directly visible to potential attackers, but the information
7484 may leak via logfiles.)
7486 Similar changes are not required for the SSL 2.0 implementation
7487 because the number of padding bytes is sent in clear for SSL 2.0,
7488 and the extra bytes are just ignored. However ssl/s2_pkt.c
7489 failed to verify that the purported number of padding bytes is in
7493 *) Add OpenUNIX-8 support including shared libraries
7494 (Boyd Lynn Gerber <gerberb@zenez.com>).
7497 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7498 'wristwatch attack' using huge encoding parameters (cf.
7499 James H. Manger's CRYPTO 2001 paper). Note that the
7500 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7501 encoding parameters and hence was not vulnerable.
7504 *) BN_sqr() bug fix.
7505 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7507 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7508 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7509 followed by modular reduction.
7510 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7512 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7513 equivalent based on BN_pseudo_rand() instead of BN_rand().
7516 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7517 This function was broken, as the check for a new client hello message
7518 to handle SGC did not allow these large messages.
7519 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7522 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7525 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7526 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7529 *) Rework the configuration and shared library support for Tru64 Unix.
7530 The configuration part makes use of modern compiler features and
7531 still retains old compiler behavior for those that run older versions
7532 of the OS. The shared library support part includes a variant that
7533 uses the RPATH feature, and is available through the special
7534 configuration target "alpha-cc-rpath", which will never be selected
7536 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7538 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7539 with the same message size as in ssl3_get_certificate_request().
7540 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7541 messages might inadvertently be reject as too long.
7542 [Petr Lampa <lampa@fee.vutbr.cz>]
7544 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7547 *) Modified SSL library such that the verify_callback that has been set
7548 specificly for an SSL object with SSL_set_verify() is actually being
7549 used. Before the change, a verify_callback set with this function was
7550 ignored and the verify_callback() set in the SSL_CTX at the time of
7551 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7552 to allow the necessary settings.
7555 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7556 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7557 done automatically (in contradiction to the requirements of the C
7558 standard). This made problems when used from OpenSSH.
7561 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7562 dh->length and always used
7564 BN_rand_range(priv_key, dh->p).
7566 BN_rand_range() is not necessary for Diffie-Hellman, and this
7567 specific range makes Diffie-Hellman unnecessarily inefficient if
7568 dh->length (recommended exponent length) is much smaller than the
7569 length of dh->p. We could use BN_rand_range() if the order of
7570 the subgroup was stored in the DH structure, but we only have
7575 BN_rand(priv_key, l, ...)
7577 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7583 RSA_eay_public_encrypt
7584 RSA_eay_private_decrypt
7585 RSA_eay_private_encrypt (signing)
7586 RSA_eay_public_decrypt (signature verification)
7588 (default implementations for RSA_public_encrypt,
7589 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7590 always reject numbers >= n.
7593 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7594 to synchronize access to 'locking_thread'. This is necessary on
7595 systems where access to 'locking_thread' (an 'unsigned long'
7596 variable) is not atomic.
7599 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7600 *before* setting the 'crypto_lock_rand' flag. The previous code had
7601 a race condition if 0 is a valid thread ID.
7602 [Travis Vitek <vitek@roguewave.com>]
7604 *) Add support for shared libraries under Irix.
7605 [Albert Chin-A-Young <china@thewrittenword.com>]
7607 *) Add configuration option to build on Linux on both big-endian and
7609 [Ralf Baechle <ralf@uni-koblenz.de>]
7611 *) Add the possibility to create shared libraries on HP-UX.
7614 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7616 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7617 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7618 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7619 PRNG state recovery was possible based on the output of
7620 one PRNG request appropriately sized to gain knowledge on
7621 'md' followed by enough consecutive 1-byte PRNG requests
7622 to traverse all of 'state'.
7624 1. When updating 'md_local' (the current thread's copy of 'md')
7625 during PRNG output generation, hash all of the previous
7626 'md_local' value, not just the half used for PRNG output.
7628 2. Make the number of bytes from 'state' included into the hash
7629 independent from the number of PRNG bytes requested.
7631 The first measure alone would be sufficient to avoid
7632 Markku-Juhani's attack. (Actually it had never occurred
7633 to me that the half of 'md_local' used for chaining was the
7634 half from which PRNG output bytes were taken -- I had always
7635 assumed that the secret half would be used.) The second
7636 measure makes sure that additional data from 'state' is never
7637 mixed into 'md_local' in small portions; this heuristically
7638 further strengthens the PRNG.
7641 *) Fix crypto/bn/asm/mips3.s.
7644 *) When only the key is given to "enc", the IV is undefined. Print out
7645 an error message in this case.
7648 *) Handle special case when X509_NAME is empty in X509 printing routines.
7651 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7652 positive and less than q.
7655 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7656 used: it isn't thread safe and the add_lock_callback should handle
7658 [Paul Rose <Paul.Rose@bridge.com>]
7660 *) Verify that incoming data obeys the block size in
7661 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7665 [Ulf Möller, Bodo Möller]
7667 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7668 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7669 when fixing the server behaviour for backwards-compatible 'client
7670 hello' messages. (Note that the attack is impractical against
7671 SSL 3.0 and TLS 1.0 anyway because length and version checking
7672 means that the probability of guessing a valid ciphertext is
7673 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7676 Before 0.9.5, the countermeasure (hide the error by generating a
7677 random 'decryption result') did not work properly because
7678 ERR_clear_error() was missing, meaning that SSL_get_error() would
7679 detect the supposedly ignored error.
7681 Both problems are now fixed.
7684 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7685 (previously it was 1024).
7688 *) Fix for compatibility mode trust settings: ignore trust settings
7689 unless some valid trust or reject settings are present.
7692 *) Fix for blowfish EVP: its a variable length cipher.
7695 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7696 parameters in DSA public key structures and return an error in the
7697 DSA routines if parameters are absent.
7700 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7701 in the current directory if neither $RANDFILE nor $HOME was set.
7702 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7703 caused some confusion to Windows users who haven't defined $HOME.
7704 Thus RAND_file_name() is changed again: e_os.h can define a
7705 DEFAULT_HOME, which will be used if $HOME is not set.
7706 For Windows, we use "C:"; on other platforms, we still require
7707 environment variables.
7709 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7710 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7711 having multiple threads call RAND_poll() concurrently.
7714 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7715 combination of a flag and a thread ID variable.
7716 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7717 flag), *other* threads can enter ssleay_add_bytes without obeying
7718 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7719 that they do not hold after the first thread unsets add_do_not_lock).
7722 *) Change bctest again: '-x' expressions are not available in all
7726 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7728 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7729 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7731 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7732 the default extension for executables, if any. Also, make the perl
7733 scripts that use symlink() to test if it really exists and use "cp"
7734 if it doesn't. All this made OpenSSL compilable and installable in
7738 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7739 If SEQUENCE is length is indefinite just set c->slen to the total
7740 amount of data available.
7741 [Steve Henson, reported by shige@FreeBSD.org]
7742 [This change does not apply to 0.9.7.]
7744 *) Change bctest to avoid here-documents inside command substitution
7745 (workaround for FreeBSD /bin/sh bug).
7746 For compatibility with Ultrix, avoid shell functions (introduced
7747 in the bctest version that searches along $PATH).
7750 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7751 with des_encrypt() defined on some operating systems, like Solaris
7755 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7756 On the Importance of Eliminating Errors in Cryptographic
7757 Computations, J. Cryptology 14 (2001) 2, 101-119,
7758 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7761 *) MIPS assembler BIGNUM division bug fix.
7764 *) Disabled incorrect Alpha assembler code.
7767 *) Fix PKCS#7 decode routines so they correctly update the length
7768 after reading an EOC for the EXPLICIT tag.
7770 [This change does not apply to 0.9.7.]
7772 *) Fix bug in PKCS#12 key generation routines. This was triggered
7773 if a 3DES key was generated with a 0 initial byte. Include
7774 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7775 (but broken) behaviour.
7778 *) Enhance bctest to search for a working bc along $PATH and print
7780 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7782 *) Fix memory leaks in err.c: free err_data string if necessary;
7783 don't write to the wrong index in ERR_set_error_data.
7786 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7790 *) Replace rdtsc with _emit statements for VC++ version 5.
7791 [Jeremy Cooper <jeremy@baymoo.org>]
7793 *) Make it possible to reuse SSLv2 sessions.
7796 *) In copy_email() check for >= 0 as a return value for
7797 X509_NAME_get_index_by_NID() since 0 is a valid index.
7798 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7800 *) Avoid coredump with unsupported or invalid public keys by checking if
7801 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7802 PKCS7_verify() fails with non detached data.
7805 *) Don't use getenv in library functions when run as setuid/setgid.
7806 New function OPENSSL_issetugid().
7809 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7810 due to incorrect handling of multi-threading:
7812 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7814 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7816 3. Count how many times MemCheck_off() has been called so that
7817 nested use can be treated correctly. This also avoids
7818 inband-signalling in the previous code (which relied on the
7819 assumption that thread ID 0 is impossible).
7822 *) Add "-rand" option also to s_client and s_server.
7825 *) Fix CPU detection on Irix 6.x.
7826 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7827 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7829 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7832 [This change does not apply to 0.9.7.]
7834 *) Use the cached encoding of an X509_NAME structure rather than
7835 copying it. This is apparently the reason for the libsafe "errors"
7836 but the code is actually correct.
7839 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7840 Bleichenbacher's DSA attack.
7841 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7842 to be set and top=0 forces the highest bit to be set; top=-1 is new
7843 and leaves the highest bit random.
7844 [Ulf Moeller, Bodo Moeller]
7846 *) In the NCONF_...-based implementations for CONF_... queries
7847 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7848 a temporary CONF structure with the data component set to NULL
7849 (which gives segmentation faults in lh_retrieve).
7850 Instead, use NULL for the CONF pointer in CONF_get_string and
7851 CONF_get_number (which may use environment variables) and directly
7852 return NULL from CONF_get_section.
7855 *) Fix potential buffer overrun for EBCDIC.
7858 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7859 keyUsage if basicConstraints absent for a CA.
7862 *) Make SMIME_write_PKCS7() write mail header values with a format that
7863 is more generally accepted (no spaces before the semicolon), since
7864 some programs can't parse those values properly otherwise. Also make
7865 sure BIO's that break lines after each write do not create invalid
7869 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7870 macros previously used would not encode an empty SEQUENCE OF
7871 and break the signature.
7873 [This change does not apply to 0.9.7.]
7875 *) Zero the premaster secret after deriving the master secret in
7879 *) Add some EVP_add_digest_alias registrations (as found in
7880 OpenSSL_add_all_digests()) to SSL_library_init()
7881 aka OpenSSL_add_ssl_algorithms(). This provides improved
7882 compatibility with peers using X.509 certificates
7883 with unconventional AlgorithmIdentifier OIDs.
7886 *) Fix for Irix with NO_ASM.
7887 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7889 *) ./config script fixes.
7890 [Ulf Moeller, Richard Levitte]
7892 *) Fix 'openssl passwd -1'.
7895 *) Change PKCS12_key_gen_asc() so it can cope with non null
7896 terminated strings whose length is passed in the passlen
7897 parameter, for example from PEM callbacks. This was done
7898 by adding an extra length parameter to asc2uni().
7899 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7901 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7902 call failed, free the DSA structure.
7905 *) Fix to uni2asc() to cope with zero length Unicode strings.
7906 These are present in some PKCS#12 files.
7909 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7910 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7911 when writing a 32767 byte record.
7912 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7914 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7915 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7917 (RSA objects have a reference count access to which is protected
7918 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7919 so they are meant to be shared between threads.)
7920 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7921 "Reddie, Steven" <Steven.Reddie@ca.com>]
7923 *) Fix a deadlock in CRYPTO_mem_leaks().
7926 *) Use better test patterns in bntest.
7929 *) rand_win.c fix for Borland C.
7932 *) BN_rshift bugfix for n == 0.
7935 *) Add a 'bctest' script that checks for some known 'bc' bugs
7936 so that 'make test' does not abort just because 'bc' is broken.
7939 *) Store verify_result within SSL_SESSION also for client side to
7940 avoid potential security hole. (Re-used sessions on the client side
7941 always resulted in verify_result==X509_V_OK, not using the original
7942 result of the server certificate verification.)
7945 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7946 SSL3_RT_APPLICATION_DATA, return 0.
7947 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7951 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7952 releases, have been re-implemented by renaming the previous
7953 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7954 and ssl3_read_internal, respectively, and adding 'peek' parameters
7955 to them. The new ssl[23]_{read,peek} functions are calls to
7956 ssl[23]_read_internal with the 'peek' flag set appropriately.
7957 A 'peek' parameter has also been added to ssl3_read_bytes, which
7958 does the actual work for ssl3_read_internal.
7961 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7962 the method-specific "init()" handler. Also clean up ex_data after
7963 calling the method-specific "finish()" handler. Previously, this was
7964 happening the other way round.
7967 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7968 The previous value, 12, was not always sufficient for BN_mod_exp().
7971 *) Make sure that shared libraries get the internal name engine with
7972 the full version number and not just 0. This should mark the
7973 shared libraries as not backward compatible. Of course, this should
7974 be changed again when we can guarantee backward binary compatibility.
7977 *) Fix typo in get_cert_by_subject() in by_dir.c
7978 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7980 *) Rework the system to generate shared libraries:
7982 - Make note of the expected extension for the shared libraries and
7983 if there is a need for symbolic links from for example libcrypto.so.0
7984 to libcrypto.so.0.9.7. There is extended info in Configure for
7987 - Make as few rebuilds of the shared libraries as possible.
7989 - Still avoid linking the OpenSSL programs with the shared libraries.
7991 - When installing, install the shared libraries separately from the
7995 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7997 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7998 and not in SSL_clear because the latter is also used by the
7999 accept/connect functions; previously, the settings made by
8000 SSL_set_read_ahead would be lost during the handshake.
8001 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
8003 *) Correct util/mkdef.pl to be selective about disabled algorithms.
8004 Previously, it would create entries for disableed algorithms no
8008 *) Added several new manual pages for SSL_* function.
8011 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
8013 *) In ssl23_get_client_hello, generate an error message when faced
8014 with an initial SSL 3.0/TLS record that is too small to contain the
8015 first two bytes of the ClientHello message, i.e. client_version.
8016 (Note that this is a pathologic case that probably has never happened
8017 in real life.) The previous approach was to use the version number
8018 from the record header as a substitute; but our protocol choice
8019 should not depend on that one because it is not authenticated
8020 by the Finished messages.
8023 *) More robust randomness gathering functions for Windows.
8024 [Jeffrey Altman <jaltman@columbia.edu>]
8026 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
8027 not set then we don't setup the error code for issuer check errors
8028 to avoid possibly overwriting other errors which the callback does
8029 handle. If an application does set the flag then we assume it knows
8030 what it is doing and can handle the new informational codes
8034 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
8035 a general "ANY" type, as such it should be able to decode anything
8036 including tagged types. However it didn't check the class so it would
8037 wrongly interpret tagged types in the same way as their universal
8038 counterpart and unknown types were just rejected. Changed so that the
8039 tagged and unknown types are handled in the same way as a SEQUENCE:
8040 that is the encoding is stored intact. There is also a new type
8041 "V_ASN1_OTHER" which is used when the class is not universal, in this
8042 case we have no idea what the actual type is so we just lump them all
8046 *) On VMS, stdout may very well lead to a file that is written to
8047 in a record-oriented fashion. That means that every write() will
8048 write a separate record, which will be read separately by the
8049 programs trying to read from it. This can be very confusing.
8051 The solution is to put a BIO filter in the way that will buffer
8052 text until a linefeed is reached, and then write everything a
8053 line at a time, so every record written will be an actual line,
8054 not chunks of lines and not (usually doesn't happen, but I've
8055 seen it once) several lines in one record. BIO_f_linebuffer() is
8058 Currently, it's a VMS-only method, because that's where it has
8059 been tested well enough.
8062 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
8063 it can return incorrect results.
8064 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
8065 but it was in 0.9.6-beta[12].)
8068 *) Disable the check for content being present when verifying detached
8069 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
8070 include zero length content when signing messages.
8073 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
8074 BIO_ctrl (for BIO pairs).
8077 *) Add DSO method for VMS.
8080 *) Bug fix: Montgomery multiplication could produce results with the
8084 *) Add RPM specification openssl.spec and modify it to build three
8085 packages. The default package contains applications, application
8086 documentation and run-time libraries. The devel package contains
8087 include files, static libraries and function documentation. The
8088 doc package contains the contents of the doc directory. The original
8089 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8092 *) Add a large number of documentation files for many SSL routines.
8093 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8095 *) Add a configuration entry for Sony News 4.
8096 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8098 *) Don't set the two most significant bits to one when generating a
8099 random number < q in the DSA library.
8102 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8103 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8104 the underlying transport is blocking) if a handshake took place.
8105 (The default behaviour is needed by applications such as s_client
8106 and s_server that use select() to determine when to use SSL_read;
8107 but for applications that know in advance when to expect data, it
8108 just makes things more complicated.)
8111 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8115 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8116 work better on such systems.
8117 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8119 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8120 Update PKCS12_parse() so it copies the friendlyName and the
8121 keyid to the certificates aux info.
8124 *) Fix bug in PKCS7_verify() which caused an infinite loop
8125 if there was more than one signature.
8126 [Sven Uszpelkat <su@celocom.de>]
8128 *) Major change in util/mkdef.pl to include extra information
8129 about each symbol, as well as presentig variables as well
8130 as functions. This change means that there's n more need
8131 to rebuild the .num files when some algorithms are excluded.
8134 *) Allow the verify time to be set by an application,
8135 rather than always using the current time.
8138 *) Phase 2 verify code reorganisation. The certificate
8139 verify code now looks up an issuer certificate by a
8140 number of criteria: subject name, authority key id
8141 and key usage. It also verifies self signed certificates
8142 by the same criteria. The main comparison function is
8143 X509_check_issued() which performs these checks.
8145 Lot of changes were necessary in order to support this
8146 without completely rewriting the lookup code.
8148 Authority and subject key identifier are now cached.
8150 The LHASH 'certs' is X509_STORE has now been replaced
8151 by a STACK_OF(X509_OBJECT). This is mainly because an
8152 LHASH can't store or retrieve multiple objects with
8153 the same hash value.
8155 As a result various functions (which were all internal
8156 use only) have changed to handle the new X509_STORE
8157 structure. This will break anything that messed round
8158 with X509_STORE internally.
8160 The functions X509_STORE_add_cert() now checks for an
8161 exact match, rather than just subject name.
8163 The X509_STORE API doesn't directly support the retrieval
8164 of multiple certificates matching a given criteria, however
8165 this can be worked round by performing a lookup first
8166 (which will fill the cache with candidate certificates)
8167 and then examining the cache for matches. This is probably
8168 the best we can do without throwing out X509_LOOKUP
8169 entirely (maybe later...).
8171 The X509_VERIFY_CTX structure has been enhanced considerably.
8173 All certificate lookup operations now go via a get_issuer()
8174 callback. Although this currently uses an X509_STORE it
8175 can be replaced by custom lookups. This is a simple way
8176 to bypass the X509_STORE hackery necessary to make this
8177 work and makes it possible to use more efficient techniques
8178 in future. A very simple version which uses a simple
8179 STACK for its trusted certificate store is also provided
8180 using X509_STORE_CTX_trusted_stack().
8182 The verify_cb() and verify() callbacks now have equivalents
8183 in the X509_STORE_CTX structure.
8185 X509_STORE_CTX also has a 'flags' field which can be used
8186 to customise the verify behaviour.
8189 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8190 excludes S/MIME capabilities.
8193 *) When a certificate request is read in keep a copy of the
8194 original encoding of the signed data and use it when outputing
8195 again. Signatures then use the original encoding rather than
8196 a decoded, encoded version which may cause problems if the
8197 request is improperly encoded.
8200 *) For consistency with other BIO_puts implementations, call
8201 buffer_write(b, ...) directly in buffer_puts instead of calling
8204 In BIO_puts, increment b->num_write as in BIO_write.
8205 [Peter.Sylvester@EdelWeb.fr]
8207 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8208 BN_zero, we may not return a BIGNUM with an array consisting of
8212 *) Avoid calling abort() from within the library when problems are
8213 detected, except if preprocessor symbols have been defined
8214 (such as REF_CHECK, BN_DEBUG etc.).
8217 *) New openssl application 'rsautl'. This utility can be
8218 used for low level RSA operations. DER public key
8219 BIO/fp routines also added.
8222 *) New Configure entry and patches for compiling on QNX 4.
8223 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8225 *) A demo state-machine implementation was sponsored by
8226 Nuron (http://www.nuron.com/) and is now available in
8227 demos/state_machine.
8230 *) New options added to the 'dgst' utility for signature
8231 generation and verification.
8234 *) Unrecognized PKCS#7 content types are now handled via a
8235 catch all ASN1_TYPE structure. This allows unsupported
8236 types to be stored as a "blob" and an application can
8237 encode and decode it manually.
8240 *) Fix various signed/unsigned issues to make a_strex.c
8242 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8244 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8245 length if passed a buffer. ASN1_INTEGER_to_BN failed
8246 if passed a NULL BN and its argument was negative.
8247 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8249 *) Modification to PKCS#7 encoding routines to output definite
8250 length encoding. Since currently the whole structures are in
8251 memory there's not real point in using indefinite length
8252 constructed encoding. However if OpenSSL is compiled with
8253 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8256 *) Added BIO_vprintf() and BIO_vsnprintf().
8259 *) Added more prefixes to parse for in the the strings written
8260 through a logging bio, to cover all the levels that are available
8261 through syslog. The prefixes are now:
8263 PANIC, EMERG, EMR => LOG_EMERG
8264 ALERT, ALR => LOG_ALERT
8265 CRIT, CRI => LOG_CRIT
8266 ERROR, ERR => LOG_ERR
8267 WARNING, WARN, WAR => LOG_WARNING
8268 NOTICE, NOTE, NOT => LOG_NOTICE
8269 INFO, INF => LOG_INFO
8270 DEBUG, DBG => LOG_DEBUG
8272 and as before, if none of those prefixes are present at the
8273 beginning of the string, LOG_ERR is chosen.
8275 On Win32, the LOG_* levels are mapped according to this:
8277 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8278 LOG_WARNING => EVENTLOG_WARNING_TYPE
8279 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8283 *) Made it possible to reconfigure with just the configuration
8284 argument "reconf" or "reconfigure". The command line arguments
8285 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8286 and are retrieved from there when reconfiguring.
8290 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8292 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8295 *) The obj_dat.pl script was messing up the sorting of object
8296 names. The reason was that it compared the quoted version
8297 of strings as a result "OCSP" > "OCSP Signing" because
8298 " > SPACE. Changed script to store unquoted versions of
8299 names and add quotes on output. It was also omitting some
8300 names from the lookup table if they were given a default
8301 value (that is if SN is missing it is given the same
8302 value as LN and vice versa), these are now added on the
8303 grounds that if an object has a name we should be able to
8304 look it up. Finally added warning output when duplicate
8305 short or long names are found.
8308 *) Changes needed for Tandem NSK.
8309 [Scott Uroff <scott@xypro.com>]
8311 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8312 RSA_padding_check_SSLv23(), special padding was never detected
8313 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8314 version rollback attacks was not effective.
8316 In s23_clnt.c, don't use special rollback-attack detection padding
8317 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8318 client; similarly, in s23_srvr.c, don't do the rollback check if
8319 SSL 2.0 is the only protocol enabled in the server.
8322 *) Make it possible to get hexdumps of unprintable data with 'openssl
8323 asn1parse'. By implication, the functions ASN1_parse_dump() and
8324 BIO_dump_indent() are added.
8327 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8328 these print out strings and name structures based on various
8329 flags including RFC2253 support and proper handling of
8330 multibyte characters. Added options to the 'x509' utility
8331 to allow the various flags to be set.
8334 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8335 Also change the functions X509_cmp_current_time() and
8336 X509_gmtime_adj() work with an ASN1_TIME structure,
8337 this will enable certificates using GeneralizedTime in validity
8338 dates to be checked.
8341 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8342 negative public key encodings) on by default,
8343 NO_NEG_PUBKEY_BUG can be set to disable it.
8346 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8347 content octets. An i2c_ASN1_OBJECT is unnecessary because
8348 the encoding can be trivially obtained from the structure.
8351 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8352 not read locks (CRYPTO_r_[un]lock).
8355 *) A first attempt at creating official support for shared
8356 libraries through configuration. I've kept it so the
8357 default is static libraries only, and the OpenSSL programs
8358 are always statically linked for now, but there are
8359 preparations for dynamic linking in place.
8360 This has been tested on Linux and Tru64.
8363 *) Randomness polling function for Win9x, as described in:
8364 Peter Gutmann, Software Generation of Practically Strong
8368 *) Fix so PRNG is seeded in req if using an already existing
8372 *) New options to smime application. -inform and -outform
8373 allow alternative formats for the S/MIME message including
8374 PEM and DER. The -content option allows the content to be
8375 specified separately. This should allow things like Netscape
8376 form signing output easier to verify.
8379 *) Fix the ASN1 encoding of tags using the 'long form'.
8382 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8383 STRING types. These convert content octets to and from the
8384 underlying type. The actual tag and length octets are
8385 already assumed to have been read in and checked. These
8386 are needed because all other string types have virtually
8387 identical handling apart from the tag. By having versions
8388 of the ASN1 functions that just operate on content octets
8389 IMPLICIT tagging can be handled properly. It also allows
8390 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8391 and ASN1_INTEGER are identical apart from the tag.
8394 *) Change the handling of OID objects as follows:
8396 - New object identifiers are inserted in objects.txt, following
8397 the syntax given in objects.README.
8398 - objects.pl is used to process obj_mac.num and create a new
8400 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8403 This is currently kind of a hack, and the perl code in objects.pl
8404 isn't very elegant, but it works as I intended. The simplest way
8405 to check that it worked correctly is to look in obj_dat.h and
8406 check the array nid_objs and make sure the objects haven't moved
8407 around (this is important!). Additions are OK, as well as
8408 consistent name changes.
8411 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8414 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8415 The given file adds to whatever has already been seeded into the
8416 random pool through the RANDFILE configuration file option or
8417 environment variable, or the default random state file.
8420 *) mkstack.pl now sorts each macro group into lexical order.
8421 Previously the output order depended on the order the files
8422 appeared in the directory, resulting in needless rewriting
8426 *) Patches to make OpenSSL compile under Win32 again. Mostly
8427 work arounds for the VC++ problem that it treats func() as
8428 func(void). Also stripped out the parts of mkdef.pl that
8429 added extra typesafe functions: these no longer exist.
8432 *) Reorganisation of the stack code. The macros are now all
8433 collected in safestack.h . Each macro is defined in terms of
8434 a "stack macro" of the form SKM_<name>(type, a, b). The
8435 DEBUG_SAFESTACK is now handled in terms of function casts,
8436 this has the advantage of retaining type safety without the
8437 use of additional functions. If DEBUG_SAFESTACK is not defined
8438 then the non typesafe macros are used instead. Also modified the
8439 mkstack.pl script to handle the new form. Needs testing to see
8440 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8441 the default if no major problems. Similar behaviour for ASN1_SET_OF
8442 and PKCS12_STACK_OF.
8445 *) When some versions of IIS use the 'NET' form of private key the
8446 key derivation algorithm is different. Normally MD5(password) is
8447 used as a 128 bit RC4 key. In the modified case
8448 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
8449 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8450 as the old Netscape_RSA functions except they have an additional
8451 'sgckey' parameter which uses the modified algorithm. Also added
8452 an -sgckey command line option to the rsa utility. Thanks to
8453 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8454 algorithm to openssl-dev.
8457 *) The evp_local.h macros were using 'c.##kname' which resulted in
8458 invalid expansion on some systems (SCO 5.0.5 for example).
8459 Corrected to 'c.kname'.
8460 [Phillip Porch <root@theporch.com>]
8462 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8463 a STACK of email addresses from a certificate or request, these look
8464 in the subject name and the subject alternative name extensions and
8465 omit any duplicate addresses.
8468 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8469 This makes DSA verification about 2 % faster.
8472 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8473 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8474 plus overhead for 1024 bit moduli).
8475 This makes exponentiations about 0.5 % faster for 1024 bit
8476 exponents (as measured by "openssl speed rsa2048").
8479 *) Rename memory handling macros to avoid conflicts with other
8481 Malloc => OPENSSL_malloc
8482 Malloc_locked => OPENSSL_malloc_locked
8483 Realloc => OPENSSL_realloc
8484 Free => OPENSSL_free
8487 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8488 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8491 *) CygWin32 support.
8492 [John Jarvie <jjarvie@newsguy.com>]
8494 *) The type-safe stack code has been rejigged. It is now only compiled
8495 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8496 by default all type-specific stack functions are "#define"d back to
8497 standard stack functions. This results in more streamlined output
8498 but retains the type-safety checking possibilities of the original
8502 *) The STACK code has been cleaned up, and certain type declarations
8503 that didn't make a lot of sense have been brought in line. This has
8504 also involved a cleanup of sorts in safestack.h to more correctly
8505 map type-safe stack functions onto their plain stack counterparts.
8506 This work has also resulted in a variety of "const"ifications of
8507 lots of the code, especially "_cmp" operations which should normally
8508 be prototyped with "const" parameters anyway.
8511 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8512 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8513 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8514 where all of 'md' is used each time the PRNG is used, but 'state'
8515 is used only indexed by a cyclic counter. As entropy may not be
8516 well distributed from the beginning, 'md' is important as a
8517 chaining variable. However, the output function chains only half
8518 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8519 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8520 in all of 'state' being rewritten, with the new values depending
8521 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8524 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8525 the handshake is continued after ssl_verify_cert_chain();
8526 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8527 can lead to 'unexplainable' connection aborts later.
8528 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8530 *) Major EVP API cipher revision.
8531 Add hooks for extra EVP features. This allows various cipher
8532 parameters to be set in the EVP interface. Support added for variable
8533 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8534 setting of RC2 and RC5 parameters.
8536 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8539 Remove lots of duplicated code from the EVP library. For example *every*
8540 cipher init() function handles the 'iv' in the same way according to the
8541 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8542 for CFB and OFB modes they zero ctx->num.
8544 New functionality allows removal of S/MIME code RC2 hack.
8546 Most of the routines have the same form and so can be declared in terms
8549 By shifting this to the top level EVP_CipherInit() it can be removed from
8550 all individual ciphers. If the cipher wants to handle IVs or keys
8551 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8554 Change lots of functions like EVP_EncryptUpdate() to now return a
8555 value: although software versions of the algorithms cannot fail
8556 any installed hardware versions can.
8559 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8560 this option is set, tolerate broken clients that send the negotiated
8561 protocol version number instead of the requested protocol version
8565 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8566 i.e. non-zero for export ciphersuites, zero otherwise.
8567 Previous versions had this flag inverted, inconsistent with
8568 rsa_tmp_cb (..._TMP_RSA_CB).
8569 [Bodo Moeller; problem reported by Amit Chopra]
8571 *) Add missing DSA library text string. Work around for some IIS
8572 key files with invalid SEQUENCE encoding.
8575 *) Add a document (doc/standards.txt) that list all kinds of standards
8576 and so on that are implemented in OpenSSL.
8579 *) Enhance c_rehash script. Old version would mishandle certificates
8580 with the same subject name hash and wouldn't handle CRLs at all.
8581 Added -fingerprint option to crl utility, to support new c_rehash
8585 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8588 *) Fix for SSL server purpose checking. Server checking was
8589 rejecting certificates which had extended key usage present
8590 but no ssl client purpose.
8591 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8593 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8594 is a little unclear about how a blank password is handled.
8595 Since the password in encoded as a BMPString with terminating
8596 double NULL a zero length password would end up as just the
8597 double NULL. However no password at all is different and is
8598 handled differently in the PKCS#12 key generation code. NS
8599 treats a blank password as zero length. MSIE treats it as no
8600 password on export: but it will try both on import. We now do
8601 the same: PKCS12_parse() tries zero length and no password if
8602 the password is set to "" or NULL (NULL is now a valid password:
8603 it wasn't before) as does the pkcs12 application.
8606 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8607 perror when PEM_read_bio_X509_REQ fails, the error message must
8608 be obtained from the error queue.
8611 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8612 it in ERR_remove_state if appropriate, and change ERR_get_state
8613 accordingly to avoid race conditions (this is necessary because
8614 thread_hash is no longer constant once set).
8617 *) Bugfix for linux-elf makefile.one.
8620 *) RSA_get_default_method() will now cause a default
8621 RSA_METHOD to be chosen if one doesn't exist already.
8622 Previously this was only set during a call to RSA_new()
8623 or RSA_new_method(NULL) meaning it was possible for
8624 RSA_get_default_method() to return NULL.
8627 *) Added native name translation to the existing DSO code
8628 that will convert (if the flag to do so is set) filenames
8629 that are sufficiently small and have no path information
8630 into a canonical native form. Eg. "blah" converted to
8631 "libblah.so" or "blah.dll" etc.
8634 *) New function ERR_error_string_n(e, buf, len) which is like
8635 ERR_error_string(e, buf), but writes at most 'len' bytes
8636 including the 0 terminator. For ERR_error_string_n, 'buf'
8638 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8640 *) CONF library reworked to become more general. A new CONF
8641 configuration file reader "class" is implemented as well as a
8642 new functions (NCONF_*, for "New CONF") to handle it. The now
8643 old CONF_* functions are still there, but are reimplemented to
8644 work in terms of the new functions. Also, a set of functions
8645 to handle the internal storage of the configuration data is
8646 provided to make it easier to write new configuration file
8647 reader "classes" (I can definitely see something reading a
8648 configuration file in XML format, for example), called _CONF_*,
8649 or "the configuration storage API"...
8651 The new configuration file reading functions are:
8653 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8654 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8656 NCONF_default, NCONF_WIN32
8658 NCONF_dump_fp, NCONF_dump_bio
8660 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8661 NCONF_new creates a new CONF object. This works in the same way
8662 as other interfaces in OpenSSL, like the BIO interface.
8663 NCONF_dump_* dump the internal storage of the configuration file,
8664 which is useful for debugging. All other functions take the same
8665 arguments as the old CONF_* functions wth the exception of the
8666 first that must be a `CONF *' instead of a `LHASH *'.
8668 To make it easer to use the new classes with the old CONF_* functions,
8669 the function CONF_set_default_method is provided.
8672 *) Add '-tls1' option to 'openssl ciphers', which was already
8673 mentioned in the documentation but had not been implemented.
8674 (This option is not yet really useful because even the additional
8675 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8678 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8679 OpenSSL-based applications) load shared libraries and bind to
8680 them in a portable way.
8681 [Geoff Thorpe, with contributions from Richard Levitte]
8683 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8685 *) Make sure _lrotl and _lrotr are only used with MSVC.
8687 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8688 (the default implementation of RAND_status).
8690 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8691 to '-clrext' (= clear extensions), as intended and documented.
8692 [Bodo Moeller; inconsistency pointed out by Michael Attili
8695 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8696 was larger than the MD block size.
8697 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8699 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8700 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8701 using the passed key: if the passed key was a private key the result
8702 of X509_print(), for example, would be to print out all the private key
8706 *) des_quad_cksum() byte order bug fix.
8707 [Ulf Möller, using the problem description in krb4-0.9.7, where
8708 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8710 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8712 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8714 *) For easily testing in shell scripts whether some command
8715 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8716 returns with exit code 0 iff no command of the given name is available.
8717 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8718 the output goes to stdout and nothing is printed to stderr.
8719 Additional arguments are always ignored.
8721 Since for each cipher there is a command of the same name,
8722 the 'no-cipher' compilation switches can be tested this way.
8724 ('openssl no-XXX' is not able to detect pseudo-commands such
8725 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8728 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8731 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8732 is set; it will be thrown away anyway because each handshake creates
8734 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8735 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8736 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
8737 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8740 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8741 'Q' and 'R' lose their special meanings (quit/renegotiate).
8742 This is part of what -quiet does; unlike -quiet, -ign_eof
8743 does not suppress any output.
8746 *) Add compatibility options to the purpose and trust code. The
8747 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8748 accepts a certificate or CA, this was the previous behaviour,
8749 with all the associated security issues.
8751 X509_TRUST_COMPAT is the old trust behaviour: only and
8752 automatically trust self signed roots in certificate store. A
8753 new trust setting X509_TRUST_DEFAULT is used to specify that
8754 a purpose has no associated trust setting and it should instead
8755 use the value in the default purpose.
8758 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8759 and fix a memory leak.
8762 *) In util/mkerr.pl (which implements 'make errors'), preserve
8763 reason strings from the previous version of the .c file, as
8764 the default to have only downcase letters (and digits) in
8765 automatically generated reasons codes is not always appropriate.
8768 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8769 using strerror. Previously, ERR_reason_error_string() returned
8770 library names as reason strings for SYSerr; but SYSerr is a special
8771 case where small numbers are errno values, not library numbers.
8774 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8775 converts DSA parameters into DH parameters. (When creating parameters,
8776 DSA_generate_parameters is used.)
8779 *) Include 'length' (recommended exponent length) in C code generated
8780 by 'openssl dhparam -C'.
8783 *) The second argument to set_label in perlasm was already being used
8784 so couldn't be used as a "file scope" flag. Moved to third argument
8788 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8789 instead of RAND_bytes for encryption IVs and salts.
8792 *) Include RAND_status() into RAND_METHOD instead of implementing
8793 it only for md_rand.c Otherwise replacing the PRNG by calling
8794 RAND_set_rand_method would be impossible.
8797 *) Don't let DSA_generate_key() enter an infinite loop if the random
8798 number generation fails.
8801 *) New 'rand' application for creating pseudo-random output.
8804 *) Added configuration support for Linux/IA64
8805 [Rolf Haberrecker <rolf@suse.de>]
8807 *) Assembler module support for Mingw32.
8810 *) Shared library support for HPUX (in shlib/).
8811 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8813 *) Shared library support for Solaris gcc.
8814 [Lutz Behnke <behnke@trustcenter.de>]
8816 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8818 *) PKCS7_encrypt() was adding text MIME headers twice because they
8819 were added manually and by SMIME_crlf_copy().
8822 *) In bntest.c don't call BN_rand with zero bits argument.
8823 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8825 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8826 case was implemented. This caused BN_div_recp() to fail occasionally.
8829 *) Add an optional second argument to the set_label() in the perl
8830 assembly language builder. If this argument exists and is set
8831 to 1 it signals that the assembler should use a symbol whose
8832 scope is the entire file, not just the current function. This
8833 is needed with MASM which uses the format label:: for this scope.
8834 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8836 *) Change the ASN1 types so they are typedefs by default. Before
8837 almost all types were #define'd to ASN1_STRING which was causing
8838 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8842 *) Change names of new functions to the new get1/get0 naming
8843 convention: After 'get1', the caller owns a reference count
8844 and has to call ..._free; 'get0' returns a pointer to some
8845 data structure without incrementing reference counters.
8846 (Some of the existing 'get' functions increment a reference
8847 counter, some don't.)
8848 Similarly, 'set1' and 'add1' functions increase reference
8849 counters or duplicate objects.
8852 *) Allow for the possibility of temp RSA key generation failure:
8853 the code used to assume it always worked and crashed on failure.
8856 *) Fix potential buffer overrun problem in BIO_printf().
8857 [Ulf Möller, using public domain code by Patrick Powell; problem
8858 pointed out by David Sacerdote <das33@cornell.edu>]
8860 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8861 RAND_egd() and RAND_status(). In the command line application,
8862 the EGD socket can be specified like a seed file using RANDFILE
8866 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8867 Some CAs (e.g. Verisign) distribute certificates in this form.
8870 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8871 list to exclude them. This means that no special compilation option
8872 is needed to use anonymous DH: it just needs to be included in the
8876 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8877 EVP_MD_type. The old functionality is available in a new macro called
8878 EVP_MD_md(). Change code that uses it and update docs.
8881 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8882 where the 'void *' argument is replaced by a function pointer argument.
8883 Previously 'void *' was abused to point to functions, which works on
8884 many platforms, but is not correct. As these functions are usually
8885 called by macros defined in OpenSSL header files, most source code
8886 should work without changes.
8889 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8890 sections with information on -D... compiler switches used for
8891 compiling the library so that applications can see them. To enable
8892 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8893 must be defined. E.g.,
8894 #define OPENSSL_ALGORITHM_DEFINES
8895 #include <openssl/opensslconf.h>
8896 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8897 [Richard Levitte, Ulf and Bodo Möller]
8899 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8903 *) Change the 'other' type in certificate aux info to a STACK_OF
8904 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8905 the required ASN1 format: arbitrary types determined by an OID.
8908 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8909 argument to 'req'. This is not because the function is newer or
8910 better than others it just uses the work 'NEW' in the certificate
8911 request header lines. Some software needs this.
8914 *) Reorganise password command line arguments: now passwords can be
8915 obtained from various sources. Delete the PEM_cb function and make
8916 it the default behaviour: i.e. if the callback is NULL and the
8917 usrdata argument is not NULL interpret it as a null terminated pass
8918 phrase. If usrdata and the callback are NULL then the pass phrase
8919 is prompted for as usual.
8922 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8923 the support is automatically enabled. The resulting binaries will
8924 autodetect the card and use it if present.
8925 [Ben Laurie and Compaq Inc.]
8927 *) Work around for Netscape hang bug. This sends certificate request
8928 and server done in one record. Since this is perfectly legal in the
8929 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8930 the bugs/SSLv3 entry for more info.
8933 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8936 *) Add -rand argument to smime and pkcs12 applications and read/write
8940 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8943 *) Add command line password options to the remaining applications.
8946 *) Bug fix for BN_div_recp() for numerators with an even number of
8950 *) More tests in bntest.c, and changed test_bn output.
8953 *) ./config recognizes MacOS X now.
8956 *) Bug fix for BN_div() when the first words of num and divsor are
8957 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8960 *) Add support for various broken PKCS#8 formats, and command line
8961 options to produce them.
8964 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8965 get temporary BIGNUMs from a BN_CTX.
8968 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8972 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8973 include a #define from the old name to the new. The original intent
8974 was that statically linked binaries could for example just call
8975 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8976 link with digests. This never worked becayse SSLeay_add_all_digests()
8977 and SSLeay_add_all_ciphers() were in the same source file so calling
8978 one would link with the other. They are now in separate source files.
8981 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8984 *) Use a less unusual form of the Miller-Rabin primality test (it used
8985 a binary algorithm for exponentiation integrated into the Miller-Rabin
8986 loop, our standard modexp algorithms are faster).
8989 *) Support for the EBCDIC character set completed.
8990 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8992 *) Source code cleanups: use const where appropriate, eliminate casts,
8993 use void * instead of char * in lhash.
8996 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8997 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8998 this the server could overwrite ephemeral keys that the client
9002 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
9003 using 50 iterations of the Rabin-Miller test.
9005 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
9006 iterations of the Rabin-Miller test as required by the appendix
9007 to FIPS PUB 186[-1]) instead of DSA_is_prime.
9008 As BN_is_prime_fasttest includes trial division, DSA parameter
9009 generation becomes much faster.
9011 This implies a change for the callback functions in DSA_is_prime
9012 and DSA_generate_parameters: The callback function is called once
9013 for each positive witness in the Rabin-Miller test, not just
9014 occasionally in the inner loop; and the parameters to the
9015 callback function now provide an iteration count for the outer
9016 loop rather than for the current invocation of the inner loop.
9017 DSA_generate_parameters additionally can call the callback
9018 function with an 'iteration count' of -1, meaning that a
9019 candidate has passed the trial division test (when q is generated
9020 from an application-provided seed, trial division is skipped).
9023 *) New function BN_is_prime_fasttest that optionally does trial
9024 division before starting the Rabin-Miller test and has
9025 an additional BN_CTX * argument (whereas BN_is_prime always
9026 has to allocate at least one BN_CTX).
9027 'callback(1, -1, cb_arg)' is called when a number has passed the
9028 trial division stage.
9031 *) Fix for bug in CRL encoding. The validity dates weren't being handled
9035 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
9038 *) New function BN_pseudo_rand().
9041 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
9042 bignum version of BN_from_montgomery() with the working code from
9043 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
9047 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
9048 made it impossible to use the same SSL_SESSION data structure in
9049 SSL2 clients in multiple threads.
9052 *) The return value of RAND_load_file() no longer counts bytes obtained
9053 by stat(). RAND_load_file(..., -1) is new and uses the complete file
9054 to seed the PRNG (previously an explicit byte count was required).
9055 [Ulf Möller, Bodo Möller]
9057 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
9058 used (char *) instead of (void *) and had casts all over the place.
9061 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
9064 *) Retain source code compatibility for BN_prime_checks macro:
9065 BN_is_prime(..., BN_prime_checks, ...) now uses
9066 BN_prime_checks_for_size to determine the appropriate number of
9067 Rabin-Miller iterations.
9070 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
9071 DH_CHECK_P_NOT_SAFE_PRIME.
9072 (Check if this is true? OpenPGP calls them "strong".)
9075 *) Merge the functionality of "dh" and "gendh" programs into a new program
9076 "dhparam". The old programs are retained for now but will handle DH keys
9077 (instead of parameters) in future.
9080 *) Make the ciphers, s_server and s_client programs check the return values
9081 when a new cipher list is set.
9084 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9085 ciphers. Before when the 56bit ciphers were enabled the sorting was
9088 The syntax for the cipher sorting has been extended to support sorting by
9089 cipher-strength (using the strength_bits hard coded in the tables).
9090 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9092 Fix a bug in the cipher-command parser: when supplying a cipher command
9093 string with an "undefined" symbol (neither command nor alphanumeric
9094 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9095 an error is flagged.
9097 Due to the strength-sorting extension, the code of the
9098 ssl_create_cipher_list() function was completely rearranged. I hope that
9099 the readability was also increased :-)
9100 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9102 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9103 for the first serial number and places 2 in the serial number file. This
9104 avoids problems when the root CA is created with serial number zero and
9105 the first user certificate has the same issuer name and serial number
9109 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9110 the new code. Add documentation for this stuff.
9113 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9114 X509_*() to X509at_*() on the grounds that they don't handle X509
9115 structures and behave in an analagous way to the X509v3 functions:
9116 they shouldn't be called directly but wrapper functions should be used
9119 So we also now have some wrapper functions that call the X509at functions
9120 when passed certificate requests. (TO DO: similar things can be done with
9121 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9122 things. Some of these need some d2i or i2d and print functionality
9123 because they handle more complex structures.)
9126 *) Add missing #ifndefs that caused missing symbols when building libssl
9127 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9128 NO_RSA in ssl/s2*.c.
9129 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9131 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9132 has a return value which indicates the quality of the random data
9133 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9134 error queue. New function RAND_pseudo_bytes() generates output that is
9135 guaranteed to be unique but not unpredictable. RAND_add is like
9136 RAND_seed, but takes an extra argument for an entropy estimate
9137 (RAND_seed always assumes full entropy).
9140 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9141 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9142 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9143 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9144 false-positive rate of at most 2^-80 for random input.
9147 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9150 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9151 in the 0.9.5 release), this returns the chain
9152 from an X509_CTX structure with a dup of the stack and all
9153 the X509 reference counts upped: so the stack will exist
9154 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9157 Also make SSL_SESSION_print() print out the verify return
9161 *) Add manpage for the pkcs12 command. Also change the default
9162 behaviour so MAC iteration counts are used unless the new
9163 -nomaciter option is used. This improves file security and
9164 only older versions of MSIE (4.0 for example) need it.
9167 *) Honor the no-xxx Configure options when creating .DEF files.
9170 *) Add PKCS#10 attributes to field table: challengePassword,
9171 unstructuredName and unstructuredAddress. These are taken from
9172 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9173 international characters are used.
9175 More changes to X509_ATTRIBUTE code: allow the setting of types
9176 based on strings. Remove the 'loc' parameter when adding
9177 attributes because these will be a SET OF encoding which is sorted
9181 *) Initial changes to the 'req' utility to allow request generation
9182 automation. This will allow an application to just generate a template
9183 file containing all the field values and have req construct the
9186 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9187 used all over the place including certificate requests and PKCS#7
9188 structures. They are currently handled manually where necessary with
9189 some primitive wrappers for PKCS#7. The new functions behave in a
9190 manner analogous to the X509 extension functions: they allow
9191 attributes to be looked up by NID and added.
9193 Later something similar to the X509V3 code would be desirable to
9194 automatically handle the encoding, decoding and printing of the
9195 more complex types. The string types like challengePassword can
9196 be handled by the string table functions.
9198 Also modified the multi byte string table handling. Now there is
9199 a 'global mask' which masks out certain types. The table itself
9200 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9201 is useful when for example there is only one permissible type
9202 (as in countryName) and using the mask might result in no valid
9206 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9207 SSL_get_peer_finished to allow applications to obtain the latest
9208 Finished messages sent to the peer or expected from the peer,
9209 respectively. (SSL_get_peer_finished is usually the Finished message
9210 actually received from the peer, otherwise the protocol will be aborted.)
9212 As the Finished message are message digests of the complete handshake
9213 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9214 be used for external authentication procedures when the authentication
9215 provided by SSL/TLS is not desired or is not enough.
9218 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9219 the host supports BWX extension and if Compaq C is present on the
9220 $PATH. Just exploiting of the BWX extension results in 20-30%
9221 performance kick for some algorithms, e.g. DES and RC4 to mention
9222 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9226 *) Add support for MS "fast SGC". This is arguably a violation of the
9227 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9228 weak crypto and after checking the certificate is SGC a second one
9229 with strong crypto. MS SGC stops the first handshake after receiving
9230 the server certificate message and sends a second client hello. Since
9231 a server will typically do all the time consuming operations before
9232 expecting any further messages from the client (server key exchange
9233 is the most expensive) there is little difference between the two.
9235 To get OpenSSL to support MS SGC we have to permit a second client
9236 hello message after we have sent server done. In addition we have to
9237 reset the MAC if we do get this second client hello.
9240 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9241 if a DER encoded private key is RSA or DSA traditional format. Changed
9242 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9243 format DER encoded private key. Newer code should use PKCS#8 format which
9244 has the key type encoded in the ASN1 structure. Added DER private key
9245 support to pkcs8 application.
9248 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9249 ciphersuites has been selected (as required by the SSL 3/TLS 1
9250 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9251 is set, we interpret this as a request to violate the specification
9252 (the worst that can happen is a handshake failure, and 'correct'
9253 behaviour would result in a handshake failure anyway).
9256 *) In SSL_CTX_add_session, take into account that there might be multiple
9257 SSL_SESSION structures with the same session ID (e.g. when two threads
9258 concurrently obtain them from an external cache).
9259 The internal cache can handle only one SSL_SESSION with a given ID,
9260 so if there's a conflict, we now throw out the old one to achieve
9264 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9265 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9266 some routines that use cipher OIDs: some ciphers do not have OIDs
9267 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9271 *) Simplify the trust setting structure and code. Now we just have
9272 two sequences of OIDs for trusted and rejected settings. These will
9273 typically have values the same as the extended key usage extension
9274 and any application specific purposes.
9276 The trust checking code now has a default behaviour: it will just
9277 check for an object with the same NID as the passed id. Functions can
9278 be provided to override either the default behaviour or the behaviour
9279 for a given id. SSL client, server and email already have functions
9280 in place for compatibility: they check the NID and also return "trusted"
9281 if the certificate is self signed.
9284 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9285 traditional format into an EVP_PKEY structure.
9288 *) Add a password callback function PEM_cb() which either prompts for
9289 a password if usr_data is NULL or otherwise assumes it is a null
9290 terminated password. Allow passwords to be passed on command line
9291 environment or config files in a few more utilities.
9294 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9295 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9296 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9297 Update documentation.
9300 *) Support for ASN1 "NULL" type. This could be handled before by using
9301 ASN1_TYPE but there wasn't any function that would try to read a NULL
9302 and produce an error if it couldn't. For compatibility we also have
9303 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9304 don't allocate anything because they don't need to.
9307 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9309 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9311 *) Rebuild of the memory allocation routines used by OpenSSL code and
9312 possibly others as well. The purpose is to make an interface that
9313 provide hooks so anyone can build a separate set of allocation and
9314 deallocation routines to be used by OpenSSL, for example memory
9315 pool implementations, or something else, which was previously hard
9316 since Malloc(), Realloc() and Free() were defined as macros having
9317 the values malloc, realloc and free, respectively (except for Win32
9318 compilations). The same is provided for memory debugging code.
9319 OpenSSL already comes with functionality to find memory leaks, but
9320 this gives people a chance to debug other memory problems.
9322 With these changes, a new set of functions and macros have appeared:
9324 CRYPTO_set_mem_debug_functions() [F]
9325 CRYPTO_get_mem_debug_functions() [F]
9326 CRYPTO_dbg_set_options() [F]
9327 CRYPTO_dbg_get_options() [F]
9328 CRYPTO_malloc_debug_init() [M]
9330 The memory debug functions are NULL by default, unless the library
9331 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9332 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9333 gives the standard debugging functions that come with OpenSSL) or
9334 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9335 provided by the library user) must be used. When the standard
9336 debugging functions are used, CRYPTO_dbg_set_options can be used to
9337 request additional information:
9338 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9339 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9341 Also, things like CRYPTO_set_mem_functions will always give the
9342 expected result (the new set of functions is used for allocation
9343 and deallocation) at all times, regardless of platform and compiler
9346 To finish it up, some functions that were never use in any other
9347 way than through macros have a new API and new semantic:
9350 CRYPTO_dbg_realloc()
9353 All macros of value have retained their old syntax.
9354 [Richard Levitte and Bodo Moeller]
9356 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9357 ordering of SMIMECapabilities wasn't in "strength order" and there
9358 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9362 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9363 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9364 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9366 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9367 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9368 functionality to handle multipart/signed properly) and a utility
9369 called 'smime' to call all this stuff. This is based on code I
9370 originally wrote for Celo who have kindly allowed it to be
9371 included in OpenSSL.
9374 *) Add variants des_set_key_checked and des_set_key_unchecked of
9375 des_set_key (aka des_key_sched). Global variable des_check_key
9376 decides which of these is called by des_set_key; this way
9377 des_check_key behaves as it always did, but applications and
9378 the library itself, which was buggy for des_check_key == 1,
9379 have a cleaner way to pick the version they need.
9382 *) New function PKCS12_newpass() which changes the password of a
9386 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9387 dynamic mix. In both cases the ids can be used as an index into the
9388 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9389 functions so they accept a list of the field values and the
9390 application doesn't need to directly manipulate the X509_TRUST
9394 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9398 *) Modify the way the V3 extension code looks up extensions. This now
9399 works in a similar way to the object code: we have some "standard"
9400 extensions in a static table which is searched with OBJ_bsearch()
9401 and the application can add dynamic ones if needed. The file
9402 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9403 updated whenever a new extension is added to the core code and kept
9404 in ext_nid order. There is a simple program 'tabtest.c' which checks
9405 this. New extensions are not added too often so this file can readily
9406 be maintained manually.
9408 There are two big advantages in doing things this way. The extensions
9409 can be looked up immediately and no longer need to be "added" using
9410 X509V3_add_standard_extensions(): this function now does nothing.
9411 [Side note: I get *lots* of email saying the extension code doesn't
9412 work because people forget to call this function]
9413 Also no dynamic allocation is done unless new extensions are added:
9414 so if we don't add custom extensions there is no need to call
9415 X509V3_EXT_cleanup().
9418 *) Modify enc utility's salting as follows: make salting the default. Add a
9419 magic header, so unsalted files fail gracefully instead of just decrypting
9420 to garbage. This is because not salting is a big security hole, so people
9421 should be discouraged from doing it.
9424 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9425 digest to be passed on the command line but it only used this
9426 parameter when signing a certificate. Modified so all relevant
9427 operations are affected by the digest parameter including the
9428 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9429 DSA key was used because it didn't fix the digest.
9432 *) Initial certificate chain verify code. Currently tests the untrusted
9433 certificates for consistency with the verify purpose (which is set
9434 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9436 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9437 this is because it will reject chains with invalid extensions whereas
9438 every previous version of OpenSSL and SSLeay made no checks at all.
9440 Trust code: checks the root CA for the relevant trust settings. Trust
9441 settings have an initial value consistent with the verify purpose: e.g.
9442 if the verify purpose is for SSL client use it expects the CA to be
9443 trusted for SSL client use. However the default value can be changed to
9444 permit custom trust settings: one example of this would be to only trust
9445 certificates from a specific "secure" set of CAs.
9447 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9448 which should be used for version portability: especially since the
9449 verify structure is likely to change more often now.
9451 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9452 to set them. If not set then assume SSL clients will verify SSL servers
9455 Two new options to the verify program: -untrusted allows a set of
9456 untrusted certificates to be passed in and -purpose which sets the
9457 intended purpose of the certificate. If a purpose is set then the
9458 new chain verify code is used to check extension consistency.
9461 *) Support for the authority information access extension.
9464 *) Modify RSA and DSA PEM read routines to transparently handle
9465 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9466 public keys in a format compatible with certificate
9467 SubjectPublicKeyInfo structures. Unfortunately there were already
9468 functions called *_PublicKey_* which used various odd formats so
9469 these are retained for compatibility: however the DSA variants were
9470 never in a public release so they have been deleted. Changed dsa/rsa
9471 utilities to handle the new format: note no releases ever handled public
9472 keys so we should be OK.
9474 The primary motivation for this change is to avoid the same fiasco
9475 that dogs private keys: there are several incompatible private key
9476 formats some of which are standard and some OpenSSL specific and
9477 require various evil hacks to allow partial transparent handling and
9478 even then it doesn't work with DER formats. Given the option anything
9479 other than PKCS#8 should be dumped: but the other formats have to
9480 stay in the name of compatibility.
9482 With public keys and the benefit of hindsight one standard format
9483 is used which works with EVP_PKEY, RSA or DSA structures: though
9484 it clearly returns an error if you try to read the wrong kind of key.
9486 Added a -pubkey option to the 'x509' utility to output the public key.
9487 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9488 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9489 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9490 that do the same as the EVP_PKEY_assign_*() except they up the
9491 reference count of the added key (they don't "swallow" the
9495 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9496 CRLs would fail if the file contained no certificates or no CRLs:
9497 added a new function to read in both types and return the number
9498 read: this means that if none are read it will be an error. The
9499 DER versions of the certificate and CRL reader would always fail
9500 because it isn't possible to mix certificates and CRLs in DER format
9501 without choking one or the other routine. Changed this to just read
9502 a certificate: this is the best we can do. Also modified the code
9503 in apps/verify.c to take notice of return codes: it was previously
9504 attempting to read in certificates from NULL pointers and ignoring
9505 any errors: this is one reason why the cert and CRL reader seemed
9506 to work. It doesn't check return codes from the default certificate
9507 routines: these may well fail if the certificates aren't installed.
9510 *) Code to support otherName option in GeneralName.
9513 *) First update to verify code. Change the verify utility
9514 so it warns if it is passed a self signed certificate:
9515 for consistency with the normal behaviour. X509_verify
9516 has been modified to it will now verify a self signed
9517 certificate if *exactly* the same certificate appears
9518 in the store: it was previously impossible to trust a
9519 single self signed certificate. This means that:
9520 openssl verify ss.pem
9521 now gives a warning about a self signed certificate but
9522 openssl verify -CAfile ss.pem ss.pem
9526 *) For servers, store verify_result in SSL_SESSION data structure
9527 (and add it to external session representation).
9528 This is needed when client certificate verifications fails,
9529 but an application-provided verification callback (set by
9530 SSL_CTX_set_cert_verify_callback) allows accepting the session
9531 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9532 but returns 1): When the session is reused, we have to set
9533 ssl->verify_result to the appropriate error code to avoid
9535 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9537 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9538 case in PKCS7_dataInit() where the signed PKCS7 structure
9539 didn't contain any existing data because it was being created.
9540 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9542 *) Add a salt to the key derivation routines in enc.c. This
9543 forms the first 8 bytes of the encrypted file. Also add a
9544 -S option to allow a salt to be input on the command line.
9547 *) New function X509_cmp(). Oddly enough there wasn't a function
9548 to compare two certificates. We do this by working out the SHA1
9549 hash and comparing that. X509_cmp() will be needed by the trust
9553 *) SSL_get1_session() is like SSL_get_session(), but increments
9554 the reference count in the SSL_SESSION returned.
9555 [Geoff Thorpe <geoff@eu.c2.net>]
9557 *) Fix for 'req': it was adding a null to request attributes.
9558 Also change the X509_LOOKUP and X509_INFO code to handle
9559 certificate auxiliary information.
9562 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9566 *) Add the possibility to add extra information to the memory leak
9567 detecting output, to form tracebacks, showing from where each
9568 allocation was originated: CRYPTO_push_info("constant string") adds
9569 the string plus current file name and line number to a per-thread
9570 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9571 is like calling CYRPTO_pop_info() until the stack is empty.
9572 Also updated memory leak detection code to be multi-thread-safe.
9575 *) Add options -text and -noout to pkcs7 utility and delete the
9576 encryption options which never did anything. Update docs.
9579 *) Add options to some of the utilities to allow the pass phrase
9580 to be included on either the command line (not recommended on
9581 OSes like Unix) or read from the environment. Update the
9582 manpages and fix a few bugs.
9585 *) Add a few manpages for some of the openssl commands.
9588 *) Fix the -revoke option in ca. It was freeing up memory twice,
9589 leaking and not finding already revoked certificates.
9592 *) Extensive changes to support certificate auxiliary information.
9593 This involves the use of X509_CERT_AUX structure and X509_AUX
9594 functions. An X509_AUX function such as PEM_read_X509_AUX()
9595 can still read in a certificate file in the usual way but it
9596 will also read in any additional "auxiliary information". By
9597 doing things this way a fair degree of compatibility can be
9598 retained: existing certificates can have this information added
9599 using the new 'x509' options.
9601 Current auxiliary information includes an "alias" and some trust
9602 settings. The trust settings will ultimately be used in enhanced
9603 certificate chain verification routines: currently a certificate
9604 can only be trusted if it is self signed and then it is trusted
9608 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9609 The problem was that one of the replacement routines had not been working
9610 since SSLeay releases. For now the offending routine has been replaced
9611 with non-optimised assembler. Even so, this now gives around 95%
9612 performance improvement for 1024 bit RSA signs.
9615 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9616 handling. Most clients have the effective key size in bits equal to
9617 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9618 A few however don't do this and instead use the size of the decrypted key
9619 to determine the RC2 key length and the AlgorithmIdentifier to determine
9620 the effective key length. In this case the effective key length can still
9621 be 40 bits but the key length can be 168 bits for example. This is fixed
9622 by manually forcing an RC2 key into the EVP_PKEY structure because the
9623 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9624 the key length and effective key length are equal.
9627 *) Add a bunch of functions that should simplify the creation of
9628 X509_NAME structures. Now you should be able to do:
9629 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9630 and have it automatically work out the correct field type and fill in
9631 the structures. The more adventurous can try:
9632 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9633 and it will (hopefully) work out the correct multibyte encoding.
9636 *) Change the 'req' utility to use the new field handling and multibyte
9637 copy routines. Before the DN field creation was handled in an ad hoc
9638 way in req, ca, and x509 which was rather broken and didn't support
9639 BMPStrings or UTF8Strings. Since some software doesn't implement
9640 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9641 using the dirstring_type option. See the new comment in the default
9642 openssl.cnf for more info.
9645 *) Make crypto/rand/md_rand.c more robust:
9646 - Assure unique random numbers after fork().
9647 - Make sure that concurrent threads access the global counter and
9648 md serializably so that we never lose entropy in them
9649 or use exactly the same state in multiple threads.
9650 Access to the large state is not always serializable because
9651 the additional locking could be a performance killer, and
9652 md should be large enough anyway.
9655 *) New file apps/app_rand.c with commonly needed functionality
9656 for handling the random seed file.
9658 Use the random seed file in some applications that previously did not:
9660 dsaparam -genkey (which also ignored its '-rand' option),
9663 x509 (when signing).
9664 Except on systems with /dev/urandom, it is crucial to have a random
9665 seed file at least for key creation, DSA signing, and for DH exchanges;
9666 for RSA signatures we could do without one.
9668 gendh and gendsa (unlike genrsa) used to read only the first byte
9669 of each file listed in the '-rand' option. The function as previously
9670 found in genrsa is now in app_rand.c and is used by all programs
9671 that support '-rand'.
9674 *) In RAND_write_file, use mode 0600 for creating files;
9675 don't just chmod when it may be too late.
9678 *) Report an error from X509_STORE_load_locations
9679 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9682 *) New function ASN1_mbstring_copy() this copies a string in either
9683 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9684 into an ASN1_STRING type. A mask of permissible types is passed
9685 and it chooses the "minimal" type to use or an error if not type
9689 *) Add function equivalents to the various macros in asn1.h. The old
9690 macros are retained with an M_ prefix. Code inside the library can
9691 use the M_ macros. External code (including the openssl utility)
9692 should *NOT* in order to be "shared library friendly".
9695 *) Add various functions that can check a certificate's extensions
9696 to see if it usable for various purposes such as SSL client,
9697 server or S/MIME and CAs of these types. This is currently
9698 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9699 verification. Also added a -purpose flag to x509 utility to
9700 print out all the purposes.
9703 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9707 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9708 for, obtain and decode and extension and obtain its critical flag.
9709 This allows all the necessary extension code to be handled in a
9710 single function call.
9713 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9714 platforms. See crypto/rc4/rc4_enc.c for further details.
9717 *) New -noout option to asn1parse. This causes no output to be produced
9718 its main use is when combined with -strparse and -out to extract data
9719 from a file (which may not be in ASN.1 format).
9722 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9723 when producing the local key id.
9724 [Richard Levitte <levitte@stacken.kth.se>]
9726 *) New option -dhparam in s_server. This allows a DH parameter file to be
9727 stated explicitly. If it is not stated then it tries the first server
9728 certificate file. The previous behaviour hard coded the filename
9732 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9733 a public key to be input or output. For example:
9734 openssl rsa -in key.pem -pubout -out pubkey.pem
9735 Also added necessary DSA public key functions to handle this.
9738 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9739 in the message. This was handled by allowing
9740 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9741 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9743 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9744 to the end of the strings whereas this didn't. This would cause problems
9745 if strings read with d2i_ASN1_bytes() were later modified.
9746 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9748 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9749 data and it contains EOF it will end up returning an error. This is
9750 caused by input 46 bytes long. The cause is due to the way base64
9751 BIOs find the start of base64 encoded data. They do this by trying a
9752 trial decode on each line until they find one that works. When they
9753 do a flag is set and it starts again knowing it can pass all the
9754 data directly through the decoder. Unfortunately it doesn't reset
9755 the context it uses. This means that if EOF is reached an attempt
9756 is made to pass two EOFs through the context and this causes the
9757 resulting error. This can also cause other problems as well. As is
9758 usual with these problems it takes *ages* to find and the fix is
9759 trivial: move one line.
9760 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9762 *) Ugly workaround to get s_client and s_server working under Windows. The
9763 old code wouldn't work because it needed to select() on sockets and the
9764 tty (for keypresses and to see if data could be written). Win32 only
9765 supports select() on sockets so we select() with a 1s timeout on the
9766 sockets and then see if any characters are waiting to be read, if none
9767 are present then we retry, we also assume we can always write data to
9768 the tty. This isn't nice because the code then blocks until we've
9769 received a complete line of data and it is effectively polling the
9770 keyboard at 1s intervals: however it's quite a bit better than not
9771 working at all :-) A dedicated Windows application might handle this
9772 with an event loop for example.
9775 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9776 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9777 will be called when RSA_sign() and RSA_verify() are used. This is useful
9778 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9779 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9780 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9781 This necessitated the support of an extra signature type NID_md5_sha1
9782 for SSL signatures and modifications to the SSL library to use it instead
9783 of calling RSA_public_decrypt() and RSA_private_encrypt().
9786 *) Add new -verify -CAfile and -CApath options to the crl program, these
9787 will lookup a CRL issuers certificate and verify the signature in a
9788 similar way to the verify program. Tidy up the crl program so it
9789 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9790 less strict. It will now permit CRL extensions even if it is not
9791 a V2 CRL: this will allow it to tolerate some broken CRLs.
9794 *) Initialize all non-automatic variables each time one of the openssl
9795 sub-programs is started (this is necessary as they may be started
9796 multiple times from the "OpenSSL>" prompt).
9797 [Lennart Bang, Bodo Moeller]
9799 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9800 removing all other RSA functionality (this is what NO_RSA does). This
9801 is so (for example) those in the US can disable those operations covered
9802 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9806 *) Non-copying interface to BIO pairs.
9807 (still largely untested)
9810 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9811 ASCII string. This was handled independently in various places before.
9814 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9815 UTF8 strings a character at a time.
9818 *) Use client_version from client hello to select the protocol
9819 (s23_srvr.c) and for RSA client key exchange verification
9820 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9823 *) Add various utility functions to handle SPKACs, these were previously
9824 handled by poking round in the structure internals. Added new function
9825 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9826 print, verify and generate SPKACs. Based on an original idea from
9827 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9830 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9833 *) Allow the config file extension section to be overwritten on the
9834 command line. Based on an original idea from Massimiliano Pala
9835 <madwolf@comune.modena.it>. The new option is called -extensions
9836 and can be applied to ca, req and x509. Also -reqexts to override
9837 the request extensions in req and -crlexts to override the crl extensions
9841 *) Add new feature to the SPKAC handling in ca. Now you can include
9842 the same field multiple times by preceding it by "XXXX." for example:
9845 this is the same syntax as used in the req config file.
9848 *) Allow certificate extensions to be added to certificate requests. These
9849 are specified in a 'req_extensions' option of the req section of the
9850 config file. They can be printed out with the -text option to req but
9851 are otherwise ignored at present.
9854 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9855 data read consists of only the final block it would not decrypted because
9856 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9857 A misplaced 'break' also meant the decrypted final block might not be
9858 copied until the next read.
9861 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9862 a few extra parameters to the DH structure: these will be useful if
9863 for example we want the value of 'q' or implement X9.42 DH.
9866 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9867 provides hooks that allow the default DSA functions or functions on a
9868 "per key" basis to be replaced. This allows hardware acceleration and
9869 hardware key storage to be handled without major modification to the
9870 library. Also added low level modexp hooks and CRYPTO_EX structure and
9871 associated functions.
9874 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9875 as "read only": it can't be written to and the buffer it points to will
9876 not be freed. Reading from a read only BIO is much more efficient than
9877 a normal memory BIO. This was added because there are several times when
9878 an area of memory needs to be read from a BIO. The previous method was
9879 to create a memory BIO and write the data to it, this results in two
9880 copies of the data and an O(n^2) reading algorithm. There is a new
9881 function BIO_new_mem_buf() which creates a read only memory BIO from
9882 an area of memory. Also modified the PKCS#7 routines to use read only
9886 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9887 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9888 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9889 but a retry condition occured while trying to read the rest.
9892 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9893 NID_pkcs7_encrypted by default: this was wrong since this should almost
9894 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9895 the encrypted data type: this is a more sensible place to put it and it
9896 allows the PKCS#12 code to be tidied up that duplicated this
9900 *) Changed obj_dat.pl script so it takes its input and output files on
9901 the command line. This should avoid shell escape redirection problems
9905 *) Initial support for certificate extension requests, these are included
9906 in things like Xenroll certificate requests. Included functions to allow
9907 extensions to be obtained and added.
9910 *) -crlf option to s_client and s_server for sending newlines as
9911 CRLF (as required by many protocols).
9914 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9916 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9917 [Ralf S. Engelschall]
9919 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9920 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9922 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9926 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9927 DH parameters/keys (q is lost during that conversion, but the resulting
9928 DH parameters contain its length).
9930 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9931 much faster than DH_generate_parameters (which creates parameters
9932 where p = 2*q + 1), and also the smaller q makes DH computations
9933 much more efficient (160-bit exponentiation instead of 1024-bit
9934 exponentiation); so this provides a convenient way to support DHE
9935 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9936 utter importance to use
9937 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9939 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9940 when such DH parameters are used, because otherwise small subgroup
9941 attacks may become possible!
9944 *) Avoid memory leak in i2d_DHparams.
9947 *) Allow the -k option to be used more than once in the enc program:
9948 this allows the same encrypted message to be read by multiple recipients.
9951 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9952 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9953 it will always use the numerical form of the OID, even if it has a short
9957 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9958 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9959 otherwise bn_mod_exp was called. In the case of hardware keys for example
9960 no private key components need be present and it might store extra data
9961 in the RSA structure, which cannot be accessed from bn_mod_exp.
9962 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9963 private key operations.
9966 *) Added support for SPARC Linux.
9969 *) pem_password_cb function type incompatibly changed from
9970 typedef int pem_password_cb(char *buf, int size, int rwflag);
9972 ....(char *buf, int size, int rwflag, void *userdata);
9973 so that applications can pass data to their callbacks:
9974 The PEM[_ASN1]_{read,write}... functions and macros now take an
9975 additional void * argument, which is just handed through whenever
9976 the password callback is called.
9977 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9979 New function SSL_CTX_set_default_passwd_cb_userdata.
9981 Compatibility note: As many C implementations push function arguments
9982 onto the stack in reverse order, the new library version is likely to
9983 interoperate with programs that have been compiled with the old
9984 pem_password_cb definition (PEM_whatever takes some data that
9985 happens to be on the stack as its last argument, and the callback
9986 just ignores this garbage); but there is no guarantee whatsoever that
9989 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9990 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9991 problems not only on Windows, but also on some Unix platforms.
9992 To avoid problematic command lines, these definitions are now in an
9993 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9994 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9997 *) MIPS III/IV assembler module is reimplemented.
10000 *) More DES library cleanups: remove references to srand/rand and
10001 delete an unused file.
10004 *) Add support for the the free Netwide assembler (NASM) under Win32,
10005 since not many people have MASM (ml) and it can be hard to obtain.
10006 This is currently experimental but it seems to work OK and pass all
10007 the tests. Check out INSTALL.W32 for info.
10010 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
10011 without temporary keys kept an extra copy of the server key,
10012 and connections with temporary keys did not free everything in case
10016 *) New function RSA_check_key and new openssl rsa option -check
10017 for verifying the consistency of RSA keys.
10018 [Ulf Moeller, Bodo Moeller]
10020 *) Various changes to make Win32 compile work:
10021 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
10022 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
10023 comparison" warnings.
10024 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
10027 *) Add a debugging option to PKCS#5 v2 key generation function: when
10028 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
10029 derived keys are printed to stderr.
10032 *) Copy the flags in ASN1_STRING_dup().
10033 [Roman E. Pavlov <pre@mo.msk.ru>]
10035 *) The x509 application mishandled signing requests containing DSA
10036 keys when the signing key was also DSA and the parameters didn't match.
10038 It was supposed to omit the parameters when they matched the signing key:
10039 the verifying software was then supposed to automatically use the CA's
10040 parameters if they were absent from the end user certificate.
10042 Omitting parameters is no longer recommended. The test was also
10043 the wrong way round! This was probably due to unusual behaviour in
10044 EVP_cmp_parameters() which returns 1 if the parameters match.
10045 This meant that parameters were omitted when they *didn't* match and
10046 the certificate was useless. Certificates signed with 'ca' didn't have
10048 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
10050 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
10051 The interface is as follows:
10052 Applications can use
10053 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
10054 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
10055 "off" is now the default.
10056 The library internally uses
10057 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
10058 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
10059 to disable memory-checking temporarily.
10061 Some inconsistent states that previously were possible (and were
10062 even the default) are now avoided.
10064 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
10065 with each memory chunk allocated; this is occasionally more helpful
10066 than just having a counter.
10068 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
10070 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
10074 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
10075 which largely parallels "options", but is for changing API behaviour,
10076 whereas "options" are about protocol behaviour.
10077 Initial "mode" flags are:
10079 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
10080 a single record has been written.
10081 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
10082 retries use the same buffer location.
10083 (But all of the contents must be
10087 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10090 *) Fix problems with no-hmac etc.
10091 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10093 *) New functions RSA_get_default_method(), RSA_set_method() and
10094 RSA_get_method(). These allows replacement of RSA_METHODs without having
10095 to mess around with the internals of an RSA structure.
10098 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10099 Also really enable memory leak checks in openssl.c and in some
10101 [Chad C. Mulligan, Bodo Moeller]
10103 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10104 up the length of negative integers. This has now been simplified to just
10105 store the length when it is first determined and use it later, rather
10106 than trying to keep track of where data is copied and updating it to
10108 [Steve Henson, reported by Brien Wheeler
10109 <bwheeler@authentica-security.com>]
10111 *) Add a new function PKCS7_signatureVerify. This allows the verification
10112 of a PKCS#7 signature but with the signing certificate passed to the
10113 function itself. This contrasts with PKCS7_dataVerify which assumes the
10114 certificate is present in the PKCS#7 structure. This isn't always the
10115 case: certificates can be omitted from a PKCS#7 structure and be
10116 distributed by "out of band" means (such as a certificate database).
10119 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10120 function prototypes in pem.h, also change util/mkdef.pl to add the
10121 necessary function names.
10124 *) mk1mf.pl (used by Windows builds) did not properly read the
10125 options set by Configure in the top level Makefile, and Configure
10126 was not even able to write more than one option correctly.
10127 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10130 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10131 file to be loaded from a BIO or FILE pointer. The BIO version will
10132 for example allow memory BIOs to contain config info.
10135 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10136 Whoever hopes to achieve shared-library compatibility across versions
10137 must use this, not the compile-time macro.
10138 (Exercise 0.9.4: Which is the minimum library version required by
10140 Note: All this applies only to multi-threaded programs, others don't
10144 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10145 through a BIO pair triggered the default case, i.e.
10146 SSLerr(...,SSL_R_UNKNOWN_STATE).
10149 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10150 can use the SSL library even if none of the specific BIOs is
10154 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10155 for the encoded length.
10156 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10158 *) Add initial documentation of the X509V3 functions.
10161 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10162 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10163 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10164 secure PKCS#8 private key format with a high iteration count.
10167 *) Fix determination of Perl interpreter: A perl or perl5
10168 _directory_ in $PATH was also accepted as the interpreter.
10169 [Ralf S. Engelschall]
10171 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10172 wrong with it but it was very old and did things like calling
10173 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10174 unusual formatting.
10177 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10178 to use the new extension code.
10181 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10182 with macros. This should make it easier to change their form, add extra
10183 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10187 *) Add to configuration table a new entry that can specify an alternative
10188 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10189 according to Mark Crispin <MRC@Panda.COM>.
10193 *) DES CBC did not update the IV. Weird.
10196 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10197 Changing the behaviour of the former might break existing programs --
10198 where IV updating is needed, des_ncbc_encrypt can be used.
10201 *) When bntest is run from "make test" it drives bc to check its
10202 calculations, as well as internally checking them. If an internal check
10203 fails, it needs to cause bc to give a non-zero result or make test carries
10204 on without noticing the failure. Fixed.
10207 *) DES library cleanups.
10210 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10211 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10212 ciphers. NOTE: although the key derivation function has been verified
10213 against some published test vectors it has not been extensively tested
10214 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10218 *) Instead of "mkdir -p", which is not fully portable, use new
10219 Perl script "util/mkdir-p.pl".
10222 *) Rewrite the way password based encryption (PBE) is handled. It used to
10223 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10224 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10225 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10226 the 'parameter' field of the AlgorithmIdentifier is passed to the
10227 underlying key generation function so it must do its own ASN1 parsing.
10228 This has also changed the EVP_PBE_CipherInit() function which now has a
10229 'parameter' argument instead of literal salt and iteration count values
10230 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10233 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10234 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10235 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10236 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10237 value was just used as a "magic string" and not used directly its
10238 value doesn't matter.
10241 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10245 *) "linux-sparc64" configuration (ultrapenguin).
10246 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10247 "linux-sparc" configuration.
10248 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10250 *) config now generates no-xxx options for missing ciphers.
10253 *) Support the EBCDIC character set (work in progress).
10254 File ebcdic.c not yet included because it has a different license.
10255 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10257 *) Support BS2000/OSD-POSIX.
10258 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10260 *) Make callbacks for key generation use void * instead of char *.
10263 *) Make S/MIME samples compile (not yet tested).
10266 *) Additional typesafe stacks.
10269 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10273 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10275 *) New configuration variant "sco5-gcc".
10277 *) Updated some demos.
10278 [Sean O Riordain, Wade Scholine]
10280 *) Add missing BIO_free at exit of pkcs12 application.
10283 *) Fix memory leak in conf.c.
10286 *) Updates for Win32 to assembler version of MD5.
10289 *) Set #! path to perl in apps/der_chop to where we found it
10290 instead of using a fixed path.
10293 *) SHA library changes for irix64-mips4-cc.
10296 *) Improvements for VMS support.
10300 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10302 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10303 This also avoids the problems with SC4.2 and unpatched SC5.
10304 [Andy Polyakov <appro@fy.chalmers.se>]
10306 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10307 These are required because of the typesafe stack would otherwise break
10308 existing code. If old code used a structure member which used to be STACK
10309 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10310 sk_num or sk_value it would produce an error because the num, data members
10311 are not present in STACK_OF. Now it just produces a warning. sk_set
10312 replaces the old method of assigning a value to sk_value
10313 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10314 that does this will no longer work (and should use sk_set instead) but
10315 this could be regarded as a "questionable" behaviour anyway.
10318 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10319 correctly handle encrypted S/MIME data.
10322 *) Change type of various DES function arguments from des_cblock
10323 (which means, in function argument declarations, pointer to char)
10324 to des_cblock * (meaning pointer to array with 8 char elements),
10325 which allows the compiler to do more typechecking; it was like
10326 that back in SSLeay, but with lots of ugly casts.
10328 Introduce new type const_des_cblock.
10331 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10332 problems: find RecipientInfo structure that matches recipient certificate
10333 and initialise the ASN1 structures properly based on passed cipher.
10336 *) Belatedly make the BN tests actually check the results.
10339 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10340 to and from BNs: it was completely broken. New compilation option
10341 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10342 key elements as negative integers.
10345 *) Reorganize and speed up MD5.
10346 [Andy Polyakov <appro@fy.chalmers.se>]
10349 [Richard Levitte <richard@levitte.org>]
10351 *) New option -out to asn1parse to allow the parsed structure to be
10352 output to a file. This is most useful when combined with the -strparse
10353 option to examine the output of things like OCTET STRINGS.
10356 *) Make SSL library a little more fool-proof by not requiring any longer
10357 that SSL_set_{accept,connect}_state be called before
10358 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10359 in many applications because usually everything *appeared* to work as
10360 intended anyway -- now it really works as intended).
10363 *) Move openssl.cnf out of lib/.
10366 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10367 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10368 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10369 [Ralf S. Engelschall]
10371 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10372 handle PKCS#7 enveloped data properly.
10373 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10375 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10376 copying pointers. The cert_st handling is changed by this in
10377 various ways (and thus what used to be known as ctx->default_cert
10378 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10379 any longer when s->cert does not give us what we need).
10380 ssl_cert_instantiate becomes obsolete by this change.
10381 As soon as we've got the new code right (possibly it already is?),
10382 we have solved a couple of bugs of the earlier code where s->cert
10383 was used as if it could not have been shared with other SSL structures.
10385 Note that using the SSL API in certain dirty ways now will result
10386 in different behaviour than observed with earlier library versions:
10387 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10388 does not influence s as it used to.
10390 In order to clean up things more thoroughly, inside SSL_SESSION
10391 we don't use CERT any longer, but a new structure SESS_CERT
10392 that holds per-session data (if available); currently, this is
10393 the peer's certificate chain and, for clients, the server's certificate
10394 and temporary key. CERT holds only those values that can have
10395 meaningful defaults in an SSL_CTX.
10398 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10399 from the internal representation. Various PKCS#7 fixes: remove some
10400 evil casts and set the enc_dig_alg field properly based on the signing
10404 *) Allow PKCS#12 password to be set from the command line or the
10405 environment. Let 'ca' get its config file name from the environment
10406 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10410 *) Allow certificate policies extension to use an IA5STRING for the
10411 organization field. This is contrary to the PKIX definition but
10412 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10416 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10417 without disallowing inline assembler and the like for non-pedantic builds.
10420 *) Support Borland C++ builder.
10421 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10423 *) Support Mingw32.
10426 *) SHA-1 cleanups and performance enhancements.
10427 [Andy Polyakov <appro@fy.chalmers.se>]
10429 *) Sparc v8plus assembler for the bignum library.
10430 [Andy Polyakov <appro@fy.chalmers.se>]
10432 *) Accept any -xxx and +xxx compiler options in Configure.
10435 *) Update HPUX configuration.
10438 *) Add missing sk_<type>_unshift() function to safestack.h
10439 [Ralf S. Engelschall]
10441 *) New function SSL_CTX_use_certificate_chain_file that sets the
10442 "extra_cert"s in addition to the certificate. (This makes sense
10443 only for "PEM" format files, as chains as a whole are not
10447 *) Support verify_depth from the SSL API.
10448 x509_vfy.c had what can be considered an off-by-one-error:
10449 Its depth (which was not part of the external interface)
10450 was actually counting the number of certificates in a chain;
10451 now it really counts the depth.
10454 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10455 instead of X509err, which often resulted in confusing error
10456 messages since the error codes are not globally unique
10457 (e.g. an alleged error in ssl3_accept when a certificate
10458 didn't match the private key).
10460 *) New function SSL_CTX_set_session_id_context that allows to set a default
10461 value (so that you don't need SSL_set_session_id_context for each
10462 connection using the SSL_CTX).
10465 *) OAEP decoding bug fix.
10468 *) Support INSTALL_PREFIX for package builders, as proposed by
10472 *) New Configure options "threads" and "no-threads". For systems
10473 where the proper compiler options are known (currently Solaris
10474 and Linux), "threads" is the default.
10477 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10480 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10481 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10482 such as /usr/local/bin.
10485 *) "make linux-shared" to build shared libraries.
10486 [Niels Poppe <niels@netbox.org>]
10488 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10491 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10492 extension adding in x509 utility.
10495 *) Remove NOPROTO sections and error code comments.
10498 *) Partial rewrite of the DEF file generator to now parse the ANSI
10502 *) New Configure options --prefix=DIR and --openssldir=DIR.
10505 *) Complete rewrite of the error code script(s). It is all now handled
10506 by one script at the top level which handles error code gathering,
10507 header rewriting and C source file generation. It should be much better
10508 than the old method: it now uses a modified version of Ulf's parser to
10509 read the ANSI prototypes in all header files (thus the old K&R definitions
10510 aren't needed for error creation any more) and do a better job of
10511 translating function codes into names. The old 'ASN1 error code imbedded
10512 in a comment' is no longer necessary and it doesn't use .err files which
10513 have now been deleted. Also the error code call doesn't have to appear all
10514 on one line (which resulted in some large lines...).
10517 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10520 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10521 0 (which usually indicates a closed connection), but continue reading.
10524 *) Fix some race conditions.
10527 *) Add support for CRL distribution points extension. Add Certificate
10528 Policies and CRL distribution points documentation.
10531 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10534 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10535 8 of keying material. Merlin has also confirmed interop with this fix
10536 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10537 [Merlin Hughes <merlin@baltimore.ie>]
10539 *) Fix lots of warnings.
10540 [Richard Levitte <levitte@stacken.kth.se>]
10542 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10543 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10544 [Richard Levitte <levitte@stacken.kth.se>]
10546 *) Fix problems with sizeof(long) == 8.
10547 [Andy Polyakov <appro@fy.chalmers.se>]
10549 *) Change functions to ANSI C.
10552 *) Fix typos in error codes.
10553 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10555 *) Remove defunct assembler files from Configure.
10558 *) SPARC v8 assembler BIGNUM implementation.
10559 [Andy Polyakov <appro@fy.chalmers.se>]
10561 *) Support for Certificate Policies extension: both print and set.
10562 Various additions to support the r2i method this uses.
10565 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10566 return a const string when you are expecting an allocated buffer.
10569 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10570 types DirectoryString and DisplayText.
10573 *) Add code to allow r2i extensions to access the configuration database,
10574 add an LHASH database driver and add several ctx helper functions.
10577 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10578 fail when they extended the size of a BIGNUM.
10581 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10582 support typesafe stack.
10585 *) Fix typo in SSL_[gs]et_options().
10586 [Nils Frostberg <nils@medcom.se>]
10588 *) Delete various functions and files that belonged to the (now obsolete)
10589 old X509V3 handling code.
10592 *) New Configure option "rsaref".
10595 *) Don't auto-generate pem.h.
10598 *) Introduce type-safe ASN.1 SETs.
10601 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10602 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10604 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10605 that links with OpenSSL (well at least cause lots of warnings), but fear
10606 not: the conversion is trivial, and it eliminates loads of evil casts. A
10607 few STACKed things have been converted already. Feel free to convert more.
10608 In the fullness of time, I'll do away with the STACK type altogether.
10611 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10612 specified in <certfile> by updating the entry in the index.txt file.
10613 This way one no longer has to edit the index.txt file manually for
10614 revoking a certificate. The -revoke option does the gory details now.
10615 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10617 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10618 `-text' option at all and this way the `-noout -text' combination was
10619 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10620 [Ralf S. Engelschall]
10622 *) Make sure a corresponding plain text error message exists for the
10623 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10624 verify callback function determined that a certificate was revoked.
10625 [Ralf S. Engelschall]
10627 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10628 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10629 all available cipers including rc5, which was forgotten until now.
10630 In order to let the testing shell script know which algorithms
10631 are available, a new (up to now undocumented) command
10632 "openssl list-cipher-commands" is used.
10635 *) Bugfix: s_client occasionally would sleep in select() when
10636 it should have checked SSL_pending() first.
10639 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10640 the raw DSA values prior to ASN.1 encoding.
10643 *) Tweaks to Configure
10644 [Niels Poppe <niels@netbox.org>]
10646 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10650 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10653 *) New config option to avoid instructions that are illegal on the 80386.
10654 The default code is faster, but requires at least a 486.
10657 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10658 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10659 same as SSL2_VERSION anyway.
10662 *) New "-showcerts" option for s_client.
10665 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10666 application. Various cleanups and fixes.
10669 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10670 modify error routines to work internally. Add error codes and PBE init
10671 to library startup routines.
10674 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10675 packing functions to asn1 and evp. Changed function names and error
10676 codes along the way.
10679 *) PKCS12 integration: and so it begins... First of several patches to
10680 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10681 objects to objects.h
10684 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10685 and display support for Thawte strong extranet extension.
10688 *) Add LinuxPPC support.
10689 [Jeff Dubrule <igor@pobox.org>]
10691 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10692 bn_div_words in alpha.s.
10693 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10695 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10696 OAEP isn't supported when OpenSSL is built with RSAref.
10697 [Ulf Moeller <ulf@fitug.de>]
10699 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10700 so they no longer are missing under -DNOPROTO.
10701 [Soren S. Jorvang <soren@t.dk>]
10704 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10706 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10707 doesn't work when the session is reused. Coming soon!
10710 *) Fix a security hole, that allows sessions to be reused in the wrong
10711 context thus bypassing client cert protection! All software that uses
10712 client certs and session caches in multiple contexts NEEDS PATCHING to
10713 allow session reuse! A fuller solution is in the works.
10714 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10716 *) Some more source tree cleanups (removed obsolete files
10717 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10718 permission on "config" script to be executable) and a fix for the INSTALL
10720 [Ulf Moeller <ulf@fitug.de>]
10722 *) Remove some legacy and erroneous uses of malloc, free instead of
10724 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10726 *) Make rsa_oaep_test return non-zero on error.
10727 [Ulf Moeller <ulf@fitug.de>]
10729 *) Add support for native Solaris shared libraries. Configure
10730 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10731 if someone would make that last step automatic.
10732 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10734 *) ctx_size was not built with the right compiler during "make links". Fixed.
10737 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10738 except NULL ciphers". This means the default cipher list will no longer
10739 enable NULL ciphers. They need to be specifically enabled e.g. with
10740 the string "DEFAULT:eNULL".
10743 *) Fix to RSA private encryption routines: if p < q then it would
10744 occasionally produce an invalid result. This will only happen with
10745 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10748 *) Be less restrictive and allow also `perl util/perlpath.pl
10749 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10750 because this way one can also use an interpreter named `perl5' (which is
10751 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10752 installed as `perl').
10753 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10755 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10756 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10758 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10759 advapi32.lib to Win32 build and change the pem test comparision
10760 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10761 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10762 and crypto/des/ede_cbcm_enc.c.
10765 *) DES quad checksum was broken on big-endian architectures. Fixed.
10768 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10769 Win32 test batch file so it (might) work again. The Win32 test batch file
10770 is horrible: I feel ill....
10773 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10774 in e_os.h. Audit of header files to check ANSI and non ANSI
10775 sections: 10 functions were absent from non ANSI section and not exported
10776 from Windows DLLs. Fixed up libeay.num for new functions.
10779 *) Make `openssl version' output lines consistent.
10780 [Ralf S. Engelschall]
10782 *) Fix Win32 symbol export lists for BIO functions: Added
10783 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10784 to ms/libeay{16,32}.def.
10785 [Ralf S. Engelschall]
10787 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10788 fine under Unix and passes some trivial tests I've now added. But the
10789 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10790 added to make sure no one expects that this stuff really works in the
10791 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10792 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10794 [Ralf S. Engelschall]
10796 *) Fix the generation of two part addresses in perl.
10797 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10799 *) Add config entry for Linux on MIPS.
10800 [John Tobey <jtobey@channel1.com>]
10802 *) Make links whenever Configure is run, unless we are on Windoze.
10805 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10806 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10810 *) Add a useful kludge to allow package maintainers to specify compiler and
10811 other platforms details on the command line without having to patch the
10812 Configure script everytime: One now can use ``perl Configure
10813 <id>:<details>'', i.e. platform ids are allowed to have details appended
10814 to them (seperated by colons). This is treated as there would be a static
10815 pre-configured entry in Configure's %table under key <id> with value
10816 <details> and ``perl Configure <id>'' is called. So, when you want to
10817 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10818 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10819 now, which overrides the FreeBSD-elf entry on-the-fly.
10820 [Ralf S. Engelschall]
10822 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10825 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10826 on the `perl Configure ...' command line. This way one can compile
10827 OpenSSL libraries with Position Independent Code (PIC) which is needed
10828 for linking it into DSOs.
10829 [Ralf S. Engelschall]
10831 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10835 *) Cleaned up the LICENSE document: The official contact for any license
10836 questions now is the OpenSSL core team under openssl-core@openssl.org.
10837 And add a paragraph about the dual-license situation to make sure people
10838 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10839 to the OpenSSL toolkit.
10840 [Ralf S. Engelschall]
10842 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10843 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10844 Additonally cleaned up the `make links' target: Remove unnecessary
10845 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10846 to speed processing and no longer clutter the display with confusing
10847 stuff. Instead only the actually done links are displayed.
10848 [Ralf S. Engelschall]
10850 *) Permit null encryption ciphersuites, used for authentication only. It used
10851 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10852 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10856 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10857 signed attributes when verifying signatures (this would break them),
10858 the detached data encoding was wrong and public keys obtained using
10859 X509_get_pubkey() weren't freed.
10862 *) Add text documentation for the BUFFER functions. Also added a work around
10863 to a Win95 console bug. This was triggered by the password read stuff: the
10864 last character typed gets carried over to the next fread(). If you were
10865 generating a new cert request using 'req' for example then the last
10866 character of the passphrase would be CR which would then enter the first
10870 *) Added the new `Includes OpenSSL Cryptography Software' button as
10871 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10872 button and can be used by applications based on OpenSSL to show the
10873 relationship to the OpenSSL project.
10874 [Ralf S. Engelschall]
10876 *) Remove confusing variables in function signatures in files
10877 ssl/ssl_lib.c and ssl/ssl.h.
10878 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10880 *) Don't install bss_file.c under PREFIX/include/
10881 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10883 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10884 functions that return function pointers and has support for NT specific
10885 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10886 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10887 unsigned to signed types: this was killing the Win32 compile.
10890 *) Add new certificate file to stack functions,
10891 SSL_add_dir_cert_subjects_to_stack() and
10892 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10893 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10894 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10895 This means that Apache-SSL and similar packages don't have to mess around
10896 to add as many CAs as they want to the preferred list.
10899 *) Experiment with doxygen documentation. Currently only partially applied to
10901 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10902 openssl.doxy as the configuration file.
10905 *) Get rid of remaining C++-style comments which strict C compilers hate.
10906 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10908 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10909 compiled in by default: it has problems with large keys.
10912 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10913 DH private keys and/or callback functions which directly correspond to
10914 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10915 is needed for applications which have to configure certificates on a
10916 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10918 For the RSA certificate situation is makes no difference, but
10919 for the DSA certificate situation this fixes the "no shared cipher"
10920 problem where the OpenSSL cipher selection procedure failed because the
10921 temporary keys were not overtaken from the context and the API provided
10922 no way to reconfigure them.
10923 The new functions now let applications reconfigure the stuff and they
10924 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10925 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10926 non-public-API function ssl_cert_instantiate() is used as a helper
10927 function and also to reduce code redundancy inside ssl_rsa.c.
10928 [Ralf S. Engelschall]
10930 *) Move s_server -dcert and -dkey options out of the undocumented feature
10931 area because they are useful for the DSA situation and should be
10932 recognized by the users.
10933 [Ralf S. Engelschall]
10935 *) Fix the cipher decision scheme for export ciphers: the export bits are
10936 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10937 SSL_EXP_MASK. So, the original variable has to be used instead of the
10938 already masked variable.
10939 [Richard Levitte <levitte@stacken.kth.se>]
10941 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10942 [Richard Levitte <levitte@stacken.kth.se>]
10944 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10945 from `int' to `unsigned int' because it's a length and initialized by
10946 EVP_DigestFinal() which expects an `unsigned int *'.
10947 [Richard Levitte <levitte@stacken.kth.se>]
10949 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10950 script. Instead use the usual Shell->Perl transition trick.
10951 [Ralf S. Engelschall]
10953 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10954 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10955 -noout -modulus' as it's already the case for `openssl rsa -noout
10956 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10957 currently the public key is printed (a decision which was already done by
10958 `openssl dsa -modulus' in the past) which serves a similar purpose.
10959 Additionally the NO_RSA no longer completely removes the whole -modulus
10960 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10962 [Ralf S. Engelschall]
10964 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10965 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10966 [Arne Ansper <arne@ats.cyber.ee>]
10968 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10969 to be added. Now both 'req' and 'ca' can use new objects defined in the
10973 *) Add cool BIO that does syslog (or event log on NT).
10974 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10976 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10977 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10978 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10979 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10982 *) Add preliminary config info for new extension code.
10985 *) Make RSA_NO_PADDING really use no padding.
10986 [Ulf Moeller <ulf@fitug.de>]
10988 *) Generate errors when private/public key check is done.
10991 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10992 for some CRL extensions and new objects added.
10995 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10996 key usage extension and fuller support for authority key id.
10999 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
11000 padding method for RSA, which is recommended for new applications in PKCS
11001 #1 v2.0 (RFC 2437, October 1998).
11002 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
11003 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
11004 against Bleichbacher's attack on RSA.
11005 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
11008 *) Updates to the new SSL compression code
11009 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11011 *) Fix so that the version number in the master secret, when passed
11012 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
11013 (because the server will not accept higher), that the version number
11014 is 0x03,0x01, not 0x03,0x00
11015 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11017 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
11018 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
11019 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
11022 *) Support for RAW extensions where an arbitrary extension can be
11023 created by including its DER encoding. See apps/openssl.cnf for
11027 *) Make sure latest Perl versions don't interpret some generated C array
11028 code as Perl array code in the crypto/err/err_genc.pl script.
11029 [Lars Weber <3weber@informatik.uni-hamburg.de>]
11031 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
11032 not many people have the assembler. Various Win32 compilation fixes and
11033 update to the INSTALL.W32 file with (hopefully) more accurate Win32
11034 build instructions.
11037 *) Modify configure script 'Configure' to automatically create crypto/date.h
11038 file under Win32 and also build pem.h from pem.org. New script
11039 util/mkfiles.pl to create the MINFO file on environments that can't do a
11040 'make files': perl util/mkfiles.pl >MINFO should work.
11043 *) Major rework of DES function declarations, in the pursuit of correctness
11044 and purity. As a result, many evil casts evaporated, and some weirdness,
11045 too. You may find this causes warnings in your code. Zapping your evil
11046 casts will probably fix them. Mostly.
11049 *) Fix for a typo in asn1.h. Bug fix to object creation script
11050 obj_dat.pl. It considered a zero in an object definition to mean
11051 "end of object": none of the objects in objects.h have any zeros
11052 so it wasn't spotted.
11053 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
11055 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
11056 Masking (CBCM). In the absence of test vectors, the best I have been able
11057 to do is check that the decrypt undoes the encrypt, so far. Send me test
11058 vectors if you have them.
11061 *) Correct calculation of key length for export ciphers (too much space was
11062 allocated for null ciphers). This has not been tested!
11065 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
11066 message is now correct (it understands "crypto" and "ssl" on its
11067 command line). There is also now an "update" option. This will update
11068 the util/ssleay.num and util/libeay.num files with any new functions.
11070 perl util/mkdef.pl crypto ssl update
11071 it will update them.
11074 *) Overhauled the Perl interface (perl/*):
11075 - ported BN stuff to OpenSSL's different BN library
11076 - made the perl/ source tree CVS-aware
11077 - renamed the package from SSLeay to OpenSSL (the files still contain
11078 their history because I've copied them in the repository)
11079 - removed obsolete files (the test scripts will be replaced
11080 by better Test::Harness variants in the future)
11081 [Ralf S. Engelschall]
11083 *) First cut for a very conservative source tree cleanup:
11084 1. merge various obsolete readme texts into doc/ssleay.txt
11085 where we collect the old documents and readme texts.
11086 2. remove the first part of files where I'm already sure that we no
11087 longer need them because of three reasons: either they are just temporary
11088 files which were left by Eric or they are preserved original files where
11089 I've verified that the diff is also available in the CVS via "cvs diff
11090 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11091 the crypto/md/ stuff).
11092 [Ralf S. Engelschall]
11094 *) More extension code. Incomplete support for subject and issuer alt
11095 name, issuer and authority key id. Change the i2v function parameters
11096 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11097 what that's for :-) Fix to ASN1 macro which messed up
11098 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11101 *) Preliminary support for ENUMERATED type. This is largely copied from the
11105 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11106 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11108 *) Make sure `make rehash' target really finds the `openssl' program.
11109 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11111 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11112 like to hear about it if this slows down other processors.
11115 *) Add CygWin32 platform information to Configure script.
11116 [Alan Batie <batie@aahz.jf.intel.com>]
11118 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11119 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11121 *) New program nseq to manipulate netscape certificate sequences
11124 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11128 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11129 but the BN code had some problems that would cause failures when
11130 doing certificate verification and some other functions.
11131 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11133 *) Add ASN1 and PEM code to support netscape certificate sequences.
11136 *) Add ASN1 and PEM code to support netscape certificate sequences.
11139 *) Add several PKIX and private extended key usage OIDs.
11142 *) Modify the 'ca' program to handle the new extension code. Modify
11143 openssl.cnf for new extension format, add comments.
11146 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11147 and add a sample to openssl.cnf so req -x509 now adds appropriate
11151 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11152 error code, add initial support to X509_print() and x509 application.
11155 *) Takes a deep breath and start addding X509 V3 extension support code. Add
11156 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11157 stuff is currently isolated and isn't even compiled yet.
11160 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11161 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11162 Removed the versions check from X509 routines when loading extensions:
11163 this allows certain broken certificates that don't set the version
11164 properly to be processed.
11167 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11168 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11169 can still be regenerated with "make depend".
11172 *) Spelling mistake in C version of CAST-128.
11173 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11175 *) Changes to the error generation code. The perl script err-code.pl
11176 now reads in the old error codes and retains the old numbers, only
11177 adding new ones if necessary. It also only changes the .err files if new
11178 codes are added. The makefiles have been modified to only insert errors
11179 when needed (to avoid needlessly modifying header files). This is done
11180 by only inserting errors if the .err file is newer than the auto generated
11181 C file. To rebuild all the error codes from scratch (the old behaviour)
11182 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11183 or delete all the .err files.
11186 *) CAST-128 was incorrectly implemented for short keys. The C version has
11187 been fixed, but is untested. The assembler versions are also fixed, but
11188 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11189 to regenerate it if needed.
11190 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11191 Hagino <itojun@kame.net>]
11193 *) File was opened incorrectly in randfile.c.
11194 [Ulf Möller <ulf@fitug.de>]
11196 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11197 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11198 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11199 al: it's just almost always a UTCTime. Note this patch adds new error
11200 codes so do a "make errors" if there are problems.
11203 *) Correct Linux 1 recognition in config.
11204 [Ulf Möller <ulf@fitug.de>]
11206 *) Remove pointless MD5 hash when using DSA keys in ca.
11207 [Anonymous <nobody@replay.com>]
11209 *) Generate an error if given an empty string as a cert directory. Also
11210 generate an error if handed NULL (previously returned 0 to indicate an
11211 error, but didn't set one).
11212 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11214 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11217 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11218 parameters. This was causing a warning which killed off the Win32 compile.
11221 *) Remove C++ style comments from crypto/bn/bn_local.h.
11222 [Neil Costigan <neil.costigan@celocom.com>]
11224 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11225 based on a text string, looking up short and long names and finally
11226 "dot" format. The "dot" format stuff didn't work. Added new function
11227 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11228 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11229 OID is not part of the table.
11232 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11233 X509_LOOKUP_by_alias().
11236 *) Sort openssl functions by name.
11239 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11240 encryption from sample DSA keys (in case anyone is interested the password
11244 *) Make _all_ *_free functions accept a NULL pointer.
11245 [Frans Heymans <fheymans@isaserver.be>]
11247 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11249 [Anonymous <nobody@replay.com>]
11251 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11252 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11254 *) Don't blow it for numeric -newkey arguments to apps/req.
11255 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11257 *) Temp key "for export" tests were wrong in s3_srvr.c.
11258 [Anonymous <nobody@replay.com>]
11260 *) Add prototype for temp key callback functions
11261 SSL_CTX_set_tmp_{rsa,dh}_callback().
11264 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11265 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11268 *) X509_name_add_entry() freed the wrong thing after an error.
11269 [Arne Ansper <arne@ats.cyber.ee>]
11271 *) rsa_eay.c would attempt to free a NULL context.
11272 [Arne Ansper <arne@ats.cyber.ee>]
11274 *) BIO_s_socket() had a broken should_retry() on Windoze.
11275 [Arne Ansper <arne@ats.cyber.ee>]
11277 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11278 [Arne Ansper <arne@ats.cyber.ee>]
11280 *) Make sure the already existing X509_STORE->depth variable is initialized
11281 in X509_STORE_new(), but document the fact that this variable is still
11282 unused in the certificate verification process.
11283 [Ralf S. Engelschall]
11285 *) Fix the various library and apps files to free up pkeys obtained from
11286 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11289 *) Fix reference counting in X509_PUBKEY_get(). This makes
11290 demos/maurice/example2.c work, amongst others, probably.
11291 [Steve Henson and Ben Laurie]
11293 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11294 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11295 are no longer created. This way we have a single and consistent command
11296 line interface `openssl <command>', similar to `cvs <command>'.
11297 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11299 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11300 BIT STRING wrapper always have zero unused bits.
11303 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11306 *) Make the top-level INSTALL documentation easier to understand.
11309 *) Makefiles updated to exit if an error occurs in a sub-directory
11310 make (including if user presses ^C) [Paul Sutton]
11312 *) Make Montgomery context stuff explicit in RSA data structure.
11315 *) Fix build order of pem and err to allow for generated pem.h.
11318 *) Fix renumbering bug in X509_NAME_delete_entry().
11321 *) Enhanced the err-ins.pl script so it makes the error library number
11322 global and can add a library name. This is needed for external ASN1 and
11323 other error libraries.
11326 *) Fixed sk_insert which never worked properly.
11329 *) Fix ASN1 macros so they can handle indefinite length construted
11330 EXPLICIT tags. Some non standard certificates use these: they can now
11334 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11335 into a single doc/ssleay.txt bundle. This way the information is still
11336 preserved but no longer messes up this directory. Now it's new room for
11337 the new set of documenation files.
11338 [Ralf S. Engelschall]
11340 *) SETs were incorrectly DER encoded. This was a major pain, because they
11341 shared code with SEQUENCEs, which aren't coded the same. This means that
11342 almost everything to do with SETs or SEQUENCEs has either changed name or
11343 number of arguments.
11344 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11346 *) Fix test data to work with the above.
11349 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11350 was already fixed by Eric for 0.9.1 it seems.
11351 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11353 *) Autodetect FreeBSD3.
11356 *) Fix various bugs in Configure. This affects the following platforms:
11360 unixware-2.0-pentium
11364 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11365 before they are needed.
11368 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11372 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11374 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11375 changed SSLeay to OpenSSL in version strings.
11376 [Ralf S. Engelschall]
11378 *) Some fixups to the top-level documents.
11381 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11382 because the symlink to include/ was missing.
11383 [Ralf S. Engelschall]
11385 *) Incorporated the popular no-RSA/DSA-only patches
11386 which allow to compile a RSA-free SSLeay.
11387 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11389 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11390 when "ssleay" is still not found.
11391 [Ralf S. Engelschall]
11393 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11394 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11396 *) Updated the README file.
11397 [Ralf S. Engelschall]
11399 *) Added various .cvsignore files in the CVS repository subdirs
11400 to make a "cvs update" really silent.
11401 [Ralf S. Engelschall]
11403 *) Recompiled the error-definition header files and added
11404 missing symbols to the Win32 linker tables.
11405 [Ralf S. Engelschall]
11407 *) Cleaned up the top-level documents;
11408 o new files: CHANGES and LICENSE
11409 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11410 o merged COPYRIGHT into LICENSE
11411 o removed obsolete TODO file
11412 o renamed MICROSOFT to INSTALL.W32
11413 [Ralf S. Engelschall]
11415 *) Removed dummy files from the 0.9.1b source tree:
11416 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11417 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11418 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11419 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11420 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11421 [Ralf S. Engelschall]
11423 *) Added various platform portability fixes.
11426 *) The Genesis of the OpenSSL rpject:
11427 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11428 Young and Tim J. Hudson created while they were working for C2Net until
11430 [The OpenSSL Project]
11433 Changes between 0.9.0b and 0.9.1b [not released]
11435 *) Updated a few CA certificates under certs/
11438 *) Changed some BIGNUM api stuff.
11441 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11442 DGUX x86, Linux Alpha, etc.
11445 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11446 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11450 *) Add -strparse option to asn1pars program which parses nested
11452 [Dr Stephen Henson <shenson@bigfoot.com>]
11454 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11457 *) DSA fix for "ca" program.
11460 *) Added "-genkey" option to "dsaparam" program.
11463 *) Added RIPE MD160 (rmd160) message digest.
11466 *) Added -a (all) option to "ssleay version" command.
11469 *) Added PLATFORM define which is the id given to Configure.
11472 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11475 *) Extended the ASN.1 parser routines.
11478 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11481 *) Added a BN_CTX to the BN library.
11484 *) Fixed the weak key values in DES library
11487 *) Changed API in EVP library for cipher aliases.
11490 *) Added support for RC2/64bit cipher.
11493 *) Converted the lhash library to the crypto/mem.c functions.
11496 *) Added more recognized ASN.1 object ids.
11499 *) Added more RSA padding checks for SSL/TLS.
11502 *) Added BIO proxy/filter functionality.
11505 *) Added extra_certs to SSL_CTX which can be used
11506 send extra CA certificates to the client in the CA cert chain sending
11507 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11510 *) Now Fortezza is denied in the authentication phase because
11511 this is key exchange mechanism is not supported by SSLeay at all.
11514 *) Additional PKCS1 checks.
11517 *) Support the string "TLSv1" for all TLS v1 ciphers.
11520 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11521 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11524 *) Fixed a few memory leaks.
11527 *) Fixed various code and comment typos.
11530 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11531 bytes sent in the client random.
11532 [Edward Bishop <ebishop@spyglass.com>]