5 Changes between 1.0.2g and 1.0.2h [3 May 2016]
7 *) Prevent padding oracle in AES-NI CBC MAC check
9 A MITM attacker can use a padding oracle attack to decrypt traffic
10 when the connection uses an AES CBC cipher and the server support
13 This issue was introduced as part of the fix for Lucky 13 padding
14 attack (CVE-2013-0169). The padding check was rewritten to be in
15 constant time by making sure that always the same bytes are read and
16 compared against either the MAC or padding bytes. But it no longer
17 checked that there was enough data to have both the MAC and padding
20 This issue was reported by Juraj Somorovsky using TLS-Attacker.
24 *) Fix EVP_EncodeUpdate overflow
26 An overflow can occur in the EVP_EncodeUpdate() function which is used for
27 Base64 encoding of binary data. If an attacker is able to supply very large
28 amounts of input data then a length check can overflow resulting in a heap
31 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
32 the PEM_write_bio* family of functions. These are mainly used within the
33 OpenSSL command line applications, so any application which processes data
34 from an untrusted source and outputs it as a PEM file should be considered
35 vulnerable to this issue. User applications that call these APIs directly
36 with large amounts of untrusted data may also be vulnerable.
38 This issue was reported by Guido Vranken.
42 *) Fix EVP_EncryptUpdate overflow
44 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
45 is able to supply very large amounts of input data after a previous call to
46 EVP_EncryptUpdate() with a partial block then a length check can overflow
47 resulting in a heap corruption. Following an analysis of all OpenSSL
48 internal usage of the EVP_EncryptUpdate() function all usage is one of two
49 forms. The first form is where the EVP_EncryptUpdate() call is known to be
50 the first called function after an EVP_EncryptInit(), and therefore that
51 specific call must be safe. The second form is where the length passed to
52 EVP_EncryptUpdate() can be seen from the code to be some small value and
53 therefore there is no possibility of an overflow. Since all instances are
54 one of these two forms, it is believed that there can be no overflows in
55 internal code due to this problem. It should be noted that
56 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
57 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
58 of these calls have also been analysed too and it is believed there are no
59 instances in internal usage where an overflow could occur.
61 This issue was reported by Guido Vranken.
65 *) Prevent ASN.1 BIO excessive memory allocation
67 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
68 a short invalid encoding can casuse allocation of large amounts of memory
69 potentially consuming excessive resources or exhausting memory.
71 Any application parsing untrusted data through d2i BIO functions is
72 affected. The memory based functions such as d2i_X509() are *not* affected.
73 Since the memory based functions are used by the TLS library, TLS
74 applications are not affected.
76 This issue was reported by Brian Carpenter.
82 ASN1 Strings that are over 1024 bytes can cause an overread in applications
83 using the X509_NAME_oneline() function on EBCDIC systems. This could result
84 in arbitrary stack data being returned in the buffer.
86 This issue was reported by Guido Vranken.
90 *) Modify behavior of ALPN to invoke callback after SNI/servername
91 callback, such that updates to the SSL_CTX affect ALPN.
94 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
98 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
99 methods are enabled and ssl2 is disabled the methods return NULL.
102 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
104 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
105 Builds that are not configured with "enable-weak-ssl-ciphers" will not
106 provide any "EXPORT" or "LOW" strength ciphers.
109 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
110 is by default disabled at build-time. Builds that are not configured with
111 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
112 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
113 will need to explicitly call either of:
115 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
117 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
119 as appropriate. Even if either of those is used, or the application
120 explicitly uses the version-specific SSLv2_method() or its client and
121 server variants, SSLv2 ciphers vulnerable to exhaustive search key
122 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
123 ciphers, and SSLv2 56-bit DES are no longer available.
127 *) Fix a double-free in DSA code
129 A double free bug was discovered when OpenSSL parses malformed DSA private
130 keys and could lead to a DoS attack or memory corruption for applications
131 that receive DSA private keys from untrusted sources. This scenario is
134 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
139 *) Disable SRP fake user seed to address a server memory leak.
141 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
143 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
144 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
145 was changed to ignore the "fake user" SRP seed, even if the seed
148 Users should use SRP_VBASE_get1_by_user instead. Note that in
149 SRP_VBASE_get1_by_user, caller must free the returned value. Note
150 also that even though configuring the SRP seed attempts to hide
151 invalid usernames by continuing the handshake with fake
152 credentials, this behaviour is not constant time and no strong
153 guarantees are made that the handshake is indistinguishable from
154 that of a valid user.
158 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
160 In the BN_hex2bn function the number of hex digits is calculated using an
161 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
162 large values of |i| this can result in |bn_expand| not allocating any
163 memory because |i * 4| is negative. This can leave the internal BIGNUM data
164 field as NULL leading to a subsequent NULL ptr deref. For very large values
165 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
166 In this case memory is allocated to the internal BIGNUM data field, but it
167 is insufficiently sized leading to heap corruption. A similar issue exists
168 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
169 is ever called by user applications with very large untrusted hex/dec data.
170 This is anticipated to be a rare occurrence.
172 All OpenSSL internal usage of these functions use data that is not expected
173 to be untrusted, e.g. config file data or application command line
174 arguments. If user developed applications generate config file data based
175 on untrusted data then it is possible that this could also lead to security
176 consequences. This is also anticipated to be rare.
178 This issue was reported to OpenSSL by Guido Vranken.
182 *) Fix memory issues in BIO_*printf functions
184 The internal |fmtstr| function used in processing a "%s" format string in
185 the BIO_*printf functions could overflow while calculating the length of a
186 string and cause an OOB read when printing very long strings.
188 Additionally the internal |doapr_outch| function can attempt to write to an
189 OOB memory location (at an offset from the NULL pointer) in the event of a
190 memory allocation failure. In 1.0.2 and below this could be caused where
191 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
192 could be in processing a very long "%s" format string. Memory leaks can
195 The first issue may mask the second issue dependent on compiler behaviour.
196 These problems could enable attacks where large amounts of untrusted data
197 is passed to the BIO_*printf functions. If applications use these functions
198 in this way then they could be vulnerable. OpenSSL itself uses these
199 functions when printing out human-readable dumps of ASN.1 data. Therefore
200 applications that print this data could be vulnerable if the data is from
201 untrusted sources. OpenSSL command line applications could also be
202 vulnerable where they print out ASN.1 data, or if untrusted data is passed
203 as command line arguments.
205 Libssl is not considered directly vulnerable. Additionally certificates etc
206 received via remote connections via libssl are also unlikely to be able to
207 trigger these issues because of message size limits enforced within libssl.
209 This issue was reported to OpenSSL Guido Vranken.
213 *) Side channel attack on modular exponentiation
215 A side-channel attack was found which makes use of cache-bank conflicts on
216 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
217 of RSA keys. The ability to exploit this issue is limited as it relies on
218 an attacker who has control of code in a thread running on the same
219 hyper-threaded core as the victim thread which is performing decryptions.
221 This issue was reported to OpenSSL by Yuval Yarom, The University of
222 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
223 Nadia Heninger, University of Pennsylvania with more information at
224 http://cachebleed.info.
228 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
229 if no keysize is specified with default_bits. This fixes an
230 omission in an earlier change that changed all RSA/DSA key generation
231 apps to use 2048 bits by default.
234 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
236 *) DH small subgroups
238 Historically OpenSSL only ever generated DH parameters based on "safe"
239 primes. More recently (in version 1.0.2) support was provided for
240 generating X9.42 style parameter files such as those required for RFC 5114
241 support. The primes used in such files may not be "safe". Where an
242 application is using DH configured with parameters based on primes that are
243 not "safe" then an attacker could use this fact to find a peer's private
244 DH exponent. This attack requires that the attacker complete multiple
245 handshakes in which the peer uses the same private DH exponent. For example
246 this could be used to discover a TLS server's private DH exponent if it's
247 reusing the private DH exponent or it's using a static DH ciphersuite.
249 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
250 TLS. It is not on by default. If the option is not set then the server
251 reuses the same private DH exponent for the life of the server process and
252 would be vulnerable to this attack. It is believed that many popular
253 applications do set this option and would therefore not be at risk.
255 The fix for this issue adds an additional check where a "q" parameter is
256 available (as is the case in X9.42 based parameters). This detects the
257 only known attack, and is the only possible defense for static DH
258 ciphersuites. This could have some performance impact.
260 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
261 default and cannot be disabled. This could have some performance impact.
263 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
267 *) SSLv2 doesn't block disabled ciphers
269 A malicious client can negotiate SSLv2 ciphers that have been disabled on
270 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
271 been disabled, provided that the SSLv2 protocol was not also disabled via
274 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
275 and Sebastian Schinzel.
279 *) Reject DH handshakes with parameters shorter than 1024 bits.
282 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
284 *) BN_mod_exp may produce incorrect results on x86_64
286 There is a carry propagating bug in the x86_64 Montgomery squaring
287 procedure. No EC algorithms are affected. Analysis suggests that attacks
288 against RSA and DSA as a result of this defect would be very difficult to
289 perform and are not believed likely. Attacks against DH are considered just
290 feasible (although very difficult) because most of the work necessary to
291 deduce information about a private key may be performed offline. The amount
292 of resources required for such an attack would be very significant and
293 likely only accessible to a limited number of attackers. An attacker would
294 additionally need online access to an unpatched system using the target
295 private key in a scenario with persistent DH parameters and a private
296 key that is shared between multiple clients. For example this can occur by
297 default in OpenSSL DHE based SSL/TLS ciphersuites.
299 This issue was reported to OpenSSL by Hanno Böck.
303 *) Certificate verify crash with missing PSS parameter
305 The signature verification routines will crash with a NULL pointer
306 dereference if presented with an ASN.1 signature using the RSA PSS
307 algorithm and absent mask generation function parameter. Since these
308 routines are used to verify certificate signature algorithms this can be
309 used to crash any certificate verification operation and exploited in a
310 DoS attack. Any application which performs certificate verification is
311 vulnerable including OpenSSL clients and servers which enable client
314 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
318 *) X509_ATTRIBUTE memory leak
320 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
321 memory. This structure is used by the PKCS#7 and CMS routines so any
322 application which reads PKCS#7 or CMS data from untrusted sources is
323 affected. SSL/TLS is not affected.
325 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
330 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
331 This changes the decoding behaviour for some invalid messages,
332 though the change is mostly in the more lenient direction, and
333 legacy behaviour is preserved as much as possible.
336 *) In DSA_generate_parameters_ex, if the provided seed is too short,
337 use a random seed, as already documented.
338 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
340 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
342 *) Alternate chains certificate forgery
344 During certificate verfification, OpenSSL will attempt to find an
345 alternative certificate chain if the first attempt to build such a chain
346 fails. An error in the implementation of this logic can mean that an
347 attacker could cause certain checks on untrusted certificates to be
348 bypassed, such as the CA flag, enabling them to use a valid leaf
349 certificate to act as a CA and "issue" an invalid certificate.
351 This issue was reported to OpenSSL by Adam Langley/David Benjamin
356 *) Race condition handling PSK identify hint
358 If PSK identity hints are received by a multi-threaded client then
359 the values are wrongly updated in the parent SSL_CTX structure. This can
360 result in a race condition potentially leading to a double free of the
365 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
367 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
368 incompatibility in the handling of HMAC. The previous ABI has now been
371 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
373 *) Malformed ECParameters causes infinite loop
375 When processing an ECParameters structure OpenSSL enters an infinite loop
376 if the curve specified is over a specially malformed binary polynomial
379 This can be used to perform denial of service against any
380 system which processes public keys, certificate requests or
381 certificates. This includes TLS clients and TLS servers with
382 client authentication enabled.
384 This issue was reported to OpenSSL by Joseph Barr-Pixton.
388 *) Exploitable out-of-bounds read in X509_cmp_time
390 X509_cmp_time does not properly check the length of the ASN1_TIME
391 string and can read a few bytes out of bounds. In addition,
392 X509_cmp_time accepts an arbitrary number of fractional seconds in the
395 An attacker can use this to craft malformed certificates and CRLs of
396 various sizes and potentially cause a segmentation fault, resulting in
397 a DoS on applications that verify certificates or CRLs. TLS clients
398 that verify CRLs are affected. TLS clients and servers with client
399 authentication enabled may be affected if they use custom verification
402 This issue was reported to OpenSSL by Robert Swiecki (Google), and
403 independently by Hanno Böck.
407 *) PKCS7 crash with missing EnvelopedContent
409 The PKCS#7 parsing code does not handle missing inner EncryptedContent
410 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
411 with missing content and trigger a NULL pointer dereference on parsing.
413 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
414 structures from untrusted sources are affected. OpenSSL clients and
415 servers are not affected.
417 This issue was reported to OpenSSL by Michal Zalewski (Google).
421 *) CMS verify infinite loop with unknown hash function
423 When verifying a signedData message the CMS code can enter an infinite loop
424 if presented with an unknown hash function OID. This can be used to perform
425 denial of service against any system which verifies signedData messages using
427 This issue was reported to OpenSSL by Johannes Bauer.
431 *) Race condition handling NewSessionTicket
433 If a NewSessionTicket is received by a multi-threaded client when attempting to
434 reuse a previous ticket then a race condition can occur potentially leading to
435 a double free of the ticket data.
439 *) Removed support for the two export grade static DH ciphersuites
440 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
441 were newly added (along with a number of other static DH ciphersuites) to
442 1.0.2. However the two export ones have *never* worked since they were
443 introduced. It seems strange in any case to be adding new export
444 ciphersuites, and given "logjam" it also does not seem correct to fix them.
447 *) Only support 256-bit or stronger elliptic curves with the
448 'ecdh_auto' setting (server) or by default (client). Of supported
449 curves, prefer P-256 (both).
452 *) Reject DH handshakes with parameters shorter than 768 bits.
453 [Kurt Roeckx and Emilia Kasper]
455 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
457 *) ClientHello sigalgs DoS fix
459 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
460 invalid signature algorithms extension a NULL pointer dereference will
461 occur. This can be exploited in a DoS attack against the server.
463 This issue was was reported to OpenSSL by David Ramos of Stanford
466 [Stephen Henson and Matt Caswell]
468 *) Multiblock corrupted pointer fix
470 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
471 feature only applies on 64 bit x86 architecture platforms that support AES
472 NI instructions. A defect in the implementation of "multiblock" can cause
473 OpenSSL's internal write buffer to become incorrectly set to NULL when
474 using non-blocking IO. Typically, when the user application is using a
475 socket BIO for writing, this will only result in a failed connection.
476 However if some other BIO is used then it is likely that a segmentation
477 fault will be triggered, thus enabling a potential DoS attack.
479 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
483 *) Segmentation fault in DTLSv1_listen fix
485 The DTLSv1_listen function is intended to be stateless and processes the
486 initial ClientHello from many peers. It is common for user code to loop
487 over the call to DTLSv1_listen until a valid ClientHello is received with
488 an associated cookie. A defect in the implementation of DTLSv1_listen means
489 that state is preserved in the SSL object from one invocation to the next
490 that can lead to a segmentation fault. Errors processing the initial
491 ClientHello can trigger this scenario. An example of such an error could be
492 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
495 This issue was reported to OpenSSL by Per Allansson.
499 *) Segmentation fault in ASN1_TYPE_cmp fix
501 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
502 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
503 certificate signature algorithm consistency this can be used to crash any
504 certificate verification operation and exploited in a DoS attack. Any
505 application which performs certificate verification is vulnerable including
506 OpenSSL clients and servers which enable client authentication.
510 *) Segmentation fault for invalid PSS parameters fix
512 The signature verification routines will crash with a NULL pointer
513 dereference if presented with an ASN.1 signature using the RSA PSS
514 algorithm and invalid parameters. Since these routines are used to verify
515 certificate signature algorithms this can be used to crash any
516 certificate verification operation and exploited in a DoS attack. Any
517 application which performs certificate verification is vulnerable including
518 OpenSSL clients and servers which enable client authentication.
520 This issue was was reported to OpenSSL by Brian Carpenter.
524 *) ASN.1 structure reuse memory corruption fix
526 Reusing a structure in ASN.1 parsing may allow an attacker to cause
527 memory corruption via an invalid write. Such reuse is and has been
528 strongly discouraged and is believed to be rare.
530 Applications that parse structures containing CHOICE or ANY DEFINED BY
531 components may be affected. Certificate parsing (d2i_X509 and related
532 functions) are however not affected. OpenSSL clients and servers are
537 *) PKCS7 NULL pointer dereferences fix
539 The PKCS#7 parsing code does not handle missing outer ContentInfo
540 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
541 missing content and trigger a NULL pointer dereference on parsing.
543 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
544 otherwise parse PKCS#7 structures from untrusted sources are
545 affected. OpenSSL clients and servers are not affected.
547 This issue was reported to OpenSSL by Michal Zalewski (Google).
551 *) DoS via reachable assert in SSLv2 servers fix
553 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
554 servers that both support SSLv2 and enable export cipher suites by sending
555 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
557 This issue was discovered by Sean Burford (Google) and Emilia Käsper
558 (OpenSSL development team).
562 *) Empty CKE with client auth and DHE fix
564 If client auth is used then a server can seg fault in the event of a DHE
565 ciphersuite being selected and a zero length ClientKeyExchange message
566 being sent by the client. This could be exploited in a DoS attack.
570 *) Handshake with unseeded PRNG fix
572 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
573 with an unseeded PRNG. The conditions are:
574 - The client is on a platform where the PRNG has not been seeded
575 automatically, and the user has not seeded manually
576 - A protocol specific client method version has been used (i.e. not
577 SSL_client_methodv23)
578 - A ciphersuite is used that does not require additional random data from
579 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
581 If the handshake succeeds then the client random that has been used will
582 have been generated from a PRNG with insufficient entropy and therefore the
583 output may be predictable.
585 For example using the following command with an unseeded openssl will
586 succeed on an unpatched platform:
588 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
592 *) Use After Free following d2i_ECPrivatekey error fix
594 A malformed EC private key file consumed via the d2i_ECPrivateKey function
595 could cause a use after free condition. This, in turn, could cause a double
596 free in several private key parsing functions (such as d2i_PrivateKey
597 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
598 for applications that receive EC private keys from untrusted
599 sources. This scenario is considered rare.
601 This issue was discovered by the BoringSSL project and fixed in their
606 *) X509_to_X509_REQ NULL pointer deref fix
608 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
609 the certificate key is invalid. This function is rarely used in practice.
611 This issue was discovered by Brian Carpenter.
615 *) Removed the export ciphers from the DEFAULT ciphers
618 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
620 *) Change RSA and DH/DSA key generation apps to generate 2048-bit
624 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
625 ARMv5 through ARMv8, as opposite to "locking" it to single one.
626 So far those who have to target multiple plaforms would compromise
627 and argue that binary targeting say ARMv5 would still execute on
628 ARMv8. "Universal" build resolves this compromise by providing
629 near-optimal performance even on newer platforms.
632 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
633 (other platforms pending).
634 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
636 *) Add support for the SignedCertificateTimestampList certificate and
637 OCSP response extensions from RFC6962.
640 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
641 for corner cases. (Certain input points at infinity could lead to
642 bogus results, with non-infinity inputs mapped to infinity too.)
645 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
646 This covers AES, SHA256/512 and GHASH. "Initial" means that most
647 common cases are optimized and there still is room for further
648 improvements. Vector Permutation AES for Altivec is also added.
651 *) Add support for little-endian ppc64 Linux target.
652 [Marcelo Cerri (IBM)]
654 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
655 SHA1, SHA256 and GHASH. "Initial" means that most common cases
656 are optimized and there still is room for further improvements.
657 Both 32- and 64-bit modes are supported.
658 [Andy Polyakov, Ard Biesheuvel (Linaro)]
660 *) Improved ARMv7 NEON support.
663 *) Support for SPARC Architecture 2011 crypto extensions, first
664 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
665 SHA256/512, MD5, GHASH and modular exponentiation.
666 [Andy Polyakov, David Miller]
668 *) Accelerated modular exponentiation for Intel processors, a.k.a.
670 [Shay Gueron & Vlad Krasnov (Intel Corp)]
672 *) Support for new and upcoming Intel processors, including AVX2,
673 BMI and SHA ISA extensions. This includes additional "stitched"
674 implementations, AESNI-SHA256 and GCM, and multi-buffer support
677 This work was sponsored by Intel Corp.
680 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
681 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
682 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
685 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
686 this fixes a limiation in previous versions of OpenSSL.
689 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
690 MGF1 digest and OAEP label.
693 *) Add EVP support for key wrapping algorithms, to avoid problems with
694 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
695 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
696 algorithms and include tests cases.
699 *) Add functions to allocate and set the fields of an ECDSA_METHOD
701 [Douglas E. Engert, Steve Henson]
703 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
704 difference in days and seconds between two tm or ASN1_TIME structures.
707 *) Add -rev test option to s_server to just reverse order of characters
708 received by client and send back to server. Also prints an abbreviated
709 summary of the connection parameters.
712 *) New option -brief for s_client and s_server to print out a brief summary
713 of connection parameters.
716 *) Add callbacks for arbitrary TLS extensions.
717 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
719 *) New option -crl_download in several openssl utilities to download CRLs
720 from CRLDP extension in certificates.
723 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
726 *) New function X509_CRL_diff to generate a delta CRL from the difference
727 of two full CRLs. Add support to "crl" utility.
730 *) New functions to set lookup_crls function and to retrieve
731 X509_STORE from X509_STORE_CTX.
734 *) Print out deprecated issuer and subject unique ID fields in
738 *) Extend OCSP I/O functions so they can be used for simple general purpose
739 HTTP as well as OCSP. New wrapper function which can be used to download
740 CRLs using the OCSP API.
743 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
746 *) SSL_CONF* functions. These provide a common framework for application
747 configuration using configuration files or command lines.
750 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
751 message callback and prints the results. Needs compile time option
752 "enable-ssl-trace". New options to s_client and s_server to enable
756 *) New ctrl and macro to retrieve supported points extensions.
757 Print out extension in s_server and s_client.
760 *) New functions to retrieve certificate signature and signature
764 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
768 *) New Suite B modes for TLS code. These use and enforce the requirements
769 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
770 only use Suite B curves. The Suite B modes can be set by using the
771 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
774 *) New chain verification flags for Suite B levels of security. Check
775 algorithms are acceptable when flags are set in X509_verify_cert.
778 *) Make tls1_check_chain return a set of flags indicating checks passed
779 by a certificate chain. Add additional tests to handle client
780 certificates: checks for matching certificate type and issuer name
784 *) If an attempt is made to use a signature algorithm not in the peer
785 preference list abort the handshake. If client has no suitable
786 signature algorithms in response to a certificate request do not
790 *) If server EC tmp key is not in client preference list abort handshake.
793 *) Add support for certificate stores in CERT structure. This makes it
794 possible to have different stores per SSL structure or one store in
795 the parent SSL_CTX. Include distint stores for certificate chain
796 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
797 to build and store a certificate chain in CERT structure: returing
798 an error if the chain cannot be built: this will allow applications
799 to test if a chain is correctly configured.
801 Note: if the CERT based stores are not set then the parent SSL_CTX
802 store is used to retain compatibility with existing behaviour.
806 *) New function ssl_set_client_disabled to set a ciphersuite disabled
807 mask based on the current session, check mask when sending client
808 hello and checking the requested ciphersuite.
811 *) New ctrls to retrieve and set certificate types in a certificate
812 request message. Print out received values in s_client. If certificate
813 types is not set with custom values set sensible values based on
814 supported signature algorithms.
817 *) Support for distinct client and server supported signature algorithms.
820 *) Add certificate callback. If set this is called whenever a certificate
821 is required by client or server. An application can decide which
822 certificate chain to present based on arbitrary criteria: for example
823 supported signature algorithms. Add very simple example to s_server.
824 This fixes many of the problems and restrictions of the existing client
825 certificate callback: for example you can now clear an existing
826 certificate and specify the whole chain.
829 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
830 the certificate can be used for (if anything). Set valid_flags field
831 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
832 to have similar checks in it.
834 Add new "cert_flags" field to CERT structure and include a "strict mode".
835 This enforces some TLS certificate requirements (such as only permitting
836 certificate signature algorithms contained in the supported algorithms
837 extension) which some implementations ignore: this option should be used
838 with caution as it could cause interoperability issues.
841 *) Update and tidy signature algorithm extension processing. Work out
842 shared signature algorithms based on preferences and peer algorithms
843 and print them out in s_client and s_server. Abort handshake if no
844 shared signature algorithms.
847 *) Add new functions to allow customised supported signature algorithms
848 for SSL and SSL_CTX structures. Add options to s_client and s_server
852 *) New function SSL_certs_clear() to delete all references to certificates
853 from an SSL structure. Before this once a certificate had been added
854 it couldn't be removed.
857 *) Integrate hostname, email address and IP address checking with certificate
858 verification. New verify options supporting checking in opensl utility.
861 *) Fixes and wildcard matching support to hostname and email checking
862 functions. Add manual page.
863 [Florian Weimer (Red Hat Product Security Team)]
865 *) New functions to check a hostname email or IP address against a
866 certificate. Add options x509 utility to print results of checks against
870 *) Fix OCSP checking.
871 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
873 *) Initial experimental support for explicitly trusted non-root CAs.
874 OpenSSL still tries to build a complete chain to a root but if an
875 intermediate CA has a trust setting included that is used. The first
876 setting is used: whether to trust (e.g., -addtrust option to the x509
880 *) Add -trusted_first option which attempts to find certificates in the
881 trusted store even if an untrusted chain is also supplied.
884 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
885 platform support for Linux and Android.
888 *) Support for linux-x32, ILP32 environment in x86_64 framework.
891 *) Experimental multi-implementation support for FIPS capable OpenSSL.
892 When in FIPS mode the approved implementations are used as normal,
893 when not in FIPS mode the internal unapproved versions are used instead.
894 This means that the FIPS capable OpenSSL isn't forced to use the
895 (often lower perfomance) FIPS implementations outside FIPS mode.
898 *) Transparently support X9.42 DH parameters when calling
899 PEM_read_bio_DHparameters. This means existing applications can handle
900 the new parameter format automatically.
903 *) Initial experimental support for X9.42 DH parameter format: mainly
904 to support use of 'q' parameter for RFC5114 parameters.
907 *) Add DH parameters from RFC5114 including test data to dhtest.
910 *) Support for automatic EC temporary key parameter selection. If enabled
911 the most preferred EC parameters are automatically used instead of
912 hardcoded fixed parameters. Now a server just has to call:
913 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
914 support ECDH and use the most appropriate parameters.
917 *) Enhance and tidy EC curve and point format TLS extension code. Use
918 static structures instead of allocation if default values are used.
919 New ctrls to set curves we wish to support and to retrieve shared curves.
920 Print out shared curves in s_server. New options to s_server and s_client
921 to set list of supported curves.
924 *) New ctrls to retrieve supported signature algorithms and
925 supported curve values as an array of NIDs. Extend openssl utility
926 to print out received values.
929 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
930 between NIDs and the more common NIST names such as "P-256". Enhance
931 ecparam utility and ECC method to recognise the NIST names for curves.
934 *) Enhance SSL/TLS certificate chain handling to support different
935 chains for each certificate instead of one chain in the parent SSL_CTX.
938 *) Support for fixed DH ciphersuite client authentication: where both
939 server and client use DH certificates with common parameters.
942 *) Support for fixed DH ciphersuites: those requiring DH server
946 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
948 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
949 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
950 X509_CINF_get_signature were reverted post internal team review.
952 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
954 *) Build fixes for the Windows and OpenVMS platforms
955 [Matt Caswell and Richard Levitte]
957 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
959 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
960 message can cause a segmentation fault in OpenSSL due to a NULL pointer
961 dereference. This could lead to a Denial Of Service attack. Thanks to
962 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
966 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
967 dtls1_buffer_record function under certain conditions. In particular this
968 could occur if an attacker sent repeated DTLS records with the same
969 sequence number but for the next epoch. The memory leak could be exploited
970 by an attacker in a Denial of Service attack through memory exhaustion.
971 Thanks to Chris Mueller for reporting this issue.
975 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
976 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
977 method would be set to NULL which could later result in a NULL pointer
978 dereference. Thanks to Frank Schmirler for reporting this issue.
982 *) Abort handshake if server key exchange message is omitted for ephemeral
985 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
986 reporting this issue.
990 *) Remove non-export ephemeral RSA code on client and server. This code
991 violated the TLS standard by allowing the use of temporary RSA keys in
992 non-export ciphersuites and could be used by a server to effectively
993 downgrade the RSA key length used to a value smaller than the server
994 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
995 INRIA or reporting this issue.
999 *) Fixed issue where DH client certificates are accepted without verification.
1000 An OpenSSL server will accept a DH certificate for client authentication
1001 without the certificate verify message. This effectively allows a client to
1002 authenticate without the use of a private key. This only affects servers
1003 which trust a client certificate authority which issues certificates
1004 containing DH keys: these are extremely rare and hardly ever encountered.
1005 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1010 *) Ensure that the session ID context of an SSL is updated when its
1011 SSL_CTX is updated via SSL_set_SSL_CTX.
1013 The session ID context is typically set from the parent SSL_CTX,
1014 and can vary with the CTX.
1017 *) Fix various certificate fingerprint issues.
1019 By using non-DER or invalid encodings outside the signed portion of a
1020 certificate the fingerprint can be changed without breaking the signature.
1021 Although no details of the signed portion of the certificate can be changed
1022 this can cause problems with some applications: e.g. those using the
1023 certificate fingerprint for blacklists.
1025 1. Reject signatures with non zero unused bits.
1027 If the BIT STRING containing the signature has non zero unused bits reject
1028 the signature. All current signature algorithms require zero unused bits.
1030 2. Check certificate algorithm consistency.
1032 Check the AlgorithmIdentifier inside TBS matches the one in the
1033 certificate signature. NB: this will result in signature failure
1034 errors for some broken certificates.
1036 Thanks to Konrad Kraszewski from Google for reporting this issue.
1038 3. Check DSA/ECDSA signatures use DER.
1040 Reencode DSA/ECDSA signatures and compare with the original received
1041 signature. Return an error if there is a mismatch.
1043 This will reject various cases including garbage after signature
1044 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1045 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1046 (negative or with leading zeroes).
1048 Further analysis was conducted and fixes were developed by Stephen Henson
1049 of the OpenSSL core team.
1054 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1055 results on some platforms, including x86_64. This bug occurs at random
1056 with a very low probability, and is not known to be exploitable in any
1057 way, though its exact impact is difficult to determine. Thanks to Pieter
1058 Wuille (Blockstream) who reported this issue and also suggested an initial
1059 fix. Further analysis was conducted by the OpenSSL development team and
1060 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1061 the OpenSSL core team.
1065 *) Do not resume sessions on the server if the negotiated protocol
1066 version does not match the session's version. Resuming with a different
1067 version, while not strictly forbidden by the RFC, is of questionable
1068 sanity and breaks all known clients.
1069 [David Benjamin, Emilia Käsper]
1071 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1072 early CCS messages during renegotiation. (Note that because
1073 renegotiation is encrypted, this early CCS was not exploitable.)
1076 *) Tighten client-side session ticket handling during renegotiation:
1077 ensure that the client only accepts a session ticket if the server sends
1078 the extension anew in the ServerHello. Previously, a TLS client would
1079 reuse the old extension state and thus accept a session ticket if one was
1080 announced in the initial ServerHello.
1082 Similarly, ensure that the client requires a session ticket if one
1083 was advertised in the ServerHello. Previously, a TLS client would
1084 ignore a missing NewSessionTicket message.
1087 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1089 *) SRTP Memory Leak.
1091 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1092 sends a carefully crafted handshake message, to cause OpenSSL to fail
1093 to free up to 64k of memory causing a memory leak. This could be
1094 exploited in a Denial Of Service attack. This issue affects OpenSSL
1095 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1096 whether SRTP is used or configured. Implementations of OpenSSL that
1097 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1099 The fix was developed by the OpenSSL team.
1103 *) Session Ticket Memory Leak.
1105 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1106 integrity of that ticket is first verified. In the event of a session
1107 ticket integrity check failing, OpenSSL will fail to free memory
1108 causing a memory leak. By sending a large number of invalid session
1109 tickets an attacker could exploit this issue in a Denial Of Service
1114 *) Build option no-ssl3 is incomplete.
1116 When OpenSSL is configured with "no-ssl3" as a build option, servers
1117 could accept and complete a SSL 3.0 handshake, and clients could be
1118 configured to send them.
1120 [Akamai and the OpenSSL team]
1122 *) Add support for TLS_FALLBACK_SCSV.
1123 Client applications doing fallback retries should call
1124 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1126 [Adam Langley, Bodo Moeller]
1128 *) Add additional DigestInfo checks.
1130 Reencode DigestInto in DER and check against the original when
1131 verifying RSA signature: this will reject any improperly encoded
1132 DigestInfo structures.
1134 Note: this is a precautionary measure and no attacks are currently known.
1138 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1140 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1141 SRP code can be overrun an internal buffer. Add sanity check that
1142 g, A, B < N to SRP code.
1144 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1145 Group for discovering this issue.
1149 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1150 TLS 1.0 instead of higher protocol versions when the ClientHello message
1151 is badly fragmented. This allows a man-in-the-middle attacker to force a
1152 downgrade to TLS 1.0 even if both the server and the client support a
1153 higher protocol version, by modifying the client's TLS records.
1155 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1156 researching this issue.
1160 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1161 to a denial of service attack. A malicious server can crash the client
1162 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1163 ciphersuite and sending carefully crafted handshake messages.
1165 Thanks to Felix Gröbert (Google) for discovering and researching this
1170 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1171 to leak memory. This can be exploited through a Denial of Service attack.
1172 Thanks to Adam Langley for discovering and researching this issue.
1176 *) An attacker can force openssl to consume large amounts of memory whilst
1177 processing DTLS handshake messages. This can be exploited through a
1178 Denial of Service attack.
1179 Thanks to Adam Langley for discovering and researching this issue.
1183 *) An attacker can force an error condition which causes openssl to crash
1184 whilst processing DTLS packets due to memory being freed twice. This
1185 can be exploited through a Denial of Service attack.
1186 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1191 *) If a multithreaded client connects to a malicious server using a resumed
1192 session and the server sends an ec point format extension it could write
1193 up to 255 bytes to freed memory.
1195 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1200 *) A malicious server can crash an OpenSSL client with a null pointer
1201 dereference (read) by specifying an SRP ciphersuite even though it was not
1202 properly negotiated with the client. This can be exploited through a
1203 Denial of Service attack.
1205 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1206 discovering and researching this issue.
1210 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1211 X509_name_oneline, X509_name_print_ex et al. to leak some information
1212 from the stack. Applications may be affected if they echo pretty printing
1213 output to the attacker.
1215 Thanks to Ivan Fratric (Google) for discovering this issue.
1217 [Emilia Käsper, and Steve Henson]
1219 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1220 for corner cases. (Certain input points at infinity could lead to
1221 bogus results, with non-infinity inputs mapped to infinity too.)
1224 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1226 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1227 handshake can force the use of weak keying material in OpenSSL
1228 SSL/TLS clients and servers.
1230 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1231 researching this issue. (CVE-2014-0224)
1232 [KIKUCHI Masashi, Steve Henson]
1234 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1235 OpenSSL DTLS client the code can be made to recurse eventually crashing
1238 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1240 [Imre Rad, Steve Henson]
1242 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1243 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1244 client or server. This is potentially exploitable to run arbitrary
1245 code on a vulnerable client or server.
1247 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1248 [Jüri Aedla, Steve Henson]
1250 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1251 are subject to a denial of service attack.
1253 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1254 this issue. (CVE-2014-3470)
1255 [Felix Gröbert, Ivan Fratric, Steve Henson]
1257 *) Harmonize version and its documentation. -f flag is used to display
1259 [mancha <mancha1@zoho.com>]
1261 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1262 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1263 [mancha <mancha1@zoho.com>]
1265 *) Fix some double frees. These are not thought to be exploitable.
1266 [mancha <mancha1@zoho.com>]
1268 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1270 *) A missing bounds check in the handling of the TLS heartbeat extension
1271 can be used to reveal up to 64k of memory to a connected client or
1274 Thanks for Neel Mehta of Google Security for discovering this bug and to
1275 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1276 preparing the fix (CVE-2014-0160)
1277 [Adam Langley, Bodo Moeller]
1279 *) Fix for the attack described in the paper "Recovering OpenSSL
1280 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1281 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1282 http://eprint.iacr.org/2014/140
1284 Thanks to Yuval Yarom and Naomi Benger for discovering this
1285 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1286 [Yuval Yarom and Naomi Benger]
1288 *) TLS pad extension: draft-agl-tls-padding-03
1290 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1291 TLS client Hello record length value would otherwise be > 255 and
1292 less that 512 pad with a dummy extension containing zeroes so it
1293 is at least 512 bytes long.
1295 [Adam Langley, Steve Henson]
1297 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1299 *) Fix for TLS record tampering bug. A carefully crafted invalid
1300 handshake could crash OpenSSL with a NULL pointer exception.
1301 Thanks to Anton Johansson for reporting this issues.
1304 *) Keep original DTLS digest and encryption contexts in retransmission
1305 structures so we can use the previous session parameters if they need
1306 to be resent. (CVE-2013-6450)
1309 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1310 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1311 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1312 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1313 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1314 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1315 [Rob Stradling, Adam Langley]
1317 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1319 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1320 supporting platforms or when small records were transferred.
1321 [Andy Polyakov, Steve Henson]
1323 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1325 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1327 This addresses the flaw in CBC record processing discovered by
1328 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1329 at: http://www.isg.rhul.ac.uk/tls/
1331 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1332 Security Group at Royal Holloway, University of London
1333 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1334 Emilia Käsper for the initial patch.
1336 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1338 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1339 ciphersuites which can be exploited in a denial of service attack.
1340 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1341 and detecting this bug and to Wolfgang Ettlinger
1342 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1346 *) Return an error when checking OCSP signatures when key is NULL.
1347 This fixes a DoS attack. (CVE-2013-0166)
1350 *) Make openssl verify return errors.
1351 [Chris Palmer <palmer@google.com> and Ben Laurie]
1353 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1354 the right response is stapled. Also change SSL_get_certificate()
1355 so it returns the certificate actually sent.
1356 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1357 [Rob Stradling <rob.stradling@comodo.com>]
1359 *) Fix possible deadlock when decoding public keys.
1362 *) Don't use TLS 1.0 record version number in initial client hello
1366 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1368 *) Sanity check record length before skipping explicit IV in TLS
1369 1.2, 1.1 and DTLS to fix DoS attack.
1371 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1372 fuzzing as a service testing platform.
1376 *) Initialise tkeylen properly when encrypting CMS messages.
1377 Thanks to Solar Designer of Openwall for reporting this issue.
1380 *) In FIPS mode don't try to use composite ciphers as they are not
1384 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1386 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1387 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1388 mean any application compiled against OpenSSL 1.0.0 headers setting
1389 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1390 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1391 0x10000000L Any application which was previously compiled against
1392 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1393 will need to be recompiled as a result. Letting be results in
1394 inability to disable specifically TLS 1.1 and in client context,
1395 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1398 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1399 disable just protocol X, but all protocols above X *if* there are
1400 protocols *below* X still enabled. In more practical terms it means
1401 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1402 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1403 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1407 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1409 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1410 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1411 in CRYPTO_realloc_clean.
1413 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1414 issue and to Adam Langley <agl@chromium.org> for fixing it.
1416 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1418 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1421 *) Workarounds for some broken servers that "hang" if a client hello
1422 record length exceeds 255 bytes.
1424 1. Do not use record version number > TLS 1.0 in initial client
1425 hello: some (but not all) hanging servers will now work.
1426 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1427 the number of ciphers sent in the client hello. This should be
1428 set to an even number, such as 50, for example by passing:
1429 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1430 Most broken servers should now work.
1431 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1432 TLS 1.2 client support entirely.
1435 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1438 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1440 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1441 STRING form instead of a DigestInfo.
1444 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1445 and the RSA_sign/RSA_verify functions. This was made more apparent when
1446 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1447 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1448 the correct format in RSA_verify so both forms transparently work.
1451 *) Some servers which support TLS 1.0 can choke if we initially indicate
1452 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1453 encrypted premaster secret. As a workaround use the maximum pemitted
1454 client version in client hello, this should keep such servers happy
1455 and still work with previous versions of OpenSSL.
1458 *) Add support for TLS/DTLS heartbeats.
1459 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1461 *) Add support for SCTP.
1462 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1464 *) Improved PRNG seeding for VOS.
1465 [Paul Green <Paul.Green@stratus.com>]
1467 *) Extensive assembler packs updates, most notably:
1469 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1470 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1471 - x86_64: bit-sliced AES implementation;
1472 - ARM: NEON support, contemporary platforms optimizations;
1473 - s390x: z196 support;
1474 - *: GHASH and GF(2^m) multiplication implementations;
1478 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1479 (removal of unnecessary code)
1480 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1482 *) Add TLS key material exporter from RFC 5705.
1485 *) Add DTLS-SRTP negotiation from RFC 5764.
1488 *) Add Next Protocol Negotiation,
1489 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1490 disabled with a no-npn flag to config or Configure. Code donated
1492 [Adam Langley <agl@google.com> and Ben Laurie]
1494 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1495 NIST-P256, NIST-P521, with constant-time single point multiplication on
1496 typical inputs. Compiler support for the nonstandard type __uint128_t is
1497 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1498 Code made available under Apache License version 2.0.
1500 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1501 line to include this in your build of OpenSSL, and run "make depend" (or
1502 "make update"). This enables the following EC_METHODs:
1504 EC_GFp_nistp224_method()
1505 EC_GFp_nistp256_method()
1506 EC_GFp_nistp521_method()
1508 EC_GROUP_new_by_curve_name() will automatically use these (while
1509 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1511 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1513 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1514 all platforms. Move ssize_t definition from e_os.h to the public
1515 header file e_os2.h as it now appears in public header file cms.h
1518 *) New -sigopt option to the ca, req and x509 utilities. Additional
1519 signature parameters can be passed using this option and in
1523 *) Add RSA PSS signing function. This will generate and set the
1524 appropriate AlgorithmIdentifiers for PSS based on those in the
1525 corresponding EVP_MD_CTX structure. No application support yet.
1528 *) Support for companion algorithm specific ASN1 signing routines.
1529 New function ASN1_item_sign_ctx() signs a pre-initialised
1530 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1531 the appropriate parameters.
1534 *) Add new algorithm specific ASN1 verification initialisation function
1535 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1536 handling will be the same no matter what EVP_PKEY_METHOD is used.
1537 Add a PSS handler to support verification of PSS signatures: checked
1538 against a number of sample certificates.
1541 *) Add signature printing for PSS. Add PSS OIDs.
1542 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1544 *) Add algorithm specific signature printing. An individual ASN1 method
1545 can now print out signatures instead of the standard hex dump.
1547 More complex signatures (e.g. PSS) can print out more meaningful
1548 information. Include DSA version that prints out the signature
1552 *) Password based recipient info support for CMS library: implementing
1556 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1557 neatly separates the code into cipher and PBE sections and is required
1558 for some algorithms that split PBES2 into separate pieces (such as
1559 password based CMS).
1562 *) Session-handling fixes:
1563 - Fix handling of connections that are resuming with a session ID,
1564 but also support Session Tickets.
1565 - Fix a bug that suppressed issuing of a new ticket if the client
1566 presented a ticket with an expired session.
1567 - Try to set the ticket lifetime hint to something reasonable.
1568 - Make tickets shorter by excluding irrelevant information.
1569 - On the client side, don't ignore renewed tickets.
1570 [Adam Langley, Bodo Moeller (Google)]
1572 *) Fix PSK session representation.
1575 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1577 This work was sponsored by Intel.
1580 *) Add GCM support to TLS library. Some custom code is needed to split
1581 the IV between the fixed (from PRF) and explicit (from TLS record)
1582 portions. This adds all GCM ciphersuites supported by RFC5288 and
1583 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
1584 add a special AESGCM string for GCM only.
1587 *) Expand range of ctrls for AES GCM. Permit setting invocation
1588 field on decrypt and retrieval of invocation field only on encrypt.
1591 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1592 As required by RFC5289 these ciphersuites cannot be used if for
1593 versions of TLS earlier than 1.2.
1596 *) For FIPS capable OpenSSL interpret a NULL default public key method
1597 as unset and return the appopriate default but do *not* set the default.
1598 This means we can return the appopriate method in applications that
1599 swicth between FIPS and non-FIPS modes.
1602 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1603 ENGINE is used then we cannot handle that in the FIPS module so we
1604 keep original code iff non-FIPS operations are allowed.
1607 *) Add -attime option to openssl utilities.
1608 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1610 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1613 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1614 FIPS EC methods unconditionally for now.
1617 *) New build option no-ec2m to disable characteristic 2 code.
1620 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1621 all cases can be covered as some introduce binary incompatibilities.
1624 *) Redirect RSA operations to FIPS module including keygen,
1625 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1628 *) Add similar low level API blocking to ciphers.
1631 *) Low level digest APIs are not approved in FIPS mode: any attempt
1632 to use these will cause a fatal error. Applications that *really* want
1633 to use them can use the private_* version instead.
1636 *) Redirect cipher operations to FIPS module for FIPS builds.
1639 *) Redirect digest operations to FIPS module for FIPS builds.
1642 *) Update build system to add "fips" flag which will link in fipscanister.o
1643 for static and shared library builds embedding a signature if needed.
1646 *) Output TLS supported curves in preference order instead of numerical
1647 order. This is currently hardcoded for the highest order curves first.
1648 This should be configurable so applications can judge speed vs strength.
1651 *) Add TLS v1.2 server support for client authentication.
1654 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1658 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1659 FIPS modules versions.
1662 *) Add TLS v1.2 client side support for client authentication. Keep cache
1663 of handshake records longer as we don't know the hash algorithm to use
1664 until after the certificate request message is received.
1667 *) Initial TLS v1.2 client support. Add a default signature algorithms
1668 extension including all the algorithms we support. Parse new signature
1669 format in client key exchange. Relax some ECC signing restrictions for
1670 TLS v1.2 as indicated in RFC5246.
1673 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1674 to new signature format when needed using client digest preference.
1675 All server ciphersuites should now work correctly in TLS v1.2. No client
1676 support yet and no support for client certificates.
1679 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1680 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1681 ciphersuites. At present only RSA key exchange ciphersuites work with
1682 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1683 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1684 and version checking.
1687 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1688 with this defined it will not be affected by any changes to ssl internal
1689 structures. Add several utility functions to allow openssl application
1690 to work with OPENSSL_NO_SSL_INTERN defined.
1694 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1696 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1699 *) Permit abbreviated handshakes when renegotiating using the function
1700 SSL_renegotiate_abbreviated().
1701 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1703 *) Add call to ENGINE_register_all_complete() to
1704 ENGINE_load_builtin_engines(), so some implementations get used
1705 automatically instead of needing explicit application support.
1708 *) Add support for TLS key exporter as described in RFC5705.
1709 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1711 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1712 a few changes are required:
1714 Add SSL_OP_NO_TLSv1_1 flag.
1715 Add TLSv1_1 methods.
1716 Update version checking logic to handle version 1.1.
1717 Add explicit IV handling (ported from DTLS code).
1718 Add command line options to s_client/s_server.
1721 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1723 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1724 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1725 content decryption and always return the same error. Note: this attack
1726 needs on average 2^20 messages so it only affects automated senders. The
1727 old behaviour can be reenabled in the CMS code by setting the
1728 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1729 an MMA defence is not necessary.
1730 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1731 this issue. (CVE-2012-0884)
1734 *) Fix CVE-2011-4619: make sure we really are receiving a
1735 client hello before rejecting multiple SGC restarts. Thanks to
1736 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1739 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1741 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1742 Thanks to Antonio Martin, Enterprise Secure Access Research and
1743 Development, Cisco Systems, Inc. for discovering this bug and
1744 preparing a fix. (CVE-2012-0050)
1747 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1749 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1750 of the Vaudenay padding oracle attack on CBC mode encryption
1751 which enables an efficient plaintext recovery attack against
1752 the OpenSSL implementation of DTLS. Their attack exploits timing
1753 differences arising during decryption processing. A research
1754 paper describing this attack can be found at:
1755 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1756 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1757 Security Group at Royal Holloway, University of London
1758 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1759 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1760 for preparing the fix. (CVE-2011-4108)
1761 [Robin Seggelmann, Michael Tuexen]
1763 *) Clear bytes used for block padding of SSL 3.0 records.
1765 [Adam Langley (Google)]
1767 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1768 Kadianakis <desnacked@gmail.com> for discovering this issue and
1769 Adam Langley for preparing the fix. (CVE-2011-4619)
1770 [Adam Langley (Google)]
1772 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1773 [Andrey Kulikov <amdeich@gmail.com>]
1775 *) Prevent malformed RFC3779 data triggering an assertion failure.
1776 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1777 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1778 [Rob Austein <sra@hactrn.net>]
1780 *) Improved PRNG seeding for VOS.
1781 [Paul Green <Paul.Green@stratus.com>]
1783 *) Fix ssl_ciph.c set-up race.
1784 [Adam Langley (Google)]
1786 *) Fix spurious failures in ecdsatest.c.
1787 [Emilia Käsper (Google)]
1789 *) Fix the BIO_f_buffer() implementation (which was mixing different
1790 interpretations of the '..._len' fields).
1791 [Adam Langley (Google)]
1793 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1794 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1795 threads won't reuse the same blinding coefficients.
1797 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1798 lock to call BN_BLINDING_invert_ex, and avoids one use of
1799 BN_BLINDING_update for each BN_BLINDING structure (previously,
1800 the last update always remained unused).
1801 [Emilia Käsper (Google)]
1803 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1804 [Bob Buckholz (Google)]
1806 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1808 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1809 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1810 [Kaspar Brand <ossl@velox.ch>]
1812 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1813 for multi-threaded use of ECDH. (CVE-2011-3210)
1814 [Adam Langley (Google)]
1816 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1819 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1820 signature public key algorithm by using OID xref utilities instead.
1821 Before this you could only use some ECC ciphersuites with SHA1 only.
1824 *) Add protection against ECDSA timing attacks as mentioned in the paper
1825 by Billy Bob Brumley and Nicola Tuveri, see:
1827 http://eprint.iacr.org/2011/232.pdf
1829 [Billy Bob Brumley and Nicola Tuveri]
1831 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1833 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1834 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1836 *) Fix bug in string printing code: if *any* escaping is enabled we must
1837 escape the escape character (backslash) or the resulting string is
1841 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1843 *) Disable code workaround for ancient and obsolete Netscape browsers
1844 and servers: an attacker can use it in a ciphersuite downgrade attack.
1845 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1848 *) Fixed J-PAKE implementation error, originally discovered by
1849 Sebastien Martini, further info and confirmation from Stefan
1850 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1853 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1855 *) Fix extension code to avoid race conditions which can result in a buffer
1856 overrun vulnerability: resumed sessions must not be modified as they can
1857 be shared by multiple threads. CVE-2010-3864
1860 *) Fix WIN32 build system to correctly link an ENGINE directory into
1864 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1866 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1868 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1870 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1872 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1873 context. The operation can be customised via the ctrl mechanism in
1874 case ENGINEs want to include additional functionality.
1877 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1880 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1881 output hashes compatible with older versions of OpenSSL.
1882 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1884 *) Fix compression algorithm handling: if resuming a session use the
1885 compression algorithm of the resumed session instead of determining
1886 it from client hello again. Don't allow server to change algorithm.
1889 *) Add load_crls() function to apps tidying load_certs() too. Add option
1890 to verify utility to allow additional CRLs to be included.
1893 *) Update OCSP request code to permit adding custom headers to the request:
1894 some responders need this.
1897 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1899 [Julia Lawall <julia@diku.dk>]
1901 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1902 needlessly dereferenced structures, used obsolete functions and
1903 didn't handle all updated verify codes correctly.
1906 *) Disable MD2 in the default configuration.
1909 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1910 indicate the initial BIO being pushed or popped. This makes it possible
1911 to determine whether the BIO is the one explicitly called or as a result
1912 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1913 it handles reference counts correctly and doesn't zero out the I/O bio
1914 when it is not being explicitly popped. WARNING: applications which
1915 included workarounds for the old buggy behaviour will need to be modified
1916 or they could free up already freed BIOs.
1919 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1920 renaming to all platforms (within the 0.9.8 branch, this was
1921 done conditionally on Netware platforms to avoid a name clash).
1922 [Guenter <lists@gknw.net>]
1924 *) Add ECDHE and PSK support to DTLS.
1925 [Michael Tuexen <tuexen@fh-muenster.de>]
1927 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1931 *) Add "missing" function EVP_MD_flags() (without this the only way to
1932 retrieve a digest flags is by accessing the structure directly. Update
1933 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1934 or cipher is registered as in the "from" argument. Print out all
1935 registered digests in the dgst usage message instead of manually
1936 attempting to work them out.
1939 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1940 this allows the use of compression and extensions. Change default cipher
1941 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1942 by default unless an application cipher string requests it.
1945 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1946 key ids to find matching certificates and keys but some PKCS#12 files
1947 don't follow the (somewhat unwritten) rules and this strategy fails.
1948 Now just gather all certificates together and the first private key
1949 then look for the first certificate that matches the key.
1952 *) Support use of registered digest and cipher names for dgst and cipher
1953 commands instead of having to add each one as a special case. So now
1960 openssl dgst -sha256 foo
1962 and this works for ENGINE based algorithms too.
1966 *) Update Gost ENGINE to support parameter files.
1967 [Victor B. Wagner <vitus@cryptocom.ru>]
1969 *) Support GeneralizedTime in ca utility.
1970 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1972 *) Enhance the hash format used for certificate directory links. The new
1973 form uses the canonical encoding (meaning equivalent names will work
1974 even if they aren't identical) and uses SHA1 instead of MD5. This form
1975 is incompatible with the older format and as a result c_rehash should
1976 be used to rebuild symbolic links.
1979 *) Make PKCS#8 the default write format for private keys, replacing the
1980 traditional format. This form is standardised, more secure and doesn't
1981 include an implicit MD5 dependency.
1984 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1985 committed to OpenSSL should pass this lot as a minimum.
1988 *) Add session ticket override functionality for use by EAP-FAST.
1989 [Jouni Malinen <j@w1.fi>]
1991 *) Modify HMAC functions to return a value. Since these can be implemented
1992 in an ENGINE errors can occur.
1995 *) Type-checked OBJ_bsearch_ex.
1998 *) Type-checked OBJ_bsearch. Also some constification necessitated
1999 by type-checking. Still to come: TXT_DB, bsearch(?),
2000 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2004 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2005 seconds to a tm structure directly, instead of going through OS
2006 specific date routines. This avoids any issues with OS routines such
2007 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2008 and X509_time_adj_ex() to cover the extended range. The existing
2009 X509_time_adj() is still usable and will no longer have any date issues.
2012 *) Delta CRL support. New use deltas option which will attempt to locate
2013 and search any appropriate delta CRLs available.
2015 This work was sponsored by Google.
2018 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2019 code and add additional score elements. Validate alternate CRL paths
2020 as part of the CRL checking and indicate a new error "CRL path validation
2021 error" in this case. Applications wanting additional details can use
2022 the verify callback and check the new "parent" field. If this is not
2023 NULL CRL path validation is taking place. Existing applications wont
2024 see this because it requires extended CRL support which is off by
2027 This work was sponsored by Google.
2030 *) Support for freshest CRL extension.
2032 This work was sponsored by Google.
2035 *) Initial indirect CRL support. Currently only supported in the CRLs
2036 passed directly and not via lookup. Process certificate issuer
2037 CRL entry extension and lookup CRL entries by bother issuer name
2038 and serial number. Check and process CRL issuer entry in IDP extension.
2040 This work was sponsored by Google.
2043 *) Add support for distinct certificate and CRL paths. The CRL issuer
2044 certificate is validated separately in this case. Only enabled if
2045 an extended CRL support flag is set: this flag will enable additional
2046 CRL functionality in future.
2048 This work was sponsored by Google.
2051 *) Add support for policy mappings extension.
2053 This work was sponsored by Google.
2056 *) Fixes to pathlength constraint, self issued certificate handling,
2057 policy processing to align with RFC3280 and PKITS tests.
2059 This work was sponsored by Google.
2062 *) Support for name constraints certificate extension. DN, email, DNS
2063 and URI types are currently supported.
2065 This work was sponsored by Google.
2068 *) To cater for systems that provide a pointer-based thread ID rather
2069 than numeric, deprecate the current numeric thread ID mechanism and
2070 replace it with a structure and associated callback type. This
2071 mechanism allows a numeric "hash" to be extracted from a thread ID in
2072 either case, and on platforms where pointers are larger than 'long',
2073 mixing is done to help ensure the numeric 'hash' is usable even if it
2074 can't be guaranteed unique. The default mechanism is to use "&errno"
2075 as a pointer-based thread ID to distinguish between threads.
2077 Applications that want to provide their own thread IDs should now use
2078 CRYPTO_THREADID_set_callback() to register a callback that will call
2079 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2081 Note that ERR_remove_state() is now deprecated, because it is tied
2082 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2083 to free the current thread's error state should be replaced by
2084 ERR_remove_thread_state(NULL).
2086 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2087 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2088 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2089 application was previously providing a numeric thread callback that
2090 was inappropriate for distinguishing threads, then uniqueness might
2091 have been obtained with &errno that happened immediately in the
2092 intermediate development versions of OpenSSL; this is no longer the
2093 case, the numeric thread callback will now override the automatic use
2095 [Geoff Thorpe, with help from Bodo Moeller]
2097 *) Initial support for different CRL issuing certificates. This covers a
2098 simple case where the self issued certificates in the chain exist and
2099 the real CRL issuer is higher in the existing chain.
2101 This work was sponsored by Google.
2104 *) Removed effectively defunct crypto/store from the build.
2107 *) Revamp of STACK to provide stronger type-checking. Still to come:
2108 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2109 ASN1_STRING, CONF_VALUE.
2112 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2113 RAM on SSL connections. This option can save about 34k per idle SSL.
2116 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2117 STACK, TXT_DB, bsearch, qsort.
2120 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2121 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2122 support for data, signedData, compressedData, digestedData and
2123 encryptedData, envelopedData types included. Scripts to check against
2124 RFC4134 examples draft and interop and consistency checks of many
2125 content types and variants.
2128 *) Add options to enc utility to support use of zlib compression BIO.
2131 *) Extend mk1mf to support importing of options and assembly language
2132 files from Configure script, currently only included in VC-WIN32.
2133 The assembly language rules can now optionally generate the source
2134 files from the associated perl scripts.
2137 *) Implement remaining functionality needed to support GOST ciphersuites.
2138 Interop testing has been performed using CryptoPro implementations.
2139 [Victor B. Wagner <vitus@cryptocom.ru>]
2141 *) s390x assembler pack.
2144 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2148 *) Implement Opaque PRF Input TLS extension as specified in
2149 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2150 official specification yet and no extension type assignment by
2151 IANA exists, this extension (for now) will have to be explicitly
2152 enabled when building OpenSSL by providing the extension number
2153 to use. For example, specify an option
2155 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2157 to the "config" or "Configure" script to enable the extension,
2158 assuming extension number 0x9527 (which is a completely arbitrary
2159 and unofficial assignment based on the MD5 hash of the Internet
2160 Draft). Note that by doing so, you potentially lose
2161 interoperability with other TLS implementations since these might
2162 be using the same extension number for other purposes.
2164 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2165 opaque PRF input value to use in the handshake. This will create
2166 an interal copy of the length-'len' string at 'src', and will
2167 return non-zero for success.
2169 To get more control and flexibility, provide a callback function
2172 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2173 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2177 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2180 Callback function 'cb' will be called in handshakes, and is
2181 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2182 Argument 'arg' is for application purposes (the value as given to
2183 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2184 be provided to the callback function). The callback function
2185 has to return non-zero to report success: usually 1 to use opaque
2186 PRF input just if possible, or 2 to enforce use of the opaque PRF
2187 input. In the latter case, the library will abort the handshake
2188 if opaque PRF input is not successfully negotiated.
2190 Arguments 'peerinput' and 'len' given to the callback function
2191 will always be NULL and 0 in the case of a client. A server will
2192 see the client's opaque PRF input through these variables if
2193 available (NULL and 0 otherwise). Note that if the server
2194 provides an opaque PRF input, the length must be the same as the
2195 length of the client's opaque PRF input.
2197 Note that the callback function will only be called when creating
2198 a new session (session resumption can resume whatever was
2199 previously negotiated), and will not be called in SSL 2.0
2200 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2201 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2202 for applications that need to enforce opaque PRF input.
2206 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2209 [Victor B. Wagner <vitus@cryptocom.ru>]
2211 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2212 RFC4507bis. The encrypted ticket format is an encrypted encoded
2213 SSL_SESSION structure, that way new session features are automatically
2216 If a client application caches session in an SSL_SESSION structure
2217 support is transparent because tickets are now stored in the encoded
2220 The SSL_CTX structure automatically generates keys for ticket
2221 protection in servers so again support should be possible
2222 with no application modification.
2224 If a client or server wishes to disable RFC4507 support then the option
2225 SSL_OP_NO_TICKET can be set.
2227 Add a TLS extension debugging callback to allow the contents of any client
2228 or server extensions to be examined.
2230 This work was sponsored by Google.
2233 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2234 OpenSSL should now compile cleanly on gcc 4.2
2235 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2237 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2238 support including streaming MAC support: this is required for GOST
2239 ciphersuite support.
2240 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2242 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2243 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2244 to output in BER and PEM format.
2247 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2248 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2249 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2250 ENGINE support for HMAC keys which are unextractable. New -mac and
2251 -macopt options to dgst utility.
2254 *) New option -sigopt to dgst utility. Update dgst to use
2255 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2256 alternative signing paramaters such as X9.31 or PSS in the dgst
2260 *) Change ssl_cipher_apply_rule(), the internal function that does
2261 the work each time a ciphersuite string requests enabling
2262 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2263 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2264 the order of disabled ciphersuites such that those ciphersuites
2265 that most recently went from enabled to disabled not only stay
2266 in order with respect to each other, but also have higher priority
2267 than other disabled ciphersuites the next time ciphersuites are
2270 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2271 the same ciphersuites as with "HIGH" alone, but in a specific
2272 order where the PSK ciphersuites come first (since they are the
2273 most recently disabled ciphersuites when "HIGH" is parsed).
2275 Also, change ssl_create_cipher_list() (using this new
2276 funcionality) such that between otherwise identical
2277 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2281 *) Change ssl_create_cipher_list() so that it automatically
2282 arranges the ciphersuites in reasonable order before starting
2283 to process the rule string. Thus, the definition for "DEFAULT"
2284 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2285 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2286 This makes it much easier to arrive at a reasonable default order
2287 in applications for which anonymous ciphers are OK (meaning
2288 that you can't actually use DEFAULT).
2289 [Bodo Moeller; suggested by Victor Duchovni]
2291 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2292 processing) into multiple integers instead of setting
2293 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2294 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2295 (These masks as well as the individual bit definitions are hidden
2296 away into the non-exported interface ssl/ssl_locl.h, so this
2297 change to the definition of the SSL_CIPHER structure shouldn't
2298 affect applications.) This give us more bits for each of these
2299 categories, so there is no longer a need to coagulate AES128 and
2300 AES256 into a single algorithm bit, and to coagulate Camellia128
2301 and Camellia256 into a single algorithm bit, which has led to all
2304 Thus, among other things, the kludge introduced in 0.9.7m and
2305 0.9.8e for masking out AES256 independently of AES128 or masking
2306 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2308 With the change, we also introduce new ciphersuite aliases that
2309 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2313 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2314 Use the leftmost N bytes of the signature input if the input is
2315 larger than the prime q (with N being the size in bytes of q).
2318 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2319 it yet and it is largely untested.
2322 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2325 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2326 some compilers (gcc 4.2 and later) reject their use. Safestack is
2327 reimplemented. Update ASN1 to avoid use of legacy functions.
2330 *) Win32/64 targets are linked with Winsock2.
2333 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2334 to external functions. This can be used to increase CRL handling
2335 efficiency especially when CRLs are very large by (for example) storing
2336 the CRL revoked certificates in a database.
2339 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2340 new CRLs added to a directory can be used. New command line option
2341 -verify_return_error to s_client and s_server. This causes real errors
2342 to be returned by the verify callback instead of carrying on no matter
2343 what. This reflects the way a "real world" verify callback would behave.
2346 *) GOST engine, supporting several GOST algorithms and public key formats.
2347 Kindly donated by Cryptocom.
2350 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2351 partitioned by DP are handled but no indirect CRL or reason partitioning
2352 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2353 selected via a scoring technique which handles IDP and AKID in CRLs.
2356 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2357 will ultimately be used for all verify operations: this will remove the
2358 X509_STORE dependency on certificate verification and allow alternative
2359 lookup methods. X509_STORE based implementations of these two callbacks.
2362 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2363 Modify get_crl() to find a valid (unexpired) CRL if possible.
2366 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2367 this would be called X509_CRL_cmp() but that name is already used by
2368 a function that just compares CRL issuer names. Cache several CRL
2369 extensions in X509_CRL structure and cache CRLDP in X509.
2372 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2373 this maps equivalent X509_NAME structures into a consistent structure.
2374 Name comparison can then be performed rapidly using memcmp().
2377 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2381 *) Allow digests to supply their own micalg string for S/MIME type using
2382 the ctrl EVP_MD_CTRL_MICALG.
2385 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2386 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2387 ctrl. It can then customise the structure before and/or after signing
2391 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2392 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2393 to free up any added signature OIDs.
2396 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2397 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2398 digest and cipher tables. New options added to openssl utility:
2399 list-message-digest-algorithms and list-cipher-algorithms.
2402 *) Change the array representation of binary polynomials: the list
2403 of degrees of non-zero coefficients is now terminated with -1.
2404 Previously it was terminated with 0, which was also part of the
2405 value; thus, the array representation was not applicable to
2406 polynomials where t^0 has coefficient zero. This change makes
2407 the array representation useful in a more general context.
2410 *) Various modifications and fixes to SSL/TLS cipher string
2411 handling. For ECC, the code now distinguishes between fixed ECDH
2412 with RSA certificates on the one hand and with ECDSA certificates
2413 on the other hand, since these are separate ciphersuites. The
2414 unused code for Fortezza ciphersuites has been removed.
2416 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2417 (not "ECDHE"). For consistency with the code for DH
2418 certificates, use of ECDH certificates is now considered ECDH
2419 authentication, not RSA or ECDSA authentication (the latter is
2420 merely the CA's signing algorithm and not actively used in the
2423 The temporary ciphersuite alias "ECCdraft" is no longer
2424 available, and ECC ciphersuites are no longer excluded from "ALL"
2425 and "DEFAULT". The following aliases now exist for RFC 4492
2426 ciphersuites, most of these by analogy with the DH case:
2428 kECDHr - ECDH cert, signed with RSA
2429 kECDHe - ECDH cert, signed with ECDSA
2430 kECDH - ECDH cert (signed with either RSA or ECDSA)
2431 kEECDH - ephemeral ECDH
2432 ECDH - ECDH cert or ephemeral ECDH
2438 AECDH - anonymous ECDH
2439 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2443 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2444 Use correct micalg parameters depending on digest(s) in signed message.
2447 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2448 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2451 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2452 an engine to register a method. Add ENGINE lookups for methods and
2453 functional reference processing.
2456 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2457 EVP_{Sign,Verify}* which allow an application to customise the signature
2461 *) New -resign option to smime utility. This adds one or more signers
2462 to an existing PKCS#7 signedData structure. Also -md option to use an
2463 alternative message digest algorithm for signing.
2466 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2467 create PKCS7 structures containing multiple signers. Update smime
2468 application to support multiple signers.
2471 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2475 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2476 Reorganize PBE internals to lookup from a static table using NIDs,
2477 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2478 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2479 PRF which will be automatically used with PBES2.
2482 *) Replace the algorithm specific calls to generate keys in "req" with the
2486 *) Update PKCS#7 enveloped data routines to use new API. This is now
2487 supported by any public key method supporting the encrypt operation. A
2488 ctrl is added to allow the public key algorithm to examine or modify
2489 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2493 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2494 a default digest type to use. In most cases this will be SHA1 but some
2495 algorithms (such as GOST) need to specify an alternative digest. The
2496 return value indicates how strong the prefernce is 1 means optional and
2497 2 is mandatory (that is it is the only supported type). Modify
2498 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2499 use the default md. Update openssl utilities to use the default digest
2500 type for signing if it is not explicitly indicated.
2503 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2504 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2505 signing method from the key type. This effectively removes the link
2506 between digests and public key types.
2509 *) Add an OID cross reference table and utility functions. Its purpose is to
2510 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2511 rsaEncryption. This will allow some of the algorithm specific hackery
2512 needed to use the correct OID to be removed.
2515 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2516 structures for PKCS7_sign(). They are now set up by the relevant public
2520 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2523 *) Add support for key derivation (agreement) in the API, DH method and
2527 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2528 public and private key formats. As a side effect these add additional
2529 command line functionality not previously available: DSA signatures can be
2530 generated and verified using pkeyutl and DH key support and generation in
2535 [Oliver Tappe <zooey@hirschkaefer.de>]
2537 *) New make target "install_html_docs" installs HTML renditions of the
2539 [Oliver Tappe <zooey@hirschkaefer.de>]
2541 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
2542 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2543 support key and parameter generation and add initial key generation
2544 functionality for RSA.
2547 *) Add functions for main EVP_PKEY_method operations. The undocumented
2548 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2549 EVP_PKEY_{encrypt,decrypt}_old.
2552 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2553 key API, doesn't do much yet.
2556 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2557 public key algorithms. New option to openssl utility:
2558 "list-public-key-algorithms" to print out info.
2561 *) Implement the Supported Elliptic Curves Extension for
2562 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2565 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2566 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2569 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2570 utilities such as rsa, dsa, dsaparam etc except they process any key
2574 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2575 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2576 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2580 *) Initial support for pluggable public key ASN1.
2581 De-spaghettify the public key ASN1 handling. Move public and private
2582 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2583 algorithm specific handling to a single module within the relevant
2584 algorithm directory. Add functions to allow (near) opaque processing
2585 of public and private key structures.
2588 *) Implement the Supported Point Formats Extension for
2589 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2592 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2593 for the psk identity [hint] and the psk callback functions to the
2594 SSL_SESSION, SSL and SSL_CTX structure.
2597 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2601 SSL_CTX_use_psk_identity_hint
2602 SSL_get_psk_identity_hint
2603 SSL_get_psk_identity
2604 SSL_use_psk_identity_hint
2606 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2608 *) Add RFC 3161 compliant time stamp request creation, response generation
2609 and response verification functionality.
2610 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2612 *) Add initial support for TLS extensions, specifically for the server_name
2613 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2614 have new members for a host name. The SSL data structure has an
2615 additional member SSL_CTX *initial_ctx so that new sessions can be
2616 stored in that context to allow for session resumption, even after the
2617 SSL has been switched to a new SSL_CTX in reaction to a client's
2618 server_name extension.
2620 New functions (subject to change):
2622 SSL_get_servername()
2623 SSL_get_servername_type()
2626 New CTRL codes and macros (subject to change):
2628 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2629 - SSL_CTX_set_tlsext_servername_callback()
2630 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2631 - SSL_CTX_set_tlsext_servername_arg()
2632 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2634 openssl s_client has a new '-servername ...' option.
2636 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2637 '-key2 ...', '-servername_fatal' (subject to change). This allows
2638 testing the HostName extension for a specific single host name ('-cert'
2639 and '-key' remain fallbacks for handshakes without HostName
2640 negotiation). If the unrecogninzed_name alert has to be sent, this by
2641 default is a warning; it becomes fatal with the '-servername_fatal'
2644 [Peter Sylvester, Remy Allais, Christophe Renou]
2646 *) Whirlpool hash implementation is added.
2649 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2650 bn(64,32). Because of instruction set limitations it doesn't have
2651 any negative impact on performance. This was done mostly in order
2652 to make it possible to share assembler modules, such as bn_mul_mont
2653 implementations, between 32- and 64-bit builds without hassle.
2656 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2657 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2661 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2662 dedicated Montgomery multiplication procedure, is introduced.
2663 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2664 "64-bit" performance on certain 32-bit targets.
2667 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2668 in SSL structures. New SSL ctrl to set maximum send fragment size.
2669 Save memory by seeting the I/O buffer sizes dynamically instead of
2670 using the maximum available value.
2673 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2674 in addition to the text details.
2677 *) Very, very preliminary EXPERIMENTAL support for printing of general
2678 ASN1 structures. This currently produces rather ugly output and doesn't
2679 handle several customised structures at all.
2682 *) Integrated support for PVK file format and some related formats such
2683 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2684 these in the 'rsa' and 'dsa' utilities.
2687 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2690 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2691 place for the (very old) "NETSCAPE" format certificates which are now
2692 handled using new ASN1 code equivalents.
2695 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2696 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2697 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2700 *) Modify CRL distribution points extension code to print out previously
2701 unsupported fields. Enhance extension setting code to allow setting of
2705 *) Add print and set support for Issuing Distribution Point CRL extension.
2708 *) Change 'Configure' script to enable Camellia by default.
2711 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2713 *) When rejecting SSL/TLS records due to an incorrect version number, never
2714 update s->server with a new major version number. As of
2715 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2716 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2717 the previous behavior could result in a read attempt at NULL when
2718 receiving specific incorrect SSL/TLS records once record payload
2719 protection is active. (CVE-2010-0740)
2720 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2722 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2723 could be crashed if the relevant tables were not present (e.g. chrooted).
2724 [Tomas Hoger <thoger@redhat.com>]
2726 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2728 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2729 [Martin Olsson, Neel Mehta]
2731 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2732 accommodate for stack sorting, always a write lock!).
2735 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2736 excessive delays in the RAND_poll(): over a minute. As a workaround
2737 include a time check in the inner Heap32Next loop too.
2740 *) The code that handled flushing of data in SSL/TLS originally used the
2741 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2742 the problem outlined in PR#1949. The fix suggested there however can
2743 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2744 of Apache). So instead simplify the code to flush unconditionally.
2745 This should be fine since flushing with no data to flush is a no op.
2748 *) Handle TLS versions 2.0 and later properly and correctly use the
2749 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2750 off ancient servers have a habit of sticking around for a while...
2753 *) Modify compression code so it frees up structures without using the
2754 ex_data callbacks. This works around a problem where some applications
2755 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2756 restarting) then use compression (e.g. SSL with compression) later.
2757 This results in significant per-connection memory leaks and
2758 has caused some security issues including CVE-2008-1678 and
2762 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2763 change when encrypting or decrypting.
2766 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2767 connect and renegotiate with servers which do not support RI.
2768 Until RI is more widely deployed this option is enabled by default.
2771 *) Add "missing" ssl ctrls to clear options and mode.
2774 *) If client attempts to renegotiate and doesn't support RI respond with
2775 a no_renegotiation alert as required by RFC5746. Some renegotiating
2776 TLS clients will continue a connection gracefully when they receive
2777 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2778 waiting for a server hello which it will never receive. Now we treat a
2779 received no_renegotiation alert as a fatal error. This is because
2780 applications requesting a renegotiation might well expect it to succeed
2781 and would have no code in place to handle the server denying it so the
2782 only safe thing to do is to terminate the connection.
2785 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2786 peer supports secure renegotiation and 0 otherwise. Print out peer
2787 renegotiation support in s_client/s_server.
2790 *) Replace the highly broken and deprecated SPKAC certification method with
2791 the updated NID creation version. This should correctly handle UTF8.
2794 *) Implement RFC5746. Re-enable renegotiation but require the extension
2795 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2796 turns out to be a bad idea. It has been replaced by
2797 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2798 SSL_CTX_set_options(). This is really not recommended unless you
2799 know what you are doing.
2800 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2802 *) Fixes to stateless session resumption handling. Use initial_ctx when
2803 issuing and attempting to decrypt tickets in case it has changed during
2804 servername handling. Use a non-zero length session ID when attempting
2805 stateless session resumption: this makes it possible to determine if
2806 a resumption has occurred immediately after receiving server hello
2807 (several places in OpenSSL subtly assume this) instead of later in
2811 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2812 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2813 fixes for a few places where the return code is not checked
2815 [Julia Lawall <julia@diku.dk>]
2817 *) Add --strict-warnings option to Configure script to include devteam
2818 warnings in other configurations.
2821 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2822 makes it possible to install openssl libraries in locations which
2823 have names other than "lib", for example "/usr/lib64" which some
2825 [Steve Henson, based on patch from Jeremy Utley]
2827 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2828 X690 8.9.12 and can produce some misleading textual output of OIDs.
2829 [Steve Henson, reported by Dan Kaminsky]
2831 *) Delete MD2 from algorithm tables. This follows the recommendation in
2832 several standards that it is not used in new applications due to
2833 several cryptographic weaknesses. For binary compatibility reasons
2834 the MD2 API is still compiled in by default.
2837 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2841 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2842 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2844 [Guenter <lists@gknw.net>]
2846 *) Fix the server certificate chain building code to use X509_verify_cert(),
2847 it used to have an ad-hoc builder which was unable to cope with anything
2848 other than a simple chain.
2849 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2851 *) Don't check self signed certificate signatures in X509_verify_cert()
2852 by default (a flag can override this): it just wastes time without
2853 adding any security. As a useful side effect self signed root CAs
2854 with non-FIPS digests are now usable in FIPS mode.
2857 *) In dtls1_process_out_of_seq_message() the check if the current message
2858 is already buffered was missing. For every new message was memory
2859 allocated, allowing an attacker to perform an denial of service attack
2860 with sending out of seq handshake messages until there is no memory
2861 left. Additionally every future messege was buffered, even if the
2862 sequence number made no sense and would be part of another handshake.
2863 So only messages with sequence numbers less than 10 in advance will be
2864 buffered. (CVE-2009-1378)
2865 [Robin Seggelmann, discovered by Daniel Mentz]
2867 *) Records are buffered if they arrive with a future epoch to be
2868 processed after finishing the corresponding handshake. There is
2869 currently no limitation to this buffer allowing an attacker to perform
2870 a DOS attack with sending records with future epochs until there is no
2871 memory left. This patch adds the pqueue_size() function to detemine
2872 the size of a buffer and limits the record buffer to 100 entries.
2874 [Robin Seggelmann, discovered by Daniel Mentz]
2876 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2877 parent structure is freed. (CVE-2009-1379)
2880 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2881 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2884 [Ilya O. <vrghost@gmail.com>]
2886 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2888 *) Disable renegotiation completely - this fixes a severe security
2889 problem (CVE-2009-3555) at the cost of breaking all
2890 renegotiation. Renegotiation can be re-enabled by setting
2891 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2892 run-time. This is really not recommended unless you know what
2896 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2898 *) Don't set val to NULL when freeing up structures, it is freed up by
2899 underlying code. If sizeof(void *) > sizeof(long) this can result in
2900 zeroing past the valid field. (CVE-2009-0789)
2901 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2903 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2904 checked correctly. This would allow some invalid signed attributes to
2905 appear to verify correctly. (CVE-2009-0591)
2906 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2908 *) Reject UniversalString and BMPString types with invalid lengths. This
2909 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2910 a legal length. (CVE-2009-0590)
2913 *) Set S/MIME signing as the default purpose rather than setting it
2914 unconditionally. This allows applications to override it at the store
2918 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2919 to handle some structures.
2922 *) Improve efficiency of mem_gets: don't search whole buffer each time
2924 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2926 *) New -hex option for openssl rand.
2929 *) Print out UTF8String and NumericString when parsing ASN1.
2932 *) Support NumericString type for name components.
2935 *) Allow CC in the environment to override the automatically chosen
2936 compiler. Note that nothing is done to ensure flags work with the
2940 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2942 *) Properly check EVP_VerifyFinal() and similar return values
2944 [Ben Laurie, Bodo Moeller, Google Security Team]
2946 *) Enable TLS extensions by default.
2949 *) Allow the CHIL engine to be loaded, whether the application is
2950 multithreaded or not. (This does not release the developer from the
2951 obligation to set up the dynamic locking callbacks.)
2952 [Sander Temme <sander@temme.net>]
2954 *) Use correct exit code if there is an error in dgst command.
2955 [Steve Henson; problem pointed out by Roland Dirlewanger]
2957 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2958 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2961 *) Add experimental JPAKE support, including demo authentication in
2962 s_client and s_server.
2965 *) Set the comparison function in v3_addr_canonize().
2966 [Rob Austein <sra@hactrn.net>]
2968 *) Add support for XMPP STARTTLS in s_client.
2969 [Philip Paeps <philip@freebsd.org>]
2971 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2972 to ensure that even with this option, only ciphersuites in the
2973 server's preference list will be accepted. (Note that the option
2974 applies only when resuming a session, so the earlier behavior was
2975 just about the algorithm choice for symmetric cryptography.)
2978 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2980 *) Fix NULL pointer dereference if a DTLS server received
2981 ChangeCipherSpec as first record (CVE-2009-1386).
2984 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2985 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2988 *) The fix in 0.9.8c that supposedly got rid of unsafe
2989 double-checked locking was incomplete for RSA blinding,
2990 addressing just one layer of what turns out to have been
2991 doubly unsafe triple-checked locking.
2993 So now fix this for real by retiring the MONT_HELPER macro
2994 in crypto/rsa/rsa_eay.c.
2996 [Bodo Moeller; problem pointed out by Marius Schilder]
2998 *) Various precautionary measures:
3000 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3002 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3003 (NB: This would require knowledge of the secret session ticket key
3004 to exploit, in which case you'd be SOL either way.)
3006 - Change bn_nist.c so that it will properly handle input BIGNUMs
3007 outside the expected range.
3009 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3012 [Neel Mehta, Bodo Moeller]
3014 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3015 the load fails. Useful for distros.
3016 [Ben Laurie and the FreeBSD team]
3018 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3021 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3024 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3026 This work was sponsored by Logica.
3029 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3030 keystores. Support for SSL/TLS client authentication too.
3031 Not compiled unless enable-capieng specified to Configure.
3033 This work was sponsored by Logica.
3036 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3037 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3038 attribute creation routines such as certifcate requests and PKCS#12
3042 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3044 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3045 handshake which could lead to a cilent crash as found using the
3046 Codenomicon TLS test suite (CVE-2008-1672)
3047 [Steve Henson, Mark Cox]
3049 *) Fix double free in TLS server name extensions which could lead to
3050 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3053 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3055 Clear the error queue to ensure that error entries left from
3056 older function calls do not interfere with the correct operation.
3057 [Lutz Jaenicke, Erik de Castro Lopo]
3059 *) Remove root CA certificates of commercial CAs:
3061 The OpenSSL project does not recommend any specific CA and does not
3062 have any policy with respect to including or excluding any CA.
3063 Therefore it does not make any sense to ship an arbitrary selection
3064 of root CA certificates with the OpenSSL software.
3067 *) RSA OAEP patches to fix two separate invalid memory reads.
3068 The first one involves inputs when 'lzero' is greater than
3069 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3070 before the beginning of from). The second one involves inputs where
3071 the 'db' section contains nothing but zeroes (there is a one-byte
3072 invalid read after the end of 'db').
3073 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3075 *) Partial backport from 0.9.9-dev:
3077 Introduce bn_mul_mont (dedicated Montgomery multiplication
3078 procedure) as a candidate for BIGNUM assembler implementation.
3079 While 0.9.9-dev uses assembler for various architectures, only
3080 x86_64 is available by default here in the 0.9.8 branch, and
3081 32-bit x86 is available through a compile-time setting.
3083 To try the 32-bit x86 assembler implementation, use Configure
3084 option "enable-montasm" (which exists only for this backport).
3086 As "enable-montasm" for 32-bit x86 disclaims code stability
3087 anyway, in this constellation we activate additional code
3088 backported from 0.9.9-dev for further performance improvements,
3089 namely BN_from_montgomery_word. (To enable this otherwise,
3090 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3092 [Andy Polyakov (backport partially by Bodo Moeller)]
3094 *) Add TLS session ticket callback. This allows an application to set
3095 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3096 values. This is useful for key rollover for example where several key
3097 sets may exist with different names.
3100 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3101 This was broken until now in 0.9.8 releases, such that the only way
3102 a registered ENGINE could be used (assuming it initialises
3103 successfully on the host) was to explicitly set it as the default
3104 for the relevant algorithms. This is in contradiction with 0.9.7
3105 behaviour and the documentation. With this fix, when an ENGINE is
3106 registered into a given algorithm's table of implementations, the
3107 'uptodate' flag is reset so that auto-discovery will be used next
3108 time a new context for that algorithm attempts to select an
3110 [Ian Lister (tweaked by Geoff Thorpe)]
3112 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3113 implemention in the following ways:
3115 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3118 Lack of BER streaming support means one pass streaming processing is
3119 only supported if data is detached: setting the streaming flag is
3120 ignored for embedded content.
3122 CMS support is disabled by default and must be explicitly enabled
3123 with the enable-cms configuration option.
3126 *) Update the GMP engine glue to do direct copies between BIGNUM and
3127 mpz_t when openssl and GMP use the same limb size. Otherwise the
3128 existing "conversion via a text string export" trick is still used.
3129 [Paul Sheer <paulsheer@gmail.com>]
3131 *) Zlib compression BIO. This is a filter BIO which compressed and
3132 uncompresses any data passed through it.
3135 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3136 RFC3394 compatible AES key wrapping.
3139 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3140 sets string data without copying. X509_ALGOR_set0() and
3141 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3142 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3143 from an X509_ATTRIBUTE structure optionally checking it occurs only
3144 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3148 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3149 to get the expected BN_FLG_CONSTTIME behavior.
3150 [Bodo Moeller (Google)]
3154 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3155 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3156 - added some more tests to do_tests.pl
3157 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3158 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3159 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3160 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3161 - various changes to netware.pl to enable gcc-cross builds on Win32
3163 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3164 - various changes to fix missing prototype warnings
3165 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3166 - added AES, WHIRLPOOL and CPUID assembler code to build files
3167 - added missing AES assembler make rules to mk1mf.pl
3168 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3169 [Guenter Knauf <eflash@gmx.net>]
3171 *) Implement certificate status request TLS extension defined in RFC3546.
3172 A client can set the appropriate parameters and receive the encoded
3173 OCSP response via a callback. A server can query the supplied parameters
3174 and set the encoded OCSP response in the callback. Add simplified examples
3175 to s_client and s_server.
3178 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3180 *) Fix various bugs:
3181 + Binary incompatibility of ssl_ctx_st structure
3182 + DTLS interoperation with non-compliant servers
3183 + Don't call get_session_cb() without proposed session
3184 + Fix ia64 assembler code
3185 [Andy Polyakov, Steve Henson]
3187 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3189 *) DTLS Handshake overhaul. There were longstanding issues with
3190 OpenSSL DTLS implementation, which were making it impossible for
3191 RFC 4347 compliant client to communicate with OpenSSL server.
3192 Unfortunately just fixing these incompatibilities would "cut off"
3193 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3194 server keeps tolerating non RFC compliant syntax. The opposite is
3195 not true, 0.9.8f client can not communicate with earlier server.
3196 This update even addresses CVE-2007-4995.
3199 *) Changes to avoid need for function casts in OpenSSL: some compilers
3200 (gcc 4.2 and later) reject their use.
3201 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3204 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3205 RFC4507bis. The encrypted ticket format is an encrypted encoded
3206 SSL_SESSION structure, that way new session features are automatically
3209 If a client application caches session in an SSL_SESSION structure
3210 support is transparent because tickets are now stored in the encoded
3213 The SSL_CTX structure automatically generates keys for ticket
3214 protection in servers so again support should be possible
3215 with no application modification.
3217 If a client or server wishes to disable RFC4507 support then the option
3218 SSL_OP_NO_TICKET can be set.
3220 Add a TLS extension debugging callback to allow the contents of any client
3221 or server extensions to be examined.
3223 This work was sponsored by Google.
3226 *) Add initial support for TLS extensions, specifically for the server_name
3227 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3228 have new members for a host name. The SSL data structure has an
3229 additional member SSL_CTX *initial_ctx so that new sessions can be
3230 stored in that context to allow for session resumption, even after the
3231 SSL has been switched to a new SSL_CTX in reaction to a client's
3232 server_name extension.
3234 New functions (subject to change):
3236 SSL_get_servername()
3237 SSL_get_servername_type()
3240 New CTRL codes and macros (subject to change):
3242 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3243 - SSL_CTX_set_tlsext_servername_callback()
3244 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3245 - SSL_CTX_set_tlsext_servername_arg()
3246 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3248 openssl s_client has a new '-servername ...' option.
3250 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3251 '-key2 ...', '-servername_fatal' (subject to change). This allows
3252 testing the HostName extension for a specific single host name ('-cert'
3253 and '-key' remain fallbacks for handshakes without HostName
3254 negotiation). If the unrecogninzed_name alert has to be sent, this by
3255 default is a warning; it becomes fatal with the '-servername_fatal'
3258 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3260 *) Add AES and SSE2 assembly language support to VC++ build.
3263 *) Mitigate attack on final subtraction in Montgomery reduction.
3266 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3267 (which previously caused an internal error).
3270 *) Squeeze another 10% out of IGE mode when in != out.
3273 *) AES IGE mode speedup.
3274 [Dean Gaudet (Google)]
3276 *) Add the Korean symmetric 128-bit cipher SEED (see
3277 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3278 add SEED ciphersuites from RFC 4162:
3280 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3281 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3282 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3283 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3285 To minimize changes between patchlevels in the OpenSSL 0.9.8
3286 series, SEED remains excluded from compilation unless OpenSSL
3287 is configured with 'enable-seed'.
3288 [KISA, Bodo Moeller]
3290 *) Mitigate branch prediction attacks, which can be practical if a
3291 single processor is shared, allowing a spy process to extract
3292 information. For detailed background information, see
3293 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3294 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3295 and Necessary Software Countermeasures"). The core of the change
3296 are new versions BN_div_no_branch() and
3297 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3298 respectively, which are slower, but avoid the security-relevant
3299 conditional branches. These are automatically called by BN_div()
3300 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3301 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3302 remove a conditional branch.
3304 BN_FLG_CONSTTIME is the new name for the previous
3305 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3306 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3307 in the exponent causes BN_mod_exp_mont() to use the alternative
3308 implementation in BN_mod_exp_mont_consttime().) The old name
3309 remains as a deprecated alias.
3311 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3312 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3313 constant-time implementations for more than just exponentiation.
3314 Here too the old name is kept as a deprecated alias.
3316 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3317 the BN_BLINDING structure gets an independent copy of the
3318 modulus. This means that the previous "BIGNUM *m" argument to
3319 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3320 essentially becomes "const BIGNUM *m", although we can't actually
3321 change this in the header file before 0.9.9. It allows
3322 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3323 enable BN_FLG_CONSTTIME.
3325 [Matthew D Wood (Intel Corp)]
3327 *) In the SSL/TLS server implementation, be strict about session ID
3328 context matching (which matters if an application uses a single
3329 external cache for different purposes). Previously,
3330 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3331 set. This did ensure strict client verification, but meant that,
3332 with applications using a single external cache for quite
3333 different requirements, clients could circumvent ciphersuite
3334 restrictions for a given session ID context by starting a session
3335 in a different context.
3338 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3339 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3340 authentication-only ciphersuites.
3343 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3344 not complete and could lead to a possible single byte overflow
3345 (CVE-2007-5135) [Ben Laurie]
3347 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3349 *) Since AES128 and AES256 (and similarly Camellia128 and
3350 Camellia256) share a single mask bit in the logic of
3351 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3352 kludge to work properly if AES128 is available and AES256 isn't
3353 (or if Camellia128 is available and Camellia256 isn't).
3356 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3357 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3358 When a point or a seed is encoded in a BIT STRING, we need to
3359 prevent the removal of trailing zero bits to get the proper DER
3360 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3361 of a NamedBitList, for which trailing 0 bits need to be removed.)
3364 *) Have SSL/TLS server implementation tolerate "mismatched" record
3365 protocol version while receiving ClientHello even if the
3366 ClientHello is fragmented. (The server can't insist on the
3367 particular protocol version it has chosen before the ServerHello
3368 message has informed the client about his choice.)
3371 *) Add RFC 3779 support.
3372 [Rob Austein for ARIN, Ben Laurie]
3374 *) Load error codes if they are not already present instead of using a
3375 static variable. This allows them to be cleanly unloaded and reloaded.
3376 Improve header file function name parsing.
3379 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3380 or CAPABILITY handshake as required by RFCs.
3383 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3385 *) Introduce limits to prevent malicious keys being able to
3386 cause a denial of service. (CVE-2006-2940)
3387 [Steve Henson, Bodo Moeller]
3389 *) Fix ASN.1 parsing of certain invalid structures that can result
3390 in a denial of service. (CVE-2006-2937) [Steve Henson]
3392 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3393 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3395 *) Fix SSL client code which could crash if connecting to a
3396 malicious SSLv2 server. (CVE-2006-4343)
3397 [Tavis Ormandy and Will Drewry, Google Security Team]
3399 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3400 match only those. Before that, "AES256-SHA" would be interpreted
3401 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3402 the same strength classification in 0.9.7h) as we currently only
3403 have a single AES bit in the ciphersuite description bitmap.
3404 That change, however, also applied to ciphersuite strings such as
3405 "RC4-MD5" that intentionally matched multiple ciphersuites --
3406 namely, SSL 2.0 ciphersuites in addition to the more common ones
3407 from SSL 3.0/TLS 1.0.
3409 So we change the selection algorithm again: Naming an explicit
3410 ciphersuite selects this one ciphersuite, and any other similar
3411 ciphersuite (same bitmap) from *other* protocol versions.
3412 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3413 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3415 Since SSL 2.0 does not have any ciphersuites for which the
3416 128/256 bit distinction would be relevant, this works for now.
3417 The proper fix will be to use different bits for AES128 and
3418 AES256, which would have avoided the problems from the beginning;
3419 however, bits are scarce, so we can only do this in a new release
3420 (not just a patchlevel) when we can change the SSL_CIPHER
3421 definition to split the single 'unsigned long mask' bitmap into
3422 multiple values to extend the available space.
3426 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3428 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3429 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3431 *) Add AES IGE and biIGE modes.
3434 *) Change the Unix randomness entropy gathering to use poll() when
3435 possible instead of select(), since the latter has some
3436 undesirable limitations.
3437 [Darryl Miles via Richard Levitte and Bodo Moeller]
3439 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3440 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3441 cannot be implicitly activated as part of, e.g., the "AES" alias.
3442 However, please upgrade to OpenSSL 0.9.9[-dev] for
3443 non-experimental use of the ECC ciphersuites to get TLS extension
3444 support, which is required for curve and point format negotiation
3445 to avoid potential handshake problems.
3448 *) Disable rogue ciphersuites:
3450 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3451 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3452 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3454 The latter two were purportedly from
3455 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3458 Also deactivate the remaining ciphersuites from
3459 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3460 unofficial, and the ID has long expired.
3463 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3464 dual-core machines) and other potential thread-safety issues.
3467 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3468 versions), which is now available for royalty-free use
3469 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3470 Also, add Camellia TLS ciphersuites from RFC 4132.
3472 To minimize changes between patchlevels in the OpenSSL 0.9.8
3473 series, Camellia remains excluded from compilation unless OpenSSL
3474 is configured with 'enable-camellia'.
3477 *) Disable the padding bug check when compression is in use. The padding
3478 bug check assumes the first packet is of even length, this is not
3479 necessarily true if compresssion is enabled and can result in false
3480 positives causing handshake failure. The actual bug test is ancient
3481 code so it is hoped that implementations will either have fixed it by
3482 now or any which still have the bug do not support compression.
3485 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3487 *) When applying a cipher rule check to see if string match is an explicit
3488 cipher suite and only match that one cipher suite if it is.
3491 *) Link in manifests for VC++ if needed.
3492 [Austin Ziegler <halostatue@gmail.com>]
3494 *) Update support for ECC-based TLS ciphersuites according to
3495 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3496 TLS extensions, which are supported starting with the 0.9.9
3497 branch, not in the OpenSSL 0.9.8 branch).
3500 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3501 opaque EVP_CIPHER_CTX handling.
3504 *) Fixes and enhancements to zlib compression code. We now only use
3505 "zlib1.dll" and use the default __cdecl calling convention on Win32
3506 to conform with the standards mentioned here:
3507 http://www.zlib.net/DLL_FAQ.txt
3508 Static zlib linking now works on Windows and the new --with-zlib-include
3509 --with-zlib-lib options to Configure can be used to supply the location
3510 of the headers and library. Gracefully handle case where zlib library
3514 *) Several fixes and enhancements to the OID generation code. The old code
3515 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3516 handle numbers larger than ULONG_MAX, truncated printing and had a
3517 non standard OBJ_obj2txt() behaviour.
3520 *) Add support for building of engines under engine/ as shared libraries
3521 under VC++ build system.
3524 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3525 Hopefully, we will not see any false combination of paths any more.
3528 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3530 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3531 (part of SSL_OP_ALL). This option used to disable the
3532 countermeasure against man-in-the-middle protocol-version
3533 rollback in the SSL 2.0 server implementation, which is a bad
3534 idea. (CVE-2005-2969)
3536 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3537 for Information Security, National Institute of Advanced Industrial
3538 Science and Technology [AIST], Japan)]
3540 *) Add two function to clear and return the verify parameter flags.
3543 *) Keep cipherlists sorted in the source instead of sorting them at
3544 runtime, thus removing the need for a lock.
3547 *) Avoid some small subgroup attacks in Diffie-Hellman.
3548 [Nick Mathewson and Ben Laurie]
3550 *) Add functions for well-known primes.
3553 *) Extended Windows CE support.
3554 [Satoshi Nakamura and Andy Polyakov]
3556 *) Initialize SSL_METHOD structures at compile time instead of during
3557 runtime, thus removing the need for a lock.
3560 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3561 attempting to decrypt each encrypted key in turn. Add support to
3565 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3567 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3570 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3573 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3574 key into the same file any more.
3577 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3580 *) Add -utf8 command line and config file option to 'ca'.
3581 [Stefan <stf@udoma.org]
3583 *) Removed the macro des_crypt(), as it seems to conflict with some
3584 libraries. Use DES_crypt().
3587 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3588 involves renaming the source and generated shared-libs for
3589 both. The engines will accept the corrected or legacy ids
3590 ('ncipher' and '4758_cca' respectively) when binding. NB,
3591 this only applies when building 'shared'.
3592 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3594 *) Add attribute functions to EVP_PKEY structure. Modify
3595 PKCS12_create() to recognize a CSP name attribute and
3596 use it. Make -CSP option work again in pkcs12 utility.
3599 *) Add new functionality to the bn blinding code:
3600 - automatic re-creation of the BN_BLINDING parameters after
3601 a fixed number of uses (currently 32)
3602 - add new function for parameter creation
3603 - introduce flags to control the update behaviour of the
3604 BN_BLINDING parameters
3605 - hide BN_BLINDING structure
3606 Add a second BN_BLINDING slot to the RSA structure to improve
3607 performance when a single RSA object is shared among several
3611 *) Add support for DTLS.
3612 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3614 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3615 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3618 *) Remove buggy and incompletet DH cert support from
3619 ssl/ssl_rsa.c and ssl/s3_both.c
3622 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3623 the apps/openssl applications.
3626 *) Compile clean with "-Wall -Wmissing-prototypes
3627 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3628 DEBUG_SAFESTACK must also be set.
3631 *) Change ./Configure so that certain algorithms can be disabled by default.
3632 The new counterpiece to "no-xxx" is "enable-xxx".
3634 The patented RC5 and MDC2 algorithms will now be disabled unless
3635 "enable-rc5" and "enable-mdc2", respectively, are specified.
3637 (IDEA remains enabled despite being patented. This is because IDEA
3638 is frequently required for interoperability, and there is no license
3639 fee for non-commercial use. As before, "no-idea" can be used to
3640 avoid this algorithm.)
3644 *) Add processing of proxy certificates (see RFC 3820). This work was
3645 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3646 EGEE (Enabling Grids for E-science in Europe).
3649 *) RC4 performance overhaul on modern architectures/implementations, such
3650 as Intel P4, IA-64 and AMD64.
3653 *) New utility extract-section.pl. This can be used specify an alternative
3654 section number in a pod file instead of having to treat each file as
3655 a separate case in Makefile. This can be done by adding two lines to the
3658 =for comment openssl_section:XXX
3660 The blank line is mandatory.
3664 *) New arguments -certform, -keyform and -pass for s_client and s_server
3665 to allow alternative format key and certificate files and passphrase
3669 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3670 update associated structures and add various utility functions.
3672 Add new policy related verify parameters, include policy checking in
3673 standard verify code. Enhance 'smime' application with extra parameters
3674 to support policy checking and print out.
3677 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3678 Nehemiah processors. These extensions support AES encryption in hardware
3679 as well as RNG (though RNG support is currently disabled).
3680 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3682 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3685 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3686 [Andy Polyakov and a number of other people]
3688 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3689 implementation contributed by IBM.
3690 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3692 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3693 exponent rather than 'unsigned long'. There is a corresponding change to
3694 the new 'rsa_keygen' element of the RSA_METHOD structure.
3695 [Jelte Jansen, Geoff Thorpe]
3697 *) Functionality for creating the initial serial number file is now
3698 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3700 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3701 number file to 1, which is bound to cause problems. To avoid
3702 the problems while respecting compatibility between different 0.9.7
3703 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3704 CA.pl for serial number initialization. With the new release 0.9.8,
3705 we can fix the problem directly in the 'ca' utility.)
3708 *) Reduced header interdepencies by declaring more opaque objects in
3709 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3710 give fewer recursive includes, which could break lazy source code - so
3711 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3712 developers should define this symbol when building and using openssl to
3713 ensure they track the recommended behaviour, interfaces, [etc], but
3714 backwards-compatible behaviour prevails when this isn't defined.
3717 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3720 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3721 This will generate a random key of the appropriate length based on the
3722 cipher context. The EVP_CIPHER can provide its own random key generation
3723 routine to support keys of a specific form. This is used in the des and
3724 3des routines to generate a key of the correct parity. Update S/MIME
3725 code to use new functions and hence generate correct parity DES keys.
3726 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3727 valid (weak or incorrect parity).
3730 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3731 as looking them up. This is useful when the verified structure may contain
3732 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3733 present unless the new PKCS7_NO_CRL flag is asserted.
3736 *) Extend ASN1 oid configuration module. It now additionally accepts the
3739 shortName = some long name, 1.2.3.4
3742 *) Reimplemented the BN_CTX implementation. There is now no more static
3743 limitation on the number of variables it can handle nor the depth of the
3744 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3745 information can now expand as required, and rather than having a single
3746 static array of bignums, BN_CTX now uses a linked-list of such arrays
3747 allowing it to expand on demand whilst maintaining the usefulness of
3748 BN_CTX's "bundling".
3751 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3752 to allow all RSA operations to function using a single BN_CTX.
3755 *) Preliminary support for certificate policy evaluation and checking. This
3756 is initially intended to pass the tests outlined in "Conformance Testing
3757 of Relying Party Client Certificate Path Processing Logic" v1.07.
3760 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3761 remained unused and not that useful. A variety of other little bignum
3762 tweaks and fixes have also been made continuing on from the audit (see
3766 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3767 associated ASN1, EVP and SSL functions and old ASN1 macros.
3770 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3771 and this should never fail. So the return value from the use of
3772 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3773 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3776 *) BN_CTX_get() should return zero-valued bignums, providing the same
3777 initialised value as BN_new().
3778 [Geoff Thorpe, suggested by Ulf Möller]
3780 *) Support for inhibitAnyPolicy certificate extension.
3783 *) An audit of the BIGNUM code is underway, for which debugging code is
3784 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3785 is considered valid when processing BIGNUMs, and causes execution to
3786 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3787 further steps are taken to deliberately pollute unused data in BIGNUM
3788 structures to try and expose faulty code further on. For now, openssl will
3789 (in its default mode of operation) continue to tolerate the inconsistent
3790 forms that it has tolerated in the past, but authors and packagers should
3791 consider trying openssl and their own applications when compiled with
3792 these debugging symbols defined. It will help highlight potential bugs in
3793 their own code, and will improve the test coverage for OpenSSL itself. At
3794 some point, these tighter rules will become openssl's default to improve
3795 maintainability, though the assert()s and other overheads will remain only
3796 in debugging configurations. See bn.h for more details.
3797 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3799 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3800 that can only be obtained through BN_CTX_new() (which implicitly
3801 initialises it). The presence of this function only made it possible
3802 to overwrite an existing structure (and cause memory leaks).
3805 *) Because of the callback-based approach for implementing LHASH as a
3806 template type, lh_insert() adds opaque objects to hash-tables and
3807 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3808 to clean up those corresponding objects before destroying the hash table
3809 (and losing the object pointers). So some over-zealous constifications in
3810 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3811 objects as "const" and the lh_doall[_arg] callback wrappers are not
3812 prototyped to have "const" restrictions on the object pointers they are
3813 given (and so aren't required to cast them away any more).
3816 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3817 (speed) prefers to use its own implementation. The two implementations
3818 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3819 its object type properly exposed (MS_TM) instead of casting to/from "char
3820 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3821 aren't necessarily the greatest nomenclatures - but this is what was used
3822 internally to the implementation so I've used that for now.
3825 *) Ensure that deprecated functions do not get compiled when
3826 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3827 the self-tests were still using deprecated key-generation functions so
3828 these have been updated also.
3831 *) Reorganise PKCS#7 code to separate the digest location functionality
3832 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
3833 New function PKCS7_set_digest() to set the digest type for PKCS#7
3834 digestedData type. Add additional code to correctly generate the
3835 digestedData type and add support for this type in PKCS7 initialization
3839 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3840 structure of type "other".
3843 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3844 sure the loop does correctly stop and breaking ("division by zero")
3845 modulus operations are not performed. The (pre-generated) prime
3846 table crypto/bn/bn_prime.h was already correct, but it could not be
3847 re-generated on some platforms because of the "division by zero"
3848 situation in the script.
3849 [Ralf S. Engelschall]
3851 *) Update support for ECC-based TLS ciphersuites according to
3852 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3853 SHA-1 now is only used for "small" curves (where the
3854 representation of a field element takes up to 24 bytes); for
3855 larger curves, the field element resulting from ECDH is directly
3856 used as premaster secret.
3857 [Douglas Stebila (Sun Microsystems Laboratories)]
3859 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3860 curve secp160r1 to the tests.
3861 [Douglas Stebila (Sun Microsystems Laboratories)]
3863 *) Add the possibility to load symbols globally with DSO.
3864 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3866 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3867 control of the error stack.
3870 *) Add support for STORE in ENGINE.
3873 *) Add the STORE type. The intention is to provide a common interface
3874 to certificate and key stores, be they simple file-based stores, or
3875 HSM-type store, or LDAP stores, or...
3876 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3879 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3880 pass a list of arguments to any function as well as provide a way
3881 for a function to pass data back to the caller.
3884 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3885 works like BUF_strdup() but can be used to duplicate a portion of
3886 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3890 *) Add the function sk_find_ex() which works like sk_find(), but will
3891 return an index to an element even if an exact match couldn't be
3892 found. The index is guaranteed to point at the element where the
3893 searched-for key would be inserted to preserve sorting order.
3896 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3897 takes an extra flags argument for optional functionality. Currently,
3898 the following flags are defined:
3900 OBJ_BSEARCH_VALUE_ON_NOMATCH
3901 This one gets OBJ_bsearch_ex() to return a pointer to the first
3902 element where the comparing function returns a negative or zero
3905 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3906 This one gets OBJ_bsearch_ex() to return a pointer to the first
3907 element where the comparing function returns zero. This is useful
3908 if there are more than one element where the comparing function
3912 *) Make it possible to create self-signed certificates with 'openssl ca'
3913 in such a way that the self-signed certificate becomes part of the
3914 CA database and uses the same mechanisms for serial number generation
3915 as all other certificate signing. The new flag '-selfsign' enables
3916 this functionality. Adapt CA.sh and CA.pl.in.
3919 *) Add functionality to check the public key of a certificate request
3920 against a given private. This is useful to check that a certificate
3921 request can be signed by that key (self-signing).
3924 *) Make it possible to have multiple active certificates with the same
3925 subject in the CA index file. This is done only if the keyword
3926 'unique_subject' is set to 'no' in the main CA section (default
3927 if 'CA_default') of the configuration file. The value is saved
3928 with the database itself in a separate index attribute file,
3929 named like the index file with '.attr' appended to the name.
3932 *) Generate muti valued AVAs using '+' notation in config files for
3936 *) Support for nameConstraints certificate extension.
3939 *) Support for policyConstraints certificate extension.
3942 *) Support for policyMappings certificate extension.
3945 *) Make sure the default DSA_METHOD implementation only uses its
3946 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3947 and change its own handlers to be NULL so as to remove unnecessary
3948 indirection. This lets alternative implementations fallback to the
3949 default implementation more easily.
3952 *) Support for directoryName in GeneralName related extensions
3956 *) Make it possible to link applications using Makefile.shared.
3957 Make that possible even when linking against static libraries!
3960 *) Support for single pass processing for S/MIME signing. This now
3961 means that S/MIME signing can be done from a pipe, in addition
3962 cleartext signing (multipart/signed type) is effectively streaming
3963 and the signed data does not need to be all held in memory.
3965 This is done with a new flag PKCS7_STREAM. When this flag is set
3966 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3967 is done after the data is output (and digests calculated) in
3968 SMIME_write_PKCS7().
3971 *) Add full support for -rpath/-R, both in shared libraries and
3972 applications, at least on the platforms where it's known how
3976 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3977 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3978 will now compute a table of multiples of the generator that
3979 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3980 faster (notably in the case of a single point multiplication,
3981 scalar * generator).
3982 [Nils Larsch, Bodo Moeller]
3984 *) IPv6 support for certificate extensions. The various extensions
3985 which use the IP:a.b.c.d can now take IPv6 addresses using the
3986 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3990 *) Added an ENGINE that implements RSA by performing private key
3991 exponentiations with the GMP library. The conversions to and from
3992 GMP's mpz_t format aren't optimised nor are any montgomery forms
3993 cached, and on x86 it appears OpenSSL's own performance has caught up.
3994 However there are likely to be other architectures where GMP could
3995 provide a boost. This ENGINE is not built in by default, but it can be
3996 specified at Configure time and should be accompanied by the necessary
3997 linker additions, eg;
3998 ./config -DOPENSSL_USE_GMP -lgmp
4001 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4002 testing availability of engines with "-t" - the old behaviour is
4003 produced by increasing the feature's verbosity with "-tt".
4006 *) ECDSA routines: under certain error conditions uninitialized BN objects
4007 could be freed. Solution: make sure initialization is performed early
4008 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4012 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4013 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4014 software implementations. For DSA and DH, parameter generation can
4015 also be overriden by providing the appropriate method callbacks.
4018 *) Change the "progress" mechanism used in key-generation and
4019 primality testing to functions that take a new BN_GENCB pointer in
4020 place of callback/argument pairs. The new API functions have "_ex"
4021 postfixes and the older functions are reimplemented as wrappers for
4022 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4023 declarations of the old functions to help (graceful) attempts to
4024 migrate to the new functions. Also, the new key-generation API
4025 functions operate on a caller-supplied key-structure and return
4026 success/failure rather than returning a key or NULL - this is to
4027 help make "keygen" another member function of RSA_METHOD etc.
4029 Example for using the new callback interface:
4031 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4035 BN_GENCB_set(&my_cb, my_callback, my_arg);
4037 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4038 /* For the meaning of a, b in calls to my_callback(), see the
4039 * documentation of the function that calls the callback.
4040 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4041 * my_callback should return 1 if it wants BN_is_prime_ex()
4042 * to continue, or 0 to stop.
4047 *) Change the ZLIB compression method to be stateful, and make it
4048 available to TLS with the number defined in
4049 draft-ietf-tls-compression-04.txt.
4052 *) Add the ASN.1 structures and functions for CertificatePair, which
4053 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4055 CertificatePair ::= SEQUENCE {
4056 forward [0] Certificate OPTIONAL,
4057 reverse [1] Certificate OPTIONAL,
4058 -- at least one of the pair shall be present -- }
4060 Also implement the PEM functions to read and write certificate
4061 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4063 This needed to be defined, mostly for the sake of the LDAP
4064 attribute crossCertificatePair, but may prove useful elsewhere as
4068 *) Make it possible to inhibit symlinking of shared libraries in
4069 Makefile.shared, for Cygwin's sake.
4072 *) Extend the BIGNUM API by creating a function
4073 void BN_set_negative(BIGNUM *a, int neg);
4074 and a macro that behave like
4075 int BN_is_negative(const BIGNUM *a);
4077 to avoid the need to access 'a->neg' directly in applications.
4080 *) Implement fast modular reduction for pseudo-Mersenne primes
4081 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4082 EC_GROUP_new_curve_GFp() will now automatically use this
4084 [Nils Larsch <nla@trustcenter.de>]
4086 *) Add new lock type (CRYPTO_LOCK_BN).
4089 *) Change the ENGINE framework to automatically load engines
4090 dynamically from specific directories unless they could be
4091 found to already be built in or loaded. Move all the
4092 current engines except for the cryptodev one to a new
4094 The engines in engines/ are built as shared libraries if
4095 the "shared" options was given to ./Configure or ./config.
4096 Otherwise, they are inserted in libcrypto.a.
4097 /usr/local/ssl/engines is the default directory for dynamic
4098 engines, but that can be overriden at configure time through
4099 the usual use of --prefix and/or --openssldir, and at run
4100 time with the environment variable OPENSSL_ENGINES.
4101 [Geoff Thorpe and Richard Levitte]
4103 *) Add Makefile.shared, a helper makefile to build shared
4104 libraries. Addapt Makefile.org.
4107 *) Add version info to Win32 DLLs.
4108 [Peter 'Luna' Runestig" <peter@runestig.com>]
4110 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4111 can be added using this API to created arbitrary PKCS#12
4112 files while avoiding the low level API.
4114 New options to PKCS12_create(), key or cert can be NULL and
4115 will then be omitted from the output file. The encryption
4116 algorithm NIDs can be set to -1 for no encryption, the mac
4117 iteration count can be set to 0 to omit the mac.
4119 Enhance pkcs12 utility by making the -nokeys and -nocerts
4120 options work when creating a PKCS#12 file. New option -nomac
4121 to omit the mac, NONE can be set for an encryption algorithm.
4122 New code is modified to use the enhanced PKCS12_create()
4123 instead of the low level API.
4126 *) Extend ASN1 encoder to support indefinite length constructed
4127 encoding. This can output sequences tags and octet strings in
4128 this form. Modify pk7_asn1.c to support indefinite length
4129 encoding. This is experimental and needs additional code to
4130 be useful, such as an ASN1 bio and some enhanced streaming
4133 Extend template encode functionality so that tagging is passed
4134 down to the template encoder.
4137 *) Let 'openssl req' fail if an argument to '-newkey' is not
4138 recognized instead of using RSA as a default.
4141 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4142 As these are not official, they are not included in "ALL";
4143 the "ECCdraft" ciphersuite group alias can be used to select them.
4144 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4146 *) Add ECDH engine support.
4147 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4149 *) Add ECDH in new directory crypto/ecdh/.
4150 [Douglas Stebila (Sun Microsystems Laboratories)]
4152 *) Let BN_rand_range() abort with an error after 100 iterations
4153 without success (which indicates a broken PRNG).
4156 *) Change BN_mod_sqrt() so that it verifies that the input value
4157 is really the square of the return value. (Previously,
4158 BN_mod_sqrt would show GIGO behaviour.)
4161 *) Add named elliptic curves over binary fields from X9.62, SECG,
4162 and WAP/WTLS; add OIDs that were still missing.
4164 [Sheueling Chang Shantz and Douglas Stebila
4165 (Sun Microsystems Laboratories)]
4167 *) Extend the EC library for elliptic curves over binary fields
4168 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4171 EC_GF2m_simple_method
4175 EC_GROUP_new_curve_GF2m
4176 EC_GROUP_set_curve_GF2m
4177 EC_GROUP_get_curve_GF2m
4178 EC_POINT_set_affine_coordinates_GF2m
4179 EC_POINT_get_affine_coordinates_GF2m
4180 EC_POINT_set_compressed_coordinates_GF2m
4182 Point compression for binary fields is disabled by default for
4183 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4186 As binary polynomials are represented as BIGNUMs, various members
4187 of the EC_GROUP and EC_POINT data structures can be shared
4188 between the implementations for prime fields and binary fields;
4189 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4190 are essentially identical to their ..._GFp counterparts.
4191 (For simplicity, the '..._GFp' prefix has been dropped from
4192 various internal method names.)
4194 An internal 'field_div' method (similar to 'field_mul' and
4195 'field_sqr') has been added; this is used only for binary fields.
4197 [Sheueling Chang Shantz and Douglas Stebila
4198 (Sun Microsystems Laboratories)]
4200 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4201 through methods ('mul', 'precompute_mult').
4203 The generic implementations (now internally called 'ec_wNAF_mul'
4204 and 'ec_wNAF_precomputed_mult') remain the default if these
4205 methods are undefined.
4207 [Sheueling Chang Shantz and Douglas Stebila
4208 (Sun Microsystems Laboratories)]
4210 *) New function EC_GROUP_get_degree, which is defined through
4211 EC_METHOD. For curves over prime fields, this returns the bit
4212 length of the modulus.
4214 [Sheueling Chang Shantz and Douglas Stebila
4215 (Sun Microsystems Laboratories)]
4217 *) New functions EC_GROUP_dup, EC_POINT_dup.
4218 (These simply call ..._new and ..._copy).
4220 [Sheueling Chang Shantz and Douglas Stebila
4221 (Sun Microsystems Laboratories)]
4223 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4224 Polynomials are represented as BIGNUMs (where the sign bit is not
4225 used) in the following functions [macros]:
4228 BN_GF2m_sub [= BN_GF2m_add]
4229 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4230 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4231 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4233 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4234 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4235 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4236 BN_GF2m_cmp [= BN_ucmp]
4238 (Note that only the 'mod' functions are actually for fields GF(2^m).
4239 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4241 For some functions, an the irreducible polynomial defining a
4242 field can be given as an 'unsigned int[]' with strictly
4243 decreasing elements giving the indices of those bits that are set;
4244 i.e., p[] represents the polynomial
4245 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4247 p[0] > p[1] > ... > p[k] = 0.
4248 This applies to the following functions:
4253 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4254 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4256 BN_GF2m_mod_sqrt_arr
4257 BN_GF2m_mod_solve_quad_arr
4261 Conversion can be performed by the following functions:
4266 bntest.c has additional tests for binary polynomial arithmetic.
4268 Two implementations for BN_GF2m_mod_div() are available.
4269 The default algorithm simply uses BN_GF2m_mod_inv() and
4270 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4271 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4272 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4274 [Sheueling Chang Shantz and Douglas Stebila
4275 (Sun Microsystems Laboratories)]
4277 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4278 functionality is disabled at compile-time.
4279 [Douglas Stebila <douglas.stebila@sun.com>]
4281 *) Change default behaviour of 'openssl asn1parse' so that more
4282 information is visible when viewing, e.g., a certificate:
4284 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4285 mode the content of non-printable OCTET STRINGs is output in a
4286 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4287 avoid the appearance of a printable string.
4288 [Nils Larsch <nla@trustcenter.de>]
4290 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4292 EC_GROUP_set_asn1_flag()
4293 EC_GROUP_get_asn1_flag()
4294 EC_GROUP_set_point_conversion_form()
4295 EC_GROUP_get_point_conversion_form()
4296 These control ASN1 encoding details:
4297 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4298 has been set to OPENSSL_EC_NAMED_CURVE.
4299 - Points are encoded in uncompressed form by default; options for
4300 asn1_for are as for point2oct, namely
4301 POINT_CONVERSION_COMPRESSED
4302 POINT_CONVERSION_UNCOMPRESSED
4303 POINT_CONVERSION_HYBRID
4305 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4308 EC_GROUP_get0_seed()
4309 EC_GROUP_get_seed_len()
4310 This is used only for ASN1 purposes (so far).
4311 [Nils Larsch <nla@trustcenter.de>]
4313 *) Add 'field_type' member to EC_METHOD, which holds the NID
4314 of the appropriate field type OID. The new function
4315 EC_METHOD_get_field_type() returns this value.
4316 [Nils Larsch <nla@trustcenter.de>]
4321 EC_POINT_point2hex()
4322 EC_POINT_hex2point()
4323 providing useful interfaces to EC_POINT_point2oct() and
4324 EC_POINT_oct2point().
4325 [Nils Larsch <nla@trustcenter.de>]
4327 *) Change internals of the EC library so that the functions
4328 EC_GROUP_set_generator()
4329 EC_GROUP_get_generator()
4330 EC_GROUP_get_order()
4331 EC_GROUP_get_cofactor()
4332 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4333 to methods, which would lead to unnecessary code duplication when
4334 adding different types of curves.
4335 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4337 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4338 arithmetic, and such that modified wNAFs are generated
4339 (which avoid length expansion in many cases).
4342 *) Add a function EC_GROUP_check_discriminant() (defined via
4343 EC_METHOD) that verifies that the curve discriminant is non-zero.
4345 Add a function EC_GROUP_check() that makes some sanity tests
4346 on a EC_GROUP, its generator and order. This includes
4347 EC_GROUP_check_discriminant().
4348 [Nils Larsch <nla@trustcenter.de>]
4350 *) Add ECDSA in new directory crypto/ecdsa/.
4352 Add applications 'openssl ecparam' and 'openssl ecdsa'
4353 (these are based on 'openssl dsaparam' and 'openssl dsa').
4355 ECDSA support is also included in various other files across the
4356 library. Most notably,
4357 - 'openssl req' now has a '-newkey ecdsa:file' option;
4358 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4359 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4360 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4361 them suitable for ECDSA where domain parameters must be
4362 extracted before the specific public key;
4363 - ECDSA engine support has been added.
4364 [Nils Larsch <nla@trustcenter.de>]
4366 *) Include some named elliptic curves, and add OIDs from X9.62,
4367 SECG, and WAP/WTLS. Each curve can be obtained from the new
4369 EC_GROUP_new_by_curve_name(),
4370 and the list of available named curves can be obtained with
4371 EC_get_builtin_curves().
4372 Also add a 'curve_name' member to EC_GROUP objects, which can be
4374 EC_GROUP_set_curve_name()
4375 EC_GROUP_get_curve_name()
4376 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4378 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4379 was actually never needed) and in BN_mul(). The removal in BN_mul()
4380 required a small change in bn_mul_part_recursive() and the addition
4381 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4382 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4383 bn_sub_words() and bn_add_words() except they take arrays with
4387 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4389 *) Cleanse PEM buffers before freeing them since they may contain
4391 [Benjamin Bennett <ben@psc.edu>]
4393 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4394 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4395 authentication-only ciphersuites.
4398 *) Since AES128 and AES256 share a single mask bit in the logic of
4399 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4400 kludge to work properly if AES128 is available and AES256 isn't.
4403 *) Expand security boundary to match 1.1.1 module.
4406 *) Remove redundant features: hash file source, editing of test vectors
4407 modify fipsld to use external fips_premain.c signature.
4410 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4411 run algorithm test programs.
4414 *) Make algorithm test programs more tolerant of whitespace.
4417 *) Have SSL/TLS server implementation tolerate "mismatched" record
4418 protocol version while receiving ClientHello even if the
4419 ClientHello is fragmented. (The server can't insist on the
4420 particular protocol version it has chosen before the ServerHello
4421 message has informed the client about his choice.)
4424 *) Load error codes if they are not already present instead of using a
4425 static variable. This allows them to be cleanly unloaded and reloaded.
4428 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4430 *) Introduce limits to prevent malicious keys being able to
4431 cause a denial of service. (CVE-2006-2940)
4432 [Steve Henson, Bodo Moeller]
4434 *) Fix ASN.1 parsing of certain invalid structures that can result
4435 in a denial of service. (CVE-2006-2937) [Steve Henson]
4437 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4438 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4440 *) Fix SSL client code which could crash if connecting to a
4441 malicious SSLv2 server. (CVE-2006-4343)
4442 [Tavis Ormandy and Will Drewry, Google Security Team]
4444 *) Change ciphersuite string processing so that an explicit
4445 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4446 will no longer include "AES128-SHA"), and any other similar
4447 ciphersuite (same bitmap) from *other* protocol versions (so that
4448 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4449 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4450 changes from 0.9.8b and 0.9.8d.
4453 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4455 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4456 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4458 *) Change the Unix randomness entropy gathering to use poll() when
4459 possible instead of select(), since the latter has some
4460 undesirable limitations.
4461 [Darryl Miles via Richard Levitte and Bodo Moeller]
4463 *) Disable rogue ciphersuites:
4465 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4466 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4467 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4469 The latter two were purportedly from
4470 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4473 Also deactive the remaining ciphersuites from
4474 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4475 unofficial, and the ID has long expired.
4478 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4479 dual-core machines) and other potential thread-safety issues.
4482 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4484 *) Adapt fipsld and the build system to link against the validated FIPS
4485 module in FIPS mode.
4488 *) Fixes for VC++ 2005 build under Windows.
4491 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4492 from a Windows bash shell such as MSYS. It is autodetected from the
4493 "config" script when run from a VC++ environment. Modify standard VC++
4494 build to use fipscanister.o from the GNU make build.
4497 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4499 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4500 The value now differs depending on if you build for FIPS or not.
4501 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4502 safely run with a non-FIPSed libcrypto, as it may crash because of
4503 the difference induced by this change.
4506 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4508 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4509 (part of SSL_OP_ALL). This option used to disable the
4510 countermeasure against man-in-the-middle protocol-version
4511 rollback in the SSL 2.0 server implementation, which is a bad
4512 idea. (CVE-2005-2969)
4514 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4515 for Information Security, National Institute of Advanced Industrial
4516 Science and Technology [AIST], Japan)]
4518 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4519 mainly for FIPS compliance and not fully integrated at this stage.
4522 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4523 the exponentiation using a fixed-length exponent. (Otherwise,
4524 the information leaked through timing could expose the secret key
4525 after many signatures; cf. Bleichenbacher's attack on DSA with
4529 *) Make a new fixed-window mod_exp implementation the default for
4530 RSA, DSA, and DH private-key operations so that the sequence of
4531 squares and multiplies and the memory access pattern are
4532 independent of the particular secret key. This will mitigate
4533 cache-timing and potential related attacks.
4535 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4536 and this is automatically used by BN_mod_exp_mont() if the new flag
4537 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4538 will use this BN flag for private exponents unless the flag
4539 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4540 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4542 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4544 *) Change the client implementation for SSLv23_method() and
4545 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4546 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4547 (Previously, the SSL 2.0 backwards compatible Client Hello
4548 message format would be used even with SSL_OP_NO_SSLv2.)
4551 *) Add support for smime-type MIME parameter in S/MIME messages which some
4555 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4556 a threadsafe manner. Modify rsa code to use new function and add calls
4557 to dsa and dh code (which had race conditions before).
4560 *) Include the fixed error library code in the C error file definitions
4561 instead of fixing them up at runtime. This keeps the error code
4562 structures constant.
4565 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4567 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4570 *) Fixes for newer kerberos headers. NB: the casts are needed because
4571 the 'length' field is signed on one version and unsigned on another
4572 with no (?) obvious way to tell the difference, without these VC++
4573 complains. Also the "definition" of FAR (blank) is no longer included
4574 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4575 some needed definitions.
4578 *) Undo Cygwin change.
4581 *) Added support for proxy certificates according to RFC 3820.
4582 Because they may be a security thread to unaware applications,
4583 they must be explicitely allowed in run-time. See
4584 docs/HOWTO/proxy_certificates.txt for further information.
4587 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4589 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4590 server and client random values. Previously
4591 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4592 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4594 This change has negligible security impact because:
4596 1. Server and client random values still have 24 bytes of pseudo random
4599 2. Server and client random values are sent in the clear in the initial
4602 3. The master secret is derived using the premaster secret (48 bytes in
4603 size for static RSA ciphersuites) as well as client server and random
4606 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4609 [Stephen Henson, reported by UK NISCC]
4611 *) Use Windows randomness collection on Cygwin.
4614 *) Fix hang in EGD/PRNGD query when communication socket is closed
4615 prematurely by EGD/PRNGD.
4616 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4618 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4621 *) Back-port of selected performance improvements from development
4622 branch, as well as improved support for PowerPC platforms.
4625 *) Add lots of checks for memory allocation failure, error codes to indicate
4626 failure and freeing up memory if a failure occurs.
4627 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4629 *) Add new -passin argument to dgst.
4632 *) Perform some character comparisons of different types in X509_NAME_cmp:
4633 this is needed for some certificates that reencode DNs into UTF8Strings
4634 (in violation of RFC3280) and can't or wont issue name rollover
4638 *) Make an explicit check during certificate validation to see that
4639 the CA setting in each certificate on the chain is correct. As a
4640 side effect always do the following basic checks on extensions,
4641 not just when there's an associated purpose to the check:
4643 - if there is an unhandled critical extension (unless the user
4644 has chosen to ignore this fault)
4645 - if the path length has been exceeded (if one is set at all)
4646 - that certain extensions fit the associated purpose (if one has
4650 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4652 *) Avoid a race condition when CRLs are checked in a multi threaded
4653 environment. This would happen due to the reordering of the revoked
4654 entries during signature checking and serial number lookup. Now the
4655 encoding is cached and the serial number sort performed under a lock.
4656 Add new STACK function sk_is_sorted().
4659 *) Add Delta CRL to the extension code.
4662 *) Various fixes to s3_pkt.c so alerts are sent properly.
4663 [David Holmes <d.holmes@f5.com>]
4665 *) Reduce the chances of duplicate issuer name and serial numbers (in
4666 violation of RFC3280) using the OpenSSL certificate creation utilities.
4667 This is done by creating a random 64 bit value for the initial serial
4668 number when a serial number file is created or when a self signed
4669 certificate is created using 'openssl req -x509'. The initial serial
4670 number file is created using 'openssl x509 -next_serial' in CA.pl
4671 rather than being initialized to 1.
4674 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4676 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4677 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4678 [Joe Orton, Steve Henson]
4680 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4682 [Joe Orton, Steve Henson]
4684 *) Make it possible to have multiple active certificates with the same
4685 subject in the CA index file. This is done only if the keyword
4686 'unique_subject' is set to 'no' in the main CA section (default
4687 if 'CA_default') of the configuration file. The value is saved
4688 with the database itself in a separate index attribute file,
4689 named like the index file with '.attr' appended to the name.
4692 *) X509 verify fixes. Disable broken certificate workarounds when
4693 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4694 keyUsage extension present. Don't accept CRLs with unhandled critical
4695 extensions: since verify currently doesn't process CRL extensions this
4696 rejects a CRL with *any* critical extensions. Add new verify error codes
4700 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4701 A clarification of RFC2560 will require the use of OCTET STRINGs and
4702 some implementations cannot handle the current raw format. Since OpenSSL
4703 copies and compares OCSP nonces as opaque blobs without any attempt at
4704 parsing them this should not create any compatibility issues.
4707 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4708 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4709 this HMAC (and other) operations are several times slower than OpenSSL
4713 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4714 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4716 *) Use the correct content when signing type "other".
4719 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4721 *) Fix various bugs revealed by running the NISCC test suite:
4723 Stop out of bounds reads in the ASN1 code when presented with
4724 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4726 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4728 If verify callback ignores invalid public key errors don't try to check
4729 certificate signature with the NULL public key.
4733 *) New -ignore_err option in ocsp application to stop the server
4734 exiting on the first error in a request.
4737 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4738 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4742 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4743 extra data after the compression methods not only for TLS 1.0
4744 but also for SSL 3.0 (as required by the specification).
4745 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4747 *) Change X509_certificate_type() to mark the key as exported/exportable
4748 when it's 512 *bits* long, not 512 bytes.
4751 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4752 blocks during encryption.
4755 *) Various fixes to base64 BIO and non blocking I/O. On write
4756 flushes were not handled properly if the BIO retried. On read
4757 data was not being buffered properly and had various logic bugs.
4758 This also affects blocking I/O when the data being decoded is a
4762 *) Various S/MIME bugfixes and compatibility changes:
4763 output correct application/pkcs7 MIME type if
4764 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4765 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4766 of files as .eml work). Correctly handle very long lines in MIME
4770 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4772 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4773 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4774 a protocol version number mismatch like a decryption error
4775 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4778 *) Turn on RSA blinding by default in the default implementation
4779 to avoid a timing attack. Applications that don't want it can call
4780 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4781 They would be ill-advised to do so in most cases.
4782 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4784 *) Change RSA blinding code so that it works when the PRNG is not
4785 seeded (in this case, the secret RSA exponent is abused as
4786 an unpredictable seed -- if it is not unpredictable, there
4787 is no point in blinding anyway). Make RSA blinding thread-safe
4788 by remembering the creator's thread ID in rsa->blinding and
4789 having all other threads use local one-time blinding factors
4790 (this requires more computation than sharing rsa->blinding, but
4791 avoids excessive locking; and if an RSA object is not shared
4792 between threads, blinding will still be very fast).
4795 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4796 ENGINE as defaults for all supported algorithms irrespective of
4797 the 'flags' parameter. 'flags' is now honoured, so applications
4798 should make sure they are passing it correctly.
4801 *) Target "mingw" now allows native Windows code to be generated in
4802 the Cygwin environment as well as with the MinGW compiler.
4805 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4807 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4808 via timing by performing a MAC computation even if incorrrect
4809 block cipher padding has been found. This is a countermeasure
4810 against active attacks where the attacker has to distinguish
4811 between bad padding and a MAC verification error. (CVE-2003-0078)
4813 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4814 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4815 Martin Vuagnoux (EPFL, Ilion)]
4817 *) Make the no-err option work as intended. The intention with no-err
4818 is not to have the whole error stack handling routines removed from
4819 libcrypto, it's only intended to remove all the function name and
4820 reason texts, thereby removing some of the footprint that may not
4821 be interesting if those errors aren't displayed anyway.
4823 NOTE: it's still possible for any application or module to have it's
4824 own set of error texts inserted. The routines are there, just not
4825 used by default when no-err is given.
4828 *) Add support for FreeBSD on IA64.
4829 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4831 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4832 Kerberos function mit_des_cbc_cksum(). Before this change,
4833 the value returned by DES_cbc_cksum() was like the one from
4834 mit_des_cbc_cksum(), except the bytes were swapped.
4835 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4837 *) Allow an application to disable the automatic SSL chain building.
4838 Before this a rather primitive chain build was always performed in
4839 ssl3_output_cert_chain(): an application had no way to send the
4840 correct chain if the automatic operation produced an incorrect result.
4842 Now the chain builder is disabled if either:
4844 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4846 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4848 The reasoning behind this is that an application would not want the
4849 auto chain building to take place if extra chain certificates are
4850 present and it might also want a means of sending no additional
4851 certificates (for example the chain has two certificates and the
4855 *) Add the possibility to build without the ENGINE framework.
4856 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4858 *) Under Win32 gmtime() can return NULL: check return value in
4859 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4862 *) DSA routines: under certain error conditions uninitialized BN objects
4863 could be freed. Solution: make sure initialization is performed early
4864 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4865 Nils Larsch <nla@trustcenter.de> via PR#459)
4868 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4869 checked on reconnect on the client side, therefore session resumption
4870 could still fail with a "ssl session id is different" error. This
4871 behaviour is masked when SSL_OP_ALL is used due to
4872 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4873 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4874 followup to PR #377.
4877 *) IA-32 assembler support enhancements: unified ELF targets, support
4878 for SCO/Caldera platforms, fix for Cygwin shared build.
4881 *) Add support for FreeBSD on sparc64. As a consequence, support for
4882 FreeBSD on non-x86 processors is separate from x86 processors on
4883 the config script, much like the NetBSD support.
4884 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4886 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4888 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4891 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4892 code (06) was taken as the first octet of the session ID and the last
4893 octet was ignored consequently. As a result SSLv2 client side session
4894 caching could not have worked due to the session ID mismatch between
4896 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4900 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4901 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4905 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4906 seems that in spite of existing for more than a year, many application
4907 author have done nothing to provide the necessary callbacks, which
4908 means that this particular engine will not work properly anywhere.
4909 This is a very unfortunate situation which forces us, in the name
4910 of usability, to give the hw_ncipher.c a static lock, which is part
4912 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4913 appear in 0.9.8 or later. We EXPECT application authors to have
4914 dealt properly with this when 0.9.8 is released (unless we actually
4915 make such changes in the libcrypto locking code that changes will
4916 have to be made anyway).
4919 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4920 octets have been read, EOF or an error occurs. Without this change
4921 some truncated ASN1 structures will not produce an error.
4924 *) Disable Heimdal support, since it hasn't been fully implemented.
4925 Still give the possibility to force the use of Heimdal, but with
4926 warnings and a request that patches get sent to openssl-dev.
4929 *) Add the VC-CE target, introduce the WINCE sysname, and add
4930 INSTALL.WCE and appropriate conditionals to make it build.
4931 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4933 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4934 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4935 edit numbers of the version.
4936 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4938 *) Introduce safe string copy and catenation functions
4939 (BUF_strlcpy() and BUF_strlcat()).
4940 [Ben Laurie (CHATS) and Richard Levitte]
4942 *) Avoid using fixed-size buffers for one-line DNs.
4943 [Ben Laurie (CHATS)]
4945 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4946 resizing buffers containing secrets, and use where appropriate.
4947 [Ben Laurie (CHATS)]
4949 *) Avoid using fixed size buffers for configuration file location.
4950 [Ben Laurie (CHATS)]
4952 *) Avoid filename truncation for various CA files.
4953 [Ben Laurie (CHATS)]
4955 *) Use sizeof in preference to magic numbers.
4956 [Ben Laurie (CHATS)]
4958 *) Avoid filename truncation in cert requests.
4959 [Ben Laurie (CHATS)]
4961 *) Add assertions to check for (supposedly impossible) buffer
4963 [Ben Laurie (CHATS)]
4965 *) Don't cache truncated DNS entries in the local cache (this could
4966 potentially lead to a spoofing attack).
4967 [Ben Laurie (CHATS)]
4969 *) Fix various buffers to be large enough for hex/decimal
4970 representations in a platform independent manner.
4971 [Ben Laurie (CHATS)]
4973 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4974 resizing buffers containing secrets, and use where appropriate.
4975 [Ben Laurie (CHATS)]
4977 *) Add BIO_indent() to avoid much slightly worrying code to do
4979 [Ben Laurie (CHATS)]
4981 *) Convert sprintf()/BIO_puts() to BIO_printf().
4982 [Ben Laurie (CHATS)]
4984 *) buffer_gets() could terminate with the buffer only half
4986 [Ben Laurie (CHATS)]
4988 *) Add assertions to prevent user-supplied crypto functions from
4989 overflowing internal buffers by having large block sizes, etc.
4990 [Ben Laurie (CHATS)]
4992 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4994 [Ben Laurie (CHATS)]
4996 *) Eliminate unused copy of key in RC4.
4997 [Ben Laurie (CHATS)]
4999 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5000 [Ben Laurie (CHATS)]
5002 *) Fix off-by-one error in EGD path.
5003 [Ben Laurie (CHATS)]
5005 *) If RANDFILE path is too long, ignore instead of truncating.
5006 [Ben Laurie (CHATS)]
5008 *) Eliminate unused and incorrectly sized X.509 structure
5010 [Ben Laurie (CHATS)]
5012 *) Eliminate unused and dangerous function knumber().
5013 [Ben Laurie (CHATS)]
5015 *) Eliminate unused and dangerous structure, KSSL_ERR.
5016 [Ben Laurie (CHATS)]
5018 *) Protect against overlong session ID context length in an encoded
5019 session object. Since these are local, this does not appear to be
5021 [Ben Laurie (CHATS)]
5023 *) Change from security patch (see 0.9.6e below) that did not affect
5024 the 0.9.6 release series:
5026 Remote buffer overflow in SSL3 protocol - an attacker could
5027 supply an oversized master key in Kerberos-enabled versions.
5029 [Ben Laurie (CHATS)]
5031 *) Change the SSL kerb5 codes to match RFC 2712.
5034 *) Make -nameopt work fully for req and add -reqopt switch.
5035 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5037 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5038 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5040 *) Make sure tests can be performed even if the corresponding algorithms
5041 have been removed entirely. This was also the last step to make
5042 OpenSSL compilable with DJGPP under all reasonable conditions.
5043 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5045 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5046 to allow version independent disabling of normally unselected ciphers,
5047 which may be activated as a side-effect of selecting a single cipher.
5049 (E.g., cipher list string "RSA" enables ciphersuites that are left
5050 out of "ALL" because they do not provide symmetric encryption.
5051 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5052 [Lutz Jaenicke, Bodo Moeller]
5054 *) Add appropriate support for separate platform-dependent build
5055 directories. The recommended way to make a platform-dependent
5056 build directory is the following (tested on Linux), maybe with
5059 # Place yourself outside of the OpenSSL source tree. In
5060 # this example, the environment variable OPENSSL_SOURCE
5061 # is assumed to contain the absolute OpenSSL source directory.
5062 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5063 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5064 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5065 mkdir -p `dirname $F`
5066 ln -s $OPENSSL_SOURCE/$F $F
5069 To be absolutely sure not to disturb the source tree, a "make clean"
5070 is a good thing. If it isn't successfull, don't worry about it,
5071 it probably means the source directory is very clean.
5074 *) Make sure any ENGINE control commands make local copies of string
5075 pointers passed to them whenever necessary. Otherwise it is possible
5076 the caller may have overwritten (or deallocated) the original string
5077 data when a later ENGINE operation tries to use the stored values.
5078 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5080 *) Improve diagnostics in file reading and command-line digests.
5081 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5083 *) Add AES modes CFB and OFB to the object database. Correct an
5084 error in AES-CFB decryption.
5087 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5088 allows existing EVP_CIPHER_CTX structures to be reused after
5089 calling EVP_*Final(). This behaviour is used by encryption
5090 BIOs and some applications. This has the side effect that
5091 applications must explicitly clean up cipher contexts with
5092 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5095 *) Check the values of dna and dnb in bn_mul_recursive before calling
5096 bn_mul_comba (a non zero value means the a or b arrays do not contain
5097 n2 elements) and fallback to bn_mul_normal if either is not zero.
5100 *) Fix escaping of non-ASCII characters when using the -subj option
5101 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5104 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5105 form for "surname", serialNumber has no short form.
5106 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5107 therefore remove "mail" short name for "internet 7".
5108 The OID for unique identifiers in X509 certificates is
5109 x500UniqueIdentifier, not uniqueIdentifier.
5110 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5113 *) Add an "init" command to the ENGINE config module and auto initialize
5114 ENGINEs. Without any "init" command the ENGINE will be initialized
5115 after all ctrl commands have been executed on it. If init=1 the
5116 ENGINE is initailized at that point (ctrls before that point are run
5117 on the uninitialized ENGINE and after on the initialized one). If
5118 init=0 then the ENGINE will not be iniatialized at all.
5121 *) Fix the 'app_verify_callback' interface so that the user-defined
5122 argument is actually passed to the callback: In the
5123 SSL_CTX_set_cert_verify_callback() prototype, the callback
5124 declaration has been changed from
5127 int (*cb)(X509_STORE_CTX *,void *);
5128 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5129 i=s->ctx->app_verify_callback(&ctx)
5130 has been changed into
5131 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5133 To update applications using SSL_CTX_set_cert_verify_callback(),
5134 a dummy argument can be added to their callback functions.
5135 [D. K. Smetters <smetters@parc.xerox.com>]
5137 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5138 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5140 *) Add and OPENSSL_LOAD_CONF define which will cause
5141 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5142 This allows older applications to transparently support certain
5143 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5144 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5145 load the config file and OPENSSL_add_all_algorithms_conf() which will
5146 always load it have also been added.
5149 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5150 Adjust NIDs and EVP layer.
5151 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5153 *) Config modules support in openssl utility.
5155 Most commands now load modules from the config file,
5156 though in a few (such as version) this isn't done
5157 because it couldn't be used for anything.
5159 In the case of ca and req the config file used is
5160 the same as the utility itself: that is the -config
5161 command line option can be used to specify an
5165 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5166 use "openssl_conf" if filename is NULL use default openssl config file.
5169 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5170 config section name. Add a new flag to tolerate a missing config file
5171 and move code to CONF_modules_load_file().
5174 *) Support for crypto accelerator cards from Accelerated Encryption
5175 Processing, www.aep.ie. (Use engine 'aep')
5176 The support was copied from 0.9.6c [engine] and adapted/corrected
5177 to work with the new engine framework.
5178 [AEP Inc. and Richard Levitte]
5180 *) Support for SureWare crypto accelerator cards from Baltimore
5181 Technologies. (Use engine 'sureware')
5182 The support was copied from 0.9.6c [engine] and adapted
5183 to work with the new engine framework.
5186 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5187 make the newer ENGINE framework commands for the CHIL engine work.
5188 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5190 *) Make it possible to produce shared libraries on ReliantUNIX.
5191 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5193 *) Add the configuration target debug-linux-ppro.
5194 Make 'openssl rsa' use the general key loading routines
5195 implemented in apps.c, and make those routines able to
5196 handle the key format FORMAT_NETSCAPE and the variant
5198 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5200 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5201 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5203 *) Add -keyform to rsautl, and document -engine.
5204 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5206 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5207 BIO_R_NO_SUCH_FILE error code rather than the generic
5208 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5211 *) Add new functions
5213 ERR_peek_last_error_line
5214 ERR_peek_last_error_line_data.
5215 These are similar to
5218 ERR_peek_error_line_data,
5219 but report on the latest error recorded rather than the first one
5220 still in the error queue.
5221 [Ben Laurie, Bodo Moeller]
5223 *) default_algorithms option in ENGINE config module. This allows things
5225 default_algorithms = ALL
5226 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5229 *) Prelminary ENGINE config module.
5232 *) New experimental application configuration code.
5235 *) Change the AES code to follow the same name structure as all other
5236 symmetric ciphers, and behave the same way. Move everything to
5237 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5238 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5240 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5241 [Ben Laurie and Theo de Raadt]
5243 *) Add option to output public keys in req command.
5244 [Massimiliano Pala madwolf@openca.org]
5246 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5247 (up to about 10% better than before for P-192 and P-224).
5250 *) New functions/macros
5252 SSL_CTX_set_msg_callback(ctx, cb)
5253 SSL_CTX_set_msg_callback_arg(ctx, arg)
5254 SSL_set_msg_callback(ssl, cb)
5255 SSL_set_msg_callback_arg(ssl, arg)
5257 to request calling a callback function
5259 void cb(int write_p, int version, int content_type,
5260 const void *buf, size_t len, SSL *ssl, void *arg)
5262 whenever a protocol message has been completely received
5263 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5264 protocol version according to which the SSL library interprets
5265 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5266 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5267 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5268 specification (change_cipher_spec(20), alert(21), handshake(22)).
5269 'buf' and 'len' point to the actual message, 'ssl' to the
5270 SSL object, and 'arg' is the application-defined value set by
5271 SSL[_CTX]_set_msg_callback_arg().
5273 'openssl s_client' and 'openssl s_server' have new '-msg' options
5274 to enable a callback that displays all protocol messages.
5277 *) Change the shared library support so shared libraries are built as
5278 soon as the corresponding static library is finished, and thereby get
5279 openssl and the test programs linked against the shared library.
5280 This still only happens when the keyword "shard" has been given to
5281 the configuration scripts.
5283 NOTE: shared library support is still an experimental thing, and
5284 backward binary compatibility is still not guaranteed.
5285 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5287 *) Add support for Subject Information Access extension.
5288 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5290 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5291 additional bytes when new memory had to be allocated, not just
5292 when reusing an existing buffer.
5295 *) New command line and configuration option 'utf8' for the req command.
5296 This allows field values to be specified as UTF8 strings.
5299 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5300 runs for the former and machine-readable output for the latter.
5303 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5304 of the e-mail address in the DN (i.e., it will go into a certificate
5305 extension only). The new configuration file option 'email_in_dn = no'
5306 has the same effect.
5307 [Massimiliano Pala madwolf@openca.org]
5309 *) Change all functions with names starting with des_ to be starting
5310 with DES_ instead. Add wrappers that are compatible with libdes,
5311 but are named _ossl_old_des_*. Finally, add macros that map the
5312 des_* symbols to the corresponding _ossl_old_des_* if libdes
5313 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5314 desired, the des_* symbols will be mapped to DES_*, with one
5317 Since we provide two compatibility mappings, the user needs to
5318 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5319 compatibility is desired. The default (i.e., when that macro
5320 isn't defined) is OpenSSL 0.9.6c compatibility.
5322 There are also macros that enable and disable the support of old
5323 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5324 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5325 are defined, the default will apply: to support the old des routines.
5327 In either case, one must include openssl/des.h to get the correct
5328 definitions. Do not try to just include openssl/des_old.h, that
5331 NOTE: This is a major break of an old API into a new one. Software
5332 authors are encouraged to switch to the DES_ style functions. Some
5333 time in the future, des_old.h and the libdes compatibility functions
5334 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5335 default), and then completely removed.
5338 *) Test for certificates which contain unsupported critical extensions.
5339 If such a certificate is found during a verify operation it is
5340 rejected by default: this behaviour can be overridden by either
5341 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5342 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5343 X509_supported_extension() has also been added which returns 1 if a
5344 particular extension is supported.
5347 *) Modify the behaviour of EVP cipher functions in similar way to digests
5348 to retain compatibility with existing code.
5351 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5352 compatibility with existing code. In particular the 'ctx' parameter does
5353 not have to be to be initialized before the call to EVP_DigestInit() and
5354 it is tidied up after a call to EVP_DigestFinal(). New function
5355 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5356 EVP_MD_CTX_copy() changed to not require the destination to be
5357 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5358 requires the destination to be valid.
5360 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5361 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5364 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5365 so that complete 'Handshake' protocol structures are kept in memory
5366 instead of overwriting 'msg_type' and 'length' with 'body' data.
5369 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5370 [Massimo Santin via Richard Levitte]
5372 *) Major restructuring to the underlying ENGINE code. This includes
5373 reduction of linker bloat, separation of pure "ENGINE" manipulation
5374 (initialisation, etc) from functionality dealing with implementations
5375 of specific crypto iterfaces. This change also introduces integrated
5376 support for symmetric ciphers and digest implementations - so ENGINEs
5377 can now accelerate these by providing EVP_CIPHER and EVP_MD
5378 implementations of their own. This is detailed in crypto/engine/README
5379 as it couldn't be adequately described here. However, there are a few
5380 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5381 were changed in the original introduction of ENGINE code have now
5382 reverted back - the hooking from this code to ENGINE is now a good
5383 deal more passive and at run-time, operations deal directly with
5384 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5385 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5386 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5387 they were not being used by the framework as there is no concept of a
5388 BIGNUM_METHOD and they could not be generalised to the new
5389 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5390 ENGINE_cpy() has been removed as it cannot be consistently defined in
5394 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5397 *) Change mkdef.pl to sort symbols that get the same entry number,
5398 and make sure the automatically generated functions ERR_load_*
5399 become part of libeay.num as well.
5402 *) New function SSL_renegotiate_pending(). This returns true once
5403 renegotiation has been requested (either SSL_renegotiate() call
5404 or HelloRequest/ClientHello receveived from the peer) and becomes
5405 false once a handshake has been completed.
5406 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5407 sends a HelloRequest, but does not ensure that a handshake takes
5408 place. SSL_renegotiate_pending() is useful for checking if the
5409 client has followed the request.)
5412 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5413 By default, clients may request session resumption even during
5414 renegotiation (if session ID contexts permit); with this option,
5415 session resumption is possible only in the first handshake.
5417 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5418 more bits available for options that should not be part of
5419 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5422 *) Add some demos for certificate and certificate request creation.
5425 *) Make maximum certificate chain size accepted from the peer application
5426 settable (SSL*_get/set_max_cert_list()), as proposed by
5427 "Douglas E. Engert" <deengert@anl.gov>.
5430 *) Add support for shared libraries for Unixware-7
5431 (Boyd Lynn Gerber <gerberb@zenez.com>).
5434 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5435 be done prior to destruction. Use this to unload error strings from
5436 ENGINEs that load their own error strings. NB: This adds two new API
5437 functions to "get" and "set" this destroy handler in an ENGINE.
5440 *) Alter all existing ENGINE implementations (except "openssl" and
5441 "openbsd") to dynamically instantiate their own error strings. This
5442 makes them more flexible to be built both as statically-linked ENGINEs
5443 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5444 Also, add stub code to each that makes building them as self-contained
5445 shared-libraries easier (see README.ENGINE).
5448 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5449 implementations into applications that are completely implemented in
5450 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5451 commands that can be used to configure what shared-library to load and
5452 to control aspects of the way it is handled. Also, made an update to
5453 the README.ENGINE file that brings its information up-to-date and
5454 provides some information and instructions on the "dynamic" ENGINE
5455 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5458 *) Make it possible to unload ranges of ERR strings with a new
5459 "ERR_unload_strings" function.
5462 *) Add a copy() function to EVP_MD.
5465 *) Make EVP_MD routines take a context pointer instead of just the
5466 md_data void pointer.
5469 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5470 that the digest can only process a single chunk of data
5471 (typically because it is provided by a piece of
5472 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5473 is only going to provide a single chunk of data, and hence the
5474 framework needn't accumulate the data for oneshot drivers.
5477 *) As with "ERR", make it possible to replace the underlying "ex_data"
5478 functions. This change also alters the storage and management of global
5479 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5480 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5481 index counters. The API functions that use this state have been changed
5482 to take a "class_index" rather than pointers to the class's local STACK
5483 and counter, and there is now an API function to dynamically create new
5484 classes. This centralisation allows us to (a) plug a lot of the
5485 thread-safety problems that existed, and (b) makes it possible to clean
5486 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5487 such data would previously have always leaked in application code and
5488 workarounds were in place to make the memory debugging turn a blind eye
5489 to it. Application code that doesn't use this new function will still
5490 leak as before, but their memory debugging output will announce it now
5491 rather than letting it slide.
5493 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5494 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5495 has a return value to indicate success or failure.
5498 *) Make it possible to replace the underlying "ERR" functions such that the
5499 global state (2 LHASH tables and 2 locks) is only used by the "default"
5500 implementation. This change also adds two functions to "get" and "set"
5501 the implementation prior to it being automatically set the first time
5502 any other ERR function takes place. Ie. an application can call "get",
5503 pass the return value to a module it has just loaded, and that module
5504 can call its own "set" function using that value. This means the
5505 module's "ERR" operations will use (and modify) the error state in the
5506 application and not in its own statically linked copy of OpenSSL code.
5509 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5510 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5511 the operation, and provides a more encapsulated way for external code
5512 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5513 to use these functions rather than manually incrementing the counts.
5515 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5518 *) Add EVP test program.
5521 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5524 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5525 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5526 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5527 These allow a CRL to be built without having to access X509_CRL fields
5528 directly. Modify 'ca' application to use new functions.
5531 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5532 bug workarounds. Rollback attack detection is a security feature.
5533 The problem will only arise on OpenSSL servers when TLSv1 is not
5534 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5535 Software authors not wanting to support TLSv1 will have special reasons
5536 for their choice and can explicitly enable this option.
5537 [Bodo Moeller, Lutz Jaenicke]
5539 *) Rationalise EVP so it can be extended: don't include a union of
5540 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5541 (similar to those existing for EVP_CIPHER_CTX).
5546 EVP_MD_CTX_init(&md); /* new function call */
5547 EVP_DigestInit(&md, EVP_sha1());
5548 EVP_DigestUpdate(&md, in, len);
5549 EVP_DigestFinal(&md, out, NULL);
5550 EVP_MD_CTX_cleanup(&md); /* new function call */
5554 *) Make DES key schedule conform to the usual scheme, as well as
5555 correcting its structure. This means that calls to DES functions
5556 now have to pass a pointer to a des_key_schedule instead of a
5557 plain des_key_schedule (which was actually always a pointer
5560 des_key_schedule ks;
5562 des_set_key_checked(..., &ks);
5563 des_ncbc_encrypt(..., &ks, ...);
5565 (Note that a later change renames 'des_...' into 'DES_...'.)
5568 *) Initial reduction of linker bloat: the use of some functions, such as
5569 PEM causes large amounts of unused functions to be linked in due to
5570 poor organisation. For example pem_all.c contains every PEM function
5571 which has a knock on effect of linking in large amounts of (unused)
5572 ASN1 code. Grouping together similar functions and splitting unrelated
5573 functions prevents this.
5576 *) Cleanup of EVP macros.
5579 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5580 correct _ecb suffix.
5583 *) Add initial OCSP responder support to ocsp application. The
5584 revocation information is handled using the text based index
5585 use by the ca application. The responder can either handle
5586 requests generated internally, supplied in files (for example
5587 via a CGI script) or using an internal minimal server.
5590 *) Add configuration choices to get zlib compression for TLS.
5593 *) Changes to Kerberos SSL for RFC 2712 compliance:
5594 1. Implemented real KerberosWrapper, instead of just using
5595 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5596 2. Implemented optional authenticator field of KerberosWrapper.
5598 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5599 and authenticator structs; see crypto/krb5/.
5601 Generalized Kerberos calls to support multiple Kerberos libraries.
5602 [Vern Staats <staatsvr@asc.hpc.mil>,
5603 Jeffrey Altman <jaltman@columbia.edu>
5604 via Richard Levitte]
5606 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5607 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5608 values for each of the key sizes rather than having just
5609 parameters (and 'speed' generating keys each time).
5612 *) Speed up EVP routines.
5615 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5616 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5617 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5618 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5620 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5621 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5622 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5625 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5627 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5630 *) Added the OS2-EMX target.
5631 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5633 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5634 to support NCONF routines in extension code. New function CONF_set_nconf()
5635 to allow functions which take an NCONF to also handle the old LHASH
5636 structure: this means that the old CONF compatible routines can be
5637 retained (in particular wrt extensions) without having to duplicate the
5638 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5641 *) Enhance the general user interface with mechanisms for inner control
5642 and with possibilities to have yes/no kind of prompts.
5645 *) Change all calls to low level digest routines in the library and
5646 applications to use EVP. Add missing calls to HMAC_cleanup() and
5647 don't assume HMAC_CTX can be copied using memcpy().
5648 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5650 *) Add the possibility to control engines through control names but with
5651 arbitrary arguments instead of just a string.
5652 Change the key loaders to take a UI_METHOD instead of a callback
5653 function pointer. NOTE: this breaks binary compatibility with earlier
5654 versions of OpenSSL [engine].
5655 Adapt the nCipher code for these new conditions and add a card insertion
5659 *) Enhance the general user interface with mechanisms to better support
5660 dialog box interfaces, application-defined prompts, the possibility
5661 to use defaults (for example default passwords from somewhere else)
5662 and interrupts/cancellations.
5665 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5666 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5669 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5670 tidy up some unnecessarily weird code in 'sk_new()').
5671 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5673 *) Change the key loading routines for ENGINEs to use the same kind
5674 callback (pem_password_cb) as all other routines that need this
5678 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5679 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5680 than this minimum value is recommended.
5683 *) New random seeder for OpenVMS, using the system process statistics
5684 that are easily reachable.
5687 *) Windows apparently can't transparently handle global
5688 variables defined in DLLs. Initialisations such as:
5690 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5692 wont compile. This is used by the any applications that need to
5693 declare their own ASN1 modules. This was fixed by adding the option
5694 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5695 needed for static libraries under Win32.
5698 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5699 setting of purpose and trust fields. New X509_STORE trust and
5700 purpose functions and tidy up setting in other SSL functions.
5703 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5704 structure. These are inherited by X509_STORE_CTX when it is
5705 initialised. This allows various defaults to be set in the
5706 X509_STORE structure (such as flags for CRL checking and custom
5707 purpose or trust settings) for functions which only use X509_STORE_CTX
5708 internally such as S/MIME.
5710 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5711 trust settings if they are not set in X509_STORE. This allows X509_STORE
5712 purposes and trust (in S/MIME for example) to override any set by default.
5714 Add command line options for CRL checking to smime, s_client and s_server
5718 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5719 are set then the CRL is looked up in the X509_STORE structure and
5720 its validity and signature checked, then if the certificate is found
5721 in the CRL the verify fails with a revoked error.
5723 Various new CRL related callbacks added to X509_STORE_CTX structure.
5725 Command line options added to 'verify' application to support this.
5727 This needs some additional work, such as being able to handle multiple
5728 CRLs with different times, extension based lookup (rather than just
5729 by subject name) and ultimately more complete V2 CRL extension
5733 *) Add a general user interface API (crypto/ui/). This is designed
5734 to replace things like des_read_password and friends (backward
5735 compatibility functions using this new API are provided).
5736 The purpose is to remove prompting functions from the DES code
5737 section as well as provide for prompting through dialog boxes in
5738 a window system and the like.
5741 *) Add "ex_data" support to ENGINE so implementations can add state at a
5742 per-structure level rather than having to store it globally.
5745 *) Make it possible for ENGINE structures to be copied when retrieved by
5746 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5747 This causes the "original" ENGINE structure to act like a template,
5748 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5749 operational state can be localised to each ENGINE structure, despite the
5750 fact they all share the same "methods". New ENGINE structures returned in
5751 this case have no functional references and the return value is the single
5752 structural reference. This matches the single structural reference returned
5753 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5757 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5758 needs to match any other type at all we need to manually clear the
5762 *) Changes to the "openssl engine" utility to include;
5763 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5764 about an ENGINE's available control commands.
5765 - executing control commands from command line arguments using the
5766 '-pre' and '-post' switches. '-post' is only used if '-t' is
5767 specified and the ENGINE is successfully initialised. The syntax for
5768 the individual commands are colon-separated, for example;
5769 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5772 *) New dynamic control command support for ENGINEs. ENGINEs can now
5773 declare their own commands (numbers), names (strings), descriptions,
5774 and input types for run-time discovery by calling applications. A
5775 subset of these commands are implicitly classed as "executable"
5776 depending on their input type, and only these can be invoked through
5777 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5778 can be based on user input, config files, etc). The distinction is
5779 that "executable" commands cannot return anything other than a boolean
5780 result and can only support numeric or string input, whereas some
5781 discoverable commands may only be for direct use through
5782 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5783 pointers, or other custom uses. The "executable" commands are to
5784 support parameterisations of ENGINE behaviour that can be
5785 unambiguously defined by ENGINEs and used consistently across any
5786 OpenSSL-based application. Commands have been added to all the
5787 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5788 control over shared-library paths without source code alterations.
5791 *) Changed all ENGINE implementations to dynamically allocate their
5792 ENGINEs rather than declaring them statically. Apart from this being
5793 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5794 this also allows the implementations to compile without using the
5795 internal engine_int.h header.
5798 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5799 'const' value. Any code that should be able to modify a RAND_METHOD
5800 should already have non-const pointers to it (ie. they should only
5801 modify their own ones).
5804 *) Made a variety of little tweaks to the ENGINE code.
5805 - "atalla" and "ubsec" string definitions were moved from header files
5806 to C code. "nuron" string definitions were placed in variables
5807 rather than hard-coded - allowing parameterisation of these values
5808 later on via ctrl() commands.
5809 - Removed unused "#if 0"'d code.
5810 - Fixed engine list iteration code so it uses ENGINE_free() to release
5811 structural references.
5812 - Constified the RAND_METHOD element of ENGINE structures.
5813 - Constified various get/set functions as appropriate and added
5814 missing functions (including a catch-all ENGINE_cpy that duplicates
5815 all ENGINE values onto a new ENGINE except reference counts/state).
5816 - Removed NULL parameter checks in get/set functions. Setting a method
5817 or function to NULL is a way of cancelling out a previously set
5818 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5819 and doesn't justify the extra error symbols and code.
5820 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5821 flags from engine_int.h to engine.h.
5822 - Changed prototypes for ENGINE handler functions (init(), finish(),
5823 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5826 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5827 to the algorithm using long division. The binary algorithm can be
5828 used only if the modulus is odd. On 32-bit systems, it is faster
5829 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5830 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5831 up to 450 bits. In 64-bit environments, the binary algorithm
5832 appears to be advantageous for much longer moduli; here we use it
5833 for moduli up to 2048 bits.
5836 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5837 could not support the combine flag in choice fields.
5840 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5841 extensions from a certificate request to the certificate.
5844 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5845 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5846 file: this allows the display of the certificate about to be
5847 signed to be customised, to allow certain fields to be included
5848 or excluded and extension details. The old system didn't display
5849 multicharacter strings properly, omitted fields not in the policy
5850 and couldn't display additional details such as extensions.
5853 *) Function EC_POINTs_mul for multiple scalar multiplication
5854 of an arbitrary number of elliptic curve points
5855 \sum scalars[i]*points[i],
5856 optionally including the generator defined for the EC_GROUP:
5857 scalar*generator + \sum scalars[i]*points[i].
5859 EC_POINT_mul is a simple wrapper function for the typical case
5860 that the point list has just one item (besides the optional
5864 *) First EC_METHODs for curves over GF(p):
5866 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5867 operations and provides various method functions that can also
5868 operate with faster implementations of modular arithmetic.
5870 EC_GFp_mont_method() reuses most functions that are part of
5871 EC_GFp_simple_method, but uses Montgomery arithmetic.
5873 [Bodo Moeller; point addition and point doubling
5874 implementation directly derived from source code provided by
5875 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5877 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5878 crypto/ec/ec_lib.c):
5880 Curves are EC_GROUP objects (with an optional group generator)
5881 based on EC_METHODs that are built into the library.
5883 Points are EC_POINT objects based on EC_GROUP objects.
5885 Most of the framework would be able to handle curves over arbitrary
5886 finite fields, but as there are no obvious types for fields other
5887 than GF(p), some functions are limited to that for now.
5890 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5891 that the file contains a complete HTTP response.
5894 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5895 change the def and num file printf format specifier from "%-40sXXX"
5896 to "%-39s XXX". The latter will always guarantee a space after the
5897 field while the former will cause them to run together if the field
5898 is 40 of more characters long.
5901 *) Constify the cipher and digest 'method' functions and structures
5902 and modify related functions to take constant EVP_MD and EVP_CIPHER
5906 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5907 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5910 *) Modify EVP_Digest*() routines so they now return values. Although the
5911 internal software routines can never fail additional hardware versions
5915 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5917 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5918 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5921 ERR_R_NESTED_ASN1_ERROR
5923 ERR_R_MISSING_ASN1_EOS
5924 were 4 .. 9, conflicting with
5925 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5927 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5928 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5930 Add new error code 'ERR_R_INTERNAL_ERROR'.
5933 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5937 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5938 sets the subject name for a new request or supersedes the
5939 subject name in a given request. Formats that can be parsed are
5940 'CN=Some Name, OU=myOU, C=IT'
5942 'CN=Some Name/OU=myOU/C=IT'.
5944 Add options '-batch' and '-verbose' to 'openssl req'.
5945 [Massimiliano Pala <madwolf@hackmasters.net>]
5947 *) Introduce the possibility to access global variables through
5948 functions on platform were that's the best way to handle exporting
5949 global variables in shared libraries. To enable this functionality,
5950 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5951 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5952 is normally done by Configure or something similar).
5954 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5955 in the source file (foo.c) like this:
5957 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5958 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5960 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5961 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5963 OPENSSL_DECLARE_GLOBAL(int,foo);
5964 #define foo OPENSSL_GLOBAL_REF(foo)
5965 OPENSSL_DECLARE_GLOBAL(double,bar);
5966 #define bar OPENSSL_GLOBAL_REF(bar)
5968 The #defines are very important, and therefore so is including the
5969 header file everywhere where the defined globals are used.
5971 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5972 of ASN.1 items, but that structure is a bit different.
5974 The largest change is in util/mkdef.pl which has been enhanced with
5975 better and easier to understand logic to choose which symbols should
5976 go into the Windows .def files as well as a number of fixes and code
5977 cleanup (among others, algorithm keywords are now sorted
5978 lexicographically to avoid constant rewrites).
5981 *) In BN_div() keep a copy of the sign of 'num' before writing the
5982 result to 'rm' because if rm==num the value will be overwritten
5983 and produce the wrong result if 'num' is negative: this caused
5984 problems with BN_mod() and BN_nnmod().
5987 *) Function OCSP_request_verify(). This checks the signature on an
5988 OCSP request and verifies the signer certificate. The signer
5989 certificate is just checked for a generic purpose and OCSP request
5993 *) Add OCSP_check_validity() function to check the validity of OCSP
5994 responses. OCSP responses are prepared in real time and may only
5995 be a few seconds old. Simply checking that the current time lies
5996 between thisUpdate and nextUpdate max reject otherwise valid responses
5997 caused by either OCSP responder or client clock inaccuracy. Instead
5998 we allow thisUpdate and nextUpdate to fall within a certain period of
5999 the current time. The age of the response can also optionally be
6000 checked. Two new options -validity_period and -status_age added to
6004 *) If signature or public key algorithm is unrecognized print out its
6005 OID rather that just UNKNOWN.
6008 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6009 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6010 ID to be generated from the issuer certificate alone which can then be
6011 passed to OCSP_id_issuer_cmp().
6014 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6015 ASN1 modules to export functions returning ASN1_ITEM pointers
6016 instead of the ASN1_ITEM structures themselves. This adds several
6017 new macros which allow the underlying ASN1 function/structure to
6018 be accessed transparently. As a result code should not use ASN1_ITEM
6019 references directly (such as &X509_it) but instead use the relevant
6020 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6021 use of the new ASN1 code on platforms where exporting structures
6022 is problematical (for example in shared libraries) but exporting
6023 functions returning pointers to structures is not.
6026 *) Add support for overriding the generation of SSL/TLS session IDs.
6027 These callbacks can be registered either in an SSL_CTX or per SSL.
6028 The purpose of this is to allow applications to control, if they wish,
6029 the arbitrary values chosen for use as session IDs, particularly as it
6030 can be useful for session caching in multiple-server environments. A
6031 command-line switch for testing this (and any client code that wishes
6032 to use such a feature) has been added to "s_server".
6033 [Geoff Thorpe, Lutz Jaenicke]
6035 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6036 of the form '#if defined(...) || defined(...) || ...' and
6037 '#if !defined(...) && !defined(...) && ...'. This also avoids
6038 the growing number of special cases it was previously handling.
6041 *) Make all configuration macros available for application by making
6042 sure they are available in opensslconf.h, by giving them names starting
6043 with "OPENSSL_" to avoid conflicts with other packages and by making
6044 sure e_os2.h will cover all platform-specific cases together with
6046 Additionally, it is now possible to define configuration/platform-
6047 specific names (called "system identities"). In the C code, these
6048 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6049 macro with the name beginning with "OPENSSL_SYS_", which is determined
6050 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6054 *) New option -set_serial to 'req' and 'x509' this allows the serial
6055 number to use to be specified on the command line. Previously self
6056 signed certificates were hard coded with serial number 0 and the
6057 CA options of 'x509' had to use a serial number in a file which was
6061 *) New options to 'ca' utility to support V2 CRL entry extensions.
6062 Currently CRL reason, invalidity date and hold instruction are
6063 supported. Add new CRL extensions to V3 code and some new objects.
6066 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6067 disable standard block padding (aka PKCS#5 padding) in the EVP
6068 API, which was previously mandatory. This means that the data is
6069 not padded in any way and so the total length much be a multiple
6070 of the block size, otherwise an error occurs.
6073 *) Initial (incomplete) OCSP SSL support.
6076 *) New function OCSP_parse_url(). This splits up a URL into its host,
6077 port and path components: primarily to parse OCSP URLs. New -url
6078 option to ocsp utility.
6081 *) New nonce behavior. The return value of OCSP_check_nonce() now
6082 reflects the various checks performed. Applications can decide
6083 whether to tolerate certain situations such as an absent nonce
6084 in a response when one was present in a request: the ocsp application
6085 just prints out a warning. New function OCSP_add1_basic_nonce()
6086 this is to allow responders to include a nonce in a response even if
6087 the request is nonce-less.
6090 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6091 skipped when using openssl x509 multiple times on a single input file,
6092 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6095 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6096 set string type: to handle setting ASN1_TIME structures. Fix ca
6097 utility to correctly initialize revocation date of CRLs.
6100 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6101 the clients preferred ciphersuites and rather use its own preferences.
6102 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6103 Internet Explorer by ensuring unchanged hash method during stepup.
6104 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6107 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6108 to aes and add a new 'exist' option to print out symbols that don't
6112 *) Additional options to ocsp utility to allow flags to be set and
6113 additional certificates supplied.
6116 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6117 OCSP client a number of certificate to only verify the response
6121 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6122 handle the new API. Currently only ECB, CBC modes supported. Add new
6125 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6126 Encryption Standard (AES) Ciphersuites for Transport Layer
6127 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6128 not enabled by default and were not part of the "ALL" ciphersuite
6129 alias because they were not yet official; they could be
6130 explicitly requested by specifying the "AESdraft" ciphersuite
6131 group alias. In the final release of OpenSSL 0.9.7, the group
6132 alias is called "AES" and is part of "ALL".)
6133 [Ben Laurie, Steve Henson, Bodo Moeller]
6135 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6136 request to response.
6139 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6140 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6141 extract information from a certificate request. OCSP_response_create()
6142 creates a response and optionally adds a basic response structure.
6143 OCSP_basic_add1_status() adds a complete single response to a basic
6144 response and returns the OCSP_SINGLERESP structure just added (to allow
6145 extensions to be included for example). OCSP_basic_add1_cert() adds a
6146 certificate to a basic response and OCSP_basic_sign() signs a basic
6147 response with various flags. New helper functions ASN1_TIME_check()
6148 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6149 (converts ASN1_TIME to GeneralizedTime).
6152 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6153 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6154 structure from a certificate. X509_pubkey_digest() digests the public_key
6155 contents: this is used in various key identifiers.
6158 *) Make sk_sort() tolerate a NULL argument.
6159 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6161 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6162 passed by the function are trusted implicitly. If any of them signed the
6163 response then it is assumed to be valid and is not verified.
6166 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6167 to data. This was previously part of the PKCS7 ASN1 code. This
6168 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6169 [Steve Henson, reported by Kenneth R. Robinette
6170 <support@securenetterm.com>]
6172 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6173 routines: without these tracing memory leaks is very painful.
6174 Fix leaks in PKCS12 and PKCS7 routines.
6177 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6178 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6179 effectively meant GeneralizedTime would never be used. Now it
6180 is initialised to -1 but X509_time_adj() now has to check the value
6181 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6182 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6183 [Steve Henson, reported by Kenneth R. Robinette
6184 <support@securenetterm.com>]
6186 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6187 result in a zero length in the ASN1_INTEGER structure which was
6188 not consistent with the structure when d2i_ASN1_INTEGER() was used
6189 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6190 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6191 where it did not print out a minus for negative ASN1_INTEGER.
6194 *) Add summary printout to ocsp utility. The various functions which
6195 convert status values to strings have been renamed to:
6196 OCSP_response_status_str(), OCSP_cert_status_str() and
6197 OCSP_crl_reason_str() and are no longer static. New options
6198 to verify nonce values and to disable verification. OCSP response
6199 printout format cleaned up.
6202 *) Add additional OCSP certificate checks. These are those specified
6203 in RFC2560. This consists of two separate checks: the CA of the
6204 certificate being checked must either be the OCSP signer certificate
6205 or the issuer of the OCSP signer certificate. In the latter case the
6206 OCSP signer certificate must contain the OCSP signing extended key
6207 usage. This check is performed by attempting to match the OCSP
6208 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6209 in the OCSP_CERTID structures of the response.
6212 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6213 and related routines. This uses the standard OpenSSL certificate
6214 verify routines to perform initial checks (just CA validity) and
6215 to obtain the certificate chain. Then additional checks will be
6216 performed on the chain. Currently the root CA is checked to see
6217 if it is explicitly trusted for OCSP signing. This is used to set
6218 a root CA as a global signing root: that is any certificate that
6219 chains to that CA is an acceptable OCSP signing certificate.
6222 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6223 extensions from a separate configuration file.
6224 As when reading extensions from the main configuration file,
6225 the '-extensions ...' option may be used for specifying the
6227 [Massimiliano Pala <madwolf@comune.modena.it>]
6229 *) New OCSP utility. Allows OCSP requests to be generated or
6230 read. The request can be sent to a responder and the output
6231 parsed, outputed or printed in text form. Not complete yet:
6232 still needs to check the OCSP response validity.
6235 *) New subcommands for 'openssl ca':
6236 'openssl ca -status <serial>' prints the status of the cert with
6237 the given serial number (according to the index file).
6238 'openssl ca -updatedb' updates the expiry status of certificates
6240 [Massimiliano Pala <madwolf@comune.modena.it>]
6242 *) New '-newreq-nodes' command option to CA.pl. This is like
6243 '-newreq', but calls 'openssl req' with the '-nodes' option
6244 so that the resulting key is not encrypted.
6245 [Damien Miller <djm@mindrot.org>]
6247 *) New configuration for the GNU Hurd.
6248 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6250 *) Initial code to implement OCSP basic response verify. This
6251 is currently incomplete. Currently just finds the signer's
6252 certificate and verifies the signature on the response.
6255 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6256 value of OPENSSLDIR. This is available via the new '-d' option
6257 to 'openssl version', and is also included in 'openssl version -a'.
6260 *) Allowing defining memory allocation callbacks that will be given
6261 file name and line number information in additional arguments
6262 (a const char* and an int). The basic functionality remains, as
6263 well as the original possibility to just replace malloc(),
6264 realloc() and free() by functions that do not know about these
6265 additional arguments. To register and find out the current
6266 settings for extended allocation functions, the following
6267 functions are provided:
6269 CRYPTO_set_mem_ex_functions
6270 CRYPTO_set_locked_mem_ex_functions
6271 CRYPTO_get_mem_ex_functions
6272 CRYPTO_get_locked_mem_ex_functions
6274 These work the same way as CRYPTO_set_mem_functions and friends.
6275 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6276 extended allocation function is enabled.
6277 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6278 a conventional allocation function is enabled.
6279 [Richard Levitte, Bodo Moeller]
6281 *) Finish off removing the remaining LHASH function pointer casts.
6282 There should no longer be any prototype-casting required when using
6283 the LHASH abstraction, and any casts that remain are "bugs". See
6284 the callback types and macros at the head of lhash.h for details
6285 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6288 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6289 If /dev/[u]random devices are not available or do not return enough
6290 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6292 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6293 /etc/entropy will be queried once each in this sequence, quering stops
6294 when enough entropy was collected without querying more sockets.
6297 *) Change the Unix RAND_poll() variant to be able to poll several
6298 random devices, as specified by DEVRANDOM, until a sufficient amount
6299 of data has been collected. We spend at most 10 ms on each file
6300 (select timeout) and read in non-blocking mode. DEVRANDOM now
6301 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6302 (previously it was just the string "/dev/urandom"), so on typical
6303 platforms the 10 ms delay will never occur.
6304 Also separate out the Unix variant to its own file, rand_unix.c.
6305 For VMS, there's a currently-empty rand_vms.c.
6308 *) Move OCSP client related routines to ocsp_cl.c. These
6309 provide utility functions which an application needing
6310 to issue a request to an OCSP responder and analyse the
6311 response will typically need: as opposed to those which an
6312 OCSP responder itself would need which will be added later.
6314 OCSP_request_sign() signs an OCSP request with an API similar
6315 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6316 response. OCSP_response_get1_basic() extracts basic response
6317 from response. OCSP_resp_find_status(): finds and extracts status
6318 information from an OCSP_CERTID structure (which will be created
6319 when the request structure is built). These are built from lower
6320 level functions which work on OCSP_SINGLERESP structures but
6321 wont normally be used unless the application wishes to examine
6322 extensions in the OCSP response for example.
6324 Replace nonce routines with a pair of functions.
6325 OCSP_request_add1_nonce() adds a nonce value and optionally
6326 generates a random value. OCSP_check_nonce() checks the
6327 validity of the nonce in an OCSP response.
6330 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6331 This doesn't copy the supplied OCSP_CERTID and avoids the
6332 need to free up the newly created id. Change return type
6333 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6334 This can then be used to add extensions to the request.
6335 Deleted OCSP_request_new(), since most of its functionality
6336 is now in OCSP_REQUEST_new() (and the case insensitive name
6337 clash) apart from the ability to set the request name which
6338 will be added elsewhere.
6341 *) Update OCSP API. Remove obsolete extensions argument from
6342 various functions. Extensions are now handled using the new
6343 OCSP extension code. New simple OCSP HTTP function which
6344 can be used to send requests and parse the response.
6347 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6348 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6349 uses the special reorder version of SET OF to sort the attributes
6350 and reorder them to match the encoded order. This resolves a long
6351 standing problem: a verify on a PKCS7 structure just after signing
6352 it used to fail because the attribute order did not match the
6353 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6354 it uses the received order. This is necessary to tolerate some broken
6355 software that does not order SET OF. This is handled by encoding
6356 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6357 to produce the required SET OF.
6360 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6361 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6362 files to get correct declarations of the ASN.1 item variables.
6365 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6366 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6367 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6368 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6369 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6370 ASN1_ITEM and no wrapper functions.
6373 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6374 replace the old function pointer based I/O routines. Change most of
6375 the *_d2i_bio() and *_d2i_fp() functions to use these.
6378 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6379 lines, recognice more "algorithms" that can be deselected, and make
6380 it complain about algorithm deselection that isn't recognised.
6383 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6384 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6385 to use new functions. Add NO_ASN1_OLD which can be set to remove
6386 some old style ASN1 functions: this can be used to determine if old
6387 code will still work when these eventually go away.
6390 *) New extension functions for OCSP structures, these follow the
6391 same conventions as certificates and CRLs.
6394 *) New function X509V3_add1_i2d(). This automatically encodes and
6395 adds an extension. Its behaviour can be customised with various
6396 flags to append, replace or delete. Various wrappers added for
6397 certifcates and CRLs.
6400 *) Fix to avoid calling the underlying ASN1 print routine when
6401 an extension cannot be parsed. Correct a typo in the
6402 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6405 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
6406 entries for variables.
6409 *) Add functionality to apps/openssl.c for detecting locking
6410 problems: As the program is single-threaded, all we have
6411 to do is register a locking callback using an array for
6412 storing which locks are currently held by the program.
6415 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6416 SSL_get_ex_data_X509_STORE_idx(), which is used in
6417 ssl_verify_cert_chain() and thus can be called at any time
6418 during TLS/SSL handshakes so that thread-safety is essential.
6419 Unfortunately, the ex_data design is not at all suited
6420 for multi-threaded use, so it probably should be abolished.
6423 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6424 [Broadcom, tweaked and integrated by Geoff Thorpe]
6426 *) Move common extension printing code to new function
6427 X509V3_print_extensions(). Reorganise OCSP print routines and
6428 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6431 *) New function X509_signature_print() to remove duplication in some
6435 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6436 set (this was treated exactly the same as SET OF previously). This
6437 is used to reorder the STACK representing the structure to match the
6438 encoding. This will be used to get round a problem where a PKCS7
6439 structure which was signed could not be verified because the STACK
6440 order did not reflect the encoded order.
6443 *) Reimplement the OCSP ASN1 module using the new code.
6446 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6447 for its ASN1 operations. The old style function pointers still exist
6448 for now but they will eventually go away.
6451 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6452 completely replaces the old ASN1 functionality with a table driven
6453 encoder and decoder which interprets an ASN1_ITEM structure describing
6454 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6455 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6456 has also been converted to the new form.
6459 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6460 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6461 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6462 for negative moduli.
6465 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6466 of not touching the result's sign bit.
6469 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6473 *) Changed the LHASH code to use prototypes for callbacks, and created
6474 macros to declare and implement thin (optionally static) functions
6475 that provide type-safety and avoid function pointer casting for the
6476 type-specific callbacks.
6479 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6481 [Veers Staats <staatsvr@asc.hpc.mil>,
6482 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6484 *) Reformat the FAQ so the different questions and answers can be divided
6485 in sections depending on the subject.
6488 *) Have the zlib compression code load ZLIB.DLL dynamically under
6492 *) New function BN_mod_sqrt for computing square roots modulo a prime
6493 (using the probabilistic Tonelli-Shanks algorithm unless
6494 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6495 be handled deterministically).
6496 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6498 *) Make BN_mod_inverse faster by explicitly handling small quotients
6499 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6500 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6503 *) New function BN_kronecker.
6506 *) Fix BN_gcd so that it works on negative inputs; the result is
6507 positive unless both parameters are zero.
6508 Previously something reasonably close to an infinite loop was
6509 possible because numbers could be growing instead of shrinking
6510 in the implementation of Euclid's algorithm.
6513 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6514 sign of the number in question.
6516 Fix BN_is_word(a,w) to work correctly for w == 0.
6518 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6519 because its test if the absolute value of 'a' equals 'w'.
6520 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6521 it exists mostly for use in the implementations of BN_is_zero(),
6522 BN_is_one(), and BN_is_word().
6525 *) New function BN_swap.
6528 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6529 the exponentiation functions are more likely to produce reasonable
6530 results on negative inputs.
6533 *) Change BN_mod_mul so that the result is always non-negative.
6534 Previously, it could be negative if one of the factors was negative;
6535 I don't think anyone really wanted that behaviour.
6538 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6539 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6540 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6541 and add new functions:
6550 BN_mod_lshift1_quick
6554 These functions always generate non-negative results.
6556 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6557 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6559 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6560 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6561 be reduced modulo m.
6562 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6565 The following entry accidentily appeared in the CHANGES file
6566 distributed with OpenSSL 0.9.7. The modifications described in
6567 it do *not* apply to OpenSSL 0.9.7.
6569 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6570 was actually never needed) and in BN_mul(). The removal in BN_mul()
6571 required a small change in bn_mul_part_recursive() and the addition
6572 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6573 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6574 bn_sub_words() and bn_add_words() except they take arrays with
6579 *) In 'openssl passwd', verify passwords read from the terminal
6580 unless the '-salt' option is used (which usually means that
6581 verification would just waste user's time since the resulting
6582 hash is going to be compared with some given password hash)
6583 or the new '-noverify' option is used.
6585 This is an incompatible change, but it does not affect
6586 non-interactive use of 'openssl passwd' (passwords on the command
6587 line, '-stdin' option, '-in ...' option) and thus should not
6591 *) Remove all references to RSAref, since there's no more need for it.
6594 *) Make DSO load along a path given through an environment variable
6595 (SHLIB_PATH) with shl_load().
6598 *) Constify the ENGINE code as a result of BIGNUM constification.
6599 Also constify the RSA code and most things related to it. In a
6600 few places, most notable in the depth of the ASN.1 code, ugly
6601 casts back to non-const were required (to be solved at a later
6605 *) Make it so the openssl application has all engines loaded by default.
6608 *) Constify the BIGNUM routines a little more.
6611 *) Add the following functions:
6613 ENGINE_load_cswift()
6615 ENGINE_load_atalla()
6617 ENGINE_load_builtin_engines()
6619 That way, an application can itself choose if external engines that
6620 are built-in in OpenSSL shall ever be used or not. The benefit is
6621 that applications won't have to be linked with libdl or other dso
6622 libraries unless it's really needed.
6624 Changed 'openssl engine' to load all engines on demand.
6625 Changed the engine header files to avoid the duplication of some
6626 declarations (they differed!).
6629 *) 'openssl engine' can now list capabilities.
6632 *) Better error reporting in 'openssl engine'.
6635 *) Never call load_dh_param(NULL) in s_server.
6638 *) Add engine application. It can currently list engines by name and
6639 identity, and test if they are actually available.
6642 *) Improve RPM specification file by forcing symbolic linking and making
6643 sure the installed documentation is also owned by root.root.
6644 [Damien Miller <djm@mindrot.org>]
6646 *) Give the OpenSSL applications more possibilities to make use of
6647 keys (public as well as private) handled by engines.
6650 *) Add OCSP code that comes from CertCo.
6653 *) Add VMS support for the Rijndael code.
6656 *) Added untested support for Nuron crypto accelerator.
6659 *) Add support for external cryptographic devices. This code was
6660 previously distributed separately as the "engine" branch.
6661 [Geoff Thorpe, Richard Levitte]
6663 *) Rework the filename-translation in the DSO code. It is now possible to
6664 have far greater control over how a "name" is turned into a filename
6665 depending on the operating environment and any oddities about the
6666 different shared library filenames on each system.
6669 *) Support threads on FreeBSD-elf in Configure.
6672 *) Fix for SHA1 assembly problem with MASM: it produces
6673 warnings about corrupt line number information when assembling
6674 with debugging information. This is caused by the overlapping
6676 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6679 NCONF_get_number() has no error checking at all. As a replacement,
6680 NCONF_get_number_e() is defined (_e for "error checking") and is
6681 promoted strongly. The old NCONF_get_number is kept around for
6682 binary backward compatibility.
6683 Make it possible for methods to load from something other than a BIO,
6684 by providing a function pointer that is given a name instead of a BIO.
6685 For example, this could be used to load configuration data from an
6689 *) Fix for non blocking accept BIOs. Added new I/O special reason
6690 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6691 with non blocking I/O was not possible because no retry code was
6692 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6696 *) Added the beginnings of Rijndael support.
6699 *) Fix for bug in DirectoryString mask setting. Add support for
6700 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6701 to allow certificate printing to more controllable, additional
6702 'certopt' option to 'x509' to allow new printing options to be
6706 *) Clean old EAY MD5 hack from e_os.h.
6709 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6711 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6712 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6713 [Joe Orton, Steve Henson]
6715 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6717 *) Fix additional bug revealed by the NISCC test suite:
6719 Stop bug triggering large recursion when presented with
6720 certain ASN.1 tags (CVE-2003-0851)
6723 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6725 *) Fix various bugs revealed by running the NISCC test suite:
6727 Stop out of bounds reads in the ASN1 code when presented with
6728 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6730 If verify callback ignores invalid public key errors don't try to check
6731 certificate signature with the NULL public key.
6735 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6736 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6740 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6741 extra data after the compression methods not only for TLS 1.0
6742 but also for SSL 3.0 (as required by the specification).
6743 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6745 *) Change X509_certificate_type() to mark the key as exported/exportable
6746 when it's 512 *bits* long, not 512 bytes.
6749 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6751 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6752 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6753 a protocol version number mismatch like a decryption error
6754 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6757 *) Turn on RSA blinding by default in the default implementation
6758 to avoid a timing attack. Applications that don't want it can call
6759 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6760 They would be ill-advised to do so in most cases.
6761 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6763 *) Change RSA blinding code so that it works when the PRNG is not
6764 seeded (in this case, the secret RSA exponent is abused as
6765 an unpredictable seed -- if it is not unpredictable, there
6766 is no point in blinding anyway). Make RSA blinding thread-safe
6767 by remembering the creator's thread ID in rsa->blinding and
6768 having all other threads use local one-time blinding factors
6769 (this requires more computation than sharing rsa->blinding, but
6770 avoids excessive locking; and if an RSA object is not shared
6771 between threads, blinding will still be very fast).
6774 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6776 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6777 via timing by performing a MAC computation even if incorrrect
6778 block cipher padding has been found. This is a countermeasure
6779 against active attacks where the attacker has to distinguish
6780 between bad padding and a MAC verification error. (CVE-2003-0078)
6782 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6783 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6784 Martin Vuagnoux (EPFL, Ilion)]
6786 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6788 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6789 memory from it's contents. This is done with a counter that will
6790 place alternating values in each byte. This can be used to solve
6791 two issues: 1) the removal of calls to memset() by highly optimizing
6792 compilers, and 2) cleansing with other values than 0, since those can
6793 be read through on certain media, for example a swap space on disk.
6796 *) Bugfix: client side session caching did not work with external caching,
6797 because the session->cipher setting was not restored when reloading
6798 from the external cache. This problem was masked, when
6799 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6800 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6803 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6804 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6805 [Zeev Lieber <zeev-l@yahoo.com>]
6807 *) Undo an undocumented change introduced in 0.9.6e which caused
6808 repeated calls to OpenSSL_add_all_ciphers() and
6809 OpenSSL_add_all_digests() to be ignored, even after calling
6813 *) Change the default configuration reader to deal with last line not
6814 being properly terminated.
6817 *) Change X509_NAME_cmp() so it applies the special rules on handling
6818 DN values that are of type PrintableString, as well as RDNs of type
6819 emailAddress where the value has the type ia5String.
6820 [stefank@valicert.com via Richard Levitte]
6822 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6823 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6824 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6825 the bitwise-OR of the two for use by the majority of applications
6826 wanting this behaviour, and update the docs. The documented
6827 behaviour and actual behaviour were inconsistent and had been
6828 changing anyway, so this is more a bug-fix than a behavioural
6830 [Geoff Thorpe, diagnosed by Nadav Har'El]
6832 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6833 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6836 *) Fix initialization code race conditions in
6837 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6838 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6839 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6840 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6841 ssl2_get_cipher_by_char(),
6842 ssl3_get_cipher_by_char().
6843 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6845 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6846 the cached sessions are flushed, as the remove_cb() might use ex_data
6847 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6848 (see [openssl.org #212]).
6849 [Geoff Thorpe, Lutz Jaenicke]
6851 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6852 length, instead of the encoding length to d2i_ASN1_OBJECT.
6855 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6857 *) [In 0.9.6g-engine release:]
6858 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6859 [Lynn Gazis <lgazis@rainbow.com>]
6861 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6863 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6864 and get fix the header length calculation.
6865 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6866 Alon Kantor <alonk@checkpoint.com> (and others),
6869 *) Use proper error handling instead of 'assertions' in buffer
6870 overflow checks added in 0.9.6e. This prevents DoS (the
6871 assertions could call abort()).
6872 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6874 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6876 *) Add various sanity checks to asn1_get_length() to reject
6877 the ASN1 length bytes if they exceed sizeof(long), will appear
6878 negative or the content length exceeds the length of the
6880 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6882 *) Fix cipher selection routines: ciphers without encryption had no flags
6883 for the cipher strength set and where therefore not handled correctly
6884 by the selection routines (PR #130).
6887 *) Fix EVP_dsa_sha macro.
6891 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6892 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6893 that was added in OpenSSL 0.9.6d.
6895 As the countermeasure turned out to be incompatible with some
6896 broken SSL implementations, the new option is part of SSL_OP_ALL.
6897 SSL_OP_ALL is usually employed when compatibility with weird SSL
6898 implementations is desired (e.g. '-bugs' option to 's_client' and
6899 's_server'), so the new option is automatically set in many
6903 *) Changes in security patch:
6905 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6906 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6907 Air Force Materiel Command, USAF, under agreement number
6910 *) Add various sanity checks to asn1_get_length() to reject
6911 the ASN1 length bytes if they exceed sizeof(long), will appear
6912 negative or the content length exceeds the length of the
6913 supplied buffer. (CVE-2002-0659)
6914 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6916 *) Assertions for various potential buffer overflows, not known to
6918 [Ben Laurie (CHATS)]
6920 *) Various temporary buffers to hold ASCII versions of integers were
6921 too small for 64 bit platforms. (CVE-2002-0655)
6922 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6924 *) Remote buffer overflow in SSL3 protocol - an attacker could
6925 supply an oversized session ID to a client. (CVE-2002-0656)
6926 [Ben Laurie (CHATS)]
6928 *) Remote buffer overflow in SSL2 protocol - an attacker could
6929 supply an oversized client master key. (CVE-2002-0656)
6930 [Ben Laurie (CHATS)]
6932 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6934 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6935 encoded as NULL) with id-dsa-with-sha1.
6936 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6938 *) Check various X509_...() return values in apps/req.c.
6939 [Nils Larsch <nla@trustcenter.de>]
6941 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6942 an end-of-file condition would erronously be flagged, when the CRLF
6943 was just at the end of a processed block. The bug was discovered when
6944 processing data through a buffering memory BIO handing the data to a
6945 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6946 <ptsekov@syntrex.com> and Nedelcho Stanev.
6949 *) Implement a countermeasure against a vulnerability recently found
6950 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6951 before application data chunks to avoid the use of known IVs
6952 with data potentially chosen by the attacker.
6955 *) Fix length checks in ssl3_get_client_hello().
6958 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6959 to prevent ssl3_read_internal() from incorrectly assuming that
6960 ssl3_read_bytes() found application data while handshake
6961 processing was enabled when in fact s->s3->in_read_app_data was
6962 merely automatically cleared during the initial handshake.
6963 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6965 *) Fix object definitions for Private and Enterprise: they were not
6966 recognized in their shortname (=lowercase) representation. Extend
6967 obj_dat.pl to issue an error when using undefined keywords instead
6968 of silently ignoring the problem (Svenning Sorensen
6969 <sss@sss.dnsalias.net>).
6972 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6973 generators, i.e. generators other than 2 and 5. (Previously, the
6974 code did not properly initialise the 'add' and 'rem' values to
6975 BN_generate_prime().)
6977 In the new general case, we do not insist that 'generator' is
6978 actually a primitive root: This requirement is rather pointless;
6979 a generator of the order-q subgroup is just as good, if not
6983 *) Map new X509 verification errors to alerts. Discovered and submitted by
6984 Tom Wu <tom@arcot.com>.
6987 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6988 returning non-zero before the data has been completely received
6989 when using non-blocking I/O.
6990 [Bodo Moeller; problem pointed out by John Hughes]
6992 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6993 [Ben Laurie, Lutz Jaenicke]
6995 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6996 Yoram Zahavi <YoramZ@gilian.com>).
6999 *) Add information about CygWin 1.3 and on, and preserve proper
7000 configuration for the versions before that.
7001 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7003 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7004 check whether we deal with a copy of a session and do not delete from
7005 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7006 <izhar@checkpoint.com>.
7009 *) Do not store session data into the internal session cache, if it
7010 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7011 flag is set). Proposed by Aslam <aslam@funk.com>.
7014 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7018 *) [In 0.9.6d-engine release:]
7019 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7020 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7022 *) Add the configuration target linux-s390x.
7023 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7025 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7026 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7027 variable as an indication that a ClientHello message has been
7028 received. As the flag value will be lost between multiple
7029 invocations of ssl3_accept when using non-blocking I/O, the
7030 function may not be aware that a handshake has actually taken
7031 place, thus preventing a new session from being added to the
7034 To avoid this problem, we now set s->new_session to 2 instead of
7035 using a local variable.
7036 [Lutz Jaenicke, Bodo Moeller]
7038 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7039 if the SSL_R_LENGTH_MISMATCH error is detected.
7040 [Geoff Thorpe, Bodo Moeller]
7042 *) New 'shared_ldflag' column in Configure platform table.
7045 *) Fix EVP_CIPHER_mode macro.
7046 ["Dan S. Camper" <dan@bti.net>]
7048 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7049 type, we must throw them away by setting rr->length to 0.
7050 [D P Chang <dpc@qualys.com>]
7052 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7054 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7055 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7056 worked incorrectly for those cases where range = 10..._2 and
7057 3*range is two bits longer than range.)
7060 *) Only add signing time to PKCS7 structures if it is not already
7064 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7065 OBJ_ld_ce should be OBJ_id_ce.
7066 Also some ip-pda OIDs in crypto/objects/objects.txt were
7067 incorrect (cf. RFC 3039).
7068 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7070 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7071 returns early because it has nothing to do.
7072 [Andy Schneider <andy.schneider@bjss.co.uk>]
7074 *) [In 0.9.6c-engine release:]
7075 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7076 [Andy Schneider <andy.schneider@bjss.co.uk>]
7078 *) [In 0.9.6c-engine release:]
7079 Add support for Cryptographic Appliance's keyserver technology.
7080 (Use engine 'keyclient')
7081 [Cryptographic Appliances and Geoff Thorpe]
7083 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7084 is called via tools/c89.sh because arguments have to be
7085 rearranged (all '-L' options must appear before the first object
7087 [Richard Shapiro <rshapiro@abinitio.com>]
7089 *) [In 0.9.6c-engine release:]
7090 Add support for Broadcom crypto accelerator cards, backported
7092 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7094 *) [In 0.9.6c-engine release:]
7095 Add support for SureWare crypto accelerator cards from
7096 Baltimore Technologies. (Use engine 'sureware')
7097 [Baltimore Technologies and Mark Cox]
7099 *) [In 0.9.6c-engine release:]
7100 Add support for crypto accelerator cards from Accelerated
7101 Encryption Processing, www.aep.ie. (Use engine 'aep')
7102 [AEP Inc. and Mark Cox]
7104 *) Add a configuration entry for gcc on UnixWare.
7105 [Gary Benson <gbenson@redhat.com>]
7107 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7108 messages are stored in a single piece (fixed-length part and
7109 variable-length part combined) and fix various bugs found on the way.
7112 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7113 instead. BIO_gethostbyname() does not know what timeouts are
7114 appropriate, so entries would stay in cache even when they have
7116 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7118 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7119 faced with a pathologically small ClientHello fragment that does
7120 not contain client_version: Instead of aborting with an error,
7121 simply choose the highest available protocol version (i.e.,
7122 TLS 1.0 unless it is disabled). In practice, ClientHello
7123 messages are never sent like this, but this change gives us
7124 strictly correct behaviour at least for TLS.
7127 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7128 never resets s->method to s->ctx->method when called from within
7129 one of the SSL handshake functions.
7130 [Bodo Moeller; problem pointed out by Niko Baric]
7132 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7133 (sent using the client's version number) if client_version is
7134 smaller than the protocol version in use. Also change
7135 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7136 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7137 the client will at least see that alert.
7140 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7144 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7145 client receives HelloRequest while in a handshake.
7146 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7148 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7149 should end in 'break', not 'goto end' which circuments various
7150 cleanups done in state SSL_ST_OK. But session related stuff
7151 must be disabled for SSL_ST_OK in the case that we just sent a
7154 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7155 before just sending a HelloRequest.
7156 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7158 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7159 reveal whether illegal block cipher padding was found or a MAC
7160 verification error occured. (Neither SSLerr() codes nor alerts
7161 are directly visible to potential attackers, but the information
7162 may leak via logfiles.)
7164 Similar changes are not required for the SSL 2.0 implementation
7165 because the number of padding bytes is sent in clear for SSL 2.0,
7166 and the extra bytes are just ignored. However ssl/s2_pkt.c
7167 failed to verify that the purported number of padding bytes is in
7171 *) Add OpenUNIX-8 support including shared libraries
7172 (Boyd Lynn Gerber <gerberb@zenez.com>).
7175 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7176 'wristwatch attack' using huge encoding parameters (cf.
7177 James H. Manger's CRYPTO 2001 paper). Note that the
7178 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7179 encoding parameters and hence was not vulnerable.
7182 *) BN_sqr() bug fix.
7183 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7185 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7186 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7187 followed by modular reduction.
7188 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7190 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7191 equivalent based on BN_pseudo_rand() instead of BN_rand().
7194 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7195 This function was broken, as the check for a new client hello message
7196 to handle SGC did not allow these large messages.
7197 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7200 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7203 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7204 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7207 *) Rework the configuration and shared library support for Tru64 Unix.
7208 The configuration part makes use of modern compiler features and
7209 still retains old compiler behavior for those that run older versions
7210 of the OS. The shared library support part includes a variant that
7211 uses the RPATH feature, and is available through the special
7212 configuration target "alpha-cc-rpath", which will never be selected
7214 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7216 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7217 with the same message size as in ssl3_get_certificate_request().
7218 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7219 messages might inadvertently be reject as too long.
7220 [Petr Lampa <lampa@fee.vutbr.cz>]
7222 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7225 *) Modified SSL library such that the verify_callback that has been set
7226 specificly for an SSL object with SSL_set_verify() is actually being
7227 used. Before the change, a verify_callback set with this function was
7228 ignored and the verify_callback() set in the SSL_CTX at the time of
7229 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7230 to allow the necessary settings.
7233 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7234 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7235 done automatically (in contradiction to the requirements of the C
7236 standard). This made problems when used from OpenSSH.
7239 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7240 dh->length and always used
7242 BN_rand_range(priv_key, dh->p).
7244 BN_rand_range() is not necessary for Diffie-Hellman, and this
7245 specific range makes Diffie-Hellman unnecessarily inefficient if
7246 dh->length (recommended exponent length) is much smaller than the
7247 length of dh->p. We could use BN_rand_range() if the order of
7248 the subgroup was stored in the DH structure, but we only have
7253 BN_rand(priv_key, l, ...)
7255 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7261 RSA_eay_public_encrypt
7262 RSA_eay_private_decrypt
7263 RSA_eay_private_encrypt (signing)
7264 RSA_eay_public_decrypt (signature verification)
7266 (default implementations for RSA_public_encrypt,
7267 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7268 always reject numbers >= n.
7271 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7272 to synchronize access to 'locking_thread'. This is necessary on
7273 systems where access to 'locking_thread' (an 'unsigned long'
7274 variable) is not atomic.
7277 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7278 *before* setting the 'crypto_lock_rand' flag. The previous code had
7279 a race condition if 0 is a valid thread ID.
7280 [Travis Vitek <vitek@roguewave.com>]
7282 *) Add support for shared libraries under Irix.
7283 [Albert Chin-A-Young <china@thewrittenword.com>]
7285 *) Add configuration option to build on Linux on both big-endian and
7287 [Ralf Baechle <ralf@uni-koblenz.de>]
7289 *) Add the possibility to create shared libraries on HP-UX.
7292 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7294 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7295 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7296 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7297 PRNG state recovery was possible based on the output of
7298 one PRNG request appropriately sized to gain knowledge on
7299 'md' followed by enough consecutive 1-byte PRNG requests
7300 to traverse all of 'state'.
7302 1. When updating 'md_local' (the current thread's copy of 'md')
7303 during PRNG output generation, hash all of the previous
7304 'md_local' value, not just the half used for PRNG output.
7306 2. Make the number of bytes from 'state' included into the hash
7307 independent from the number of PRNG bytes requested.
7309 The first measure alone would be sufficient to avoid
7310 Markku-Juhani's attack. (Actually it had never occurred
7311 to me that the half of 'md_local' used for chaining was the
7312 half from which PRNG output bytes were taken -- I had always
7313 assumed that the secret half would be used.) The second
7314 measure makes sure that additional data from 'state' is never
7315 mixed into 'md_local' in small portions; this heuristically
7316 further strengthens the PRNG.
7319 *) Fix crypto/bn/asm/mips3.s.
7322 *) When only the key is given to "enc", the IV is undefined. Print out
7323 an error message in this case.
7326 *) Handle special case when X509_NAME is empty in X509 printing routines.
7329 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7330 positive and less than q.
7333 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7334 used: it isn't thread safe and the add_lock_callback should handle
7336 [Paul Rose <Paul.Rose@bridge.com>]
7338 *) Verify that incoming data obeys the block size in
7339 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7343 [Ulf Möller, Bodo Möller]
7345 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7346 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7347 when fixing the server behaviour for backwards-compatible 'client
7348 hello' messages. (Note that the attack is impractical against
7349 SSL 3.0 and TLS 1.0 anyway because length and version checking
7350 means that the probability of guessing a valid ciphertext is
7351 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7354 Before 0.9.5, the countermeasure (hide the error by generating a
7355 random 'decryption result') did not work properly because
7356 ERR_clear_error() was missing, meaning that SSL_get_error() would
7357 detect the supposedly ignored error.
7359 Both problems are now fixed.
7362 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7363 (previously it was 1024).
7366 *) Fix for compatibility mode trust settings: ignore trust settings
7367 unless some valid trust or reject settings are present.
7370 *) Fix for blowfish EVP: its a variable length cipher.
7373 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7374 parameters in DSA public key structures and return an error in the
7375 DSA routines if parameters are absent.
7378 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7379 in the current directory if neither $RANDFILE nor $HOME was set.
7380 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7381 caused some confusion to Windows users who haven't defined $HOME.
7382 Thus RAND_file_name() is changed again: e_os.h can define a
7383 DEFAULT_HOME, which will be used if $HOME is not set.
7384 For Windows, we use "C:"; on other platforms, we still require
7385 environment variables.
7387 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7388 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7389 having multiple threads call RAND_poll() concurrently.
7392 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7393 combination of a flag and a thread ID variable.
7394 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7395 flag), *other* threads can enter ssleay_add_bytes without obeying
7396 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7397 that they do not hold after the first thread unsets add_do_not_lock).
7400 *) Change bctest again: '-x' expressions are not available in all
7404 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7406 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7407 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7409 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7410 the default extension for executables, if any. Also, make the perl
7411 scripts that use symlink() to test if it really exists and use "cp"
7412 if it doesn't. All this made OpenSSL compilable and installable in
7416 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7417 If SEQUENCE is length is indefinite just set c->slen to the total
7418 amount of data available.
7419 [Steve Henson, reported by shige@FreeBSD.org]
7420 [This change does not apply to 0.9.7.]
7422 *) Change bctest to avoid here-documents inside command substitution
7423 (workaround for FreeBSD /bin/sh bug).
7424 For compatibility with Ultrix, avoid shell functions (introduced
7425 in the bctest version that searches along $PATH).
7428 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7429 with des_encrypt() defined on some operating systems, like Solaris
7433 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7434 On the Importance of Eliminating Errors in Cryptographic
7435 Computations, J. Cryptology 14 (2001) 2, 101-119,
7436 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7439 *) MIPS assembler BIGNUM division bug fix.
7442 *) Disabled incorrect Alpha assembler code.
7445 *) Fix PKCS#7 decode routines so they correctly update the length
7446 after reading an EOC for the EXPLICIT tag.
7448 [This change does not apply to 0.9.7.]
7450 *) Fix bug in PKCS#12 key generation routines. This was triggered
7451 if a 3DES key was generated with a 0 initial byte. Include
7452 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7453 (but broken) behaviour.
7456 *) Enhance bctest to search for a working bc along $PATH and print
7458 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7460 *) Fix memory leaks in err.c: free err_data string if necessary;
7461 don't write to the wrong index in ERR_set_error_data.
7464 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7468 *) Replace rdtsc with _emit statements for VC++ version 5.
7469 [Jeremy Cooper <jeremy@baymoo.org>]
7471 *) Make it possible to reuse SSLv2 sessions.
7474 *) In copy_email() check for >= 0 as a return value for
7475 X509_NAME_get_index_by_NID() since 0 is a valid index.
7476 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7478 *) Avoid coredump with unsupported or invalid public keys by checking if
7479 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7480 PKCS7_verify() fails with non detached data.
7483 *) Don't use getenv in library functions when run as setuid/setgid.
7484 New function OPENSSL_issetugid().
7487 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7488 due to incorrect handling of multi-threading:
7490 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7492 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7494 3. Count how many times MemCheck_off() has been called so that
7495 nested use can be treated correctly. This also avoids
7496 inband-signalling in the previous code (which relied on the
7497 assumption that thread ID 0 is impossible).
7500 *) Add "-rand" option also to s_client and s_server.
7503 *) Fix CPU detection on Irix 6.x.
7504 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7505 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7507 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7510 [This change does not apply to 0.9.7.]
7512 *) Use the cached encoding of an X509_NAME structure rather than
7513 copying it. This is apparently the reason for the libsafe "errors"
7514 but the code is actually correct.
7517 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7518 Bleichenbacher's DSA attack.
7519 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7520 to be set and top=0 forces the highest bit to be set; top=-1 is new
7521 and leaves the highest bit random.
7522 [Ulf Moeller, Bodo Moeller]
7524 *) In the NCONF_...-based implementations for CONF_... queries
7525 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7526 a temporary CONF structure with the data component set to NULL
7527 (which gives segmentation faults in lh_retrieve).
7528 Instead, use NULL for the CONF pointer in CONF_get_string and
7529 CONF_get_number (which may use environment variables) and directly
7530 return NULL from CONF_get_section.
7533 *) Fix potential buffer overrun for EBCDIC.
7536 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7537 keyUsage if basicConstraints absent for a CA.
7540 *) Make SMIME_write_PKCS7() write mail header values with a format that
7541 is more generally accepted (no spaces before the semicolon), since
7542 some programs can't parse those values properly otherwise. Also make
7543 sure BIO's that break lines after each write do not create invalid
7547 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7548 macros previously used would not encode an empty SEQUENCE OF
7549 and break the signature.
7551 [This change does not apply to 0.9.7.]
7553 *) Zero the premaster secret after deriving the master secret in
7557 *) Add some EVP_add_digest_alias registrations (as found in
7558 OpenSSL_add_all_digests()) to SSL_library_init()
7559 aka OpenSSL_add_ssl_algorithms(). This provides improved
7560 compatibility with peers using X.509 certificates
7561 with unconventional AlgorithmIdentifier OIDs.
7564 *) Fix for Irix with NO_ASM.
7565 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7567 *) ./config script fixes.
7568 [Ulf Moeller, Richard Levitte]
7570 *) Fix 'openssl passwd -1'.
7573 *) Change PKCS12_key_gen_asc() so it can cope with non null
7574 terminated strings whose length is passed in the passlen
7575 parameter, for example from PEM callbacks. This was done
7576 by adding an extra length parameter to asc2uni().
7577 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7579 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7580 call failed, free the DSA structure.
7583 *) Fix to uni2asc() to cope with zero length Unicode strings.
7584 These are present in some PKCS#12 files.
7587 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7588 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7589 when writing a 32767 byte record.
7590 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7592 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7593 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7595 (RSA objects have a reference count access to which is protected
7596 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7597 so they are meant to be shared between threads.)
7598 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7599 "Reddie, Steven" <Steven.Reddie@ca.com>]
7601 *) Fix a deadlock in CRYPTO_mem_leaks().
7604 *) Use better test patterns in bntest.
7607 *) rand_win.c fix for Borland C.
7610 *) BN_rshift bugfix for n == 0.
7613 *) Add a 'bctest' script that checks for some known 'bc' bugs
7614 so that 'make test' does not abort just because 'bc' is broken.
7617 *) Store verify_result within SSL_SESSION also for client side to
7618 avoid potential security hole. (Re-used sessions on the client side
7619 always resulted in verify_result==X509_V_OK, not using the original
7620 result of the server certificate verification.)
7623 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7624 SSL3_RT_APPLICATION_DATA, return 0.
7625 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7629 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7630 releases, have been re-implemented by renaming the previous
7631 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7632 and ssl3_read_internal, respectively, and adding 'peek' parameters
7633 to them. The new ssl[23]_{read,peek} functions are calls to
7634 ssl[23]_read_internal with the 'peek' flag set appropriately.
7635 A 'peek' parameter has also been added to ssl3_read_bytes, which
7636 does the actual work for ssl3_read_internal.
7639 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7640 the method-specific "init()" handler. Also clean up ex_data after
7641 calling the method-specific "finish()" handler. Previously, this was
7642 happening the other way round.
7645 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7646 The previous value, 12, was not always sufficient for BN_mod_exp().
7649 *) Make sure that shared libraries get the internal name engine with
7650 the full version number and not just 0. This should mark the
7651 shared libraries as not backward compatible. Of course, this should
7652 be changed again when we can guarantee backward binary compatibility.
7655 *) Fix typo in get_cert_by_subject() in by_dir.c
7656 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7658 *) Rework the system to generate shared libraries:
7660 - Make note of the expected extension for the shared libraries and
7661 if there is a need for symbolic links from for example libcrypto.so.0
7662 to libcrypto.so.0.9.7. There is extended info in Configure for
7665 - Make as few rebuilds of the shared libraries as possible.
7667 - Still avoid linking the OpenSSL programs with the shared libraries.
7669 - When installing, install the shared libraries separately from the
7673 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7675 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7676 and not in SSL_clear because the latter is also used by the
7677 accept/connect functions; previously, the settings made by
7678 SSL_set_read_ahead would be lost during the handshake.
7679 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7681 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7682 Previously, it would create entries for disableed algorithms no
7686 *) Added several new manual pages for SSL_* function.
7689 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7691 *) In ssl23_get_client_hello, generate an error message when faced
7692 with an initial SSL 3.0/TLS record that is too small to contain the
7693 first two bytes of the ClientHello message, i.e. client_version.
7694 (Note that this is a pathologic case that probably has never happened
7695 in real life.) The previous approach was to use the version number
7696 from the record header as a substitute; but our protocol choice
7697 should not depend on that one because it is not authenticated
7698 by the Finished messages.
7701 *) More robust randomness gathering functions for Windows.
7702 [Jeffrey Altman <jaltman@columbia.edu>]
7704 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7705 not set then we don't setup the error code for issuer check errors
7706 to avoid possibly overwriting other errors which the callback does
7707 handle. If an application does set the flag then we assume it knows
7708 what it is doing and can handle the new informational codes
7712 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7713 a general "ANY" type, as such it should be able to decode anything
7714 including tagged types. However it didn't check the class so it would
7715 wrongly interpret tagged types in the same way as their universal
7716 counterpart and unknown types were just rejected. Changed so that the
7717 tagged and unknown types are handled in the same way as a SEQUENCE:
7718 that is the encoding is stored intact. There is also a new type
7719 "V_ASN1_OTHER" which is used when the class is not universal, in this
7720 case we have no idea what the actual type is so we just lump them all
7724 *) On VMS, stdout may very well lead to a file that is written to
7725 in a record-oriented fashion. That means that every write() will
7726 write a separate record, which will be read separately by the
7727 programs trying to read from it. This can be very confusing.
7729 The solution is to put a BIO filter in the way that will buffer
7730 text until a linefeed is reached, and then write everything a
7731 line at a time, so every record written will be an actual line,
7732 not chunks of lines and not (usually doesn't happen, but I've
7733 seen it once) several lines in one record. BIO_f_linebuffer() is
7736 Currently, it's a VMS-only method, because that's where it has
7737 been tested well enough.
7740 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7741 it can return incorrect results.
7742 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7743 but it was in 0.9.6-beta[12].)
7746 *) Disable the check for content being present when verifying detached
7747 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7748 include zero length content when signing messages.
7751 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7752 BIO_ctrl (for BIO pairs).
7755 *) Add DSO method for VMS.
7758 *) Bug fix: Montgomery multiplication could produce results with the
7762 *) Add RPM specification openssl.spec and modify it to build three
7763 packages. The default package contains applications, application
7764 documentation and run-time libraries. The devel package contains
7765 include files, static libraries and function documentation. The
7766 doc package contains the contents of the doc directory. The original
7767 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7770 *) Add a large number of documentation files for many SSL routines.
7771 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7773 *) Add a configuration entry for Sony News 4.
7774 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7776 *) Don't set the two most significant bits to one when generating a
7777 random number < q in the DSA library.
7780 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7781 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7782 the underlying transport is blocking) if a handshake took place.
7783 (The default behaviour is needed by applications such as s_client
7784 and s_server that use select() to determine when to use SSL_read;
7785 but for applications that know in advance when to expect data, it
7786 just makes things more complicated.)
7789 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7793 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7794 work better on such systems.
7795 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7797 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7798 Update PKCS12_parse() so it copies the friendlyName and the
7799 keyid to the certificates aux info.
7802 *) Fix bug in PKCS7_verify() which caused an infinite loop
7803 if there was more than one signature.
7804 [Sven Uszpelkat <su@celocom.de>]
7806 *) Major change in util/mkdef.pl to include extra information
7807 about each symbol, as well as presentig variables as well
7808 as functions. This change means that there's n more need
7809 to rebuild the .num files when some algorithms are excluded.
7812 *) Allow the verify time to be set by an application,
7813 rather than always using the current time.
7816 *) Phase 2 verify code reorganisation. The certificate
7817 verify code now looks up an issuer certificate by a
7818 number of criteria: subject name, authority key id
7819 and key usage. It also verifies self signed certificates
7820 by the same criteria. The main comparison function is
7821 X509_check_issued() which performs these checks.
7823 Lot of changes were necessary in order to support this
7824 without completely rewriting the lookup code.
7826 Authority and subject key identifier are now cached.
7828 The LHASH 'certs' is X509_STORE has now been replaced
7829 by a STACK_OF(X509_OBJECT). This is mainly because an
7830 LHASH can't store or retrieve multiple objects with
7831 the same hash value.
7833 As a result various functions (which were all internal
7834 use only) have changed to handle the new X509_STORE
7835 structure. This will break anything that messed round
7836 with X509_STORE internally.
7838 The functions X509_STORE_add_cert() now checks for an
7839 exact match, rather than just subject name.
7841 The X509_STORE API doesn't directly support the retrieval
7842 of multiple certificates matching a given criteria, however
7843 this can be worked round by performing a lookup first
7844 (which will fill the cache with candidate certificates)
7845 and then examining the cache for matches. This is probably
7846 the best we can do without throwing out X509_LOOKUP
7847 entirely (maybe later...).
7849 The X509_VERIFY_CTX structure has been enhanced considerably.
7851 All certificate lookup operations now go via a get_issuer()
7852 callback. Although this currently uses an X509_STORE it
7853 can be replaced by custom lookups. This is a simple way
7854 to bypass the X509_STORE hackery necessary to make this
7855 work and makes it possible to use more efficient techniques
7856 in future. A very simple version which uses a simple
7857 STACK for its trusted certificate store is also provided
7858 using X509_STORE_CTX_trusted_stack().
7860 The verify_cb() and verify() callbacks now have equivalents
7861 in the X509_STORE_CTX structure.
7863 X509_STORE_CTX also has a 'flags' field which can be used
7864 to customise the verify behaviour.
7867 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7868 excludes S/MIME capabilities.
7871 *) When a certificate request is read in keep a copy of the
7872 original encoding of the signed data and use it when outputing
7873 again. Signatures then use the original encoding rather than
7874 a decoded, encoded version which may cause problems if the
7875 request is improperly encoded.
7878 *) For consistency with other BIO_puts implementations, call
7879 buffer_write(b, ...) directly in buffer_puts instead of calling
7882 In BIO_puts, increment b->num_write as in BIO_write.
7883 [Peter.Sylvester@EdelWeb.fr]
7885 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7886 BN_zero, we may not return a BIGNUM with an array consisting of
7890 *) Avoid calling abort() from within the library when problems are
7891 detected, except if preprocessor symbols have been defined
7892 (such as REF_CHECK, BN_DEBUG etc.).
7895 *) New openssl application 'rsautl'. This utility can be
7896 used for low level RSA operations. DER public key
7897 BIO/fp routines also added.
7900 *) New Configure entry and patches for compiling on QNX 4.
7901 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7903 *) A demo state-machine implementation was sponsored by
7904 Nuron (http://www.nuron.com/) and is now available in
7905 demos/state_machine.
7908 *) New options added to the 'dgst' utility for signature
7909 generation and verification.
7912 *) Unrecognized PKCS#7 content types are now handled via a
7913 catch all ASN1_TYPE structure. This allows unsupported
7914 types to be stored as a "blob" and an application can
7915 encode and decode it manually.
7918 *) Fix various signed/unsigned issues to make a_strex.c
7920 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7922 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7923 length if passed a buffer. ASN1_INTEGER_to_BN failed
7924 if passed a NULL BN and its argument was negative.
7925 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7927 *) Modification to PKCS#7 encoding routines to output definite
7928 length encoding. Since currently the whole structures are in
7929 memory there's not real point in using indefinite length
7930 constructed encoding. However if OpenSSL is compiled with
7931 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7934 *) Added BIO_vprintf() and BIO_vsnprintf().
7937 *) Added more prefixes to parse for in the the strings written
7938 through a logging bio, to cover all the levels that are available
7939 through syslog. The prefixes are now:
7941 PANIC, EMERG, EMR => LOG_EMERG
7942 ALERT, ALR => LOG_ALERT
7943 CRIT, CRI => LOG_CRIT
7944 ERROR, ERR => LOG_ERR
7945 WARNING, WARN, WAR => LOG_WARNING
7946 NOTICE, NOTE, NOT => LOG_NOTICE
7947 INFO, INF => LOG_INFO
7948 DEBUG, DBG => LOG_DEBUG
7950 and as before, if none of those prefixes are present at the
7951 beginning of the string, LOG_ERR is chosen.
7953 On Win32, the LOG_* levels are mapped according to this:
7955 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7956 LOG_WARNING => EVENTLOG_WARNING_TYPE
7957 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7961 *) Made it possible to reconfigure with just the configuration
7962 argument "reconf" or "reconfigure". The command line arguments
7963 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7964 and are retrieved from there when reconfiguring.
7968 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7970 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7973 *) The obj_dat.pl script was messing up the sorting of object
7974 names. The reason was that it compared the quoted version
7975 of strings as a result "OCSP" > "OCSP Signing" because
7976 " > SPACE. Changed script to store unquoted versions of
7977 names and add quotes on output. It was also omitting some
7978 names from the lookup table if they were given a default
7979 value (that is if SN is missing it is given the same
7980 value as LN and vice versa), these are now added on the
7981 grounds that if an object has a name we should be able to
7982 look it up. Finally added warning output when duplicate
7983 short or long names are found.
7986 *) Changes needed for Tandem NSK.
7987 [Scott Uroff <scott@xypro.com>]
7989 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7990 RSA_padding_check_SSLv23(), special padding was never detected
7991 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7992 version rollback attacks was not effective.
7994 In s23_clnt.c, don't use special rollback-attack detection padding
7995 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7996 client; similarly, in s23_srvr.c, don't do the rollback check if
7997 SSL 2.0 is the only protocol enabled in the server.
8000 *) Make it possible to get hexdumps of unprintable data with 'openssl
8001 asn1parse'. By implication, the functions ASN1_parse_dump() and
8002 BIO_dump_indent() are added.
8005 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8006 these print out strings and name structures based on various
8007 flags including RFC2253 support and proper handling of
8008 multibyte characters. Added options to the 'x509' utility
8009 to allow the various flags to be set.
8012 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8013 Also change the functions X509_cmp_current_time() and
8014 X509_gmtime_adj() work with an ASN1_TIME structure,
8015 this will enable certificates using GeneralizedTime in validity
8016 dates to be checked.
8019 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8020 negative public key encodings) on by default,
8021 NO_NEG_PUBKEY_BUG can be set to disable it.
8024 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8025 content octets. An i2c_ASN1_OBJECT is unnecessary because
8026 the encoding can be trivially obtained from the structure.
8029 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8030 not read locks (CRYPTO_r_[un]lock).
8033 *) A first attempt at creating official support for shared
8034 libraries through configuration. I've kept it so the
8035 default is static libraries only, and the OpenSSL programs
8036 are always statically linked for now, but there are
8037 preparations for dynamic linking in place.
8038 This has been tested on Linux and Tru64.
8041 *) Randomness polling function for Win9x, as described in:
8042 Peter Gutmann, Software Generation of Practically Strong
8046 *) Fix so PRNG is seeded in req if using an already existing
8050 *) New options to smime application. -inform and -outform
8051 allow alternative formats for the S/MIME message including
8052 PEM and DER. The -content option allows the content to be
8053 specified separately. This should allow things like Netscape
8054 form signing output easier to verify.
8057 *) Fix the ASN1 encoding of tags using the 'long form'.
8060 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8061 STRING types. These convert content octets to and from the
8062 underlying type. The actual tag and length octets are
8063 already assumed to have been read in and checked. These
8064 are needed because all other string types have virtually
8065 identical handling apart from the tag. By having versions
8066 of the ASN1 functions that just operate on content octets
8067 IMPLICIT tagging can be handled properly. It also allows
8068 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8069 and ASN1_INTEGER are identical apart from the tag.
8072 *) Change the handling of OID objects as follows:
8074 - New object identifiers are inserted in objects.txt, following
8075 the syntax given in objects.README.
8076 - objects.pl is used to process obj_mac.num and create a new
8078 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8081 This is currently kind of a hack, and the perl code in objects.pl
8082 isn't very elegant, but it works as I intended. The simplest way
8083 to check that it worked correctly is to look in obj_dat.h and
8084 check the array nid_objs and make sure the objects haven't moved
8085 around (this is important!). Additions are OK, as well as
8086 consistent name changes.
8089 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8092 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8093 The given file adds to whatever has already been seeded into the
8094 random pool through the RANDFILE configuration file option or
8095 environment variable, or the default random state file.
8098 *) mkstack.pl now sorts each macro group into lexical order.
8099 Previously the output order depended on the order the files
8100 appeared in the directory, resulting in needless rewriting
8104 *) Patches to make OpenSSL compile under Win32 again. Mostly
8105 work arounds for the VC++ problem that it treats func() as
8106 func(void). Also stripped out the parts of mkdef.pl that
8107 added extra typesafe functions: these no longer exist.
8110 *) Reorganisation of the stack code. The macros are now all
8111 collected in safestack.h . Each macro is defined in terms of
8112 a "stack macro" of the form SKM_<name>(type, a, b). The
8113 DEBUG_SAFESTACK is now handled in terms of function casts,
8114 this has the advantage of retaining type safety without the
8115 use of additional functions. If DEBUG_SAFESTACK is not defined
8116 then the non typesafe macros are used instead. Also modified the
8117 mkstack.pl script to handle the new form. Needs testing to see
8118 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8119 the default if no major problems. Similar behaviour for ASN1_SET_OF
8120 and PKCS12_STACK_OF.
8123 *) When some versions of IIS use the 'NET' form of private key the
8124 key derivation algorithm is different. Normally MD5(password) is
8125 used as a 128 bit RC4 key. In the modified case
8126 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
8127 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8128 as the old Netscape_RSA functions except they have an additional
8129 'sgckey' parameter which uses the modified algorithm. Also added
8130 an -sgckey command line option to the rsa utility. Thanks to
8131 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8132 algorithm to openssl-dev.
8135 *) The evp_local.h macros were using 'c.##kname' which resulted in
8136 invalid expansion on some systems (SCO 5.0.5 for example).
8137 Corrected to 'c.kname'.
8138 [Phillip Porch <root@theporch.com>]
8140 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8141 a STACK of email addresses from a certificate or request, these look
8142 in the subject name and the subject alternative name extensions and
8143 omit any duplicate addresses.
8146 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8147 This makes DSA verification about 2 % faster.
8150 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8151 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8152 plus overhead for 1024 bit moduli).
8153 This makes exponentiations about 0.5 % faster for 1024 bit
8154 exponents (as measured by "openssl speed rsa2048").
8157 *) Rename memory handling macros to avoid conflicts with other
8159 Malloc => OPENSSL_malloc
8160 Malloc_locked => OPENSSL_malloc_locked
8161 Realloc => OPENSSL_realloc
8162 Free => OPENSSL_free
8165 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8166 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8169 *) CygWin32 support.
8170 [John Jarvie <jjarvie@newsguy.com>]
8172 *) The type-safe stack code has been rejigged. It is now only compiled
8173 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8174 by default all type-specific stack functions are "#define"d back to
8175 standard stack functions. This results in more streamlined output
8176 but retains the type-safety checking possibilities of the original
8180 *) The STACK code has been cleaned up, and certain type declarations
8181 that didn't make a lot of sense have been brought in line. This has
8182 also involved a cleanup of sorts in safestack.h to more correctly
8183 map type-safe stack functions onto their plain stack counterparts.
8184 This work has also resulted in a variety of "const"ifications of
8185 lots of the code, especially "_cmp" operations which should normally
8186 be prototyped with "const" parameters anyway.
8189 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8190 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8191 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8192 where all of 'md' is used each time the PRNG is used, but 'state'
8193 is used only indexed by a cyclic counter. As entropy may not be
8194 well distributed from the beginning, 'md' is important as a
8195 chaining variable. However, the output function chains only half
8196 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8197 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8198 in all of 'state' being rewritten, with the new values depending
8199 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8202 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8203 the handshake is continued after ssl_verify_cert_chain();
8204 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8205 can lead to 'unexplainable' connection aborts later.
8206 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8208 *) Major EVP API cipher revision.
8209 Add hooks for extra EVP features. This allows various cipher
8210 parameters to be set in the EVP interface. Support added for variable
8211 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8212 setting of RC2 and RC5 parameters.
8214 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8217 Remove lots of duplicated code from the EVP library. For example *every*
8218 cipher init() function handles the 'iv' in the same way according to the
8219 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8220 for CFB and OFB modes they zero ctx->num.
8222 New functionality allows removal of S/MIME code RC2 hack.
8224 Most of the routines have the same form and so can be declared in terms
8227 By shifting this to the top level EVP_CipherInit() it can be removed from
8228 all individual ciphers. If the cipher wants to handle IVs or keys
8229 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8232 Change lots of functions like EVP_EncryptUpdate() to now return a
8233 value: although software versions of the algorithms cannot fail
8234 any installed hardware versions can.
8237 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8238 this option is set, tolerate broken clients that send the negotiated
8239 protocol version number instead of the requested protocol version
8243 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8244 i.e. non-zero for export ciphersuites, zero otherwise.
8245 Previous versions had this flag inverted, inconsistent with
8246 rsa_tmp_cb (..._TMP_RSA_CB).
8247 [Bodo Moeller; problem reported by Amit Chopra]
8249 *) Add missing DSA library text string. Work around for some IIS
8250 key files with invalid SEQUENCE encoding.
8253 *) Add a document (doc/standards.txt) that list all kinds of standards
8254 and so on that are implemented in OpenSSL.
8257 *) Enhance c_rehash script. Old version would mishandle certificates
8258 with the same subject name hash and wouldn't handle CRLs at all.
8259 Added -fingerprint option to crl utility, to support new c_rehash
8263 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8266 *) Fix for SSL server purpose checking. Server checking was
8267 rejecting certificates which had extended key usage present
8268 but no ssl client purpose.
8269 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8271 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8272 is a little unclear about how a blank password is handled.
8273 Since the password in encoded as a BMPString with terminating
8274 double NULL a zero length password would end up as just the
8275 double NULL. However no password at all is different and is
8276 handled differently in the PKCS#12 key generation code. NS
8277 treats a blank password as zero length. MSIE treats it as no
8278 password on export: but it will try both on import. We now do
8279 the same: PKCS12_parse() tries zero length and no password if
8280 the password is set to "" or NULL (NULL is now a valid password:
8281 it wasn't before) as does the pkcs12 application.
8284 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8285 perror when PEM_read_bio_X509_REQ fails, the error message must
8286 be obtained from the error queue.
8289 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8290 it in ERR_remove_state if appropriate, and change ERR_get_state
8291 accordingly to avoid race conditions (this is necessary because
8292 thread_hash is no longer constant once set).
8295 *) Bugfix for linux-elf makefile.one.
8298 *) RSA_get_default_method() will now cause a default
8299 RSA_METHOD to be chosen if one doesn't exist already.
8300 Previously this was only set during a call to RSA_new()
8301 or RSA_new_method(NULL) meaning it was possible for
8302 RSA_get_default_method() to return NULL.
8305 *) Added native name translation to the existing DSO code
8306 that will convert (if the flag to do so is set) filenames
8307 that are sufficiently small and have no path information
8308 into a canonical native form. Eg. "blah" converted to
8309 "libblah.so" or "blah.dll" etc.
8312 *) New function ERR_error_string_n(e, buf, len) which is like
8313 ERR_error_string(e, buf), but writes at most 'len' bytes
8314 including the 0 terminator. For ERR_error_string_n, 'buf'
8316 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8318 *) CONF library reworked to become more general. A new CONF
8319 configuration file reader "class" is implemented as well as a
8320 new functions (NCONF_*, for "New CONF") to handle it. The now
8321 old CONF_* functions are still there, but are reimplemented to
8322 work in terms of the new functions. Also, a set of functions
8323 to handle the internal storage of the configuration data is
8324 provided to make it easier to write new configuration file
8325 reader "classes" (I can definitely see something reading a
8326 configuration file in XML format, for example), called _CONF_*,
8327 or "the configuration storage API"...
8329 The new configuration file reading functions are:
8331 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8332 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8334 NCONF_default, NCONF_WIN32
8336 NCONF_dump_fp, NCONF_dump_bio
8338 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8339 NCONF_new creates a new CONF object. This works in the same way
8340 as other interfaces in OpenSSL, like the BIO interface.
8341 NCONF_dump_* dump the internal storage of the configuration file,
8342 which is useful for debugging. All other functions take the same
8343 arguments as the old CONF_* functions wth the exception of the
8344 first that must be a `CONF *' instead of a `LHASH *'.
8346 To make it easer to use the new classes with the old CONF_* functions,
8347 the function CONF_set_default_method is provided.
8350 *) Add '-tls1' option to 'openssl ciphers', which was already
8351 mentioned in the documentation but had not been implemented.
8352 (This option is not yet really useful because even the additional
8353 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8356 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8357 OpenSSL-based applications) load shared libraries and bind to
8358 them in a portable way.
8359 [Geoff Thorpe, with contributions from Richard Levitte]
8361 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8363 *) Make sure _lrotl and _lrotr are only used with MSVC.
8365 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8366 (the default implementation of RAND_status).
8368 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8369 to '-clrext' (= clear extensions), as intended and documented.
8370 [Bodo Moeller; inconsistency pointed out by Michael Attili
8373 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8374 was larger than the MD block size.
8375 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8377 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8378 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8379 using the passed key: if the passed key was a private key the result
8380 of X509_print(), for example, would be to print out all the private key
8384 *) des_quad_cksum() byte order bug fix.
8385 [Ulf Möller, using the problem description in krb4-0.9.7, where
8386 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8388 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8390 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8392 *) For easily testing in shell scripts whether some command
8393 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8394 returns with exit code 0 iff no command of the given name is available.
8395 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8396 the output goes to stdout and nothing is printed to stderr.
8397 Additional arguments are always ignored.
8399 Since for each cipher there is a command of the same name,
8400 the 'no-cipher' compilation switches can be tested this way.
8402 ('openssl no-XXX' is not able to detect pseudo-commands such
8403 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8406 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8409 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8410 is set; it will be thrown away anyway because each handshake creates
8412 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8413 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8414 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
8415 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8418 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8419 'Q' and 'R' lose their special meanings (quit/renegotiate).
8420 This is part of what -quiet does; unlike -quiet, -ign_eof
8421 does not suppress any output.
8424 *) Add compatibility options to the purpose and trust code. The
8425 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8426 accepts a certificate or CA, this was the previous behaviour,
8427 with all the associated security issues.
8429 X509_TRUST_COMPAT is the old trust behaviour: only and
8430 automatically trust self signed roots in certificate store. A
8431 new trust setting X509_TRUST_DEFAULT is used to specify that
8432 a purpose has no associated trust setting and it should instead
8433 use the value in the default purpose.
8436 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8437 and fix a memory leak.
8440 *) In util/mkerr.pl (which implements 'make errors'), preserve
8441 reason strings from the previous version of the .c file, as
8442 the default to have only downcase letters (and digits) in
8443 automatically generated reasons codes is not always appropriate.
8446 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8447 using strerror. Previously, ERR_reason_error_string() returned
8448 library names as reason strings for SYSerr; but SYSerr is a special
8449 case where small numbers are errno values, not library numbers.
8452 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8453 converts DSA parameters into DH parameters. (When creating parameters,
8454 DSA_generate_parameters is used.)
8457 *) Include 'length' (recommended exponent length) in C code generated
8458 by 'openssl dhparam -C'.
8461 *) The second argument to set_label in perlasm was already being used
8462 so couldn't be used as a "file scope" flag. Moved to third argument
8466 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8467 instead of RAND_bytes for encryption IVs and salts.
8470 *) Include RAND_status() into RAND_METHOD instead of implementing
8471 it only for md_rand.c Otherwise replacing the PRNG by calling
8472 RAND_set_rand_method would be impossible.
8475 *) Don't let DSA_generate_key() enter an infinite loop if the random
8476 number generation fails.
8479 *) New 'rand' application for creating pseudo-random output.
8482 *) Added configuration support for Linux/IA64
8483 [Rolf Haberrecker <rolf@suse.de>]
8485 *) Assembler module support for Mingw32.
8488 *) Shared library support for HPUX (in shlib/).
8489 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8491 *) Shared library support for Solaris gcc.
8492 [Lutz Behnke <behnke@trustcenter.de>]
8494 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8496 *) PKCS7_encrypt() was adding text MIME headers twice because they
8497 were added manually and by SMIME_crlf_copy().
8500 *) In bntest.c don't call BN_rand with zero bits argument.
8501 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8503 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8504 case was implemented. This caused BN_div_recp() to fail occasionally.
8507 *) Add an optional second argument to the set_label() in the perl
8508 assembly language builder. If this argument exists and is set
8509 to 1 it signals that the assembler should use a symbol whose
8510 scope is the entire file, not just the current function. This
8511 is needed with MASM which uses the format label:: for this scope.
8512 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8514 *) Change the ASN1 types so they are typedefs by default. Before
8515 almost all types were #define'd to ASN1_STRING which was causing
8516 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8520 *) Change names of new functions to the new get1/get0 naming
8521 convention: After 'get1', the caller owns a reference count
8522 and has to call ..._free; 'get0' returns a pointer to some
8523 data structure without incrementing reference counters.
8524 (Some of the existing 'get' functions increment a reference
8525 counter, some don't.)
8526 Similarly, 'set1' and 'add1' functions increase reference
8527 counters or duplicate objects.
8530 *) Allow for the possibility of temp RSA key generation failure:
8531 the code used to assume it always worked and crashed on failure.
8534 *) Fix potential buffer overrun problem in BIO_printf().
8535 [Ulf Möller, using public domain code by Patrick Powell; problem
8536 pointed out by David Sacerdote <das33@cornell.edu>]
8538 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8539 RAND_egd() and RAND_status(). In the command line application,
8540 the EGD socket can be specified like a seed file using RANDFILE
8544 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8545 Some CAs (e.g. Verisign) distribute certificates in this form.
8548 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8549 list to exclude them. This means that no special compilation option
8550 is needed to use anonymous DH: it just needs to be included in the
8554 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8555 EVP_MD_type. The old functionality is available in a new macro called
8556 EVP_MD_md(). Change code that uses it and update docs.
8559 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8560 where the 'void *' argument is replaced by a function pointer argument.
8561 Previously 'void *' was abused to point to functions, which works on
8562 many platforms, but is not correct. As these functions are usually
8563 called by macros defined in OpenSSL header files, most source code
8564 should work without changes.
8567 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8568 sections with information on -D... compiler switches used for
8569 compiling the library so that applications can see them. To enable
8570 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8571 must be defined. E.g.,
8572 #define OPENSSL_ALGORITHM_DEFINES
8573 #include <openssl/opensslconf.h>
8574 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8575 [Richard Levitte, Ulf and Bodo Möller]
8577 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8581 *) Change the 'other' type in certificate aux info to a STACK_OF
8582 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8583 the required ASN1 format: arbitrary types determined by an OID.
8586 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8587 argument to 'req'. This is not because the function is newer or
8588 better than others it just uses the work 'NEW' in the certificate
8589 request header lines. Some software needs this.
8592 *) Reorganise password command line arguments: now passwords can be
8593 obtained from various sources. Delete the PEM_cb function and make
8594 it the default behaviour: i.e. if the callback is NULL and the
8595 usrdata argument is not NULL interpret it as a null terminated pass
8596 phrase. If usrdata and the callback are NULL then the pass phrase
8597 is prompted for as usual.
8600 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8601 the support is automatically enabled. The resulting binaries will
8602 autodetect the card and use it if present.
8603 [Ben Laurie and Compaq Inc.]
8605 *) Work around for Netscape hang bug. This sends certificate request
8606 and server done in one record. Since this is perfectly legal in the
8607 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8608 the bugs/SSLv3 entry for more info.
8611 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8614 *) Add -rand argument to smime and pkcs12 applications and read/write
8618 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8621 *) Add command line password options to the remaining applications.
8624 *) Bug fix for BN_div_recp() for numerators with an even number of
8628 *) More tests in bntest.c, and changed test_bn output.
8631 *) ./config recognizes MacOS X now.
8634 *) Bug fix for BN_div() when the first words of num and divsor are
8635 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8638 *) Add support for various broken PKCS#8 formats, and command line
8639 options to produce them.
8642 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8643 get temporary BIGNUMs from a BN_CTX.
8646 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8650 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8651 include a #define from the old name to the new. The original intent
8652 was that statically linked binaries could for example just call
8653 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8654 link with digests. This never worked becayse SSLeay_add_all_digests()
8655 and SSLeay_add_all_ciphers() were in the same source file so calling
8656 one would link with the other. They are now in separate source files.
8659 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8662 *) Use a less unusual form of the Miller-Rabin primality test (it used
8663 a binary algorithm for exponentiation integrated into the Miller-Rabin
8664 loop, our standard modexp algorithms are faster).
8667 *) Support for the EBCDIC character set completed.
8668 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8670 *) Source code cleanups: use const where appropriate, eliminate casts,
8671 use void * instead of char * in lhash.
8674 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8675 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8676 this the server could overwrite ephemeral keys that the client
8680 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8681 using 50 iterations of the Rabin-Miller test.
8683 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8684 iterations of the Rabin-Miller test as required by the appendix
8685 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8686 As BN_is_prime_fasttest includes trial division, DSA parameter
8687 generation becomes much faster.
8689 This implies a change for the callback functions in DSA_is_prime
8690 and DSA_generate_parameters: The callback function is called once
8691 for each positive witness in the Rabin-Miller test, not just
8692 occasionally in the inner loop; and the parameters to the
8693 callback function now provide an iteration count for the outer
8694 loop rather than for the current invocation of the inner loop.
8695 DSA_generate_parameters additionally can call the callback
8696 function with an 'iteration count' of -1, meaning that a
8697 candidate has passed the trial division test (when q is generated
8698 from an application-provided seed, trial division is skipped).
8701 *) New function BN_is_prime_fasttest that optionally does trial
8702 division before starting the Rabin-Miller test and has
8703 an additional BN_CTX * argument (whereas BN_is_prime always
8704 has to allocate at least one BN_CTX).
8705 'callback(1, -1, cb_arg)' is called when a number has passed the
8706 trial division stage.
8709 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8713 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8716 *) New function BN_pseudo_rand().
8719 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8720 bignum version of BN_from_montgomery() with the working code from
8721 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8725 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8726 made it impossible to use the same SSL_SESSION data structure in
8727 SSL2 clients in multiple threads.
8730 *) The return value of RAND_load_file() no longer counts bytes obtained
8731 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8732 to seed the PRNG (previously an explicit byte count was required).
8733 [Ulf Möller, Bodo Möller]
8735 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8736 used (char *) instead of (void *) and had casts all over the place.
8739 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8742 *) Retain source code compatibility for BN_prime_checks macro:
8743 BN_is_prime(..., BN_prime_checks, ...) now uses
8744 BN_prime_checks_for_size to determine the appropriate number of
8745 Rabin-Miller iterations.
8748 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8749 DH_CHECK_P_NOT_SAFE_PRIME.
8750 (Check if this is true? OpenPGP calls them "strong".)
8753 *) Merge the functionality of "dh" and "gendh" programs into a new program
8754 "dhparam". The old programs are retained for now but will handle DH keys
8755 (instead of parameters) in future.
8758 *) Make the ciphers, s_server and s_client programs check the return values
8759 when a new cipher list is set.
8762 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8763 ciphers. Before when the 56bit ciphers were enabled the sorting was
8766 The syntax for the cipher sorting has been extended to support sorting by
8767 cipher-strength (using the strength_bits hard coded in the tables).
8768 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8770 Fix a bug in the cipher-command parser: when supplying a cipher command
8771 string with an "undefined" symbol (neither command nor alphanumeric
8772 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8773 an error is flagged.
8775 Due to the strength-sorting extension, the code of the
8776 ssl_create_cipher_list() function was completely rearranged. I hope that
8777 the readability was also increased :-)
8778 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8780 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8781 for the first serial number and places 2 in the serial number file. This
8782 avoids problems when the root CA is created with serial number zero and
8783 the first user certificate has the same issuer name and serial number
8787 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8788 the new code. Add documentation for this stuff.
8791 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8792 X509_*() to X509at_*() on the grounds that they don't handle X509
8793 structures and behave in an analagous way to the X509v3 functions:
8794 they shouldn't be called directly but wrapper functions should be used
8797 So we also now have some wrapper functions that call the X509at functions
8798 when passed certificate requests. (TO DO: similar things can be done with
8799 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8800 things. Some of these need some d2i or i2d and print functionality
8801 because they handle more complex structures.)
8804 *) Add missing #ifndefs that caused missing symbols when building libssl
8805 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8806 NO_RSA in ssl/s2*.c.
8807 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8809 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8810 has a return value which indicates the quality of the random data
8811 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8812 error queue. New function RAND_pseudo_bytes() generates output that is
8813 guaranteed to be unique but not unpredictable. RAND_add is like
8814 RAND_seed, but takes an extra argument for an entropy estimate
8815 (RAND_seed always assumes full entropy).
8818 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8819 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8820 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8821 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8822 false-positive rate of at most 2^-80 for random input.
8825 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8828 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8829 in the 0.9.5 release), this returns the chain
8830 from an X509_CTX structure with a dup of the stack and all
8831 the X509 reference counts upped: so the stack will exist
8832 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8835 Also make SSL_SESSION_print() print out the verify return
8839 *) Add manpage for the pkcs12 command. Also change the default
8840 behaviour so MAC iteration counts are used unless the new
8841 -nomaciter option is used. This improves file security and
8842 only older versions of MSIE (4.0 for example) need it.
8845 *) Honor the no-xxx Configure options when creating .DEF files.
8848 *) Add PKCS#10 attributes to field table: challengePassword,
8849 unstructuredName and unstructuredAddress. These are taken from
8850 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8851 international characters are used.
8853 More changes to X509_ATTRIBUTE code: allow the setting of types
8854 based on strings. Remove the 'loc' parameter when adding
8855 attributes because these will be a SET OF encoding which is sorted
8859 *) Initial changes to the 'req' utility to allow request generation
8860 automation. This will allow an application to just generate a template
8861 file containing all the field values and have req construct the
8864 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8865 used all over the place including certificate requests and PKCS#7
8866 structures. They are currently handled manually where necessary with
8867 some primitive wrappers for PKCS#7. The new functions behave in a
8868 manner analogous to the X509 extension functions: they allow
8869 attributes to be looked up by NID and added.
8871 Later something similar to the X509V3 code would be desirable to
8872 automatically handle the encoding, decoding and printing of the
8873 more complex types. The string types like challengePassword can
8874 be handled by the string table functions.
8876 Also modified the multi byte string table handling. Now there is
8877 a 'global mask' which masks out certain types. The table itself
8878 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8879 is useful when for example there is only one permissible type
8880 (as in countryName) and using the mask might result in no valid
8884 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8885 SSL_get_peer_finished to allow applications to obtain the latest
8886 Finished messages sent to the peer or expected from the peer,
8887 respectively. (SSL_get_peer_finished is usually the Finished message
8888 actually received from the peer, otherwise the protocol will be aborted.)
8890 As the Finished message are message digests of the complete handshake
8891 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8892 be used for external authentication procedures when the authentication
8893 provided by SSL/TLS is not desired or is not enough.
8896 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8897 the host supports BWX extension and if Compaq C is present on the
8898 $PATH. Just exploiting of the BWX extension results in 20-30%
8899 performance kick for some algorithms, e.g. DES and RC4 to mention
8900 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8904 *) Add support for MS "fast SGC". This is arguably a violation of the
8905 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8906 weak crypto and after checking the certificate is SGC a second one
8907 with strong crypto. MS SGC stops the first handshake after receiving
8908 the server certificate message and sends a second client hello. Since
8909 a server will typically do all the time consuming operations before
8910 expecting any further messages from the client (server key exchange
8911 is the most expensive) there is little difference between the two.
8913 To get OpenSSL to support MS SGC we have to permit a second client
8914 hello message after we have sent server done. In addition we have to
8915 reset the MAC if we do get this second client hello.
8918 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8919 if a DER encoded private key is RSA or DSA traditional format. Changed
8920 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8921 format DER encoded private key. Newer code should use PKCS#8 format which
8922 has the key type encoded in the ASN1 structure. Added DER private key
8923 support to pkcs8 application.
8926 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8927 ciphersuites has been selected (as required by the SSL 3/TLS 1
8928 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8929 is set, we interpret this as a request to violate the specification
8930 (the worst that can happen is a handshake failure, and 'correct'
8931 behaviour would result in a handshake failure anyway).
8934 *) In SSL_CTX_add_session, take into account that there might be multiple
8935 SSL_SESSION structures with the same session ID (e.g. when two threads
8936 concurrently obtain them from an external cache).
8937 The internal cache can handle only one SSL_SESSION with a given ID,
8938 so if there's a conflict, we now throw out the old one to achieve
8942 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8943 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8944 some routines that use cipher OIDs: some ciphers do not have OIDs
8945 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8949 *) Simplify the trust setting structure and code. Now we just have
8950 two sequences of OIDs for trusted and rejected settings. These will
8951 typically have values the same as the extended key usage extension
8952 and any application specific purposes.
8954 The trust checking code now has a default behaviour: it will just
8955 check for an object with the same NID as the passed id. Functions can
8956 be provided to override either the default behaviour or the behaviour
8957 for a given id. SSL client, server and email already have functions
8958 in place for compatibility: they check the NID and also return "trusted"
8959 if the certificate is self signed.
8962 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8963 traditional format into an EVP_PKEY structure.
8966 *) Add a password callback function PEM_cb() which either prompts for
8967 a password if usr_data is NULL or otherwise assumes it is a null
8968 terminated password. Allow passwords to be passed on command line
8969 environment or config files in a few more utilities.
8972 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8973 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8974 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8975 Update documentation.
8978 *) Support for ASN1 "NULL" type. This could be handled before by using
8979 ASN1_TYPE but there wasn't any function that would try to read a NULL
8980 and produce an error if it couldn't. For compatibility we also have
8981 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8982 don't allocate anything because they don't need to.
8985 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8987 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8989 *) Rebuild of the memory allocation routines used by OpenSSL code and
8990 possibly others as well. The purpose is to make an interface that
8991 provide hooks so anyone can build a separate set of allocation and
8992 deallocation routines to be used by OpenSSL, for example memory
8993 pool implementations, or something else, which was previously hard
8994 since Malloc(), Realloc() and Free() were defined as macros having
8995 the values malloc, realloc and free, respectively (except for Win32
8996 compilations). The same is provided for memory debugging code.
8997 OpenSSL already comes with functionality to find memory leaks, but
8998 this gives people a chance to debug other memory problems.
9000 With these changes, a new set of functions and macros have appeared:
9002 CRYPTO_set_mem_debug_functions() [F]
9003 CRYPTO_get_mem_debug_functions() [F]
9004 CRYPTO_dbg_set_options() [F]
9005 CRYPTO_dbg_get_options() [F]
9006 CRYPTO_malloc_debug_init() [M]
9008 The memory debug functions are NULL by default, unless the library
9009 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9010 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9011 gives the standard debugging functions that come with OpenSSL) or
9012 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9013 provided by the library user) must be used. When the standard
9014 debugging functions are used, CRYPTO_dbg_set_options can be used to
9015 request additional information:
9016 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9017 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9019 Also, things like CRYPTO_set_mem_functions will always give the
9020 expected result (the new set of functions is used for allocation
9021 and deallocation) at all times, regardless of platform and compiler
9024 To finish it up, some functions that were never use in any other
9025 way than through macros have a new API and new semantic:
9028 CRYPTO_dbg_realloc()
9031 All macros of value have retained their old syntax.
9032 [Richard Levitte and Bodo Moeller]
9034 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9035 ordering of SMIMECapabilities wasn't in "strength order" and there
9036 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9040 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9041 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9042 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9044 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9045 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9046 functionality to handle multipart/signed properly) and a utility
9047 called 'smime' to call all this stuff. This is based on code I
9048 originally wrote for Celo who have kindly allowed it to be
9049 included in OpenSSL.
9052 *) Add variants des_set_key_checked and des_set_key_unchecked of
9053 des_set_key (aka des_key_sched). Global variable des_check_key
9054 decides which of these is called by des_set_key; this way
9055 des_check_key behaves as it always did, but applications and
9056 the library itself, which was buggy for des_check_key == 1,
9057 have a cleaner way to pick the version they need.
9060 *) New function PKCS12_newpass() which changes the password of a
9064 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9065 dynamic mix. In both cases the ids can be used as an index into the
9066 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9067 functions so they accept a list of the field values and the
9068 application doesn't need to directly manipulate the X509_TRUST
9072 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9076 *) Modify the way the V3 extension code looks up extensions. This now
9077 works in a similar way to the object code: we have some "standard"
9078 extensions in a static table which is searched with OBJ_bsearch()
9079 and the application can add dynamic ones if needed. The file
9080 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9081 updated whenever a new extension is added to the core code and kept
9082 in ext_nid order. There is a simple program 'tabtest.c' which checks
9083 this. New extensions are not added too often so this file can readily
9084 be maintained manually.
9086 There are two big advantages in doing things this way. The extensions
9087 can be looked up immediately and no longer need to be "added" using
9088 X509V3_add_standard_extensions(): this function now does nothing.
9089 [Side note: I get *lots* of email saying the extension code doesn't
9090 work because people forget to call this function]
9091 Also no dynamic allocation is done unless new extensions are added:
9092 so if we don't add custom extensions there is no need to call
9093 X509V3_EXT_cleanup().
9096 *) Modify enc utility's salting as follows: make salting the default. Add a
9097 magic header, so unsalted files fail gracefully instead of just decrypting
9098 to garbage. This is because not salting is a big security hole, so people
9099 should be discouraged from doing it.
9102 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9103 digest to be passed on the command line but it only used this
9104 parameter when signing a certificate. Modified so all relevant
9105 operations are affected by the digest parameter including the
9106 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9107 DSA key was used because it didn't fix the digest.
9110 *) Initial certificate chain verify code. Currently tests the untrusted
9111 certificates for consistency with the verify purpose (which is set
9112 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9114 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9115 this is because it will reject chains with invalid extensions whereas
9116 every previous version of OpenSSL and SSLeay made no checks at all.
9118 Trust code: checks the root CA for the relevant trust settings. Trust
9119 settings have an initial value consistent with the verify purpose: e.g.
9120 if the verify purpose is for SSL client use it expects the CA to be
9121 trusted for SSL client use. However the default value can be changed to
9122 permit custom trust settings: one example of this would be to only trust
9123 certificates from a specific "secure" set of CAs.
9125 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9126 which should be used for version portability: especially since the
9127 verify structure is likely to change more often now.
9129 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9130 to set them. If not set then assume SSL clients will verify SSL servers
9133 Two new options to the verify program: -untrusted allows a set of
9134 untrusted certificates to be passed in and -purpose which sets the
9135 intended purpose of the certificate. If a purpose is set then the
9136 new chain verify code is used to check extension consistency.
9139 *) Support for the authority information access extension.
9142 *) Modify RSA and DSA PEM read routines to transparently handle
9143 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9144 public keys in a format compatible with certificate
9145 SubjectPublicKeyInfo structures. Unfortunately there were already
9146 functions called *_PublicKey_* which used various odd formats so
9147 these are retained for compatibility: however the DSA variants were
9148 never in a public release so they have been deleted. Changed dsa/rsa
9149 utilities to handle the new format: note no releases ever handled public
9150 keys so we should be OK.
9152 The primary motivation for this change is to avoid the same fiasco
9153 that dogs private keys: there are several incompatible private key
9154 formats some of which are standard and some OpenSSL specific and
9155 require various evil hacks to allow partial transparent handling and
9156 even then it doesn't work with DER formats. Given the option anything
9157 other than PKCS#8 should be dumped: but the other formats have to
9158 stay in the name of compatibility.
9160 With public keys and the benefit of hindsight one standard format
9161 is used which works with EVP_PKEY, RSA or DSA structures: though
9162 it clearly returns an error if you try to read the wrong kind of key.
9164 Added a -pubkey option to the 'x509' utility to output the public key.
9165 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9166 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9167 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9168 that do the same as the EVP_PKEY_assign_*() except they up the
9169 reference count of the added key (they don't "swallow" the
9173 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9174 CRLs would fail if the file contained no certificates or no CRLs:
9175 added a new function to read in both types and return the number
9176 read: this means that if none are read it will be an error. The
9177 DER versions of the certificate and CRL reader would always fail
9178 because it isn't possible to mix certificates and CRLs in DER format
9179 without choking one or the other routine. Changed this to just read
9180 a certificate: this is the best we can do. Also modified the code
9181 in apps/verify.c to take notice of return codes: it was previously
9182 attempting to read in certificates from NULL pointers and ignoring
9183 any errors: this is one reason why the cert and CRL reader seemed
9184 to work. It doesn't check return codes from the default certificate
9185 routines: these may well fail if the certificates aren't installed.
9188 *) Code to support otherName option in GeneralName.
9191 *) First update to verify code. Change the verify utility
9192 so it warns if it is passed a self signed certificate:
9193 for consistency with the normal behaviour. X509_verify
9194 has been modified to it will now verify a self signed
9195 certificate if *exactly* the same certificate appears
9196 in the store: it was previously impossible to trust a
9197 single self signed certificate. This means that:
9198 openssl verify ss.pem
9199 now gives a warning about a self signed certificate but
9200 openssl verify -CAfile ss.pem ss.pem
9204 *) For servers, store verify_result in SSL_SESSION data structure
9205 (and add it to external session representation).
9206 This is needed when client certificate verifications fails,
9207 but an application-provided verification callback (set by
9208 SSL_CTX_set_cert_verify_callback) allows accepting the session
9209 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9210 but returns 1): When the session is reused, we have to set
9211 ssl->verify_result to the appropriate error code to avoid
9213 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9215 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9216 case in PKCS7_dataInit() where the signed PKCS7 structure
9217 didn't contain any existing data because it was being created.
9218 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9220 *) Add a salt to the key derivation routines in enc.c. This
9221 forms the first 8 bytes of the encrypted file. Also add a
9222 -S option to allow a salt to be input on the command line.
9225 *) New function X509_cmp(). Oddly enough there wasn't a function
9226 to compare two certificates. We do this by working out the SHA1
9227 hash and comparing that. X509_cmp() will be needed by the trust
9231 *) SSL_get1_session() is like SSL_get_session(), but increments
9232 the reference count in the SSL_SESSION returned.
9233 [Geoff Thorpe <geoff@eu.c2.net>]
9235 *) Fix for 'req': it was adding a null to request attributes.
9236 Also change the X509_LOOKUP and X509_INFO code to handle
9237 certificate auxiliary information.
9240 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9244 *) Add the possibility to add extra information to the memory leak
9245 detecting output, to form tracebacks, showing from where each
9246 allocation was originated: CRYPTO_push_info("constant string") adds
9247 the string plus current file name and line number to a per-thread
9248 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9249 is like calling CYRPTO_pop_info() until the stack is empty.
9250 Also updated memory leak detection code to be multi-thread-safe.
9253 *) Add options -text and -noout to pkcs7 utility and delete the
9254 encryption options which never did anything. Update docs.
9257 *) Add options to some of the utilities to allow the pass phrase
9258 to be included on either the command line (not recommended on
9259 OSes like Unix) or read from the environment. Update the
9260 manpages and fix a few bugs.
9263 *) Add a few manpages for some of the openssl commands.
9266 *) Fix the -revoke option in ca. It was freeing up memory twice,
9267 leaking and not finding already revoked certificates.
9270 *) Extensive changes to support certificate auxiliary information.
9271 This involves the use of X509_CERT_AUX structure and X509_AUX
9272 functions. An X509_AUX function such as PEM_read_X509_AUX()
9273 can still read in a certificate file in the usual way but it
9274 will also read in any additional "auxiliary information". By
9275 doing things this way a fair degree of compatibility can be
9276 retained: existing certificates can have this information added
9277 using the new 'x509' options.
9279 Current auxiliary information includes an "alias" and some trust
9280 settings. The trust settings will ultimately be used in enhanced
9281 certificate chain verification routines: currently a certificate
9282 can only be trusted if it is self signed and then it is trusted
9286 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9287 The problem was that one of the replacement routines had not been working
9288 since SSLeay releases. For now the offending routine has been replaced
9289 with non-optimised assembler. Even so, this now gives around 95%
9290 performance improvement for 1024 bit RSA signs.
9293 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9294 handling. Most clients have the effective key size in bits equal to
9295 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9296 A few however don't do this and instead use the size of the decrypted key
9297 to determine the RC2 key length and the AlgorithmIdentifier to determine
9298 the effective key length. In this case the effective key length can still
9299 be 40 bits but the key length can be 168 bits for example. This is fixed
9300 by manually forcing an RC2 key into the EVP_PKEY structure because the
9301 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9302 the key length and effective key length are equal.
9305 *) Add a bunch of functions that should simplify the creation of
9306 X509_NAME structures. Now you should be able to do:
9307 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9308 and have it automatically work out the correct field type and fill in
9309 the structures. The more adventurous can try:
9310 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9311 and it will (hopefully) work out the correct multibyte encoding.
9314 *) Change the 'req' utility to use the new field handling and multibyte
9315 copy routines. Before the DN field creation was handled in an ad hoc
9316 way in req, ca, and x509 which was rather broken and didn't support
9317 BMPStrings or UTF8Strings. Since some software doesn't implement
9318 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9319 using the dirstring_type option. See the new comment in the default
9320 openssl.cnf for more info.
9323 *) Make crypto/rand/md_rand.c more robust:
9324 - Assure unique random numbers after fork().
9325 - Make sure that concurrent threads access the global counter and
9326 md serializably so that we never lose entropy in them
9327 or use exactly the same state in multiple threads.
9328 Access to the large state is not always serializable because
9329 the additional locking could be a performance killer, and
9330 md should be large enough anyway.
9333 *) New file apps/app_rand.c with commonly needed functionality
9334 for handling the random seed file.
9336 Use the random seed file in some applications that previously did not:
9338 dsaparam -genkey (which also ignored its '-rand' option),
9341 x509 (when signing).
9342 Except on systems with /dev/urandom, it is crucial to have a random
9343 seed file at least for key creation, DSA signing, and for DH exchanges;
9344 for RSA signatures we could do without one.
9346 gendh and gendsa (unlike genrsa) used to read only the first byte
9347 of each file listed in the '-rand' option. The function as previously
9348 found in genrsa is now in app_rand.c and is used by all programs
9349 that support '-rand'.
9352 *) In RAND_write_file, use mode 0600 for creating files;
9353 don't just chmod when it may be too late.
9356 *) Report an error from X509_STORE_load_locations
9357 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9360 *) New function ASN1_mbstring_copy() this copies a string in either
9361 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9362 into an ASN1_STRING type. A mask of permissible types is passed
9363 and it chooses the "minimal" type to use or an error if not type
9367 *) Add function equivalents to the various macros in asn1.h. The old
9368 macros are retained with an M_ prefix. Code inside the library can
9369 use the M_ macros. External code (including the openssl utility)
9370 should *NOT* in order to be "shared library friendly".
9373 *) Add various functions that can check a certificate's extensions
9374 to see if it usable for various purposes such as SSL client,
9375 server or S/MIME and CAs of these types. This is currently
9376 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9377 verification. Also added a -purpose flag to x509 utility to
9378 print out all the purposes.
9381 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9385 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9386 for, obtain and decode and extension and obtain its critical flag.
9387 This allows all the necessary extension code to be handled in a
9388 single function call.
9391 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9392 platforms. See crypto/rc4/rc4_enc.c for further details.
9395 *) New -noout option to asn1parse. This causes no output to be produced
9396 its main use is when combined with -strparse and -out to extract data
9397 from a file (which may not be in ASN.1 format).
9400 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9401 when producing the local key id.
9402 [Richard Levitte <levitte@stacken.kth.se>]
9404 *) New option -dhparam in s_server. This allows a DH parameter file to be
9405 stated explicitly. If it is not stated then it tries the first server
9406 certificate file. The previous behaviour hard coded the filename
9410 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9411 a public key to be input or output. For example:
9412 openssl rsa -in key.pem -pubout -out pubkey.pem
9413 Also added necessary DSA public key functions to handle this.
9416 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9417 in the message. This was handled by allowing
9418 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9419 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9421 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9422 to the end of the strings whereas this didn't. This would cause problems
9423 if strings read with d2i_ASN1_bytes() were later modified.
9424 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9426 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9427 data and it contains EOF it will end up returning an error. This is
9428 caused by input 46 bytes long. The cause is due to the way base64
9429 BIOs find the start of base64 encoded data. They do this by trying a
9430 trial decode on each line until they find one that works. When they
9431 do a flag is set and it starts again knowing it can pass all the
9432 data directly through the decoder. Unfortunately it doesn't reset
9433 the context it uses. This means that if EOF is reached an attempt
9434 is made to pass two EOFs through the context and this causes the
9435 resulting error. This can also cause other problems as well. As is
9436 usual with these problems it takes *ages* to find and the fix is
9437 trivial: move one line.
9438 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9440 *) Ugly workaround to get s_client and s_server working under Windows. The
9441 old code wouldn't work because it needed to select() on sockets and the
9442 tty (for keypresses and to see if data could be written). Win32 only
9443 supports select() on sockets so we select() with a 1s timeout on the
9444 sockets and then see if any characters are waiting to be read, if none
9445 are present then we retry, we also assume we can always write data to
9446 the tty. This isn't nice because the code then blocks until we've
9447 received a complete line of data and it is effectively polling the
9448 keyboard at 1s intervals: however it's quite a bit better than not
9449 working at all :-) A dedicated Windows application might handle this
9450 with an event loop for example.
9453 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9454 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9455 will be called when RSA_sign() and RSA_verify() are used. This is useful
9456 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9457 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9458 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9459 This necessitated the support of an extra signature type NID_md5_sha1
9460 for SSL signatures and modifications to the SSL library to use it instead
9461 of calling RSA_public_decrypt() and RSA_private_encrypt().
9464 *) Add new -verify -CAfile and -CApath options to the crl program, these
9465 will lookup a CRL issuers certificate and verify the signature in a
9466 similar way to the verify program. Tidy up the crl program so it
9467 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9468 less strict. It will now permit CRL extensions even if it is not
9469 a V2 CRL: this will allow it to tolerate some broken CRLs.
9472 *) Initialize all non-automatic variables each time one of the openssl
9473 sub-programs is started (this is necessary as they may be started
9474 multiple times from the "OpenSSL>" prompt).
9475 [Lennart Bang, Bodo Moeller]
9477 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9478 removing all other RSA functionality (this is what NO_RSA does). This
9479 is so (for example) those in the US can disable those operations covered
9480 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9484 *) Non-copying interface to BIO pairs.
9485 (still largely untested)
9488 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9489 ASCII string. This was handled independently in various places before.
9492 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9493 UTF8 strings a character at a time.
9496 *) Use client_version from client hello to select the protocol
9497 (s23_srvr.c) and for RSA client key exchange verification
9498 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9501 *) Add various utility functions to handle SPKACs, these were previously
9502 handled by poking round in the structure internals. Added new function
9503 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9504 print, verify and generate SPKACs. Based on an original idea from
9505 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9508 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9511 *) Allow the config file extension section to be overwritten on the
9512 command line. Based on an original idea from Massimiliano Pala
9513 <madwolf@comune.modena.it>. The new option is called -extensions
9514 and can be applied to ca, req and x509. Also -reqexts to override
9515 the request extensions in req and -crlexts to override the crl extensions
9519 *) Add new feature to the SPKAC handling in ca. Now you can include
9520 the same field multiple times by preceding it by "XXXX." for example:
9523 this is the same syntax as used in the req config file.
9526 *) Allow certificate extensions to be added to certificate requests. These
9527 are specified in a 'req_extensions' option of the req section of the
9528 config file. They can be printed out with the -text option to req but
9529 are otherwise ignored at present.
9532 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9533 data read consists of only the final block it would not decrypted because
9534 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9535 A misplaced 'break' also meant the decrypted final block might not be
9536 copied until the next read.
9539 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9540 a few extra parameters to the DH structure: these will be useful if
9541 for example we want the value of 'q' or implement X9.42 DH.
9544 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9545 provides hooks that allow the default DSA functions or functions on a
9546 "per key" basis to be replaced. This allows hardware acceleration and
9547 hardware key storage to be handled without major modification to the
9548 library. Also added low level modexp hooks and CRYPTO_EX structure and
9549 associated functions.
9552 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9553 as "read only": it can't be written to and the buffer it points to will
9554 not be freed. Reading from a read only BIO is much more efficient than
9555 a normal memory BIO. This was added because there are several times when
9556 an area of memory needs to be read from a BIO. The previous method was
9557 to create a memory BIO and write the data to it, this results in two
9558 copies of the data and an O(n^2) reading algorithm. There is a new
9559 function BIO_new_mem_buf() which creates a read only memory BIO from
9560 an area of memory. Also modified the PKCS#7 routines to use read only
9564 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9565 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9566 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9567 but a retry condition occured while trying to read the rest.
9570 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9571 NID_pkcs7_encrypted by default: this was wrong since this should almost
9572 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9573 the encrypted data type: this is a more sensible place to put it and it
9574 allows the PKCS#12 code to be tidied up that duplicated this
9578 *) Changed obj_dat.pl script so it takes its input and output files on
9579 the command line. This should avoid shell escape redirection problems
9583 *) Initial support for certificate extension requests, these are included
9584 in things like Xenroll certificate requests. Included functions to allow
9585 extensions to be obtained and added.
9588 *) -crlf option to s_client and s_server for sending newlines as
9589 CRLF (as required by many protocols).
9592 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9594 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9595 [Ralf S. Engelschall]
9597 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9598 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9600 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9604 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9605 DH parameters/keys (q is lost during that conversion, but the resulting
9606 DH parameters contain its length).
9608 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9609 much faster than DH_generate_parameters (which creates parameters
9610 where p = 2*q + 1), and also the smaller q makes DH computations
9611 much more efficient (160-bit exponentiation instead of 1024-bit
9612 exponentiation); so this provides a convenient way to support DHE
9613 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9614 utter importance to use
9615 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9617 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9618 when such DH parameters are used, because otherwise small subgroup
9619 attacks may become possible!
9622 *) Avoid memory leak in i2d_DHparams.
9625 *) Allow the -k option to be used more than once in the enc program:
9626 this allows the same encrypted message to be read by multiple recipients.
9629 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9630 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9631 it will always use the numerical form of the OID, even if it has a short
9635 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9636 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9637 otherwise bn_mod_exp was called. In the case of hardware keys for example
9638 no private key components need be present and it might store extra data
9639 in the RSA structure, which cannot be accessed from bn_mod_exp.
9640 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9641 private key operations.
9644 *) Added support for SPARC Linux.
9647 *) pem_password_cb function type incompatibly changed from
9648 typedef int pem_password_cb(char *buf, int size, int rwflag);
9650 ....(char *buf, int size, int rwflag, void *userdata);
9651 so that applications can pass data to their callbacks:
9652 The PEM[_ASN1]_{read,write}... functions and macros now take an
9653 additional void * argument, which is just handed through whenever
9654 the password callback is called.
9655 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9657 New function SSL_CTX_set_default_passwd_cb_userdata.
9659 Compatibility note: As many C implementations push function arguments
9660 onto the stack in reverse order, the new library version is likely to
9661 interoperate with programs that have been compiled with the old
9662 pem_password_cb definition (PEM_whatever takes some data that
9663 happens to be on the stack as its last argument, and the callback
9664 just ignores this garbage); but there is no guarantee whatsoever that
9667 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9668 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9669 problems not only on Windows, but also on some Unix platforms.
9670 To avoid problematic command lines, these definitions are now in an
9671 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9672 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9675 *) MIPS III/IV assembler module is reimplemented.
9678 *) More DES library cleanups: remove references to srand/rand and
9679 delete an unused file.
9682 *) Add support for the the free Netwide assembler (NASM) under Win32,
9683 since not many people have MASM (ml) and it can be hard to obtain.
9684 This is currently experimental but it seems to work OK and pass all
9685 the tests. Check out INSTALL.W32 for info.
9688 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9689 without temporary keys kept an extra copy of the server key,
9690 and connections with temporary keys did not free everything in case
9694 *) New function RSA_check_key and new openssl rsa option -check
9695 for verifying the consistency of RSA keys.
9696 [Ulf Moeller, Bodo Moeller]
9698 *) Various changes to make Win32 compile work:
9699 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9700 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9701 comparison" warnings.
9702 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9705 *) Add a debugging option to PKCS#5 v2 key generation function: when
9706 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9707 derived keys are printed to stderr.
9710 *) Copy the flags in ASN1_STRING_dup().
9711 [Roman E. Pavlov <pre@mo.msk.ru>]
9713 *) The x509 application mishandled signing requests containing DSA
9714 keys when the signing key was also DSA and the parameters didn't match.
9716 It was supposed to omit the parameters when they matched the signing key:
9717 the verifying software was then supposed to automatically use the CA's
9718 parameters if they were absent from the end user certificate.
9720 Omitting parameters is no longer recommended. The test was also
9721 the wrong way round! This was probably due to unusual behaviour in
9722 EVP_cmp_parameters() which returns 1 if the parameters match.
9723 This meant that parameters were omitted when they *didn't* match and
9724 the certificate was useless. Certificates signed with 'ca' didn't have
9726 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9728 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9729 The interface is as follows:
9730 Applications can use
9731 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9732 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9733 "off" is now the default.
9734 The library internally uses
9735 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9736 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9737 to disable memory-checking temporarily.
9739 Some inconsistent states that previously were possible (and were
9740 even the default) are now avoided.
9742 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9743 with each memory chunk allocated; this is occasionally more helpful
9744 than just having a counter.
9746 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9748 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9752 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9753 which largely parallels "options", but is for changing API behaviour,
9754 whereas "options" are about protocol behaviour.
9755 Initial "mode" flags are:
9757 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9758 a single record has been written.
9759 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9760 retries use the same buffer location.
9761 (But all of the contents must be
9765 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9768 *) Fix problems with no-hmac etc.
9769 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9771 *) New functions RSA_get_default_method(), RSA_set_method() and
9772 RSA_get_method(). These allows replacement of RSA_METHODs without having
9773 to mess around with the internals of an RSA structure.
9776 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9777 Also really enable memory leak checks in openssl.c and in some
9779 [Chad C. Mulligan, Bodo Moeller]
9781 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9782 up the length of negative integers. This has now been simplified to just
9783 store the length when it is first determined and use it later, rather
9784 than trying to keep track of where data is copied and updating it to
9786 [Steve Henson, reported by Brien Wheeler
9787 <bwheeler@authentica-security.com>]
9789 *) Add a new function PKCS7_signatureVerify. This allows the verification
9790 of a PKCS#7 signature but with the signing certificate passed to the
9791 function itself. This contrasts with PKCS7_dataVerify which assumes the
9792 certificate is present in the PKCS#7 structure. This isn't always the
9793 case: certificates can be omitted from a PKCS#7 structure and be
9794 distributed by "out of band" means (such as a certificate database).
9797 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9798 function prototypes in pem.h, also change util/mkdef.pl to add the
9799 necessary function names.
9802 *) mk1mf.pl (used by Windows builds) did not properly read the
9803 options set by Configure in the top level Makefile, and Configure
9804 was not even able to write more than one option correctly.
9805 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9808 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9809 file to be loaded from a BIO or FILE pointer. The BIO version will
9810 for example allow memory BIOs to contain config info.
9813 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9814 Whoever hopes to achieve shared-library compatibility across versions
9815 must use this, not the compile-time macro.
9816 (Exercise 0.9.4: Which is the minimum library version required by
9818 Note: All this applies only to multi-threaded programs, others don't
9822 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9823 through a BIO pair triggered the default case, i.e.
9824 SSLerr(...,SSL_R_UNKNOWN_STATE).
9827 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9828 can use the SSL library even if none of the specific BIOs is
9832 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9833 for the encoded length.
9834 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9836 *) Add initial documentation of the X509V3 functions.
9839 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9840 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9841 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9842 secure PKCS#8 private key format with a high iteration count.
9845 *) Fix determination of Perl interpreter: A perl or perl5
9846 _directory_ in $PATH was also accepted as the interpreter.
9847 [Ralf S. Engelschall]
9849 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9850 wrong with it but it was very old and did things like calling
9851 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9855 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9856 to use the new extension code.
9859 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9860 with macros. This should make it easier to change their form, add extra
9861 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9865 *) Add to configuration table a new entry that can specify an alternative
9866 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9867 according to Mark Crispin <MRC@Panda.COM>.
9871 *) DES CBC did not update the IV. Weird.
9874 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9875 Changing the behaviour of the former might break existing programs --
9876 where IV updating is needed, des_ncbc_encrypt can be used.
9879 *) When bntest is run from "make test" it drives bc to check its
9880 calculations, as well as internally checking them. If an internal check
9881 fails, it needs to cause bc to give a non-zero result or make test carries
9882 on without noticing the failure. Fixed.
9885 *) DES library cleanups.
9888 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9889 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9890 ciphers. NOTE: although the key derivation function has been verified
9891 against some published test vectors it has not been extensively tested
9892 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9896 *) Instead of "mkdir -p", which is not fully portable, use new
9897 Perl script "util/mkdir-p.pl".
9900 *) Rewrite the way password based encryption (PBE) is handled. It used to
9901 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9902 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9903 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9904 the 'parameter' field of the AlgorithmIdentifier is passed to the
9905 underlying key generation function so it must do its own ASN1 parsing.
9906 This has also changed the EVP_PBE_CipherInit() function which now has a
9907 'parameter' argument instead of literal salt and iteration count values
9908 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9911 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9912 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9913 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9914 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9915 value was just used as a "magic string" and not used directly its
9916 value doesn't matter.
9919 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9923 *) "linux-sparc64" configuration (ultrapenguin).
9924 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9925 "linux-sparc" configuration.
9926 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9928 *) config now generates no-xxx options for missing ciphers.
9931 *) Support the EBCDIC character set (work in progress).
9932 File ebcdic.c not yet included because it has a different license.
9933 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9935 *) Support BS2000/OSD-POSIX.
9936 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9938 *) Make callbacks for key generation use void * instead of char *.
9941 *) Make S/MIME samples compile (not yet tested).
9944 *) Additional typesafe stacks.
9947 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9951 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9953 *) New configuration variant "sco5-gcc".
9955 *) Updated some demos.
9956 [Sean O Riordain, Wade Scholine]
9958 *) Add missing BIO_free at exit of pkcs12 application.
9961 *) Fix memory leak in conf.c.
9964 *) Updates for Win32 to assembler version of MD5.
9967 *) Set #! path to perl in apps/der_chop to where we found it
9968 instead of using a fixed path.
9971 *) SHA library changes for irix64-mips4-cc.
9974 *) Improvements for VMS support.
9978 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9980 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9981 This also avoids the problems with SC4.2 and unpatched SC5.
9982 [Andy Polyakov <appro@fy.chalmers.se>]
9984 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9985 These are required because of the typesafe stack would otherwise break
9986 existing code. If old code used a structure member which used to be STACK
9987 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9988 sk_num or sk_value it would produce an error because the num, data members
9989 are not present in STACK_OF. Now it just produces a warning. sk_set
9990 replaces the old method of assigning a value to sk_value
9991 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9992 that does this will no longer work (and should use sk_set instead) but
9993 this could be regarded as a "questionable" behaviour anyway.
9996 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9997 correctly handle encrypted S/MIME data.
10000 *) Change type of various DES function arguments from des_cblock
10001 (which means, in function argument declarations, pointer to char)
10002 to des_cblock * (meaning pointer to array with 8 char elements),
10003 which allows the compiler to do more typechecking; it was like
10004 that back in SSLeay, but with lots of ugly casts.
10006 Introduce new type const_des_cblock.
10009 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10010 problems: find RecipientInfo structure that matches recipient certificate
10011 and initialise the ASN1 structures properly based on passed cipher.
10014 *) Belatedly make the BN tests actually check the results.
10017 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10018 to and from BNs: it was completely broken. New compilation option
10019 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10020 key elements as negative integers.
10023 *) Reorganize and speed up MD5.
10024 [Andy Polyakov <appro@fy.chalmers.se>]
10027 [Richard Levitte <richard@levitte.org>]
10029 *) New option -out to asn1parse to allow the parsed structure to be
10030 output to a file. This is most useful when combined with the -strparse
10031 option to examine the output of things like OCTET STRINGS.
10034 *) Make SSL library a little more fool-proof by not requiring any longer
10035 that SSL_set_{accept,connect}_state be called before
10036 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10037 in many applications because usually everything *appeared* to work as
10038 intended anyway -- now it really works as intended).
10041 *) Move openssl.cnf out of lib/.
10044 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10045 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10046 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10047 [Ralf S. Engelschall]
10049 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10050 handle PKCS#7 enveloped data properly.
10051 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10053 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10054 copying pointers. The cert_st handling is changed by this in
10055 various ways (and thus what used to be known as ctx->default_cert
10056 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10057 any longer when s->cert does not give us what we need).
10058 ssl_cert_instantiate becomes obsolete by this change.
10059 As soon as we've got the new code right (possibly it already is?),
10060 we have solved a couple of bugs of the earlier code where s->cert
10061 was used as if it could not have been shared with other SSL structures.
10063 Note that using the SSL API in certain dirty ways now will result
10064 in different behaviour than observed with earlier library versions:
10065 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10066 does not influence s as it used to.
10068 In order to clean up things more thoroughly, inside SSL_SESSION
10069 we don't use CERT any longer, but a new structure SESS_CERT
10070 that holds per-session data (if available); currently, this is
10071 the peer's certificate chain and, for clients, the server's certificate
10072 and temporary key. CERT holds only those values that can have
10073 meaningful defaults in an SSL_CTX.
10076 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10077 from the internal representation. Various PKCS#7 fixes: remove some
10078 evil casts and set the enc_dig_alg field properly based on the signing
10082 *) Allow PKCS#12 password to be set from the command line or the
10083 environment. Let 'ca' get its config file name from the environment
10084 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10088 *) Allow certificate policies extension to use an IA5STRING for the
10089 organization field. This is contrary to the PKIX definition but
10090 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10094 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10095 without disallowing inline assembler and the like for non-pedantic builds.
10098 *) Support Borland C++ builder.
10099 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10101 *) Support Mingw32.
10104 *) SHA-1 cleanups and performance enhancements.
10105 [Andy Polyakov <appro@fy.chalmers.se>]
10107 *) Sparc v8plus assembler for the bignum library.
10108 [Andy Polyakov <appro@fy.chalmers.se>]
10110 *) Accept any -xxx and +xxx compiler options in Configure.
10113 *) Update HPUX configuration.
10116 *) Add missing sk_<type>_unshift() function to safestack.h
10117 [Ralf S. Engelschall]
10119 *) New function SSL_CTX_use_certificate_chain_file that sets the
10120 "extra_cert"s in addition to the certificate. (This makes sense
10121 only for "PEM" format files, as chains as a whole are not
10125 *) Support verify_depth from the SSL API.
10126 x509_vfy.c had what can be considered an off-by-one-error:
10127 Its depth (which was not part of the external interface)
10128 was actually counting the number of certificates in a chain;
10129 now it really counts the depth.
10132 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10133 instead of X509err, which often resulted in confusing error
10134 messages since the error codes are not globally unique
10135 (e.g. an alleged error in ssl3_accept when a certificate
10136 didn't match the private key).
10138 *) New function SSL_CTX_set_session_id_context that allows to set a default
10139 value (so that you don't need SSL_set_session_id_context for each
10140 connection using the SSL_CTX).
10143 *) OAEP decoding bug fix.
10146 *) Support INSTALL_PREFIX for package builders, as proposed by
10150 *) New Configure options "threads" and "no-threads". For systems
10151 where the proper compiler options are known (currently Solaris
10152 and Linux), "threads" is the default.
10155 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10158 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10159 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10160 such as /usr/local/bin.
10163 *) "make linux-shared" to build shared libraries.
10164 [Niels Poppe <niels@netbox.org>]
10166 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10169 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10170 extension adding in x509 utility.
10173 *) Remove NOPROTO sections and error code comments.
10176 *) Partial rewrite of the DEF file generator to now parse the ANSI
10180 *) New Configure options --prefix=DIR and --openssldir=DIR.
10183 *) Complete rewrite of the error code script(s). It is all now handled
10184 by one script at the top level which handles error code gathering,
10185 header rewriting and C source file generation. It should be much better
10186 than the old method: it now uses a modified version of Ulf's parser to
10187 read the ANSI prototypes in all header files (thus the old K&R definitions
10188 aren't needed for error creation any more) and do a better job of
10189 translating function codes into names. The old 'ASN1 error code imbedded
10190 in a comment' is no longer necessary and it doesn't use .err files which
10191 have now been deleted. Also the error code call doesn't have to appear all
10192 on one line (which resulted in some large lines...).
10195 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10198 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10199 0 (which usually indicates a closed connection), but continue reading.
10202 *) Fix some race conditions.
10205 *) Add support for CRL distribution points extension. Add Certificate
10206 Policies and CRL distribution points documentation.
10209 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10212 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10213 8 of keying material. Merlin has also confirmed interop with this fix
10214 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10215 [Merlin Hughes <merlin@baltimore.ie>]
10217 *) Fix lots of warnings.
10218 [Richard Levitte <levitte@stacken.kth.se>]
10220 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10221 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10222 [Richard Levitte <levitte@stacken.kth.se>]
10224 *) Fix problems with sizeof(long) == 8.
10225 [Andy Polyakov <appro@fy.chalmers.se>]
10227 *) Change functions to ANSI C.
10230 *) Fix typos in error codes.
10231 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10233 *) Remove defunct assembler files from Configure.
10236 *) SPARC v8 assembler BIGNUM implementation.
10237 [Andy Polyakov <appro@fy.chalmers.se>]
10239 *) Support for Certificate Policies extension: both print and set.
10240 Various additions to support the r2i method this uses.
10243 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10244 return a const string when you are expecting an allocated buffer.
10247 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10248 types DirectoryString and DisplayText.
10251 *) Add code to allow r2i extensions to access the configuration database,
10252 add an LHASH database driver and add several ctx helper functions.
10255 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10256 fail when they extended the size of a BIGNUM.
10259 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10260 support typesafe stack.
10263 *) Fix typo in SSL_[gs]et_options().
10264 [Nils Frostberg <nils@medcom.se>]
10266 *) Delete various functions and files that belonged to the (now obsolete)
10267 old X509V3 handling code.
10270 *) New Configure option "rsaref".
10273 *) Don't auto-generate pem.h.
10276 *) Introduce type-safe ASN.1 SETs.
10279 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10280 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10282 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10283 that links with OpenSSL (well at least cause lots of warnings), but fear
10284 not: the conversion is trivial, and it eliminates loads of evil casts. A
10285 few STACKed things have been converted already. Feel free to convert more.
10286 In the fullness of time, I'll do away with the STACK type altogether.
10289 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10290 specified in <certfile> by updating the entry in the index.txt file.
10291 This way one no longer has to edit the index.txt file manually for
10292 revoking a certificate. The -revoke option does the gory details now.
10293 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10295 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10296 `-text' option at all and this way the `-noout -text' combination was
10297 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10298 [Ralf S. Engelschall]
10300 *) Make sure a corresponding plain text error message exists for the
10301 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10302 verify callback function determined that a certificate was revoked.
10303 [Ralf S. Engelschall]
10305 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10306 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10307 all available cipers including rc5, which was forgotten until now.
10308 In order to let the testing shell script know which algorithms
10309 are available, a new (up to now undocumented) command
10310 "openssl list-cipher-commands" is used.
10313 *) Bugfix: s_client occasionally would sleep in select() when
10314 it should have checked SSL_pending() first.
10317 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10318 the raw DSA values prior to ASN.1 encoding.
10321 *) Tweaks to Configure
10322 [Niels Poppe <niels@netbox.org>]
10324 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10328 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10331 *) New config option to avoid instructions that are illegal on the 80386.
10332 The default code is faster, but requires at least a 486.
10335 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10336 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10337 same as SSL2_VERSION anyway.
10340 *) New "-showcerts" option for s_client.
10343 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10344 application. Various cleanups and fixes.
10347 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10348 modify error routines to work internally. Add error codes and PBE init
10349 to library startup routines.
10352 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10353 packing functions to asn1 and evp. Changed function names and error
10354 codes along the way.
10357 *) PKCS12 integration: and so it begins... First of several patches to
10358 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10359 objects to objects.h
10362 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10363 and display support for Thawte strong extranet extension.
10366 *) Add LinuxPPC support.
10367 [Jeff Dubrule <igor@pobox.org>]
10369 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10370 bn_div_words in alpha.s.
10371 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10373 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10374 OAEP isn't supported when OpenSSL is built with RSAref.
10375 [Ulf Moeller <ulf@fitug.de>]
10377 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10378 so they no longer are missing under -DNOPROTO.
10379 [Soren S. Jorvang <soren@t.dk>]
10382 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10384 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10385 doesn't work when the session is reused. Coming soon!
10388 *) Fix a security hole, that allows sessions to be reused in the wrong
10389 context thus bypassing client cert protection! All software that uses
10390 client certs and session caches in multiple contexts NEEDS PATCHING to
10391 allow session reuse! A fuller solution is in the works.
10392 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10394 *) Some more source tree cleanups (removed obsolete files
10395 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10396 permission on "config" script to be executable) and a fix for the INSTALL
10398 [Ulf Moeller <ulf@fitug.de>]
10400 *) Remove some legacy and erroneous uses of malloc, free instead of
10402 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10404 *) Make rsa_oaep_test return non-zero on error.
10405 [Ulf Moeller <ulf@fitug.de>]
10407 *) Add support for native Solaris shared libraries. Configure
10408 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10409 if someone would make that last step automatic.
10410 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10412 *) ctx_size was not built with the right compiler during "make links". Fixed.
10415 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10416 except NULL ciphers". This means the default cipher list will no longer
10417 enable NULL ciphers. They need to be specifically enabled e.g. with
10418 the string "DEFAULT:eNULL".
10421 *) Fix to RSA private encryption routines: if p < q then it would
10422 occasionally produce an invalid result. This will only happen with
10423 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10426 *) Be less restrictive and allow also `perl util/perlpath.pl
10427 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10428 because this way one can also use an interpreter named `perl5' (which is
10429 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10430 installed as `perl').
10431 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10433 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10434 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10436 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10437 advapi32.lib to Win32 build and change the pem test comparision
10438 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10439 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10440 and crypto/des/ede_cbcm_enc.c.
10443 *) DES quad checksum was broken on big-endian architectures. Fixed.
10446 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10447 Win32 test batch file so it (might) work again. The Win32 test batch file
10448 is horrible: I feel ill....
10451 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10452 in e_os.h. Audit of header files to check ANSI and non ANSI
10453 sections: 10 functions were absent from non ANSI section and not exported
10454 from Windows DLLs. Fixed up libeay.num for new functions.
10457 *) Make `openssl version' output lines consistent.
10458 [Ralf S. Engelschall]
10460 *) Fix Win32 symbol export lists for BIO functions: Added
10461 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10462 to ms/libeay{16,32}.def.
10463 [Ralf S. Engelschall]
10465 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10466 fine under Unix and passes some trivial tests I've now added. But the
10467 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10468 added to make sure no one expects that this stuff really works in the
10469 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10470 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10472 [Ralf S. Engelschall]
10474 *) Fix the generation of two part addresses in perl.
10475 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10477 *) Add config entry for Linux on MIPS.
10478 [John Tobey <jtobey@channel1.com>]
10480 *) Make links whenever Configure is run, unless we are on Windoze.
10483 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10484 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10488 *) Add a useful kludge to allow package maintainers to specify compiler and
10489 other platforms details on the command line without having to patch the
10490 Configure script everytime: One now can use ``perl Configure
10491 <id>:<details>'', i.e. platform ids are allowed to have details appended
10492 to them (seperated by colons). This is treated as there would be a static
10493 pre-configured entry in Configure's %table under key <id> with value
10494 <details> and ``perl Configure <id>'' is called. So, when you want to
10495 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10496 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10497 now, which overrides the FreeBSD-elf entry on-the-fly.
10498 [Ralf S. Engelschall]
10500 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10503 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10504 on the `perl Configure ...' command line. This way one can compile
10505 OpenSSL libraries with Position Independent Code (PIC) which is needed
10506 for linking it into DSOs.
10507 [Ralf S. Engelschall]
10509 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10513 *) Cleaned up the LICENSE document: The official contact for any license
10514 questions now is the OpenSSL core team under openssl-core@openssl.org.
10515 And add a paragraph about the dual-license situation to make sure people
10516 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10517 to the OpenSSL toolkit.
10518 [Ralf S. Engelschall]
10520 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10521 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10522 Additonally cleaned up the `make links' target: Remove unnecessary
10523 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10524 to speed processing and no longer clutter the display with confusing
10525 stuff. Instead only the actually done links are displayed.
10526 [Ralf S. Engelschall]
10528 *) Permit null encryption ciphersuites, used for authentication only. It used
10529 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10530 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10534 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10535 signed attributes when verifying signatures (this would break them),
10536 the detached data encoding was wrong and public keys obtained using
10537 X509_get_pubkey() weren't freed.
10540 *) Add text documentation for the BUFFER functions. Also added a work around
10541 to a Win95 console bug. This was triggered by the password read stuff: the
10542 last character typed gets carried over to the next fread(). If you were
10543 generating a new cert request using 'req' for example then the last
10544 character of the passphrase would be CR which would then enter the first
10548 *) Added the new `Includes OpenSSL Cryptography Software' button as
10549 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10550 button and can be used by applications based on OpenSSL to show the
10551 relationship to the OpenSSL project.
10552 [Ralf S. Engelschall]
10554 *) Remove confusing variables in function signatures in files
10555 ssl/ssl_lib.c and ssl/ssl.h.
10556 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10558 *) Don't install bss_file.c under PREFIX/include/
10559 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10561 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10562 functions that return function pointers and has support for NT specific
10563 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10564 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10565 unsigned to signed types: this was killing the Win32 compile.
10568 *) Add new certificate file to stack functions,
10569 SSL_add_dir_cert_subjects_to_stack() and
10570 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10571 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10572 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10573 This means that Apache-SSL and similar packages don't have to mess around
10574 to add as many CAs as they want to the preferred list.
10577 *) Experiment with doxygen documentation. Currently only partially applied to
10579 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10580 openssl.doxy as the configuration file.
10583 *) Get rid of remaining C++-style comments which strict C compilers hate.
10584 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10586 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10587 compiled in by default: it has problems with large keys.
10590 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10591 DH private keys and/or callback functions which directly correspond to
10592 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10593 is needed for applications which have to configure certificates on a
10594 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10596 For the RSA certificate situation is makes no difference, but
10597 for the DSA certificate situation this fixes the "no shared cipher"
10598 problem where the OpenSSL cipher selection procedure failed because the
10599 temporary keys were not overtaken from the context and the API provided
10600 no way to reconfigure them.
10601 The new functions now let applications reconfigure the stuff and they
10602 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10603 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10604 non-public-API function ssl_cert_instantiate() is used as a helper
10605 function and also to reduce code redundancy inside ssl_rsa.c.
10606 [Ralf S. Engelschall]
10608 *) Move s_server -dcert and -dkey options out of the undocumented feature
10609 area because they are useful for the DSA situation and should be
10610 recognized by the users.
10611 [Ralf S. Engelschall]
10613 *) Fix the cipher decision scheme for export ciphers: the export bits are
10614 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10615 SSL_EXP_MASK. So, the original variable has to be used instead of the
10616 already masked variable.
10617 [Richard Levitte <levitte@stacken.kth.se>]
10619 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10620 [Richard Levitte <levitte@stacken.kth.se>]
10622 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10623 from `int' to `unsigned int' because it's a length and initialized by
10624 EVP_DigestFinal() which expects an `unsigned int *'.
10625 [Richard Levitte <levitte@stacken.kth.se>]
10627 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10628 script. Instead use the usual Shell->Perl transition trick.
10629 [Ralf S. Engelschall]
10631 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10632 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10633 -noout -modulus' as it's already the case for `openssl rsa -noout
10634 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10635 currently the public key is printed (a decision which was already done by
10636 `openssl dsa -modulus' in the past) which serves a similar purpose.
10637 Additionally the NO_RSA no longer completely removes the whole -modulus
10638 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10640 [Ralf S. Engelschall]
10642 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10643 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10644 [Arne Ansper <arne@ats.cyber.ee>]
10646 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10647 to be added. Now both 'req' and 'ca' can use new objects defined in the
10651 *) Add cool BIO that does syslog (or event log on NT).
10652 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10654 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10655 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10656 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10657 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10660 *) Add preliminary config info for new extension code.
10663 *) Make RSA_NO_PADDING really use no padding.
10664 [Ulf Moeller <ulf@fitug.de>]
10666 *) Generate errors when private/public key check is done.
10669 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10670 for some CRL extensions and new objects added.
10673 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10674 key usage extension and fuller support for authority key id.
10677 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10678 padding method for RSA, which is recommended for new applications in PKCS
10679 #1 v2.0 (RFC 2437, October 1998).
10680 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10681 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10682 against Bleichbacher's attack on RSA.
10683 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10686 *) Updates to the new SSL compression code
10687 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10689 *) Fix so that the version number in the master secret, when passed
10690 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10691 (because the server will not accept higher), that the version number
10692 is 0x03,0x01, not 0x03,0x00
10693 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10695 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10696 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10697 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10700 *) Support for RAW extensions where an arbitrary extension can be
10701 created by including its DER encoding. See apps/openssl.cnf for
10705 *) Make sure latest Perl versions don't interpret some generated C array
10706 code as Perl array code in the crypto/err/err_genc.pl script.
10707 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10709 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10710 not many people have the assembler. Various Win32 compilation fixes and
10711 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10712 build instructions.
10715 *) Modify configure script 'Configure' to automatically create crypto/date.h
10716 file under Win32 and also build pem.h from pem.org. New script
10717 util/mkfiles.pl to create the MINFO file on environments that can't do a
10718 'make files': perl util/mkfiles.pl >MINFO should work.
10721 *) Major rework of DES function declarations, in the pursuit of correctness
10722 and purity. As a result, many evil casts evaporated, and some weirdness,
10723 too. You may find this causes warnings in your code. Zapping your evil
10724 casts will probably fix them. Mostly.
10727 *) Fix for a typo in asn1.h. Bug fix to object creation script
10728 obj_dat.pl. It considered a zero in an object definition to mean
10729 "end of object": none of the objects in objects.h have any zeros
10730 so it wasn't spotted.
10731 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10733 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10734 Masking (CBCM). In the absence of test vectors, the best I have been able
10735 to do is check that the decrypt undoes the encrypt, so far. Send me test
10736 vectors if you have them.
10739 *) Correct calculation of key length for export ciphers (too much space was
10740 allocated for null ciphers). This has not been tested!
10743 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10744 message is now correct (it understands "crypto" and "ssl" on its
10745 command line). There is also now an "update" option. This will update
10746 the util/ssleay.num and util/libeay.num files with any new functions.
10748 perl util/mkdef.pl crypto ssl update
10749 it will update them.
10752 *) Overhauled the Perl interface (perl/*):
10753 - ported BN stuff to OpenSSL's different BN library
10754 - made the perl/ source tree CVS-aware
10755 - renamed the package from SSLeay to OpenSSL (the files still contain
10756 their history because I've copied them in the repository)
10757 - removed obsolete files (the test scripts will be replaced
10758 by better Test::Harness variants in the future)
10759 [Ralf S. Engelschall]
10761 *) First cut for a very conservative source tree cleanup:
10762 1. merge various obsolete readme texts into doc/ssleay.txt
10763 where we collect the old documents and readme texts.
10764 2. remove the first part of files where I'm already sure that we no
10765 longer need them because of three reasons: either they are just temporary
10766 files which were left by Eric or they are preserved original files where
10767 I've verified that the diff is also available in the CVS via "cvs diff
10768 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10769 the crypto/md/ stuff).
10770 [Ralf S. Engelschall]
10772 *) More extension code. Incomplete support for subject and issuer alt
10773 name, issuer and authority key id. Change the i2v function parameters
10774 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10775 what that's for :-) Fix to ASN1 macro which messed up
10776 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10779 *) Preliminary support for ENUMERATED type. This is largely copied from the
10783 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10784 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10786 *) Make sure `make rehash' target really finds the `openssl' program.
10787 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10789 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10790 like to hear about it if this slows down other processors.
10793 *) Add CygWin32 platform information to Configure script.
10794 [Alan Batie <batie@aahz.jf.intel.com>]
10796 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10797 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10799 *) New program nseq to manipulate netscape certificate sequences
10802 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10806 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10807 but the BN code had some problems that would cause failures when
10808 doing certificate verification and some other functions.
10809 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10811 *) Add ASN1 and PEM code to support netscape certificate sequences.
10814 *) Add ASN1 and PEM code to support netscape certificate sequences.
10817 *) Add several PKIX and private extended key usage OIDs.
10820 *) Modify the 'ca' program to handle the new extension code. Modify
10821 openssl.cnf for new extension format, add comments.
10824 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10825 and add a sample to openssl.cnf so req -x509 now adds appropriate
10829 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10830 error code, add initial support to X509_print() and x509 application.
10833 *) Takes a deep breath and start addding X509 V3 extension support code. Add
10834 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10835 stuff is currently isolated and isn't even compiled yet.
10838 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10839 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10840 Removed the versions check from X509 routines when loading extensions:
10841 this allows certain broken certificates that don't set the version
10842 properly to be processed.
10845 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10846 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10847 can still be regenerated with "make depend".
10850 *) Spelling mistake in C version of CAST-128.
10851 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10853 *) Changes to the error generation code. The perl script err-code.pl
10854 now reads in the old error codes and retains the old numbers, only
10855 adding new ones if necessary. It also only changes the .err files if new
10856 codes are added. The makefiles have been modified to only insert errors
10857 when needed (to avoid needlessly modifying header files). This is done
10858 by only inserting errors if the .err file is newer than the auto generated
10859 C file. To rebuild all the error codes from scratch (the old behaviour)
10860 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10861 or delete all the .err files.
10864 *) CAST-128 was incorrectly implemented for short keys. The C version has
10865 been fixed, but is untested. The assembler versions are also fixed, but
10866 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10867 to regenerate it if needed.
10868 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10869 Hagino <itojun@kame.net>]
10871 *) File was opened incorrectly in randfile.c.
10872 [Ulf Möller <ulf@fitug.de>]
10874 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10875 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10876 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10877 al: it's just almost always a UTCTime. Note this patch adds new error
10878 codes so do a "make errors" if there are problems.
10881 *) Correct Linux 1 recognition in config.
10882 [Ulf Möller <ulf@fitug.de>]
10884 *) Remove pointless MD5 hash when using DSA keys in ca.
10885 [Anonymous <nobody@replay.com>]
10887 *) Generate an error if given an empty string as a cert directory. Also
10888 generate an error if handed NULL (previously returned 0 to indicate an
10889 error, but didn't set one).
10890 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10892 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10895 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10896 parameters. This was causing a warning which killed off the Win32 compile.
10899 *) Remove C++ style comments from crypto/bn/bn_local.h.
10900 [Neil Costigan <neil.costigan@celocom.com>]
10902 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10903 based on a text string, looking up short and long names and finally
10904 "dot" format. The "dot" format stuff didn't work. Added new function
10905 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10906 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10907 OID is not part of the table.
10910 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10911 X509_LOOKUP_by_alias().
10914 *) Sort openssl functions by name.
10917 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10918 encryption from sample DSA keys (in case anyone is interested the password
10922 *) Make _all_ *_free functions accept a NULL pointer.
10923 [Frans Heymans <fheymans@isaserver.be>]
10925 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10927 [Anonymous <nobody@replay.com>]
10929 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10930 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10932 *) Don't blow it for numeric -newkey arguments to apps/req.
10933 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10935 *) Temp key "for export" tests were wrong in s3_srvr.c.
10936 [Anonymous <nobody@replay.com>]
10938 *) Add prototype for temp key callback functions
10939 SSL_CTX_set_tmp_{rsa,dh}_callback().
10942 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10943 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10946 *) X509_name_add_entry() freed the wrong thing after an error.
10947 [Arne Ansper <arne@ats.cyber.ee>]
10949 *) rsa_eay.c would attempt to free a NULL context.
10950 [Arne Ansper <arne@ats.cyber.ee>]
10952 *) BIO_s_socket() had a broken should_retry() on Windoze.
10953 [Arne Ansper <arne@ats.cyber.ee>]
10955 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10956 [Arne Ansper <arne@ats.cyber.ee>]
10958 *) Make sure the already existing X509_STORE->depth variable is initialized
10959 in X509_STORE_new(), but document the fact that this variable is still
10960 unused in the certificate verification process.
10961 [Ralf S. Engelschall]
10963 *) Fix the various library and apps files to free up pkeys obtained from
10964 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10967 *) Fix reference counting in X509_PUBKEY_get(). This makes
10968 demos/maurice/example2.c work, amongst others, probably.
10969 [Steve Henson and Ben Laurie]
10971 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10972 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10973 are no longer created. This way we have a single and consistent command
10974 line interface `openssl <command>', similar to `cvs <command>'.
10975 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10977 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10978 BIT STRING wrapper always have zero unused bits.
10981 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10984 *) Make the top-level INSTALL documentation easier to understand.
10987 *) Makefiles updated to exit if an error occurs in a sub-directory
10988 make (including if user presses ^C) [Paul Sutton]
10990 *) Make Montgomery context stuff explicit in RSA data structure.
10993 *) Fix build order of pem and err to allow for generated pem.h.
10996 *) Fix renumbering bug in X509_NAME_delete_entry().
10999 *) Enhanced the err-ins.pl script so it makes the error library number
11000 global and can add a library name. This is needed for external ASN1 and
11001 other error libraries.
11004 *) Fixed sk_insert which never worked properly.
11007 *) Fix ASN1 macros so they can handle indefinite length construted
11008 EXPLICIT tags. Some non standard certificates use these: they can now
11012 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11013 into a single doc/ssleay.txt bundle. This way the information is still
11014 preserved but no longer messes up this directory. Now it's new room for
11015 the new set of documenation files.
11016 [Ralf S. Engelschall]
11018 *) SETs were incorrectly DER encoded. This was a major pain, because they
11019 shared code with SEQUENCEs, which aren't coded the same. This means that
11020 almost everything to do with SETs or SEQUENCEs has either changed name or
11021 number of arguments.
11022 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11024 *) Fix test data to work with the above.
11027 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11028 was already fixed by Eric for 0.9.1 it seems.
11029 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11031 *) Autodetect FreeBSD3.
11034 *) Fix various bugs in Configure. This affects the following platforms:
11038 unixware-2.0-pentium
11042 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11043 before they are needed.
11046 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11050 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11052 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11053 changed SSLeay to OpenSSL in version strings.
11054 [Ralf S. Engelschall]
11056 *) Some fixups to the top-level documents.
11059 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11060 because the symlink to include/ was missing.
11061 [Ralf S. Engelschall]
11063 *) Incorporated the popular no-RSA/DSA-only patches
11064 which allow to compile a RSA-free SSLeay.
11065 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11067 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11068 when "ssleay" is still not found.
11069 [Ralf S. Engelschall]
11071 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11072 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11074 *) Updated the README file.
11075 [Ralf S. Engelschall]
11077 *) Added various .cvsignore files in the CVS repository subdirs
11078 to make a "cvs update" really silent.
11079 [Ralf S. Engelschall]
11081 *) Recompiled the error-definition header files and added
11082 missing symbols to the Win32 linker tables.
11083 [Ralf S. Engelschall]
11085 *) Cleaned up the top-level documents;
11086 o new files: CHANGES and LICENSE
11087 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11088 o merged COPYRIGHT into LICENSE
11089 o removed obsolete TODO file
11090 o renamed MICROSOFT to INSTALL.W32
11091 [Ralf S. Engelschall]
11093 *) Removed dummy files from the 0.9.1b source tree:
11094 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11095 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11096 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11097 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11098 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11099 [Ralf S. Engelschall]
11101 *) Added various platform portability fixes.
11104 *) The Genesis of the OpenSSL rpject:
11105 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11106 Young and Tim J. Hudson created while they were working for C2Net until
11108 [The OpenSSL Project]
11111 Changes between 0.9.0b and 0.9.1b [not released]
11113 *) Updated a few CA certificates under certs/
11116 *) Changed some BIGNUM api stuff.
11119 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11120 DGUX x86, Linux Alpha, etc.
11123 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11124 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11128 *) Add -strparse option to asn1pars program which parses nested
11130 [Dr Stephen Henson <shenson@bigfoot.com>]
11132 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11135 *) DSA fix for "ca" program.
11138 *) Added "-genkey" option to "dsaparam" program.
11141 *) Added RIPE MD160 (rmd160) message digest.
11144 *) Added -a (all) option to "ssleay version" command.
11147 *) Added PLATFORM define which is the id given to Configure.
11150 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11153 *) Extended the ASN.1 parser routines.
11156 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11159 *) Added a BN_CTX to the BN library.
11162 *) Fixed the weak key values in DES library
11165 *) Changed API in EVP library for cipher aliases.
11168 *) Added support for RC2/64bit cipher.
11171 *) Converted the lhash library to the crypto/mem.c functions.
11174 *) Added more recognized ASN.1 object ids.
11177 *) Added more RSA padding checks for SSL/TLS.
11180 *) Added BIO proxy/filter functionality.
11183 *) Added extra_certs to SSL_CTX which can be used
11184 send extra CA certificates to the client in the CA cert chain sending
11185 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11188 *) Now Fortezza is denied in the authentication phase because
11189 this is key exchange mechanism is not supported by SSLeay at all.
11192 *) Additional PKCS1 checks.
11195 *) Support the string "TLSv1" for all TLS v1 ciphers.
11198 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11199 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11202 *) Fixed a few memory leaks.
11205 *) Fixed various code and comment typos.
11208 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11209 bytes sent in the client random.
11210 [Edward Bishop <ebishop@spyglass.com>]