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 1.0.2q and 1.0.2r [26 Feb 2019]
12 *) 0-byte record padding oracle
14 If an application encounters a fatal protocol error and then calls
15 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
16 then OpenSSL can respond differently to the calling application if a 0 byte
17 record is received with invalid padding compared to if a 0 byte record is
18 received with an invalid MAC. If the application then behaves differently
19 based on that in a way that is detectable to the remote peer, then this
20 amounts to a padding oracle that could be used to decrypt data.
22 In order for this to be exploitable "non-stitched" ciphersuites must be in
23 use. Stitched ciphersuites are optimised implementations of certain
24 commonly used ciphersuites. Also the application must call SSL_shutdown()
25 twice even if a protocol error has occurred (applications should not do
26 this but some do anyway).
28 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
29 Aviram, with additional investigation by Steven Collison and Andrew
30 Hourselt. It was reported to OpenSSL on 10th December 2018.
34 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
37 Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
39 *) Microarchitecture timing vulnerability in ECC scalar multiplication
41 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
42 shown to be vulnerable to a microarchitecture timing side channel attack.
43 An attacker with sufficient access to mount local timing attacks during
44 ECDSA signature generation could recover the private key.
46 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
47 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
52 *) Timing vulnerability in DSA signature generation
54 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
55 timing side channel attack. An attacker could use variations in the signing
56 algorithm to recover the private key.
58 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
62 *) Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
63 Module, accidentally introduced while backporting security fixes from the
64 development branch and hindering the use of ECC in FIPS mode.
67 Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
69 *) Client DoS due to large DH parameter
71 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
72 malicious server can send a very large prime value to the client. This will
73 cause the client to spend an unreasonably long period of time generating a
74 key for this prime resulting in a hang until the client has finished. This
75 could be exploited in a Denial Of Service attack.
77 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
81 *) Cache timing vulnerability in RSA Key Generation
83 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
84 a cache timing side channel attack. An attacker with sufficient access to
85 mount cache timing attacks during the RSA key generation process could
86 recover the private key.
88 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
89 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
93 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
94 parameter is no longer accepted, as it leads to a corrupt table. NULL
95 pem_str is reserved for alias entries only.
98 *) Revert blinding in ECDSA sign and instead make problematic addition
99 length-invariant. Switch even to fixed-length Montgomery multiplication.
102 *) Change generating and checking of primes so that the error rate of not
103 being prime depends on the intended use based on the size of the input.
104 For larger primes this will result in more rounds of Miller-Rabin.
105 The maximal error rate for primes with more than 1080 bits is lowered
107 [Kurt Roeckx, Annie Yousar]
109 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
112 *) Add blinding to ECDSA and DSA signatures to protect against side channel
113 attacks discovered by Keegan Ryan (NCC Group).
116 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
117 now allow empty (zero character) pass phrases.
120 *) Certificate time validation (X509_cmp_time) enforces stricter
121 compliance with RFC 5280. Fractional seconds and timezone offsets
122 are no longer allowed.
125 Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
127 *) Constructed ASN.1 types with a recursive definition could exceed the stack
129 Constructed ASN.1 types with a recursive definition (such as can be found
130 in PKCS7) could eventually exceed the stack given malicious input with
131 excessive recursion. This could result in a Denial Of Service attack. There
132 are no such structures used within SSL/TLS that come from untrusted sources
133 so this is considered safe.
135 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
140 Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
142 *) Read/write after SSL object in error state
144 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
145 mechanism. The intent was that if a fatal error occurred during a handshake
146 then OpenSSL would move into the error state and would immediately fail if
147 you attempted to continue the handshake. This works as designed for the
148 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
149 SSL_connect()), however due to a bug it does not work correctly if
150 SSL_read() or SSL_write() is called directly. In that scenario, if the
151 handshake fails then a fatal error will be returned in the initial function
152 call. If SSL_read()/SSL_write() is subsequently called by the application
153 for the same SSL object then it will succeed and the data is passed without
154 being decrypted/encrypted directly from the SSL/TLS record layer.
156 In order to exploit this issue an application bug would have to be present
157 that resulted in a call to SSL_read()/SSL_write() being issued after having
158 already received a fatal error.
160 This issue was reported to OpenSSL by David Benjamin (Google).
164 *) rsaz_1024_mul_avx2 overflow bug on x86_64
166 There is an overflow bug in the AVX2 Montgomery multiplication procedure
167 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
168 Analysis suggests that attacks against RSA and DSA as a result of this
169 defect would be very difficult to perform and are not believed likely.
170 Attacks against DH1024 are considered just feasible, because most of the
171 work necessary to deduce information about a private key may be performed
172 offline. The amount of resources required for such an attack would be
173 significant. However, for an attack on TLS to be meaningful, the server
174 would have to share the DH1024 private key among multiple clients, which is
175 no longer an option since CVE-2016-0701.
177 This only affects processors that support the AVX2 but not ADX extensions
178 like Intel Haswell (4th generation).
180 This issue was reported to OpenSSL by David Benjamin (Google). The issue
181 was originally found via the OSS-Fuzz project.
185 Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
187 *) bn_sqrx8x_internal carry bug on x86_64
189 There is a carry propagating bug in the x86_64 Montgomery squaring
190 procedure. No EC algorithms are affected. Analysis suggests that attacks
191 against RSA and DSA as a result of this defect would be very difficult to
192 perform and are not believed likely. Attacks against DH are considered just
193 feasible (although very difficult) because most of the work necessary to
194 deduce information about a private key may be performed offline. The amount
195 of resources required for such an attack would be very significant and
196 likely only accessible to a limited number of attackers. An attacker would
197 additionally need online access to an unpatched system using the target
198 private key in a scenario with persistent DH parameters and a private
199 key that is shared between multiple clients.
201 This only affects processors that support the BMI1, BMI2 and ADX extensions
202 like Intel Broadwell (5th generation) and later or AMD Ryzen.
204 This issue was reported to OpenSSL by the OSS-Fuzz project.
208 *) Malformed X.509 IPAddressFamily could cause OOB read
210 If an X.509 certificate has a malformed IPAddressFamily extension,
211 OpenSSL could do a one-byte buffer overread. The most likely result
212 would be an erroneous display of the certificate in text format.
214 This issue was reported to OpenSSL by the OSS-Fuzz project.
218 Changes between 1.0.2k and 1.0.2l [25 May 2017]
220 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
221 platform rather than 'mingw'.
224 Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
226 *) Truncated packet could crash via OOB read
228 If one side of an SSL/TLS path is running on a 32-bit host and a specific
229 cipher is being used, then a truncated packet can cause that host to
230 perform an out-of-bounds read, usually resulting in a crash.
232 This issue was reported to OpenSSL by Robert Święcki of Google.
236 *) BN_mod_exp may produce incorrect results on x86_64
238 There is a carry propagating bug in the x86_64 Montgomery squaring
239 procedure. No EC algorithms are affected. Analysis suggests that attacks
240 against RSA and DSA as a result of this defect would be very difficult to
241 perform and are not believed likely. Attacks against DH are considered just
242 feasible (although very difficult) because most of the work necessary to
243 deduce information about a private key may be performed offline. The amount
244 of resources required for such an attack would be very significant and
245 likely only accessible to a limited number of attackers. An attacker would
246 additionally need online access to an unpatched system using the target
247 private key in a scenario with persistent DH parameters and a private
248 key that is shared between multiple clients. For example this can occur by
249 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
250 similar to CVE-2015-3193 but must be treated as a separate problem.
252 This issue was reported to OpenSSL by the OSS-Fuzz project.
256 *) Montgomery multiplication may produce incorrect results
258 There is a carry propagating bug in the Broadwell-specific Montgomery
259 multiplication procedure that handles input lengths divisible by, but
260 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
261 and DH private keys are impossible. This is because the subroutine in
262 question is not used in operations with the private key itself and an input
263 of the attacker's direct choice. Otherwise the bug can manifest itself as
264 transient authentication and key negotiation failures or reproducible
265 erroneous outcome of public-key operations with specially crafted input.
266 Among EC algorithms only Brainpool P-512 curves are affected and one
267 presumably can attack ECDH key negotiation. Impact was not analyzed in
268 detail, because pre-requisites for attack are considered unlikely. Namely
269 multiple clients have to choose the curve in question and the server has to
270 share the private key among them, neither of which is default behaviour.
271 Even then only clients that chose the curve will be affected.
273 This issue was publicly reported as transient failures and was not
274 initially recognized as a security issue. Thanks to Richard Morgan for
275 providing reproducible case.
279 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
280 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
281 prevent issues where no progress is being made and the peer continually
282 sends unrecognised record types, using up resources processing them.
285 Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
287 *) Missing CRL sanity check
289 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
290 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
291 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
293 This issue only affects the OpenSSL 1.0.2i
297 Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
299 *) OCSP Status Request extension unbounded memory growth
301 A malicious client can send an excessively large OCSP Status Request
302 extension. If that client continually requests renegotiation, sending a
303 large OCSP Status Request extension each time, then there will be unbounded
304 memory growth on the server. This will eventually lead to a Denial Of
305 Service attack through memory exhaustion. Servers with a default
306 configuration are vulnerable even if they do not support OCSP. Builds using
307 the "no-ocsp" build time option are not affected.
309 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
313 *) In order to mitigate the SWEET32 attack, the DES ciphers were moved from
316 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
321 *) OOB write in MDC2_Update()
323 An overflow can occur in MDC2_Update() either if called directly or
324 through the EVP_DigestUpdate() function using MDC2. If an attacker
325 is able to supply very large amounts of input data after a previous
326 call to EVP_EncryptUpdate() with a partial block then a length check
327 can overflow resulting in a heap corruption.
329 The amount of data needed is comparable to SIZE_MAX which is impractical
332 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
336 *) Malformed SHA512 ticket DoS
338 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
339 DoS attack where a malformed ticket will result in an OOB read which will
342 The use of SHA512 in TLS session tickets is comparatively rare as it requires
343 a custom server callback and ticket lookup mechanism.
345 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
349 *) OOB write in BN_bn2dec()
351 The function BN_bn2dec() does not check the return value of BN_div_word().
352 This can cause an OOB write if an application uses this function with an
353 overly large BIGNUM. This could be a problem if an overly large certificate
354 or CRL is printed out from an untrusted source. TLS is not affected because
355 record limits will reject an oversized certificate before it is parsed.
357 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
361 *) OOB read in TS_OBJ_print_bio()
363 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
364 the total length the OID text representation would use and not the amount
365 of data written. This will result in OOB reads when large OIDs are
368 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
372 *) Pointer arithmetic undefined behaviour
374 Avoid some undefined pointer arithmetic
376 A common idiom in the codebase is to check limits in the following manner:
379 Where "p" points to some malloc'd data of SIZE bytes and
382 "len" here could be from some externally supplied data (e.g. from a TLS
385 The rules of C pointer arithmetic are such that "p + len" is only well
386 defined where len <= SIZE. Therefore the above idiom is actually
389 For example this could cause problems if some malloc implementation
390 provides an address for "p" such that "p + len" actually overflows for
391 values of len that are too big and therefore p + len < limit.
393 This issue was reported to OpenSSL by Guido Vranken
397 *) Constant time flag not preserved in DSA signing
399 Operations in the DSA signing algorithm should run in constant time in
400 order to avoid side channel attacks. A flaw in the OpenSSL DSA
401 implementation means that a non-constant time codepath is followed for
402 certain operations. This has been demonstrated through a cache-timing
403 attack to be sufficient for an attacker to recover the private DSA key.
405 This issue was reported by César Pereida (Aalto University), Billy Brumley
406 (Tampere University of Technology), and Yuval Yarom (The University of
411 *) DTLS buffered message DoS
413 In a DTLS connection where handshake messages are delivered out-of-order
414 those messages that OpenSSL is not yet ready to process will be buffered
415 for later use. Under certain circumstances, a flaw in the logic means that
416 those messages do not get removed from the buffer even though the handshake
417 has been completed. An attacker could force up to approx. 15 messages to
418 remain in the buffer when they are no longer required. These messages will
419 be cleared when the DTLS connection is closed. The default maximum size for
420 a message is 100k. Therefore the attacker could force an additional 1500k
421 to be consumed per connection. By opening many simulataneous connections an
422 attacker could cause a DoS attack through memory exhaustion.
424 This issue was reported to OpenSSL by Quan Luo.
428 *) DTLS replay protection DoS
430 A flaw in the DTLS replay attack protection mechanism means that records
431 that arrive for future epochs update the replay protection "window" before
432 the MAC for the record has been validated. This could be exploited by an
433 attacker by sending a record for the next epoch (which does not have to
434 decrypt or have a valid MAC), with a very large sequence number. This means
435 that all subsequent legitimate packets are dropped causing a denial of
436 service for a specific DTLS connection.
438 This issue was reported to OpenSSL by the OCAP audit team.
442 *) Certificate message OOB reads
444 In OpenSSL 1.0.2 and earlier some missing message length checks can result
445 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
446 theoretical DoS risk but this has not been observed in practice on common
449 The messages affected are client certificate, client certificate request
450 and server certificate. As a result the attack can only be performed
451 against a client or a server which enables client authentication.
453 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
457 Changes between 1.0.2g and 1.0.2h [3 May 2016]
459 *) Prevent padding oracle in AES-NI CBC MAC check
461 A MITM attacker can use a padding oracle attack to decrypt traffic
462 when the connection uses an AES CBC cipher and the server support
465 This issue was introduced as part of the fix for Lucky 13 padding
466 attack (CVE-2013-0169). The padding check was rewritten to be in
467 constant time by making sure that always the same bytes are read and
468 compared against either the MAC or padding bytes. But it no longer
469 checked that there was enough data to have both the MAC and padding
472 This issue was reported by Juraj Somorovsky using TLS-Attacker.
476 *) Fix EVP_EncodeUpdate overflow
478 An overflow can occur in the EVP_EncodeUpdate() function which is used for
479 Base64 encoding of binary data. If an attacker is able to supply very large
480 amounts of input data then a length check can overflow resulting in a heap
483 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
484 the PEM_write_bio* family of functions. These are mainly used within the
485 OpenSSL command line applications, so any application which processes data
486 from an untrusted source and outputs it as a PEM file should be considered
487 vulnerable to this issue. User applications that call these APIs directly
488 with large amounts of untrusted data may also be vulnerable.
490 This issue was reported by Guido Vranken.
494 *) Fix EVP_EncryptUpdate overflow
496 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
497 is able to supply very large amounts of input data after a previous call to
498 EVP_EncryptUpdate() with a partial block then a length check can overflow
499 resulting in a heap corruption. Following an analysis of all OpenSSL
500 internal usage of the EVP_EncryptUpdate() function all usage is one of two
501 forms. The first form is where the EVP_EncryptUpdate() call is known to be
502 the first called function after an EVP_EncryptInit(), and therefore that
503 specific call must be safe. The second form is where the length passed to
504 EVP_EncryptUpdate() can be seen from the code to be some small value and
505 therefore there is no possibility of an overflow. Since all instances are
506 one of these two forms, it is believed that there can be no overflows in
507 internal code due to this problem. It should be noted that
508 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
509 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
510 of these calls have also been analysed too and it is believed there are no
511 instances in internal usage where an overflow could occur.
513 This issue was reported by Guido Vranken.
517 *) Prevent ASN.1 BIO excessive memory allocation
519 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
520 a short invalid encoding can casuse allocation of large amounts of memory
521 potentially consuming excessive resources or exhausting memory.
523 Any application parsing untrusted data through d2i BIO functions is
524 affected. The memory based functions such as d2i_X509() are *not* affected.
525 Since the memory based functions are used by the TLS library, TLS
526 applications are not affected.
528 This issue was reported by Brian Carpenter.
534 ASN1 Strings that are over 1024 bytes can cause an overread in applications
535 using the X509_NAME_oneline() function on EBCDIC systems. This could result
536 in arbitrary stack data being returned in the buffer.
538 This issue was reported by Guido Vranken.
542 *) Modify behavior of ALPN to invoke callback after SNI/servername
543 callback, such that updates to the SSL_CTX affect ALPN.
546 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
550 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
551 methods are enabled and ssl2 is disabled the methods return NULL.
554 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
556 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
557 Builds that are not configured with "enable-weak-ssl-ciphers" will not
558 provide any "EXPORT" or "LOW" strength ciphers.
561 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
562 is by default disabled at build-time. Builds that are not configured with
563 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
564 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
565 will need to explicitly call either of:
567 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
569 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
571 as appropriate. Even if either of those is used, or the application
572 explicitly uses the version-specific SSLv2_method() or its client and
573 server variants, SSLv2 ciphers vulnerable to exhaustive search key
574 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
575 ciphers, and SSLv2 56-bit DES are no longer available.
579 *) Fix a double-free in DSA code
581 A double free bug was discovered when OpenSSL parses malformed DSA private
582 keys and could lead to a DoS attack or memory corruption for applications
583 that receive DSA private keys from untrusted sources. This scenario is
586 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
591 *) Disable SRP fake user seed to address a server memory leak.
593 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
595 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
596 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
597 was changed to ignore the "fake user" SRP seed, even if the seed
600 Users should use SRP_VBASE_get1_by_user instead. Note that in
601 SRP_VBASE_get1_by_user, caller must free the returned value. Note
602 also that even though configuring the SRP seed attempts to hide
603 invalid usernames by continuing the handshake with fake
604 credentials, this behaviour is not constant time and no strong
605 guarantees are made that the handshake is indistinguishable from
606 that of a valid user.
610 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
612 In the BN_hex2bn function the number of hex digits is calculated using an
613 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
614 large values of |i| this can result in |bn_expand| not allocating any
615 memory because |i * 4| is negative. This can leave the internal BIGNUM data
616 field as NULL leading to a subsequent NULL ptr deref. For very large values
617 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
618 In this case memory is allocated to the internal BIGNUM data field, but it
619 is insufficiently sized leading to heap corruption. A similar issue exists
620 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
621 is ever called by user applications with very large untrusted hex/dec data.
622 This is anticipated to be a rare occurrence.
624 All OpenSSL internal usage of these functions use data that is not expected
625 to be untrusted, e.g. config file data or application command line
626 arguments. If user developed applications generate config file data based
627 on untrusted data then it is possible that this could also lead to security
628 consequences. This is also anticipated to be rare.
630 This issue was reported to OpenSSL by Guido Vranken.
634 *) Fix memory issues in BIO_*printf functions
636 The internal |fmtstr| function used in processing a "%s" format string in
637 the BIO_*printf functions could overflow while calculating the length of a
638 string and cause an OOB read when printing very long strings.
640 Additionally the internal |doapr_outch| function can attempt to write to an
641 OOB memory location (at an offset from the NULL pointer) in the event of a
642 memory allocation failure. In 1.0.2 and below this could be caused where
643 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
644 could be in processing a very long "%s" format string. Memory leaks can
647 The first issue may mask the second issue dependent on compiler behaviour.
648 These problems could enable attacks where large amounts of untrusted data
649 is passed to the BIO_*printf functions. If applications use these functions
650 in this way then they could be vulnerable. OpenSSL itself uses these
651 functions when printing out human-readable dumps of ASN.1 data. Therefore
652 applications that print this data could be vulnerable if the data is from
653 untrusted sources. OpenSSL command line applications could also be
654 vulnerable where they print out ASN.1 data, or if untrusted data is passed
655 as command line arguments.
657 Libssl is not considered directly vulnerable. Additionally certificates etc
658 received via remote connections via libssl are also unlikely to be able to
659 trigger these issues because of message size limits enforced within libssl.
661 This issue was reported to OpenSSL Guido Vranken.
665 *) Side channel attack on modular exponentiation
667 A side-channel attack was found which makes use of cache-bank conflicts on
668 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
669 of RSA keys. The ability to exploit this issue is limited as it relies on
670 an attacker who has control of code in a thread running on the same
671 hyper-threaded core as the victim thread which is performing decryptions.
673 This issue was reported to OpenSSL by Yuval Yarom, The University of
674 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
675 Nadia Heninger, University of Pennsylvania with more information at
676 http://cachebleed.info.
680 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
681 if no keysize is specified with default_bits. This fixes an
682 omission in an earlier change that changed all RSA/DSA key generation
683 apps to use 2048 bits by default.
686 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
688 *) DH small subgroups
690 Historically OpenSSL only ever generated DH parameters based on "safe"
691 primes. More recently (in version 1.0.2) support was provided for
692 generating X9.42 style parameter files such as those required for RFC 5114
693 support. The primes used in such files may not be "safe". Where an
694 application is using DH configured with parameters based on primes that are
695 not "safe" then an attacker could use this fact to find a peer's private
696 DH exponent. This attack requires that the attacker complete multiple
697 handshakes in which the peer uses the same private DH exponent. For example
698 this could be used to discover a TLS server's private DH exponent if it's
699 reusing the private DH exponent or it's using a static DH ciphersuite.
701 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
702 TLS. It is not on by default. If the option is not set then the server
703 reuses the same private DH exponent for the life of the server process and
704 would be vulnerable to this attack. It is believed that many popular
705 applications do set this option and would therefore not be at risk.
707 The fix for this issue adds an additional check where a "q" parameter is
708 available (as is the case in X9.42 based parameters). This detects the
709 only known attack, and is the only possible defense for static DH
710 ciphersuites. This could have some performance impact.
712 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
713 default and cannot be disabled. This could have some performance impact.
715 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
719 *) SSLv2 doesn't block disabled ciphers
721 A malicious client can negotiate SSLv2 ciphers that have been disabled on
722 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
723 been disabled, provided that the SSLv2 protocol was not also disabled via
726 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
727 and Sebastian Schinzel.
731 *) Reject DH handshakes with parameters shorter than 1024 bits.
734 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
736 *) BN_mod_exp may produce incorrect results on x86_64
738 There is a carry propagating bug in the x86_64 Montgomery squaring
739 procedure. No EC algorithms are affected. Analysis suggests that attacks
740 against RSA and DSA as a result of this defect would be very difficult to
741 perform and are not believed likely. Attacks against DH are considered just
742 feasible (although very difficult) because most of the work necessary to
743 deduce information about a private key may be performed offline. The amount
744 of resources required for such an attack would be very significant and
745 likely only accessible to a limited number of attackers. An attacker would
746 additionally need online access to an unpatched system using the target
747 private key in a scenario with persistent DH parameters and a private
748 key that is shared between multiple clients. For example this can occur by
749 default in OpenSSL DHE based SSL/TLS ciphersuites.
751 This issue was reported to OpenSSL by Hanno Böck.
755 *) Certificate verify crash with missing PSS parameter
757 The signature verification routines will crash with a NULL pointer
758 dereference if presented with an ASN.1 signature using the RSA PSS
759 algorithm and absent mask generation function parameter. Since these
760 routines are used to verify certificate signature algorithms this can be
761 used to crash any certificate verification operation and exploited in a
762 DoS attack. Any application which performs certificate verification is
763 vulnerable including OpenSSL clients and servers which enable client
766 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
770 *) X509_ATTRIBUTE memory leak
772 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
773 memory. This structure is used by the PKCS#7 and CMS routines so any
774 application which reads PKCS#7 or CMS data from untrusted sources is
775 affected. SSL/TLS is not affected.
777 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
782 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
783 This changes the decoding behaviour for some invalid messages,
784 though the change is mostly in the more lenient direction, and
785 legacy behaviour is preserved as much as possible.
788 *) In DSA_generate_parameters_ex, if the provided seed is too short,
789 use a random seed, as already documented.
790 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
792 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
794 *) Alternate chains certificate forgery
796 During certificate verfification, OpenSSL will attempt to find an
797 alternative certificate chain if the first attempt to build such a chain
798 fails. An error in the implementation of this logic can mean that an
799 attacker could cause certain checks on untrusted certificates to be
800 bypassed, such as the CA flag, enabling them to use a valid leaf
801 certificate to act as a CA and "issue" an invalid certificate.
803 This issue was reported to OpenSSL by Adam Langley/David Benjamin
808 *) Race condition handling PSK identify hint
810 If PSK identity hints are received by a multi-threaded client then
811 the values are wrongly updated in the parent SSL_CTX structure. This can
812 result in a race condition potentially leading to a double free of the
817 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
819 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
820 incompatibility in the handling of HMAC. The previous ABI has now been
823 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
825 *) Malformed ECParameters causes infinite loop
827 When processing an ECParameters structure OpenSSL enters an infinite loop
828 if the curve specified is over a specially malformed binary polynomial
831 This can be used to perform denial of service against any
832 system which processes public keys, certificate requests or
833 certificates. This includes TLS clients and TLS servers with
834 client authentication enabled.
836 This issue was reported to OpenSSL by Joseph Barr-Pixton.
840 *) Exploitable out-of-bounds read in X509_cmp_time
842 X509_cmp_time does not properly check the length of the ASN1_TIME
843 string and can read a few bytes out of bounds. In addition,
844 X509_cmp_time accepts an arbitrary number of fractional seconds in the
847 An attacker can use this to craft malformed certificates and CRLs of
848 various sizes and potentially cause a segmentation fault, resulting in
849 a DoS on applications that verify certificates or CRLs. TLS clients
850 that verify CRLs are affected. TLS clients and servers with client
851 authentication enabled may be affected if they use custom verification
854 This issue was reported to OpenSSL by Robert Swiecki (Google), and
855 independently by Hanno Böck.
859 *) PKCS7 crash with missing EnvelopedContent
861 The PKCS#7 parsing code does not handle missing inner EncryptedContent
862 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
863 with missing content and trigger a NULL pointer dereference on parsing.
865 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
866 structures from untrusted sources are affected. OpenSSL clients and
867 servers are not affected.
869 This issue was reported to OpenSSL by Michal Zalewski (Google).
873 *) CMS verify infinite loop with unknown hash function
875 When verifying a signedData message the CMS code can enter an infinite loop
876 if presented with an unknown hash function OID. This can be used to perform
877 denial of service against any system which verifies signedData messages using
879 This issue was reported to OpenSSL by Johannes Bauer.
883 *) Race condition handling NewSessionTicket
885 If a NewSessionTicket is received by a multi-threaded client when attempting to
886 reuse a previous ticket then a race condition can occur potentially leading to
887 a double free of the ticket data.
891 *) Removed support for the two export grade static DH ciphersuites
892 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
893 were newly added (along with a number of other static DH ciphersuites) to
894 1.0.2. However the two export ones have *never* worked since they were
895 introduced. It seems strange in any case to be adding new export
896 ciphersuites, and given "logjam" it also does not seem correct to fix them.
899 *) Only support 256-bit or stronger elliptic curves with the
900 'ecdh_auto' setting (server) or by default (client). Of supported
901 curves, prefer P-256 (both).
904 *) Reject DH handshakes with parameters shorter than 768 bits.
905 [Kurt Roeckx and Emilia Kasper]
907 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
909 *) ClientHello sigalgs DoS fix
911 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
912 invalid signature algorithms extension a NULL pointer dereference will
913 occur. This can be exploited in a DoS attack against the server.
915 This issue was was reported to OpenSSL by David Ramos of Stanford
918 [Stephen Henson and Matt Caswell]
920 *) Multiblock corrupted pointer fix
922 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
923 feature only applies on 64 bit x86 architecture platforms that support AES
924 NI instructions. A defect in the implementation of "multiblock" can cause
925 OpenSSL's internal write buffer to become incorrectly set to NULL when
926 using non-blocking IO. Typically, when the user application is using a
927 socket BIO for writing, this will only result in a failed connection.
928 However if some other BIO is used then it is likely that a segmentation
929 fault will be triggered, thus enabling a potential DoS attack.
931 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
935 *) Segmentation fault in DTLSv1_listen fix
937 The DTLSv1_listen function is intended to be stateless and processes the
938 initial ClientHello from many peers. It is common for user code to loop
939 over the call to DTLSv1_listen until a valid ClientHello is received with
940 an associated cookie. A defect in the implementation of DTLSv1_listen means
941 that state is preserved in the SSL object from one invocation to the next
942 that can lead to a segmentation fault. Errors processing the initial
943 ClientHello can trigger this scenario. An example of such an error could be
944 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
947 This issue was reported to OpenSSL by Per Allansson.
951 *) Segmentation fault in ASN1_TYPE_cmp fix
953 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
954 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
955 certificate signature algorithm consistency this can be used to crash any
956 certificate verification operation and exploited in a DoS attack. Any
957 application which performs certificate verification is vulnerable including
958 OpenSSL clients and servers which enable client authentication.
962 *) Segmentation fault for invalid PSS parameters fix
964 The signature verification routines will crash with a NULL pointer
965 dereference if presented with an ASN.1 signature using the RSA PSS
966 algorithm and invalid parameters. Since these routines are used to verify
967 certificate signature algorithms this can be used to crash any
968 certificate verification operation and exploited in a DoS attack. Any
969 application which performs certificate verification is vulnerable including
970 OpenSSL clients and servers which enable client authentication.
972 This issue was was reported to OpenSSL by Brian Carpenter.
976 *) ASN.1 structure reuse memory corruption fix
978 Reusing a structure in ASN.1 parsing may allow an attacker to cause
979 memory corruption via an invalid write. Such reuse is and has been
980 strongly discouraged and is believed to be rare.
982 Applications that parse structures containing CHOICE or ANY DEFINED BY
983 components may be affected. Certificate parsing (d2i_X509 and related
984 functions) are however not affected. OpenSSL clients and servers are
989 *) PKCS7 NULL pointer dereferences fix
991 The PKCS#7 parsing code does not handle missing outer ContentInfo
992 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
993 missing content and trigger a NULL pointer dereference on parsing.
995 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
996 otherwise parse PKCS#7 structures from untrusted sources are
997 affected. OpenSSL clients and servers are not affected.
999 This issue was reported to OpenSSL by Michal Zalewski (Google).
1003 *) DoS via reachable assert in SSLv2 servers fix
1005 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
1006 servers that both support SSLv2 and enable export cipher suites by sending
1007 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
1009 This issue was discovered by Sean Burford (Google) and Emilia Käsper
1010 (OpenSSL development team).
1014 *) Empty CKE with client auth and DHE fix
1016 If client auth is used then a server can seg fault in the event of a DHE
1017 ciphersuite being selected and a zero length ClientKeyExchange message
1018 being sent by the client. This could be exploited in a DoS attack.
1022 *) Handshake with unseeded PRNG fix
1024 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
1025 with an unseeded PRNG. The conditions are:
1026 - The client is on a platform where the PRNG has not been seeded
1027 automatically, and the user has not seeded manually
1028 - A protocol specific client method version has been used (i.e. not
1029 SSL_client_methodv23)
1030 - A ciphersuite is used that does not require additional random data from
1031 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
1033 If the handshake succeeds then the client random that has been used will
1034 have been generated from a PRNG with insufficient entropy and therefore the
1035 output may be predictable.
1037 For example using the following command with an unseeded openssl will
1038 succeed on an unpatched platform:
1040 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
1044 *) Use After Free following d2i_ECPrivatekey error fix
1046 A malformed EC private key file consumed via the d2i_ECPrivateKey function
1047 could cause a use after free condition. This, in turn, could cause a double
1048 free in several private key parsing functions (such as d2i_PrivateKey
1049 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
1050 for applications that receive EC private keys from untrusted
1051 sources. This scenario is considered rare.
1053 This issue was discovered by the BoringSSL project and fixed in their
1058 *) X509_to_X509_REQ NULL pointer deref fix
1060 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1061 the certificate key is invalid. This function is rarely used in practice.
1063 This issue was discovered by Brian Carpenter.
1067 *) Removed the export ciphers from the DEFAULT ciphers
1070 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1072 *) Change RSA and DH/DSA key generation apps to generate 2048-bit
1076 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1077 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1078 So far those who have to target multiple plaforms would compromise
1079 and argue that binary targeting say ARMv5 would still execute on
1080 ARMv8. "Universal" build resolves this compromise by providing
1081 near-optimal performance even on newer platforms.
1084 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1085 (other platforms pending).
1086 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1088 *) Add support for the SignedCertificateTimestampList certificate and
1089 OCSP response extensions from RFC6962.
1092 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1093 for corner cases. (Certain input points at infinity could lead to
1094 bogus results, with non-infinity inputs mapped to infinity too.)
1097 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1098 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1099 common cases are optimized and there still is room for further
1100 improvements. Vector Permutation AES for Altivec is also added.
1103 *) Add support for little-endian ppc64 Linux target.
1104 [Marcelo Cerri (IBM)]
1106 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1107 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1108 are optimized and there still is room for further improvements.
1109 Both 32- and 64-bit modes are supported.
1110 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1112 *) Improved ARMv7 NEON support.
1115 *) Support for SPARC Architecture 2011 crypto extensions, first
1116 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1117 SHA256/512, MD5, GHASH and modular exponentiation.
1118 [Andy Polyakov, David Miller]
1120 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1122 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1124 *) Support for new and upcoming Intel processors, including AVX2,
1125 BMI and SHA ISA extensions. This includes additional "stitched"
1126 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1129 This work was sponsored by Intel Corp.
1132 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1133 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1134 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1137 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1138 this fixes a limiation in previous versions of OpenSSL.
1141 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1142 MGF1 digest and OAEP label.
1145 *) Add EVP support for key wrapping algorithms, to avoid problems with
1146 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1147 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1148 algorithms and include tests cases.
1151 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1153 [Douglas E. Engert, Steve Henson]
1155 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1156 difference in days and seconds between two tm or ASN1_TIME structures.
1159 *) Add -rev test option to s_server to just reverse order of characters
1160 received by client and send back to server. Also prints an abbreviated
1161 summary of the connection parameters.
1164 *) New option -brief for s_client and s_server to print out a brief summary
1165 of connection parameters.
1168 *) Add callbacks for arbitrary TLS extensions.
1169 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1171 *) New option -crl_download in several openssl utilities to download CRLs
1172 from CRLDP extension in certificates.
1175 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1178 *) New function X509_CRL_diff to generate a delta CRL from the difference
1179 of two full CRLs. Add support to "crl" utility.
1182 *) New functions to set lookup_crls function and to retrieve
1183 X509_STORE from X509_STORE_CTX.
1186 *) Print out deprecated issuer and subject unique ID fields in
1190 *) Extend OCSP I/O functions so they can be used for simple general purpose
1191 HTTP as well as OCSP. New wrapper function which can be used to download
1192 CRLs using the OCSP API.
1195 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1198 *) SSL_CONF* functions. These provide a common framework for application
1199 configuration using configuration files or command lines.
1202 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1203 message callback and prints the results. Needs compile time option
1204 "enable-ssl-trace". New options to s_client and s_server to enable
1208 *) New ctrl and macro to retrieve supported points extensions.
1209 Print out extension in s_server and s_client.
1212 *) New functions to retrieve certificate signature and signature
1216 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1220 *) New Suite B modes for TLS code. These use and enforce the requirements
1221 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1222 only use Suite B curves. The Suite B modes can be set by using the
1223 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1226 *) New chain verification flags for Suite B levels of security. Check
1227 algorithms are acceptable when flags are set in X509_verify_cert.
1230 *) Make tls1_check_chain return a set of flags indicating checks passed
1231 by a certificate chain. Add additional tests to handle client
1232 certificates: checks for matching certificate type and issuer name
1236 *) If an attempt is made to use a signature algorithm not in the peer
1237 preference list abort the handshake. If client has no suitable
1238 signature algorithms in response to a certificate request do not
1239 use the certificate.
1242 *) If server EC tmp key is not in client preference list abort handshake.
1245 *) Add support for certificate stores in CERT structure. This makes it
1246 possible to have different stores per SSL structure or one store in
1247 the parent SSL_CTX. Include distint stores for certificate chain
1248 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1249 to build and store a certificate chain in CERT structure: returing
1250 an error if the chain cannot be built: this will allow applications
1251 to test if a chain is correctly configured.
1253 Note: if the CERT based stores are not set then the parent SSL_CTX
1254 store is used to retain compatibility with existing behaviour.
1258 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1259 mask based on the current session, check mask when sending client
1260 hello and checking the requested ciphersuite.
1263 *) New ctrls to retrieve and set certificate types in a certificate
1264 request message. Print out received values in s_client. If certificate
1265 types is not set with custom values set sensible values based on
1266 supported signature algorithms.
1269 *) Support for distinct client and server supported signature algorithms.
1272 *) Add certificate callback. If set this is called whenever a certificate
1273 is required by client or server. An application can decide which
1274 certificate chain to present based on arbitrary criteria: for example
1275 supported signature algorithms. Add very simple example to s_server.
1276 This fixes many of the problems and restrictions of the existing client
1277 certificate callback: for example you can now clear an existing
1278 certificate and specify the whole chain.
1281 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1282 the certificate can be used for (if anything). Set valid_flags field
1283 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1284 to have similar checks in it.
1286 Add new "cert_flags" field to CERT structure and include a "strict mode".
1287 This enforces some TLS certificate requirements (such as only permitting
1288 certificate signature algorithms contained in the supported algorithms
1289 extension) which some implementations ignore: this option should be used
1290 with caution as it could cause interoperability issues.
1293 *) Update and tidy signature algorithm extension processing. Work out
1294 shared signature algorithms based on preferences and peer algorithms
1295 and print them out in s_client and s_server. Abort handshake if no
1296 shared signature algorithms.
1299 *) Add new functions to allow customised supported signature algorithms
1300 for SSL and SSL_CTX structures. Add options to s_client and s_server
1304 *) New function SSL_certs_clear() to delete all references to certificates
1305 from an SSL structure. Before this once a certificate had been added
1306 it couldn't be removed.
1309 *) Integrate hostname, email address and IP address checking with certificate
1310 verification. New verify options supporting checking in opensl utility.
1313 *) Fixes and wildcard matching support to hostname and email checking
1314 functions. Add manual page.
1315 [Florian Weimer (Red Hat Product Security Team)]
1317 *) New functions to check a hostname email or IP address against a
1318 certificate. Add options x509 utility to print results of checks against
1322 *) Fix OCSP checking.
1323 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1325 *) Initial experimental support for explicitly trusted non-root CAs.
1326 OpenSSL still tries to build a complete chain to a root but if an
1327 intermediate CA has a trust setting included that is used. The first
1328 setting is used: whether to trust (e.g., -addtrust option to the x509
1332 *) Add -trusted_first option which attempts to find certificates in the
1333 trusted store even if an untrusted chain is also supplied.
1336 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1337 platform support for Linux and Android.
1340 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1343 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1344 When in FIPS mode the approved implementations are used as normal,
1345 when not in FIPS mode the internal unapproved versions are used instead.
1346 This means that the FIPS capable OpenSSL isn't forced to use the
1347 (often lower perfomance) FIPS implementations outside FIPS mode.
1350 *) Transparently support X9.42 DH parameters when calling
1351 PEM_read_bio_DHparameters. This means existing applications can handle
1352 the new parameter format automatically.
1355 *) Initial experimental support for X9.42 DH parameter format: mainly
1356 to support use of 'q' parameter for RFC5114 parameters.
1359 *) Add DH parameters from RFC5114 including test data to dhtest.
1362 *) Support for automatic EC temporary key parameter selection. If enabled
1363 the most preferred EC parameters are automatically used instead of
1364 hardcoded fixed parameters. Now a server just has to call:
1365 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1366 support ECDH and use the most appropriate parameters.
1369 *) Enhance and tidy EC curve and point format TLS extension code. Use
1370 static structures instead of allocation if default values are used.
1371 New ctrls to set curves we wish to support and to retrieve shared curves.
1372 Print out shared curves in s_server. New options to s_server and s_client
1373 to set list of supported curves.
1376 *) New ctrls to retrieve supported signature algorithms and
1377 supported curve values as an array of NIDs. Extend openssl utility
1378 to print out received values.
1381 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1382 between NIDs and the more common NIST names such as "P-256". Enhance
1383 ecparam utility and ECC method to recognise the NIST names for curves.
1386 *) Enhance SSL/TLS certificate chain handling to support different
1387 chains for each certificate instead of one chain in the parent SSL_CTX.
1390 *) Support for fixed DH ciphersuite client authentication: where both
1391 server and client use DH certificates with common parameters.
1394 *) Support for fixed DH ciphersuites: those requiring DH server
1398 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1400 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1401 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1402 X509_CINF_get_signature were reverted post internal team review.
1404 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1406 *) Build fixes for the Windows and OpenVMS platforms
1407 [Matt Caswell and Richard Levitte]
1409 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1411 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1412 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1413 dereference. This could lead to a Denial Of Service attack. Thanks to
1414 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1418 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1419 dtls1_buffer_record function under certain conditions. In particular this
1420 could occur if an attacker sent repeated DTLS records with the same
1421 sequence number but for the next epoch. The memory leak could be exploited
1422 by an attacker in a Denial of Service attack through memory exhaustion.
1423 Thanks to Chris Mueller for reporting this issue.
1427 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1428 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1429 method would be set to NULL which could later result in a NULL pointer
1430 dereference. Thanks to Frank Schmirler for reporting this issue.
1434 *) Abort handshake if server key exchange message is omitted for ephemeral
1437 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1438 reporting this issue.
1442 *) Remove non-export ephemeral RSA code on client and server. This code
1443 violated the TLS standard by allowing the use of temporary RSA keys in
1444 non-export ciphersuites and could be used by a server to effectively
1445 downgrade the RSA key length used to a value smaller than the server
1446 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1447 INRIA or reporting this issue.
1451 *) Fixed issue where DH client certificates are accepted without verification.
1452 An OpenSSL server will accept a DH certificate for client authentication
1453 without the certificate verify message. This effectively allows a client to
1454 authenticate without the use of a private key. This only affects servers
1455 which trust a client certificate authority which issues certificates
1456 containing DH keys: these are extremely rare and hardly ever encountered.
1457 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1462 *) Ensure that the session ID context of an SSL is updated when its
1463 SSL_CTX is updated via SSL_set_SSL_CTX.
1465 The session ID context is typically set from the parent SSL_CTX,
1466 and can vary with the CTX.
1469 *) Fix various certificate fingerprint issues.
1471 By using non-DER or invalid encodings outside the signed portion of a
1472 certificate the fingerprint can be changed without breaking the signature.
1473 Although no details of the signed portion of the certificate can be changed
1474 this can cause problems with some applications: e.g. those using the
1475 certificate fingerprint for blacklists.
1477 1. Reject signatures with non zero unused bits.
1479 If the BIT STRING containing the signature has non zero unused bits reject
1480 the signature. All current signature algorithms require zero unused bits.
1482 2. Check certificate algorithm consistency.
1484 Check the AlgorithmIdentifier inside TBS matches the one in the
1485 certificate signature. NB: this will result in signature failure
1486 errors for some broken certificates.
1488 Thanks to Konrad Kraszewski from Google for reporting this issue.
1490 3. Check DSA/ECDSA signatures use DER.
1492 Reencode DSA/ECDSA signatures and compare with the original received
1493 signature. Return an error if there is a mismatch.
1495 This will reject various cases including garbage after signature
1496 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1497 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1498 (negative or with leading zeroes).
1500 Further analysis was conducted and fixes were developed by Stephen Henson
1501 of the OpenSSL core team.
1506 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1507 results on some platforms, including x86_64. This bug occurs at random
1508 with a very low probability, and is not known to be exploitable in any
1509 way, though its exact impact is difficult to determine. Thanks to Pieter
1510 Wuille (Blockstream) who reported this issue and also suggested an initial
1511 fix. Further analysis was conducted by the OpenSSL development team and
1512 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1513 the OpenSSL core team.
1517 *) Do not resume sessions on the server if the negotiated protocol
1518 version does not match the session's version. Resuming with a different
1519 version, while not strictly forbidden by the RFC, is of questionable
1520 sanity and breaks all known clients.
1521 [David Benjamin, Emilia Käsper]
1523 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1524 early CCS messages during renegotiation. (Note that because
1525 renegotiation is encrypted, this early CCS was not exploitable.)
1528 *) Tighten client-side session ticket handling during renegotiation:
1529 ensure that the client only accepts a session ticket if the server sends
1530 the extension anew in the ServerHello. Previously, a TLS client would
1531 reuse the old extension state and thus accept a session ticket if one was
1532 announced in the initial ServerHello.
1534 Similarly, ensure that the client requires a session ticket if one
1535 was advertised in the ServerHello. Previously, a TLS client would
1536 ignore a missing NewSessionTicket message.
1539 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1541 *) SRTP Memory Leak.
1543 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1544 sends a carefully crafted handshake message, to cause OpenSSL to fail
1545 to free up to 64k of memory causing a memory leak. This could be
1546 exploited in a Denial Of Service attack. This issue affects OpenSSL
1547 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1548 whether SRTP is used or configured. Implementations of OpenSSL that
1549 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1551 The fix was developed by the OpenSSL team.
1555 *) Session Ticket Memory Leak.
1557 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1558 integrity of that ticket is first verified. In the event of a session
1559 ticket integrity check failing, OpenSSL will fail to free memory
1560 causing a memory leak. By sending a large number of invalid session
1561 tickets an attacker could exploit this issue in a Denial Of Service
1566 *) Build option no-ssl3 is incomplete.
1568 When OpenSSL is configured with "no-ssl3" as a build option, servers
1569 could accept and complete a SSL 3.0 handshake, and clients could be
1570 configured to send them.
1572 [Akamai and the OpenSSL team]
1574 *) Add support for TLS_FALLBACK_SCSV.
1575 Client applications doing fallback retries should call
1576 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1578 [Adam Langley, Bodo Moeller]
1580 *) Add additional DigestInfo checks.
1582 Reencode DigestInto in DER and check against the original when
1583 verifying RSA signature: this will reject any improperly encoded
1584 DigestInfo structures.
1586 Note: this is a precautionary measure and no attacks are currently known.
1590 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1592 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1593 SRP code can be overrun an internal buffer. Add sanity check that
1594 g, A, B < N to SRP code.
1596 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1597 Group for discovering this issue.
1601 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1602 TLS 1.0 instead of higher protocol versions when the ClientHello message
1603 is badly fragmented. This allows a man-in-the-middle attacker to force a
1604 downgrade to TLS 1.0 even if both the server and the client support a
1605 higher protocol version, by modifying the client's TLS records.
1607 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1608 researching this issue.
1612 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1613 to a denial of service attack. A malicious server can crash the client
1614 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1615 ciphersuite and sending carefully crafted handshake messages.
1617 Thanks to Felix Gröbert (Google) for discovering and researching this
1622 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1623 to leak memory. This can be exploited through a Denial of Service attack.
1624 Thanks to Adam Langley for discovering and researching this issue.
1628 *) An attacker can force openssl to consume large amounts of memory whilst
1629 processing DTLS handshake messages. This can be exploited through a
1630 Denial of Service attack.
1631 Thanks to Adam Langley for discovering and researching this issue.
1635 *) An attacker can force an error condition which causes openssl to crash
1636 whilst processing DTLS packets due to memory being freed twice. This
1637 can be exploited through a Denial of Service attack.
1638 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1643 *) If a multithreaded client connects to a malicious server using a resumed
1644 session and the server sends an ec point format extension it could write
1645 up to 255 bytes to freed memory.
1647 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1652 *) A malicious server can crash an OpenSSL client with a null pointer
1653 dereference (read) by specifying an SRP ciphersuite even though it was not
1654 properly negotiated with the client. This can be exploited through a
1655 Denial of Service attack.
1657 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1658 discovering and researching this issue.
1662 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1663 X509_name_oneline, X509_name_print_ex et al. to leak some information
1664 from the stack. Applications may be affected if they echo pretty printing
1665 output to the attacker.
1667 Thanks to Ivan Fratric (Google) for discovering this issue.
1669 [Emilia Käsper, and Steve Henson]
1671 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1672 for corner cases. (Certain input points at infinity could lead to
1673 bogus results, with non-infinity inputs mapped to infinity too.)
1676 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1678 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1679 handshake can force the use of weak keying material in OpenSSL
1680 SSL/TLS clients and servers.
1682 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1683 researching this issue. (CVE-2014-0224)
1684 [KIKUCHI Masashi, Steve Henson]
1686 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1687 OpenSSL DTLS client the code can be made to recurse eventually crashing
1690 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1692 [Imre Rad, Steve Henson]
1694 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1695 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1696 client or server. This is potentially exploitable to run arbitrary
1697 code on a vulnerable client or server.
1699 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1700 [Jüri Aedla, Steve Henson]
1702 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1703 are subject to a denial of service attack.
1705 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1706 this issue. (CVE-2014-3470)
1707 [Felix Gröbert, Ivan Fratric, Steve Henson]
1709 *) Harmonize version and its documentation. -f flag is used to display
1711 [mancha <mancha1@zoho.com>]
1713 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1714 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1715 [mancha <mancha1@zoho.com>]
1717 *) Fix some double frees. These are not thought to be exploitable.
1718 [mancha <mancha1@zoho.com>]
1720 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1722 *) A missing bounds check in the handling of the TLS heartbeat extension
1723 can be used to reveal up to 64k of memory to a connected client or
1726 Thanks for Neel Mehta of Google Security for discovering this bug and to
1727 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1728 preparing the fix (CVE-2014-0160)
1729 [Adam Langley, Bodo Moeller]
1731 *) Fix for the attack described in the paper "Recovering OpenSSL
1732 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1733 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1734 http://eprint.iacr.org/2014/140
1736 Thanks to Yuval Yarom and Naomi Benger for discovering this
1737 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1738 [Yuval Yarom and Naomi Benger]
1740 *) TLS pad extension: draft-agl-tls-padding-03
1742 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1743 TLS client Hello record length value would otherwise be > 255 and
1744 less that 512 pad with a dummy extension containing zeroes so it
1745 is at least 512 bytes long.
1747 [Adam Langley, Steve Henson]
1749 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1751 *) Fix for TLS record tampering bug. A carefully crafted invalid
1752 handshake could crash OpenSSL with a NULL pointer exception.
1753 Thanks to Anton Johansson for reporting this issues.
1756 *) Keep original DTLS digest and encryption contexts in retransmission
1757 structures so we can use the previous session parameters if they need
1758 to be resent. (CVE-2013-6450)
1761 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1762 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1763 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1764 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1765 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1766 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1767 [Rob Stradling, Adam Langley]
1769 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1771 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1772 supporting platforms or when small records were transferred.
1773 [Andy Polyakov, Steve Henson]
1775 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1777 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1779 This addresses the flaw in CBC record processing discovered by
1780 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1781 at: http://www.isg.rhul.ac.uk/tls/
1783 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1784 Security Group at Royal Holloway, University of London
1785 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1786 Emilia Käsper for the initial patch.
1788 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1790 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1791 ciphersuites which can be exploited in a denial of service attack.
1792 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1793 and detecting this bug and to Wolfgang Ettlinger
1794 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1798 *) Return an error when checking OCSP signatures when key is NULL.
1799 This fixes a DoS attack. (CVE-2013-0166)
1802 *) Make openssl verify return errors.
1803 [Chris Palmer <palmer@google.com> and Ben Laurie]
1805 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1806 the right response is stapled. Also change SSL_get_certificate()
1807 so it returns the certificate actually sent.
1808 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1809 [Rob Stradling <rob.stradling@comodo.com>]
1811 *) Fix possible deadlock when decoding public keys.
1814 *) Don't use TLS 1.0 record version number in initial client hello
1818 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1820 *) Sanity check record length before skipping explicit IV in TLS
1821 1.2, 1.1 and DTLS to fix DoS attack.
1823 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1824 fuzzing as a service testing platform.
1828 *) Initialise tkeylen properly when encrypting CMS messages.
1829 Thanks to Solar Designer of Openwall for reporting this issue.
1832 *) In FIPS mode don't try to use composite ciphers as they are not
1836 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1838 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1839 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1840 mean any application compiled against OpenSSL 1.0.0 headers setting
1841 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1842 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1843 0x10000000L Any application which was previously compiled against
1844 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1845 will need to be recompiled as a result. Letting be results in
1846 inability to disable specifically TLS 1.1 and in client context,
1847 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1850 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1851 disable just protocol X, but all protocols above X *if* there are
1852 protocols *below* X still enabled. In more practical terms it means
1853 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1854 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1855 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1859 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1861 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1862 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1863 in CRYPTO_realloc_clean.
1865 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1866 issue and to Adam Langley <agl@chromium.org> for fixing it.
1868 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1870 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1873 *) Workarounds for some broken servers that "hang" if a client hello
1874 record length exceeds 255 bytes.
1876 1. Do not use record version number > TLS 1.0 in initial client
1877 hello: some (but not all) hanging servers will now work.
1878 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1879 the number of ciphers sent in the client hello. This should be
1880 set to an even number, such as 50, for example by passing:
1881 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1882 Most broken servers should now work.
1883 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1884 TLS 1.2 client support entirely.
1887 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1890 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1892 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1893 STRING form instead of a DigestInfo.
1896 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1897 and the RSA_sign/RSA_verify functions. This was made more apparent when
1898 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1899 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1900 the correct format in RSA_verify so both forms transparently work.
1903 *) Some servers which support TLS 1.0 can choke if we initially indicate
1904 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1905 encrypted premaster secret. As a workaround use the maximum pemitted
1906 client version in client hello, this should keep such servers happy
1907 and still work with previous versions of OpenSSL.
1910 *) Add support for TLS/DTLS heartbeats.
1911 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1913 *) Add support for SCTP.
1914 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1916 *) Improved PRNG seeding for VOS.
1917 [Paul Green <Paul.Green@stratus.com>]
1919 *) Extensive assembler packs updates, most notably:
1921 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1922 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1923 - x86_64: bit-sliced AES implementation;
1924 - ARM: NEON support, contemporary platforms optimizations;
1925 - s390x: z196 support;
1926 - *: GHASH and GF(2^m) multiplication implementations;
1930 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1931 (removal of unnecessary code)
1932 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1934 *) Add TLS key material exporter from RFC 5705.
1937 *) Add DTLS-SRTP negotiation from RFC 5764.
1940 *) Add Next Protocol Negotiation,
1941 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1942 disabled with a no-npn flag to config or Configure. Code donated
1944 [Adam Langley <agl@google.com> and Ben Laurie]
1946 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1947 NIST-P256, NIST-P521, with constant-time single point multiplication on
1948 typical inputs. Compiler support for the nonstandard type __uint128_t is
1949 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1950 Code made available under Apache License version 2.0.
1952 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1953 line to include this in your build of OpenSSL, and run "make depend" (or
1954 "make update"). This enables the following EC_METHODs:
1956 EC_GFp_nistp224_method()
1957 EC_GFp_nistp256_method()
1958 EC_GFp_nistp521_method()
1960 EC_GROUP_new_by_curve_name() will automatically use these (while
1961 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1963 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1965 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1966 all platforms. Move ssize_t definition from e_os.h to the public
1967 header file e_os2.h as it now appears in public header file cms.h
1970 *) New -sigopt option to the ca, req and x509 utilities. Additional
1971 signature parameters can be passed using this option and in
1975 *) Add RSA PSS signing function. This will generate and set the
1976 appropriate AlgorithmIdentifiers for PSS based on those in the
1977 corresponding EVP_MD_CTX structure. No application support yet.
1980 *) Support for companion algorithm specific ASN1 signing routines.
1981 New function ASN1_item_sign_ctx() signs a pre-initialised
1982 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1983 the appropriate parameters.
1986 *) Add new algorithm specific ASN1 verification initialisation function
1987 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1988 handling will be the same no matter what EVP_PKEY_METHOD is used.
1989 Add a PSS handler to support verification of PSS signatures: checked
1990 against a number of sample certificates.
1993 *) Add signature printing for PSS. Add PSS OIDs.
1994 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1996 *) Add algorithm specific signature printing. An individual ASN1 method
1997 can now print out signatures instead of the standard hex dump.
1999 More complex signatures (e.g. PSS) can print out more meaningful
2000 information. Include DSA version that prints out the signature
2004 *) Password based recipient info support for CMS library: implementing
2008 *) Split password based encryption into PBES2 and PBKDF2 functions. This
2009 neatly separates the code into cipher and PBE sections and is required
2010 for some algorithms that split PBES2 into separate pieces (such as
2011 password based CMS).
2014 *) Session-handling fixes:
2015 - Fix handling of connections that are resuming with a session ID,
2016 but also support Session Tickets.
2017 - Fix a bug that suppressed issuing of a new ticket if the client
2018 presented a ticket with an expired session.
2019 - Try to set the ticket lifetime hint to something reasonable.
2020 - Make tickets shorter by excluding irrelevant information.
2021 - On the client side, don't ignore renewed tickets.
2022 [Adam Langley, Bodo Moeller (Google)]
2024 *) Fix PSK session representation.
2027 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
2029 This work was sponsored by Intel.
2032 *) Add GCM support to TLS library. Some custom code is needed to split
2033 the IV between the fixed (from PRF) and explicit (from TLS record)
2034 portions. This adds all GCM ciphersuites supported by RFC5288 and
2035 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
2036 add a special AESGCM string for GCM only.
2039 *) Expand range of ctrls for AES GCM. Permit setting invocation
2040 field on decrypt and retrieval of invocation field only on encrypt.
2043 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
2044 As required by RFC5289 these ciphersuites cannot be used if for
2045 versions of TLS earlier than 1.2.
2048 *) For FIPS capable OpenSSL interpret a NULL default public key method
2049 as unset and return the appopriate default but do *not* set the default.
2050 This means we can return the appopriate method in applications that
2051 swicth between FIPS and non-FIPS modes.
2054 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
2055 ENGINE is used then we cannot handle that in the FIPS module so we
2056 keep original code iff non-FIPS operations are allowed.
2059 *) Add -attime option to openssl utilities.
2060 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2062 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2065 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2066 FIPS EC methods unconditionally for now.
2069 *) New build option no-ec2m to disable characteristic 2 code.
2072 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2073 all cases can be covered as some introduce binary incompatibilities.
2076 *) Redirect RSA operations to FIPS module including keygen,
2077 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2080 *) Add similar low level API blocking to ciphers.
2083 *) Low level digest APIs are not approved in FIPS mode: any attempt
2084 to use these will cause a fatal error. Applications that *really* want
2085 to use them can use the private_* version instead.
2088 *) Redirect cipher operations to FIPS module for FIPS builds.
2091 *) Redirect digest operations to FIPS module for FIPS builds.
2094 *) Update build system to add "fips" flag which will link in fipscanister.o
2095 for static and shared library builds embedding a signature if needed.
2098 *) Output TLS supported curves in preference order instead of numerical
2099 order. This is currently hardcoded for the highest order curves first.
2100 This should be configurable so applications can judge speed vs strength.
2103 *) Add TLS v1.2 server support for client authentication.
2106 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2110 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2111 FIPS modules versions.
2114 *) Add TLS v1.2 client side support for client authentication. Keep cache
2115 of handshake records longer as we don't know the hash algorithm to use
2116 until after the certificate request message is received.
2119 *) Initial TLS v1.2 client support. Add a default signature algorithms
2120 extension including all the algorithms we support. Parse new signature
2121 format in client key exchange. Relax some ECC signing restrictions for
2122 TLS v1.2 as indicated in RFC5246.
2125 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2126 to new signature format when needed using client digest preference.
2127 All server ciphersuites should now work correctly in TLS v1.2. No client
2128 support yet and no support for client certificates.
2131 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2132 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2133 ciphersuites. At present only RSA key exchange ciphersuites work with
2134 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2135 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2136 and version checking.
2139 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2140 with this defined it will not be affected by any changes to ssl internal
2141 structures. Add several utility functions to allow openssl application
2142 to work with OPENSSL_NO_SSL_INTERN defined.
2145 *) A long standing patch to add support for SRP from EdelWeb (Peter
2146 Sylvester and Christophe Renou) was integrated.
2147 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
2148 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
2151 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2154 *) Permit abbreviated handshakes when renegotiating using the function
2155 SSL_renegotiate_abbreviated().
2156 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2158 *) Add call to ENGINE_register_all_complete() to
2159 ENGINE_load_builtin_engines(), so some implementations get used
2160 automatically instead of needing explicit application support.
2163 *) Add support for TLS key exporter as described in RFC5705.
2164 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2166 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2167 a few changes are required:
2169 Add SSL_OP_NO_TLSv1_1 flag.
2170 Add TLSv1_1 methods.
2171 Update version checking logic to handle version 1.1.
2172 Add explicit IV handling (ported from DTLS code).
2173 Add command line options to s_client/s_server.
2176 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2178 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2179 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2180 content decryption and always return the same error. Note: this attack
2181 needs on average 2^20 messages so it only affects automated senders. The
2182 old behaviour can be reenabled in the CMS code by setting the
2183 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2184 an MMA defence is not necessary.
2185 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2186 this issue. (CVE-2012-0884)
2189 *) Fix CVE-2011-4619: make sure we really are receiving a
2190 client hello before rejecting multiple SGC restarts. Thanks to
2191 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2194 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2196 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2197 Thanks to Antonio Martin, Enterprise Secure Access Research and
2198 Development, Cisco Systems, Inc. for discovering this bug and
2199 preparing a fix. (CVE-2012-0050)
2202 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2204 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2205 of the Vaudenay padding oracle attack on CBC mode encryption
2206 which enables an efficient plaintext recovery attack against
2207 the OpenSSL implementation of DTLS. Their attack exploits timing
2208 differences arising during decryption processing. A research
2209 paper describing this attack can be found at:
2210 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2211 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2212 Security Group at Royal Holloway, University of London
2213 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2214 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2215 for preparing the fix. (CVE-2011-4108)
2216 [Robin Seggelmann, Michael Tuexen]
2218 *) Clear bytes used for block padding of SSL 3.0 records.
2220 [Adam Langley (Google)]
2222 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2223 Kadianakis <desnacked@gmail.com> for discovering this issue and
2224 Adam Langley for preparing the fix. (CVE-2011-4619)
2225 [Adam Langley (Google)]
2227 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2228 [Andrey Kulikov <amdeich@gmail.com>]
2230 *) Prevent malformed RFC3779 data triggering an assertion failure.
2231 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2232 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2233 [Rob Austein <sra@hactrn.net>]
2235 *) Improved PRNG seeding for VOS.
2236 [Paul Green <Paul.Green@stratus.com>]
2238 *) Fix ssl_ciph.c set-up race.
2239 [Adam Langley (Google)]
2241 *) Fix spurious failures in ecdsatest.c.
2242 [Emilia Käsper (Google)]
2244 *) Fix the BIO_f_buffer() implementation (which was mixing different
2245 interpretations of the '..._len' fields).
2246 [Adam Langley (Google)]
2248 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2249 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2250 threads won't reuse the same blinding coefficients.
2252 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2253 lock to call BN_BLINDING_invert_ex, and avoids one use of
2254 BN_BLINDING_update for each BN_BLINDING structure (previously,
2255 the last update always remained unused).
2256 [Emilia Käsper (Google)]
2258 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2259 [Bob Buckholz (Google)]
2261 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2263 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2264 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2265 [Kaspar Brand <ossl@velox.ch>]
2267 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2268 for multi-threaded use of ECDH. (CVE-2011-3210)
2269 [Adam Langley (Google)]
2271 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2274 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2275 signature public key algorithm by using OID xref utilities instead.
2276 Before this you could only use some ECC ciphersuites with SHA1 only.
2279 *) Add protection against ECDSA timing attacks as mentioned in the paper
2280 by Billy Bob Brumley and Nicola Tuveri, see:
2282 http://eprint.iacr.org/2011/232.pdf
2284 [Billy Bob Brumley and Nicola Tuveri]
2286 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2288 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2289 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2291 *) Fix bug in string printing code: if *any* escaping is enabled we must
2292 escape the escape character (backslash) or the resulting string is
2296 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2298 *) Disable code workaround for ancient and obsolete Netscape browsers
2299 and servers: an attacker can use it in a ciphersuite downgrade attack.
2300 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2303 *) Fixed J-PAKE implementation error, originally discovered by
2304 Sebastien Martini, further info and confirmation from Stefan
2305 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2308 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2310 *) Fix extension code to avoid race conditions which can result in a buffer
2311 overrun vulnerability: resumed sessions must not be modified as they can
2312 be shared by multiple threads. CVE-2010-3864
2315 *) Fix WIN32 build system to correctly link an ENGINE directory into
2319 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2321 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2323 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2325 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2327 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2328 context. The operation can be customised via the ctrl mechanism in
2329 case ENGINEs want to include additional functionality.
2332 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2335 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2336 output hashes compatible with older versions of OpenSSL.
2337 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2339 *) Fix compression algorithm handling: if resuming a session use the
2340 compression algorithm of the resumed session instead of determining
2341 it from client hello again. Don't allow server to change algorithm.
2344 *) Add load_crls() function to apps tidying load_certs() too. Add option
2345 to verify utility to allow additional CRLs to be included.
2348 *) Update OCSP request code to permit adding custom headers to the request:
2349 some responders need this.
2352 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2354 [Julia Lawall <julia@diku.dk>]
2356 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2357 needlessly dereferenced structures, used obsolete functions and
2358 didn't handle all updated verify codes correctly.
2361 *) Disable MD2 in the default configuration.
2364 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2365 indicate the initial BIO being pushed or popped. This makes it possible
2366 to determine whether the BIO is the one explicitly called or as a result
2367 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2368 it handles reference counts correctly and doesn't zero out the I/O bio
2369 when it is not being explicitly popped. WARNING: applications which
2370 included workarounds for the old buggy behaviour will need to be modified
2371 or they could free up already freed BIOs.
2374 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2375 renaming to all platforms (within the 0.9.8 branch, this was
2376 done conditionally on Netware platforms to avoid a name clash).
2377 [Guenter <lists@gknw.net>]
2379 *) Add ECDHE and PSK support to DTLS.
2380 [Michael Tuexen <tuexen@fh-muenster.de>]
2382 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2386 *) Add "missing" function EVP_MD_flags() (without this the only way to
2387 retrieve a digest flags is by accessing the structure directly. Update
2388 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2389 or cipher is registered as in the "from" argument. Print out all
2390 registered digests in the dgst usage message instead of manually
2391 attempting to work them out.
2394 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2395 this allows the use of compression and extensions. Change default cipher
2396 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2397 by default unless an application cipher string requests it.
2400 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2401 key ids to find matching certificates and keys but some PKCS#12 files
2402 don't follow the (somewhat unwritten) rules and this strategy fails.
2403 Now just gather all certificates together and the first private key
2404 then look for the first certificate that matches the key.
2407 *) Support use of registered digest and cipher names for dgst and cipher
2408 commands instead of having to add each one as a special case. So now
2415 openssl dgst -sha256 foo
2417 and this works for ENGINE based algorithms too.
2421 *) Update Gost ENGINE to support parameter files.
2422 [Victor B. Wagner <vitus@cryptocom.ru>]
2424 *) Support GeneralizedTime in ca utility.
2425 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2427 *) Enhance the hash format used for certificate directory links. The new
2428 form uses the canonical encoding (meaning equivalent names will work
2429 even if they aren't identical) and uses SHA1 instead of MD5. This form
2430 is incompatible with the older format and as a result c_rehash should
2431 be used to rebuild symbolic links.
2434 *) Make PKCS#8 the default write format for private keys, replacing the
2435 traditional format. This form is standardised, more secure and doesn't
2436 include an implicit MD5 dependency.
2439 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2440 committed to OpenSSL should pass this lot as a minimum.
2443 *) Add session ticket override functionality for use by EAP-FAST.
2444 [Jouni Malinen <j@w1.fi>]
2446 *) Modify HMAC functions to return a value. Since these can be implemented
2447 in an ENGINE errors can occur.
2450 *) Type-checked OBJ_bsearch_ex.
2453 *) Type-checked OBJ_bsearch. Also some constification necessitated
2454 by type-checking. Still to come: TXT_DB, bsearch(?),
2455 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2459 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2460 seconds to a tm structure directly, instead of going through OS
2461 specific date routines. This avoids any issues with OS routines such
2462 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2463 and X509_time_adj_ex() to cover the extended range. The existing
2464 X509_time_adj() is still usable and will no longer have any date issues.
2467 *) Delta CRL support. New use deltas option which will attempt to locate
2468 and search any appropriate delta CRLs available.
2470 This work was sponsored by Google.
2473 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2474 code and add additional score elements. Validate alternate CRL paths
2475 as part of the CRL checking and indicate a new error "CRL path validation
2476 error" in this case. Applications wanting additional details can use
2477 the verify callback and check the new "parent" field. If this is not
2478 NULL CRL path validation is taking place. Existing applications wont
2479 see this because it requires extended CRL support which is off by
2482 This work was sponsored by Google.
2485 *) Support for freshest CRL extension.
2487 This work was sponsored by Google.
2490 *) Initial indirect CRL support. Currently only supported in the CRLs
2491 passed directly and not via lookup. Process certificate issuer
2492 CRL entry extension and lookup CRL entries by bother issuer name
2493 and serial number. Check and process CRL issuer entry in IDP extension.
2495 This work was sponsored by Google.
2498 *) Add support for distinct certificate and CRL paths. The CRL issuer
2499 certificate is validated separately in this case. Only enabled if
2500 an extended CRL support flag is set: this flag will enable additional
2501 CRL functionality in future.
2503 This work was sponsored by Google.
2506 *) Add support for policy mappings extension.
2508 This work was sponsored by Google.
2511 *) Fixes to pathlength constraint, self issued certificate handling,
2512 policy processing to align with RFC3280 and PKITS tests.
2514 This work was sponsored by Google.
2517 *) Support for name constraints certificate extension. DN, email, DNS
2518 and URI types are currently supported.
2520 This work was sponsored by Google.
2523 *) To cater for systems that provide a pointer-based thread ID rather
2524 than numeric, deprecate the current numeric thread ID mechanism and
2525 replace it with a structure and associated callback type. This
2526 mechanism allows a numeric "hash" to be extracted from a thread ID in
2527 either case, and on platforms where pointers are larger than 'long',
2528 mixing is done to help ensure the numeric 'hash' is usable even if it
2529 can't be guaranteed unique. The default mechanism is to use "&errno"
2530 as a pointer-based thread ID to distinguish between threads.
2532 Applications that want to provide their own thread IDs should now use
2533 CRYPTO_THREADID_set_callback() to register a callback that will call
2534 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2536 Note that ERR_remove_state() is now deprecated, because it is tied
2537 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2538 to free the current thread's error state should be replaced by
2539 ERR_remove_thread_state(NULL).
2541 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2542 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2543 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2544 application was previously providing a numeric thread callback that
2545 was inappropriate for distinguishing threads, then uniqueness might
2546 have been obtained with &errno that happened immediately in the
2547 intermediate development versions of OpenSSL; this is no longer the
2548 case, the numeric thread callback will now override the automatic use
2550 [Geoff Thorpe, with help from Bodo Moeller]
2552 *) Initial support for different CRL issuing certificates. This covers a
2553 simple case where the self issued certificates in the chain exist and
2554 the real CRL issuer is higher in the existing chain.
2556 This work was sponsored by Google.
2559 *) Removed effectively defunct crypto/store from the build.
2562 *) Revamp of STACK to provide stronger type-checking. Still to come:
2563 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2564 ASN1_STRING, CONF_VALUE.
2567 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2568 RAM on SSL connections. This option can save about 34k per idle SSL.
2571 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2572 STACK, TXT_DB, bsearch, qsort.
2575 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2576 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2577 support for data, signedData, compressedData, digestedData and
2578 encryptedData, envelopedData types included. Scripts to check against
2579 RFC4134 examples draft and interop and consistency checks of many
2580 content types and variants.
2583 *) Add options to enc utility to support use of zlib compression BIO.
2586 *) Extend mk1mf to support importing of options and assembly language
2587 files from Configure script, currently only included in VC-WIN32.
2588 The assembly language rules can now optionally generate the source
2589 files from the associated perl scripts.
2592 *) Implement remaining functionality needed to support GOST ciphersuites.
2593 Interop testing has been performed using CryptoPro implementations.
2594 [Victor B. Wagner <vitus@cryptocom.ru>]
2596 *) s390x assembler pack.
2599 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2603 *) Implement Opaque PRF Input TLS extension as specified in
2604 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2605 official specification yet and no extension type assignment by
2606 IANA exists, this extension (for now) will have to be explicitly
2607 enabled when building OpenSSL by providing the extension number
2608 to use. For example, specify an option
2610 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2612 to the "config" or "Configure" script to enable the extension,
2613 assuming extension number 0x9527 (which is a completely arbitrary
2614 and unofficial assignment based on the MD5 hash of the Internet
2615 Draft). Note that by doing so, you potentially lose
2616 interoperability with other TLS implementations since these might
2617 be using the same extension number for other purposes.
2619 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2620 opaque PRF input value to use in the handshake. This will create
2621 an interal copy of the length-'len' string at 'src', and will
2622 return non-zero for success.
2624 To get more control and flexibility, provide a callback function
2627 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2628 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2632 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2635 Callback function 'cb' will be called in handshakes, and is
2636 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2637 Argument 'arg' is for application purposes (the value as given to
2638 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2639 be provided to the callback function). The callback function
2640 has to return non-zero to report success: usually 1 to use opaque
2641 PRF input just if possible, or 2 to enforce use of the opaque PRF
2642 input. In the latter case, the library will abort the handshake
2643 if opaque PRF input is not successfully negotiated.
2645 Arguments 'peerinput' and 'len' given to the callback function
2646 will always be NULL and 0 in the case of a client. A server will
2647 see the client's opaque PRF input through these variables if
2648 available (NULL and 0 otherwise). Note that if the server
2649 provides an opaque PRF input, the length must be the same as the
2650 length of the client's opaque PRF input.
2652 Note that the callback function will only be called when creating
2653 a new session (session resumption can resume whatever was
2654 previously negotiated), and will not be called in SSL 2.0
2655 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2656 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2657 for applications that need to enforce opaque PRF input.
2661 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2664 [Victor B. Wagner <vitus@cryptocom.ru>]
2666 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2667 RFC4507bis. The encrypted ticket format is an encrypted encoded
2668 SSL_SESSION structure, that way new session features are automatically
2671 If a client application caches session in an SSL_SESSION structure
2672 support is transparent because tickets are now stored in the encoded
2675 The SSL_CTX structure automatically generates keys for ticket
2676 protection in servers so again support should be possible
2677 with no application modification.
2679 If a client or server wishes to disable RFC4507 support then the option
2680 SSL_OP_NO_TICKET can be set.
2682 Add a TLS extension debugging callback to allow the contents of any client
2683 or server extensions to be examined.
2685 This work was sponsored by Google.
2688 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2689 OpenSSL should now compile cleanly on gcc 4.2
2690 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2692 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2693 support including streaming MAC support: this is required for GOST
2694 ciphersuite support.
2695 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2697 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2698 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2699 to output in BER and PEM format.
2702 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2703 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2704 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2705 ENGINE support for HMAC keys which are unextractable. New -mac and
2706 -macopt options to dgst utility.
2709 *) New option -sigopt to dgst utility. Update dgst to use
2710 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2711 alternative signing paramaters such as X9.31 or PSS in the dgst
2715 *) Change ssl_cipher_apply_rule(), the internal function that does
2716 the work each time a ciphersuite string requests enabling
2717 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2718 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2719 the order of disabled ciphersuites such that those ciphersuites
2720 that most recently went from enabled to disabled not only stay
2721 in order with respect to each other, but also have higher priority
2722 than other disabled ciphersuites the next time ciphersuites are
2725 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2726 the same ciphersuites as with "HIGH" alone, but in a specific
2727 order where the PSK ciphersuites come first (since they are the
2728 most recently disabled ciphersuites when "HIGH" is parsed).
2730 Also, change ssl_create_cipher_list() (using this new
2731 funcionality) such that between otherwise identical
2732 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2736 *) Change ssl_create_cipher_list() so that it automatically
2737 arranges the ciphersuites in reasonable order before starting
2738 to process the rule string. Thus, the definition for "DEFAULT"
2739 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2740 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2741 This makes it much easier to arrive at a reasonable default order
2742 in applications for which anonymous ciphers are OK (meaning
2743 that you can't actually use DEFAULT).
2744 [Bodo Moeller; suggested by Victor Duchovni]
2746 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2747 processing) into multiple integers instead of setting
2748 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2749 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2750 (These masks as well as the individual bit definitions are hidden
2751 away into the non-exported interface ssl/ssl_locl.h, so this
2752 change to the definition of the SSL_CIPHER structure shouldn't
2753 affect applications.) This give us more bits for each of these
2754 categories, so there is no longer a need to coagulate AES128 and
2755 AES256 into a single algorithm bit, and to coagulate Camellia128
2756 and Camellia256 into a single algorithm bit, which has led to all
2759 Thus, among other things, the kludge introduced in 0.9.7m and
2760 0.9.8e for masking out AES256 independently of AES128 or masking
2761 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2763 With the change, we also introduce new ciphersuite aliases that
2764 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2768 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2769 Use the leftmost N bytes of the signature input if the input is
2770 larger than the prime q (with N being the size in bytes of q).
2773 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2774 it yet and it is largely untested.
2777 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2780 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2781 some compilers (gcc 4.2 and later) reject their use. Safestack is
2782 reimplemented. Update ASN1 to avoid use of legacy functions.
2785 *) Win32/64 targets are linked with Winsock2.
2788 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2789 to external functions. This can be used to increase CRL handling
2790 efficiency especially when CRLs are very large by (for example) storing
2791 the CRL revoked certificates in a database.
2794 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2795 new CRLs added to a directory can be used. New command line option
2796 -verify_return_error to s_client and s_server. This causes real errors
2797 to be returned by the verify callback instead of carrying on no matter
2798 what. This reflects the way a "real world" verify callback would behave.
2801 *) GOST engine, supporting several GOST algorithms and public key formats.
2802 Kindly donated by Cryptocom.
2805 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2806 partitioned by DP are handled but no indirect CRL or reason partitioning
2807 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2808 selected via a scoring technique which handles IDP and AKID in CRLs.
2811 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2812 will ultimately be used for all verify operations: this will remove the
2813 X509_STORE dependency on certificate verification and allow alternative
2814 lookup methods. X509_STORE based implementations of these two callbacks.
2817 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2818 Modify get_crl() to find a valid (unexpired) CRL if possible.
2821 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2822 this would be called X509_CRL_cmp() but that name is already used by
2823 a function that just compares CRL issuer names. Cache several CRL
2824 extensions in X509_CRL structure and cache CRLDP in X509.
2827 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2828 this maps equivalent X509_NAME structures into a consistent structure.
2829 Name comparison can then be performed rapidly using memcmp().
2832 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2836 *) Allow digests to supply their own micalg string for S/MIME type using
2837 the ctrl EVP_MD_CTRL_MICALG.
2840 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2841 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2842 ctrl. It can then customise the structure before and/or after signing
2846 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2847 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2848 to free up any added signature OIDs.
2851 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2852 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2853 digest and cipher tables. New options added to openssl utility:
2854 list-message-digest-algorithms and list-cipher-algorithms.
2857 *) Change the array representation of binary polynomials: the list
2858 of degrees of non-zero coefficients is now terminated with -1.
2859 Previously it was terminated with 0, which was also part of the
2860 value; thus, the array representation was not applicable to
2861 polynomials where t^0 has coefficient zero. This change makes
2862 the array representation useful in a more general context.
2865 *) Various modifications and fixes to SSL/TLS cipher string
2866 handling. For ECC, the code now distinguishes between fixed ECDH
2867 with RSA certificates on the one hand and with ECDSA certificates
2868 on the other hand, since these are separate ciphersuites. The
2869 unused code for Fortezza ciphersuites has been removed.
2871 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2872 (not "ECDHE"). For consistency with the code for DH
2873 certificates, use of ECDH certificates is now considered ECDH
2874 authentication, not RSA or ECDSA authentication (the latter is
2875 merely the CA's signing algorithm and not actively used in the
2878 The temporary ciphersuite alias "ECCdraft" is no longer
2879 available, and ECC ciphersuites are no longer excluded from "ALL"
2880 and "DEFAULT". The following aliases now exist for RFC 4492
2881 ciphersuites, most of these by analogy with the DH case:
2883 kECDHr - ECDH cert, signed with RSA
2884 kECDHe - ECDH cert, signed with ECDSA
2885 kECDH - ECDH cert (signed with either RSA or ECDSA)
2886 kEECDH - ephemeral ECDH
2887 ECDH - ECDH cert or ephemeral ECDH
2893 AECDH - anonymous ECDH
2894 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2898 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2899 Use correct micalg parameters depending on digest(s) in signed message.
2902 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2903 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2906 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2907 an engine to register a method. Add ENGINE lookups for methods and
2908 functional reference processing.
2911 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2912 EVP_{Sign,Verify}* which allow an application to customise the signature
2916 *) New -resign option to smime utility. This adds one or more signers
2917 to an existing PKCS#7 signedData structure. Also -md option to use an
2918 alternative message digest algorithm for signing.
2921 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2922 create PKCS7 structures containing multiple signers. Update smime
2923 application to support multiple signers.
2926 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2930 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2931 Reorganize PBE internals to lookup from a static table using NIDs,
2932 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2933 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2934 PRF which will be automatically used with PBES2.
2937 *) Replace the algorithm specific calls to generate keys in "req" with the
2941 *) Update PKCS#7 enveloped data routines to use new API. This is now
2942 supported by any public key method supporting the encrypt operation. A
2943 ctrl is added to allow the public key algorithm to examine or modify
2944 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2948 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2949 a default digest type to use. In most cases this will be SHA1 but some
2950 algorithms (such as GOST) need to specify an alternative digest. The
2951 return value indicates how strong the prefernce is 1 means optional and
2952 2 is mandatory (that is it is the only supported type). Modify
2953 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2954 use the default md. Update openssl utilities to use the default digest
2955 type for signing if it is not explicitly indicated.
2958 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2959 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2960 signing method from the key type. This effectively removes the link
2961 between digests and public key types.
2964 *) Add an OID cross reference table and utility functions. Its purpose is to
2965 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2966 rsaEncryption. This will allow some of the algorithm specific hackery
2967 needed to use the correct OID to be removed.
2970 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2971 structures for PKCS7_sign(). They are now set up by the relevant public
2975 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2978 *) Add support for key derivation (agreement) in the API, DH method and
2982 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2983 public and private key formats. As a side effect these add additional
2984 command line functionality not previously available: DSA signatures can be
2985 generated and verified using pkeyutl and DH key support and generation in
2990 [Oliver Tappe <zooey@hirschkaefer.de>]
2992 *) New make target "install_html_docs" installs HTML renditions of the
2994 [Oliver Tappe <zooey@hirschkaefer.de>]
2996 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
2997 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2998 support key and parameter generation and add initial key generation
2999 functionality for RSA.
3002 *) Add functions for main EVP_PKEY_method operations. The undocumented
3003 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
3004 EVP_PKEY_{encrypt,decrypt}_old.
3007 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
3008 key API, doesn't do much yet.
3011 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
3012 public key algorithms. New option to openssl utility:
3013 "list-public-key-algorithms" to print out info.
3016 *) Implement the Supported Elliptic Curves Extension for
3017 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3020 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
3021 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
3024 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
3025 utilities such as rsa, dsa, dsaparam etc except they process any key
3029 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
3030 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
3031 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
3035 *) Initial support for pluggable public key ASN1.
3036 De-spaghettify the public key ASN1 handling. Move public and private
3037 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
3038 algorithm specific handling to a single module within the relevant
3039 algorithm directory. Add functions to allow (near) opaque processing
3040 of public and private key structures.
3043 *) Implement the Supported Point Formats Extension for
3044 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3047 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
3048 for the psk identity [hint] and the psk callback functions to the
3049 SSL_SESSION, SSL and SSL_CTX structure.
3052 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
3056 SSL_CTX_use_psk_identity_hint
3057 SSL_get_psk_identity_hint
3058 SSL_get_psk_identity
3059 SSL_use_psk_identity_hint
3061 [Mika Kousa and Pasi Eronen of Nokia Corporation]
3063 *) Add RFC 3161 compliant time stamp request creation, response generation
3064 and response verification functionality.
3065 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
3067 *) Add initial support for TLS extensions, specifically for the server_name
3068 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3069 have new members for a host name. The SSL data structure has an
3070 additional member SSL_CTX *initial_ctx so that new sessions can be
3071 stored in that context to allow for session resumption, even after the
3072 SSL has been switched to a new SSL_CTX in reaction to a client's
3073 server_name extension.
3075 New functions (subject to change):
3077 SSL_get_servername()
3078 SSL_get_servername_type()
3081 New CTRL codes and macros (subject to change):
3083 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3084 - SSL_CTX_set_tlsext_servername_callback()
3085 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3086 - SSL_CTX_set_tlsext_servername_arg()
3087 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3089 openssl s_client has a new '-servername ...' option.
3091 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3092 '-key2 ...', '-servername_fatal' (subject to change). This allows
3093 testing the HostName extension for a specific single host name ('-cert'
3094 and '-key' remain fallbacks for handshakes without HostName
3095 negotiation). If the unrecogninzed_name alert has to be sent, this by
3096 default is a warning; it becomes fatal with the '-servername_fatal'
3099 [Peter Sylvester, Remy Allais, Christophe Renou]
3101 *) Whirlpool hash implementation is added.
3104 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3105 bn(64,32). Because of instruction set limitations it doesn't have
3106 any negative impact on performance. This was done mostly in order
3107 to make it possible to share assembler modules, such as bn_mul_mont
3108 implementations, between 32- and 64-bit builds without hassle.
3111 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3112 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3116 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3117 dedicated Montgomery multiplication procedure, is introduced.
3118 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3119 "64-bit" performance on certain 32-bit targets.
3122 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3123 in SSL structures. New SSL ctrl to set maximum send fragment size.
3124 Save memory by seeting the I/O buffer sizes dynamically instead of
3125 using the maximum available value.
3128 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3129 in addition to the text details.
3132 *) Very, very preliminary EXPERIMENTAL support for printing of general
3133 ASN1 structures. This currently produces rather ugly output and doesn't
3134 handle several customised structures at all.
3137 *) Integrated support for PVK file format and some related formats such
3138 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3139 these in the 'rsa' and 'dsa' utilities.
3142 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3145 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3146 place for the (very old) "NETSCAPE" format certificates which are now
3147 handled using new ASN1 code equivalents.
3150 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3151 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3152 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3155 *) Modify CRL distribution points extension code to print out previously
3156 unsupported fields. Enhance extension setting code to allow setting of
3160 *) Add print and set support for Issuing Distribution Point CRL extension.
3163 *) Change 'Configure' script to enable Camellia by default.
3166 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3168 *) When rejecting SSL/TLS records due to an incorrect version number, never
3169 update s->server with a new major version number. As of
3170 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3171 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3172 the previous behavior could result in a read attempt at NULL when
3173 receiving specific incorrect SSL/TLS records once record payload
3174 protection is active. (CVE-2010-0740)
3175 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3177 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3178 could be crashed if the relevant tables were not present (e.g. chrooted).
3179 [Tomas Hoger <thoger@redhat.com>]
3181 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3183 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3184 [Martin Olsson, Neel Mehta]
3186 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3187 accommodate for stack sorting, always a write lock!).
3190 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3191 excessive delays in the RAND_poll(): over a minute. As a workaround
3192 include a time check in the inner Heap32Next loop too.
3195 *) The code that handled flushing of data in SSL/TLS originally used the
3196 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3197 the problem outlined in PR#1949. The fix suggested there however can
3198 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3199 of Apache). So instead simplify the code to flush unconditionally.
3200 This should be fine since flushing with no data to flush is a no op.
3203 *) Handle TLS versions 2.0 and later properly and correctly use the
3204 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
3205 off ancient servers have a habit of sticking around for a while...
3208 *) Modify compression code so it frees up structures without using the
3209 ex_data callbacks. This works around a problem where some applications
3210 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
3211 restarting) then use compression (e.g. SSL with compression) later.
3212 This results in significant per-connection memory leaks and
3213 has caused some security issues including CVE-2008-1678 and
3217 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3218 change when encrypting or decrypting.
3221 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3222 connect and renegotiate with servers which do not support RI.
3223 Until RI is more widely deployed this option is enabled by default.
3226 *) Add "missing" ssl ctrls to clear options and mode.
3229 *) If client attempts to renegotiate and doesn't support RI respond with
3230 a no_renegotiation alert as required by RFC5746. Some renegotiating
3231 TLS clients will continue a connection gracefully when they receive
3232 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3233 waiting for a server hello which it will never receive. Now we treat a
3234 received no_renegotiation alert as a fatal error. This is because
3235 applications requesting a renegotiation might well expect it to succeed
3236 and would have no code in place to handle the server denying it so the
3237 only safe thing to do is to terminate the connection.
3240 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3241 peer supports secure renegotiation and 0 otherwise. Print out peer
3242 renegotiation support in s_client/s_server.
3245 *) Replace the highly broken and deprecated SPKAC certification method with
3246 the updated NID creation version. This should correctly handle UTF8.
3249 *) Implement RFC5746. Re-enable renegotiation but require the extension
3250 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3251 turns out to be a bad idea. It has been replaced by
3252 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3253 SSL_CTX_set_options(). This is really not recommended unless you
3254 know what you are doing.
3255 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3257 *) Fixes to stateless session resumption handling. Use initial_ctx when
3258 issuing and attempting to decrypt tickets in case it has changed during
3259 servername handling. Use a non-zero length session ID when attempting
3260 stateless session resumption: this makes it possible to determine if
3261 a resumption has occurred immediately after receiving server hello
3262 (several places in OpenSSL subtly assume this) instead of later in
3266 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3267 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3268 fixes for a few places where the return code is not checked
3270 [Julia Lawall <julia@diku.dk>]
3272 *) Add --strict-warnings option to Configure script to include devteam
3273 warnings in other configurations.
3276 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3277 makes it possible to install openssl libraries in locations which
3278 have names other than "lib", for example "/usr/lib64" which some
3280 [Steve Henson, based on patch from Jeremy Utley]
3282 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3283 X690 8.9.12 and can produce some misleading textual output of OIDs.
3284 [Steve Henson, reported by Dan Kaminsky]
3286 *) Delete MD2 from algorithm tables. This follows the recommendation in
3287 several standards that it is not used in new applications due to
3288 several cryptographic weaknesses. For binary compatibility reasons
3289 the MD2 API is still compiled in by default.
3292 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3296 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3297 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3299 [Guenter <lists@gknw.net>]
3301 *) Fix the server certificate chain building code to use X509_verify_cert(),
3302 it used to have an ad-hoc builder which was unable to cope with anything
3303 other than a simple chain.
3304 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3306 *) Don't check self signed certificate signatures in X509_verify_cert()
3307 by default (a flag can override this): it just wastes time without
3308 adding any security. As a useful side effect self signed root CAs
3309 with non-FIPS digests are now usable in FIPS mode.
3312 *) In dtls1_process_out_of_seq_message() the check if the current message
3313 is already buffered was missing. For every new message was memory
3314 allocated, allowing an attacker to perform an denial of service attack
3315 with sending out of seq handshake messages until there is no memory
3316 left. Additionally every future messege was buffered, even if the
3317 sequence number made no sense and would be part of another handshake.
3318 So only messages with sequence numbers less than 10 in advance will be
3319 buffered. (CVE-2009-1378)
3320 [Robin Seggelmann, discovered by Daniel Mentz]
3322 *) Records are buffered if they arrive with a future epoch to be
3323 processed after finishing the corresponding handshake. There is
3324 currently no limitation to this buffer allowing an attacker to perform
3325 a DOS attack with sending records with future epochs until there is no
3326 memory left. This patch adds the pqueue_size() function to detemine
3327 the size of a buffer and limits the record buffer to 100 entries.
3329 [Robin Seggelmann, discovered by Daniel Mentz]
3331 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3332 parent structure is freed. (CVE-2009-1379)
3335 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3336 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3339 [Ilya O. <vrghost@gmail.com>]
3341 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3343 *) Disable renegotiation completely - this fixes a severe security
3344 problem (CVE-2009-3555) at the cost of breaking all
3345 renegotiation. Renegotiation can be re-enabled by setting
3346 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3347 run-time. This is really not recommended unless you know what
3351 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3353 *) Don't set val to NULL when freeing up structures, it is freed up by
3354 underlying code. If sizeof(void *) > sizeof(long) this can result in
3355 zeroing past the valid field. (CVE-2009-0789)
3356 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3358 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3359 checked correctly. This would allow some invalid signed attributes to
3360 appear to verify correctly. (CVE-2009-0591)
3361 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3363 *) Reject UniversalString and BMPString types with invalid lengths. This
3364 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3365 a legal length. (CVE-2009-0590)
3368 *) Set S/MIME signing as the default purpose rather than setting it
3369 unconditionally. This allows applications to override it at the store
3373 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3374 to handle some structures.
3377 *) Improve efficiency of mem_gets: don't search whole buffer each time
3379 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3381 *) New -hex option for openssl rand.
3384 *) Print out UTF8String and NumericString when parsing ASN1.
3387 *) Support NumericString type for name components.
3390 *) Allow CC in the environment to override the automatically chosen
3391 compiler. Note that nothing is done to ensure flags work with the
3395 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3397 *) Properly check EVP_VerifyFinal() and similar return values
3399 [Ben Laurie, Bodo Moeller, Google Security Team]
3401 *) Enable TLS extensions by default.
3404 *) Allow the CHIL engine to be loaded, whether the application is
3405 multithreaded or not. (This does not release the developer from the
3406 obligation to set up the dynamic locking callbacks.)
3407 [Sander Temme <sander@temme.net>]
3409 *) Use correct exit code if there is an error in dgst command.
3410 [Steve Henson; problem pointed out by Roland Dirlewanger]
3412 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3413 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3416 *) Add experimental JPAKE support, including demo authentication in
3417 s_client and s_server.
3420 *) Set the comparison function in v3_addr_canonize().
3421 [Rob Austein <sra@hactrn.net>]
3423 *) Add support for XMPP STARTTLS in s_client.
3424 [Philip Paeps <philip@freebsd.org>]
3426 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3427 to ensure that even with this option, only ciphersuites in the
3428 server's preference list will be accepted. (Note that the option
3429 applies only when resuming a session, so the earlier behavior was
3430 just about the algorithm choice for symmetric cryptography.)
3433 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3435 *) Fix NULL pointer dereference if a DTLS server received
3436 ChangeCipherSpec as first record (CVE-2009-1386).
3439 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
3440 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3443 *) The fix in 0.9.8c that supposedly got rid of unsafe
3444 double-checked locking was incomplete for RSA blinding,
3445 addressing just one layer of what turns out to have been
3446 doubly unsafe triple-checked locking.
3448 So now fix this for real by retiring the MONT_HELPER macro
3449 in crypto/rsa/rsa_eay.c.
3451 [Bodo Moeller; problem pointed out by Marius Schilder]
3453 *) Various precautionary measures:
3455 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3457 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3458 (NB: This would require knowledge of the secret session ticket key
3459 to exploit, in which case you'd be SOL either way.)
3461 - Change bn_nist.c so that it will properly handle input BIGNUMs
3462 outside the expected range.
3464 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3467 [Neel Mehta, Bodo Moeller]
3469 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3470 the load fails. Useful for distros.
3471 [Ben Laurie and the FreeBSD team]
3473 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3476 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3479 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3481 This work was sponsored by Logica.
3484 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3485 keystores. Support for SSL/TLS client authentication too.
3486 Not compiled unless enable-capieng specified to Configure.
3488 This work was sponsored by Logica.
3491 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3492 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3493 attribute creation routines such as certifcate requests and PKCS#12
3497 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3499 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3500 handshake which could lead to a cilent crash as found using the
3501 Codenomicon TLS test suite (CVE-2008-1672)
3502 [Steve Henson, Mark Cox]
3504 *) Fix double free in TLS server name extensions which could lead to
3505 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3508 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3510 Clear the error queue to ensure that error entries left from
3511 older function calls do not interfere with the correct operation.
3512 [Lutz Jaenicke, Erik de Castro Lopo]
3514 *) Remove root CA certificates of commercial CAs:
3516 The OpenSSL project does not recommend any specific CA and does not
3517 have any policy with respect to including or excluding any CA.
3518 Therefore it does not make any sense to ship an arbitrary selection
3519 of root CA certificates with the OpenSSL software.
3522 *) RSA OAEP patches to fix two separate invalid memory reads.
3523 The first one involves inputs when 'lzero' is greater than
3524 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3525 before the beginning of from). The second one involves inputs where
3526 the 'db' section contains nothing but zeroes (there is a one-byte
3527 invalid read after the end of 'db').
3528 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3530 *) Partial backport from 0.9.9-dev:
3532 Introduce bn_mul_mont (dedicated Montgomery multiplication
3533 procedure) as a candidate for BIGNUM assembler implementation.
3534 While 0.9.9-dev uses assembler for various architectures, only
3535 x86_64 is available by default here in the 0.9.8 branch, and
3536 32-bit x86 is available through a compile-time setting.
3538 To try the 32-bit x86 assembler implementation, use Configure
3539 option "enable-montasm" (which exists only for this backport).
3541 As "enable-montasm" for 32-bit x86 disclaims code stability
3542 anyway, in this constellation we activate additional code
3543 backported from 0.9.9-dev for further performance improvements,
3544 namely BN_from_montgomery_word. (To enable this otherwise,
3545 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3547 [Andy Polyakov (backport partially by Bodo Moeller)]
3549 *) Add TLS session ticket callback. This allows an application to set
3550 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3551 values. This is useful for key rollover for example where several key
3552 sets may exist with different names.
3555 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3556 This was broken until now in 0.9.8 releases, such that the only way
3557 a registered ENGINE could be used (assuming it initialises
3558 successfully on the host) was to explicitly set it as the default
3559 for the relevant algorithms. This is in contradiction with 0.9.7
3560 behaviour and the documentation. With this fix, when an ENGINE is
3561 registered into a given algorithm's table of implementations, the
3562 'uptodate' flag is reset so that auto-discovery will be used next
3563 time a new context for that algorithm attempts to select an
3565 [Ian Lister (tweaked by Geoff Thorpe)]
3567 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3568 implemention in the following ways:
3570 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3573 Lack of BER streaming support means one pass streaming processing is
3574 only supported if data is detached: setting the streaming flag is
3575 ignored for embedded content.
3577 CMS support is disabled by default and must be explicitly enabled
3578 with the enable-cms configuration option.
3581 *) Update the GMP engine glue to do direct copies between BIGNUM and
3582 mpz_t when openssl and GMP use the same limb size. Otherwise the
3583 existing "conversion via a text string export" trick is still used.
3584 [Paul Sheer <paulsheer@gmail.com>]
3586 *) Zlib compression BIO. This is a filter BIO which compressed and
3587 uncompresses any data passed through it.
3590 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3591 RFC3394 compatible AES key wrapping.
3594 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3595 sets string data without copying. X509_ALGOR_set0() and
3596 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3597 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3598 from an X509_ATTRIBUTE structure optionally checking it occurs only
3599 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3603 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3604 to get the expected BN_FLG_CONSTTIME behavior.
3605 [Bodo Moeller (Google)]
3609 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3610 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3611 - added some more tests to do_tests.pl
3612 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3613 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3614 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3615 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3616 - various changes to netware.pl to enable gcc-cross builds on Win32
3618 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3619 - various changes to fix missing prototype warnings
3620 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3621 - added AES, WHIRLPOOL and CPUID assembler code to build files
3622 - added missing AES assembler make rules to mk1mf.pl
3623 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3624 [Guenter Knauf <eflash@gmx.net>]
3626 *) Implement certificate status request TLS extension defined in RFC3546.
3627 A client can set the appropriate parameters and receive the encoded
3628 OCSP response via a callback. A server can query the supplied parameters
3629 and set the encoded OCSP response in the callback. Add simplified examples
3630 to s_client and s_server.
3633 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3635 *) Fix various bugs:
3636 + Binary incompatibility of ssl_ctx_st structure
3637 + DTLS interoperation with non-compliant servers
3638 + Don't call get_session_cb() without proposed session
3639 + Fix ia64 assembler code
3640 [Andy Polyakov, Steve Henson]
3642 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3644 *) DTLS Handshake overhaul. There were longstanding issues with
3645 OpenSSL DTLS implementation, which were making it impossible for
3646 RFC 4347 compliant client to communicate with OpenSSL server.
3647 Unfortunately just fixing these incompatibilities would "cut off"
3648 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3649 server keeps tolerating non RFC compliant syntax. The opposite is
3650 not true, 0.9.8f client can not communicate with earlier server.
3651 This update even addresses CVE-2007-4995.
3654 *) Changes to avoid need for function casts in OpenSSL: some compilers
3655 (gcc 4.2 and later) reject their use.
3656 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3659 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3660 RFC4507bis. The encrypted ticket format is an encrypted encoded
3661 SSL_SESSION structure, that way new session features are automatically
3664 If a client application caches session in an SSL_SESSION structure
3665 support is transparent because tickets are now stored in the encoded
3668 The SSL_CTX structure automatically generates keys for ticket
3669 protection in servers so again support should be possible
3670 with no application modification.
3672 If a client or server wishes to disable RFC4507 support then the option
3673 SSL_OP_NO_TICKET can be set.
3675 Add a TLS extension debugging callback to allow the contents of any client
3676 or server extensions to be examined.
3678 This work was sponsored by Google.
3681 *) Add initial support for TLS extensions, specifically for the server_name
3682 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3683 have new members for a host name. The SSL data structure has an
3684 additional member SSL_CTX *initial_ctx so that new sessions can be
3685 stored in that context to allow for session resumption, even after the
3686 SSL has been switched to a new SSL_CTX in reaction to a client's
3687 server_name extension.
3689 New functions (subject to change):
3691 SSL_get_servername()
3692 SSL_get_servername_type()
3695 New CTRL codes and macros (subject to change):
3697 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3698 - SSL_CTX_set_tlsext_servername_callback()
3699 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3700 - SSL_CTX_set_tlsext_servername_arg()
3701 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3703 openssl s_client has a new '-servername ...' option.
3705 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3706 '-key2 ...', '-servername_fatal' (subject to change). This allows
3707 testing the HostName extension for a specific single host name ('-cert'
3708 and '-key' remain fallbacks for handshakes without HostName
3709 negotiation). If the unrecogninzed_name alert has to be sent, this by
3710 default is a warning; it becomes fatal with the '-servername_fatal'
3713 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3715 *) Add AES and SSE2 assembly language support to VC++ build.
3718 *) Mitigate attack on final subtraction in Montgomery reduction.
3721 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3722 (which previously caused an internal error).
3725 *) Squeeze another 10% out of IGE mode when in != out.
3728 *) AES IGE mode speedup.
3729 [Dean Gaudet (Google)]
3731 *) Add the Korean symmetric 128-bit cipher SEED (see
3732 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3733 add SEED ciphersuites from RFC 4162:
3735 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3736 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3737 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3738 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3740 To minimize changes between patchlevels in the OpenSSL 0.9.8
3741 series, SEED remains excluded from compilation unless OpenSSL
3742 is configured with 'enable-seed'.
3743 [KISA, Bodo Moeller]
3745 *) Mitigate branch prediction attacks, which can be practical if a
3746 single processor is shared, allowing a spy process to extract
3747 information. For detailed background information, see
3748 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3749 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3750 and Necessary Software Countermeasures"). The core of the change
3751 are new versions BN_div_no_branch() and
3752 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3753 respectively, which are slower, but avoid the security-relevant
3754 conditional branches. These are automatically called by BN_div()
3755 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3756 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3757 remove a conditional branch.
3759 BN_FLG_CONSTTIME is the new name for the previous
3760 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3761 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3762 in the exponent causes BN_mod_exp_mont() to use the alternative
3763 implementation in BN_mod_exp_mont_consttime().) The old name
3764 remains as a deprecated alias.
3766 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3767 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3768 constant-time implementations for more than just exponentiation.
3769 Here too the old name is kept as a deprecated alias.
3771 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3772 the BN_BLINDING structure gets an independent copy of the
3773 modulus. This means that the previous "BIGNUM *m" argument to
3774 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3775 essentially becomes "const BIGNUM *m", although we can't actually
3776 change this in the header file before 0.9.9. It allows
3777 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3778 enable BN_FLG_CONSTTIME.
3780 [Matthew D Wood (Intel Corp)]
3782 *) In the SSL/TLS server implementation, be strict about session ID
3783 context matching (which matters if an application uses a single
3784 external cache for different purposes). Previously,
3785 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3786 set. This did ensure strict client verification, but meant that,
3787 with applications using a single external cache for quite
3788 different requirements, clients could circumvent ciphersuite
3789 restrictions for a given session ID context by starting a session
3790 in a different context.
3793 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3794 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3795 authentication-only ciphersuites.
3798 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3799 not complete and could lead to a possible single byte overflow
3800 (CVE-2007-5135) [Ben Laurie]
3802 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3804 *) Since AES128 and AES256 (and similarly Camellia128 and
3805 Camellia256) share a single mask bit in the logic of
3806 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3807 kludge to work properly if AES128 is available and AES256 isn't
3808 (or if Camellia128 is available and Camellia256 isn't).
3811 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3812 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3813 When a point or a seed is encoded in a BIT STRING, we need to
3814 prevent the removal of trailing zero bits to get the proper DER
3815 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3816 of a NamedBitList, for which trailing 0 bits need to be removed.)
3819 *) Have SSL/TLS server implementation tolerate "mismatched" record
3820 protocol version while receiving ClientHello even if the
3821 ClientHello is fragmented. (The server can't insist on the
3822 particular protocol version it has chosen before the ServerHello
3823 message has informed the client about his choice.)
3826 *) Add RFC 3779 support.
3827 [Rob Austein for ARIN, Ben Laurie]
3829 *) Load error codes if they are not already present instead of using a
3830 static variable. This allows them to be cleanly unloaded and reloaded.
3831 Improve header file function name parsing.
3834 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3835 or CAPABILITY handshake as required by RFCs.
3838 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3840 *) Introduce limits to prevent malicious keys being able to
3841 cause a denial of service. (CVE-2006-2940)
3842 [Steve Henson, Bodo Moeller]
3844 *) Fix ASN.1 parsing of certain invalid structures that can result
3845 in a denial of service. (CVE-2006-2937) [Steve Henson]
3847 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3848 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3850 *) Fix SSL client code which could crash if connecting to a
3851 malicious SSLv2 server. (CVE-2006-4343)
3852 [Tavis Ormandy and Will Drewry, Google Security Team]
3854 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3855 match only those. Before that, "AES256-SHA" would be interpreted
3856 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3857 the same strength classification in 0.9.7h) as we currently only
3858 have a single AES bit in the ciphersuite description bitmap.
3859 That change, however, also applied to ciphersuite strings such as
3860 "RC4-MD5" that intentionally matched multiple ciphersuites --
3861 namely, SSL 2.0 ciphersuites in addition to the more common ones
3862 from SSL 3.0/TLS 1.0.
3864 So we change the selection algorithm again: Naming an explicit
3865 ciphersuite selects this one ciphersuite, and any other similar
3866 ciphersuite (same bitmap) from *other* protocol versions.
3867 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3868 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3870 Since SSL 2.0 does not have any ciphersuites for which the
3871 128/256 bit distinction would be relevant, this works for now.
3872 The proper fix will be to use different bits for AES128 and
3873 AES256, which would have avoided the problems from the beginning;
3874 however, bits are scarce, so we can only do this in a new release
3875 (not just a patchlevel) when we can change the SSL_CIPHER
3876 definition to split the single 'unsigned long mask' bitmap into
3877 multiple values to extend the available space.
3881 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3883 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3884 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3886 *) Add AES IGE and biIGE modes.
3889 *) Change the Unix randomness entropy gathering to use poll() when
3890 possible instead of select(), since the latter has some
3891 undesirable limitations.
3892 [Darryl Miles via Richard Levitte and Bodo Moeller]
3894 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3895 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3896 cannot be implicitly activated as part of, e.g., the "AES" alias.
3897 However, please upgrade to OpenSSL 0.9.9[-dev] for
3898 non-experimental use of the ECC ciphersuites to get TLS extension
3899 support, which is required for curve and point format negotiation
3900 to avoid potential handshake problems.
3903 *) Disable rogue ciphersuites:
3905 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3906 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3907 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3909 The latter two were purportedly from
3910 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3913 Also deactivate the remaining ciphersuites from
3914 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3915 unofficial, and the ID has long expired.
3918 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3919 dual-core machines) and other potential thread-safety issues.
3922 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3923 versions), which is now available for royalty-free use
3924 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3925 Also, add Camellia TLS ciphersuites from RFC 4132.
3927 To minimize changes between patchlevels in the OpenSSL 0.9.8
3928 series, Camellia remains excluded from compilation unless OpenSSL
3929 is configured with 'enable-camellia'.
3932 *) Disable the padding bug check when compression is in use. The padding
3933 bug check assumes the first packet is of even length, this is not
3934 necessarily true if compresssion is enabled and can result in false
3935 positives causing handshake failure. The actual bug test is ancient
3936 code so it is hoped that implementations will either have fixed it by
3937 now or any which still have the bug do not support compression.
3940 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3942 *) When applying a cipher rule check to see if string match is an explicit
3943 cipher suite and only match that one cipher suite if it is.
3946 *) Link in manifests for VC++ if needed.
3947 [Austin Ziegler <halostatue@gmail.com>]
3949 *) Update support for ECC-based TLS ciphersuites according to
3950 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3951 TLS extensions, which are supported starting with the 0.9.9
3952 branch, not in the OpenSSL 0.9.8 branch).
3955 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3956 opaque EVP_CIPHER_CTX handling.
3959 *) Fixes and enhancements to zlib compression code. We now only use
3960 "zlib1.dll" and use the default __cdecl calling convention on Win32
3961 to conform with the standards mentioned here:
3962 http://www.zlib.net/DLL_FAQ.txt
3963 Static zlib linking now works on Windows and the new --with-zlib-include
3964 --with-zlib-lib options to Configure can be used to supply the location
3965 of the headers and library. Gracefully handle case where zlib library
3969 *) Several fixes and enhancements to the OID generation code. The old code
3970 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3971 handle numbers larger than ULONG_MAX, truncated printing and had a
3972 non standard OBJ_obj2txt() behaviour.
3975 *) Add support for building of engines under engine/ as shared libraries
3976 under VC++ build system.
3979 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3980 Hopefully, we will not see any false combination of paths any more.
3983 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3985 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3986 (part of SSL_OP_ALL). This option used to disable the
3987 countermeasure against man-in-the-middle protocol-version
3988 rollback in the SSL 2.0 server implementation, which is a bad
3989 idea. (CVE-2005-2969)
3991 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3992 for Information Security, National Institute of Advanced Industrial
3993 Science and Technology [AIST], Japan)]
3995 *) Add two function to clear and return the verify parameter flags.
3998 *) Keep cipherlists sorted in the source instead of sorting them at
3999 runtime, thus removing the need for a lock.
4002 *) Avoid some small subgroup attacks in Diffie-Hellman.
4003 [Nick Mathewson and Ben Laurie]
4005 *) Add functions for well-known primes.
4008 *) Extended Windows CE support.
4009 [Satoshi Nakamura and Andy Polyakov]
4011 *) Initialize SSL_METHOD structures at compile time instead of during
4012 runtime, thus removing the need for a lock.
4015 *) Make PKCS7_decrypt() work even if no certificate is supplied by
4016 attempting to decrypt each encrypted key in turn. Add support to
4020 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
4022 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
4025 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
4028 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
4029 key into the same file any more.
4032 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
4035 *) Add -utf8 command line and config file option to 'ca'.
4036 [Stefan <stf@udoma.org]
4038 *) Removed the macro des_crypt(), as it seems to conflict with some
4039 libraries. Use DES_crypt().
4042 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
4043 involves renaming the source and generated shared-libs for
4044 both. The engines will accept the corrected or legacy ids
4045 ('ncipher' and '4758_cca' respectively) when binding. NB,
4046 this only applies when building 'shared'.
4047 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
4049 *) Add attribute functions to EVP_PKEY structure. Modify
4050 PKCS12_create() to recognize a CSP name attribute and
4051 use it. Make -CSP option work again in pkcs12 utility.
4054 *) Add new functionality to the bn blinding code:
4055 - automatic re-creation of the BN_BLINDING parameters after
4056 a fixed number of uses (currently 32)
4057 - add new function for parameter creation
4058 - introduce flags to control the update behaviour of the
4059 BN_BLINDING parameters
4060 - hide BN_BLINDING structure
4061 Add a second BN_BLINDING slot to the RSA structure to improve
4062 performance when a single RSA object is shared among several
4066 *) Add support for DTLS.
4067 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
4069 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
4070 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
4073 *) Remove buggy and incompletet DH cert support from
4074 ssl/ssl_rsa.c and ssl/s3_both.c
4077 *) Use SHA-1 instead of MD5 as the default digest algorithm for
4078 the apps/openssl applications.
4081 *) Compile clean with "-Wall -Wmissing-prototypes
4082 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
4083 DEBUG_SAFESTACK must also be set.
4086 *) Change ./Configure so that certain algorithms can be disabled by default.
4087 The new counterpiece to "no-xxx" is "enable-xxx".
4089 The patented RC5 and MDC2 algorithms will now be disabled unless
4090 "enable-rc5" and "enable-mdc2", respectively, are specified.
4092 (IDEA remains enabled despite being patented. This is because IDEA
4093 is frequently required for interoperability, and there is no license
4094 fee for non-commercial use. As before, "no-idea" can be used to
4095 avoid this algorithm.)
4099 *) Add processing of proxy certificates (see RFC 3820). This work was
4100 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
4101 EGEE (Enabling Grids for E-science in Europe).
4104 *) RC4 performance overhaul on modern architectures/implementations, such
4105 as Intel P4, IA-64 and AMD64.
4108 *) New utility extract-section.pl. This can be used specify an alternative
4109 section number in a pod file instead of having to treat each file as
4110 a separate case in Makefile. This can be done by adding two lines to the
4113 =for comment openssl_section:XXX
4115 The blank line is mandatory.
4119 *) New arguments -certform, -keyform and -pass for s_client and s_server
4120 to allow alternative format key and certificate files and passphrase
4124 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
4125 update associated structures and add various utility functions.
4127 Add new policy related verify parameters, include policy checking in
4128 standard verify code. Enhance 'smime' application with extra parameters
4129 to support policy checking and print out.
4132 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
4133 Nehemiah processors. These extensions support AES encryption in hardware
4134 as well as RNG (though RNG support is currently disabled).
4135 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
4137 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
4140 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
4141 [Andy Polyakov and a number of other people]
4143 *) Improved PowerPC platform support. Most notably BIGNUM assembler
4144 implementation contributed by IBM.
4145 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
4147 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
4148 exponent rather than 'unsigned long'. There is a corresponding change to
4149 the new 'rsa_keygen' element of the RSA_METHOD structure.
4150 [Jelte Jansen, Geoff Thorpe]
4152 *) Functionality for creating the initial serial number file is now
4153 moved from CA.pl to the 'ca' utility with a new option -create_serial.
4155 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
4156 number file to 1, which is bound to cause problems. To avoid
4157 the problems while respecting compatibility between different 0.9.7
4158 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
4159 CA.pl for serial number initialization. With the new release 0.9.8,
4160 we can fix the problem directly in the 'ca' utility.)
4163 *) Reduced header interdepencies by declaring more opaque objects in
4164 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
4165 give fewer recursive includes, which could break lazy source code - so
4166 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
4167 developers should define this symbol when building and using openssl to
4168 ensure they track the recommended behaviour, interfaces, [etc], but
4169 backwards-compatible behaviour prevails when this isn't defined.
4172 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
4175 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
4176 This will generate a random key of the appropriate length based on the
4177 cipher context. The EVP_CIPHER can provide its own random key generation
4178 routine to support keys of a specific form. This is used in the des and
4179 3des routines to generate a key of the correct parity. Update S/MIME
4180 code to use new functions and hence generate correct parity DES keys.
4181 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
4182 valid (weak or incorrect parity).
4185 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
4186 as looking them up. This is useful when the verified structure may contain
4187 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
4188 present unless the new PKCS7_NO_CRL flag is asserted.
4191 *) Extend ASN1 oid configuration module. It now additionally accepts the
4194 shortName = some long name, 1.2.3.4
4197 *) Reimplemented the BN_CTX implementation. There is now no more static
4198 limitation on the number of variables it can handle nor the depth of the
4199 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
4200 information can now expand as required, and rather than having a single
4201 static array of bignums, BN_CTX now uses a linked-list of such arrays
4202 allowing it to expand on demand whilst maintaining the usefulness of
4203 BN_CTX's "bundling".
4206 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
4207 to allow all RSA operations to function using a single BN_CTX.
4210 *) Preliminary support for certificate policy evaluation and checking. This
4211 is initially intended to pass the tests outlined in "Conformance Testing
4212 of Relying Party Client Certificate Path Processing Logic" v1.07.
4215 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
4216 remained unused and not that useful. A variety of other little bignum
4217 tweaks and fixes have also been made continuing on from the audit (see
4221 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4222 associated ASN1, EVP and SSL functions and old ASN1 macros.
4225 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4226 and this should never fail. So the return value from the use of
4227 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4228 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4231 *) BN_CTX_get() should return zero-valued bignums, providing the same
4232 initialised value as BN_new().
4233 [Geoff Thorpe, suggested by Ulf Möller]
4235 *) Support for inhibitAnyPolicy certificate extension.
4238 *) An audit of the BIGNUM code is underway, for which debugging code is
4239 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4240 is considered valid when processing BIGNUMs, and causes execution to
4241 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4242 further steps are taken to deliberately pollute unused data in BIGNUM
4243 structures to try and expose faulty code further on. For now, openssl will
4244 (in its default mode of operation) continue to tolerate the inconsistent
4245 forms that it has tolerated in the past, but authors and packagers should
4246 consider trying openssl and their own applications when compiled with
4247 these debugging symbols defined. It will help highlight potential bugs in
4248 their own code, and will improve the test coverage for OpenSSL itself. At
4249 some point, these tighter rules will become openssl's default to improve
4250 maintainability, though the assert()s and other overheads will remain only
4251 in debugging configurations. See bn.h for more details.
4252 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4254 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4255 that can only be obtained through BN_CTX_new() (which implicitly
4256 initialises it). The presence of this function only made it possible
4257 to overwrite an existing structure (and cause memory leaks).
4260 *) Because of the callback-based approach for implementing LHASH as a
4261 template type, lh_insert() adds opaque objects to hash-tables and
4262 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4263 to clean up those corresponding objects before destroying the hash table
4264 (and losing the object pointers). So some over-zealous constifications in
4265 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4266 objects as "const" and the lh_doall[_arg] callback wrappers are not
4267 prototyped to have "const" restrictions on the object pointers they are
4268 given (and so aren't required to cast them away any more).
4271 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4272 (speed) prefers to use its own implementation. The two implementations
4273 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4274 its object type properly exposed (MS_TM) instead of casting to/from "char
4275 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4276 aren't necessarily the greatest nomenclatures - but this is what was used
4277 internally to the implementation so I've used that for now.
4280 *) Ensure that deprecated functions do not get compiled when
4281 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4282 the self-tests were still using deprecated key-generation functions so
4283 these have been updated also.
4286 *) Reorganise PKCS#7 code to separate the digest location functionality
4287 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
4288 New function PKCS7_set_digest() to set the digest type for PKCS#7
4289 digestedData type. Add additional code to correctly generate the
4290 digestedData type and add support for this type in PKCS7 initialization
4294 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4295 structure of type "other".
4298 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4299 sure the loop does correctly stop and breaking ("division by zero")
4300 modulus operations are not performed. The (pre-generated) prime
4301 table crypto/bn/bn_prime.h was already correct, but it could not be
4302 re-generated on some platforms because of the "division by zero"
4303 situation in the script.
4304 [Ralf S. Engelschall]
4306 *) Update support for ECC-based TLS ciphersuites according to
4307 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4308 SHA-1 now is only used for "small" curves (where the
4309 representation of a field element takes up to 24 bytes); for
4310 larger curves, the field element resulting from ECDH is directly
4311 used as premaster secret.
4312 [Douglas Stebila (Sun Microsystems Laboratories)]
4314 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4315 curve secp160r1 to the tests.
4316 [Douglas Stebila (Sun Microsystems Laboratories)]
4318 *) Add the possibility to load symbols globally with DSO.
4319 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4321 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4322 control of the error stack.
4325 *) Add support for STORE in ENGINE.
4328 *) Add the STORE type. The intention is to provide a common interface
4329 to certificate and key stores, be they simple file-based stores, or
4330 HSM-type store, or LDAP stores, or...
4331 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4334 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4335 pass a list of arguments to any function as well as provide a way
4336 for a function to pass data back to the caller.
4339 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4340 works like BUF_strdup() but can be used to duplicate a portion of
4341 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4345 *) Add the function sk_find_ex() which works like sk_find(), but will
4346 return an index to an element even if an exact match couldn't be
4347 found. The index is guaranteed to point at the element where the
4348 searched-for key would be inserted to preserve sorting order.
4351 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4352 takes an extra flags argument for optional functionality. Currently,
4353 the following flags are defined:
4355 OBJ_BSEARCH_VALUE_ON_NOMATCH
4356 This one gets OBJ_bsearch_ex() to return a pointer to the first
4357 element where the comparing function returns a negative or zero
4360 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4361 This one gets OBJ_bsearch_ex() to return a pointer to the first
4362 element where the comparing function returns zero. This is useful
4363 if there are more than one element where the comparing function
4367 *) Make it possible to create self-signed certificates with 'openssl ca'
4368 in such a way that the self-signed certificate becomes part of the
4369 CA database and uses the same mechanisms for serial number generation
4370 as all other certificate signing. The new flag '-selfsign' enables
4371 this functionality. Adapt CA.sh and CA.pl.in.
4374 *) Add functionality to check the public key of a certificate request
4375 against a given private. This is useful to check that a certificate
4376 request can be signed by that key (self-signing).
4379 *) Make it possible to have multiple active certificates with the same
4380 subject in the CA index file. This is done only if the keyword
4381 'unique_subject' is set to 'no' in the main CA section (default
4382 if 'CA_default') of the configuration file. The value is saved
4383 with the database itself in a separate index attribute file,
4384 named like the index file with '.attr' appended to the name.
4387 *) Generate muti valued AVAs using '+' notation in config files for
4391 *) Support for nameConstraints certificate extension.
4394 *) Support for policyConstraints certificate extension.
4397 *) Support for policyMappings certificate extension.
4400 *) Make sure the default DSA_METHOD implementation only uses its
4401 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4402 and change its own handlers to be NULL so as to remove unnecessary
4403 indirection. This lets alternative implementations fallback to the
4404 default implementation more easily.
4407 *) Support for directoryName in GeneralName related extensions
4411 *) Make it possible to link applications using Makefile.shared.
4412 Make that possible even when linking against static libraries!
4415 *) Support for single pass processing for S/MIME signing. This now
4416 means that S/MIME signing can be done from a pipe, in addition
4417 cleartext signing (multipart/signed type) is effectively streaming
4418 and the signed data does not need to be all held in memory.
4420 This is done with a new flag PKCS7_STREAM. When this flag is set
4421 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4422 is done after the data is output (and digests calculated) in
4423 SMIME_write_PKCS7().
4426 *) Add full support for -rpath/-R, both in shared libraries and
4427 applications, at least on the platforms where it's known how
4431 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4432 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4433 will now compute a table of multiples of the generator that
4434 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4435 faster (notably in the case of a single point multiplication,
4436 scalar * generator).
4437 [Nils Larsch, Bodo Moeller]
4439 *) IPv6 support for certificate extensions. The various extensions
4440 which use the IP:a.b.c.d can now take IPv6 addresses using the
4441 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4445 *) Added an ENGINE that implements RSA by performing private key
4446 exponentiations with the GMP library. The conversions to and from
4447 GMP's mpz_t format aren't optimised nor are any montgomery forms
4448 cached, and on x86 it appears OpenSSL's own performance has caught up.
4449 However there are likely to be other architectures where GMP could
4450 provide a boost. This ENGINE is not built in by default, but it can be
4451 specified at Configure time and should be accompanied by the necessary
4452 linker additions, eg;
4453 ./config -DOPENSSL_USE_GMP -lgmp
4456 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4457 testing availability of engines with "-t" - the old behaviour is
4458 produced by increasing the feature's verbosity with "-tt".
4461 *) ECDSA routines: under certain error conditions uninitialized BN objects
4462 could be freed. Solution: make sure initialization is performed early
4463 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4467 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4468 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4469 software implementations. For DSA and DH, parameter generation can
4470 also be overriden by providing the appropriate method callbacks.
4473 *) Change the "progress" mechanism used in key-generation and
4474 primality testing to functions that take a new BN_GENCB pointer in
4475 place of callback/argument pairs. The new API functions have "_ex"
4476 postfixes and the older functions are reimplemented as wrappers for
4477 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4478 declarations of the old functions to help (graceful) attempts to
4479 migrate to the new functions. Also, the new key-generation API
4480 functions operate on a caller-supplied key-structure and return
4481 success/failure rather than returning a key or NULL - this is to
4482 help make "keygen" another member function of RSA_METHOD etc.
4484 Example for using the new callback interface:
4486 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4490 BN_GENCB_set(&my_cb, my_callback, my_arg);
4492 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4493 /* For the meaning of a, b in calls to my_callback(), see the
4494 * documentation of the function that calls the callback.
4495 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4496 * my_callback should return 1 if it wants BN_is_prime_ex()
4497 * to continue, or 0 to stop.
4502 *) Change the ZLIB compression method to be stateful, and make it
4503 available to TLS with the number defined in
4504 draft-ietf-tls-compression-04.txt.
4507 *) Add the ASN.1 structures and functions for CertificatePair, which
4508 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4510 CertificatePair ::= SEQUENCE {
4511 forward [0] Certificate OPTIONAL,
4512 reverse [1] Certificate OPTIONAL,
4513 -- at least one of the pair shall be present -- }
4515 Also implement the PEM functions to read and write certificate
4516 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4518 This needed to be defined, mostly for the sake of the LDAP
4519 attribute crossCertificatePair, but may prove useful elsewhere as
4523 *) Make it possible to inhibit symlinking of shared libraries in
4524 Makefile.shared, for Cygwin's sake.
4527 *) Extend the BIGNUM API by creating a function
4528 void BN_set_negative(BIGNUM *a, int neg);
4529 and a macro that behave like
4530 int BN_is_negative(const BIGNUM *a);
4532 to avoid the need to access 'a->neg' directly in applications.
4535 *) Implement fast modular reduction for pseudo-Mersenne primes
4536 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4537 EC_GROUP_new_curve_GFp() will now automatically use this
4539 [Nils Larsch <nla@trustcenter.de>]
4541 *) Add new lock type (CRYPTO_LOCK_BN).
4544 *) Change the ENGINE framework to automatically load engines
4545 dynamically from specific directories unless they could be
4546 found to already be built in or loaded. Move all the
4547 current engines except for the cryptodev one to a new
4549 The engines in engines/ are built as shared libraries if
4550 the "shared" options was given to ./Configure or ./config.
4551 Otherwise, they are inserted in libcrypto.a.
4552 /usr/local/ssl/engines is the default directory for dynamic
4553 engines, but that can be overriden at configure time through
4554 the usual use of --prefix and/or --openssldir, and at run
4555 time with the environment variable OPENSSL_ENGINES.
4556 [Geoff Thorpe and Richard Levitte]
4558 *) Add Makefile.shared, a helper makefile to build shared
4559 libraries. Addapt Makefile.org.
4562 *) Add version info to Win32 DLLs.
4563 [Peter 'Luna' Runestig" <peter@runestig.com>]
4565 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4566 can be added using this API to created arbitrary PKCS#12
4567 files while avoiding the low level API.
4569 New options to PKCS12_create(), key or cert can be NULL and
4570 will then be omitted from the output file. The encryption
4571 algorithm NIDs can be set to -1 for no encryption, the mac
4572 iteration count can be set to 0 to omit the mac.
4574 Enhance pkcs12 utility by making the -nokeys and -nocerts
4575 options work when creating a PKCS#12 file. New option -nomac
4576 to omit the mac, NONE can be set for an encryption algorithm.
4577 New code is modified to use the enhanced PKCS12_create()
4578 instead of the low level API.
4581 *) Extend ASN1 encoder to support indefinite length constructed
4582 encoding. This can output sequences tags and octet strings in
4583 this form. Modify pk7_asn1.c to support indefinite length
4584 encoding. This is experimental and needs additional code to
4585 be useful, such as an ASN1 bio and some enhanced streaming
4588 Extend template encode functionality so that tagging is passed
4589 down to the template encoder.
4592 *) Let 'openssl req' fail if an argument to '-newkey' is not
4593 recognized instead of using RSA as a default.
4596 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4597 As these are not official, they are not included in "ALL";
4598 the "ECCdraft" ciphersuite group alias can be used to select them.
4599 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4601 *) Add ECDH engine support.
4602 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4604 *) Add ECDH in new directory crypto/ecdh/.
4605 [Douglas Stebila (Sun Microsystems Laboratories)]
4607 *) Let BN_rand_range() abort with an error after 100 iterations
4608 without success (which indicates a broken PRNG).
4611 *) Change BN_mod_sqrt() so that it verifies that the input value
4612 is really the square of the return value. (Previously,
4613 BN_mod_sqrt would show GIGO behaviour.)
4616 *) Add named elliptic curves over binary fields from X9.62, SECG,
4617 and WAP/WTLS; add OIDs that were still missing.
4619 [Sheueling Chang Shantz and Douglas Stebila
4620 (Sun Microsystems Laboratories)]
4622 *) Extend the EC library for elliptic curves over binary fields
4623 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4626 EC_GF2m_simple_method
4630 EC_GROUP_new_curve_GF2m
4631 EC_GROUP_set_curve_GF2m
4632 EC_GROUP_get_curve_GF2m
4633 EC_POINT_set_affine_coordinates_GF2m
4634 EC_POINT_get_affine_coordinates_GF2m
4635 EC_POINT_set_compressed_coordinates_GF2m
4637 Point compression for binary fields is disabled by default for
4638 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4641 As binary polynomials are represented as BIGNUMs, various members
4642 of the EC_GROUP and EC_POINT data structures can be shared
4643 between the implementations for prime fields and binary fields;
4644 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4645 are essentially identical to their ..._GFp counterparts.
4646 (For simplicity, the '..._GFp' prefix has been dropped from
4647 various internal method names.)
4649 An internal 'field_div' method (similar to 'field_mul' and
4650 'field_sqr') has been added; this is used only for binary fields.
4652 [Sheueling Chang Shantz and Douglas Stebila
4653 (Sun Microsystems Laboratories)]
4655 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4656 through methods ('mul', 'precompute_mult').
4658 The generic implementations (now internally called 'ec_wNAF_mul'
4659 and 'ec_wNAF_precomputed_mult') remain the default if these
4660 methods are undefined.
4662 [Sheueling Chang Shantz and Douglas Stebila
4663 (Sun Microsystems Laboratories)]
4665 *) New function EC_GROUP_get_degree, which is defined through
4666 EC_METHOD. For curves over prime fields, this returns the bit
4667 length of the modulus.
4669 [Sheueling Chang Shantz and Douglas Stebila
4670 (Sun Microsystems Laboratories)]
4672 *) New functions EC_GROUP_dup, EC_POINT_dup.
4673 (These simply call ..._new and ..._copy).
4675 [Sheueling Chang Shantz and Douglas Stebila
4676 (Sun Microsystems Laboratories)]
4678 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4679 Polynomials are represented as BIGNUMs (where the sign bit is not
4680 used) in the following functions [macros]:
4683 BN_GF2m_sub [= BN_GF2m_add]
4684 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4685 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4686 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4688 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4689 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4690 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4691 BN_GF2m_cmp [= BN_ucmp]
4693 (Note that only the 'mod' functions are actually for fields GF(2^m).
4694 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4696 For some functions, an the irreducible polynomial defining a
4697 field can be given as an 'unsigned int[]' with strictly
4698 decreasing elements giving the indices of those bits that are set;
4699 i.e., p[] represents the polynomial
4700 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4702 p[0] > p[1] > ... > p[k] = 0.
4703 This applies to the following functions:
4708 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4709 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4711 BN_GF2m_mod_sqrt_arr
4712 BN_GF2m_mod_solve_quad_arr
4716 Conversion can be performed by the following functions:
4721 bntest.c has additional tests for binary polynomial arithmetic.
4723 Two implementations for BN_GF2m_mod_div() are available.
4724 The default algorithm simply uses BN_GF2m_mod_inv() and
4725 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4726 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4727 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4729 [Sheueling Chang Shantz and Douglas Stebila
4730 (Sun Microsystems Laboratories)]
4732 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4733 functionality is disabled at compile-time.
4734 [Douglas Stebila <douglas.stebila@sun.com>]
4736 *) Change default behaviour of 'openssl asn1parse' so that more
4737 information is visible when viewing, e.g., a certificate:
4739 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4740 mode the content of non-printable OCTET STRINGs is output in a
4741 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4742 avoid the appearance of a printable string.
4743 [Nils Larsch <nla@trustcenter.de>]
4745 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4747 EC_GROUP_set_asn1_flag()
4748 EC_GROUP_get_asn1_flag()
4749 EC_GROUP_set_point_conversion_form()
4750 EC_GROUP_get_point_conversion_form()
4751 These control ASN1 encoding details:
4752 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4753 has been set to OPENSSL_EC_NAMED_CURVE.
4754 - Points are encoded in uncompressed form by default; options for
4755 asn1_for are as for point2oct, namely
4756 POINT_CONVERSION_COMPRESSED
4757 POINT_CONVERSION_UNCOMPRESSED
4758 POINT_CONVERSION_HYBRID
4760 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4763 EC_GROUP_get0_seed()
4764 EC_GROUP_get_seed_len()
4765 This is used only for ASN1 purposes (so far).
4766 [Nils Larsch <nla@trustcenter.de>]
4768 *) Add 'field_type' member to EC_METHOD, which holds the NID
4769 of the appropriate field type OID. The new function
4770 EC_METHOD_get_field_type() returns this value.
4771 [Nils Larsch <nla@trustcenter.de>]
4776 EC_POINT_point2hex()
4777 EC_POINT_hex2point()
4778 providing useful interfaces to EC_POINT_point2oct() and
4779 EC_POINT_oct2point().
4780 [Nils Larsch <nla@trustcenter.de>]
4782 *) Change internals of the EC library so that the functions
4783 EC_GROUP_set_generator()
4784 EC_GROUP_get_generator()
4785 EC_GROUP_get_order()
4786 EC_GROUP_get_cofactor()
4787 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4788 to methods, which would lead to unnecessary code duplication when
4789 adding different types of curves.
4790 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4792 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4793 arithmetic, and such that modified wNAFs are generated
4794 (which avoid length expansion in many cases).
4797 *) Add a function EC_GROUP_check_discriminant() (defined via
4798 EC_METHOD) that verifies that the curve discriminant is non-zero.
4800 Add a function EC_GROUP_check() that makes some sanity tests
4801 on a EC_GROUP, its generator and order. This includes
4802 EC_GROUP_check_discriminant().
4803 [Nils Larsch <nla@trustcenter.de>]
4805 *) Add ECDSA in new directory crypto/ecdsa/.
4807 Add applications 'openssl ecparam' and 'openssl ecdsa'
4808 (these are based on 'openssl dsaparam' and 'openssl dsa').
4810 ECDSA support is also included in various other files across the
4811 library. Most notably,
4812 - 'openssl req' now has a '-newkey ecdsa:file' option;
4813 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4814 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4815 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4816 them suitable for ECDSA where domain parameters must be
4817 extracted before the specific public key;
4818 - ECDSA engine support has been added.
4819 [Nils Larsch <nla@trustcenter.de>]
4821 *) Include some named elliptic curves, and add OIDs from X9.62,
4822 SECG, and WAP/WTLS. Each curve can be obtained from the new
4824 EC_GROUP_new_by_curve_name(),
4825 and the list of available named curves can be obtained with
4826 EC_get_builtin_curves().
4827 Also add a 'curve_name' member to EC_GROUP objects, which can be
4829 EC_GROUP_set_curve_name()
4830 EC_GROUP_get_curve_name()
4831 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4833 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4834 was actually never needed) and in BN_mul(). The removal in BN_mul()
4835 required a small change in bn_mul_part_recursive() and the addition
4836 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4837 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4838 bn_sub_words() and bn_add_words() except they take arrays with
4842 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4844 *) Cleanse PEM buffers before freeing them since they may contain
4846 [Benjamin Bennett <ben@psc.edu>]
4848 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4849 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4850 authentication-only ciphersuites.
4853 *) Since AES128 and AES256 share a single mask bit in the logic of
4854 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4855 kludge to work properly if AES128 is available and AES256 isn't.
4858 *) Expand security boundary to match 1.1.1 module.
4861 *) Remove redundant features: hash file source, editing of test vectors
4862 modify fipsld to use external fips_premain.c signature.
4865 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4866 run algorithm test programs.
4869 *) Make algorithm test programs more tolerant of whitespace.
4872 *) Have SSL/TLS server implementation tolerate "mismatched" record
4873 protocol version while receiving ClientHello even if the
4874 ClientHello is fragmented. (The server can't insist on the
4875 particular protocol version it has chosen before the ServerHello
4876 message has informed the client about his choice.)
4879 *) Load error codes if they are not already present instead of using a
4880 static variable. This allows them to be cleanly unloaded and reloaded.
4883 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4885 *) Introduce limits to prevent malicious keys being able to
4886 cause a denial of service. (CVE-2006-2940)
4887 [Steve Henson, Bodo Moeller]
4889 *) Fix ASN.1 parsing of certain invalid structures that can result
4890 in a denial of service. (CVE-2006-2937) [Steve Henson]
4892 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4893 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4895 *) Fix SSL client code which could crash if connecting to a
4896 malicious SSLv2 server. (CVE-2006-4343)
4897 [Tavis Ormandy and Will Drewry, Google Security Team]
4899 *) Change ciphersuite string processing so that an explicit
4900 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4901 will no longer include "AES128-SHA"), and any other similar
4902 ciphersuite (same bitmap) from *other* protocol versions (so that
4903 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4904 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4905 changes from 0.9.8b and 0.9.8d.
4908 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4910 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4911 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4913 *) Change the Unix randomness entropy gathering to use poll() when
4914 possible instead of select(), since the latter has some
4915 undesirable limitations.
4916 [Darryl Miles via Richard Levitte and Bodo Moeller]
4918 *) Disable rogue ciphersuites:
4920 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4921 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4922 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4924 The latter two were purportedly from
4925 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4928 Also deactive the remaining ciphersuites from
4929 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4930 unofficial, and the ID has long expired.
4933 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4934 dual-core machines) and other potential thread-safety issues.
4937 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4939 *) Adapt fipsld and the build system to link against the validated FIPS
4940 module in FIPS mode.
4943 *) Fixes for VC++ 2005 build under Windows.
4946 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4947 from a Windows bash shell such as MSYS. It is autodetected from the
4948 "config" script when run from a VC++ environment. Modify standard VC++
4949 build to use fipscanister.o from the GNU make build.
4952 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4954 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4955 The value now differs depending on if you build for FIPS or not.
4956 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4957 safely run with a non-FIPSed libcrypto, as it may crash because of
4958 the difference induced by this change.
4961 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4963 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4964 (part of SSL_OP_ALL). This option used to disable the
4965 countermeasure against man-in-the-middle protocol-version
4966 rollback in the SSL 2.0 server implementation, which is a bad
4967 idea. (CVE-2005-2969)
4969 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4970 for Information Security, National Institute of Advanced Industrial
4971 Science and Technology [AIST], Japan)]
4973 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4974 mainly for FIPS compliance and not fully integrated at this stage.
4977 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4978 the exponentiation using a fixed-length exponent. (Otherwise,
4979 the information leaked through timing could expose the secret key
4980 after many signatures; cf. Bleichenbacher's attack on DSA with
4984 *) Make a new fixed-window mod_exp implementation the default for
4985 RSA, DSA, and DH private-key operations so that the sequence of
4986 squares and multiplies and the memory access pattern are
4987 independent of the particular secret key. This will mitigate
4988 cache-timing and potential related attacks.
4990 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4991 and this is automatically used by BN_mod_exp_mont() if the new flag
4992 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4993 will use this BN flag for private exponents unless the flag
4994 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4995 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4997 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4999 *) Change the client implementation for SSLv23_method() and
5000 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
5001 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
5002 (Previously, the SSL 2.0 backwards compatible Client Hello
5003 message format would be used even with SSL_OP_NO_SSLv2.)
5006 *) Add support for smime-type MIME parameter in S/MIME messages which some
5010 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
5011 a threadsafe manner. Modify rsa code to use new function and add calls
5012 to dsa and dh code (which had race conditions before).
5015 *) Include the fixed error library code in the C error file definitions
5016 instead of fixing them up at runtime. This keeps the error code
5017 structures constant.
5020 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
5022 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
5025 *) Fixes for newer kerberos headers. NB: the casts are needed because
5026 the 'length' field is signed on one version and unsigned on another
5027 with no (?) obvious way to tell the difference, without these VC++
5028 complains. Also the "definition" of FAR (blank) is no longer included
5029 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
5030 some needed definitions.
5033 *) Undo Cygwin change.
5036 *) Added support for proxy certificates according to RFC 3820.
5037 Because they may be a security thread to unaware applications,
5038 they must be explicitely allowed in run-time. See
5039 docs/HOWTO/proxy_certificates.txt for further information.
5042 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
5044 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
5045 server and client random values. Previously
5046 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
5047 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
5049 This change has negligible security impact because:
5051 1. Server and client random values still have 24 bytes of pseudo random
5054 2. Server and client random values are sent in the clear in the initial
5057 3. The master secret is derived using the premaster secret (48 bytes in
5058 size for static RSA ciphersuites) as well as client server and random
5061 The OpenSSL team would like to thank the UK NISCC for bringing this issue
5064 [Stephen Henson, reported by UK NISCC]
5066 *) Use Windows randomness collection on Cygwin.
5069 *) Fix hang in EGD/PRNGD query when communication socket is closed
5070 prematurely by EGD/PRNGD.
5071 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
5073 *) Prompt for pass phrases when appropriate for PKCS12 input format.
5076 *) Back-port of selected performance improvements from development
5077 branch, as well as improved support for PowerPC platforms.
5080 *) Add lots of checks for memory allocation failure, error codes to indicate
5081 failure and freeing up memory if a failure occurs.
5082 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
5084 *) Add new -passin argument to dgst.
5087 *) Perform some character comparisons of different types in X509_NAME_cmp:
5088 this is needed for some certificates that reencode DNs into UTF8Strings
5089 (in violation of RFC3280) and can't or wont issue name rollover
5093 *) Make an explicit check during certificate validation to see that
5094 the CA setting in each certificate on the chain is correct. As a
5095 side effect always do the following basic checks on extensions,
5096 not just when there's an associated purpose to the check:
5098 - if there is an unhandled critical extension (unless the user
5099 has chosen to ignore this fault)
5100 - if the path length has been exceeded (if one is set at all)
5101 - that certain extensions fit the associated purpose (if one has
5105 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
5107 *) Avoid a race condition when CRLs are checked in a multi threaded
5108 environment. This would happen due to the reordering of the revoked
5109 entries during signature checking and serial number lookup. Now the
5110 encoding is cached and the serial number sort performed under a lock.
5111 Add new STACK function sk_is_sorted().
5114 *) Add Delta CRL to the extension code.
5117 *) Various fixes to s3_pkt.c so alerts are sent properly.
5118 [David Holmes <d.holmes@f5.com>]
5120 *) Reduce the chances of duplicate issuer name and serial numbers (in
5121 violation of RFC3280) using the OpenSSL certificate creation utilities.
5122 This is done by creating a random 64 bit value for the initial serial
5123 number when a serial number file is created or when a self signed
5124 certificate is created using 'openssl req -x509'. The initial serial
5125 number file is created using 'openssl x509 -next_serial' in CA.pl
5126 rather than being initialized to 1.
5129 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
5131 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5132 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5133 [Joe Orton, Steve Henson]
5135 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
5137 [Joe Orton, Steve Henson]
5139 *) Make it possible to have multiple active certificates with the same
5140 subject in the CA index file. This is done only if the keyword
5141 'unique_subject' is set to 'no' in the main CA section (default
5142 if 'CA_default') of the configuration file. The value is saved
5143 with the database itself in a separate index attribute file,
5144 named like the index file with '.attr' appended to the name.
5147 *) X509 verify fixes. Disable broken certificate workarounds when
5148 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
5149 keyUsage extension present. Don't accept CRLs with unhandled critical
5150 extensions: since verify currently doesn't process CRL extensions this
5151 rejects a CRL with *any* critical extensions. Add new verify error codes
5155 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
5156 A clarification of RFC2560 will require the use of OCTET STRINGs and
5157 some implementations cannot handle the current raw format. Since OpenSSL
5158 copies and compares OCSP nonces as opaque blobs without any attempt at
5159 parsing them this should not create any compatibility issues.
5162 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
5163 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
5164 this HMAC (and other) operations are several times slower than OpenSSL
5168 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
5169 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5171 *) Use the correct content when signing type "other".
5174 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5176 *) Fix various bugs revealed by running the NISCC test suite:
5178 Stop out of bounds reads in the ASN1 code when presented with
5179 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5181 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
5183 If verify callback ignores invalid public key errors don't try to check
5184 certificate signature with the NULL public key.
5188 *) New -ignore_err option in ocsp application to stop the server
5189 exiting on the first error in a request.
5192 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5193 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5197 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5198 extra data after the compression methods not only for TLS 1.0
5199 but also for SSL 3.0 (as required by the specification).
5200 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5202 *) Change X509_certificate_type() to mark the key as exported/exportable
5203 when it's 512 *bits* long, not 512 bytes.
5206 *) Change AES_cbc_encrypt() so it outputs exact multiple of
5207 blocks during encryption.
5210 *) Various fixes to base64 BIO and non blocking I/O. On write
5211 flushes were not handled properly if the BIO retried. On read
5212 data was not being buffered properly and had various logic bugs.
5213 This also affects blocking I/O when the data being decoded is a
5217 *) Various S/MIME bugfixes and compatibility changes:
5218 output correct application/pkcs7 MIME type if
5219 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5220 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5221 of files as .eml work). Correctly handle very long lines in MIME
5225 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5227 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5228 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5229 a protocol version number mismatch like a decryption error
5230 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5233 *) Turn on RSA blinding by default in the default implementation
5234 to avoid a timing attack. Applications that don't want it can call
5235 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5236 They would be ill-advised to do so in most cases.
5237 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5239 *) Change RSA blinding code so that it works when the PRNG is not
5240 seeded (in this case, the secret RSA exponent is abused as
5241 an unpredictable seed -- if it is not unpredictable, there
5242 is no point in blinding anyway). Make RSA blinding thread-safe
5243 by remembering the creator's thread ID in rsa->blinding and
5244 having all other threads use local one-time blinding factors
5245 (this requires more computation than sharing rsa->blinding, but
5246 avoids excessive locking; and if an RSA object is not shared
5247 between threads, blinding will still be very fast).
5250 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5251 ENGINE as defaults for all supported algorithms irrespective of
5252 the 'flags' parameter. 'flags' is now honoured, so applications
5253 should make sure they are passing it correctly.
5256 *) Target "mingw" now allows native Windows code to be generated in
5257 the Cygwin environment as well as with the MinGW compiler.
5260 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5262 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5263 via timing by performing a MAC computation even if incorrrect
5264 block cipher padding has been found. This is a countermeasure
5265 against active attacks where the attacker has to distinguish
5266 between bad padding and a MAC verification error. (CVE-2003-0078)
5268 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5269 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5270 Martin Vuagnoux (EPFL, Ilion)]
5272 *) Make the no-err option work as intended. The intention with no-err
5273 is not to have the whole error stack handling routines removed from
5274 libcrypto, it's only intended to remove all the function name and
5275 reason texts, thereby removing some of the footprint that may not
5276 be interesting if those errors aren't displayed anyway.
5278 NOTE: it's still possible for any application or module to have it's
5279 own set of error texts inserted. The routines are there, just not
5280 used by default when no-err is given.
5283 *) Add support for FreeBSD on IA64.
5284 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5286 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5287 Kerberos function mit_des_cbc_cksum(). Before this change,
5288 the value returned by DES_cbc_cksum() was like the one from
5289 mit_des_cbc_cksum(), except the bytes were swapped.
5290 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5292 *) Allow an application to disable the automatic SSL chain building.
5293 Before this a rather primitive chain build was always performed in
5294 ssl3_output_cert_chain(): an application had no way to send the
5295 correct chain if the automatic operation produced an incorrect result.
5297 Now the chain builder is disabled if either:
5299 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5301 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5303 The reasoning behind this is that an application would not want the
5304 auto chain building to take place if extra chain certificates are
5305 present and it might also want a means of sending no additional
5306 certificates (for example the chain has two certificates and the
5310 *) Add the possibility to build without the ENGINE framework.
5311 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5313 *) Under Win32 gmtime() can return NULL: check return value in
5314 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5317 *) DSA routines: under certain error conditions uninitialized BN objects
5318 could be freed. Solution: make sure initialization is performed early
5319 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5320 Nils Larsch <nla@trustcenter.de> via PR#459)
5323 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5324 checked on reconnect on the client side, therefore session resumption
5325 could still fail with a "ssl session id is different" error. This
5326 behaviour is masked when SSL_OP_ALL is used due to
5327 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5328 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5329 followup to PR #377.
5332 *) IA-32 assembler support enhancements: unified ELF targets, support
5333 for SCO/Caldera platforms, fix for Cygwin shared build.
5336 *) Add support for FreeBSD on sparc64. As a consequence, support for
5337 FreeBSD on non-x86 processors is separate from x86 processors on
5338 the config script, much like the NetBSD support.
5339 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5341 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5343 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5346 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5347 code (06) was taken as the first octet of the session ID and the last
5348 octet was ignored consequently. As a result SSLv2 client side session
5349 caching could not have worked due to the session ID mismatch between
5351 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5355 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5356 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5360 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5361 seems that in spite of existing for more than a year, many application
5362 author have done nothing to provide the necessary callbacks, which
5363 means that this particular engine will not work properly anywhere.
5364 This is a very unfortunate situation which forces us, in the name
5365 of usability, to give the hw_ncipher.c a static lock, which is part
5367 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5368 appear in 0.9.8 or later. We EXPECT application authors to have
5369 dealt properly with this when 0.9.8 is released (unless we actually
5370 make such changes in the libcrypto locking code that changes will
5371 have to be made anyway).
5374 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5375 octets have been read, EOF or an error occurs. Without this change
5376 some truncated ASN1 structures will not produce an error.
5379 *) Disable Heimdal support, since it hasn't been fully implemented.
5380 Still give the possibility to force the use of Heimdal, but with
5381 warnings and a request that patches get sent to openssl-dev.
5384 *) Add the VC-CE target, introduce the WINCE sysname, and add
5385 INSTALL.WCE and appropriate conditionals to make it build.
5386 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5388 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5389 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5390 edit numbers of the version.
5391 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5393 *) Introduce safe string copy and catenation functions
5394 (BUF_strlcpy() and BUF_strlcat()).
5395 [Ben Laurie (CHATS) and Richard Levitte]
5397 *) Avoid using fixed-size buffers for one-line DNs.
5398 [Ben Laurie (CHATS)]
5400 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5401 resizing buffers containing secrets, and use where appropriate.
5402 [Ben Laurie (CHATS)]
5404 *) Avoid using fixed size buffers for configuration file location.
5405 [Ben Laurie (CHATS)]
5407 *) Avoid filename truncation for various CA files.
5408 [Ben Laurie (CHATS)]
5410 *) Use sizeof in preference to magic numbers.
5411 [Ben Laurie (CHATS)]
5413 *) Avoid filename truncation in cert requests.
5414 [Ben Laurie (CHATS)]
5416 *) Add assertions to check for (supposedly impossible) buffer
5418 [Ben Laurie (CHATS)]
5420 *) Don't cache truncated DNS entries in the local cache (this could
5421 potentially lead to a spoofing attack).
5422 [Ben Laurie (CHATS)]
5424 *) Fix various buffers to be large enough for hex/decimal
5425 representations in a platform independent manner.
5426 [Ben Laurie (CHATS)]
5428 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5429 resizing buffers containing secrets, and use where appropriate.
5430 [Ben Laurie (CHATS)]
5432 *) Add BIO_indent() to avoid much slightly worrying code to do
5434 [Ben Laurie (CHATS)]
5436 *) Convert sprintf()/BIO_puts() to BIO_printf().
5437 [Ben Laurie (CHATS)]
5439 *) buffer_gets() could terminate with the buffer only half
5441 [Ben Laurie (CHATS)]
5443 *) Add assertions to prevent user-supplied crypto functions from
5444 overflowing internal buffers by having large block sizes, etc.
5445 [Ben Laurie (CHATS)]
5447 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5449 [Ben Laurie (CHATS)]
5451 *) Eliminate unused copy of key in RC4.
5452 [Ben Laurie (CHATS)]
5454 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5455 [Ben Laurie (CHATS)]
5457 *) Fix off-by-one error in EGD path.
5458 [Ben Laurie (CHATS)]
5460 *) If RANDFILE path is too long, ignore instead of truncating.
5461 [Ben Laurie (CHATS)]
5463 *) Eliminate unused and incorrectly sized X.509 structure
5465 [Ben Laurie (CHATS)]
5467 *) Eliminate unused and dangerous function knumber().
5468 [Ben Laurie (CHATS)]
5470 *) Eliminate unused and dangerous structure, KSSL_ERR.
5471 [Ben Laurie (CHATS)]
5473 *) Protect against overlong session ID context length in an encoded
5474 session object. Since these are local, this does not appear to be
5476 [Ben Laurie (CHATS)]
5478 *) Change from security patch (see 0.9.6e below) that did not affect
5479 the 0.9.6 release series:
5481 Remote buffer overflow in SSL3 protocol - an attacker could
5482 supply an oversized master key in Kerberos-enabled versions.
5484 [Ben Laurie (CHATS)]
5486 *) Change the SSL kerb5 codes to match RFC 2712.
5489 *) Make -nameopt work fully for req and add -reqopt switch.
5490 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5492 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5493 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5495 *) Make sure tests can be performed even if the corresponding algorithms
5496 have been removed entirely. This was also the last step to make
5497 OpenSSL compilable with DJGPP under all reasonable conditions.
5498 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5500 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5501 to allow version independent disabling of normally unselected ciphers,
5502 which may be activated as a side-effect of selecting a single cipher.
5504 (E.g., cipher list string "RSA" enables ciphersuites that are left
5505 out of "ALL" because they do not provide symmetric encryption.
5506 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5507 [Lutz Jaenicke, Bodo Moeller]
5509 *) Add appropriate support for separate platform-dependent build
5510 directories. The recommended way to make a platform-dependent
5511 build directory is the following (tested on Linux), maybe with
5514 # Place yourself outside of the OpenSSL source tree. In
5515 # this example, the environment variable OPENSSL_SOURCE
5516 # is assumed to contain the absolute OpenSSL source directory.
5517 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5518 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5519 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5520 mkdir -p `dirname $F`
5521 ln -s $OPENSSL_SOURCE/$F $F
5524 To be absolutely sure not to disturb the source tree, a "make clean"
5525 is a good thing. If it isn't successfull, don't worry about it,
5526 it probably means the source directory is very clean.
5529 *) Make sure any ENGINE control commands make local copies of string
5530 pointers passed to them whenever necessary. Otherwise it is possible
5531 the caller may have overwritten (or deallocated) the original string
5532 data when a later ENGINE operation tries to use the stored values.
5533 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5535 *) Improve diagnostics in file reading and command-line digests.
5536 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5538 *) Add AES modes CFB and OFB to the object database. Correct an
5539 error in AES-CFB decryption.
5542 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5543 allows existing EVP_CIPHER_CTX structures to be reused after
5544 calling EVP_*Final(). This behaviour is used by encryption
5545 BIOs and some applications. This has the side effect that
5546 applications must explicitly clean up cipher contexts with
5547 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5550 *) Check the values of dna and dnb in bn_mul_recursive before calling
5551 bn_mul_comba (a non zero value means the a or b arrays do not contain
5552 n2 elements) and fallback to bn_mul_normal if either is not zero.
5555 *) Fix escaping of non-ASCII characters when using the -subj option
5556 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5559 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5560 form for "surname", serialNumber has no short form.
5561 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5562 therefore remove "mail" short name for "internet 7".
5563 The OID for unique identifiers in X509 certificates is
5564 x500UniqueIdentifier, not uniqueIdentifier.
5565 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5568 *) Add an "init" command to the ENGINE config module and auto initialize
5569 ENGINEs. Without any "init" command the ENGINE will be initialized
5570 after all ctrl commands have been executed on it. If init=1 the
5571 ENGINE is initailized at that point (ctrls before that point are run
5572 on the uninitialized ENGINE and after on the initialized one). If
5573 init=0 then the ENGINE will not be iniatialized at all.
5576 *) Fix the 'app_verify_callback' interface so that the user-defined
5577 argument is actually passed to the callback: In the
5578 SSL_CTX_set_cert_verify_callback() prototype, the callback
5579 declaration has been changed from
5582 int (*cb)(X509_STORE_CTX *,void *);
5583 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5584 i=s->ctx->app_verify_callback(&ctx)
5585 has been changed into
5586 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5588 To update applications using SSL_CTX_set_cert_verify_callback(),
5589 a dummy argument can be added to their callback functions.
5590 [D. K. Smetters <smetters@parc.xerox.com>]
5592 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5593 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5595 *) Add and OPENSSL_LOAD_CONF define which will cause
5596 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5597 This allows older applications to transparently support certain
5598 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5599 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5600 load the config file and OPENSSL_add_all_algorithms_conf() which will
5601 always load it have also been added.
5604 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5605 Adjust NIDs and EVP layer.
5606 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5608 *) Config modules support in openssl utility.
5610 Most commands now load modules from the config file,
5611 though in a few (such as version) this isn't done
5612 because it couldn't be used for anything.
5614 In the case of ca and req the config file used is
5615 the same as the utility itself: that is the -config
5616 command line option can be used to specify an
5620 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5621 use "openssl_conf" if filename is NULL use default openssl config file.
5624 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5625 config section name. Add a new flag to tolerate a missing config file
5626 and move code to CONF_modules_load_file().
5629 *) Support for crypto accelerator cards from Accelerated Encryption
5630 Processing, www.aep.ie. (Use engine 'aep')
5631 The support was copied from 0.9.6c [engine] and adapted/corrected
5632 to work with the new engine framework.
5633 [AEP Inc. and Richard Levitte]
5635 *) Support for SureWare crypto accelerator cards from Baltimore
5636 Technologies. (Use engine 'sureware')
5637 The support was copied from 0.9.6c [engine] and adapted
5638 to work with the new engine framework.
5641 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5642 make the newer ENGINE framework commands for the CHIL engine work.
5643 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5645 *) Make it possible to produce shared libraries on ReliantUNIX.
5646 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5648 *) Add the configuration target debug-linux-ppro.
5649 Make 'openssl rsa' use the general key loading routines
5650 implemented in apps.c, and make those routines able to
5651 handle the key format FORMAT_NETSCAPE and the variant
5653 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5655 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5656 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5658 *) Add -keyform to rsautl, and document -engine.
5659 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5661 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5662 BIO_R_NO_SUCH_FILE error code rather than the generic
5663 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5666 *) Add new functions
5668 ERR_peek_last_error_line
5669 ERR_peek_last_error_line_data.
5670 These are similar to
5673 ERR_peek_error_line_data,
5674 but report on the latest error recorded rather than the first one
5675 still in the error queue.
5676 [Ben Laurie, Bodo Moeller]
5678 *) default_algorithms option in ENGINE config module. This allows things
5680 default_algorithms = ALL
5681 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5684 *) Prelminary ENGINE config module.
5687 *) New experimental application configuration code.
5690 *) Change the AES code to follow the same name structure as all other
5691 symmetric ciphers, and behave the same way. Move everything to
5692 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5693 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5695 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5696 [Ben Laurie and Theo de Raadt]
5698 *) Add option to output public keys in req command.
5699 [Massimiliano Pala madwolf@openca.org]
5701 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5702 (up to about 10% better than before for P-192 and P-224).
5705 *) New functions/macros
5707 SSL_CTX_set_msg_callback(ctx, cb)
5708 SSL_CTX_set_msg_callback_arg(ctx, arg)
5709 SSL_set_msg_callback(ssl, cb)
5710 SSL_set_msg_callback_arg(ssl, arg)
5712 to request calling a callback function
5714 void cb(int write_p, int version, int content_type,
5715 const void *buf, size_t len, SSL *ssl, void *arg)
5717 whenever a protocol message has been completely received
5718 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5719 protocol version according to which the SSL library interprets
5720 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5721 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5722 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5723 specification (change_cipher_spec(20), alert(21), handshake(22)).
5724 'buf' and 'len' point to the actual message, 'ssl' to the
5725 SSL object, and 'arg' is the application-defined value set by
5726 SSL[_CTX]_set_msg_callback_arg().
5728 'openssl s_client' and 'openssl s_server' have new '-msg' options
5729 to enable a callback that displays all protocol messages.
5732 *) Change the shared library support so shared libraries are built as
5733 soon as the corresponding static library is finished, and thereby get
5734 openssl and the test programs linked against the shared library.
5735 This still only happens when the keyword "shard" has been given to
5736 the configuration scripts.
5738 NOTE: shared library support is still an experimental thing, and
5739 backward binary compatibility is still not guaranteed.
5740 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5742 *) Add support for Subject Information Access extension.
5743 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5745 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5746 additional bytes when new memory had to be allocated, not just
5747 when reusing an existing buffer.
5750 *) New command line and configuration option 'utf8' for the req command.
5751 This allows field values to be specified as UTF8 strings.
5754 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5755 runs for the former and machine-readable output for the latter.
5758 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5759 of the e-mail address in the DN (i.e., it will go into a certificate
5760 extension only). The new configuration file option 'email_in_dn = no'
5761 has the same effect.
5762 [Massimiliano Pala madwolf@openca.org]
5764 *) Change all functions with names starting with des_ to be starting
5765 with DES_ instead. Add wrappers that are compatible with libdes,
5766 but are named _ossl_old_des_*. Finally, add macros that map the
5767 des_* symbols to the corresponding _ossl_old_des_* if libdes
5768 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5769 desired, the des_* symbols will be mapped to DES_*, with one
5772 Since we provide two compatibility mappings, the user needs to
5773 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5774 compatibility is desired. The default (i.e., when that macro
5775 isn't defined) is OpenSSL 0.9.6c compatibility.
5777 There are also macros that enable and disable the support of old
5778 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5779 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5780 are defined, the default will apply: to support the old des routines.
5782 In either case, one must include openssl/des.h to get the correct
5783 definitions. Do not try to just include openssl/des_old.h, that
5786 NOTE: This is a major break of an old API into a new one. Software
5787 authors are encouraged to switch to the DES_ style functions. Some
5788 time in the future, des_old.h and the libdes compatibility functions
5789 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5790 default), and then completely removed.
5793 *) Test for certificates which contain unsupported critical extensions.
5794 If such a certificate is found during a verify operation it is
5795 rejected by default: this behaviour can be overridden by either
5796 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5797 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5798 X509_supported_extension() has also been added which returns 1 if a
5799 particular extension is supported.
5802 *) Modify the behaviour of EVP cipher functions in similar way to digests
5803 to retain compatibility with existing code.
5806 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5807 compatibility with existing code. In particular the 'ctx' parameter does
5808 not have to be to be initialized before the call to EVP_DigestInit() and
5809 it is tidied up after a call to EVP_DigestFinal(). New function
5810 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5811 EVP_MD_CTX_copy() changed to not require the destination to be
5812 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5813 requires the destination to be valid.
5815 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5816 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5819 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5820 so that complete 'Handshake' protocol structures are kept in memory
5821 instead of overwriting 'msg_type' and 'length' with 'body' data.
5824 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5825 [Massimo Santin via Richard Levitte]
5827 *) Major restructuring to the underlying ENGINE code. This includes
5828 reduction of linker bloat, separation of pure "ENGINE" manipulation
5829 (initialisation, etc) from functionality dealing with implementations
5830 of specific crypto iterfaces. This change also introduces integrated
5831 support for symmetric ciphers and digest implementations - so ENGINEs
5832 can now accelerate these by providing EVP_CIPHER and EVP_MD
5833 implementations of their own. This is detailed in crypto/engine/README
5834 as it couldn't be adequately described here. However, there are a few
5835 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5836 were changed in the original introduction of ENGINE code have now
5837 reverted back - the hooking from this code to ENGINE is now a good
5838 deal more passive and at run-time, operations deal directly with
5839 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5840 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5841 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5842 they were not being used by the framework as there is no concept of a
5843 BIGNUM_METHOD and they could not be generalised to the new
5844 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5845 ENGINE_cpy() has been removed as it cannot be consistently defined in
5849 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5852 *) Change mkdef.pl to sort symbols that get the same entry number,
5853 and make sure the automatically generated functions ERR_load_*
5854 become part of libeay.num as well.
5857 *) New function SSL_renegotiate_pending(). This returns true once
5858 renegotiation has been requested (either SSL_renegotiate() call
5859 or HelloRequest/ClientHello receveived from the peer) and becomes
5860 false once a handshake has been completed.
5861 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5862 sends a HelloRequest, but does not ensure that a handshake takes
5863 place. SSL_renegotiate_pending() is useful for checking if the
5864 client has followed the request.)
5867 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5868 By default, clients may request session resumption even during
5869 renegotiation (if session ID contexts permit); with this option,
5870 session resumption is possible only in the first handshake.
5872 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5873 more bits available for options that should not be part of
5874 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5877 *) Add some demos for certificate and certificate request creation.
5880 *) Make maximum certificate chain size accepted from the peer application
5881 settable (SSL*_get/set_max_cert_list()), as proposed by
5882 "Douglas E. Engert" <deengert@anl.gov>.
5885 *) Add support for shared libraries for Unixware-7
5886 (Boyd Lynn Gerber <gerberb@zenez.com>).
5889 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5890 be done prior to destruction. Use this to unload error strings from
5891 ENGINEs that load their own error strings. NB: This adds two new API
5892 functions to "get" and "set" this destroy handler in an ENGINE.
5895 *) Alter all existing ENGINE implementations (except "openssl" and
5896 "openbsd") to dynamically instantiate their own error strings. This
5897 makes them more flexible to be built both as statically-linked ENGINEs
5898 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5899 Also, add stub code to each that makes building them as self-contained
5900 shared-libraries easier (see README.ENGINE).
5903 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5904 implementations into applications that are completely implemented in
5905 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5906 commands that can be used to configure what shared-library to load and
5907 to control aspects of the way it is handled. Also, made an update to
5908 the README.ENGINE file that brings its information up-to-date and
5909 provides some information and instructions on the "dynamic" ENGINE
5910 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5913 *) Make it possible to unload ranges of ERR strings with a new
5914 "ERR_unload_strings" function.
5917 *) Add a copy() function to EVP_MD.
5920 *) Make EVP_MD routines take a context pointer instead of just the
5921 md_data void pointer.
5924 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5925 that the digest can only process a single chunk of data
5926 (typically because it is provided by a piece of
5927 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5928 is only going to provide a single chunk of data, and hence the
5929 framework needn't accumulate the data for oneshot drivers.
5932 *) As with "ERR", make it possible to replace the underlying "ex_data"
5933 functions. This change also alters the storage and management of global
5934 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5935 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5936 index counters. The API functions that use this state have been changed
5937 to take a "class_index" rather than pointers to the class's local STACK
5938 and counter, and there is now an API function to dynamically create new
5939 classes. This centralisation allows us to (a) plug a lot of the
5940 thread-safety problems that existed, and (b) makes it possible to clean
5941 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5942 such data would previously have always leaked in application code and
5943 workarounds were in place to make the memory debugging turn a blind eye
5944 to it. Application code that doesn't use this new function will still
5945 leak as before, but their memory debugging output will announce it now
5946 rather than letting it slide.
5948 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5949 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5950 has a return value to indicate success or failure.
5953 *) Make it possible to replace the underlying "ERR" functions such that the
5954 global state (2 LHASH tables and 2 locks) is only used by the "default"
5955 implementation. This change also adds two functions to "get" and "set"
5956 the implementation prior to it being automatically set the first time
5957 any other ERR function takes place. Ie. an application can call "get",
5958 pass the return value to a module it has just loaded, and that module
5959 can call its own "set" function using that value. This means the
5960 module's "ERR" operations will use (and modify) the error state in the
5961 application and not in its own statically linked copy of OpenSSL code.
5964 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5965 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5966 the operation, and provides a more encapsulated way for external code
5967 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5968 to use these functions rather than manually incrementing the counts.
5970 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5973 *) Add EVP test program.
5976 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5979 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5980 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5981 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5982 These allow a CRL to be built without having to access X509_CRL fields
5983 directly. Modify 'ca' application to use new functions.
5986 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5987 bug workarounds. Rollback attack detection is a security feature.
5988 The problem will only arise on OpenSSL servers when TLSv1 is not
5989 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5990 Software authors not wanting to support TLSv1 will have special reasons
5991 for their choice and can explicitly enable this option.
5992 [Bodo Moeller, Lutz Jaenicke]
5994 *) Rationalise EVP so it can be extended: don't include a union of
5995 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5996 (similar to those existing for EVP_CIPHER_CTX).
6001 EVP_MD_CTX_init(&md); /* new function call */
6002 EVP_DigestInit(&md, EVP_sha1());
6003 EVP_DigestUpdate(&md, in, len);
6004 EVP_DigestFinal(&md, out, NULL);
6005 EVP_MD_CTX_cleanup(&md); /* new function call */
6009 *) Make DES key schedule conform to the usual scheme, as well as
6010 correcting its structure. This means that calls to DES functions
6011 now have to pass a pointer to a des_key_schedule instead of a
6012 plain des_key_schedule (which was actually always a pointer
6015 des_key_schedule ks;
6017 des_set_key_checked(..., &ks);
6018 des_ncbc_encrypt(..., &ks, ...);
6020 (Note that a later change renames 'des_...' into 'DES_...'.)
6023 *) Initial reduction of linker bloat: the use of some functions, such as
6024 PEM causes large amounts of unused functions to be linked in due to
6025 poor organisation. For example pem_all.c contains every PEM function
6026 which has a knock on effect of linking in large amounts of (unused)
6027 ASN1 code. Grouping together similar functions and splitting unrelated
6028 functions prevents this.
6031 *) Cleanup of EVP macros.
6034 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
6035 correct _ecb suffix.
6038 *) Add initial OCSP responder support to ocsp application. The
6039 revocation information is handled using the text based index
6040 use by the ca application. The responder can either handle
6041 requests generated internally, supplied in files (for example
6042 via a CGI script) or using an internal minimal server.
6045 *) Add configuration choices to get zlib compression for TLS.
6048 *) Changes to Kerberos SSL for RFC 2712 compliance:
6049 1. Implemented real KerberosWrapper, instead of just using
6050 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
6051 2. Implemented optional authenticator field of KerberosWrapper.
6053 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
6054 and authenticator structs; see crypto/krb5/.
6056 Generalized Kerberos calls to support multiple Kerberos libraries.
6057 [Vern Staats <staatsvr@asc.hpc.mil>,
6058 Jeffrey Altman <jaltman@columbia.edu>
6059 via Richard Levitte]
6061 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
6062 already does with RSA. testdsa.h now has 'priv_key/pub_key'
6063 values for each of the key sizes rather than having just
6064 parameters (and 'speed' generating keys each time).
6067 *) Speed up EVP routines.
6070 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
6071 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
6072 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
6073 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
6075 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
6076 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
6077 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
6080 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
6082 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
6085 *) Added the OS2-EMX target.
6086 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
6088 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
6089 to support NCONF routines in extension code. New function CONF_set_nconf()
6090 to allow functions which take an NCONF to also handle the old LHASH
6091 structure: this means that the old CONF compatible routines can be
6092 retained (in particular wrt extensions) without having to duplicate the
6093 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
6096 *) Enhance the general user interface with mechanisms for inner control
6097 and with possibilities to have yes/no kind of prompts.
6100 *) Change all calls to low level digest routines in the library and
6101 applications to use EVP. Add missing calls to HMAC_cleanup() and
6102 don't assume HMAC_CTX can be copied using memcpy().
6103 [Verdon Walker <VWalker@novell.com>, Steve Henson]
6105 *) Add the possibility to control engines through control names but with
6106 arbitrary arguments instead of just a string.
6107 Change the key loaders to take a UI_METHOD instead of a callback
6108 function pointer. NOTE: this breaks binary compatibility with earlier
6109 versions of OpenSSL [engine].
6110 Adapt the nCipher code for these new conditions and add a card insertion
6114 *) Enhance the general user interface with mechanisms to better support
6115 dialog box interfaces, application-defined prompts, the possibility
6116 to use defaults (for example default passwords from somewhere else)
6117 and interrupts/cancellations.
6120 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
6121 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
6124 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
6125 tidy up some unnecessarily weird code in 'sk_new()').
6126 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
6128 *) Change the key loading routines for ENGINEs to use the same kind
6129 callback (pem_password_cb) as all other routines that need this
6133 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
6134 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
6135 than this minimum value is recommended.
6138 *) New random seeder for OpenVMS, using the system process statistics
6139 that are easily reachable.
6142 *) Windows apparently can't transparently handle global
6143 variables defined in DLLs. Initialisations such as:
6145 const ASN1_ITEM *it = &ASN1_INTEGER_it;
6147 wont compile. This is used by the any applications that need to
6148 declare their own ASN1 modules. This was fixed by adding the option
6149 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
6150 needed for static libraries under Win32.
6153 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
6154 setting of purpose and trust fields. New X509_STORE trust and
6155 purpose functions and tidy up setting in other SSL functions.
6158 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
6159 structure. These are inherited by X509_STORE_CTX when it is
6160 initialised. This allows various defaults to be set in the
6161 X509_STORE structure (such as flags for CRL checking and custom
6162 purpose or trust settings) for functions which only use X509_STORE_CTX
6163 internally such as S/MIME.
6165 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
6166 trust settings if they are not set in X509_STORE. This allows X509_STORE
6167 purposes and trust (in S/MIME for example) to override any set by default.
6169 Add command line options for CRL checking to smime, s_client and s_server
6173 *) Initial CRL based revocation checking. If the CRL checking flag(s)
6174 are set then the CRL is looked up in the X509_STORE structure and
6175 its validity and signature checked, then if the certificate is found
6176 in the CRL the verify fails with a revoked error.
6178 Various new CRL related callbacks added to X509_STORE_CTX structure.
6180 Command line options added to 'verify' application to support this.
6182 This needs some additional work, such as being able to handle multiple
6183 CRLs with different times, extension based lookup (rather than just
6184 by subject name) and ultimately more complete V2 CRL extension
6188 *) Add a general user interface API (crypto/ui/). This is designed
6189 to replace things like des_read_password and friends (backward
6190 compatibility functions using this new API are provided).
6191 The purpose is to remove prompting functions from the DES code
6192 section as well as provide for prompting through dialog boxes in
6193 a window system and the like.
6196 *) Add "ex_data" support to ENGINE so implementations can add state at a
6197 per-structure level rather than having to store it globally.
6200 *) Make it possible for ENGINE structures to be copied when retrieved by
6201 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
6202 This causes the "original" ENGINE structure to act like a template,
6203 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
6204 operational state can be localised to each ENGINE structure, despite the
6205 fact they all share the same "methods". New ENGINE structures returned in
6206 this case have no functional references and the return value is the single
6207 structural reference. This matches the single structural reference returned
6208 by ENGINE_by_id() normally, when it is incremented on the pre-existing
6212 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
6213 needs to match any other type at all we need to manually clear the
6217 *) Changes to the "openssl engine" utility to include;
6218 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6219 about an ENGINE's available control commands.
6220 - executing control commands from command line arguments using the
6221 '-pre' and '-post' switches. '-post' is only used if '-t' is
6222 specified and the ENGINE is successfully initialised. The syntax for
6223 the individual commands are colon-separated, for example;
6224 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6227 *) New dynamic control command support for ENGINEs. ENGINEs can now
6228 declare their own commands (numbers), names (strings), descriptions,
6229 and input types for run-time discovery by calling applications. A
6230 subset of these commands are implicitly classed as "executable"
6231 depending on their input type, and only these can be invoked through
6232 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6233 can be based on user input, config files, etc). The distinction is
6234 that "executable" commands cannot return anything other than a boolean
6235 result and can only support numeric or string input, whereas some
6236 discoverable commands may only be for direct use through
6237 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6238 pointers, or other custom uses. The "executable" commands are to
6239 support parameterisations of ENGINE behaviour that can be
6240 unambiguously defined by ENGINEs and used consistently across any
6241 OpenSSL-based application. Commands have been added to all the
6242 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6243 control over shared-library paths without source code alterations.
6246 *) Changed all ENGINE implementations to dynamically allocate their
6247 ENGINEs rather than declaring them statically. Apart from this being
6248 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6249 this also allows the implementations to compile without using the
6250 internal engine_int.h header.
6253 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6254 'const' value. Any code that should be able to modify a RAND_METHOD
6255 should already have non-const pointers to it (ie. they should only
6256 modify their own ones).
6259 *) Made a variety of little tweaks to the ENGINE code.
6260 - "atalla" and "ubsec" string definitions were moved from header files
6261 to C code. "nuron" string definitions were placed in variables
6262 rather than hard-coded - allowing parameterisation of these values
6263 later on via ctrl() commands.
6264 - Removed unused "#if 0"'d code.
6265 - Fixed engine list iteration code so it uses ENGINE_free() to release
6266 structural references.
6267 - Constified the RAND_METHOD element of ENGINE structures.
6268 - Constified various get/set functions as appropriate and added
6269 missing functions (including a catch-all ENGINE_cpy that duplicates
6270 all ENGINE values onto a new ENGINE except reference counts/state).
6271 - Removed NULL parameter checks in get/set functions. Setting a method
6272 or function to NULL is a way of cancelling out a previously set
6273 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6274 and doesn't justify the extra error symbols and code.
6275 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6276 flags from engine_int.h to engine.h.
6277 - Changed prototypes for ENGINE handler functions (init(), finish(),
6278 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6281 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6282 to the algorithm using long division. The binary algorithm can be
6283 used only if the modulus is odd. On 32-bit systems, it is faster
6284 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6285 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6286 up to 450 bits. In 64-bit environments, the binary algorithm
6287 appears to be advantageous for much longer moduli; here we use it
6288 for moduli up to 2048 bits.
6291 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6292 could not support the combine flag in choice fields.
6295 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6296 extensions from a certificate request to the certificate.
6299 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6300 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6301 file: this allows the display of the certificate about to be
6302 signed to be customised, to allow certain fields to be included
6303 or excluded and extension details. The old system didn't display
6304 multicharacter strings properly, omitted fields not in the policy
6305 and couldn't display additional details such as extensions.
6308 *) Function EC_POINTs_mul for multiple scalar multiplication
6309 of an arbitrary number of elliptic curve points
6310 \sum scalars[i]*points[i],
6311 optionally including the generator defined for the EC_GROUP:
6312 scalar*generator + \sum scalars[i]*points[i].
6314 EC_POINT_mul is a simple wrapper function for the typical case
6315 that the point list has just one item (besides the optional
6319 *) First EC_METHODs for curves over GF(p):
6321 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6322 operations and provides various method functions that can also
6323 operate with faster implementations of modular arithmetic.
6325 EC_GFp_mont_method() reuses most functions that are part of
6326 EC_GFp_simple_method, but uses Montgomery arithmetic.
6328 [Bodo Moeller; point addition and point doubling
6329 implementation directly derived from source code provided by
6330 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6332 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6333 crypto/ec/ec_lib.c):
6335 Curves are EC_GROUP objects (with an optional group generator)
6336 based on EC_METHODs that are built into the library.
6338 Points are EC_POINT objects based on EC_GROUP objects.
6340 Most of the framework would be able to handle curves over arbitrary
6341 finite fields, but as there are no obvious types for fields other
6342 than GF(p), some functions are limited to that for now.
6345 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6346 that the file contains a complete HTTP response.
6349 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6350 change the def and num file printf format specifier from "%-40sXXX"
6351 to "%-39s XXX". The latter will always guarantee a space after the
6352 field while the former will cause them to run together if the field
6353 is 40 of more characters long.
6356 *) Constify the cipher and digest 'method' functions and structures
6357 and modify related functions to take constant EVP_MD and EVP_CIPHER
6361 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6362 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6365 *) Modify EVP_Digest*() routines so they now return values. Although the
6366 internal software routines can never fail additional hardware versions
6370 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6372 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6373 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6376 ERR_R_NESTED_ASN1_ERROR
6378 ERR_R_MISSING_ASN1_EOS
6379 were 4 .. 9, conflicting with
6380 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6382 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6383 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6385 Add new error code 'ERR_R_INTERNAL_ERROR'.
6388 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6392 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6393 sets the subject name for a new request or supersedes the
6394 subject name in a given request. Formats that can be parsed are
6395 'CN=Some Name, OU=myOU, C=IT'
6397 'CN=Some Name/OU=myOU/C=IT'.
6399 Add options '-batch' and '-verbose' to 'openssl req'.
6400 [Massimiliano Pala <madwolf@hackmasters.net>]
6402 *) Introduce the possibility to access global variables through
6403 functions on platform were that's the best way to handle exporting
6404 global variables in shared libraries. To enable this functionality,
6405 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6406 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6407 is normally done by Configure or something similar).
6409 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6410 in the source file (foo.c) like this:
6412 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6413 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6415 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6416 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6418 OPENSSL_DECLARE_GLOBAL(int,foo);
6419 #define foo OPENSSL_GLOBAL_REF(foo)
6420 OPENSSL_DECLARE_GLOBAL(double,bar);
6421 #define bar OPENSSL_GLOBAL_REF(bar)
6423 The #defines are very important, and therefore so is including the
6424 header file everywhere where the defined globals are used.
6426 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6427 of ASN.1 items, but that structure is a bit different.
6429 The largest change is in util/mkdef.pl which has been enhanced with
6430 better and easier to understand logic to choose which symbols should
6431 go into the Windows .def files as well as a number of fixes and code
6432 cleanup (among others, algorithm keywords are now sorted
6433 lexicographically to avoid constant rewrites).
6436 *) In BN_div() keep a copy of the sign of 'num' before writing the
6437 result to 'rm' because if rm==num the value will be overwritten
6438 and produce the wrong result if 'num' is negative: this caused
6439 problems with BN_mod() and BN_nnmod().
6442 *) Function OCSP_request_verify(). This checks the signature on an
6443 OCSP request and verifies the signer certificate. The signer
6444 certificate is just checked for a generic purpose and OCSP request
6448 *) Add OCSP_check_validity() function to check the validity of OCSP
6449 responses. OCSP responses are prepared in real time and may only
6450 be a few seconds old. Simply checking that the current time lies
6451 between thisUpdate and nextUpdate max reject otherwise valid responses
6452 caused by either OCSP responder or client clock inaccuracy. Instead
6453 we allow thisUpdate and nextUpdate to fall within a certain period of
6454 the current time. The age of the response can also optionally be
6455 checked. Two new options -validity_period and -status_age added to
6459 *) If signature or public key algorithm is unrecognized print out its
6460 OID rather that just UNKNOWN.
6463 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6464 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6465 ID to be generated from the issuer certificate alone which can then be
6466 passed to OCSP_id_issuer_cmp().
6469 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6470 ASN1 modules to export functions returning ASN1_ITEM pointers
6471 instead of the ASN1_ITEM structures themselves. This adds several
6472 new macros which allow the underlying ASN1 function/structure to
6473 be accessed transparently. As a result code should not use ASN1_ITEM
6474 references directly (such as &X509_it) but instead use the relevant
6475 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6476 use of the new ASN1 code on platforms where exporting structures
6477 is problematical (for example in shared libraries) but exporting
6478 functions returning pointers to structures is not.
6481 *) Add support for overriding the generation of SSL/TLS session IDs.
6482 These callbacks can be registered either in an SSL_CTX or per SSL.
6483 The purpose of this is to allow applications to control, if they wish,
6484 the arbitrary values chosen for use as session IDs, particularly as it
6485 can be useful for session caching in multiple-server environments. A
6486 command-line switch for testing this (and any client code that wishes
6487 to use such a feature) has been added to "s_server".
6488 [Geoff Thorpe, Lutz Jaenicke]
6490 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6491 of the form '#if defined(...) || defined(...) || ...' and
6492 '#if !defined(...) && !defined(...) && ...'. This also avoids
6493 the growing number of special cases it was previously handling.
6496 *) Make all configuration macros available for application by making
6497 sure they are available in opensslconf.h, by giving them names starting
6498 with "OPENSSL_" to avoid conflicts with other packages and by making
6499 sure e_os2.h will cover all platform-specific cases together with
6501 Additionally, it is now possible to define configuration/platform-
6502 specific names (called "system identities"). In the C code, these
6503 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6504 macro with the name beginning with "OPENSSL_SYS_", which is determined
6505 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6509 *) New option -set_serial to 'req' and 'x509' this allows the serial
6510 number to use to be specified on the command line. Previously self
6511 signed certificates were hard coded with serial number 0 and the
6512 CA options of 'x509' had to use a serial number in a file which was
6516 *) New options to 'ca' utility to support V2 CRL entry extensions.
6517 Currently CRL reason, invalidity date and hold instruction are
6518 supported. Add new CRL extensions to V3 code and some new objects.
6521 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6522 disable standard block padding (aka PKCS#5 padding) in the EVP
6523 API, which was previously mandatory. This means that the data is
6524 not padded in any way and so the total length much be a multiple
6525 of the block size, otherwise an error occurs.
6528 *) Initial (incomplete) OCSP SSL support.
6531 *) New function OCSP_parse_url(). This splits up a URL into its host,
6532 port and path components: primarily to parse OCSP URLs. New -url
6533 option to ocsp utility.
6536 *) New nonce behavior. The return value of OCSP_check_nonce() now
6537 reflects the various checks performed. Applications can decide
6538 whether to tolerate certain situations such as an absent nonce
6539 in a response when one was present in a request: the ocsp application
6540 just prints out a warning. New function OCSP_add1_basic_nonce()
6541 this is to allow responders to include a nonce in a response even if
6542 the request is nonce-less.
6545 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6546 skipped when using openssl x509 multiple times on a single input file,
6547 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6550 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6551 set string type: to handle setting ASN1_TIME structures. Fix ca
6552 utility to correctly initialize revocation date of CRLs.
6555 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6556 the clients preferred ciphersuites and rather use its own preferences.
6557 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6558 Internet Explorer by ensuring unchanged hash method during stepup.
6559 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6562 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6563 to aes and add a new 'exist' option to print out symbols that don't
6567 *) Additional options to ocsp utility to allow flags to be set and
6568 additional certificates supplied.
6571 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6572 OCSP client a number of certificate to only verify the response
6576 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6577 handle the new API. Currently only ECB, CBC modes supported. Add new
6580 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6581 Encryption Standard (AES) Ciphersuites for Transport Layer
6582 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6583 not enabled by default and were not part of the "ALL" ciphersuite
6584 alias because they were not yet official; they could be
6585 explicitly requested by specifying the "AESdraft" ciphersuite
6586 group alias. In the final release of OpenSSL 0.9.7, the group
6587 alias is called "AES" and is part of "ALL".)
6588 [Ben Laurie, Steve Henson, Bodo Moeller]
6590 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6591 request to response.
6594 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6595 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6596 extract information from a certificate request. OCSP_response_create()
6597 creates a response and optionally adds a basic response structure.
6598 OCSP_basic_add1_status() adds a complete single response to a basic
6599 response and returns the OCSP_SINGLERESP structure just added (to allow
6600 extensions to be included for example). OCSP_basic_add1_cert() adds a
6601 certificate to a basic response and OCSP_basic_sign() signs a basic
6602 response with various flags. New helper functions ASN1_TIME_check()
6603 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6604 (converts ASN1_TIME to GeneralizedTime).
6607 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6608 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6609 structure from a certificate. X509_pubkey_digest() digests the public_key
6610 contents: this is used in various key identifiers.
6613 *) Make sk_sort() tolerate a NULL argument.
6614 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6616 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6617 passed by the function are trusted implicitly. If any of them signed the
6618 response then it is assumed to be valid and is not verified.
6621 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6622 to data. This was previously part of the PKCS7 ASN1 code. This
6623 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6624 [Steve Henson, reported by Kenneth R. Robinette
6625 <support@securenetterm.com>]
6627 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6628 routines: without these tracing memory leaks is very painful.
6629 Fix leaks in PKCS12 and PKCS7 routines.
6632 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6633 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6634 effectively meant GeneralizedTime would never be used. Now it
6635 is initialised to -1 but X509_time_adj() now has to check the value
6636 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6637 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6638 [Steve Henson, reported by Kenneth R. Robinette
6639 <support@securenetterm.com>]
6641 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6642 result in a zero length in the ASN1_INTEGER structure which was
6643 not consistent with the structure when d2i_ASN1_INTEGER() was used
6644 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6645 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6646 where it did not print out a minus for negative ASN1_INTEGER.
6649 *) Add summary printout to ocsp utility. The various functions which
6650 convert status values to strings have been renamed to:
6651 OCSP_response_status_str(), OCSP_cert_status_str() and
6652 OCSP_crl_reason_str() and are no longer static. New options
6653 to verify nonce values and to disable verification. OCSP response
6654 printout format cleaned up.
6657 *) Add additional OCSP certificate checks. These are those specified
6658 in RFC2560. This consists of two separate checks: the CA of the
6659 certificate being checked must either be the OCSP signer certificate
6660 or the issuer of the OCSP signer certificate. In the latter case the
6661 OCSP signer certificate must contain the OCSP signing extended key
6662 usage. This check is performed by attempting to match the OCSP
6663 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6664 in the OCSP_CERTID structures of the response.
6667 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6668 and related routines. This uses the standard OpenSSL certificate
6669 verify routines to perform initial checks (just CA validity) and
6670 to obtain the certificate chain. Then additional checks will be
6671 performed on the chain. Currently the root CA is checked to see
6672 if it is explicitly trusted for OCSP signing. This is used to set
6673 a root CA as a global signing root: that is any certificate that
6674 chains to that CA is an acceptable OCSP signing certificate.
6677 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6678 extensions from a separate configuration file.
6679 As when reading extensions from the main configuration file,
6680 the '-extensions ...' option may be used for specifying the
6682 [Massimiliano Pala <madwolf@comune.modena.it>]
6684 *) New OCSP utility. Allows OCSP requests to be generated or
6685 read. The request can be sent to a responder and the output
6686 parsed, outputed or printed in text form. Not complete yet:
6687 still needs to check the OCSP response validity.
6690 *) New subcommands for 'openssl ca':
6691 'openssl ca -status <serial>' prints the status of the cert with
6692 the given serial number (according to the index file).
6693 'openssl ca -updatedb' updates the expiry status of certificates
6695 [Massimiliano Pala <madwolf@comune.modena.it>]
6697 *) New '-newreq-nodes' command option to CA.pl. This is like
6698 '-newreq', but calls 'openssl req' with the '-nodes' option
6699 so that the resulting key is not encrypted.
6700 [Damien Miller <djm@mindrot.org>]
6702 *) New configuration for the GNU Hurd.
6703 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6705 *) Initial code to implement OCSP basic response verify. This
6706 is currently incomplete. Currently just finds the signer's
6707 certificate and verifies the signature on the response.
6710 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6711 value of OPENSSLDIR. This is available via the new '-d' option
6712 to 'openssl version', and is also included in 'openssl version -a'.
6715 *) Allowing defining memory allocation callbacks that will be given
6716 file name and line number information in additional arguments
6717 (a const char* and an int). The basic functionality remains, as
6718 well as the original possibility to just replace malloc(),
6719 realloc() and free() by functions that do not know about these
6720 additional arguments. To register and find out the current
6721 settings for extended allocation functions, the following
6722 functions are provided:
6724 CRYPTO_set_mem_ex_functions
6725 CRYPTO_set_locked_mem_ex_functions
6726 CRYPTO_get_mem_ex_functions
6727 CRYPTO_get_locked_mem_ex_functions
6729 These work the same way as CRYPTO_set_mem_functions and friends.
6730 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6731 extended allocation function is enabled.
6732 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6733 a conventional allocation function is enabled.
6734 [Richard Levitte, Bodo Moeller]
6736 *) Finish off removing the remaining LHASH function pointer casts.
6737 There should no longer be any prototype-casting required when using
6738 the LHASH abstraction, and any casts that remain are "bugs". See
6739 the callback types and macros at the head of lhash.h for details
6740 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6743 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6744 If /dev/[u]random devices are not available or do not return enough
6745 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6747 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6748 /etc/entropy will be queried once each in this sequence, quering stops
6749 when enough entropy was collected without querying more sockets.
6752 *) Change the Unix RAND_poll() variant to be able to poll several
6753 random devices, as specified by DEVRANDOM, until a sufficient amount
6754 of data has been collected. We spend at most 10 ms on each file
6755 (select timeout) and read in non-blocking mode. DEVRANDOM now
6756 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6757 (previously it was just the string "/dev/urandom"), so on typical
6758 platforms the 10 ms delay will never occur.
6759 Also separate out the Unix variant to its own file, rand_unix.c.
6760 For VMS, there's a currently-empty rand_vms.c.
6763 *) Move OCSP client related routines to ocsp_cl.c. These
6764 provide utility functions which an application needing
6765 to issue a request to an OCSP responder and analyse the
6766 response will typically need: as opposed to those which an
6767 OCSP responder itself would need which will be added later.
6769 OCSP_request_sign() signs an OCSP request with an API similar
6770 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6771 response. OCSP_response_get1_basic() extracts basic response
6772 from response. OCSP_resp_find_status(): finds and extracts status
6773 information from an OCSP_CERTID structure (which will be created
6774 when the request structure is built). These are built from lower
6775 level functions which work on OCSP_SINGLERESP structures but
6776 wont normally be used unless the application wishes to examine
6777 extensions in the OCSP response for example.
6779 Replace nonce routines with a pair of functions.
6780 OCSP_request_add1_nonce() adds a nonce value and optionally
6781 generates a random value. OCSP_check_nonce() checks the
6782 validity of the nonce in an OCSP response.
6785 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6786 This doesn't copy the supplied OCSP_CERTID and avoids the
6787 need to free up the newly created id. Change return type
6788 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6789 This can then be used to add extensions to the request.
6790 Deleted OCSP_request_new(), since most of its functionality
6791 is now in OCSP_REQUEST_new() (and the case insensitive name
6792 clash) apart from the ability to set the request name which
6793 will be added elsewhere.
6796 *) Update OCSP API. Remove obsolete extensions argument from
6797 various functions. Extensions are now handled using the new
6798 OCSP extension code. New simple OCSP HTTP function which
6799 can be used to send requests and parse the response.
6802 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6803 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6804 uses the special reorder version of SET OF to sort the attributes
6805 and reorder them to match the encoded order. This resolves a long
6806 standing problem: a verify on a PKCS7 structure just after signing
6807 it used to fail because the attribute order did not match the
6808 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6809 it uses the received order. This is necessary to tolerate some broken
6810 software that does not order SET OF. This is handled by encoding
6811 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6812 to produce the required SET OF.
6815 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6816 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6817 files to get correct declarations of the ASN.1 item variables.
6820 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6821 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6822 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6823 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6824 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6825 ASN1_ITEM and no wrapper functions.
6828 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6829 replace the old function pointer based I/O routines. Change most of
6830 the *_d2i_bio() and *_d2i_fp() functions to use these.
6833 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6834 lines, recognice more "algorithms" that can be deselected, and make
6835 it complain about algorithm deselection that isn't recognised.
6838 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6839 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6840 to use new functions. Add NO_ASN1_OLD which can be set to remove
6841 some old style ASN1 functions: this can be used to determine if old
6842 code will still work when these eventually go away.
6845 *) New extension functions for OCSP structures, these follow the
6846 same conventions as certificates and CRLs.
6849 *) New function X509V3_add1_i2d(). This automatically encodes and
6850 adds an extension. Its behaviour can be customised with various
6851 flags to append, replace or delete. Various wrappers added for
6852 certifcates and CRLs.
6855 *) Fix to avoid calling the underlying ASN1 print routine when
6856 an extension cannot be parsed. Correct a typo in the
6857 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6860 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
6861 entries for variables.
6864 *) Add functionality to apps/openssl.c for detecting locking
6865 problems: As the program is single-threaded, all we have
6866 to do is register a locking callback using an array for
6867 storing which locks are currently held by the program.
6870 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6871 SSL_get_ex_data_X509_STORE_idx(), which is used in
6872 ssl_verify_cert_chain() and thus can be called at any time
6873 during TLS/SSL handshakes so that thread-safety is essential.
6874 Unfortunately, the ex_data design is not at all suited
6875 for multi-threaded use, so it probably should be abolished.
6878 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6879 [Broadcom, tweaked and integrated by Geoff Thorpe]
6881 *) Move common extension printing code to new function
6882 X509V3_print_extensions(). Reorganise OCSP print routines and
6883 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6886 *) New function X509_signature_print() to remove duplication in some
6890 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6891 set (this was treated exactly the same as SET OF previously). This
6892 is used to reorder the STACK representing the structure to match the
6893 encoding. This will be used to get round a problem where a PKCS7
6894 structure which was signed could not be verified because the STACK
6895 order did not reflect the encoded order.
6898 *) Reimplement the OCSP ASN1 module using the new code.
6901 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6902 for its ASN1 operations. The old style function pointers still exist
6903 for now but they will eventually go away.
6906 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6907 completely replaces the old ASN1 functionality with a table driven
6908 encoder and decoder which interprets an ASN1_ITEM structure describing
6909 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6910 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6911 has also been converted to the new form.
6914 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6915 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6916 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6917 for negative moduli.
6920 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6921 of not touching the result's sign bit.
6924 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6928 *) Changed the LHASH code to use prototypes for callbacks, and created
6929 macros to declare and implement thin (optionally static) functions
6930 that provide type-safety and avoid function pointer casting for the
6931 type-specific callbacks.
6934 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6936 [Veers Staats <staatsvr@asc.hpc.mil>,
6937 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6939 *) Reformat the FAQ so the different questions and answers can be divided
6940 in sections depending on the subject.
6943 *) Have the zlib compression code load ZLIB.DLL dynamically under
6947 *) New function BN_mod_sqrt for computing square roots modulo a prime
6948 (using the probabilistic Tonelli-Shanks algorithm unless
6949 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6950 be handled deterministically).
6951 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6953 *) Make BN_mod_inverse faster by explicitly handling small quotients
6954 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6955 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6958 *) New function BN_kronecker.
6961 *) Fix BN_gcd so that it works on negative inputs; the result is
6962 positive unless both parameters are zero.
6963 Previously something reasonably close to an infinite loop was
6964 possible because numbers could be growing instead of shrinking
6965 in the implementation of Euclid's algorithm.
6968 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6969 sign of the number in question.
6971 Fix BN_is_word(a,w) to work correctly for w == 0.
6973 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6974 because its test if the absolute value of 'a' equals 'w'.
6975 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6976 it exists mostly for use in the implementations of BN_is_zero(),
6977 BN_is_one(), and BN_is_word().
6980 *) New function BN_swap.
6983 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6984 the exponentiation functions are more likely to produce reasonable
6985 results on negative inputs.
6988 *) Change BN_mod_mul so that the result is always non-negative.
6989 Previously, it could be negative if one of the factors was negative;
6990 I don't think anyone really wanted that behaviour.
6993 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6994 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6995 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6996 and add new functions:
7005 BN_mod_lshift1_quick
7009 These functions always generate non-negative results.
7011 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
7012 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
7014 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
7015 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
7016 be reduced modulo m.
7017 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7020 The following entry accidentily appeared in the CHANGES file
7021 distributed with OpenSSL 0.9.7. The modifications described in
7022 it do *not* apply to OpenSSL 0.9.7.
7024 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
7025 was actually never needed) and in BN_mul(). The removal in BN_mul()
7026 required a small change in bn_mul_part_recursive() and the addition
7027 of the functions bn_cmp_part_words(), bn_sub_part_words() and
7028 bn_add_part_words(), which do the same thing as bn_cmp_words(),
7029 bn_sub_words() and bn_add_words() except they take arrays with
7034 *) In 'openssl passwd', verify passwords read from the terminal
7035 unless the '-salt' option is used (which usually means that
7036 verification would just waste user's time since the resulting
7037 hash is going to be compared with some given password hash)
7038 or the new '-noverify' option is used.
7040 This is an incompatible change, but it does not affect
7041 non-interactive use of 'openssl passwd' (passwords on the command
7042 line, '-stdin' option, '-in ...' option) and thus should not
7046 *) Remove all references to RSAref, since there's no more need for it.
7049 *) Make DSO load along a path given through an environment variable
7050 (SHLIB_PATH) with shl_load().
7053 *) Constify the ENGINE code as a result of BIGNUM constification.
7054 Also constify the RSA code and most things related to it. In a
7055 few places, most notable in the depth of the ASN.1 code, ugly
7056 casts back to non-const were required (to be solved at a later
7060 *) Make it so the openssl application has all engines loaded by default.
7063 *) Constify the BIGNUM routines a little more.
7066 *) Add the following functions:
7068 ENGINE_load_cswift()
7070 ENGINE_load_atalla()
7072 ENGINE_load_builtin_engines()
7074 That way, an application can itself choose if external engines that
7075 are built-in in OpenSSL shall ever be used or not. The benefit is
7076 that applications won't have to be linked with libdl or other dso
7077 libraries unless it's really needed.
7079 Changed 'openssl engine' to load all engines on demand.
7080 Changed the engine header files to avoid the duplication of some
7081 declarations (they differed!).
7084 *) 'openssl engine' can now list capabilities.
7087 *) Better error reporting in 'openssl engine'.
7090 *) Never call load_dh_param(NULL) in s_server.
7093 *) Add engine application. It can currently list engines by name and
7094 identity, and test if they are actually available.
7097 *) Improve RPM specification file by forcing symbolic linking and making
7098 sure the installed documentation is also owned by root.root.
7099 [Damien Miller <djm@mindrot.org>]
7101 *) Give the OpenSSL applications more possibilities to make use of
7102 keys (public as well as private) handled by engines.
7105 *) Add OCSP code that comes from CertCo.
7108 *) Add VMS support for the Rijndael code.
7111 *) Added untested support for Nuron crypto accelerator.
7114 *) Add support for external cryptographic devices. This code was
7115 previously distributed separately as the "engine" branch.
7116 [Geoff Thorpe, Richard Levitte]
7118 *) Rework the filename-translation in the DSO code. It is now possible to
7119 have far greater control over how a "name" is turned into a filename
7120 depending on the operating environment and any oddities about the
7121 different shared library filenames on each system.
7124 *) Support threads on FreeBSD-elf in Configure.
7127 *) Fix for SHA1 assembly problem with MASM: it produces
7128 warnings about corrupt line number information when assembling
7129 with debugging information. This is caused by the overlapping
7131 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
7134 NCONF_get_number() has no error checking at all. As a replacement,
7135 NCONF_get_number_e() is defined (_e for "error checking") and is
7136 promoted strongly. The old NCONF_get_number is kept around for
7137 binary backward compatibility.
7138 Make it possible for methods to load from something other than a BIO,
7139 by providing a function pointer that is given a name instead of a BIO.
7140 For example, this could be used to load configuration data from an
7144 *) Fix for non blocking accept BIOs. Added new I/O special reason
7145 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
7146 with non blocking I/O was not possible because no retry code was
7147 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
7151 *) Added the beginnings of Rijndael support.
7154 *) Fix for bug in DirectoryString mask setting. Add support for
7155 X509_NAME_print_ex() in 'req' and X509_print_ex() function
7156 to allow certificate printing to more controllable, additional
7157 'certopt' option to 'x509' to allow new printing options to be
7161 *) Clean old EAY MD5 hack from e_os.h.
7164 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
7166 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7167 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7168 [Joe Orton, Steve Henson]
7170 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
7172 *) Fix additional bug revealed by the NISCC test suite:
7174 Stop bug triggering large recursion when presented with
7175 certain ASN.1 tags (CVE-2003-0851)
7178 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
7180 *) Fix various bugs revealed by running the NISCC test suite:
7182 Stop out of bounds reads in the ASN1 code when presented with
7183 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7185 If verify callback ignores invalid public key errors don't try to check
7186 certificate signature with the NULL public key.
7190 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7191 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7195 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7196 extra data after the compression methods not only for TLS 1.0
7197 but also for SSL 3.0 (as required by the specification).
7198 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7200 *) Change X509_certificate_type() to mark the key as exported/exportable
7201 when it's 512 *bits* long, not 512 bytes.
7204 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
7206 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7207 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7208 a protocol version number mismatch like a decryption error
7209 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7212 *) Turn on RSA blinding by default in the default implementation
7213 to avoid a timing attack. Applications that don't want it can call
7214 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7215 They would be ill-advised to do so in most cases.
7216 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7218 *) Change RSA blinding code so that it works when the PRNG is not
7219 seeded (in this case, the secret RSA exponent is abused as
7220 an unpredictable seed -- if it is not unpredictable, there
7221 is no point in blinding anyway). Make RSA blinding thread-safe
7222 by remembering the creator's thread ID in rsa->blinding and
7223 having all other threads use local one-time blinding factors
7224 (this requires more computation than sharing rsa->blinding, but
7225 avoids excessive locking; and if an RSA object is not shared
7226 between threads, blinding will still be very fast).
7229 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7231 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7232 via timing by performing a MAC computation even if incorrrect
7233 block cipher padding has been found. This is a countermeasure
7234 against active attacks where the attacker has to distinguish
7235 between bad padding and a MAC verification error. (CVE-2003-0078)
7237 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7238 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7239 Martin Vuagnoux (EPFL, Ilion)]
7241 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7243 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7244 memory from it's contents. This is done with a counter that will
7245 place alternating values in each byte. This can be used to solve
7246 two issues: 1) the removal of calls to memset() by highly optimizing
7247 compilers, and 2) cleansing with other values than 0, since those can
7248 be read through on certain media, for example a swap space on disk.
7251 *) Bugfix: client side session caching did not work with external caching,
7252 because the session->cipher setting was not restored when reloading
7253 from the external cache. This problem was masked, when
7254 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7255 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7258 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7259 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7260 [Zeev Lieber <zeev-l@yahoo.com>]
7262 *) Undo an undocumented change introduced in 0.9.6e which caused
7263 repeated calls to OpenSSL_add_all_ciphers() and
7264 OpenSSL_add_all_digests() to be ignored, even after calling
7268 *) Change the default configuration reader to deal with last line not
7269 being properly terminated.
7272 *) Change X509_NAME_cmp() so it applies the special rules on handling
7273 DN values that are of type PrintableString, as well as RDNs of type
7274 emailAddress where the value has the type ia5String.
7275 [stefank@valicert.com via Richard Levitte]
7277 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7278 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7279 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7280 the bitwise-OR of the two for use by the majority of applications
7281 wanting this behaviour, and update the docs. The documented
7282 behaviour and actual behaviour were inconsistent and had been
7283 changing anyway, so this is more a bug-fix than a behavioural
7285 [Geoff Thorpe, diagnosed by Nadav Har'El]
7287 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7288 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7291 *) Fix initialization code race conditions in
7292 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7293 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7294 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7295 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7296 ssl2_get_cipher_by_char(),
7297 ssl3_get_cipher_by_char().
7298 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7300 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7301 the cached sessions are flushed, as the remove_cb() might use ex_data
7302 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7303 (see [openssl.org #212]).
7304 [Geoff Thorpe, Lutz Jaenicke]
7306 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7307 length, instead of the encoding length to d2i_ASN1_OBJECT.
7310 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7312 *) [In 0.9.6g-engine release:]
7313 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7314 [Lynn Gazis <lgazis@rainbow.com>]
7316 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7318 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7319 and get fix the header length calculation.
7320 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7321 Alon Kantor <alonk@checkpoint.com> (and others),
7324 *) Use proper error handling instead of 'assertions' in buffer
7325 overflow checks added in 0.9.6e. This prevents DoS (the
7326 assertions could call abort()).
7327 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7329 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7331 *) Add various sanity checks to asn1_get_length() to reject
7332 the ASN1 length bytes if they exceed sizeof(long), will appear
7333 negative or the content length exceeds the length of the
7335 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7337 *) Fix cipher selection routines: ciphers without encryption had no flags
7338 for the cipher strength set and where therefore not handled correctly
7339 by the selection routines (PR #130).
7342 *) Fix EVP_dsa_sha macro.
7346 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7347 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7348 that was added in OpenSSL 0.9.6d.
7350 As the countermeasure turned out to be incompatible with some
7351 broken SSL implementations, the new option is part of SSL_OP_ALL.
7352 SSL_OP_ALL is usually employed when compatibility with weird SSL
7353 implementations is desired (e.g. '-bugs' option to 's_client' and
7354 's_server'), so the new option is automatically set in many
7358 *) Changes in security patch:
7360 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7361 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7362 Air Force Materiel Command, USAF, under agreement number
7365 *) Add various sanity checks to asn1_get_length() to reject
7366 the ASN1 length bytes if they exceed sizeof(long), will appear
7367 negative or the content length exceeds the length of the
7368 supplied buffer. (CVE-2002-0659)
7369 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7371 *) Assertions for various potential buffer overflows, not known to
7373 [Ben Laurie (CHATS)]
7375 *) Various temporary buffers to hold ASCII versions of integers were
7376 too small for 64 bit platforms. (CVE-2002-0655)
7377 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7379 *) Remote buffer overflow in SSL3 protocol - an attacker could
7380 supply an oversized session ID to a client. (CVE-2002-0656)
7381 [Ben Laurie (CHATS)]
7383 *) Remote buffer overflow in SSL2 protocol - an attacker could
7384 supply an oversized client master key. (CVE-2002-0656)
7385 [Ben Laurie (CHATS)]
7387 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7389 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7390 encoded as NULL) with id-dsa-with-sha1.
7391 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7393 *) Check various X509_...() return values in apps/req.c.
7394 [Nils Larsch <nla@trustcenter.de>]
7396 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7397 an end-of-file condition would erronously be flagged, when the CRLF
7398 was just at the end of a processed block. The bug was discovered when
7399 processing data through a buffering memory BIO handing the data to a
7400 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7401 <ptsekov@syntrex.com> and Nedelcho Stanev.
7404 *) Implement a countermeasure against a vulnerability recently found
7405 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7406 before application data chunks to avoid the use of known IVs
7407 with data potentially chosen by the attacker.
7410 *) Fix length checks in ssl3_get_client_hello().
7413 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7414 to prevent ssl3_read_internal() from incorrectly assuming that
7415 ssl3_read_bytes() found application data while handshake
7416 processing was enabled when in fact s->s3->in_read_app_data was
7417 merely automatically cleared during the initial handshake.
7418 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7420 *) Fix object definitions for Private and Enterprise: they were not
7421 recognized in their shortname (=lowercase) representation. Extend
7422 obj_dat.pl to issue an error when using undefined keywords instead
7423 of silently ignoring the problem (Svenning Sorensen
7424 <sss@sss.dnsalias.net>).
7427 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7428 generators, i.e. generators other than 2 and 5. (Previously, the
7429 code did not properly initialise the 'add' and 'rem' values to
7430 BN_generate_prime().)
7432 In the new general case, we do not insist that 'generator' is
7433 actually a primitive root: This requirement is rather pointless;
7434 a generator of the order-q subgroup is just as good, if not
7438 *) Map new X509 verification errors to alerts. Discovered and submitted by
7439 Tom Wu <tom@arcot.com>.
7442 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7443 returning non-zero before the data has been completely received
7444 when using non-blocking I/O.
7445 [Bodo Moeller; problem pointed out by John Hughes]
7447 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7448 [Ben Laurie, Lutz Jaenicke]
7450 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7451 Yoram Zahavi <YoramZ@gilian.com>).
7454 *) Add information about CygWin 1.3 and on, and preserve proper
7455 configuration for the versions before that.
7456 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7458 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7459 check whether we deal with a copy of a session and do not delete from
7460 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7461 <izhar@checkpoint.com>.
7464 *) Do not store session data into the internal session cache, if it
7465 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7466 flag is set). Proposed by Aslam <aslam@funk.com>.
7469 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7473 *) [In 0.9.6d-engine release:]
7474 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7475 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7477 *) Add the configuration target linux-s390x.
7478 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7480 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7481 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7482 variable as an indication that a ClientHello message has been
7483 received. As the flag value will be lost between multiple
7484 invocations of ssl3_accept when using non-blocking I/O, the
7485 function may not be aware that a handshake has actually taken
7486 place, thus preventing a new session from being added to the
7489 To avoid this problem, we now set s->new_session to 2 instead of
7490 using a local variable.
7491 [Lutz Jaenicke, Bodo Moeller]
7493 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7494 if the SSL_R_LENGTH_MISMATCH error is detected.
7495 [Geoff Thorpe, Bodo Moeller]
7497 *) New 'shared_ldflag' column in Configure platform table.
7500 *) Fix EVP_CIPHER_mode macro.
7501 ["Dan S. Camper" <dan@bti.net>]
7503 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7504 type, we must throw them away by setting rr->length to 0.
7505 [D P Chang <dpc@qualys.com>]
7507 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7509 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7510 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7511 worked incorrectly for those cases where range = 10..._2 and
7512 3*range is two bits longer than range.)
7515 *) Only add signing time to PKCS7 structures if it is not already
7519 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7520 OBJ_ld_ce should be OBJ_id_ce.
7521 Also some ip-pda OIDs in crypto/objects/objects.txt were
7522 incorrect (cf. RFC 3039).
7523 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7525 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7526 returns early because it has nothing to do.
7527 [Andy Schneider <andy.schneider@bjss.co.uk>]
7529 *) [In 0.9.6c-engine release:]
7530 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7531 [Andy Schneider <andy.schneider@bjss.co.uk>]
7533 *) [In 0.9.6c-engine release:]
7534 Add support for Cryptographic Appliance's keyserver technology.
7535 (Use engine 'keyclient')
7536 [Cryptographic Appliances and Geoff Thorpe]
7538 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7539 is called via tools/c89.sh because arguments have to be
7540 rearranged (all '-L' options must appear before the first object
7542 [Richard Shapiro <rshapiro@abinitio.com>]
7544 *) [In 0.9.6c-engine release:]
7545 Add support for Broadcom crypto accelerator cards, backported
7547 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7549 *) [In 0.9.6c-engine release:]
7550 Add support for SureWare crypto accelerator cards from
7551 Baltimore Technologies. (Use engine 'sureware')
7552 [Baltimore Technologies and Mark Cox]
7554 *) [In 0.9.6c-engine release:]
7555 Add support for crypto accelerator cards from Accelerated
7556 Encryption Processing, www.aep.ie. (Use engine 'aep')
7557 [AEP Inc. and Mark Cox]
7559 *) Add a configuration entry for gcc on UnixWare.
7560 [Gary Benson <gbenson@redhat.com>]
7562 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7563 messages are stored in a single piece (fixed-length part and
7564 variable-length part combined) and fix various bugs found on the way.
7567 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7568 instead. BIO_gethostbyname() does not know what timeouts are
7569 appropriate, so entries would stay in cache even when they have
7571 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7573 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7574 faced with a pathologically small ClientHello fragment that does
7575 not contain client_version: Instead of aborting with an error,
7576 simply choose the highest available protocol version (i.e.,
7577 TLS 1.0 unless it is disabled). In practice, ClientHello
7578 messages are never sent like this, but this change gives us
7579 strictly correct behaviour at least for TLS.
7582 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7583 never resets s->method to s->ctx->method when called from within
7584 one of the SSL handshake functions.
7585 [Bodo Moeller; problem pointed out by Niko Baric]
7587 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7588 (sent using the client's version number) if client_version is
7589 smaller than the protocol version in use. Also change
7590 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7591 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7592 the client will at least see that alert.
7595 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7599 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7600 client receives HelloRequest while in a handshake.
7601 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7603 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7604 should end in 'break', not 'goto end' which circuments various
7605 cleanups done in state SSL_ST_OK. But session related stuff
7606 must be disabled for SSL_ST_OK in the case that we just sent a
7609 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7610 before just sending a HelloRequest.
7611 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7613 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7614 reveal whether illegal block cipher padding was found or a MAC
7615 verification error occured. (Neither SSLerr() codes nor alerts
7616 are directly visible to potential attackers, but the information
7617 may leak via logfiles.)
7619 Similar changes are not required for the SSL 2.0 implementation
7620 because the number of padding bytes is sent in clear for SSL 2.0,
7621 and the extra bytes are just ignored. However ssl/s2_pkt.c
7622 failed to verify that the purported number of padding bytes is in
7626 *) Add OpenUNIX-8 support including shared libraries
7627 (Boyd Lynn Gerber <gerberb@zenez.com>).
7630 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7631 'wristwatch attack' using huge encoding parameters (cf.
7632 James H. Manger's CRYPTO 2001 paper). Note that the
7633 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7634 encoding parameters and hence was not vulnerable.
7637 *) BN_sqr() bug fix.
7638 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7640 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7641 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7642 followed by modular reduction.
7643 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7645 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7646 equivalent based on BN_pseudo_rand() instead of BN_rand().
7649 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7650 This function was broken, as the check for a new client hello message
7651 to handle SGC did not allow these large messages.
7652 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7655 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7658 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7659 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7662 *) Rework the configuration and shared library support for Tru64 Unix.
7663 The configuration part makes use of modern compiler features and
7664 still retains old compiler behavior for those that run older versions
7665 of the OS. The shared library support part includes a variant that
7666 uses the RPATH feature, and is available through the special
7667 configuration target "alpha-cc-rpath", which will never be selected
7669 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7671 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7672 with the same message size as in ssl3_get_certificate_request().
7673 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7674 messages might inadvertently be reject as too long.
7675 [Petr Lampa <lampa@fee.vutbr.cz>]
7677 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7680 *) Modified SSL library such that the verify_callback that has been set
7681 specificly for an SSL object with SSL_set_verify() is actually being
7682 used. Before the change, a verify_callback set with this function was
7683 ignored and the verify_callback() set in the SSL_CTX at the time of
7684 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7685 to allow the necessary settings.
7688 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7689 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7690 done automatically (in contradiction to the requirements of the C
7691 standard). This made problems when used from OpenSSH.
7694 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7695 dh->length and always used
7697 BN_rand_range(priv_key, dh->p).
7699 BN_rand_range() is not necessary for Diffie-Hellman, and this
7700 specific range makes Diffie-Hellman unnecessarily inefficient if
7701 dh->length (recommended exponent length) is much smaller than the
7702 length of dh->p. We could use BN_rand_range() if the order of
7703 the subgroup was stored in the DH structure, but we only have
7708 BN_rand(priv_key, l, ...)
7710 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7716 RSA_eay_public_encrypt
7717 RSA_eay_private_decrypt
7718 RSA_eay_private_encrypt (signing)
7719 RSA_eay_public_decrypt (signature verification)
7721 (default implementations for RSA_public_encrypt,
7722 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7723 always reject numbers >= n.
7726 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7727 to synchronize access to 'locking_thread'. This is necessary on
7728 systems where access to 'locking_thread' (an 'unsigned long'
7729 variable) is not atomic.
7732 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7733 *before* setting the 'crypto_lock_rand' flag. The previous code had
7734 a race condition if 0 is a valid thread ID.
7735 [Travis Vitek <vitek@roguewave.com>]
7737 *) Add support for shared libraries under Irix.
7738 [Albert Chin-A-Young <china@thewrittenword.com>]
7740 *) Add configuration option to build on Linux on both big-endian and
7742 [Ralf Baechle <ralf@uni-koblenz.de>]
7744 *) Add the possibility to create shared libraries on HP-UX.
7747 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7749 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7750 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7751 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7752 PRNG state recovery was possible based on the output of
7753 one PRNG request appropriately sized to gain knowledge on
7754 'md' followed by enough consecutive 1-byte PRNG requests
7755 to traverse all of 'state'.
7757 1. When updating 'md_local' (the current thread's copy of 'md')
7758 during PRNG output generation, hash all of the previous
7759 'md_local' value, not just the half used for PRNG output.
7761 2. Make the number of bytes from 'state' included into the hash
7762 independent from the number of PRNG bytes requested.
7764 The first measure alone would be sufficient to avoid
7765 Markku-Juhani's attack. (Actually it had never occurred
7766 to me that the half of 'md_local' used for chaining was the
7767 half from which PRNG output bytes were taken -- I had always
7768 assumed that the secret half would be used.) The second
7769 measure makes sure that additional data from 'state' is never
7770 mixed into 'md_local' in small portions; this heuristically
7771 further strengthens the PRNG.
7774 *) Fix crypto/bn/asm/mips3.s.
7777 *) When only the key is given to "enc", the IV is undefined. Print out
7778 an error message in this case.
7781 *) Handle special case when X509_NAME is empty in X509 printing routines.
7784 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7785 positive and less than q.
7788 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7789 used: it isn't thread safe and the add_lock_callback should handle
7791 [Paul Rose <Paul.Rose@bridge.com>]
7793 *) Verify that incoming data obeys the block size in
7794 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7798 [Ulf Möller, Bodo Möller]
7800 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7801 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7802 when fixing the server behaviour for backwards-compatible 'client
7803 hello' messages. (Note that the attack is impractical against
7804 SSL 3.0 and TLS 1.0 anyway because length and version checking
7805 means that the probability of guessing a valid ciphertext is
7806 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7809 Before 0.9.5, the countermeasure (hide the error by generating a
7810 random 'decryption result') did not work properly because
7811 ERR_clear_error() was missing, meaning that SSL_get_error() would
7812 detect the supposedly ignored error.
7814 Both problems are now fixed.
7817 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7818 (previously it was 1024).
7821 *) Fix for compatibility mode trust settings: ignore trust settings
7822 unless some valid trust or reject settings are present.
7825 *) Fix for blowfish EVP: its a variable length cipher.
7828 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7829 parameters in DSA public key structures and return an error in the
7830 DSA routines if parameters are absent.
7833 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7834 in the current directory if neither $RANDFILE nor $HOME was set.
7835 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7836 caused some confusion to Windows users who haven't defined $HOME.
7837 Thus RAND_file_name() is changed again: e_os.h can define a
7838 DEFAULT_HOME, which will be used if $HOME is not set.
7839 For Windows, we use "C:"; on other platforms, we still require
7840 environment variables.
7842 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7843 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7844 having multiple threads call RAND_poll() concurrently.
7847 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7848 combination of a flag and a thread ID variable.
7849 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7850 flag), *other* threads can enter ssleay_add_bytes without obeying
7851 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7852 that they do not hold after the first thread unsets add_do_not_lock).
7855 *) Change bctest again: '-x' expressions are not available in all
7859 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7861 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7862 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7864 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7865 the default extension for executables, if any. Also, make the perl
7866 scripts that use symlink() to test if it really exists and use "cp"
7867 if it doesn't. All this made OpenSSL compilable and installable in
7871 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7872 If SEQUENCE is length is indefinite just set c->slen to the total
7873 amount of data available.
7874 [Steve Henson, reported by shige@FreeBSD.org]
7875 [This change does not apply to 0.9.7.]
7877 *) Change bctest to avoid here-documents inside command substitution
7878 (workaround for FreeBSD /bin/sh bug).
7879 For compatibility with Ultrix, avoid shell functions (introduced
7880 in the bctest version that searches along $PATH).
7883 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7884 with des_encrypt() defined on some operating systems, like Solaris
7888 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7889 On the Importance of Eliminating Errors in Cryptographic
7890 Computations, J. Cryptology 14 (2001) 2, 101-119,
7891 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7894 *) MIPS assembler BIGNUM division bug fix.
7897 *) Disabled incorrect Alpha assembler code.
7900 *) Fix PKCS#7 decode routines so they correctly update the length
7901 after reading an EOC for the EXPLICIT tag.
7903 [This change does not apply to 0.9.7.]
7905 *) Fix bug in PKCS#12 key generation routines. This was triggered
7906 if a 3DES key was generated with a 0 initial byte. Include
7907 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7908 (but broken) behaviour.
7911 *) Enhance bctest to search for a working bc along $PATH and print
7913 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7915 *) Fix memory leaks in err.c: free err_data string if necessary;
7916 don't write to the wrong index in ERR_set_error_data.
7919 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7923 *) Replace rdtsc with _emit statements for VC++ version 5.
7924 [Jeremy Cooper <jeremy@baymoo.org>]
7926 *) Make it possible to reuse SSLv2 sessions.
7929 *) In copy_email() check for >= 0 as a return value for
7930 X509_NAME_get_index_by_NID() since 0 is a valid index.
7931 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7933 *) Avoid coredump with unsupported or invalid public keys by checking if
7934 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7935 PKCS7_verify() fails with non detached data.
7938 *) Don't use getenv in library functions when run as setuid/setgid.
7939 New function OPENSSL_issetugid().
7942 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7943 due to incorrect handling of multi-threading:
7945 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7947 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7949 3. Count how many times MemCheck_off() has been called so that
7950 nested use can be treated correctly. This also avoids
7951 inband-signalling in the previous code (which relied on the
7952 assumption that thread ID 0 is impossible).
7955 *) Add "-rand" option also to s_client and s_server.
7958 *) Fix CPU detection on Irix 6.x.
7959 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7960 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7962 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7965 [This change does not apply to 0.9.7.]
7967 *) Use the cached encoding of an X509_NAME structure rather than
7968 copying it. This is apparently the reason for the libsafe "errors"
7969 but the code is actually correct.
7972 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7973 Bleichenbacher's DSA attack.
7974 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7975 to be set and top=0 forces the highest bit to be set; top=-1 is new
7976 and leaves the highest bit random.
7977 [Ulf Moeller, Bodo Moeller]
7979 *) In the NCONF_...-based implementations for CONF_... queries
7980 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7981 a temporary CONF structure with the data component set to NULL
7982 (which gives segmentation faults in lh_retrieve).
7983 Instead, use NULL for the CONF pointer in CONF_get_string and
7984 CONF_get_number (which may use environment variables) and directly
7985 return NULL from CONF_get_section.
7988 *) Fix potential buffer overrun for EBCDIC.
7991 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7992 keyUsage if basicConstraints absent for a CA.
7995 *) Make SMIME_write_PKCS7() write mail header values with a format that
7996 is more generally accepted (no spaces before the semicolon), since
7997 some programs can't parse those values properly otherwise. Also make
7998 sure BIO's that break lines after each write do not create invalid
8002 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
8003 macros previously used would not encode an empty SEQUENCE OF
8004 and break the signature.
8006 [This change does not apply to 0.9.7.]
8008 *) Zero the premaster secret after deriving the master secret in
8012 *) Add some EVP_add_digest_alias registrations (as found in
8013 OpenSSL_add_all_digests()) to SSL_library_init()
8014 aka OpenSSL_add_ssl_algorithms(). This provides improved
8015 compatibility with peers using X.509 certificates
8016 with unconventional AlgorithmIdentifier OIDs.
8019 *) Fix for Irix with NO_ASM.
8020 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8022 *) ./config script fixes.
8023 [Ulf Moeller, Richard Levitte]
8025 *) Fix 'openssl passwd -1'.
8028 *) Change PKCS12_key_gen_asc() so it can cope with non null
8029 terminated strings whose length is passed in the passlen
8030 parameter, for example from PEM callbacks. This was done
8031 by adding an extra length parameter to asc2uni().
8032 [Steve Henson, reported by <oddissey@samsung.co.kr>]
8034 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
8035 call failed, free the DSA structure.
8038 *) Fix to uni2asc() to cope with zero length Unicode strings.
8039 These are present in some PKCS#12 files.
8042 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
8043 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
8044 when writing a 32767 byte record.
8045 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
8047 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
8048 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
8050 (RSA objects have a reference count access to which is protected
8051 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
8052 so they are meant to be shared between threads.)
8053 [Bodo Moeller, Geoff Thorpe; original patch submitted by
8054 "Reddie, Steven" <Steven.Reddie@ca.com>]
8056 *) Fix a deadlock in CRYPTO_mem_leaks().
8059 *) Use better test patterns in bntest.
8062 *) rand_win.c fix for Borland C.
8065 *) BN_rshift bugfix for n == 0.
8068 *) Add a 'bctest' script that checks for some known 'bc' bugs
8069 so that 'make test' does not abort just because 'bc' is broken.
8072 *) Store verify_result within SSL_SESSION also for client side to
8073 avoid potential security hole. (Re-used sessions on the client side
8074 always resulted in verify_result==X509_V_OK, not using the original
8075 result of the server certificate verification.)
8078 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
8079 SSL3_RT_APPLICATION_DATA, return 0.
8080 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
8084 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
8085 releases, have been re-implemented by renaming the previous
8086 implementations of ssl2_read and ssl3_read to ssl2_read_internal
8087 and ssl3_read_internal, respectively, and adding 'peek' parameters
8088 to them. The new ssl[23]_{read,peek} functions are calls to
8089 ssl[23]_read_internal with the 'peek' flag set appropriately.
8090 A 'peek' parameter has also been added to ssl3_read_bytes, which
8091 does the actual work for ssl3_read_internal.
8094 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
8095 the method-specific "init()" handler. Also clean up ex_data after
8096 calling the method-specific "finish()" handler. Previously, this was
8097 happening the other way round.
8100 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
8101 The previous value, 12, was not always sufficient for BN_mod_exp().
8104 *) Make sure that shared libraries get the internal name engine with
8105 the full version number and not just 0. This should mark the
8106 shared libraries as not backward compatible. Of course, this should
8107 be changed again when we can guarantee backward binary compatibility.
8110 *) Fix typo in get_cert_by_subject() in by_dir.c
8111 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
8113 *) Rework the system to generate shared libraries:
8115 - Make note of the expected extension for the shared libraries and
8116 if there is a need for symbolic links from for example libcrypto.so.0
8117 to libcrypto.so.0.9.7. There is extended info in Configure for
8120 - Make as few rebuilds of the shared libraries as possible.
8122 - Still avoid linking the OpenSSL programs with the shared libraries.
8124 - When installing, install the shared libraries separately from the
8128 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
8130 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
8131 and not in SSL_clear because the latter is also used by the
8132 accept/connect functions; previously, the settings made by
8133 SSL_set_read_ahead would be lost during the handshake.
8134 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
8136 *) Correct util/mkdef.pl to be selective about disabled algorithms.
8137 Previously, it would create entries for disableed algorithms no
8141 *) Added several new manual pages for SSL_* function.
8144 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
8146 *) In ssl23_get_client_hello, generate an error message when faced
8147 with an initial SSL 3.0/TLS record that is too small to contain the
8148 first two bytes of the ClientHello message, i.e. client_version.
8149 (Note that this is a pathologic case that probably has never happened
8150 in real life.) The previous approach was to use the version number
8151 from the record header as a substitute; but our protocol choice
8152 should not depend on that one because it is not authenticated
8153 by the Finished messages.
8156 *) More robust randomness gathering functions for Windows.
8157 [Jeffrey Altman <jaltman@columbia.edu>]
8159 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
8160 not set then we don't setup the error code for issuer check errors
8161 to avoid possibly overwriting other errors which the callback does
8162 handle. If an application does set the flag then we assume it knows
8163 what it is doing and can handle the new informational codes
8167 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
8168 a general "ANY" type, as such it should be able to decode anything
8169 including tagged types. However it didn't check the class so it would
8170 wrongly interpret tagged types in the same way as their universal
8171 counterpart and unknown types were just rejected. Changed so that the
8172 tagged and unknown types are handled in the same way as a SEQUENCE:
8173 that is the encoding is stored intact. There is also a new type
8174 "V_ASN1_OTHER" which is used when the class is not universal, in this
8175 case we have no idea what the actual type is so we just lump them all
8179 *) On VMS, stdout may very well lead to a file that is written to
8180 in a record-oriented fashion. That means that every write() will
8181 write a separate record, which will be read separately by the
8182 programs trying to read from it. This can be very confusing.
8184 The solution is to put a BIO filter in the way that will buffer
8185 text until a linefeed is reached, and then write everything a
8186 line at a time, so every record written will be an actual line,
8187 not chunks of lines and not (usually doesn't happen, but I've
8188 seen it once) several lines in one record. BIO_f_linebuffer() is
8191 Currently, it's a VMS-only method, because that's where it has
8192 been tested well enough.
8195 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
8196 it can return incorrect results.
8197 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
8198 but it was in 0.9.6-beta[12].)
8201 *) Disable the check for content being present when verifying detached
8202 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
8203 include zero length content when signing messages.
8206 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
8207 BIO_ctrl (for BIO pairs).
8210 *) Add DSO method for VMS.
8213 *) Bug fix: Montgomery multiplication could produce results with the
8217 *) Add RPM specification openssl.spec and modify it to build three
8218 packages. The default package contains applications, application
8219 documentation and run-time libraries. The devel package contains
8220 include files, static libraries and function documentation. The
8221 doc package contains the contents of the doc directory. The original
8222 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8225 *) Add a large number of documentation files for many SSL routines.
8226 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8228 *) Add a configuration entry for Sony News 4.
8229 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8231 *) Don't set the two most significant bits to one when generating a
8232 random number < q in the DSA library.
8235 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8236 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8237 the underlying transport is blocking) if a handshake took place.
8238 (The default behaviour is needed by applications such as s_client
8239 and s_server that use select() to determine when to use SSL_read;
8240 but for applications that know in advance when to expect data, it
8241 just makes things more complicated.)
8244 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8248 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8249 work better on such systems.
8250 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8252 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8253 Update PKCS12_parse() so it copies the friendlyName and the
8254 keyid to the certificates aux info.
8257 *) Fix bug in PKCS7_verify() which caused an infinite loop
8258 if there was more than one signature.
8259 [Sven Uszpelkat <su@celocom.de>]
8261 *) Major change in util/mkdef.pl to include extra information
8262 about each symbol, as well as presentig variables as well
8263 as functions. This change means that there's n more need
8264 to rebuild the .num files when some algorithms are excluded.
8267 *) Allow the verify time to be set by an application,
8268 rather than always using the current time.
8271 *) Phase 2 verify code reorganisation. The certificate
8272 verify code now looks up an issuer certificate by a
8273 number of criteria: subject name, authority key id
8274 and key usage. It also verifies self signed certificates
8275 by the same criteria. The main comparison function is
8276 X509_check_issued() which performs these checks.
8278 Lot of changes were necessary in order to support this
8279 without completely rewriting the lookup code.
8281 Authority and subject key identifier are now cached.
8283 The LHASH 'certs' is X509_STORE has now been replaced
8284 by a STACK_OF(X509_OBJECT). This is mainly because an
8285 LHASH can't store or retrieve multiple objects with
8286 the same hash value.
8288 As a result various functions (which were all internal
8289 use only) have changed to handle the new X509_STORE
8290 structure. This will break anything that messed round
8291 with X509_STORE internally.
8293 The functions X509_STORE_add_cert() now checks for an
8294 exact match, rather than just subject name.
8296 The X509_STORE API doesn't directly support the retrieval
8297 of multiple certificates matching a given criteria, however
8298 this can be worked round by performing a lookup first
8299 (which will fill the cache with candidate certificates)
8300 and then examining the cache for matches. This is probably
8301 the best we can do without throwing out X509_LOOKUP
8302 entirely (maybe later...).
8304 The X509_VERIFY_CTX structure has been enhanced considerably.
8306 All certificate lookup operations now go via a get_issuer()
8307 callback. Although this currently uses an X509_STORE it
8308 can be replaced by custom lookups. This is a simple way
8309 to bypass the X509_STORE hackery necessary to make this
8310 work and makes it possible to use more efficient techniques
8311 in future. A very simple version which uses a simple
8312 STACK for its trusted certificate store is also provided
8313 using X509_STORE_CTX_trusted_stack().
8315 The verify_cb() and verify() callbacks now have equivalents
8316 in the X509_STORE_CTX structure.
8318 X509_STORE_CTX also has a 'flags' field which can be used
8319 to customise the verify behaviour.
8322 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8323 excludes S/MIME capabilities.
8326 *) When a certificate request is read in keep a copy of the
8327 original encoding of the signed data and use it when outputing
8328 again. Signatures then use the original encoding rather than
8329 a decoded, encoded version which may cause problems if the
8330 request is improperly encoded.
8333 *) For consistency with other BIO_puts implementations, call
8334 buffer_write(b, ...) directly in buffer_puts instead of calling
8337 In BIO_puts, increment b->num_write as in BIO_write.
8338 [Peter.Sylvester@EdelWeb.fr]
8340 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8341 BN_zero, we may not return a BIGNUM with an array consisting of
8345 *) Avoid calling abort() from within the library when problems are
8346 detected, except if preprocessor symbols have been defined
8347 (such as REF_CHECK, BN_DEBUG etc.).
8350 *) New openssl application 'rsautl'. This utility can be
8351 used for low level RSA operations. DER public key
8352 BIO/fp routines also added.
8355 *) New Configure entry and patches for compiling on QNX 4.
8356 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8358 *) A demo state-machine implementation was sponsored by
8359 Nuron (http://www.nuron.com/) and is now available in
8360 demos/state_machine.
8363 *) New options added to the 'dgst' utility for signature
8364 generation and verification.
8367 *) Unrecognized PKCS#7 content types are now handled via a
8368 catch all ASN1_TYPE structure. This allows unsupported
8369 types to be stored as a "blob" and an application can
8370 encode and decode it manually.
8373 *) Fix various signed/unsigned issues to make a_strex.c
8375 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8377 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8378 length if passed a buffer. ASN1_INTEGER_to_BN failed
8379 if passed a NULL BN and its argument was negative.
8380 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8382 *) Modification to PKCS#7 encoding routines to output definite
8383 length encoding. Since currently the whole structures are in
8384 memory there's not real point in using indefinite length
8385 constructed encoding. However if OpenSSL is compiled with
8386 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8389 *) Added BIO_vprintf() and BIO_vsnprintf().
8392 *) Added more prefixes to parse for in the the strings written
8393 through a logging bio, to cover all the levels that are available
8394 through syslog. The prefixes are now:
8396 PANIC, EMERG, EMR => LOG_EMERG
8397 ALERT, ALR => LOG_ALERT
8398 CRIT, CRI => LOG_CRIT
8399 ERROR, ERR => LOG_ERR
8400 WARNING, WARN, WAR => LOG_WARNING
8401 NOTICE, NOTE, NOT => LOG_NOTICE
8402 INFO, INF => LOG_INFO
8403 DEBUG, DBG => LOG_DEBUG
8405 and as before, if none of those prefixes are present at the
8406 beginning of the string, LOG_ERR is chosen.
8408 On Win32, the LOG_* levels are mapped according to this:
8410 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8411 LOG_WARNING => EVENTLOG_WARNING_TYPE
8412 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8416 *) Made it possible to reconfigure with just the configuration
8417 argument "reconf" or "reconfigure". The command line arguments
8418 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8419 and are retrieved from there when reconfiguring.
8423 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8425 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8428 *) The obj_dat.pl script was messing up the sorting of object
8429 names. The reason was that it compared the quoted version
8430 of strings as a result "OCSP" > "OCSP Signing" because
8431 " > SPACE. Changed script to store unquoted versions of
8432 names and add quotes on output. It was also omitting some
8433 names from the lookup table if they were given a default
8434 value (that is if SN is missing it is given the same
8435 value as LN and vice versa), these are now added on the
8436 grounds that if an object has a name we should be able to
8437 look it up. Finally added warning output when duplicate
8438 short or long names are found.
8441 *) Changes needed for Tandem NSK.
8442 [Scott Uroff <scott@xypro.com>]
8444 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8445 RSA_padding_check_SSLv23(), special padding was never detected
8446 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8447 version rollback attacks was not effective.
8449 In s23_clnt.c, don't use special rollback-attack detection padding
8450 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8451 client; similarly, in s23_srvr.c, don't do the rollback check if
8452 SSL 2.0 is the only protocol enabled in the server.
8455 *) Make it possible to get hexdumps of unprintable data with 'openssl
8456 asn1parse'. By implication, the functions ASN1_parse_dump() and
8457 BIO_dump_indent() are added.
8460 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8461 these print out strings and name structures based on various
8462 flags including RFC2253 support and proper handling of
8463 multibyte characters. Added options to the 'x509' utility
8464 to allow the various flags to be set.
8467 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8468 Also change the functions X509_cmp_current_time() and
8469 X509_gmtime_adj() work with an ASN1_TIME structure,
8470 this will enable certificates using GeneralizedTime in validity
8471 dates to be checked.
8474 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8475 negative public key encodings) on by default,
8476 NO_NEG_PUBKEY_BUG can be set to disable it.
8479 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8480 content octets. An i2c_ASN1_OBJECT is unnecessary because
8481 the encoding can be trivially obtained from the structure.
8484 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8485 not read locks (CRYPTO_r_[un]lock).
8488 *) A first attempt at creating official support for shared
8489 libraries through configuration. I've kept it so the
8490 default is static libraries only, and the OpenSSL programs
8491 are always statically linked for now, but there are
8492 preparations for dynamic linking in place.
8493 This has been tested on Linux and Tru64.
8496 *) Randomness polling function for Win9x, as described in:
8497 Peter Gutmann, Software Generation of Practically Strong
8501 *) Fix so PRNG is seeded in req if using an already existing
8505 *) New options to smime application. -inform and -outform
8506 allow alternative formats for the S/MIME message including
8507 PEM and DER. The -content option allows the content to be
8508 specified separately. This should allow things like Netscape
8509 form signing output easier to verify.
8512 *) Fix the ASN1 encoding of tags using the 'long form'.
8515 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8516 STRING types. These convert content octets to and from the
8517 underlying type. The actual tag and length octets are
8518 already assumed to have been read in and checked. These
8519 are needed because all other string types have virtually
8520 identical handling apart from the tag. By having versions
8521 of the ASN1 functions that just operate on content octets
8522 IMPLICIT tagging can be handled properly. It also allows
8523 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8524 and ASN1_INTEGER are identical apart from the tag.
8527 *) Change the handling of OID objects as follows:
8529 - New object identifiers are inserted in objects.txt, following
8530 the syntax given in objects.README.
8531 - objects.pl is used to process obj_mac.num and create a new
8533 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8536 This is currently kind of a hack, and the perl code in objects.pl
8537 isn't very elegant, but it works as I intended. The simplest way
8538 to check that it worked correctly is to look in obj_dat.h and
8539 check the array nid_objs and make sure the objects haven't moved
8540 around (this is important!). Additions are OK, as well as
8541 consistent name changes.
8544 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8547 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8548 The given file adds to whatever has already been seeded into the
8549 random pool through the RANDFILE configuration file option or
8550 environment variable, or the default random state file.
8553 *) mkstack.pl now sorts each macro group into lexical order.
8554 Previously the output order depended on the order the files
8555 appeared in the directory, resulting in needless rewriting
8559 *) Patches to make OpenSSL compile under Win32 again. Mostly
8560 work arounds for the VC++ problem that it treats func() as
8561 func(void). Also stripped out the parts of mkdef.pl that
8562 added extra typesafe functions: these no longer exist.
8565 *) Reorganisation of the stack code. The macros are now all
8566 collected in safestack.h . Each macro is defined in terms of
8567 a "stack macro" of the form SKM_<name>(type, a, b). The
8568 DEBUG_SAFESTACK is now handled in terms of function casts,
8569 this has the advantage of retaining type safety without the
8570 use of additional functions. If DEBUG_SAFESTACK is not defined
8571 then the non typesafe macros are used instead. Also modified the
8572 mkstack.pl script to handle the new form. Needs testing to see
8573 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8574 the default if no major problems. Similar behaviour for ASN1_SET_OF
8575 and PKCS12_STACK_OF.
8578 *) When some versions of IIS use the 'NET' form of private key the
8579 key derivation algorithm is different. Normally MD5(password) is
8580 used as a 128 bit RC4 key. In the modified case
8581 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
8582 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8583 as the old Netscape_RSA functions except they have an additional
8584 'sgckey' parameter which uses the modified algorithm. Also added
8585 an -sgckey command line option to the rsa utility. Thanks to
8586 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8587 algorithm to openssl-dev.
8590 *) The evp_local.h macros were using 'c.##kname' which resulted in
8591 invalid expansion on some systems (SCO 5.0.5 for example).
8592 Corrected to 'c.kname'.
8593 [Phillip Porch <root@theporch.com>]
8595 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8596 a STACK of email addresses from a certificate or request, these look
8597 in the subject name and the subject alternative name extensions and
8598 omit any duplicate addresses.
8601 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8602 This makes DSA verification about 2 % faster.
8605 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8606 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8607 plus overhead for 1024 bit moduli).
8608 This makes exponentiations about 0.5 % faster for 1024 bit
8609 exponents (as measured by "openssl speed rsa2048").
8612 *) Rename memory handling macros to avoid conflicts with other
8614 Malloc => OPENSSL_malloc
8615 Malloc_locked => OPENSSL_malloc_locked
8616 Realloc => OPENSSL_realloc
8617 Free => OPENSSL_free
8620 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8621 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8624 *) CygWin32 support.
8625 [John Jarvie <jjarvie@newsguy.com>]
8627 *) The type-safe stack code has been rejigged. It is now only compiled
8628 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8629 by default all type-specific stack functions are "#define"d back to
8630 standard stack functions. This results in more streamlined output
8631 but retains the type-safety checking possibilities of the original
8635 *) The STACK code has been cleaned up, and certain type declarations
8636 that didn't make a lot of sense have been brought in line. This has
8637 also involved a cleanup of sorts in safestack.h to more correctly
8638 map type-safe stack functions onto their plain stack counterparts.
8639 This work has also resulted in a variety of "const"ifications of
8640 lots of the code, especially "_cmp" operations which should normally
8641 be prototyped with "const" parameters anyway.
8644 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8645 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8646 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8647 where all of 'md' is used each time the PRNG is used, but 'state'
8648 is used only indexed by a cyclic counter. As entropy may not be
8649 well distributed from the beginning, 'md' is important as a
8650 chaining variable. However, the output function chains only half
8651 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8652 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8653 in all of 'state' being rewritten, with the new values depending
8654 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8657 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8658 the handshake is continued after ssl_verify_cert_chain();
8659 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8660 can lead to 'unexplainable' connection aborts later.
8661 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8663 *) Major EVP API cipher revision.
8664 Add hooks for extra EVP features. This allows various cipher
8665 parameters to be set in the EVP interface. Support added for variable
8666 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8667 setting of RC2 and RC5 parameters.
8669 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8672 Remove lots of duplicated code from the EVP library. For example *every*
8673 cipher init() function handles the 'iv' in the same way according to the
8674 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8675 for CFB and OFB modes they zero ctx->num.
8677 New functionality allows removal of S/MIME code RC2 hack.
8679 Most of the routines have the same form and so can be declared in terms
8682 By shifting this to the top level EVP_CipherInit() it can be removed from
8683 all individual ciphers. If the cipher wants to handle IVs or keys
8684 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8687 Change lots of functions like EVP_EncryptUpdate() to now return a
8688 value: although software versions of the algorithms cannot fail
8689 any installed hardware versions can.
8692 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8693 this option is set, tolerate broken clients that send the negotiated
8694 protocol version number instead of the requested protocol version
8698 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8699 i.e. non-zero for export ciphersuites, zero otherwise.
8700 Previous versions had this flag inverted, inconsistent with
8701 rsa_tmp_cb (..._TMP_RSA_CB).
8702 [Bodo Moeller; problem reported by Amit Chopra]
8704 *) Add missing DSA library text string. Work around for some IIS
8705 key files with invalid SEQUENCE encoding.
8708 *) Add a document (doc/standards.txt) that list all kinds of standards
8709 and so on that are implemented in OpenSSL.
8712 *) Enhance c_rehash script. Old version would mishandle certificates
8713 with the same subject name hash and wouldn't handle CRLs at all.
8714 Added -fingerprint option to crl utility, to support new c_rehash
8718 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8721 *) Fix for SSL server purpose checking. Server checking was
8722 rejecting certificates which had extended key usage present
8723 but no ssl client purpose.
8724 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8726 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8727 is a little unclear about how a blank password is handled.
8728 Since the password in encoded as a BMPString with terminating
8729 double NULL a zero length password would end up as just the
8730 double NULL. However no password at all is different and is
8731 handled differently in the PKCS#12 key generation code. NS
8732 treats a blank password as zero length. MSIE treats it as no
8733 password on export: but it will try both on import. We now do
8734 the same: PKCS12_parse() tries zero length and no password if
8735 the password is set to "" or NULL (NULL is now a valid password:
8736 it wasn't before) as does the pkcs12 application.
8739 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8740 perror when PEM_read_bio_X509_REQ fails, the error message must
8741 be obtained from the error queue.
8744 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8745 it in ERR_remove_state if appropriate, and change ERR_get_state
8746 accordingly to avoid race conditions (this is necessary because
8747 thread_hash is no longer constant once set).
8750 *) Bugfix for linux-elf makefile.one.
8753 *) RSA_get_default_method() will now cause a default
8754 RSA_METHOD to be chosen if one doesn't exist already.
8755 Previously this was only set during a call to RSA_new()
8756 or RSA_new_method(NULL) meaning it was possible for
8757 RSA_get_default_method() to return NULL.
8760 *) Added native name translation to the existing DSO code
8761 that will convert (if the flag to do so is set) filenames
8762 that are sufficiently small and have no path information
8763 into a canonical native form. Eg. "blah" converted to
8764 "libblah.so" or "blah.dll" etc.
8767 *) New function ERR_error_string_n(e, buf, len) which is like
8768 ERR_error_string(e, buf), but writes at most 'len' bytes
8769 including the 0 terminator. For ERR_error_string_n, 'buf'
8771 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8773 *) CONF library reworked to become more general. A new CONF
8774 configuration file reader "class" is implemented as well as a
8775 new functions (NCONF_*, for "New CONF") to handle it. The now
8776 old CONF_* functions are still there, but are reimplemented to
8777 work in terms of the new functions. Also, a set of functions
8778 to handle the internal storage of the configuration data is
8779 provided to make it easier to write new configuration file
8780 reader "classes" (I can definitely see something reading a
8781 configuration file in XML format, for example), called _CONF_*,
8782 or "the configuration storage API"...
8784 The new configuration file reading functions are:
8786 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8787 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8789 NCONF_default, NCONF_WIN32
8791 NCONF_dump_fp, NCONF_dump_bio
8793 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8794 NCONF_new creates a new CONF object. This works in the same way
8795 as other interfaces in OpenSSL, like the BIO interface.
8796 NCONF_dump_* dump the internal storage of the configuration file,
8797 which is useful for debugging. All other functions take the same
8798 arguments as the old CONF_* functions wth the exception of the
8799 first that must be a `CONF *' instead of a `LHASH *'.
8801 To make it easer to use the new classes with the old CONF_* functions,
8802 the function CONF_set_default_method is provided.
8805 *) Add '-tls1' option to 'openssl ciphers', which was already
8806 mentioned in the documentation but had not been implemented.
8807 (This option is not yet really useful because even the additional
8808 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8811 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8812 OpenSSL-based applications) load shared libraries and bind to
8813 them in a portable way.
8814 [Geoff Thorpe, with contributions from Richard Levitte]
8816 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8818 *) Make sure _lrotl and _lrotr are only used with MSVC.
8820 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8821 (the default implementation of RAND_status).
8823 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8824 to '-clrext' (= clear extensions), as intended and documented.
8825 [Bodo Moeller; inconsistency pointed out by Michael Attili
8828 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8829 was larger than the MD block size.
8830 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8832 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8833 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8834 using the passed key: if the passed key was a private key the result
8835 of X509_print(), for example, would be to print out all the private key
8839 *) des_quad_cksum() byte order bug fix.
8840 [Ulf Möller, using the problem description in krb4-0.9.7, where
8841 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8843 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8845 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8847 *) For easily testing in shell scripts whether some command
8848 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8849 returns with exit code 0 iff no command of the given name is available.
8850 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8851 the output goes to stdout and nothing is printed to stderr.
8852 Additional arguments are always ignored.
8854 Since for each cipher there is a command of the same name,
8855 the 'no-cipher' compilation switches can be tested this way.
8857 ('openssl no-XXX' is not able to detect pseudo-commands such
8858 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8861 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8864 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8865 is set; it will be thrown away anyway because each handshake creates
8867 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8868 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8869 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
8870 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8873 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8874 'Q' and 'R' lose their special meanings (quit/renegotiate).
8875 This is part of what -quiet does; unlike -quiet, -ign_eof
8876 does not suppress any output.
8879 *) Add compatibility options to the purpose and trust code. The
8880 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8881 accepts a certificate or CA, this was the previous behaviour,
8882 with all the associated security issues.
8884 X509_TRUST_COMPAT is the old trust behaviour: only and
8885 automatically trust self signed roots in certificate store. A
8886 new trust setting X509_TRUST_DEFAULT is used to specify that
8887 a purpose has no associated trust setting and it should instead
8888 use the value in the default purpose.
8891 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8892 and fix a memory leak.
8895 *) In util/mkerr.pl (which implements 'make errors'), preserve
8896 reason strings from the previous version of the .c file, as
8897 the default to have only downcase letters (and digits) in
8898 automatically generated reasons codes is not always appropriate.
8901 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8902 using strerror. Previously, ERR_reason_error_string() returned
8903 library names as reason strings for SYSerr; but SYSerr is a special
8904 case where small numbers are errno values, not library numbers.
8907 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8908 converts DSA parameters into DH parameters. (When creating parameters,
8909 DSA_generate_parameters is used.)
8912 *) Include 'length' (recommended exponent length) in C code generated
8913 by 'openssl dhparam -C'.
8916 *) The second argument to set_label in perlasm was already being used
8917 so couldn't be used as a "file scope" flag. Moved to third argument
8921 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8922 instead of RAND_bytes for encryption IVs and salts.
8925 *) Include RAND_status() into RAND_METHOD instead of implementing
8926 it only for md_rand.c Otherwise replacing the PRNG by calling
8927 RAND_set_rand_method would be impossible.
8930 *) Don't let DSA_generate_key() enter an infinite loop if the random
8931 number generation fails.
8934 *) New 'rand' application for creating pseudo-random output.
8937 *) Added configuration support for Linux/IA64
8938 [Rolf Haberrecker <rolf@suse.de>]
8940 *) Assembler module support for Mingw32.
8943 *) Shared library support for HPUX (in shlib/).
8944 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8946 *) Shared library support for Solaris gcc.
8947 [Lutz Behnke <behnke@trustcenter.de>]
8949 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8951 *) PKCS7_encrypt() was adding text MIME headers twice because they
8952 were added manually and by SMIME_crlf_copy().
8955 *) In bntest.c don't call BN_rand with zero bits argument.
8956 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8958 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8959 case was implemented. This caused BN_div_recp() to fail occasionally.
8962 *) Add an optional second argument to the set_label() in the perl
8963 assembly language builder. If this argument exists and is set
8964 to 1 it signals that the assembler should use a symbol whose
8965 scope is the entire file, not just the current function. This
8966 is needed with MASM which uses the format label:: for this scope.
8967 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8969 *) Change the ASN1 types so they are typedefs by default. Before
8970 almost all types were #define'd to ASN1_STRING which was causing
8971 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8975 *) Change names of new functions to the new get1/get0 naming
8976 convention: After 'get1', the caller owns a reference count
8977 and has to call ..._free; 'get0' returns a pointer to some
8978 data structure without incrementing reference counters.
8979 (Some of the existing 'get' functions increment a reference
8980 counter, some don't.)
8981 Similarly, 'set1' and 'add1' functions increase reference
8982 counters or duplicate objects.
8985 *) Allow for the possibility of temp RSA key generation failure:
8986 the code used to assume it always worked and crashed on failure.
8989 *) Fix potential buffer overrun problem in BIO_printf().
8990 [Ulf Möller, using public domain code by Patrick Powell; problem
8991 pointed out by David Sacerdote <das33@cornell.edu>]
8993 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8994 RAND_egd() and RAND_status(). In the command line application,
8995 the EGD socket can be specified like a seed file using RANDFILE
8999 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
9000 Some CAs (e.g. Verisign) distribute certificates in this form.
9003 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
9004 list to exclude them. This means that no special compilation option
9005 is needed to use anonymous DH: it just needs to be included in the
9009 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
9010 EVP_MD_type. The old functionality is available in a new macro called
9011 EVP_MD_md(). Change code that uses it and update docs.
9014 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
9015 where the 'void *' argument is replaced by a function pointer argument.
9016 Previously 'void *' was abused to point to functions, which works on
9017 many platforms, but is not correct. As these functions are usually
9018 called by macros defined in OpenSSL header files, most source code
9019 should work without changes.
9022 *) <openssl/opensslconf.h> (which is created by Configure) now contains
9023 sections with information on -D... compiler switches used for
9024 compiling the library so that applications can see them. To enable
9025 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
9026 must be defined. E.g.,
9027 #define OPENSSL_ALGORITHM_DEFINES
9028 #include <openssl/opensslconf.h>
9029 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
9030 [Richard Levitte, Ulf and Bodo Möller]
9032 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
9036 *) Change the 'other' type in certificate aux info to a STACK_OF
9037 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
9038 the required ASN1 format: arbitrary types determined by an OID.
9041 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
9042 argument to 'req'. This is not because the function is newer or
9043 better than others it just uses the work 'NEW' in the certificate
9044 request header lines. Some software needs this.
9047 *) Reorganise password command line arguments: now passwords can be
9048 obtained from various sources. Delete the PEM_cb function and make
9049 it the default behaviour: i.e. if the callback is NULL and the
9050 usrdata argument is not NULL interpret it as a null terminated pass
9051 phrase. If usrdata and the callback are NULL then the pass phrase
9052 is prompted for as usual.
9055 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
9056 the support is automatically enabled. The resulting binaries will
9057 autodetect the card and use it if present.
9058 [Ben Laurie and Compaq Inc.]
9060 *) Work around for Netscape hang bug. This sends certificate request
9061 and server done in one record. Since this is perfectly legal in the
9062 SSL/TLS protocol it isn't a "bug" option and is on by default. See
9063 the bugs/SSLv3 entry for more info.
9066 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
9069 *) Add -rand argument to smime and pkcs12 applications and read/write
9073 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
9076 *) Add command line password options to the remaining applications.
9079 *) Bug fix for BN_div_recp() for numerators with an even number of
9083 *) More tests in bntest.c, and changed test_bn output.
9086 *) ./config recognizes MacOS X now.
9089 *) Bug fix for BN_div() when the first words of num and divsor are
9090 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
9093 *) Add support for various broken PKCS#8 formats, and command line
9094 options to produce them.
9097 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
9098 get temporary BIGNUMs from a BN_CTX.
9101 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
9105 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
9106 include a #define from the old name to the new. The original intent
9107 was that statically linked binaries could for example just call
9108 SSLeay_add_all_ciphers() to just add ciphers to the table and not
9109 link with digests. This never worked becayse SSLeay_add_all_digests()
9110 and SSLeay_add_all_ciphers() were in the same source file so calling
9111 one would link with the other. They are now in separate source files.
9114 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
9117 *) Use a less unusual form of the Miller-Rabin primality test (it used
9118 a binary algorithm for exponentiation integrated into the Miller-Rabin
9119 loop, our standard modexp algorithms are faster).
9122 *) Support for the EBCDIC character set completed.
9123 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
9125 *) Source code cleanups: use const where appropriate, eliminate casts,
9126 use void * instead of char * in lhash.
9129 *) Bugfix: ssl3_send_server_key_exchange was not restartable
9130 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
9131 this the server could overwrite ephemeral keys that the client
9135 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
9136 using 50 iterations of the Rabin-Miller test.
9138 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
9139 iterations of the Rabin-Miller test as required by the appendix
9140 to FIPS PUB 186[-1]) instead of DSA_is_prime.
9141 As BN_is_prime_fasttest includes trial division, DSA parameter
9142 generation becomes much faster.
9144 This implies a change for the callback functions in DSA_is_prime
9145 and DSA_generate_parameters: The callback function is called once
9146 for each positive witness in the Rabin-Miller test, not just
9147 occasionally in the inner loop; and the parameters to the
9148 callback function now provide an iteration count for the outer
9149 loop rather than for the current invocation of the inner loop.
9150 DSA_generate_parameters additionally can call the callback
9151 function with an 'iteration count' of -1, meaning that a
9152 candidate has passed the trial division test (when q is generated
9153 from an application-provided seed, trial division is skipped).
9156 *) New function BN_is_prime_fasttest that optionally does trial
9157 division before starting the Rabin-Miller test and has
9158 an additional BN_CTX * argument (whereas BN_is_prime always
9159 has to allocate at least one BN_CTX).
9160 'callback(1, -1, cb_arg)' is called when a number has passed the
9161 trial division stage.
9164 *) Fix for bug in CRL encoding. The validity dates weren't being handled
9168 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
9171 *) New function BN_pseudo_rand().
9174 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
9175 bignum version of BN_from_montgomery() with the working code from
9176 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
9180 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
9181 made it impossible to use the same SSL_SESSION data structure in
9182 SSL2 clients in multiple threads.
9185 *) The return value of RAND_load_file() no longer counts bytes obtained
9186 by stat(). RAND_load_file(..., -1) is new and uses the complete file
9187 to seed the PRNG (previously an explicit byte count was required).
9188 [Ulf Möller, Bodo Möller]
9190 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
9191 used (char *) instead of (void *) and had casts all over the place.
9194 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
9197 *) Retain source code compatibility for BN_prime_checks macro:
9198 BN_is_prime(..., BN_prime_checks, ...) now uses
9199 BN_prime_checks_for_size to determine the appropriate number of
9200 Rabin-Miller iterations.
9203 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
9204 DH_CHECK_P_NOT_SAFE_PRIME.
9205 (Check if this is true? OpenPGP calls them "strong".)
9208 *) Merge the functionality of "dh" and "gendh" programs into a new program
9209 "dhparam". The old programs are retained for now but will handle DH keys
9210 (instead of parameters) in future.
9213 *) Make the ciphers, s_server and s_client programs check the return values
9214 when a new cipher list is set.
9217 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9218 ciphers. Before when the 56bit ciphers were enabled the sorting was
9221 The syntax for the cipher sorting has been extended to support sorting by
9222 cipher-strength (using the strength_bits hard coded in the tables).
9223 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9225 Fix a bug in the cipher-command parser: when supplying a cipher command
9226 string with an "undefined" symbol (neither command nor alphanumeric
9227 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9228 an error is flagged.
9230 Due to the strength-sorting extension, the code of the
9231 ssl_create_cipher_list() function was completely rearranged. I hope that
9232 the readability was also increased :-)
9233 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9235 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9236 for the first serial number and places 2 in the serial number file. This
9237 avoids problems when the root CA is created with serial number zero and
9238 the first user certificate has the same issuer name and serial number
9242 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9243 the new code. Add documentation for this stuff.
9246 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9247 X509_*() to X509at_*() on the grounds that they don't handle X509
9248 structures and behave in an analagous way to the X509v3 functions:
9249 they shouldn't be called directly but wrapper functions should be used
9252 So we also now have some wrapper functions that call the X509at functions
9253 when passed certificate requests. (TO DO: similar things can be done with
9254 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9255 things. Some of these need some d2i or i2d and print functionality
9256 because they handle more complex structures.)
9259 *) Add missing #ifndefs that caused missing symbols when building libssl
9260 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9261 NO_RSA in ssl/s2*.c.
9262 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9264 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9265 has a return value which indicates the quality of the random data
9266 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9267 error queue. New function RAND_pseudo_bytes() generates output that is
9268 guaranteed to be unique but not unpredictable. RAND_add is like
9269 RAND_seed, but takes an extra argument for an entropy estimate
9270 (RAND_seed always assumes full entropy).
9273 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9274 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9275 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9276 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9277 false-positive rate of at most 2^-80 for random input.
9280 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9283 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9284 in the 0.9.5 release), this returns the chain
9285 from an X509_CTX structure with a dup of the stack and all
9286 the X509 reference counts upped: so the stack will exist
9287 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9290 Also make SSL_SESSION_print() print out the verify return
9294 *) Add manpage for the pkcs12 command. Also change the default
9295 behaviour so MAC iteration counts are used unless the new
9296 -nomaciter option is used. This improves file security and
9297 only older versions of MSIE (4.0 for example) need it.
9300 *) Honor the no-xxx Configure options when creating .DEF files.
9303 *) Add PKCS#10 attributes to field table: challengePassword,
9304 unstructuredName and unstructuredAddress. These are taken from
9305 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9306 international characters are used.
9308 More changes to X509_ATTRIBUTE code: allow the setting of types
9309 based on strings. Remove the 'loc' parameter when adding
9310 attributes because these will be a SET OF encoding which is sorted
9314 *) Initial changes to the 'req' utility to allow request generation
9315 automation. This will allow an application to just generate a template
9316 file containing all the field values and have req construct the
9319 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9320 used all over the place including certificate requests and PKCS#7
9321 structures. They are currently handled manually where necessary with
9322 some primitive wrappers for PKCS#7. The new functions behave in a
9323 manner analogous to the X509 extension functions: they allow
9324 attributes to be looked up by NID and added.
9326 Later something similar to the X509V3 code would be desirable to
9327 automatically handle the encoding, decoding and printing of the
9328 more complex types. The string types like challengePassword can
9329 be handled by the string table functions.
9331 Also modified the multi byte string table handling. Now there is
9332 a 'global mask' which masks out certain types. The table itself
9333 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9334 is useful when for example there is only one permissible type
9335 (as in countryName) and using the mask might result in no valid
9339 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9340 SSL_get_peer_finished to allow applications to obtain the latest
9341 Finished messages sent to the peer or expected from the peer,
9342 respectively. (SSL_get_peer_finished is usually the Finished message
9343 actually received from the peer, otherwise the protocol will be aborted.)
9345 As the Finished message are message digests of the complete handshake
9346 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9347 be used for external authentication procedures when the authentication
9348 provided by SSL/TLS is not desired or is not enough.
9351 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9352 the host supports BWX extension and if Compaq C is present on the
9353 $PATH. Just exploiting of the BWX extension results in 20-30%
9354 performance kick for some algorithms, e.g. DES and RC4 to mention
9355 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9359 *) Add support for MS "fast SGC". This is arguably a violation of the
9360 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9361 weak crypto and after checking the certificate is SGC a second one
9362 with strong crypto. MS SGC stops the first handshake after receiving
9363 the server certificate message and sends a second client hello. Since
9364 a server will typically do all the time consuming operations before
9365 expecting any further messages from the client (server key exchange
9366 is the most expensive) there is little difference between the two.
9368 To get OpenSSL to support MS SGC we have to permit a second client
9369 hello message after we have sent server done. In addition we have to
9370 reset the MAC if we do get this second client hello.
9373 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9374 if a DER encoded private key is RSA or DSA traditional format. Changed
9375 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9376 format DER encoded private key. Newer code should use PKCS#8 format which
9377 has the key type encoded in the ASN1 structure. Added DER private key
9378 support to pkcs8 application.
9381 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9382 ciphersuites has been selected (as required by the SSL 3/TLS 1
9383 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9384 is set, we interpret this as a request to violate the specification
9385 (the worst that can happen is a handshake failure, and 'correct'
9386 behaviour would result in a handshake failure anyway).
9389 *) In SSL_CTX_add_session, take into account that there might be multiple
9390 SSL_SESSION structures with the same session ID (e.g. when two threads
9391 concurrently obtain them from an external cache).
9392 The internal cache can handle only one SSL_SESSION with a given ID,
9393 so if there's a conflict, we now throw out the old one to achieve
9397 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9398 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9399 some routines that use cipher OIDs: some ciphers do not have OIDs
9400 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9404 *) Simplify the trust setting structure and code. Now we just have
9405 two sequences of OIDs for trusted and rejected settings. These will
9406 typically have values the same as the extended key usage extension
9407 and any application specific purposes.
9409 The trust checking code now has a default behaviour: it will just
9410 check for an object with the same NID as the passed id. Functions can
9411 be provided to override either the default behaviour or the behaviour
9412 for a given id. SSL client, server and email already have functions
9413 in place for compatibility: they check the NID and also return "trusted"
9414 if the certificate is self signed.
9417 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9418 traditional format into an EVP_PKEY structure.
9421 *) Add a password callback function PEM_cb() which either prompts for
9422 a password if usr_data is NULL or otherwise assumes it is a null
9423 terminated password. Allow passwords to be passed on command line
9424 environment or config files in a few more utilities.
9427 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9428 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9429 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9430 Update documentation.
9433 *) Support for ASN1 "NULL" type. This could be handled before by using
9434 ASN1_TYPE but there wasn't any function that would try to read a NULL
9435 and produce an error if it couldn't. For compatibility we also have
9436 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9437 don't allocate anything because they don't need to.
9440 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9442 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9444 *) Rebuild of the memory allocation routines used by OpenSSL code and
9445 possibly others as well. The purpose is to make an interface that
9446 provide hooks so anyone can build a separate set of allocation and
9447 deallocation routines to be used by OpenSSL, for example memory
9448 pool implementations, or something else, which was previously hard
9449 since Malloc(), Realloc() and Free() were defined as macros having
9450 the values malloc, realloc and free, respectively (except for Win32
9451 compilations). The same is provided for memory debugging code.
9452 OpenSSL already comes with functionality to find memory leaks, but
9453 this gives people a chance to debug other memory problems.
9455 With these changes, a new set of functions and macros have appeared:
9457 CRYPTO_set_mem_debug_functions() [F]
9458 CRYPTO_get_mem_debug_functions() [F]
9459 CRYPTO_dbg_set_options() [F]
9460 CRYPTO_dbg_get_options() [F]
9461 CRYPTO_malloc_debug_init() [M]
9463 The memory debug functions are NULL by default, unless the library
9464 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9465 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9466 gives the standard debugging functions that come with OpenSSL) or
9467 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9468 provided by the library user) must be used. When the standard
9469 debugging functions are used, CRYPTO_dbg_set_options can be used to
9470 request additional information:
9471 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9472 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9474 Also, things like CRYPTO_set_mem_functions will always give the
9475 expected result (the new set of functions is used for allocation
9476 and deallocation) at all times, regardless of platform and compiler
9479 To finish it up, some functions that were never use in any other
9480 way than through macros have a new API and new semantic:
9483 CRYPTO_dbg_realloc()
9486 All macros of value have retained their old syntax.
9487 [Richard Levitte and Bodo Moeller]
9489 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9490 ordering of SMIMECapabilities wasn't in "strength order" and there
9491 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9495 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9496 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9497 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9499 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9500 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9501 functionality to handle multipart/signed properly) and a utility
9502 called 'smime' to call all this stuff. This is based on code I
9503 originally wrote for Celo who have kindly allowed it to be
9504 included in OpenSSL.
9507 *) Add variants des_set_key_checked and des_set_key_unchecked of
9508 des_set_key (aka des_key_sched). Global variable des_check_key
9509 decides which of these is called by des_set_key; this way
9510 des_check_key behaves as it always did, but applications and
9511 the library itself, which was buggy for des_check_key == 1,
9512 have a cleaner way to pick the version they need.
9515 *) New function PKCS12_newpass() which changes the password of a
9519 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9520 dynamic mix. In both cases the ids can be used as an index into the
9521 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9522 functions so they accept a list of the field values and the
9523 application doesn't need to directly manipulate the X509_TRUST
9527 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9531 *) Modify the way the V3 extension code looks up extensions. This now
9532 works in a similar way to the object code: we have some "standard"
9533 extensions in a static table which is searched with OBJ_bsearch()
9534 and the application can add dynamic ones if needed. The file
9535 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9536 updated whenever a new extension is added to the core code and kept
9537 in ext_nid order. There is a simple program 'tabtest.c' which checks
9538 this. New extensions are not added too often so this file can readily
9539 be maintained manually.
9541 There are two big advantages in doing things this way. The extensions
9542 can be looked up immediately and no longer need to be "added" using
9543 X509V3_add_standard_extensions(): this function now does nothing.
9544 [Side note: I get *lots* of email saying the extension code doesn't
9545 work because people forget to call this function]
9546 Also no dynamic allocation is done unless new extensions are added:
9547 so if we don't add custom extensions there is no need to call
9548 X509V3_EXT_cleanup().
9551 *) Modify enc utility's salting as follows: make salting the default. Add a
9552 magic header, so unsalted files fail gracefully instead of just decrypting
9553 to garbage. This is because not salting is a big security hole, so people
9554 should be discouraged from doing it.
9557 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9558 digest to be passed on the command line but it only used this
9559 parameter when signing a certificate. Modified so all relevant
9560 operations are affected by the digest parameter including the
9561 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9562 DSA key was used because it didn't fix the digest.
9565 *) Initial certificate chain verify code. Currently tests the untrusted
9566 certificates for consistency with the verify purpose (which is set
9567 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9569 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9570 this is because it will reject chains with invalid extensions whereas
9571 every previous version of OpenSSL and SSLeay made no checks at all.
9573 Trust code: checks the root CA for the relevant trust settings. Trust
9574 settings have an initial value consistent with the verify purpose: e.g.
9575 if the verify purpose is for SSL client use it expects the CA to be
9576 trusted for SSL client use. However the default value can be changed to
9577 permit custom trust settings: one example of this would be to only trust
9578 certificates from a specific "secure" set of CAs.
9580 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9581 which should be used for version portability: especially since the
9582 verify structure is likely to change more often now.
9584 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9585 to set them. If not set then assume SSL clients will verify SSL servers
9588 Two new options to the verify program: -untrusted allows a set of
9589 untrusted certificates to be passed in and -purpose which sets the
9590 intended purpose of the certificate. If a purpose is set then the
9591 new chain verify code is used to check extension consistency.
9594 *) Support for the authority information access extension.
9597 *) Modify RSA and DSA PEM read routines to transparently handle
9598 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9599 public keys in a format compatible with certificate
9600 SubjectPublicKeyInfo structures. Unfortunately there were already
9601 functions called *_PublicKey_* which used various odd formats so
9602 these are retained for compatibility: however the DSA variants were
9603 never in a public release so they have been deleted. Changed dsa/rsa
9604 utilities to handle the new format: note no releases ever handled public
9605 keys so we should be OK.
9607 The primary motivation for this change is to avoid the same fiasco
9608 that dogs private keys: there are several incompatible private key
9609 formats some of which are standard and some OpenSSL specific and
9610 require various evil hacks to allow partial transparent handling and
9611 even then it doesn't work with DER formats. Given the option anything
9612 other than PKCS#8 should be dumped: but the other formats have to
9613 stay in the name of compatibility.
9615 With public keys and the benefit of hindsight one standard format
9616 is used which works with EVP_PKEY, RSA or DSA structures: though
9617 it clearly returns an error if you try to read the wrong kind of key.
9619 Added a -pubkey option to the 'x509' utility to output the public key.
9620 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9621 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9622 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9623 that do the same as the EVP_PKEY_assign_*() except they up the
9624 reference count of the added key (they don't "swallow" the
9628 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9629 CRLs would fail if the file contained no certificates or no CRLs:
9630 added a new function to read in both types and return the number
9631 read: this means that if none are read it will be an error. The
9632 DER versions of the certificate and CRL reader would always fail
9633 because it isn't possible to mix certificates and CRLs in DER format
9634 without choking one or the other routine. Changed this to just read
9635 a certificate: this is the best we can do. Also modified the code
9636 in apps/verify.c to take notice of return codes: it was previously
9637 attempting to read in certificates from NULL pointers and ignoring
9638 any errors: this is one reason why the cert and CRL reader seemed
9639 to work. It doesn't check return codes from the default certificate
9640 routines: these may well fail if the certificates aren't installed.
9643 *) Code to support otherName option in GeneralName.
9646 *) First update to verify code. Change the verify utility
9647 so it warns if it is passed a self signed certificate:
9648 for consistency with the normal behaviour. X509_verify
9649 has been modified to it will now verify a self signed
9650 certificate if *exactly* the same certificate appears
9651 in the store: it was previously impossible to trust a
9652 single self signed certificate. This means that:
9653 openssl verify ss.pem
9654 now gives a warning about a self signed certificate but
9655 openssl verify -CAfile ss.pem ss.pem
9659 *) For servers, store verify_result in SSL_SESSION data structure
9660 (and add it to external session representation).
9661 This is needed when client certificate verifications fails,
9662 but an application-provided verification callback (set by
9663 SSL_CTX_set_cert_verify_callback) allows accepting the session
9664 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9665 but returns 1): When the session is reused, we have to set
9666 ssl->verify_result to the appropriate error code to avoid
9668 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9670 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9671 case in PKCS7_dataInit() where the signed PKCS7 structure
9672 didn't contain any existing data because it was being created.
9673 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9675 *) Add a salt to the key derivation routines in enc.c. This
9676 forms the first 8 bytes of the encrypted file. Also add a
9677 -S option to allow a salt to be input on the command line.
9680 *) New function X509_cmp(). Oddly enough there wasn't a function
9681 to compare two certificates. We do this by working out the SHA1
9682 hash and comparing that. X509_cmp() will be needed by the trust
9686 *) SSL_get1_session() is like SSL_get_session(), but increments
9687 the reference count in the SSL_SESSION returned.
9688 [Geoff Thorpe <geoff@eu.c2.net>]
9690 *) Fix for 'req': it was adding a null to request attributes.
9691 Also change the X509_LOOKUP and X509_INFO code to handle
9692 certificate auxiliary information.
9695 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9699 *) Add the possibility to add extra information to the memory leak
9700 detecting output, to form tracebacks, showing from where each
9701 allocation was originated: CRYPTO_push_info("constant string") adds
9702 the string plus current file name and line number to a per-thread
9703 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9704 is like calling CYRPTO_pop_info() until the stack is empty.
9705 Also updated memory leak detection code to be multi-thread-safe.
9708 *) Add options -text and -noout to pkcs7 utility and delete the
9709 encryption options which never did anything. Update docs.
9712 *) Add options to some of the utilities to allow the pass phrase
9713 to be included on either the command line (not recommended on
9714 OSes like Unix) or read from the environment. Update the
9715 manpages and fix a few bugs.
9718 *) Add a few manpages for some of the openssl commands.
9721 *) Fix the -revoke option in ca. It was freeing up memory twice,
9722 leaking and not finding already revoked certificates.
9725 *) Extensive changes to support certificate auxiliary information.
9726 This involves the use of X509_CERT_AUX structure and X509_AUX
9727 functions. An X509_AUX function such as PEM_read_X509_AUX()
9728 can still read in a certificate file in the usual way but it
9729 will also read in any additional "auxiliary information". By
9730 doing things this way a fair degree of compatibility can be
9731 retained: existing certificates can have this information added
9732 using the new 'x509' options.
9734 Current auxiliary information includes an "alias" and some trust
9735 settings. The trust settings will ultimately be used in enhanced
9736 certificate chain verification routines: currently a certificate
9737 can only be trusted if it is self signed and then it is trusted
9741 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9742 The problem was that one of the replacement routines had not been working
9743 since SSLeay releases. For now the offending routine has been replaced
9744 with non-optimised assembler. Even so, this now gives around 95%
9745 performance improvement for 1024 bit RSA signs.
9748 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9749 handling. Most clients have the effective key size in bits equal to
9750 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9751 A few however don't do this and instead use the size of the decrypted key
9752 to determine the RC2 key length and the AlgorithmIdentifier to determine
9753 the effective key length. In this case the effective key length can still
9754 be 40 bits but the key length can be 168 bits for example. This is fixed
9755 by manually forcing an RC2 key into the EVP_PKEY structure because the
9756 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9757 the key length and effective key length are equal.
9760 *) Add a bunch of functions that should simplify the creation of
9761 X509_NAME structures. Now you should be able to do:
9762 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9763 and have it automatically work out the correct field type and fill in
9764 the structures. The more adventurous can try:
9765 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9766 and it will (hopefully) work out the correct multibyte encoding.
9769 *) Change the 'req' utility to use the new field handling and multibyte
9770 copy routines. Before the DN field creation was handled in an ad hoc
9771 way in req, ca, and x509 which was rather broken and didn't support
9772 BMPStrings or UTF8Strings. Since some software doesn't implement
9773 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9774 using the dirstring_type option. See the new comment in the default
9775 openssl.cnf for more info.
9778 *) Make crypto/rand/md_rand.c more robust:
9779 - Assure unique random numbers after fork().
9780 - Make sure that concurrent threads access the global counter and
9781 md serializably so that we never lose entropy in them
9782 or use exactly the same state in multiple threads.
9783 Access to the large state is not always serializable because
9784 the additional locking could be a performance killer, and
9785 md should be large enough anyway.
9788 *) New file apps/app_rand.c with commonly needed functionality
9789 for handling the random seed file.
9791 Use the random seed file in some applications that previously did not:
9793 dsaparam -genkey (which also ignored its '-rand' option),
9796 x509 (when signing).
9797 Except on systems with /dev/urandom, it is crucial to have a random
9798 seed file at least for key creation, DSA signing, and for DH exchanges;
9799 for RSA signatures we could do without one.
9801 gendh and gendsa (unlike genrsa) used to read only the first byte
9802 of each file listed in the '-rand' option. The function as previously
9803 found in genrsa is now in app_rand.c and is used by all programs
9804 that support '-rand'.
9807 *) In RAND_write_file, use mode 0600 for creating files;
9808 don't just chmod when it may be too late.
9811 *) Report an error from X509_STORE_load_locations
9812 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9815 *) New function ASN1_mbstring_copy() this copies a string in either
9816 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9817 into an ASN1_STRING type. A mask of permissible types is passed
9818 and it chooses the "minimal" type to use or an error if not type
9822 *) Add function equivalents to the various macros in asn1.h. The old
9823 macros are retained with an M_ prefix. Code inside the library can
9824 use the M_ macros. External code (including the openssl utility)
9825 should *NOT* in order to be "shared library friendly".
9828 *) Add various functions that can check a certificate's extensions
9829 to see if it usable for various purposes such as SSL client,
9830 server or S/MIME and CAs of these types. This is currently
9831 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9832 verification. Also added a -purpose flag to x509 utility to
9833 print out all the purposes.
9836 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9840 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9841 for, obtain and decode and extension and obtain its critical flag.
9842 This allows all the necessary extension code to be handled in a
9843 single function call.
9846 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9847 platforms. See crypto/rc4/rc4_enc.c for further details.
9850 *) New -noout option to asn1parse. This causes no output to be produced
9851 its main use is when combined with -strparse and -out to extract data
9852 from a file (which may not be in ASN.1 format).
9855 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9856 when producing the local key id.
9857 [Richard Levitte <levitte@stacken.kth.se>]
9859 *) New option -dhparam in s_server. This allows a DH parameter file to be
9860 stated explicitly. If it is not stated then it tries the first server
9861 certificate file. The previous behaviour hard coded the filename
9865 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9866 a public key to be input or output. For example:
9867 openssl rsa -in key.pem -pubout -out pubkey.pem
9868 Also added necessary DSA public key functions to handle this.
9871 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9872 in the message. This was handled by allowing
9873 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9874 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9876 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9877 to the end of the strings whereas this didn't. This would cause problems
9878 if strings read with d2i_ASN1_bytes() were later modified.
9879 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9881 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9882 data and it contains EOF it will end up returning an error. This is
9883 caused by input 46 bytes long. The cause is due to the way base64
9884 BIOs find the start of base64 encoded data. They do this by trying a
9885 trial decode on each line until they find one that works. When they
9886 do a flag is set and it starts again knowing it can pass all the
9887 data directly through the decoder. Unfortunately it doesn't reset
9888 the context it uses. This means that if EOF is reached an attempt
9889 is made to pass two EOFs through the context and this causes the
9890 resulting error. This can also cause other problems as well. As is
9891 usual with these problems it takes *ages* to find and the fix is
9892 trivial: move one line.
9893 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9895 *) Ugly workaround to get s_client and s_server working under Windows. The
9896 old code wouldn't work because it needed to select() on sockets and the
9897 tty (for keypresses and to see if data could be written). Win32 only
9898 supports select() on sockets so we select() with a 1s timeout on the
9899 sockets and then see if any characters are waiting to be read, if none
9900 are present then we retry, we also assume we can always write data to
9901 the tty. This isn't nice because the code then blocks until we've
9902 received a complete line of data and it is effectively polling the
9903 keyboard at 1s intervals: however it's quite a bit better than not
9904 working at all :-) A dedicated Windows application might handle this
9905 with an event loop for example.
9908 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9909 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9910 will be called when RSA_sign() and RSA_verify() are used. This is useful
9911 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9912 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9913 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9914 This necessitated the support of an extra signature type NID_md5_sha1
9915 for SSL signatures and modifications to the SSL library to use it instead
9916 of calling RSA_public_decrypt() and RSA_private_encrypt().
9919 *) Add new -verify -CAfile and -CApath options to the crl program, these
9920 will lookup a CRL issuers certificate and verify the signature in a
9921 similar way to the verify program. Tidy up the crl program so it
9922 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9923 less strict. It will now permit CRL extensions even if it is not
9924 a V2 CRL: this will allow it to tolerate some broken CRLs.
9927 *) Initialize all non-automatic variables each time one of the openssl
9928 sub-programs is started (this is necessary as they may be started
9929 multiple times from the "OpenSSL>" prompt).
9930 [Lennart Bang, Bodo Moeller]
9932 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9933 removing all other RSA functionality (this is what NO_RSA does). This
9934 is so (for example) those in the US can disable those operations covered
9935 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9939 *) Non-copying interface to BIO pairs.
9940 (still largely untested)
9943 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9944 ASCII string. This was handled independently in various places before.
9947 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9948 UTF8 strings a character at a time.
9951 *) Use client_version from client hello to select the protocol
9952 (s23_srvr.c) and for RSA client key exchange verification
9953 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9956 *) Add various utility functions to handle SPKACs, these were previously
9957 handled by poking round in the structure internals. Added new function
9958 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9959 print, verify and generate SPKACs. Based on an original idea from
9960 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9963 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9966 *) Allow the config file extension section to be overwritten on the
9967 command line. Based on an original idea from Massimiliano Pala
9968 <madwolf@comune.modena.it>. The new option is called -extensions
9969 and can be applied to ca, req and x509. Also -reqexts to override
9970 the request extensions in req and -crlexts to override the crl extensions
9974 *) Add new feature to the SPKAC handling in ca. Now you can include
9975 the same field multiple times by preceding it by "XXXX." for example:
9978 this is the same syntax as used in the req config file.
9981 *) Allow certificate extensions to be added to certificate requests. These
9982 are specified in a 'req_extensions' option of the req section of the
9983 config file. They can be printed out with the -text option to req but
9984 are otherwise ignored at present.
9987 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9988 data read consists of only the final block it would not decrypted because
9989 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9990 A misplaced 'break' also meant the decrypted final block might not be
9991 copied until the next read.
9994 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9995 a few extra parameters to the DH structure: these will be useful if
9996 for example we want the value of 'q' or implement X9.42 DH.
9999 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
10000 provides hooks that allow the default DSA functions or functions on a
10001 "per key" basis to be replaced. This allows hardware acceleration and
10002 hardware key storage to be handled without major modification to the
10003 library. Also added low level modexp hooks and CRYPTO_EX structure and
10004 associated functions.
10007 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
10008 as "read only": it can't be written to and the buffer it points to will
10009 not be freed. Reading from a read only BIO is much more efficient than
10010 a normal memory BIO. This was added because there are several times when
10011 an area of memory needs to be read from a BIO. The previous method was
10012 to create a memory BIO and write the data to it, this results in two
10013 copies of the data and an O(n^2) reading algorithm. There is a new
10014 function BIO_new_mem_buf() which creates a read only memory BIO from
10015 an area of memory. Also modified the PKCS#7 routines to use read only
10019 *) Bugfix: ssl23_get_client_hello did not work properly when called in
10020 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
10021 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
10022 but a retry condition occured while trying to read the rest.
10025 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
10026 NID_pkcs7_encrypted by default: this was wrong since this should almost
10027 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
10028 the encrypted data type: this is a more sensible place to put it and it
10029 allows the PKCS#12 code to be tidied up that duplicated this
10033 *) Changed obj_dat.pl script so it takes its input and output files on
10034 the command line. This should avoid shell escape redirection problems
10038 *) Initial support for certificate extension requests, these are included
10039 in things like Xenroll certificate requests. Included functions to allow
10040 extensions to be obtained and added.
10043 *) -crlf option to s_client and s_server for sending newlines as
10044 CRLF (as required by many protocols).
10047 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
10049 *) Install libRSAglue.a when OpenSSL is built with RSAref.
10050 [Ralf S. Engelschall]
10052 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
10053 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
10055 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
10059 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
10060 DH parameters/keys (q is lost during that conversion, but the resulting
10061 DH parameters contain its length).
10063 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
10064 much faster than DH_generate_parameters (which creates parameters
10065 where p = 2*q + 1), and also the smaller q makes DH computations
10066 much more efficient (160-bit exponentiation instead of 1024-bit
10067 exponentiation); so this provides a convenient way to support DHE
10068 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
10069 utter importance to use
10070 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10072 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10073 when such DH parameters are used, because otherwise small subgroup
10074 attacks may become possible!
10077 *) Avoid memory leak in i2d_DHparams.
10080 *) Allow the -k option to be used more than once in the enc program:
10081 this allows the same encrypted message to be read by multiple recipients.
10084 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
10085 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
10086 it will always use the numerical form of the OID, even if it has a short
10090 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
10091 method only got called if p,q,dmp1,dmq1,iqmp components were present,
10092 otherwise bn_mod_exp was called. In the case of hardware keys for example
10093 no private key components need be present and it might store extra data
10094 in the RSA structure, which cannot be accessed from bn_mod_exp.
10095 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
10096 private key operations.
10099 *) Added support for SPARC Linux.
10102 *) pem_password_cb function type incompatibly changed from
10103 typedef int pem_password_cb(char *buf, int size, int rwflag);
10105 ....(char *buf, int size, int rwflag, void *userdata);
10106 so that applications can pass data to their callbacks:
10107 The PEM[_ASN1]_{read,write}... functions and macros now take an
10108 additional void * argument, which is just handed through whenever
10109 the password callback is called.
10110 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
10112 New function SSL_CTX_set_default_passwd_cb_userdata.
10114 Compatibility note: As many C implementations push function arguments
10115 onto the stack in reverse order, the new library version is likely to
10116 interoperate with programs that have been compiled with the old
10117 pem_password_cb definition (PEM_whatever takes some data that
10118 happens to be on the stack as its last argument, and the callback
10119 just ignores this garbage); but there is no guarantee whatsoever that
10122 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
10123 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
10124 problems not only on Windows, but also on some Unix platforms.
10125 To avoid problematic command lines, these definitions are now in an
10126 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
10127 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
10130 *) MIPS III/IV assembler module is reimplemented.
10133 *) More DES library cleanups: remove references to srand/rand and
10134 delete an unused file.
10137 *) Add support for the the free Netwide assembler (NASM) under Win32,
10138 since not many people have MASM (ml) and it can be hard to obtain.
10139 This is currently experimental but it seems to work OK and pass all
10140 the tests. Check out INSTALL.W32 for info.
10143 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
10144 without temporary keys kept an extra copy of the server key,
10145 and connections with temporary keys did not free everything in case
10149 *) New function RSA_check_key and new openssl rsa option -check
10150 for verifying the consistency of RSA keys.
10151 [Ulf Moeller, Bodo Moeller]
10153 *) Various changes to make Win32 compile work:
10154 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
10155 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
10156 comparison" warnings.
10157 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
10160 *) Add a debugging option to PKCS#5 v2 key generation function: when
10161 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
10162 derived keys are printed to stderr.
10165 *) Copy the flags in ASN1_STRING_dup().
10166 [Roman E. Pavlov <pre@mo.msk.ru>]
10168 *) The x509 application mishandled signing requests containing DSA
10169 keys when the signing key was also DSA and the parameters didn't match.
10171 It was supposed to omit the parameters when they matched the signing key:
10172 the verifying software was then supposed to automatically use the CA's
10173 parameters if they were absent from the end user certificate.
10175 Omitting parameters is no longer recommended. The test was also
10176 the wrong way round! This was probably due to unusual behaviour in
10177 EVP_cmp_parameters() which returns 1 if the parameters match.
10178 This meant that parameters were omitted when they *didn't* match and
10179 the certificate was useless. Certificates signed with 'ca' didn't have
10181 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
10183 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
10184 The interface is as follows:
10185 Applications can use
10186 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
10187 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
10188 "off" is now the default.
10189 The library internally uses
10190 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
10191 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
10192 to disable memory-checking temporarily.
10194 Some inconsistent states that previously were possible (and were
10195 even the default) are now avoided.
10197 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
10198 with each memory chunk allocated; this is occasionally more helpful
10199 than just having a counter.
10201 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
10203 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
10207 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
10208 which largely parallels "options", but is for changing API behaviour,
10209 whereas "options" are about protocol behaviour.
10210 Initial "mode" flags are:
10212 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
10213 a single record has been written.
10214 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
10215 retries use the same buffer location.
10216 (But all of the contents must be
10220 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10223 *) Fix problems with no-hmac etc.
10224 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10226 *) New functions RSA_get_default_method(), RSA_set_method() and
10227 RSA_get_method(). These allows replacement of RSA_METHODs without having
10228 to mess around with the internals of an RSA structure.
10231 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10232 Also really enable memory leak checks in openssl.c and in some
10234 [Chad C. Mulligan, Bodo Moeller]
10236 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10237 up the length of negative integers. This has now been simplified to just
10238 store the length when it is first determined and use it later, rather
10239 than trying to keep track of where data is copied and updating it to
10241 [Steve Henson, reported by Brien Wheeler
10242 <bwheeler@authentica-security.com>]
10244 *) Add a new function PKCS7_signatureVerify. This allows the verification
10245 of a PKCS#7 signature but with the signing certificate passed to the
10246 function itself. This contrasts with PKCS7_dataVerify which assumes the
10247 certificate is present in the PKCS#7 structure. This isn't always the
10248 case: certificates can be omitted from a PKCS#7 structure and be
10249 distributed by "out of band" means (such as a certificate database).
10252 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10253 function prototypes in pem.h, also change util/mkdef.pl to add the
10254 necessary function names.
10257 *) mk1mf.pl (used by Windows builds) did not properly read the
10258 options set by Configure in the top level Makefile, and Configure
10259 was not even able to write more than one option correctly.
10260 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10263 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10264 file to be loaded from a BIO or FILE pointer. The BIO version will
10265 for example allow memory BIOs to contain config info.
10268 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10269 Whoever hopes to achieve shared-library compatibility across versions
10270 must use this, not the compile-time macro.
10271 (Exercise 0.9.4: Which is the minimum library version required by
10273 Note: All this applies only to multi-threaded programs, others don't
10277 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10278 through a BIO pair triggered the default case, i.e.
10279 SSLerr(...,SSL_R_UNKNOWN_STATE).
10282 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10283 can use the SSL library even if none of the specific BIOs is
10287 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10288 for the encoded length.
10289 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10291 *) Add initial documentation of the X509V3 functions.
10294 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10295 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10296 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10297 secure PKCS#8 private key format with a high iteration count.
10300 *) Fix determination of Perl interpreter: A perl or perl5
10301 _directory_ in $PATH was also accepted as the interpreter.
10302 [Ralf S. Engelschall]
10304 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10305 wrong with it but it was very old and did things like calling
10306 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10307 unusual formatting.
10310 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10311 to use the new extension code.
10314 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10315 with macros. This should make it easier to change their form, add extra
10316 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10320 *) Add to configuration table a new entry that can specify an alternative
10321 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10322 according to Mark Crispin <MRC@Panda.COM>.
10326 *) DES CBC did not update the IV. Weird.
10329 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10330 Changing the behaviour of the former might break existing programs --
10331 where IV updating is needed, des_ncbc_encrypt can be used.
10334 *) When bntest is run from "make test" it drives bc to check its
10335 calculations, as well as internally checking them. If an internal check
10336 fails, it needs to cause bc to give a non-zero result or make test carries
10337 on without noticing the failure. Fixed.
10340 *) DES library cleanups.
10343 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10344 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10345 ciphers. NOTE: although the key derivation function has been verified
10346 against some published test vectors it has not been extensively tested
10347 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10351 *) Instead of "mkdir -p", which is not fully portable, use new
10352 Perl script "util/mkdir-p.pl".
10355 *) Rewrite the way password based encryption (PBE) is handled. It used to
10356 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10357 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10358 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10359 the 'parameter' field of the AlgorithmIdentifier is passed to the
10360 underlying key generation function so it must do its own ASN1 parsing.
10361 This has also changed the EVP_PBE_CipherInit() function which now has a
10362 'parameter' argument instead of literal salt and iteration count values
10363 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10366 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10367 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10368 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10369 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10370 value was just used as a "magic string" and not used directly its
10371 value doesn't matter.
10374 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10378 *) "linux-sparc64" configuration (ultrapenguin).
10379 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10380 "linux-sparc" configuration.
10381 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10383 *) config now generates no-xxx options for missing ciphers.
10386 *) Support the EBCDIC character set (work in progress).
10387 File ebcdic.c not yet included because it has a different license.
10388 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10390 *) Support BS2000/OSD-POSIX.
10391 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10393 *) Make callbacks for key generation use void * instead of char *.
10396 *) Make S/MIME samples compile (not yet tested).
10399 *) Additional typesafe stacks.
10402 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10406 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10408 *) New configuration variant "sco5-gcc".
10410 *) Updated some demos.
10411 [Sean O Riordain, Wade Scholine]
10413 *) Add missing BIO_free at exit of pkcs12 application.
10416 *) Fix memory leak in conf.c.
10419 *) Updates for Win32 to assembler version of MD5.
10422 *) Set #! path to perl in apps/der_chop to where we found it
10423 instead of using a fixed path.
10426 *) SHA library changes for irix64-mips4-cc.
10429 *) Improvements for VMS support.
10433 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10435 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10436 This also avoids the problems with SC4.2 and unpatched SC5.
10437 [Andy Polyakov <appro@fy.chalmers.se>]
10439 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10440 These are required because of the typesafe stack would otherwise break
10441 existing code. If old code used a structure member which used to be STACK
10442 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10443 sk_num or sk_value it would produce an error because the num, data members
10444 are not present in STACK_OF. Now it just produces a warning. sk_set
10445 replaces the old method of assigning a value to sk_value
10446 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10447 that does this will no longer work (and should use sk_set instead) but
10448 this could be regarded as a "questionable" behaviour anyway.
10451 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10452 correctly handle encrypted S/MIME data.
10455 *) Change type of various DES function arguments from des_cblock
10456 (which means, in function argument declarations, pointer to char)
10457 to des_cblock * (meaning pointer to array with 8 char elements),
10458 which allows the compiler to do more typechecking; it was like
10459 that back in SSLeay, but with lots of ugly casts.
10461 Introduce new type const_des_cblock.
10464 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10465 problems: find RecipientInfo structure that matches recipient certificate
10466 and initialise the ASN1 structures properly based on passed cipher.
10469 *) Belatedly make the BN tests actually check the results.
10472 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10473 to and from BNs: it was completely broken. New compilation option
10474 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10475 key elements as negative integers.
10478 *) Reorganize and speed up MD5.
10479 [Andy Polyakov <appro@fy.chalmers.se>]
10482 [Richard Levitte <richard@levitte.org>]
10484 *) New option -out to asn1parse to allow the parsed structure to be
10485 output to a file. This is most useful when combined with the -strparse
10486 option to examine the output of things like OCTET STRINGS.
10489 *) Make SSL library a little more fool-proof by not requiring any longer
10490 that SSL_set_{accept,connect}_state be called before
10491 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10492 in many applications because usually everything *appeared* to work as
10493 intended anyway -- now it really works as intended).
10496 *) Move openssl.cnf out of lib/.
10499 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10500 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10501 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10502 [Ralf S. Engelschall]
10504 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10505 handle PKCS#7 enveloped data properly.
10506 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10508 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10509 copying pointers. The cert_st handling is changed by this in
10510 various ways (and thus what used to be known as ctx->default_cert
10511 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10512 any longer when s->cert does not give us what we need).
10513 ssl_cert_instantiate becomes obsolete by this change.
10514 As soon as we've got the new code right (possibly it already is?),
10515 we have solved a couple of bugs of the earlier code where s->cert
10516 was used as if it could not have been shared with other SSL structures.
10518 Note that using the SSL API in certain dirty ways now will result
10519 in different behaviour than observed with earlier library versions:
10520 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10521 does not influence s as it used to.
10523 In order to clean up things more thoroughly, inside SSL_SESSION
10524 we don't use CERT any longer, but a new structure SESS_CERT
10525 that holds per-session data (if available); currently, this is
10526 the peer's certificate chain and, for clients, the server's certificate
10527 and temporary key. CERT holds only those values that can have
10528 meaningful defaults in an SSL_CTX.
10531 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10532 from the internal representation. Various PKCS#7 fixes: remove some
10533 evil casts and set the enc_dig_alg field properly based on the signing
10537 *) Allow PKCS#12 password to be set from the command line or the
10538 environment. Let 'ca' get its config file name from the environment
10539 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10543 *) Allow certificate policies extension to use an IA5STRING for the
10544 organization field. This is contrary to the PKIX definition but
10545 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10549 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10550 without disallowing inline assembler and the like for non-pedantic builds.
10553 *) Support Borland C++ builder.
10554 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10556 *) Support Mingw32.
10559 *) SHA-1 cleanups and performance enhancements.
10560 [Andy Polyakov <appro@fy.chalmers.se>]
10562 *) Sparc v8plus assembler for the bignum library.
10563 [Andy Polyakov <appro@fy.chalmers.se>]
10565 *) Accept any -xxx and +xxx compiler options in Configure.
10568 *) Update HPUX configuration.
10571 *) Add missing sk_<type>_unshift() function to safestack.h
10572 [Ralf S. Engelschall]
10574 *) New function SSL_CTX_use_certificate_chain_file that sets the
10575 "extra_cert"s in addition to the certificate. (This makes sense
10576 only for "PEM" format files, as chains as a whole are not
10580 *) Support verify_depth from the SSL API.
10581 x509_vfy.c had what can be considered an off-by-one-error:
10582 Its depth (which was not part of the external interface)
10583 was actually counting the number of certificates in a chain;
10584 now it really counts the depth.
10587 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10588 instead of X509err, which often resulted in confusing error
10589 messages since the error codes are not globally unique
10590 (e.g. an alleged error in ssl3_accept when a certificate
10591 didn't match the private key).
10593 *) New function SSL_CTX_set_session_id_context that allows to set a default
10594 value (so that you don't need SSL_set_session_id_context for each
10595 connection using the SSL_CTX).
10598 *) OAEP decoding bug fix.
10601 *) Support INSTALL_PREFIX for package builders, as proposed by
10605 *) New Configure options "threads" and "no-threads". For systems
10606 where the proper compiler options are known (currently Solaris
10607 and Linux), "threads" is the default.
10610 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10613 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10614 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10615 such as /usr/local/bin.
10618 *) "make linux-shared" to build shared libraries.
10619 [Niels Poppe <niels@netbox.org>]
10621 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10624 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10625 extension adding in x509 utility.
10628 *) Remove NOPROTO sections and error code comments.
10631 *) Partial rewrite of the DEF file generator to now parse the ANSI
10635 *) New Configure options --prefix=DIR and --openssldir=DIR.
10638 *) Complete rewrite of the error code script(s). It is all now handled
10639 by one script at the top level which handles error code gathering,
10640 header rewriting and C source file generation. It should be much better
10641 than the old method: it now uses a modified version of Ulf's parser to
10642 read the ANSI prototypes in all header files (thus the old K&R definitions
10643 aren't needed for error creation any more) and do a better job of
10644 translating function codes into names. The old 'ASN1 error code imbedded
10645 in a comment' is no longer necessary and it doesn't use .err files which
10646 have now been deleted. Also the error code call doesn't have to appear all
10647 on one line (which resulted in some large lines...).
10650 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10653 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10654 0 (which usually indicates a closed connection), but continue reading.
10657 *) Fix some race conditions.
10660 *) Add support for CRL distribution points extension. Add Certificate
10661 Policies and CRL distribution points documentation.
10664 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10667 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10668 8 of keying material. Merlin has also confirmed interop with this fix
10669 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10670 [Merlin Hughes <merlin@baltimore.ie>]
10672 *) Fix lots of warnings.
10673 [Richard Levitte <levitte@stacken.kth.se>]
10675 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10676 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10677 [Richard Levitte <levitte@stacken.kth.se>]
10679 *) Fix problems with sizeof(long) == 8.
10680 [Andy Polyakov <appro@fy.chalmers.se>]
10682 *) Change functions to ANSI C.
10685 *) Fix typos in error codes.
10686 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10688 *) Remove defunct assembler files from Configure.
10691 *) SPARC v8 assembler BIGNUM implementation.
10692 [Andy Polyakov <appro@fy.chalmers.se>]
10694 *) Support for Certificate Policies extension: both print and set.
10695 Various additions to support the r2i method this uses.
10698 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10699 return a const string when you are expecting an allocated buffer.
10702 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10703 types DirectoryString and DisplayText.
10706 *) Add code to allow r2i extensions to access the configuration database,
10707 add an LHASH database driver and add several ctx helper functions.
10710 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10711 fail when they extended the size of a BIGNUM.
10714 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10715 support typesafe stack.
10718 *) Fix typo in SSL_[gs]et_options().
10719 [Nils Frostberg <nils@medcom.se>]
10721 *) Delete various functions and files that belonged to the (now obsolete)
10722 old X509V3 handling code.
10725 *) New Configure option "rsaref".
10728 *) Don't auto-generate pem.h.
10731 *) Introduce type-safe ASN.1 SETs.
10734 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10735 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10737 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10738 that links with OpenSSL (well at least cause lots of warnings), but fear
10739 not: the conversion is trivial, and it eliminates loads of evil casts. A
10740 few STACKed things have been converted already. Feel free to convert more.
10741 In the fullness of time, I'll do away with the STACK type altogether.
10744 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10745 specified in <certfile> by updating the entry in the index.txt file.
10746 This way one no longer has to edit the index.txt file manually for
10747 revoking a certificate. The -revoke option does the gory details now.
10748 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10750 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10751 `-text' option at all and this way the `-noout -text' combination was
10752 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10753 [Ralf S. Engelschall]
10755 *) Make sure a corresponding plain text error message exists for the
10756 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10757 verify callback function determined that a certificate was revoked.
10758 [Ralf S. Engelschall]
10760 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10761 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10762 all available cipers including rc5, which was forgotten until now.
10763 In order to let the testing shell script know which algorithms
10764 are available, a new (up to now undocumented) command
10765 "openssl list-cipher-commands" is used.
10768 *) Bugfix: s_client occasionally would sleep in select() when
10769 it should have checked SSL_pending() first.
10772 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10773 the raw DSA values prior to ASN.1 encoding.
10776 *) Tweaks to Configure
10777 [Niels Poppe <niels@netbox.org>]
10779 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10783 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10786 *) New config option to avoid instructions that are illegal on the 80386.
10787 The default code is faster, but requires at least a 486.
10790 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10791 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10792 same as SSL2_VERSION anyway.
10795 *) New "-showcerts" option for s_client.
10798 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10799 application. Various cleanups and fixes.
10802 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10803 modify error routines to work internally. Add error codes and PBE init
10804 to library startup routines.
10807 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10808 packing functions to asn1 and evp. Changed function names and error
10809 codes along the way.
10812 *) PKCS12 integration: and so it begins... First of several patches to
10813 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10814 objects to objects.h
10817 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10818 and display support for Thawte strong extranet extension.
10821 *) Add LinuxPPC support.
10822 [Jeff Dubrule <igor@pobox.org>]
10824 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10825 bn_div_words in alpha.s.
10826 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10828 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10829 OAEP isn't supported when OpenSSL is built with RSAref.
10830 [Ulf Moeller <ulf@fitug.de>]
10832 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10833 so they no longer are missing under -DNOPROTO.
10834 [Soren S. Jorvang <soren@t.dk>]
10837 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10839 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10840 doesn't work when the session is reused. Coming soon!
10843 *) Fix a security hole, that allows sessions to be reused in the wrong
10844 context thus bypassing client cert protection! All software that uses
10845 client certs and session caches in multiple contexts NEEDS PATCHING to
10846 allow session reuse! A fuller solution is in the works.
10847 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10849 *) Some more source tree cleanups (removed obsolete files
10850 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10851 permission on "config" script to be executable) and a fix for the INSTALL
10853 [Ulf Moeller <ulf@fitug.de>]
10855 *) Remove some legacy and erroneous uses of malloc, free instead of
10857 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10859 *) Make rsa_oaep_test return non-zero on error.
10860 [Ulf Moeller <ulf@fitug.de>]
10862 *) Add support for native Solaris shared libraries. Configure
10863 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10864 if someone would make that last step automatic.
10865 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10867 *) ctx_size was not built with the right compiler during "make links". Fixed.
10870 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10871 except NULL ciphers". This means the default cipher list will no longer
10872 enable NULL ciphers. They need to be specifically enabled e.g. with
10873 the string "DEFAULT:eNULL".
10876 *) Fix to RSA private encryption routines: if p < q then it would
10877 occasionally produce an invalid result. This will only happen with
10878 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10881 *) Be less restrictive and allow also `perl util/perlpath.pl
10882 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10883 because this way one can also use an interpreter named `perl5' (which is
10884 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10885 installed as `perl').
10886 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10888 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10889 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10891 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10892 advapi32.lib to Win32 build and change the pem test comparision
10893 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10894 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10895 and crypto/des/ede_cbcm_enc.c.
10898 *) DES quad checksum was broken on big-endian architectures. Fixed.
10901 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10902 Win32 test batch file so it (might) work again. The Win32 test batch file
10903 is horrible: I feel ill....
10906 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10907 in e_os.h. Audit of header files to check ANSI and non ANSI
10908 sections: 10 functions were absent from non ANSI section and not exported
10909 from Windows DLLs. Fixed up libeay.num for new functions.
10912 *) Make `openssl version' output lines consistent.
10913 [Ralf S. Engelschall]
10915 *) Fix Win32 symbol export lists for BIO functions: Added
10916 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10917 to ms/libeay{16,32}.def.
10918 [Ralf S. Engelschall]
10920 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10921 fine under Unix and passes some trivial tests I've now added. But the
10922 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10923 added to make sure no one expects that this stuff really works in the
10924 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10925 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10927 [Ralf S. Engelschall]
10929 *) Fix the generation of two part addresses in perl.
10930 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10932 *) Add config entry for Linux on MIPS.
10933 [John Tobey <jtobey@channel1.com>]
10935 *) Make links whenever Configure is run, unless we are on Windoze.
10938 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10939 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10943 *) Add a useful kludge to allow package maintainers to specify compiler and
10944 other platforms details on the command line without having to patch the
10945 Configure script everytime: One now can use ``perl Configure
10946 <id>:<details>'', i.e. platform ids are allowed to have details appended
10947 to them (seperated by colons). This is treated as there would be a static
10948 pre-configured entry in Configure's %table under key <id> with value
10949 <details> and ``perl Configure <id>'' is called. So, when you want to
10950 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10951 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10952 now, which overrides the FreeBSD-elf entry on-the-fly.
10953 [Ralf S. Engelschall]
10955 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10958 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10959 on the `perl Configure ...' command line. This way one can compile
10960 OpenSSL libraries with Position Independent Code (PIC) which is needed
10961 for linking it into DSOs.
10962 [Ralf S. Engelschall]
10964 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10968 *) Cleaned up the LICENSE document: The official contact for any license
10969 questions now is the OpenSSL core team under openssl-core@openssl.org.
10970 And add a paragraph about the dual-license situation to make sure people
10971 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10972 to the OpenSSL toolkit.
10973 [Ralf S. Engelschall]
10975 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10976 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10977 Additonally cleaned up the `make links' target: Remove unnecessary
10978 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10979 to speed processing and no longer clutter the display with confusing
10980 stuff. Instead only the actually done links are displayed.
10981 [Ralf S. Engelschall]
10983 *) Permit null encryption ciphersuites, used for authentication only. It used
10984 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10985 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10989 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10990 signed attributes when verifying signatures (this would break them),
10991 the detached data encoding was wrong and public keys obtained using
10992 X509_get_pubkey() weren't freed.
10995 *) Add text documentation for the BUFFER functions. Also added a work around
10996 to a Win95 console bug. This was triggered by the password read stuff: the
10997 last character typed gets carried over to the next fread(). If you were
10998 generating a new cert request using 'req' for example then the last
10999 character of the passphrase would be CR which would then enter the first
11003 *) Added the new `Includes OpenSSL Cryptography Software' button as
11004 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
11005 button and can be used by applications based on OpenSSL to show the
11006 relationship to the OpenSSL project.
11007 [Ralf S. Engelschall]
11009 *) Remove confusing variables in function signatures in files
11010 ssl/ssl_lib.c and ssl/ssl.h.
11011 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11013 *) Don't install bss_file.c under PREFIX/include/
11014 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11016 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
11017 functions that return function pointers and has support for NT specific
11018 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
11019 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
11020 unsigned to signed types: this was killing the Win32 compile.
11023 *) Add new certificate file to stack functions,
11024 SSL_add_dir_cert_subjects_to_stack() and
11025 SSL_add_file_cert_subjects_to_stack(). These largely supplant
11026 SSL_load_client_CA_file(), and can be used to add multiple certs easily
11027 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
11028 This means that Apache-SSL and similar packages don't have to mess around
11029 to add as many CAs as they want to the preferred list.
11032 *) Experiment with doxygen documentation. Currently only partially applied to
11034 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
11035 openssl.doxy as the configuration file.
11038 *) Get rid of remaining C++-style comments which strict C compilers hate.
11039 [Ralf S. Engelschall, pointed out by Carlos Amengual]
11041 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
11042 compiled in by default: it has problems with large keys.
11045 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
11046 DH private keys and/or callback functions which directly correspond to
11047 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
11048 is needed for applications which have to configure certificates on a
11049 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
11051 For the RSA certificate situation is makes no difference, but
11052 for the DSA certificate situation this fixes the "no shared cipher"
11053 problem where the OpenSSL cipher selection procedure failed because the
11054 temporary keys were not overtaken from the context and the API provided
11055 no way to reconfigure them.
11056 The new functions now let applications reconfigure the stuff and they
11057 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
11058 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
11059 non-public-API function ssl_cert_instantiate() is used as a helper
11060 function and also to reduce code redundancy inside ssl_rsa.c.
11061 [Ralf S. Engelschall]
11063 *) Move s_server -dcert and -dkey options out of the undocumented feature
11064 area because they are useful for the DSA situation and should be
11065 recognized by the users.
11066 [Ralf S. Engelschall]
11068 *) Fix the cipher decision scheme for export ciphers: the export bits are
11069 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
11070 SSL_EXP_MASK. So, the original variable has to be used instead of the
11071 already masked variable.
11072 [Richard Levitte <levitte@stacken.kth.se>]
11074 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
11075 [Richard Levitte <levitte@stacken.kth.se>]
11077 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
11078 from `int' to `unsigned int' because it's a length and initialized by
11079 EVP_DigestFinal() which expects an `unsigned int *'.
11080 [Richard Levitte <levitte@stacken.kth.se>]
11082 *) Don't hard-code path to Perl interpreter on shebang line of Configure
11083 script. Instead use the usual Shell->Perl transition trick.
11084 [Ralf S. Engelschall]
11086 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
11087 (in addition to RSA certificates) to match the behaviour of `openssl dsa
11088 -noout -modulus' as it's already the case for `openssl rsa -noout
11089 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
11090 currently the public key is printed (a decision which was already done by
11091 `openssl dsa -modulus' in the past) which serves a similar purpose.
11092 Additionally the NO_RSA no longer completely removes the whole -modulus
11093 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
11095 [Ralf S. Engelschall]
11097 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
11098 BIO. See the source (crypto/evp/bio_ok.c) for more info.
11099 [Arne Ansper <arne@ats.cyber.ee>]
11101 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
11102 to be added. Now both 'req' and 'ca' can use new objects defined in the
11106 *) Add cool BIO that does syslog (or event log on NT).
11107 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
11109 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
11110 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
11111 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
11112 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
11115 *) Add preliminary config info for new extension code.
11118 *) Make RSA_NO_PADDING really use no padding.
11119 [Ulf Moeller <ulf@fitug.de>]
11121 *) Generate errors when private/public key check is done.
11124 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
11125 for some CRL extensions and new objects added.
11128 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
11129 key usage extension and fuller support for authority key id.
11132 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
11133 padding method for RSA, which is recommended for new applications in PKCS
11134 #1 v2.0 (RFC 2437, October 1998).
11135 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
11136 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
11137 against Bleichbacher's attack on RSA.
11138 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
11141 *) Updates to the new SSL compression code
11142 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11144 *) Fix so that the version number in the master secret, when passed
11145 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
11146 (because the server will not accept higher), that the version number
11147 is 0x03,0x01, not 0x03,0x00
11148 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11150 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
11151 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
11152 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
11155 *) Support for RAW extensions where an arbitrary extension can be
11156 created by including its DER encoding. See apps/openssl.cnf for
11160 *) Make sure latest Perl versions don't interpret some generated C array
11161 code as Perl array code in the crypto/err/err_genc.pl script.
11162 [Lars Weber <3weber@informatik.uni-hamburg.de>]
11164 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
11165 not many people have the assembler. Various Win32 compilation fixes and
11166 update to the INSTALL.W32 file with (hopefully) more accurate Win32
11167 build instructions.
11170 *) Modify configure script 'Configure' to automatically create crypto/date.h
11171 file under Win32 and also build pem.h from pem.org. New script
11172 util/mkfiles.pl to create the MINFO file on environments that can't do a
11173 'make files': perl util/mkfiles.pl >MINFO should work.
11176 *) Major rework of DES function declarations, in the pursuit of correctness
11177 and purity. As a result, many evil casts evaporated, and some weirdness,
11178 too. You may find this causes warnings in your code. Zapping your evil
11179 casts will probably fix them. Mostly.
11182 *) Fix for a typo in asn1.h. Bug fix to object creation script
11183 obj_dat.pl. It considered a zero in an object definition to mean
11184 "end of object": none of the objects in objects.h have any zeros
11185 so it wasn't spotted.
11186 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
11188 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
11189 Masking (CBCM). In the absence of test vectors, the best I have been able
11190 to do is check that the decrypt undoes the encrypt, so far. Send me test
11191 vectors if you have them.
11194 *) Correct calculation of key length for export ciphers (too much space was
11195 allocated for null ciphers). This has not been tested!
11198 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
11199 message is now correct (it understands "crypto" and "ssl" on its
11200 command line). There is also now an "update" option. This will update
11201 the util/ssleay.num and util/libeay.num files with any new functions.
11203 perl util/mkdef.pl crypto ssl update
11204 it will update them.
11207 *) Overhauled the Perl interface (perl/*):
11208 - ported BN stuff to OpenSSL's different BN library
11209 - made the perl/ source tree CVS-aware
11210 - renamed the package from SSLeay to OpenSSL (the files still contain
11211 their history because I've copied them in the repository)
11212 - removed obsolete files (the test scripts will be replaced
11213 by better Test::Harness variants in the future)
11214 [Ralf S. Engelschall]
11216 *) First cut for a very conservative source tree cleanup:
11217 1. merge various obsolete readme texts into doc/ssleay.txt
11218 where we collect the old documents and readme texts.
11219 2. remove the first part of files where I'm already sure that we no
11220 longer need them because of three reasons: either they are just temporary
11221 files which were left by Eric or they are preserved original files where
11222 I've verified that the diff is also available in the CVS via "cvs diff
11223 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11224 the crypto/md/ stuff).
11225 [Ralf S. Engelschall]
11227 *) More extension code. Incomplete support for subject and issuer alt
11228 name, issuer and authority key id. Change the i2v function parameters
11229 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11230 what that's for :-) Fix to ASN1 macro which messed up
11231 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11234 *) Preliminary support for ENUMERATED type. This is largely copied from the
11238 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11239 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11241 *) Make sure `make rehash' target really finds the `openssl' program.
11242 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11244 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11245 like to hear about it if this slows down other processors.
11248 *) Add CygWin32 platform information to Configure script.
11249 [Alan Batie <batie@aahz.jf.intel.com>]
11251 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11252 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11254 *) New program nseq to manipulate netscape certificate sequences
11257 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11261 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11262 but the BN code had some problems that would cause failures when
11263 doing certificate verification and some other functions.
11264 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11266 *) Add ASN1 and PEM code to support netscape certificate sequences.
11269 *) Add ASN1 and PEM code to support netscape certificate sequences.
11272 *) Add several PKIX and private extended key usage OIDs.
11275 *) Modify the 'ca' program to handle the new extension code. Modify
11276 openssl.cnf for new extension format, add comments.
11279 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11280 and add a sample to openssl.cnf so req -x509 now adds appropriate
11284 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11285 error code, add initial support to X509_print() and x509 application.
11288 *) Takes a deep breath and start addding X509 V3 extension support code. Add
11289 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11290 stuff is currently isolated and isn't even compiled yet.
11293 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11294 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11295 Removed the versions check from X509 routines when loading extensions:
11296 this allows certain broken certificates that don't set the version
11297 properly to be processed.
11300 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11301 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11302 can still be regenerated with "make depend".
11305 *) Spelling mistake in C version of CAST-128.
11306 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11308 *) Changes to the error generation code. The perl script err-code.pl
11309 now reads in the old error codes and retains the old numbers, only
11310 adding new ones if necessary. It also only changes the .err files if new
11311 codes are added. The makefiles have been modified to only insert errors
11312 when needed (to avoid needlessly modifying header files). This is done
11313 by only inserting errors if the .err file is newer than the auto generated
11314 C file. To rebuild all the error codes from scratch (the old behaviour)
11315 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11316 or delete all the .err files.
11319 *) CAST-128 was incorrectly implemented for short keys. The C version has
11320 been fixed, but is untested. The assembler versions are also fixed, but
11321 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11322 to regenerate it if needed.
11323 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11324 Hagino <itojun@kame.net>]
11326 *) File was opened incorrectly in randfile.c.
11327 [Ulf Möller <ulf@fitug.de>]
11329 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11330 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11331 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11332 al: it's just almost always a UTCTime. Note this patch adds new error
11333 codes so do a "make errors" if there are problems.
11336 *) Correct Linux 1 recognition in config.
11337 [Ulf Möller <ulf@fitug.de>]
11339 *) Remove pointless MD5 hash when using DSA keys in ca.
11340 [Anonymous <nobody@replay.com>]
11342 *) Generate an error if given an empty string as a cert directory. Also
11343 generate an error if handed NULL (previously returned 0 to indicate an
11344 error, but didn't set one).
11345 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11347 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11350 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11351 parameters. This was causing a warning which killed off the Win32 compile.
11354 *) Remove C++ style comments from crypto/bn/bn_local.h.
11355 [Neil Costigan <neil.costigan@celocom.com>]
11357 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11358 based on a text string, looking up short and long names and finally
11359 "dot" format. The "dot" format stuff didn't work. Added new function
11360 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11361 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11362 OID is not part of the table.
11365 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11366 X509_LOOKUP_by_alias().
11369 *) Sort openssl functions by name.
11372 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11373 encryption from sample DSA keys (in case anyone is interested the password
11377 *) Make _all_ *_free functions accept a NULL pointer.
11378 [Frans Heymans <fheymans@isaserver.be>]
11380 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11382 [Anonymous <nobody@replay.com>]
11384 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11385 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11387 *) Don't blow it for numeric -newkey arguments to apps/req.
11388 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11390 *) Temp key "for export" tests were wrong in s3_srvr.c.
11391 [Anonymous <nobody@replay.com>]
11393 *) Add prototype for temp key callback functions
11394 SSL_CTX_set_tmp_{rsa,dh}_callback().
11397 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11398 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11401 *) X509_name_add_entry() freed the wrong thing after an error.
11402 [Arne Ansper <arne@ats.cyber.ee>]
11404 *) rsa_eay.c would attempt to free a NULL context.
11405 [Arne Ansper <arne@ats.cyber.ee>]
11407 *) BIO_s_socket() had a broken should_retry() on Windoze.
11408 [Arne Ansper <arne@ats.cyber.ee>]
11410 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11411 [Arne Ansper <arne@ats.cyber.ee>]
11413 *) Make sure the already existing X509_STORE->depth variable is initialized
11414 in X509_STORE_new(), but document the fact that this variable is still
11415 unused in the certificate verification process.
11416 [Ralf S. Engelschall]
11418 *) Fix the various library and apps files to free up pkeys obtained from
11419 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11422 *) Fix reference counting in X509_PUBKEY_get(). This makes
11423 demos/maurice/example2.c work, amongst others, probably.
11424 [Steve Henson and Ben Laurie]
11426 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11427 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11428 are no longer created. This way we have a single and consistent command
11429 line interface `openssl <command>', similar to `cvs <command>'.
11430 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11432 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11433 BIT STRING wrapper always have zero unused bits.
11436 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11439 *) Make the top-level INSTALL documentation easier to understand.
11442 *) Makefiles updated to exit if an error occurs in a sub-directory
11443 make (including if user presses ^C) [Paul Sutton]
11445 *) Make Montgomery context stuff explicit in RSA data structure.
11448 *) Fix build order of pem and err to allow for generated pem.h.
11451 *) Fix renumbering bug in X509_NAME_delete_entry().
11454 *) Enhanced the err-ins.pl script so it makes the error library number
11455 global and can add a library name. This is needed for external ASN1 and
11456 other error libraries.
11459 *) Fixed sk_insert which never worked properly.
11462 *) Fix ASN1 macros so they can handle indefinite length construted
11463 EXPLICIT tags. Some non standard certificates use these: they can now
11467 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11468 into a single doc/ssleay.txt bundle. This way the information is still
11469 preserved but no longer messes up this directory. Now it's new room for
11470 the new set of documenation files.
11471 [Ralf S. Engelschall]
11473 *) SETs were incorrectly DER encoded. This was a major pain, because they
11474 shared code with SEQUENCEs, which aren't coded the same. This means that
11475 almost everything to do with SETs or SEQUENCEs has either changed name or
11476 number of arguments.
11477 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11479 *) Fix test data to work with the above.
11482 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11483 was already fixed by Eric for 0.9.1 it seems.
11484 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11486 *) Autodetect FreeBSD3.
11489 *) Fix various bugs in Configure. This affects the following platforms:
11493 unixware-2.0-pentium
11497 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11498 before they are needed.
11501 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11505 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11507 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11508 changed SSLeay to OpenSSL in version strings.
11509 [Ralf S. Engelschall]
11511 *) Some fixups to the top-level documents.
11514 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11515 because the symlink to include/ was missing.
11516 [Ralf S. Engelschall]
11518 *) Incorporated the popular no-RSA/DSA-only patches
11519 which allow to compile a RSA-free SSLeay.
11520 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11522 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11523 when "ssleay" is still not found.
11524 [Ralf S. Engelschall]
11526 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11527 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11529 *) Updated the README file.
11530 [Ralf S. Engelschall]
11532 *) Added various .cvsignore files in the CVS repository subdirs
11533 to make a "cvs update" really silent.
11534 [Ralf S. Engelschall]
11536 *) Recompiled the error-definition header files and added
11537 missing symbols to the Win32 linker tables.
11538 [Ralf S. Engelschall]
11540 *) Cleaned up the top-level documents;
11541 o new files: CHANGES and LICENSE
11542 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11543 o merged COPYRIGHT into LICENSE
11544 o removed obsolete TODO file
11545 o renamed MICROSOFT to INSTALL.W32
11546 [Ralf S. Engelschall]
11548 *) Removed dummy files from the 0.9.1b source tree:
11549 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11550 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11551 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11552 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11553 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11554 [Ralf S. Engelschall]
11556 *) Added various platform portability fixes.
11559 *) The Genesis of the OpenSSL rpject:
11560 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11561 Young and Tim J. Hudson created while they were working for C2Net until
11563 [The OpenSSL Project]
11566 Changes between 0.9.0b and 0.9.1b [not released]
11568 *) Updated a few CA certificates under certs/
11571 *) Changed some BIGNUM api stuff.
11574 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11575 DGUX x86, Linux Alpha, etc.
11578 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11579 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11583 *) Add -strparse option to asn1pars program which parses nested
11585 [Dr Stephen Henson <shenson@bigfoot.com>]
11587 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11590 *) DSA fix for "ca" program.
11593 *) Added "-genkey" option to "dsaparam" program.
11596 *) Added RIPE MD160 (rmd160) message digest.
11599 *) Added -a (all) option to "ssleay version" command.
11602 *) Added PLATFORM define which is the id given to Configure.
11605 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11608 *) Extended the ASN.1 parser routines.
11611 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11614 *) Added a BN_CTX to the BN library.
11617 *) Fixed the weak key values in DES library
11620 *) Changed API in EVP library for cipher aliases.
11623 *) Added support for RC2/64bit cipher.
11626 *) Converted the lhash library to the crypto/mem.c functions.
11629 *) Added more recognized ASN.1 object ids.
11632 *) Added more RSA padding checks for SSL/TLS.
11635 *) Added BIO proxy/filter functionality.
11638 *) Added extra_certs to SSL_CTX which can be used
11639 send extra CA certificates to the client in the CA cert chain sending
11640 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11643 *) Now Fortezza is denied in the authentication phase because
11644 this is key exchange mechanism is not supported by SSLeay at all.
11647 *) Additional PKCS1 checks.
11650 *) Support the string "TLSv1" for all TLS v1 ciphers.
11653 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11654 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11657 *) Fixed a few memory leaks.
11660 *) Fixed various code and comment typos.
11663 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11664 bytes sent in the client random.
11665 [Edward Bishop <ebishop@spyglass.com>]