5 Changes between 1.0.2c and 1.0.2d [xx XXX xxxx]
7 *) Alternate chains certificate forgery
9 During certificate verfification, OpenSSL will attempt to find an
10 alternative certificate chain if the first attempt to build such a chain
11 fails. An error in the implementation of this logic can mean that an
12 attacker could cause certain checks on untrusted certificates to be
13 bypassed, such as the CA flag, enabling them to use a valid leaf
14 certificate to act as a CA and "issue" an invalid certificate.
16 This issue was reported to OpenSSL by Adam Langley/David Benjamin
20 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
22 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
23 incompatibility in the handling of HMAC. The previous ABI has now been
26 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
28 *) Malformed ECParameters causes infinite loop
30 When processing an ECParameters structure OpenSSL enters an infinite loop
31 if the curve specified is over a specially malformed binary polynomial
34 This can be used to perform denial of service against any
35 system which processes public keys, certificate requests or
36 certificates. This includes TLS clients and TLS servers with
37 client authentication enabled.
39 This issue was reported to OpenSSL by Joseph Barr-Pixton.
43 *) Exploitable out-of-bounds read in X509_cmp_time
45 X509_cmp_time does not properly check the length of the ASN1_TIME
46 string and can read a few bytes out of bounds. In addition,
47 X509_cmp_time accepts an arbitrary number of fractional seconds in the
50 An attacker can use this to craft malformed certificates and CRLs of
51 various sizes and potentially cause a segmentation fault, resulting in
52 a DoS on applications that verify certificates or CRLs. TLS clients
53 that verify CRLs are affected. TLS clients and servers with client
54 authentication enabled may be affected if they use custom verification
57 This issue was reported to OpenSSL by Robert Swiecki (Google), and
58 independently by Hanno Böck.
62 *) PKCS7 crash with missing EnvelopedContent
64 The PKCS#7 parsing code does not handle missing inner EncryptedContent
65 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
66 with missing content and trigger a NULL pointer dereference on parsing.
68 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
69 structures from untrusted sources are affected. OpenSSL clients and
70 servers are not affected.
72 This issue was reported to OpenSSL by Michal Zalewski (Google).
76 *) CMS verify infinite loop with unknown hash function
78 When verifying a signedData message the CMS code can enter an infinite loop
79 if presented with an unknown hash function OID. This can be used to perform
80 denial of service against any system which verifies signedData messages using
82 This issue was reported to OpenSSL by Johannes Bauer.
86 *) Race condition handling NewSessionTicket
88 If a NewSessionTicket is received by a multi-threaded client when attempting to
89 reuse a previous ticket then a race condition can occur potentially leading to
90 a double free of the ticket data.
94 *) Removed support for the two export grade static DH ciphersuites
95 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
96 were newly added (along with a number of other static DH ciphersuites) to
97 1.0.2. However the two export ones have *never* worked since they were
98 introduced. It seems strange in any case to be adding new export
99 ciphersuites, and given "logjam" it also does not seem correct to fix them.
102 *) Only support 256-bit or stronger elliptic curves with the
103 'ecdh_auto' setting (server) or by default (client). Of supported
104 curves, prefer P-256 (both).
107 *) Reject DH handshakes with parameters shorter than 768 bits.
108 [Kurt Roeckx and Emilia Kasper]
110 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
112 *) ClientHello sigalgs DoS fix
114 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
115 invalid signature algorithms extension a NULL pointer dereference will
116 occur. This can be exploited in a DoS attack against the server.
118 This issue was was reported to OpenSSL by David Ramos of Stanford
121 [Stephen Henson and Matt Caswell]
123 *) Multiblock corrupted pointer fix
125 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
126 feature only applies on 64 bit x86 architecture platforms that support AES
127 NI instructions. A defect in the implementation of "multiblock" can cause
128 OpenSSL's internal write buffer to become incorrectly set to NULL when
129 using non-blocking IO. Typically, when the user application is using a
130 socket BIO for writing, this will only result in a failed connection.
131 However if some other BIO is used then it is likely that a segmentation
132 fault will be triggered, thus enabling a potential DoS attack.
134 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
138 *) Segmentation fault in DTLSv1_listen fix
140 The DTLSv1_listen function is intended to be stateless and processes the
141 initial ClientHello from many peers. It is common for user code to loop
142 over the call to DTLSv1_listen until a valid ClientHello is received with
143 an associated cookie. A defect in the implementation of DTLSv1_listen means
144 that state is preserved in the SSL object from one invocation to the next
145 that can lead to a segmentation fault. Errors processing the initial
146 ClientHello can trigger this scenario. An example of such an error could be
147 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
150 This issue was reported to OpenSSL by Per Allansson.
154 *) Segmentation fault in ASN1_TYPE_cmp fix
156 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
157 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
158 certificate signature algorithm consistency this can be used to crash any
159 certificate verification operation and exploited in a DoS attack. Any
160 application which performs certificate verification is vulnerable including
161 OpenSSL clients and servers which enable client authentication.
165 *) Segmentation fault for invalid PSS parameters fix
167 The signature verification routines will crash with a NULL pointer
168 dereference if presented with an ASN.1 signature using the RSA PSS
169 algorithm and invalid parameters. Since these routines are used to verify
170 certificate signature algorithms this can be used to crash any
171 certificate verification operation and exploited in a DoS attack. Any
172 application which performs certificate verification is vulnerable including
173 OpenSSL clients and servers which enable client authentication.
175 This issue was was reported to OpenSSL by Brian Carpenter.
179 *) ASN.1 structure reuse memory corruption fix
181 Reusing a structure in ASN.1 parsing may allow an attacker to cause
182 memory corruption via an invalid write. Such reuse is and has been
183 strongly discouraged and is believed to be rare.
185 Applications that parse structures containing CHOICE or ANY DEFINED BY
186 components may be affected. Certificate parsing (d2i_X509 and related
187 functions) are however not affected. OpenSSL clients and servers are
192 *) PKCS7 NULL pointer dereferences fix
194 The PKCS#7 parsing code does not handle missing outer ContentInfo
195 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
196 missing content and trigger a NULL pointer dereference on parsing.
198 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
199 otherwise parse PKCS#7 structures from untrusted sources are
200 affected. OpenSSL clients and servers are not affected.
202 This issue was reported to OpenSSL by Michal Zalewski (Google).
206 *) DoS via reachable assert in SSLv2 servers fix
208 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
209 servers that both support SSLv2 and enable export cipher suites by sending
210 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
212 This issue was discovered by Sean Burford (Google) and Emilia Käsper
213 (OpenSSL development team).
217 *) Empty CKE with client auth and DHE fix
219 If client auth is used then a server can seg fault in the event of a DHE
220 ciphersuite being selected and a zero length ClientKeyExchange message
221 being sent by the client. This could be exploited in a DoS attack.
225 *) Handshake with unseeded PRNG fix
227 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
228 with an unseeded PRNG. The conditions are:
229 - The client is on a platform where the PRNG has not been seeded
230 automatically, and the user has not seeded manually
231 - A protocol specific client method version has been used (i.e. not
232 SSL_client_methodv23)
233 - A ciphersuite is used that does not require additional random data from
234 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
236 If the handshake succeeds then the client random that has been used will
237 have been generated from a PRNG with insufficient entropy and therefore the
238 output may be predictable.
240 For example using the following command with an unseeded openssl will
241 succeed on an unpatched platform:
243 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
247 *) Use After Free following d2i_ECPrivatekey error fix
249 A malformed EC private key file consumed via the d2i_ECPrivateKey function
250 could cause a use after free condition. This, in turn, could cause a double
251 free in several private key parsing functions (such as d2i_PrivateKey
252 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
253 for applications that receive EC private keys from untrusted
254 sources. This scenario is considered rare.
256 This issue was discovered by the BoringSSL project and fixed in their
261 *) X509_to_X509_REQ NULL pointer deref fix
263 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
264 the certificate key is invalid. This function is rarely used in practice.
266 This issue was discovered by Brian Carpenter.
270 *) Removed the export ciphers from the DEFAULT ciphers
273 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
275 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
276 ARMv5 through ARMv8, as opposite to "locking" it to single one.
277 So far those who have to target multiple plaforms would compromise
278 and argue that binary targeting say ARMv5 would still execute on
279 ARMv8. "Universal" build resolves this compromise by providing
280 near-optimal performance even on newer platforms.
283 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
284 (other platforms pending).
285 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
287 *) Add support for the SignedCertificateTimestampList certificate and
288 OCSP response extensions from RFC6962.
291 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
292 for corner cases. (Certain input points at infinity could lead to
293 bogus results, with non-infinity inputs mapped to infinity too.)
296 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
297 This covers AES, SHA256/512 and GHASH. "Initial" means that most
298 common cases are optimized and there still is room for further
299 improvements. Vector Permutation AES for Altivec is also added.
302 *) Add support for little-endian ppc64 Linux target.
303 [Marcelo Cerri (IBM)]
305 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
306 SHA1, SHA256 and GHASH. "Initial" means that most common cases
307 are optimized and there still is room for further improvements.
308 Both 32- and 64-bit modes are supported.
309 [Andy Polyakov, Ard Biesheuvel (Linaro)]
311 *) Improved ARMv7 NEON support.
314 *) Support for SPARC Architecture 2011 crypto extensions, first
315 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
316 SHA256/512, MD5, GHASH and modular exponentiation.
317 [Andy Polyakov, David Miller]
319 *) Accelerated modular exponentiation for Intel processors, a.k.a.
321 [Shay Gueron & Vlad Krasnov (Intel Corp)]
323 *) Support for new and upcoming Intel processors, including AVX2,
324 BMI and SHA ISA extensions. This includes additional "stitched"
325 implementations, AESNI-SHA256 and GCM, and multi-buffer support
328 This work was sponsored by Intel Corp.
331 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
332 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
333 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
336 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
337 this fixes a limiation in previous versions of OpenSSL.
340 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
341 MGF1 digest and OAEP label.
344 *) Add EVP support for key wrapping algorithms, to avoid problems with
345 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
346 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
347 algorithms and include tests cases.
350 *) Add functions to allocate and set the fields of an ECDSA_METHOD
352 [Douglas E. Engert, Steve Henson]
354 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
355 difference in days and seconds between two tm or ASN1_TIME structures.
358 *) Add -rev test option to s_server to just reverse order of characters
359 received by client and send back to server. Also prints an abbreviated
360 summary of the connection parameters.
363 *) New option -brief for s_client and s_server to print out a brief summary
364 of connection parameters.
367 *) Add callbacks for arbitrary TLS extensions.
368 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
370 *) New option -crl_download in several openssl utilities to download CRLs
371 from CRLDP extension in certificates.
374 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
377 *) New function X509_CRL_diff to generate a delta CRL from the difference
378 of two full CRLs. Add support to "crl" utility.
381 *) New functions to set lookup_crls function and to retrieve
382 X509_STORE from X509_STORE_CTX.
385 *) Print out deprecated issuer and subject unique ID fields in
389 *) Extend OCSP I/O functions so they can be used for simple general purpose
390 HTTP as well as OCSP. New wrapper function which can be used to download
391 CRLs using the OCSP API.
394 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
397 *) SSL_CONF* functions. These provide a common framework for application
398 configuration using configuration files or command lines.
401 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
402 message callback and prints the results. Needs compile time option
403 "enable-ssl-trace". New options to s_client and s_server to enable
407 *) New ctrl and macro to retrieve supported points extensions.
408 Print out extension in s_server and s_client.
411 *) New functions to retrieve certificate signature and signature
415 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
419 *) New Suite B modes for TLS code. These use and enforce the requirements
420 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
421 only use Suite B curves. The Suite B modes can be set by using the
422 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
425 *) New chain verification flags for Suite B levels of security. Check
426 algorithms are acceptable when flags are set in X509_verify_cert.
429 *) Make tls1_check_chain return a set of flags indicating checks passed
430 by a certificate chain. Add additional tests to handle client
431 certificates: checks for matching certificate type and issuer name
435 *) If an attempt is made to use a signature algorithm not in the peer
436 preference list abort the handshake. If client has no suitable
437 signature algorithms in response to a certificate request do not
441 *) If server EC tmp key is not in client preference list abort handshake.
444 *) Add support for certificate stores in CERT structure. This makes it
445 possible to have different stores per SSL structure or one store in
446 the parent SSL_CTX. Include distint stores for certificate chain
447 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
448 to build and store a certificate chain in CERT structure: returing
449 an error if the chain cannot be built: this will allow applications
450 to test if a chain is correctly configured.
452 Note: if the CERT based stores are not set then the parent SSL_CTX
453 store is used to retain compatibility with existing behaviour.
457 *) New function ssl_set_client_disabled to set a ciphersuite disabled
458 mask based on the current session, check mask when sending client
459 hello and checking the requested ciphersuite.
462 *) New ctrls to retrieve and set certificate types in a certificate
463 request message. Print out received values in s_client. If certificate
464 types is not set with custom values set sensible values based on
465 supported signature algorithms.
468 *) Support for distinct client and server supported signature algorithms.
471 *) Add certificate callback. If set this is called whenever a certificate
472 is required by client or server. An application can decide which
473 certificate chain to present based on arbitrary criteria: for example
474 supported signature algorithms. Add very simple example to s_server.
475 This fixes many of the problems and restrictions of the existing client
476 certificate callback: for example you can now clear an existing
477 certificate and specify the whole chain.
480 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
481 the certificate can be used for (if anything). Set valid_flags field
482 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
483 to have similar checks in it.
485 Add new "cert_flags" field to CERT structure and include a "strict mode".
486 This enforces some TLS certificate requirements (such as only permitting
487 certificate signature algorithms contained in the supported algorithms
488 extension) which some implementations ignore: this option should be used
489 with caution as it could cause interoperability issues.
492 *) Update and tidy signature algorithm extension processing. Work out
493 shared signature algorithms based on preferences and peer algorithms
494 and print them out in s_client and s_server. Abort handshake if no
495 shared signature algorithms.
498 *) Add new functions to allow customised supported signature algorithms
499 for SSL and SSL_CTX structures. Add options to s_client and s_server
503 *) New function SSL_certs_clear() to delete all references to certificates
504 from an SSL structure. Before this once a certificate had been added
505 it couldn't be removed.
508 *) Integrate hostname, email address and IP address checking with certificate
509 verification. New verify options supporting checking in opensl utility.
512 *) Fixes and wildcard matching support to hostname and email checking
513 functions. Add manual page.
514 [Florian Weimer (Red Hat Product Security Team)]
516 *) New functions to check a hostname email or IP address against a
517 certificate. Add options x509 utility to print results of checks against
521 *) Fix OCSP checking.
522 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
524 *) Initial experimental support for explicitly trusted non-root CAs.
525 OpenSSL still tries to build a complete chain to a root but if an
526 intermediate CA has a trust setting included that is used. The first
527 setting is used: whether to trust (e.g., -addtrust option to the x509
531 *) Add -trusted_first option which attempts to find certificates in the
532 trusted store even if an untrusted chain is also supplied.
535 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
536 platform support for Linux and Android.
539 *) Support for linux-x32, ILP32 environment in x86_64 framework.
542 *) Experimental multi-implementation support for FIPS capable OpenSSL.
543 When in FIPS mode the approved implementations are used as normal,
544 when not in FIPS mode the internal unapproved versions are used instead.
545 This means that the FIPS capable OpenSSL isn't forced to use the
546 (often lower perfomance) FIPS implementations outside FIPS mode.
549 *) Transparently support X9.42 DH parameters when calling
550 PEM_read_bio_DHparameters. This means existing applications can handle
551 the new parameter format automatically.
554 *) Initial experimental support for X9.42 DH parameter format: mainly
555 to support use of 'q' parameter for RFC5114 parameters.
558 *) Add DH parameters from RFC5114 including test data to dhtest.
561 *) Support for automatic EC temporary key parameter selection. If enabled
562 the most preferred EC parameters are automatically used instead of
563 hardcoded fixed parameters. Now a server just has to call:
564 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
565 support ECDH and use the most appropriate parameters.
568 *) Enhance and tidy EC curve and point format TLS extension code. Use
569 static structures instead of allocation if default values are used.
570 New ctrls to set curves we wish to support and to retrieve shared curves.
571 Print out shared curves in s_server. New options to s_server and s_client
572 to set list of supported curves.
575 *) New ctrls to retrieve supported signature algorithms and
576 supported curve values as an array of NIDs. Extend openssl utility
577 to print out received values.
580 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
581 between NIDs and the more common NIST names such as "P-256". Enhance
582 ecparam utility and ECC method to recognise the NIST names for curves.
585 *) Enhance SSL/TLS certificate chain handling to support different
586 chains for each certificate instead of one chain in the parent SSL_CTX.
589 *) Support for fixed DH ciphersuite client authentication: where both
590 server and client use DH certificates with common parameters.
593 *) Support for fixed DH ciphersuites: those requiring DH server
597 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
599 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
600 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
601 X509_CINF_get_signature were reverted post internal team review.
603 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
605 *) Build fixes for the Windows and OpenVMS platforms
606 [Matt Caswell and Richard Levitte]
608 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
610 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
611 message can cause a segmentation fault in OpenSSL due to a NULL pointer
612 dereference. This could lead to a Denial Of Service attack. Thanks to
613 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
617 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
618 dtls1_buffer_record function under certain conditions. In particular this
619 could occur if an attacker sent repeated DTLS records with the same
620 sequence number but for the next epoch. The memory leak could be exploited
621 by an attacker in a Denial of Service attack through memory exhaustion.
622 Thanks to Chris Mueller for reporting this issue.
626 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
627 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
628 method would be set to NULL which could later result in a NULL pointer
629 dereference. Thanks to Frank Schmirler for reporting this issue.
633 *) Abort handshake if server key exchange message is omitted for ephemeral
636 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
637 reporting this issue.
641 *) Remove non-export ephemeral RSA code on client and server. This code
642 violated the TLS standard by allowing the use of temporary RSA keys in
643 non-export ciphersuites and could be used by a server to effectively
644 downgrade the RSA key length used to a value smaller than the server
645 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
646 INRIA or reporting this issue.
650 *) Fixed issue where DH client certificates are accepted without verification.
651 An OpenSSL server will accept a DH certificate for client authentication
652 without the certificate verify message. This effectively allows a client to
653 authenticate without the use of a private key. This only affects servers
654 which trust a client certificate authority which issues certificates
655 containing DH keys: these are extremely rare and hardly ever encountered.
656 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
661 *) Ensure that the session ID context of an SSL is updated when its
662 SSL_CTX is updated via SSL_set_SSL_CTX.
664 The session ID context is typically set from the parent SSL_CTX,
665 and can vary with the CTX.
668 *) Fix various certificate fingerprint issues.
670 By using non-DER or invalid encodings outside the signed portion of a
671 certificate the fingerprint can be changed without breaking the signature.
672 Although no details of the signed portion of the certificate can be changed
673 this can cause problems with some applications: e.g. those using the
674 certificate fingerprint for blacklists.
676 1. Reject signatures with non zero unused bits.
678 If the BIT STRING containing the signature has non zero unused bits reject
679 the signature. All current signature algorithms require zero unused bits.
681 2. Check certificate algorithm consistency.
683 Check the AlgorithmIdentifier inside TBS matches the one in the
684 certificate signature. NB: this will result in signature failure
685 errors for some broken certificates.
687 Thanks to Konrad Kraszewski from Google for reporting this issue.
689 3. Check DSA/ECDSA signatures use DER.
691 Reencode DSA/ECDSA signatures and compare with the original received
692 signature. Return an error if there is a mismatch.
694 This will reject various cases including garbage after signature
695 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
696 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
697 (negative or with leading zeroes).
699 Further analysis was conducted and fixes were developed by Stephen Henson
700 of the OpenSSL core team.
705 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
706 results on some platforms, including x86_64. This bug occurs at random
707 with a very low probability, and is not known to be exploitable in any
708 way, though its exact impact is difficult to determine. Thanks to Pieter
709 Wuille (Blockstream) who reported this issue and also suggested an initial
710 fix. Further analysis was conducted by the OpenSSL development team and
711 Adam Langley of Google. The final fix was developed by Andy Polyakov of
712 the OpenSSL core team.
716 *) Do not resume sessions on the server if the negotiated protocol
717 version does not match the session's version. Resuming with a different
718 version, while not strictly forbidden by the RFC, is of questionable
719 sanity and breaks all known clients.
720 [David Benjamin, Emilia Käsper]
722 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
723 early CCS messages during renegotiation. (Note that because
724 renegotiation is encrypted, this early CCS was not exploitable.)
727 *) Tighten client-side session ticket handling during renegotiation:
728 ensure that the client only accepts a session ticket if the server sends
729 the extension anew in the ServerHello. Previously, a TLS client would
730 reuse the old extension state and thus accept a session ticket if one was
731 announced in the initial ServerHello.
733 Similarly, ensure that the client requires a session ticket if one
734 was advertised in the ServerHello. Previously, a TLS client would
735 ignore a missing NewSessionTicket message.
738 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
742 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
743 sends a carefully crafted handshake message, to cause OpenSSL to fail
744 to free up to 64k of memory causing a memory leak. This could be
745 exploited in a Denial Of Service attack. This issue affects OpenSSL
746 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
747 whether SRTP is used or configured. Implementations of OpenSSL that
748 have been compiled with OPENSSL_NO_SRTP defined are not affected.
750 The fix was developed by the OpenSSL team.
754 *) Session Ticket Memory Leak.
756 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
757 integrity of that ticket is first verified. In the event of a session
758 ticket integrity check failing, OpenSSL will fail to free memory
759 causing a memory leak. By sending a large number of invalid session
760 tickets an attacker could exploit this issue in a Denial Of Service
765 *) Build option no-ssl3 is incomplete.
767 When OpenSSL is configured with "no-ssl3" as a build option, servers
768 could accept and complete a SSL 3.0 handshake, and clients could be
769 configured to send them.
771 [Akamai and the OpenSSL team]
773 *) Add support for TLS_FALLBACK_SCSV.
774 Client applications doing fallback retries should call
775 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
777 [Adam Langley, Bodo Moeller]
779 *) Add additional DigestInfo checks.
781 Reencode DigestInto in DER and check against the original when
782 verifying RSA signature: this will reject any improperly encoded
783 DigestInfo structures.
785 Note: this is a precautionary measure and no attacks are currently known.
789 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
791 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
792 SRP code can be overrun an internal buffer. Add sanity check that
793 g, A, B < N to SRP code.
795 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
796 Group for discovering this issue.
800 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
801 TLS 1.0 instead of higher protocol versions when the ClientHello message
802 is badly fragmented. This allows a man-in-the-middle attacker to force a
803 downgrade to TLS 1.0 even if both the server and the client support a
804 higher protocol version, by modifying the client's TLS records.
806 Thanks to David Benjamin and Adam Langley (Google) for discovering and
807 researching this issue.
811 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
812 to a denial of service attack. A malicious server can crash the client
813 with a null pointer dereference (read) by specifying an anonymous (EC)DH
814 ciphersuite and sending carefully crafted handshake messages.
816 Thanks to Felix Gröbert (Google) for discovering and researching this
821 *) By sending carefully crafted DTLS packets an attacker could cause openssl
822 to leak memory. This can be exploited through a Denial of Service attack.
823 Thanks to Adam Langley for discovering and researching this issue.
827 *) An attacker can force openssl to consume large amounts of memory whilst
828 processing DTLS handshake messages. This can be exploited through a
829 Denial of Service attack.
830 Thanks to Adam Langley for discovering and researching this issue.
834 *) An attacker can force an error condition which causes openssl to crash
835 whilst processing DTLS packets due to memory being freed twice. This
836 can be exploited through a Denial of Service attack.
837 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
842 *) If a multithreaded client connects to a malicious server using a resumed
843 session and the server sends an ec point format extension it could write
844 up to 255 bytes to freed memory.
846 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
851 *) A malicious server can crash an OpenSSL client with a null pointer
852 dereference (read) by specifying an SRP ciphersuite even though it was not
853 properly negotiated with the client. This can be exploited through a
854 Denial of Service attack.
856 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
857 discovering and researching this issue.
861 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
862 X509_name_oneline, X509_name_print_ex et al. to leak some information
863 from the stack. Applications may be affected if they echo pretty printing
864 output to the attacker.
866 Thanks to Ivan Fratric (Google) for discovering this issue.
868 [Emilia Käsper, and Steve Henson]
870 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
871 for corner cases. (Certain input points at infinity could lead to
872 bogus results, with non-infinity inputs mapped to infinity too.)
875 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
877 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
878 handshake can force the use of weak keying material in OpenSSL
879 SSL/TLS clients and servers.
881 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
882 researching this issue. (CVE-2014-0224)
883 [KIKUCHI Masashi, Steve Henson]
885 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
886 OpenSSL DTLS client the code can be made to recurse eventually crashing
889 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
891 [Imre Rad, Steve Henson]
893 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
894 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
895 client or server. This is potentially exploitable to run arbitrary
896 code on a vulnerable client or server.
898 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
899 [Jüri Aedla, Steve Henson]
901 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
902 are subject to a denial of service attack.
904 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
905 this issue. (CVE-2014-3470)
906 [Felix Gröbert, Ivan Fratric, Steve Henson]
908 *) Harmonize version and its documentation. -f flag is used to display
910 [mancha <mancha1@zoho.com>]
912 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
914 [mancha <mancha1@zoho.com>]
916 *) Fix some double frees. These are not thought to be exploitable.
917 [mancha <mancha1@zoho.com>]
919 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
921 *) A missing bounds check in the handling of the TLS heartbeat extension
922 can be used to reveal up to 64k of memory to a connected client or
925 Thanks for Neel Mehta of Google Security for discovering this bug and to
926 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
927 preparing the fix (CVE-2014-0160)
928 [Adam Langley, Bodo Moeller]
930 *) Fix for the attack described in the paper "Recovering OpenSSL
931 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
932 by Yuval Yarom and Naomi Benger. Details can be obtained from:
933 http://eprint.iacr.org/2014/140
935 Thanks to Yuval Yarom and Naomi Benger for discovering this
936 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
937 [Yuval Yarom and Naomi Benger]
939 *) TLS pad extension: draft-agl-tls-padding-03
941 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
942 TLS client Hello record length value would otherwise be > 255 and
943 less that 512 pad with a dummy extension containing zeroes so it
944 is at least 512 bytes long.
946 [Adam Langley, Steve Henson]
948 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
950 *) Fix for TLS record tampering bug. A carefully crafted invalid
951 handshake could crash OpenSSL with a NULL pointer exception.
952 Thanks to Anton Johansson for reporting this issues.
955 *) Keep original DTLS digest and encryption contexts in retransmission
956 structures so we can use the previous session parameters if they need
957 to be resent. (CVE-2013-6450)
960 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
961 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
962 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
963 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
964 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
965 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
966 [Rob Stradling, Adam Langley]
968 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
970 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
971 supporting platforms or when small records were transferred.
972 [Andy Polyakov, Steve Henson]
974 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
976 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
978 This addresses the flaw in CBC record processing discovered by
979 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
980 at: http://www.isg.rhul.ac.uk/tls/
982 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
983 Security Group at Royal Holloway, University of London
984 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
985 Emilia Käsper for the initial patch.
987 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
989 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
990 ciphersuites which can be exploited in a denial of service attack.
991 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
992 and detecting this bug and to Wolfgang Ettlinger
993 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
997 *) Return an error when checking OCSP signatures when key is NULL.
998 This fixes a DoS attack. (CVE-2013-0166)
1001 *) Make openssl verify return errors.
1002 [Chris Palmer <palmer@google.com> and Ben Laurie]
1004 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1005 the right response is stapled. Also change SSL_get_certificate()
1006 so it returns the certificate actually sent.
1007 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1008 [Rob Stradling <rob.stradling@comodo.com>]
1010 *) Fix possible deadlock when decoding public keys.
1013 *) Don't use TLS 1.0 record version number in initial client hello
1017 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1019 *) Sanity check record length before skipping explicit IV in TLS
1020 1.2, 1.1 and DTLS to fix DoS attack.
1022 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1023 fuzzing as a service testing platform.
1027 *) Initialise tkeylen properly when encrypting CMS messages.
1028 Thanks to Solar Designer of Openwall for reporting this issue.
1031 *) In FIPS mode don't try to use composite ciphers as they are not
1035 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1037 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1038 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1039 mean any application compiled against OpenSSL 1.0.0 headers setting
1040 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1041 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1042 0x10000000L Any application which was previously compiled against
1043 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1044 will need to be recompiled as a result. Letting be results in
1045 inability to disable specifically TLS 1.1 and in client context,
1046 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1049 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1050 disable just protocol X, but all protocols above X *if* there are
1051 protocols *below* X still enabled. In more practical terms it means
1052 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1053 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1054 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1058 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1060 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1061 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1062 in CRYPTO_realloc_clean.
1064 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1065 issue and to Adam Langley <agl@chromium.org> for fixing it.
1067 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1069 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1072 *) Workarounds for some broken servers that "hang" if a client hello
1073 record length exceeds 255 bytes.
1075 1. Do not use record version number > TLS 1.0 in initial client
1076 hello: some (but not all) hanging servers will now work.
1077 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1078 the number of ciphers sent in the client hello. This should be
1079 set to an even number, such as 50, for example by passing:
1080 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1081 Most broken servers should now work.
1082 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1083 TLS 1.2 client support entirely.
1086 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1089 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1091 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1092 STRING form instead of a DigestInfo.
1095 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1096 and the RSA_sign/RSA_verify functions. This was made more apparent when
1097 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1098 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1099 the correct format in RSA_verify so both forms transparently work.
1102 *) Some servers which support TLS 1.0 can choke if we initially indicate
1103 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1104 encrypted premaster secret. As a workaround use the maximum pemitted
1105 client version in client hello, this should keep such servers happy
1106 and still work with previous versions of OpenSSL.
1109 *) Add support for TLS/DTLS heartbeats.
1110 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1112 *) Add support for SCTP.
1113 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1115 *) Improved PRNG seeding for VOS.
1116 [Paul Green <Paul.Green@stratus.com>]
1118 *) Extensive assembler packs updates, most notably:
1120 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1121 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1122 - x86_64: bit-sliced AES implementation;
1123 - ARM: NEON support, contemporary platforms optimizations;
1124 - s390x: z196 support;
1125 - *: GHASH and GF(2^m) multiplication implementations;
1129 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1130 (removal of unnecessary code)
1131 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1133 *) Add TLS key material exporter from RFC 5705.
1136 *) Add DTLS-SRTP negotiation from RFC 5764.
1139 *) Add Next Protocol Negotiation,
1140 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1141 disabled with a no-npn flag to config or Configure. Code donated
1143 [Adam Langley <agl@google.com> and Ben Laurie]
1145 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1146 NIST-P256, NIST-P521, with constant-time single point multiplication on
1147 typical inputs. Compiler support for the nonstandard type __uint128_t is
1148 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1149 Code made available under Apache License version 2.0.
1151 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1152 line to include this in your build of OpenSSL, and run "make depend" (or
1153 "make update"). This enables the following EC_METHODs:
1155 EC_GFp_nistp224_method()
1156 EC_GFp_nistp256_method()
1157 EC_GFp_nistp521_method()
1159 EC_GROUP_new_by_curve_name() will automatically use these (while
1160 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1162 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1164 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1165 all platforms. Move ssize_t definition from e_os.h to the public
1166 header file e_os2.h as it now appears in public header file cms.h
1169 *) New -sigopt option to the ca, req and x509 utilities. Additional
1170 signature parameters can be passed using this option and in
1174 *) Add RSA PSS signing function. This will generate and set the
1175 appropriate AlgorithmIdentifiers for PSS based on those in the
1176 corresponding EVP_MD_CTX structure. No application support yet.
1179 *) Support for companion algorithm specific ASN1 signing routines.
1180 New function ASN1_item_sign_ctx() signs a pre-initialised
1181 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1182 the appropriate parameters.
1185 *) Add new algorithm specific ASN1 verification initialisation function
1186 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1187 handling will be the same no matter what EVP_PKEY_METHOD is used.
1188 Add a PSS handler to support verification of PSS signatures: checked
1189 against a number of sample certificates.
1192 *) Add signature printing for PSS. Add PSS OIDs.
1193 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1195 *) Add algorithm specific signature printing. An individual ASN1 method
1196 can now print out signatures instead of the standard hex dump.
1198 More complex signatures (e.g. PSS) can print out more meaningful
1199 information. Include DSA version that prints out the signature
1203 *) Password based recipient info support for CMS library: implementing
1207 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1208 neatly separates the code into cipher and PBE sections and is required
1209 for some algorithms that split PBES2 into separate pieces (such as
1210 password based CMS).
1213 *) Session-handling fixes:
1214 - Fix handling of connections that are resuming with a session ID,
1215 but also support Session Tickets.
1216 - Fix a bug that suppressed issuing of a new ticket if the client
1217 presented a ticket with an expired session.
1218 - Try to set the ticket lifetime hint to something reasonable.
1219 - Make tickets shorter by excluding irrelevant information.
1220 - On the client side, don't ignore renewed tickets.
1221 [Adam Langley, Bodo Moeller (Google)]
1223 *) Fix PSK session representation.
1226 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1228 This work was sponsored by Intel.
1231 *) Add GCM support to TLS library. Some custom code is needed to split
1232 the IV between the fixed (from PRF) and explicit (from TLS record)
1233 portions. This adds all GCM ciphersuites supported by RFC5288 and
1234 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
1235 add a special AESGCM string for GCM only.
1238 *) Expand range of ctrls for AES GCM. Permit setting invocation
1239 field on decrypt and retrieval of invocation field only on encrypt.
1242 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1243 As required by RFC5289 these ciphersuites cannot be used if for
1244 versions of TLS earlier than 1.2.
1247 *) For FIPS capable OpenSSL interpret a NULL default public key method
1248 as unset and return the appopriate default but do *not* set the default.
1249 This means we can return the appopriate method in applications that
1250 swicth between FIPS and non-FIPS modes.
1253 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1254 ENGINE is used then we cannot handle that in the FIPS module so we
1255 keep original code iff non-FIPS operations are allowed.
1258 *) Add -attime option to openssl utilities.
1259 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1261 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1264 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1265 FIPS EC methods unconditionally for now.
1268 *) New build option no-ec2m to disable characteristic 2 code.
1271 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1272 all cases can be covered as some introduce binary incompatibilities.
1275 *) Redirect RSA operations to FIPS module including keygen,
1276 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1279 *) Add similar low level API blocking to ciphers.
1282 *) Low level digest APIs are not approved in FIPS mode: any attempt
1283 to use these will cause a fatal error. Applications that *really* want
1284 to use them can use the private_* version instead.
1287 *) Redirect cipher operations to FIPS module for FIPS builds.
1290 *) Redirect digest operations to FIPS module for FIPS builds.
1293 *) Update build system to add "fips" flag which will link in fipscanister.o
1294 for static and shared library builds embedding a signature if needed.
1297 *) Output TLS supported curves in preference order instead of numerical
1298 order. This is currently hardcoded for the highest order curves first.
1299 This should be configurable so applications can judge speed vs strength.
1302 *) Add TLS v1.2 server support for client authentication.
1305 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1309 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1310 FIPS modules versions.
1313 *) Add TLS v1.2 client side support for client authentication. Keep cache
1314 of handshake records longer as we don't know the hash algorithm to use
1315 until after the certificate request message is received.
1318 *) Initial TLS v1.2 client support. Add a default signature algorithms
1319 extension including all the algorithms we support. Parse new signature
1320 format in client key exchange. Relax some ECC signing restrictions for
1321 TLS v1.2 as indicated in RFC5246.
1324 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1325 to new signature format when needed using client digest preference.
1326 All server ciphersuites should now work correctly in TLS v1.2. No client
1327 support yet and no support for client certificates.
1330 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1331 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1332 ciphersuites. At present only RSA key exchange ciphersuites work with
1333 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1334 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1335 and version checking.
1338 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1339 with this defined it will not be affected by any changes to ssl internal
1340 structures. Add several utility functions to allow openssl application
1341 to work with OPENSSL_NO_SSL_INTERN defined.
1345 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1347 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1350 *) Permit abbreviated handshakes when renegotiating using the function
1351 SSL_renegotiate_abbreviated().
1352 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1354 *) Add call to ENGINE_register_all_complete() to
1355 ENGINE_load_builtin_engines(), so some implementations get used
1356 automatically instead of needing explicit application support.
1359 *) Add support for TLS key exporter as described in RFC5705.
1360 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1362 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1363 a few changes are required:
1365 Add SSL_OP_NO_TLSv1_1 flag.
1366 Add TLSv1_1 methods.
1367 Update version checking logic to handle version 1.1.
1368 Add explicit IV handling (ported from DTLS code).
1369 Add command line options to s_client/s_server.
1372 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1374 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1375 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1376 content decryption and always return the same error. Note: this attack
1377 needs on average 2^20 messages so it only affects automated senders. The
1378 old behaviour can be reenabled in the CMS code by setting the
1379 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1380 an MMA defence is not necessary.
1381 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1382 this issue. (CVE-2012-0884)
1385 *) Fix CVE-2011-4619: make sure we really are receiving a
1386 client hello before rejecting multiple SGC restarts. Thanks to
1387 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1390 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1392 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1393 Thanks to Antonio Martin, Enterprise Secure Access Research and
1394 Development, Cisco Systems, Inc. for discovering this bug and
1395 preparing a fix. (CVE-2012-0050)
1398 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1400 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1401 of the Vaudenay padding oracle attack on CBC mode encryption
1402 which enables an efficient plaintext recovery attack against
1403 the OpenSSL implementation of DTLS. Their attack exploits timing
1404 differences arising during decryption processing. A research
1405 paper describing this attack can be found at:
1406 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1407 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1408 Security Group at Royal Holloway, University of London
1409 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1410 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1411 for preparing the fix. (CVE-2011-4108)
1412 [Robin Seggelmann, Michael Tuexen]
1414 *) Clear bytes used for block padding of SSL 3.0 records.
1416 [Adam Langley (Google)]
1418 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1419 Kadianakis <desnacked@gmail.com> for discovering this issue and
1420 Adam Langley for preparing the fix. (CVE-2011-4619)
1421 [Adam Langley (Google)]
1423 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1424 [Andrey Kulikov <amdeich@gmail.com>]
1426 *) Prevent malformed RFC3779 data triggering an assertion failure.
1427 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1428 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1429 [Rob Austein <sra@hactrn.net>]
1431 *) Improved PRNG seeding for VOS.
1432 [Paul Green <Paul.Green@stratus.com>]
1434 *) Fix ssl_ciph.c set-up race.
1435 [Adam Langley (Google)]
1437 *) Fix spurious failures in ecdsatest.c.
1438 [Emilia Käsper (Google)]
1440 *) Fix the BIO_f_buffer() implementation (which was mixing different
1441 interpretations of the '..._len' fields).
1442 [Adam Langley (Google)]
1444 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1445 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1446 threads won't reuse the same blinding coefficients.
1448 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1449 lock to call BN_BLINDING_invert_ex, and avoids one use of
1450 BN_BLINDING_update for each BN_BLINDING structure (previously,
1451 the last update always remained unused).
1452 [Emilia Käsper (Google)]
1454 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1455 [Bob Buckholz (Google)]
1457 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1459 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1460 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1461 [Kaspar Brand <ossl@velox.ch>]
1463 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1464 for multi-threaded use of ECDH. (CVE-2011-3210)
1465 [Adam Langley (Google)]
1467 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1470 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1471 signature public key algorithm by using OID xref utilities instead.
1472 Before this you could only use some ECC ciphersuites with SHA1 only.
1475 *) Add protection against ECDSA timing attacks as mentioned in the paper
1476 by Billy Bob Brumley and Nicola Tuveri, see:
1478 http://eprint.iacr.org/2011/232.pdf
1480 [Billy Bob Brumley and Nicola Tuveri]
1482 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1484 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1485 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1487 *) Fix bug in string printing code: if *any* escaping is enabled we must
1488 escape the escape character (backslash) or the resulting string is
1492 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1494 *) Disable code workaround for ancient and obsolete Netscape browsers
1495 and servers: an attacker can use it in a ciphersuite downgrade attack.
1496 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1499 *) Fixed J-PAKE implementation error, originally discovered by
1500 Sebastien Martini, further info and confirmation from Stefan
1501 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1504 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1506 *) Fix extension code to avoid race conditions which can result in a buffer
1507 overrun vulnerability: resumed sessions must not be modified as they can
1508 be shared by multiple threads. CVE-2010-3864
1511 *) Fix WIN32 build system to correctly link an ENGINE directory into
1515 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1517 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1519 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1521 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1523 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1524 context. The operation can be customised via the ctrl mechanism in
1525 case ENGINEs want to include additional functionality.
1528 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1531 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1532 output hashes compatible with older versions of OpenSSL.
1533 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1535 *) Fix compression algorithm handling: if resuming a session use the
1536 compression algorithm of the resumed session instead of determining
1537 it from client hello again. Don't allow server to change algorithm.
1540 *) Add load_crls() function to apps tidying load_certs() too. Add option
1541 to verify utility to allow additional CRLs to be included.
1544 *) Update OCSP request code to permit adding custom headers to the request:
1545 some responders need this.
1548 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1550 [Julia Lawall <julia@diku.dk>]
1552 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1553 needlessly dereferenced structures, used obsolete functions and
1554 didn't handle all updated verify codes correctly.
1557 *) Disable MD2 in the default configuration.
1560 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1561 indicate the initial BIO being pushed or popped. This makes it possible
1562 to determine whether the BIO is the one explicitly called or as a result
1563 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1564 it handles reference counts correctly and doesn't zero out the I/O bio
1565 when it is not being explicitly popped. WARNING: applications which
1566 included workarounds for the old buggy behaviour will need to be modified
1567 or they could free up already freed BIOs.
1570 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1571 renaming to all platforms (within the 0.9.8 branch, this was
1572 done conditionally on Netware platforms to avoid a name clash).
1573 [Guenter <lists@gknw.net>]
1575 *) Add ECDHE and PSK support to DTLS.
1576 [Michael Tuexen <tuexen@fh-muenster.de>]
1578 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1582 *) Add "missing" function EVP_MD_flags() (without this the only way to
1583 retrieve a digest flags is by accessing the structure directly. Update
1584 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1585 or cipher is registered as in the "from" argument. Print out all
1586 registered digests in the dgst usage message instead of manually
1587 attempting to work them out.
1590 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1591 this allows the use of compression and extensions. Change default cipher
1592 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1593 by default unless an application cipher string requests it.
1596 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1597 key ids to find matching certificates and keys but some PKCS#12 files
1598 don't follow the (somewhat unwritten) rules and this strategy fails.
1599 Now just gather all certificates together and the first private key
1600 then look for the first certificate that matches the key.
1603 *) Support use of registered digest and cipher names for dgst and cipher
1604 commands instead of having to add each one as a special case. So now
1611 openssl dgst -sha256 foo
1613 and this works for ENGINE based algorithms too.
1617 *) Update Gost ENGINE to support parameter files.
1618 [Victor B. Wagner <vitus@cryptocom.ru>]
1620 *) Support GeneralizedTime in ca utility.
1621 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1623 *) Enhance the hash format used for certificate directory links. The new
1624 form uses the canonical encoding (meaning equivalent names will work
1625 even if they aren't identical) and uses SHA1 instead of MD5. This form
1626 is incompatible with the older format and as a result c_rehash should
1627 be used to rebuild symbolic links.
1630 *) Make PKCS#8 the default write format for private keys, replacing the
1631 traditional format. This form is standardised, more secure and doesn't
1632 include an implicit MD5 dependency.
1635 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1636 committed to OpenSSL should pass this lot as a minimum.
1639 *) Add session ticket override functionality for use by EAP-FAST.
1640 [Jouni Malinen <j@w1.fi>]
1642 *) Modify HMAC functions to return a value. Since these can be implemented
1643 in an ENGINE errors can occur.
1646 *) Type-checked OBJ_bsearch_ex.
1649 *) Type-checked OBJ_bsearch. Also some constification necessitated
1650 by type-checking. Still to come: TXT_DB, bsearch(?),
1651 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1655 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1656 seconds to a tm structure directly, instead of going through OS
1657 specific date routines. This avoids any issues with OS routines such
1658 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1659 and X509_time_adj_ex() to cover the extended range. The existing
1660 X509_time_adj() is still usable and will no longer have any date issues.
1663 *) Delta CRL support. New use deltas option which will attempt to locate
1664 and search any appropriate delta CRLs available.
1666 This work was sponsored by Google.
1669 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1670 code and add additional score elements. Validate alternate CRL paths
1671 as part of the CRL checking and indicate a new error "CRL path validation
1672 error" in this case. Applications wanting additional details can use
1673 the verify callback and check the new "parent" field. If this is not
1674 NULL CRL path validation is taking place. Existing applications wont
1675 see this because it requires extended CRL support which is off by
1678 This work was sponsored by Google.
1681 *) Support for freshest CRL extension.
1683 This work was sponsored by Google.
1686 *) Initial indirect CRL support. Currently only supported in the CRLs
1687 passed directly and not via lookup. Process certificate issuer
1688 CRL entry extension and lookup CRL entries by bother issuer name
1689 and serial number. Check and process CRL issuer entry in IDP extension.
1691 This work was sponsored by Google.
1694 *) Add support for distinct certificate and CRL paths. The CRL issuer
1695 certificate is validated separately in this case. Only enabled if
1696 an extended CRL support flag is set: this flag will enable additional
1697 CRL functionality in future.
1699 This work was sponsored by Google.
1702 *) Add support for policy mappings extension.
1704 This work was sponsored by Google.
1707 *) Fixes to pathlength constraint, self issued certificate handling,
1708 policy processing to align with RFC3280 and PKITS tests.
1710 This work was sponsored by Google.
1713 *) Support for name constraints certificate extension. DN, email, DNS
1714 and URI types are currently supported.
1716 This work was sponsored by Google.
1719 *) To cater for systems that provide a pointer-based thread ID rather
1720 than numeric, deprecate the current numeric thread ID mechanism and
1721 replace it with a structure and associated callback type. This
1722 mechanism allows a numeric "hash" to be extracted from a thread ID in
1723 either case, and on platforms where pointers are larger than 'long',
1724 mixing is done to help ensure the numeric 'hash' is usable even if it
1725 can't be guaranteed unique. The default mechanism is to use "&errno"
1726 as a pointer-based thread ID to distinguish between threads.
1728 Applications that want to provide their own thread IDs should now use
1729 CRYPTO_THREADID_set_callback() to register a callback that will call
1730 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1732 Note that ERR_remove_state() is now deprecated, because it is tied
1733 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1734 to free the current thread's error state should be replaced by
1735 ERR_remove_thread_state(NULL).
1737 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1738 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1739 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1740 application was previously providing a numeric thread callback that
1741 was inappropriate for distinguishing threads, then uniqueness might
1742 have been obtained with &errno that happened immediately in the
1743 intermediate development versions of OpenSSL; this is no longer the
1744 case, the numeric thread callback will now override the automatic use
1746 [Geoff Thorpe, with help from Bodo Moeller]
1748 *) Initial support for different CRL issuing certificates. This covers a
1749 simple case where the self issued certificates in the chain exist and
1750 the real CRL issuer is higher in the existing chain.
1752 This work was sponsored by Google.
1755 *) Removed effectively defunct crypto/store from the build.
1758 *) Revamp of STACK to provide stronger type-checking. Still to come:
1759 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1760 ASN1_STRING, CONF_VALUE.
1763 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1764 RAM on SSL connections. This option can save about 34k per idle SSL.
1767 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1768 STACK, TXT_DB, bsearch, qsort.
1771 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1772 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1773 support for data, signedData, compressedData, digestedData and
1774 encryptedData, envelopedData types included. Scripts to check against
1775 RFC4134 examples draft and interop and consistency checks of many
1776 content types and variants.
1779 *) Add options to enc utility to support use of zlib compression BIO.
1782 *) Extend mk1mf to support importing of options and assembly language
1783 files from Configure script, currently only included in VC-WIN32.
1784 The assembly language rules can now optionally generate the source
1785 files from the associated perl scripts.
1788 *) Implement remaining functionality needed to support GOST ciphersuites.
1789 Interop testing has been performed using CryptoPro implementations.
1790 [Victor B. Wagner <vitus@cryptocom.ru>]
1792 *) s390x assembler pack.
1795 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1799 *) Implement Opaque PRF Input TLS extension as specified in
1800 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1801 official specification yet and no extension type assignment by
1802 IANA exists, this extension (for now) will have to be explicitly
1803 enabled when building OpenSSL by providing the extension number
1804 to use. For example, specify an option
1806 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1808 to the "config" or "Configure" script to enable the extension,
1809 assuming extension number 0x9527 (which is a completely arbitrary
1810 and unofficial assignment based on the MD5 hash of the Internet
1811 Draft). Note that by doing so, you potentially lose
1812 interoperability with other TLS implementations since these might
1813 be using the same extension number for other purposes.
1815 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1816 opaque PRF input value to use in the handshake. This will create
1817 an interal copy of the length-'len' string at 'src', and will
1818 return non-zero for success.
1820 To get more control and flexibility, provide a callback function
1823 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1824 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1828 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1831 Callback function 'cb' will be called in handshakes, and is
1832 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1833 Argument 'arg' is for application purposes (the value as given to
1834 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1835 be provided to the callback function). The callback function
1836 has to return non-zero to report success: usually 1 to use opaque
1837 PRF input just if possible, or 2 to enforce use of the opaque PRF
1838 input. In the latter case, the library will abort the handshake
1839 if opaque PRF input is not successfully negotiated.
1841 Arguments 'peerinput' and 'len' given to the callback function
1842 will always be NULL and 0 in the case of a client. A server will
1843 see the client's opaque PRF input through these variables if
1844 available (NULL and 0 otherwise). Note that if the server
1845 provides an opaque PRF input, the length must be the same as the
1846 length of the client's opaque PRF input.
1848 Note that the callback function will only be called when creating
1849 a new session (session resumption can resume whatever was
1850 previously negotiated), and will not be called in SSL 2.0
1851 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1852 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1853 for applications that need to enforce opaque PRF input.
1857 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1860 [Victor B. Wagner <vitus@cryptocom.ru>]
1862 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1863 RFC4507bis. The encrypted ticket format is an encrypted encoded
1864 SSL_SESSION structure, that way new session features are automatically
1867 If a client application caches session in an SSL_SESSION structure
1868 support is transparent because tickets are now stored in the encoded
1871 The SSL_CTX structure automatically generates keys for ticket
1872 protection in servers so again support should be possible
1873 with no application modification.
1875 If a client or server wishes to disable RFC4507 support then the option
1876 SSL_OP_NO_TICKET can be set.
1878 Add a TLS extension debugging callback to allow the contents of any client
1879 or server extensions to be examined.
1881 This work was sponsored by Google.
1884 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1885 OpenSSL should now compile cleanly on gcc 4.2
1886 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1888 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1889 support including streaming MAC support: this is required for GOST
1890 ciphersuite support.
1891 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1893 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1894 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1895 to output in BER and PEM format.
1898 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1899 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1900 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1901 ENGINE support for HMAC keys which are unextractable. New -mac and
1902 -macopt options to dgst utility.
1905 *) New option -sigopt to dgst utility. Update dgst to use
1906 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1907 alternative signing paramaters such as X9.31 or PSS in the dgst
1911 *) Change ssl_cipher_apply_rule(), the internal function that does
1912 the work each time a ciphersuite string requests enabling
1913 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1914 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1915 the order of disabled ciphersuites such that those ciphersuites
1916 that most recently went from enabled to disabled not only stay
1917 in order with respect to each other, but also have higher priority
1918 than other disabled ciphersuites the next time ciphersuites are
1921 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1922 the same ciphersuites as with "HIGH" alone, but in a specific
1923 order where the PSK ciphersuites come first (since they are the
1924 most recently disabled ciphersuites when "HIGH" is parsed).
1926 Also, change ssl_create_cipher_list() (using this new
1927 funcionality) such that between otherwise identical
1928 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1932 *) Change ssl_create_cipher_list() so that it automatically
1933 arranges the ciphersuites in reasonable order before starting
1934 to process the rule string. Thus, the definition for "DEFAULT"
1935 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1936 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1937 This makes it much easier to arrive at a reasonable default order
1938 in applications for which anonymous ciphers are OK (meaning
1939 that you can't actually use DEFAULT).
1940 [Bodo Moeller; suggested by Victor Duchovni]
1942 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1943 processing) into multiple integers instead of setting
1944 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1945 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1946 (These masks as well as the individual bit definitions are hidden
1947 away into the non-exported interface ssl/ssl_locl.h, so this
1948 change to the definition of the SSL_CIPHER structure shouldn't
1949 affect applications.) This give us more bits for each of these
1950 categories, so there is no longer a need to coagulate AES128 and
1951 AES256 into a single algorithm bit, and to coagulate Camellia128
1952 and Camellia256 into a single algorithm bit, which has led to all
1955 Thus, among other things, the kludge introduced in 0.9.7m and
1956 0.9.8e for masking out AES256 independently of AES128 or masking
1957 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1959 With the change, we also introduce new ciphersuite aliases that
1960 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1964 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1965 Use the leftmost N bytes of the signature input if the input is
1966 larger than the prime q (with N being the size in bytes of q).
1969 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1970 it yet and it is largely untested.
1973 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1976 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1977 some compilers (gcc 4.2 and later) reject their use. Safestack is
1978 reimplemented. Update ASN1 to avoid use of legacy functions.
1981 *) Win32/64 targets are linked with Winsock2.
1984 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1985 to external functions. This can be used to increase CRL handling
1986 efficiency especially when CRLs are very large by (for example) storing
1987 the CRL revoked certificates in a database.
1990 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1991 new CRLs added to a directory can be used. New command line option
1992 -verify_return_error to s_client and s_server. This causes real errors
1993 to be returned by the verify callback instead of carrying on no matter
1994 what. This reflects the way a "real world" verify callback would behave.
1997 *) GOST engine, supporting several GOST algorithms and public key formats.
1998 Kindly donated by Cryptocom.
2001 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2002 partitioned by DP are handled but no indirect CRL or reason partitioning
2003 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2004 selected via a scoring technique which handles IDP and AKID in CRLs.
2007 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2008 will ultimately be used for all verify operations: this will remove the
2009 X509_STORE dependency on certificate verification and allow alternative
2010 lookup methods. X509_STORE based implementations of these two callbacks.
2013 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2014 Modify get_crl() to find a valid (unexpired) CRL if possible.
2017 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2018 this would be called X509_CRL_cmp() but that name is already used by
2019 a function that just compares CRL issuer names. Cache several CRL
2020 extensions in X509_CRL structure and cache CRLDP in X509.
2023 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2024 this maps equivalent X509_NAME structures into a consistent structure.
2025 Name comparison can then be performed rapidly using memcmp().
2028 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2032 *) Allow digests to supply their own micalg string for S/MIME type using
2033 the ctrl EVP_MD_CTRL_MICALG.
2036 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2037 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2038 ctrl. It can then customise the structure before and/or after signing
2042 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2043 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2044 to free up any added signature OIDs.
2047 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2048 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2049 digest and cipher tables. New options added to openssl utility:
2050 list-message-digest-algorithms and list-cipher-algorithms.
2053 *) Change the array representation of binary polynomials: the list
2054 of degrees of non-zero coefficients is now terminated with -1.
2055 Previously it was terminated with 0, which was also part of the
2056 value; thus, the array representation was not applicable to
2057 polynomials where t^0 has coefficient zero. This change makes
2058 the array representation useful in a more general context.
2061 *) Various modifications and fixes to SSL/TLS cipher string
2062 handling. For ECC, the code now distinguishes between fixed ECDH
2063 with RSA certificates on the one hand and with ECDSA certificates
2064 on the other hand, since these are separate ciphersuites. The
2065 unused code for Fortezza ciphersuites has been removed.
2067 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2068 (not "ECDHE"). For consistency with the code for DH
2069 certificates, use of ECDH certificates is now considered ECDH
2070 authentication, not RSA or ECDSA authentication (the latter is
2071 merely the CA's signing algorithm and not actively used in the
2074 The temporary ciphersuite alias "ECCdraft" is no longer
2075 available, and ECC ciphersuites are no longer excluded from "ALL"
2076 and "DEFAULT". The following aliases now exist for RFC 4492
2077 ciphersuites, most of these by analogy with the DH case:
2079 kECDHr - ECDH cert, signed with RSA
2080 kECDHe - ECDH cert, signed with ECDSA
2081 kECDH - ECDH cert (signed with either RSA or ECDSA)
2082 kEECDH - ephemeral ECDH
2083 ECDH - ECDH cert or ephemeral ECDH
2089 AECDH - anonymous ECDH
2090 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2094 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2095 Use correct micalg parameters depending on digest(s) in signed message.
2098 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2099 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2102 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2103 an engine to register a method. Add ENGINE lookups for methods and
2104 functional reference processing.
2107 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2108 EVP_{Sign,Verify}* which allow an application to customise the signature
2112 *) New -resign option to smime utility. This adds one or more signers
2113 to an existing PKCS#7 signedData structure. Also -md option to use an
2114 alternative message digest algorithm for signing.
2117 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2118 create PKCS7 structures containing multiple signers. Update smime
2119 application to support multiple signers.
2122 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2126 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2127 Reorganize PBE internals to lookup from a static table using NIDs,
2128 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2129 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2130 PRF which will be automatically used with PBES2.
2133 *) Replace the algorithm specific calls to generate keys in "req" with the
2137 *) Update PKCS#7 enveloped data routines to use new API. This is now
2138 supported by any public key method supporting the encrypt operation. A
2139 ctrl is added to allow the public key algorithm to examine or modify
2140 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2144 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2145 a default digest type to use. In most cases this will be SHA1 but some
2146 algorithms (such as GOST) need to specify an alternative digest. The
2147 return value indicates how strong the prefernce is 1 means optional and
2148 2 is mandatory (that is it is the only supported type). Modify
2149 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2150 use the default md. Update openssl utilities to use the default digest
2151 type for signing if it is not explicitly indicated.
2154 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2155 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2156 signing method from the key type. This effectively removes the link
2157 between digests and public key types.
2160 *) Add an OID cross reference table and utility functions. Its purpose is to
2161 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2162 rsaEncryption. This will allow some of the algorithm specific hackery
2163 needed to use the correct OID to be removed.
2166 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2167 structures for PKCS7_sign(). They are now set up by the relevant public
2171 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2174 *) Add support for key derivation (agreement) in the API, DH method and
2178 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2179 public and private key formats. As a side effect these add additional
2180 command line functionality not previously available: DSA signatures can be
2181 generated and verified using pkeyutl and DH key support and generation in
2186 [Oliver Tappe <zooey@hirschkaefer.de>]
2188 *) New make target "install_html_docs" installs HTML renditions of the
2190 [Oliver Tappe <zooey@hirschkaefer.de>]
2192 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
2193 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2194 support key and parameter generation and add initial key generation
2195 functionality for RSA.
2198 *) Add functions for main EVP_PKEY_method operations. The undocumented
2199 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2200 EVP_PKEY_{encrypt,decrypt}_old.
2203 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2204 key API, doesn't do much yet.
2207 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2208 public key algorithms. New option to openssl utility:
2209 "list-public-key-algorithms" to print out info.
2212 *) Implement the Supported Elliptic Curves Extension for
2213 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2216 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2217 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2220 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2221 utilities such as rsa, dsa, dsaparam etc except they process any key
2225 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2226 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2227 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2231 *) Initial support for pluggable public key ASN1.
2232 De-spaghettify the public key ASN1 handling. Move public and private
2233 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2234 algorithm specific handling to a single module within the relevant
2235 algorithm directory. Add functions to allow (near) opaque processing
2236 of public and private key structures.
2239 *) Implement the Supported Point Formats Extension for
2240 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2243 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2244 for the psk identity [hint] and the psk callback functions to the
2245 SSL_SESSION, SSL and SSL_CTX structure.
2248 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2252 SSL_CTX_use_psk_identity_hint
2253 SSL_get_psk_identity_hint
2254 SSL_get_psk_identity
2255 SSL_use_psk_identity_hint
2257 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2259 *) Add RFC 3161 compliant time stamp request creation, response generation
2260 and response verification functionality.
2261 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2263 *) Add initial support for TLS extensions, specifically for the server_name
2264 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2265 have new members for a host name. The SSL data structure has an
2266 additional member SSL_CTX *initial_ctx so that new sessions can be
2267 stored in that context to allow for session resumption, even after the
2268 SSL has been switched to a new SSL_CTX in reaction to a client's
2269 server_name extension.
2271 New functions (subject to change):
2273 SSL_get_servername()
2274 SSL_get_servername_type()
2277 New CTRL codes and macros (subject to change):
2279 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2280 - SSL_CTX_set_tlsext_servername_callback()
2281 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2282 - SSL_CTX_set_tlsext_servername_arg()
2283 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2285 openssl s_client has a new '-servername ...' option.
2287 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2288 '-key2 ...', '-servername_fatal' (subject to change). This allows
2289 testing the HostName extension for a specific single host name ('-cert'
2290 and '-key' remain fallbacks for handshakes without HostName
2291 negotiation). If the unrecogninzed_name alert has to be sent, this by
2292 default is a warning; it becomes fatal with the '-servername_fatal'
2295 [Peter Sylvester, Remy Allais, Christophe Renou]
2297 *) Whirlpool hash implementation is added.
2300 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2301 bn(64,32). Because of instruction set limitations it doesn't have
2302 any negative impact on performance. This was done mostly in order
2303 to make it possible to share assembler modules, such as bn_mul_mont
2304 implementations, between 32- and 64-bit builds without hassle.
2307 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2308 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2312 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2313 dedicated Montgomery multiplication procedure, is introduced.
2314 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2315 "64-bit" performance on certain 32-bit targets.
2318 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2319 in SSL structures. New SSL ctrl to set maximum send fragment size.
2320 Save memory by seeting the I/O buffer sizes dynamically instead of
2321 using the maximum available value.
2324 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2325 in addition to the text details.
2328 *) Very, very preliminary EXPERIMENTAL support for printing of general
2329 ASN1 structures. This currently produces rather ugly output and doesn't
2330 handle several customised structures at all.
2333 *) Integrated support for PVK file format and some related formats such
2334 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2335 these in the 'rsa' and 'dsa' utilities.
2338 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2341 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2342 place for the (very old) "NETSCAPE" format certificates which are now
2343 handled using new ASN1 code equivalents.
2346 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2347 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2348 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2351 *) Modify CRL distribution points extension code to print out previously
2352 unsupported fields. Enhance extension setting code to allow setting of
2356 *) Add print and set support for Issuing Distribution Point CRL extension.
2359 *) Change 'Configure' script to enable Camellia by default.
2362 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2364 *) When rejecting SSL/TLS records due to an incorrect version number, never
2365 update s->server with a new major version number. As of
2366 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2367 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2368 the previous behavior could result in a read attempt at NULL when
2369 receiving specific incorrect SSL/TLS records once record payload
2370 protection is active. (CVE-2010-0740)
2371 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2373 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2374 could be crashed if the relevant tables were not present (e.g. chrooted).
2375 [Tomas Hoger <thoger@redhat.com>]
2377 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2379 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2380 [Martin Olsson, Neel Mehta]
2382 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2383 accommodate for stack sorting, always a write lock!).
2386 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2387 excessive delays in the RAND_poll(): over a minute. As a workaround
2388 include a time check in the inner Heap32Next loop too.
2391 *) The code that handled flushing of data in SSL/TLS originally used the
2392 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2393 the problem outlined in PR#1949. The fix suggested there however can
2394 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2395 of Apache). So instead simplify the code to flush unconditionally.
2396 This should be fine since flushing with no data to flush is a no op.
2399 *) Handle TLS versions 2.0 and later properly and correctly use the
2400 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2401 off ancient servers have a habit of sticking around for a while...
2404 *) Modify compression code so it frees up structures without using the
2405 ex_data callbacks. This works around a problem where some applications
2406 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2407 restarting) then use compression (e.g. SSL with compression) later.
2408 This results in significant per-connection memory leaks and
2409 has caused some security issues including CVE-2008-1678 and
2413 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2414 change when encrypting or decrypting.
2417 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2418 connect and renegotiate with servers which do not support RI.
2419 Until RI is more widely deployed this option is enabled by default.
2422 *) Add "missing" ssl ctrls to clear options and mode.
2425 *) If client attempts to renegotiate and doesn't support RI respond with
2426 a no_renegotiation alert as required by RFC5746. Some renegotiating
2427 TLS clients will continue a connection gracefully when they receive
2428 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2429 waiting for a server hello which it will never receive. Now we treat a
2430 received no_renegotiation alert as a fatal error. This is because
2431 applications requesting a renegotiation might well expect it to succeed
2432 and would have no code in place to handle the server denying it so the
2433 only safe thing to do is to terminate the connection.
2436 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2437 peer supports secure renegotiation and 0 otherwise. Print out peer
2438 renegotiation support in s_client/s_server.
2441 *) Replace the highly broken and deprecated SPKAC certification method with
2442 the updated NID creation version. This should correctly handle UTF8.
2445 *) Implement RFC5746. Re-enable renegotiation but require the extension
2446 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2447 turns out to be a bad idea. It has been replaced by
2448 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2449 SSL_CTX_set_options(). This is really not recommended unless you
2450 know what you are doing.
2451 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2453 *) Fixes to stateless session resumption handling. Use initial_ctx when
2454 issuing and attempting to decrypt tickets in case it has changed during
2455 servername handling. Use a non-zero length session ID when attempting
2456 stateless session resumption: this makes it possible to determine if
2457 a resumption has occurred immediately after receiving server hello
2458 (several places in OpenSSL subtly assume this) instead of later in
2462 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2463 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2464 fixes for a few places where the return code is not checked
2466 [Julia Lawall <julia@diku.dk>]
2468 *) Add --strict-warnings option to Configure script to include devteam
2469 warnings in other configurations.
2472 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2473 makes it possible to install openssl libraries in locations which
2474 have names other than "lib", for example "/usr/lib64" which some
2476 [Steve Henson, based on patch from Jeremy Utley]
2478 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2479 X690 8.9.12 and can produce some misleading textual output of OIDs.
2480 [Steve Henson, reported by Dan Kaminsky]
2482 *) Delete MD2 from algorithm tables. This follows the recommendation in
2483 several standards that it is not used in new applications due to
2484 several cryptographic weaknesses. For binary compatibility reasons
2485 the MD2 API is still compiled in by default.
2488 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2492 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2493 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2495 [Guenter <lists@gknw.net>]
2497 *) Fix the server certificate chain building code to use X509_verify_cert(),
2498 it used to have an ad-hoc builder which was unable to cope with anything
2499 other than a simple chain.
2500 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2502 *) Don't check self signed certificate signatures in X509_verify_cert()
2503 by default (a flag can override this): it just wastes time without
2504 adding any security. As a useful side effect self signed root CAs
2505 with non-FIPS digests are now usable in FIPS mode.
2508 *) In dtls1_process_out_of_seq_message() the check if the current message
2509 is already buffered was missing. For every new message was memory
2510 allocated, allowing an attacker to perform an denial of service attack
2511 with sending out of seq handshake messages until there is no memory
2512 left. Additionally every future messege was buffered, even if the
2513 sequence number made no sense and would be part of another handshake.
2514 So only messages with sequence numbers less than 10 in advance will be
2515 buffered. (CVE-2009-1378)
2516 [Robin Seggelmann, discovered by Daniel Mentz]
2518 *) Records are buffered if they arrive with a future epoch to be
2519 processed after finishing the corresponding handshake. There is
2520 currently no limitation to this buffer allowing an attacker to perform
2521 a DOS attack with sending records with future epochs until there is no
2522 memory left. This patch adds the pqueue_size() function to detemine
2523 the size of a buffer and limits the record buffer to 100 entries.
2525 [Robin Seggelmann, discovered by Daniel Mentz]
2527 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2528 parent structure is freed. (CVE-2009-1379)
2531 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2532 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2535 [Ilya O. <vrghost@gmail.com>]
2537 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2539 *) Disable renegotiation completely - this fixes a severe security
2540 problem (CVE-2009-3555) at the cost of breaking all
2541 renegotiation. Renegotiation can be re-enabled by setting
2542 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2543 run-time. This is really not recommended unless you know what
2547 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2549 *) Don't set val to NULL when freeing up structures, it is freed up by
2550 underlying code. If sizeof(void *) > sizeof(long) this can result in
2551 zeroing past the valid field. (CVE-2009-0789)
2552 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2554 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2555 checked correctly. This would allow some invalid signed attributes to
2556 appear to verify correctly. (CVE-2009-0591)
2557 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2559 *) Reject UniversalString and BMPString types with invalid lengths. This
2560 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2561 a legal length. (CVE-2009-0590)
2564 *) Set S/MIME signing as the default purpose rather than setting it
2565 unconditionally. This allows applications to override it at the store
2569 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2570 to handle some structures.
2573 *) Improve efficiency of mem_gets: don't search whole buffer each time
2575 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2577 *) New -hex option for openssl rand.
2580 *) Print out UTF8String and NumericString when parsing ASN1.
2583 *) Support NumericString type for name components.
2586 *) Allow CC in the environment to override the automatically chosen
2587 compiler. Note that nothing is done to ensure flags work with the
2591 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2593 *) Properly check EVP_VerifyFinal() and similar return values
2595 [Ben Laurie, Bodo Moeller, Google Security Team]
2597 *) Enable TLS extensions by default.
2600 *) Allow the CHIL engine to be loaded, whether the application is
2601 multithreaded or not. (This does not release the developer from the
2602 obligation to set up the dynamic locking callbacks.)
2603 [Sander Temme <sander@temme.net>]
2605 *) Use correct exit code if there is an error in dgst command.
2606 [Steve Henson; problem pointed out by Roland Dirlewanger]
2608 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2609 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2612 *) Add experimental JPAKE support, including demo authentication in
2613 s_client and s_server.
2616 *) Set the comparison function in v3_addr_canonize().
2617 [Rob Austein <sra@hactrn.net>]
2619 *) Add support for XMPP STARTTLS in s_client.
2620 [Philip Paeps <philip@freebsd.org>]
2622 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2623 to ensure that even with this option, only ciphersuites in the
2624 server's preference list will be accepted. (Note that the option
2625 applies only when resuming a session, so the earlier behavior was
2626 just about the algorithm choice for symmetric cryptography.)
2629 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2631 *) Fix NULL pointer dereference if a DTLS server received
2632 ChangeCipherSpec as first record (CVE-2009-1386).
2635 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2636 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2639 *) The fix in 0.9.8c that supposedly got rid of unsafe
2640 double-checked locking was incomplete for RSA blinding,
2641 addressing just one layer of what turns out to have been
2642 doubly unsafe triple-checked locking.
2644 So now fix this for real by retiring the MONT_HELPER macro
2645 in crypto/rsa/rsa_eay.c.
2647 [Bodo Moeller; problem pointed out by Marius Schilder]
2649 *) Various precautionary measures:
2651 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2653 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2654 (NB: This would require knowledge of the secret session ticket key
2655 to exploit, in which case you'd be SOL either way.)
2657 - Change bn_nist.c so that it will properly handle input BIGNUMs
2658 outside the expected range.
2660 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2663 [Neel Mehta, Bodo Moeller]
2665 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2666 the load fails. Useful for distros.
2667 [Ben Laurie and the FreeBSD team]
2669 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2672 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2675 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2677 This work was sponsored by Logica.
2680 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2681 keystores. Support for SSL/TLS client authentication too.
2682 Not compiled unless enable-capieng specified to Configure.
2684 This work was sponsored by Logica.
2687 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2688 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2689 attribute creation routines such as certifcate requests and PKCS#12
2693 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2695 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2696 handshake which could lead to a cilent crash as found using the
2697 Codenomicon TLS test suite (CVE-2008-1672)
2698 [Steve Henson, Mark Cox]
2700 *) Fix double free in TLS server name extensions which could lead to
2701 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2704 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2706 Clear the error queue to ensure that error entries left from
2707 older function calls do not interfere with the correct operation.
2708 [Lutz Jaenicke, Erik de Castro Lopo]
2710 *) Remove root CA certificates of commercial CAs:
2712 The OpenSSL project does not recommend any specific CA and does not
2713 have any policy with respect to including or excluding any CA.
2714 Therefore it does not make any sense to ship an arbitrary selection
2715 of root CA certificates with the OpenSSL software.
2718 *) RSA OAEP patches to fix two separate invalid memory reads.
2719 The first one involves inputs when 'lzero' is greater than
2720 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2721 before the beginning of from). The second one involves inputs where
2722 the 'db' section contains nothing but zeroes (there is a one-byte
2723 invalid read after the end of 'db').
2724 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2726 *) Partial backport from 0.9.9-dev:
2728 Introduce bn_mul_mont (dedicated Montgomery multiplication
2729 procedure) as a candidate for BIGNUM assembler implementation.
2730 While 0.9.9-dev uses assembler for various architectures, only
2731 x86_64 is available by default here in the 0.9.8 branch, and
2732 32-bit x86 is available through a compile-time setting.
2734 To try the 32-bit x86 assembler implementation, use Configure
2735 option "enable-montasm" (which exists only for this backport).
2737 As "enable-montasm" for 32-bit x86 disclaims code stability
2738 anyway, in this constellation we activate additional code
2739 backported from 0.9.9-dev for further performance improvements,
2740 namely BN_from_montgomery_word. (To enable this otherwise,
2741 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2743 [Andy Polyakov (backport partially by Bodo Moeller)]
2745 *) Add TLS session ticket callback. This allows an application to set
2746 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2747 values. This is useful for key rollover for example where several key
2748 sets may exist with different names.
2751 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2752 This was broken until now in 0.9.8 releases, such that the only way
2753 a registered ENGINE could be used (assuming it initialises
2754 successfully on the host) was to explicitly set it as the default
2755 for the relevant algorithms. This is in contradiction with 0.9.7
2756 behaviour and the documentation. With this fix, when an ENGINE is
2757 registered into a given algorithm's table of implementations, the
2758 'uptodate' flag is reset so that auto-discovery will be used next
2759 time a new context for that algorithm attempts to select an
2761 [Ian Lister (tweaked by Geoff Thorpe)]
2763 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2764 implemention in the following ways:
2766 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2769 Lack of BER streaming support means one pass streaming processing is
2770 only supported if data is detached: setting the streaming flag is
2771 ignored for embedded content.
2773 CMS support is disabled by default and must be explicitly enabled
2774 with the enable-cms configuration option.
2777 *) Update the GMP engine glue to do direct copies between BIGNUM and
2778 mpz_t when openssl and GMP use the same limb size. Otherwise the
2779 existing "conversion via a text string export" trick is still used.
2780 [Paul Sheer <paulsheer@gmail.com>]
2782 *) Zlib compression BIO. This is a filter BIO which compressed and
2783 uncompresses any data passed through it.
2786 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2787 RFC3394 compatible AES key wrapping.
2790 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2791 sets string data without copying. X509_ALGOR_set0() and
2792 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2793 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2794 from an X509_ATTRIBUTE structure optionally checking it occurs only
2795 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2799 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2800 to get the expected BN_FLG_CONSTTIME behavior.
2801 [Bodo Moeller (Google)]
2805 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2806 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2807 - added some more tests to do_tests.pl
2808 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2809 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2810 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2811 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2812 - various changes to netware.pl to enable gcc-cross builds on Win32
2814 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2815 - various changes to fix missing prototype warnings
2816 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2817 - added AES, WHIRLPOOL and CPUID assembler code to build files
2818 - added missing AES assembler make rules to mk1mf.pl
2819 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2820 [Guenter Knauf <eflash@gmx.net>]
2822 *) Implement certificate status request TLS extension defined in RFC3546.
2823 A client can set the appropriate parameters and receive the encoded
2824 OCSP response via a callback. A server can query the supplied parameters
2825 and set the encoded OCSP response in the callback. Add simplified examples
2826 to s_client and s_server.
2829 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2831 *) Fix various bugs:
2832 + Binary incompatibility of ssl_ctx_st structure
2833 + DTLS interoperation with non-compliant servers
2834 + Don't call get_session_cb() without proposed session
2835 + Fix ia64 assembler code
2836 [Andy Polyakov, Steve Henson]
2838 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2840 *) DTLS Handshake overhaul. There were longstanding issues with
2841 OpenSSL DTLS implementation, which were making it impossible for
2842 RFC 4347 compliant client to communicate with OpenSSL server.
2843 Unfortunately just fixing these incompatibilities would "cut off"
2844 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2845 server keeps tolerating non RFC compliant syntax. The opposite is
2846 not true, 0.9.8f client can not communicate with earlier server.
2847 This update even addresses CVE-2007-4995.
2850 *) Changes to avoid need for function casts in OpenSSL: some compilers
2851 (gcc 4.2 and later) reject their use.
2852 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2855 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2856 RFC4507bis. The encrypted ticket format is an encrypted encoded
2857 SSL_SESSION structure, that way new session features are automatically
2860 If a client application caches session in an SSL_SESSION structure
2861 support is transparent because tickets are now stored in the encoded
2864 The SSL_CTX structure automatically generates keys for ticket
2865 protection in servers so again support should be possible
2866 with no application modification.
2868 If a client or server wishes to disable RFC4507 support then the option
2869 SSL_OP_NO_TICKET can be set.
2871 Add a TLS extension debugging callback to allow the contents of any client
2872 or server extensions to be examined.
2874 This work was sponsored by Google.
2877 *) Add initial support for TLS extensions, specifically for the server_name
2878 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2879 have new members for a host name. The SSL data structure has an
2880 additional member SSL_CTX *initial_ctx so that new sessions can be
2881 stored in that context to allow for session resumption, even after the
2882 SSL has been switched to a new SSL_CTX in reaction to a client's
2883 server_name extension.
2885 New functions (subject to change):
2887 SSL_get_servername()
2888 SSL_get_servername_type()
2891 New CTRL codes and macros (subject to change):
2893 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2894 - SSL_CTX_set_tlsext_servername_callback()
2895 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2896 - SSL_CTX_set_tlsext_servername_arg()
2897 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2899 openssl s_client has a new '-servername ...' option.
2901 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2902 '-key2 ...', '-servername_fatal' (subject to change). This allows
2903 testing the HostName extension for a specific single host name ('-cert'
2904 and '-key' remain fallbacks for handshakes without HostName
2905 negotiation). If the unrecogninzed_name alert has to be sent, this by
2906 default is a warning; it becomes fatal with the '-servername_fatal'
2909 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2911 *) Add AES and SSE2 assembly language support to VC++ build.
2914 *) Mitigate attack on final subtraction in Montgomery reduction.
2917 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2918 (which previously caused an internal error).
2921 *) Squeeze another 10% out of IGE mode when in != out.
2924 *) AES IGE mode speedup.
2925 [Dean Gaudet (Google)]
2927 *) Add the Korean symmetric 128-bit cipher SEED (see
2928 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2929 add SEED ciphersuites from RFC 4162:
2931 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2932 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2933 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2934 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2936 To minimize changes between patchlevels in the OpenSSL 0.9.8
2937 series, SEED remains excluded from compilation unless OpenSSL
2938 is configured with 'enable-seed'.
2939 [KISA, Bodo Moeller]
2941 *) Mitigate branch prediction attacks, which can be practical if a
2942 single processor is shared, allowing a spy process to extract
2943 information. For detailed background information, see
2944 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2945 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2946 and Necessary Software Countermeasures"). The core of the change
2947 are new versions BN_div_no_branch() and
2948 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2949 respectively, which are slower, but avoid the security-relevant
2950 conditional branches. These are automatically called by BN_div()
2951 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2952 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2953 remove a conditional branch.
2955 BN_FLG_CONSTTIME is the new name for the previous
2956 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2957 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2958 in the exponent causes BN_mod_exp_mont() to use the alternative
2959 implementation in BN_mod_exp_mont_consttime().) The old name
2960 remains as a deprecated alias.
2962 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2963 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2964 constant-time implementations for more than just exponentiation.
2965 Here too the old name is kept as a deprecated alias.
2967 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2968 the BN_BLINDING structure gets an independent copy of the
2969 modulus. This means that the previous "BIGNUM *m" argument to
2970 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2971 essentially becomes "const BIGNUM *m", although we can't actually
2972 change this in the header file before 0.9.9. It allows
2973 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2974 enable BN_FLG_CONSTTIME.
2976 [Matthew D Wood (Intel Corp)]
2978 *) In the SSL/TLS server implementation, be strict about session ID
2979 context matching (which matters if an application uses a single
2980 external cache for different purposes). Previously,
2981 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2982 set. This did ensure strict client verification, but meant that,
2983 with applications using a single external cache for quite
2984 different requirements, clients could circumvent ciphersuite
2985 restrictions for a given session ID context by starting a session
2986 in a different context.
2989 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2990 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2991 authentication-only ciphersuites.
2994 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2995 not complete and could lead to a possible single byte overflow
2996 (CVE-2007-5135) [Ben Laurie]
2998 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3000 *) Since AES128 and AES256 (and similarly Camellia128 and
3001 Camellia256) share a single mask bit in the logic of
3002 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3003 kludge to work properly if AES128 is available and AES256 isn't
3004 (or if Camellia128 is available and Camellia256 isn't).
3007 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3008 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3009 When a point or a seed is encoded in a BIT STRING, we need to
3010 prevent the removal of trailing zero bits to get the proper DER
3011 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3012 of a NamedBitList, for which trailing 0 bits need to be removed.)
3015 *) Have SSL/TLS server implementation tolerate "mismatched" record
3016 protocol version while receiving ClientHello even if the
3017 ClientHello is fragmented. (The server can't insist on the
3018 particular protocol version it has chosen before the ServerHello
3019 message has informed the client about his choice.)
3022 *) Add RFC 3779 support.
3023 [Rob Austein for ARIN, Ben Laurie]
3025 *) Load error codes if they are not already present instead of using a
3026 static variable. This allows them to be cleanly unloaded and reloaded.
3027 Improve header file function name parsing.
3030 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3031 or CAPABILITY handshake as required by RFCs.
3034 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3036 *) Introduce limits to prevent malicious keys being able to
3037 cause a denial of service. (CVE-2006-2940)
3038 [Steve Henson, Bodo Moeller]
3040 *) Fix ASN.1 parsing of certain invalid structures that can result
3041 in a denial of service. (CVE-2006-2937) [Steve Henson]
3043 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3044 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3046 *) Fix SSL client code which could crash if connecting to a
3047 malicious SSLv2 server. (CVE-2006-4343)
3048 [Tavis Ormandy and Will Drewry, Google Security Team]
3050 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3051 match only those. Before that, "AES256-SHA" would be interpreted
3052 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3053 the same strength classification in 0.9.7h) as we currently only
3054 have a single AES bit in the ciphersuite description bitmap.
3055 That change, however, also applied to ciphersuite strings such as
3056 "RC4-MD5" that intentionally matched multiple ciphersuites --
3057 namely, SSL 2.0 ciphersuites in addition to the more common ones
3058 from SSL 3.0/TLS 1.0.
3060 So we change the selection algorithm again: Naming an explicit
3061 ciphersuite selects this one ciphersuite, and any other similar
3062 ciphersuite (same bitmap) from *other* protocol versions.
3063 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3064 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3066 Since SSL 2.0 does not have any ciphersuites for which the
3067 128/256 bit distinction would be relevant, this works for now.
3068 The proper fix will be to use different bits for AES128 and
3069 AES256, which would have avoided the problems from the beginning;
3070 however, bits are scarce, so we can only do this in a new release
3071 (not just a patchlevel) when we can change the SSL_CIPHER
3072 definition to split the single 'unsigned long mask' bitmap into
3073 multiple values to extend the available space.
3077 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3079 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3080 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3082 *) Add AES IGE and biIGE modes.
3085 *) Change the Unix randomness entropy gathering to use poll() when
3086 possible instead of select(), since the latter has some
3087 undesirable limitations.
3088 [Darryl Miles via Richard Levitte and Bodo Moeller]
3090 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3091 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3092 cannot be implicitly activated as part of, e.g., the "AES" alias.
3093 However, please upgrade to OpenSSL 0.9.9[-dev] for
3094 non-experimental use of the ECC ciphersuites to get TLS extension
3095 support, which is required for curve and point format negotiation
3096 to avoid potential handshake problems.
3099 *) Disable rogue ciphersuites:
3101 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3102 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3103 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3105 The latter two were purportedly from
3106 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3109 Also deactivate the remaining ciphersuites from
3110 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3111 unofficial, and the ID has long expired.
3114 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3115 dual-core machines) and other potential thread-safety issues.
3118 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3119 versions), which is now available for royalty-free use
3120 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3121 Also, add Camellia TLS ciphersuites from RFC 4132.
3123 To minimize changes between patchlevels in the OpenSSL 0.9.8
3124 series, Camellia remains excluded from compilation unless OpenSSL
3125 is configured with 'enable-camellia'.
3128 *) Disable the padding bug check when compression is in use. The padding
3129 bug check assumes the first packet is of even length, this is not
3130 necessarily true if compresssion is enabled and can result in false
3131 positives causing handshake failure. The actual bug test is ancient
3132 code so it is hoped that implementations will either have fixed it by
3133 now or any which still have the bug do not support compression.
3136 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3138 *) When applying a cipher rule check to see if string match is an explicit
3139 cipher suite and only match that one cipher suite if it is.
3142 *) Link in manifests for VC++ if needed.
3143 [Austin Ziegler <halostatue@gmail.com>]
3145 *) Update support for ECC-based TLS ciphersuites according to
3146 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3147 TLS extensions, which are supported starting with the 0.9.9
3148 branch, not in the OpenSSL 0.9.8 branch).
3151 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3152 opaque EVP_CIPHER_CTX handling.
3155 *) Fixes and enhancements to zlib compression code. We now only use
3156 "zlib1.dll" and use the default __cdecl calling convention on Win32
3157 to conform with the standards mentioned here:
3158 http://www.zlib.net/DLL_FAQ.txt
3159 Static zlib linking now works on Windows and the new --with-zlib-include
3160 --with-zlib-lib options to Configure can be used to supply the location
3161 of the headers and library. Gracefully handle case where zlib library
3165 *) Several fixes and enhancements to the OID generation code. The old code
3166 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3167 handle numbers larger than ULONG_MAX, truncated printing and had a
3168 non standard OBJ_obj2txt() behaviour.
3171 *) Add support for building of engines under engine/ as shared libraries
3172 under VC++ build system.
3175 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3176 Hopefully, we will not see any false combination of paths any more.
3179 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3181 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3182 (part of SSL_OP_ALL). This option used to disable the
3183 countermeasure against man-in-the-middle protocol-version
3184 rollback in the SSL 2.0 server implementation, which is a bad
3185 idea. (CVE-2005-2969)
3187 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3188 for Information Security, National Institute of Advanced Industrial
3189 Science and Technology [AIST], Japan)]
3191 *) Add two function to clear and return the verify parameter flags.
3194 *) Keep cipherlists sorted in the source instead of sorting them at
3195 runtime, thus removing the need for a lock.
3198 *) Avoid some small subgroup attacks in Diffie-Hellman.
3199 [Nick Mathewson and Ben Laurie]
3201 *) Add functions for well-known primes.
3204 *) Extended Windows CE support.
3205 [Satoshi Nakamura and Andy Polyakov]
3207 *) Initialize SSL_METHOD structures at compile time instead of during
3208 runtime, thus removing the need for a lock.
3211 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3212 attempting to decrypt each encrypted key in turn. Add support to
3216 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3218 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3221 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3224 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3225 key into the same file any more.
3228 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3231 *) Add -utf8 command line and config file option to 'ca'.
3232 [Stefan <stf@udoma.org]
3234 *) Removed the macro des_crypt(), as it seems to conflict with some
3235 libraries. Use DES_crypt().
3238 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3239 involves renaming the source and generated shared-libs for
3240 both. The engines will accept the corrected or legacy ids
3241 ('ncipher' and '4758_cca' respectively) when binding. NB,
3242 this only applies when building 'shared'.
3243 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3245 *) Add attribute functions to EVP_PKEY structure. Modify
3246 PKCS12_create() to recognize a CSP name attribute and
3247 use it. Make -CSP option work again in pkcs12 utility.
3250 *) Add new functionality to the bn blinding code:
3251 - automatic re-creation of the BN_BLINDING parameters after
3252 a fixed number of uses (currently 32)
3253 - add new function for parameter creation
3254 - introduce flags to control the update behaviour of the
3255 BN_BLINDING parameters
3256 - hide BN_BLINDING structure
3257 Add a second BN_BLINDING slot to the RSA structure to improve
3258 performance when a single RSA object is shared among several
3262 *) Add support for DTLS.
3263 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3265 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3266 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3269 *) Remove buggy and incompletet DH cert support from
3270 ssl/ssl_rsa.c and ssl/s3_both.c
3273 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3274 the apps/openssl applications.
3277 *) Compile clean with "-Wall -Wmissing-prototypes
3278 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3279 DEBUG_SAFESTACK must also be set.
3282 *) Change ./Configure so that certain algorithms can be disabled by default.
3283 The new counterpiece to "no-xxx" is "enable-xxx".
3285 The patented RC5 and MDC2 algorithms will now be disabled unless
3286 "enable-rc5" and "enable-mdc2", respectively, are specified.
3288 (IDEA remains enabled despite being patented. This is because IDEA
3289 is frequently required for interoperability, and there is no license
3290 fee for non-commercial use. As before, "no-idea" can be used to
3291 avoid this algorithm.)
3295 *) Add processing of proxy certificates (see RFC 3820). This work was
3296 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3297 EGEE (Enabling Grids for E-science in Europe).
3300 *) RC4 performance overhaul on modern architectures/implementations, such
3301 as Intel P4, IA-64 and AMD64.
3304 *) New utility extract-section.pl. This can be used specify an alternative
3305 section number in a pod file instead of having to treat each file as
3306 a separate case in Makefile. This can be done by adding two lines to the
3309 =for comment openssl_section:XXX
3311 The blank line is mandatory.
3315 *) New arguments -certform, -keyform and -pass for s_client and s_server
3316 to allow alternative format key and certificate files and passphrase
3320 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3321 update associated structures and add various utility functions.
3323 Add new policy related verify parameters, include policy checking in
3324 standard verify code. Enhance 'smime' application with extra parameters
3325 to support policy checking and print out.
3328 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3329 Nehemiah processors. These extensions support AES encryption in hardware
3330 as well as RNG (though RNG support is currently disabled).
3331 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3333 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3336 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3337 [Andy Polyakov and a number of other people]
3339 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3340 implementation contributed by IBM.
3341 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3343 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3344 exponent rather than 'unsigned long'. There is a corresponding change to
3345 the new 'rsa_keygen' element of the RSA_METHOD structure.
3346 [Jelte Jansen, Geoff Thorpe]
3348 *) Functionality for creating the initial serial number file is now
3349 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3351 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3352 number file to 1, which is bound to cause problems. To avoid
3353 the problems while respecting compatibility between different 0.9.7
3354 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3355 CA.pl for serial number initialization. With the new release 0.9.8,
3356 we can fix the problem directly in the 'ca' utility.)
3359 *) Reduced header interdepencies by declaring more opaque objects in
3360 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3361 give fewer recursive includes, which could break lazy source code - so
3362 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3363 developers should define this symbol when building and using openssl to
3364 ensure they track the recommended behaviour, interfaces, [etc], but
3365 backwards-compatible behaviour prevails when this isn't defined.
3368 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3371 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3372 This will generate a random key of the appropriate length based on the
3373 cipher context. The EVP_CIPHER can provide its own random key generation
3374 routine to support keys of a specific form. This is used in the des and
3375 3des routines to generate a key of the correct parity. Update S/MIME
3376 code to use new functions and hence generate correct parity DES keys.
3377 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3378 valid (weak or incorrect parity).
3381 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3382 as looking them up. This is useful when the verified structure may contain
3383 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3384 present unless the new PKCS7_NO_CRL flag is asserted.
3387 *) Extend ASN1 oid configuration module. It now additionally accepts the
3390 shortName = some long name, 1.2.3.4
3393 *) Reimplemented the BN_CTX implementation. There is now no more static
3394 limitation on the number of variables it can handle nor the depth of the
3395 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3396 information can now expand as required, and rather than having a single
3397 static array of bignums, BN_CTX now uses a linked-list of such arrays
3398 allowing it to expand on demand whilst maintaining the usefulness of
3399 BN_CTX's "bundling".
3402 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3403 to allow all RSA operations to function using a single BN_CTX.
3406 *) Preliminary support for certificate policy evaluation and checking. This
3407 is initially intended to pass the tests outlined in "Conformance Testing
3408 of Relying Party Client Certificate Path Processing Logic" v1.07.
3411 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3412 remained unused and not that useful. A variety of other little bignum
3413 tweaks and fixes have also been made continuing on from the audit (see
3417 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3418 associated ASN1, EVP and SSL functions and old ASN1 macros.
3421 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3422 and this should never fail. So the return value from the use of
3423 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3424 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3427 *) BN_CTX_get() should return zero-valued bignums, providing the same
3428 initialised value as BN_new().
3429 [Geoff Thorpe, suggested by Ulf Möller]
3431 *) Support for inhibitAnyPolicy certificate extension.
3434 *) An audit of the BIGNUM code is underway, for which debugging code is
3435 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3436 is considered valid when processing BIGNUMs, and causes execution to
3437 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3438 further steps are taken to deliberately pollute unused data in BIGNUM
3439 structures to try and expose faulty code further on. For now, openssl will
3440 (in its default mode of operation) continue to tolerate the inconsistent
3441 forms that it has tolerated in the past, but authors and packagers should
3442 consider trying openssl and their own applications when compiled with
3443 these debugging symbols defined. It will help highlight potential bugs in
3444 their own code, and will improve the test coverage for OpenSSL itself. At
3445 some point, these tighter rules will become openssl's default to improve
3446 maintainability, though the assert()s and other overheads will remain only
3447 in debugging configurations. See bn.h for more details.
3448 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3450 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3451 that can only be obtained through BN_CTX_new() (which implicitly
3452 initialises it). The presence of this function only made it possible
3453 to overwrite an existing structure (and cause memory leaks).
3456 *) Because of the callback-based approach for implementing LHASH as a
3457 template type, lh_insert() adds opaque objects to hash-tables and
3458 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3459 to clean up those corresponding objects before destroying the hash table
3460 (and losing the object pointers). So some over-zealous constifications in
3461 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3462 objects as "const" and the lh_doall[_arg] callback wrappers are not
3463 prototyped to have "const" restrictions on the object pointers they are
3464 given (and so aren't required to cast them away any more).
3467 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3468 (speed) prefers to use its own implementation. The two implementations
3469 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3470 its object type properly exposed (MS_TM) instead of casting to/from "char
3471 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3472 aren't necessarily the greatest nomenclatures - but this is what was used
3473 internally to the implementation so I've used that for now.
3476 *) Ensure that deprecated functions do not get compiled when
3477 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3478 the self-tests were still using deprecated key-generation functions so
3479 these have been updated also.
3482 *) Reorganise PKCS#7 code to separate the digest location functionality
3483 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
3484 New function PKCS7_set_digest() to set the digest type for PKCS#7
3485 digestedData type. Add additional code to correctly generate the
3486 digestedData type and add support for this type in PKCS7 initialization
3490 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3491 structure of type "other".
3494 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3495 sure the loop does correctly stop and breaking ("division by zero")
3496 modulus operations are not performed. The (pre-generated) prime
3497 table crypto/bn/bn_prime.h was already correct, but it could not be
3498 re-generated on some platforms because of the "division by zero"
3499 situation in the script.
3500 [Ralf S. Engelschall]
3502 *) Update support for ECC-based TLS ciphersuites according to
3503 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3504 SHA-1 now is only used for "small" curves (where the
3505 representation of a field element takes up to 24 bytes); for
3506 larger curves, the field element resulting from ECDH is directly
3507 used as premaster secret.
3508 [Douglas Stebila (Sun Microsystems Laboratories)]
3510 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3511 curve secp160r1 to the tests.
3512 [Douglas Stebila (Sun Microsystems Laboratories)]
3514 *) Add the possibility to load symbols globally with DSO.
3515 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3517 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3518 control of the error stack.
3521 *) Add support for STORE in ENGINE.
3524 *) Add the STORE type. The intention is to provide a common interface
3525 to certificate and key stores, be they simple file-based stores, or
3526 HSM-type store, or LDAP stores, or...
3527 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3530 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3531 pass a list of arguments to any function as well as provide a way
3532 for a function to pass data back to the caller.
3535 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3536 works like BUF_strdup() but can be used to duplicate a portion of
3537 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3541 *) Add the function sk_find_ex() which works like sk_find(), but will
3542 return an index to an element even if an exact match couldn't be
3543 found. The index is guaranteed to point at the element where the
3544 searched-for key would be inserted to preserve sorting order.
3547 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3548 takes an extra flags argument for optional functionality. Currently,
3549 the following flags are defined:
3551 OBJ_BSEARCH_VALUE_ON_NOMATCH
3552 This one gets OBJ_bsearch_ex() to return a pointer to the first
3553 element where the comparing function returns a negative or zero
3556 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3557 This one gets OBJ_bsearch_ex() to return a pointer to the first
3558 element where the comparing function returns zero. This is useful
3559 if there are more than one element where the comparing function
3563 *) Make it possible to create self-signed certificates with 'openssl ca'
3564 in such a way that the self-signed certificate becomes part of the
3565 CA database and uses the same mechanisms for serial number generation
3566 as all other certificate signing. The new flag '-selfsign' enables
3567 this functionality. Adapt CA.sh and CA.pl.in.
3570 *) Add functionality to check the public key of a certificate request
3571 against a given private. This is useful to check that a certificate
3572 request can be signed by that key (self-signing).
3575 *) Make it possible to have multiple active certificates with the same
3576 subject in the CA index file. This is done only if the keyword
3577 'unique_subject' is set to 'no' in the main CA section (default
3578 if 'CA_default') of the configuration file. The value is saved
3579 with the database itself in a separate index attribute file,
3580 named like the index file with '.attr' appended to the name.
3583 *) Generate muti valued AVAs using '+' notation in config files for
3587 *) Support for nameConstraints certificate extension.
3590 *) Support for policyConstraints certificate extension.
3593 *) Support for policyMappings certificate extension.
3596 *) Make sure the default DSA_METHOD implementation only uses its
3597 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3598 and change its own handlers to be NULL so as to remove unnecessary
3599 indirection. This lets alternative implementations fallback to the
3600 default implementation more easily.
3603 *) Support for directoryName in GeneralName related extensions
3607 *) Make it possible to link applications using Makefile.shared.
3608 Make that possible even when linking against static libraries!
3611 *) Support for single pass processing for S/MIME signing. This now
3612 means that S/MIME signing can be done from a pipe, in addition
3613 cleartext signing (multipart/signed type) is effectively streaming
3614 and the signed data does not need to be all held in memory.
3616 This is done with a new flag PKCS7_STREAM. When this flag is set
3617 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3618 is done after the data is output (and digests calculated) in
3619 SMIME_write_PKCS7().
3622 *) Add full support for -rpath/-R, both in shared libraries and
3623 applications, at least on the platforms where it's known how
3627 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3628 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3629 will now compute a table of multiples of the generator that
3630 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3631 faster (notably in the case of a single point multiplication,
3632 scalar * generator).
3633 [Nils Larsch, Bodo Moeller]
3635 *) IPv6 support for certificate extensions. The various extensions
3636 which use the IP:a.b.c.d can now take IPv6 addresses using the
3637 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3641 *) Added an ENGINE that implements RSA by performing private key
3642 exponentiations with the GMP library. The conversions to and from
3643 GMP's mpz_t format aren't optimised nor are any montgomery forms
3644 cached, and on x86 it appears OpenSSL's own performance has caught up.
3645 However there are likely to be other architectures where GMP could
3646 provide a boost. This ENGINE is not built in by default, but it can be
3647 specified at Configure time and should be accompanied by the necessary
3648 linker additions, eg;
3649 ./config -DOPENSSL_USE_GMP -lgmp
3652 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3653 testing availability of engines with "-t" - the old behaviour is
3654 produced by increasing the feature's verbosity with "-tt".
3657 *) ECDSA routines: under certain error conditions uninitialized BN objects
3658 could be freed. Solution: make sure initialization is performed early
3659 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3663 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3664 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3665 software implementations. For DSA and DH, parameter generation can
3666 also be overriden by providing the appropriate method callbacks.
3669 *) Change the "progress" mechanism used in key-generation and
3670 primality testing to functions that take a new BN_GENCB pointer in
3671 place of callback/argument pairs. The new API functions have "_ex"
3672 postfixes and the older functions are reimplemented as wrappers for
3673 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3674 declarations of the old functions to help (graceful) attempts to
3675 migrate to the new functions. Also, the new key-generation API
3676 functions operate on a caller-supplied key-structure and return
3677 success/failure rather than returning a key or NULL - this is to
3678 help make "keygen" another member function of RSA_METHOD etc.
3680 Example for using the new callback interface:
3682 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3686 BN_GENCB_set(&my_cb, my_callback, my_arg);
3688 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3689 /* For the meaning of a, b in calls to my_callback(), see the
3690 * documentation of the function that calls the callback.
3691 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3692 * my_callback should return 1 if it wants BN_is_prime_ex()
3693 * to continue, or 0 to stop.
3698 *) Change the ZLIB compression method to be stateful, and make it
3699 available to TLS with the number defined in
3700 draft-ietf-tls-compression-04.txt.
3703 *) Add the ASN.1 structures and functions for CertificatePair, which
3704 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3706 CertificatePair ::= SEQUENCE {
3707 forward [0] Certificate OPTIONAL,
3708 reverse [1] Certificate OPTIONAL,
3709 -- at least one of the pair shall be present -- }
3711 Also implement the PEM functions to read and write certificate
3712 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3714 This needed to be defined, mostly for the sake of the LDAP
3715 attribute crossCertificatePair, but may prove useful elsewhere as
3719 *) Make it possible to inhibit symlinking of shared libraries in
3720 Makefile.shared, for Cygwin's sake.
3723 *) Extend the BIGNUM API by creating a function
3724 void BN_set_negative(BIGNUM *a, int neg);
3725 and a macro that behave like
3726 int BN_is_negative(const BIGNUM *a);
3728 to avoid the need to access 'a->neg' directly in applications.
3731 *) Implement fast modular reduction for pseudo-Mersenne primes
3732 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3733 EC_GROUP_new_curve_GFp() will now automatically use this
3735 [Nils Larsch <nla@trustcenter.de>]
3737 *) Add new lock type (CRYPTO_LOCK_BN).
3740 *) Change the ENGINE framework to automatically load engines
3741 dynamically from specific directories unless they could be
3742 found to already be built in or loaded. Move all the
3743 current engines except for the cryptodev one to a new
3745 The engines in engines/ are built as shared libraries if
3746 the "shared" options was given to ./Configure or ./config.
3747 Otherwise, they are inserted in libcrypto.a.
3748 /usr/local/ssl/engines is the default directory for dynamic
3749 engines, but that can be overriden at configure time through
3750 the usual use of --prefix and/or --openssldir, and at run
3751 time with the environment variable OPENSSL_ENGINES.
3752 [Geoff Thorpe and Richard Levitte]
3754 *) Add Makefile.shared, a helper makefile to build shared
3755 libraries. Addapt Makefile.org.
3758 *) Add version info to Win32 DLLs.
3759 [Peter 'Luna' Runestig" <peter@runestig.com>]
3761 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3762 can be added using this API to created arbitrary PKCS#12
3763 files while avoiding the low level API.
3765 New options to PKCS12_create(), key or cert can be NULL and
3766 will then be omitted from the output file. The encryption
3767 algorithm NIDs can be set to -1 for no encryption, the mac
3768 iteration count can be set to 0 to omit the mac.
3770 Enhance pkcs12 utility by making the -nokeys and -nocerts
3771 options work when creating a PKCS#12 file. New option -nomac
3772 to omit the mac, NONE can be set for an encryption algorithm.
3773 New code is modified to use the enhanced PKCS12_create()
3774 instead of the low level API.
3777 *) Extend ASN1 encoder to support indefinite length constructed
3778 encoding. This can output sequences tags and octet strings in
3779 this form. Modify pk7_asn1.c to support indefinite length
3780 encoding. This is experimental and needs additional code to
3781 be useful, such as an ASN1 bio and some enhanced streaming
3784 Extend template encode functionality so that tagging is passed
3785 down to the template encoder.
3788 *) Let 'openssl req' fail if an argument to '-newkey' is not
3789 recognized instead of using RSA as a default.
3792 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3793 As these are not official, they are not included in "ALL";
3794 the "ECCdraft" ciphersuite group alias can be used to select them.
3795 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3797 *) Add ECDH engine support.
3798 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3800 *) Add ECDH in new directory crypto/ecdh/.
3801 [Douglas Stebila (Sun Microsystems Laboratories)]
3803 *) Let BN_rand_range() abort with an error after 100 iterations
3804 without success (which indicates a broken PRNG).
3807 *) Change BN_mod_sqrt() so that it verifies that the input value
3808 is really the square of the return value. (Previously,
3809 BN_mod_sqrt would show GIGO behaviour.)
3812 *) Add named elliptic curves over binary fields from X9.62, SECG,
3813 and WAP/WTLS; add OIDs that were still missing.
3815 [Sheueling Chang Shantz and Douglas Stebila
3816 (Sun Microsystems Laboratories)]
3818 *) Extend the EC library for elliptic curves over binary fields
3819 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3822 EC_GF2m_simple_method
3826 EC_GROUP_new_curve_GF2m
3827 EC_GROUP_set_curve_GF2m
3828 EC_GROUP_get_curve_GF2m
3829 EC_POINT_set_affine_coordinates_GF2m
3830 EC_POINT_get_affine_coordinates_GF2m
3831 EC_POINT_set_compressed_coordinates_GF2m
3833 Point compression for binary fields is disabled by default for
3834 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3837 As binary polynomials are represented as BIGNUMs, various members
3838 of the EC_GROUP and EC_POINT data structures can be shared
3839 between the implementations for prime fields and binary fields;
3840 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3841 are essentially identical to their ..._GFp counterparts.
3842 (For simplicity, the '..._GFp' prefix has been dropped from
3843 various internal method names.)
3845 An internal 'field_div' method (similar to 'field_mul' and
3846 'field_sqr') has been added; this is used only for binary fields.
3848 [Sheueling Chang Shantz and Douglas Stebila
3849 (Sun Microsystems Laboratories)]
3851 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3852 through methods ('mul', 'precompute_mult').
3854 The generic implementations (now internally called 'ec_wNAF_mul'
3855 and 'ec_wNAF_precomputed_mult') remain the default if these
3856 methods are undefined.
3858 [Sheueling Chang Shantz and Douglas Stebila
3859 (Sun Microsystems Laboratories)]
3861 *) New function EC_GROUP_get_degree, which is defined through
3862 EC_METHOD. For curves over prime fields, this returns the bit
3863 length of the modulus.
3865 [Sheueling Chang Shantz and Douglas Stebila
3866 (Sun Microsystems Laboratories)]
3868 *) New functions EC_GROUP_dup, EC_POINT_dup.
3869 (These simply call ..._new and ..._copy).
3871 [Sheueling Chang Shantz and Douglas Stebila
3872 (Sun Microsystems Laboratories)]
3874 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3875 Polynomials are represented as BIGNUMs (where the sign bit is not
3876 used) in the following functions [macros]:
3879 BN_GF2m_sub [= BN_GF2m_add]
3880 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3881 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3882 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3884 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3885 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3886 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3887 BN_GF2m_cmp [= BN_ucmp]
3889 (Note that only the 'mod' functions are actually for fields GF(2^m).
3890 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3892 For some functions, an the irreducible polynomial defining a
3893 field can be given as an 'unsigned int[]' with strictly
3894 decreasing elements giving the indices of those bits that are set;
3895 i.e., p[] represents the polynomial
3896 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3898 p[0] > p[1] > ... > p[k] = 0.
3899 This applies to the following functions:
3904 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3905 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3907 BN_GF2m_mod_sqrt_arr
3908 BN_GF2m_mod_solve_quad_arr
3912 Conversion can be performed by the following functions:
3917 bntest.c has additional tests for binary polynomial arithmetic.
3919 Two implementations for BN_GF2m_mod_div() are available.
3920 The default algorithm simply uses BN_GF2m_mod_inv() and
3921 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3922 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3923 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3925 [Sheueling Chang Shantz and Douglas Stebila
3926 (Sun Microsystems Laboratories)]
3928 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3929 functionality is disabled at compile-time.
3930 [Douglas Stebila <douglas.stebila@sun.com>]
3932 *) Change default behaviour of 'openssl asn1parse' so that more
3933 information is visible when viewing, e.g., a certificate:
3935 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3936 mode the content of non-printable OCTET STRINGs is output in a
3937 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3938 avoid the appearance of a printable string.
3939 [Nils Larsch <nla@trustcenter.de>]
3941 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3943 EC_GROUP_set_asn1_flag()
3944 EC_GROUP_get_asn1_flag()
3945 EC_GROUP_set_point_conversion_form()
3946 EC_GROUP_get_point_conversion_form()
3947 These control ASN1 encoding details:
3948 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3949 has been set to OPENSSL_EC_NAMED_CURVE.
3950 - Points are encoded in uncompressed form by default; options for
3951 asn1_for are as for point2oct, namely
3952 POINT_CONVERSION_COMPRESSED
3953 POINT_CONVERSION_UNCOMPRESSED
3954 POINT_CONVERSION_HYBRID
3956 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3959 EC_GROUP_get0_seed()
3960 EC_GROUP_get_seed_len()
3961 This is used only for ASN1 purposes (so far).
3962 [Nils Larsch <nla@trustcenter.de>]
3964 *) Add 'field_type' member to EC_METHOD, which holds the NID
3965 of the appropriate field type OID. The new function
3966 EC_METHOD_get_field_type() returns this value.
3967 [Nils Larsch <nla@trustcenter.de>]
3972 EC_POINT_point2hex()
3973 EC_POINT_hex2point()
3974 providing useful interfaces to EC_POINT_point2oct() and
3975 EC_POINT_oct2point().
3976 [Nils Larsch <nla@trustcenter.de>]
3978 *) Change internals of the EC library so that the functions
3979 EC_GROUP_set_generator()
3980 EC_GROUP_get_generator()
3981 EC_GROUP_get_order()
3982 EC_GROUP_get_cofactor()
3983 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3984 to methods, which would lead to unnecessary code duplication when
3985 adding different types of curves.
3986 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3988 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3989 arithmetic, and such that modified wNAFs are generated
3990 (which avoid length expansion in many cases).
3993 *) Add a function EC_GROUP_check_discriminant() (defined via
3994 EC_METHOD) that verifies that the curve discriminant is non-zero.
3996 Add a function EC_GROUP_check() that makes some sanity tests
3997 on a EC_GROUP, its generator and order. This includes
3998 EC_GROUP_check_discriminant().
3999 [Nils Larsch <nla@trustcenter.de>]
4001 *) Add ECDSA in new directory crypto/ecdsa/.
4003 Add applications 'openssl ecparam' and 'openssl ecdsa'
4004 (these are based on 'openssl dsaparam' and 'openssl dsa').
4006 ECDSA support is also included in various other files across the
4007 library. Most notably,
4008 - 'openssl req' now has a '-newkey ecdsa:file' option;
4009 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4010 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4011 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4012 them suitable for ECDSA where domain parameters must be
4013 extracted before the specific public key;
4014 - ECDSA engine support has been added.
4015 [Nils Larsch <nla@trustcenter.de>]
4017 *) Include some named elliptic curves, and add OIDs from X9.62,
4018 SECG, and WAP/WTLS. Each curve can be obtained from the new
4020 EC_GROUP_new_by_curve_name(),
4021 and the list of available named curves can be obtained with
4022 EC_get_builtin_curves().
4023 Also add a 'curve_name' member to EC_GROUP objects, which can be
4025 EC_GROUP_set_curve_name()
4026 EC_GROUP_get_curve_name()
4027 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4029 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4030 was actually never needed) and in BN_mul(). The removal in BN_mul()
4031 required a small change in bn_mul_part_recursive() and the addition
4032 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4033 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4034 bn_sub_words() and bn_add_words() except they take arrays with
4038 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4040 *) Cleanse PEM buffers before freeing them since they may contain
4042 [Benjamin Bennett <ben@psc.edu>]
4044 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4045 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4046 authentication-only ciphersuites.
4049 *) Since AES128 and AES256 share a single mask bit in the logic of
4050 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4051 kludge to work properly if AES128 is available and AES256 isn't.
4054 *) Expand security boundary to match 1.1.1 module.
4057 *) Remove redundant features: hash file source, editing of test vectors
4058 modify fipsld to use external fips_premain.c signature.
4061 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4062 run algorithm test programs.
4065 *) Make algorithm test programs more tolerant of whitespace.
4068 *) Have SSL/TLS server implementation tolerate "mismatched" record
4069 protocol version while receiving ClientHello even if the
4070 ClientHello is fragmented. (The server can't insist on the
4071 particular protocol version it has chosen before the ServerHello
4072 message has informed the client about his choice.)
4075 *) Load error codes if they are not already present instead of using a
4076 static variable. This allows them to be cleanly unloaded and reloaded.
4079 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4081 *) Introduce limits to prevent malicious keys being able to
4082 cause a denial of service. (CVE-2006-2940)
4083 [Steve Henson, Bodo Moeller]
4085 *) Fix ASN.1 parsing of certain invalid structures that can result
4086 in a denial of service. (CVE-2006-2937) [Steve Henson]
4088 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4089 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4091 *) Fix SSL client code which could crash if connecting to a
4092 malicious SSLv2 server. (CVE-2006-4343)
4093 [Tavis Ormandy and Will Drewry, Google Security Team]
4095 *) Change ciphersuite string processing so that an explicit
4096 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4097 will no longer include "AES128-SHA"), and any other similar
4098 ciphersuite (same bitmap) from *other* protocol versions (so that
4099 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4100 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4101 changes from 0.9.8b and 0.9.8d.
4104 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4106 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4107 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4109 *) Change the Unix randomness entropy gathering to use poll() when
4110 possible instead of select(), since the latter has some
4111 undesirable limitations.
4112 [Darryl Miles via Richard Levitte and Bodo Moeller]
4114 *) Disable rogue ciphersuites:
4116 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4117 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4118 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4120 The latter two were purportedly from
4121 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4124 Also deactive the remaining ciphersuites from
4125 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4126 unofficial, and the ID has long expired.
4129 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4130 dual-core machines) and other potential thread-safety issues.
4133 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4135 *) Adapt fipsld and the build system to link against the validated FIPS
4136 module in FIPS mode.
4139 *) Fixes for VC++ 2005 build under Windows.
4142 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4143 from a Windows bash shell such as MSYS. It is autodetected from the
4144 "config" script when run from a VC++ environment. Modify standard VC++
4145 build to use fipscanister.o from the GNU make build.
4148 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4150 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4151 The value now differs depending on if you build for FIPS or not.
4152 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4153 safely run with a non-FIPSed libcrypto, as it may crash because of
4154 the difference induced by this change.
4157 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4159 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4160 (part of SSL_OP_ALL). This option used to disable the
4161 countermeasure against man-in-the-middle protocol-version
4162 rollback in the SSL 2.0 server implementation, which is a bad
4163 idea. (CVE-2005-2969)
4165 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4166 for Information Security, National Institute of Advanced Industrial
4167 Science and Technology [AIST], Japan)]
4169 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4170 mainly for FIPS compliance and not fully integrated at this stage.
4173 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4174 the exponentiation using a fixed-length exponent. (Otherwise,
4175 the information leaked through timing could expose the secret key
4176 after many signatures; cf. Bleichenbacher's attack on DSA with
4180 *) Make a new fixed-window mod_exp implementation the default for
4181 RSA, DSA, and DH private-key operations so that the sequence of
4182 squares and multiplies and the memory access pattern are
4183 independent of the particular secret key. This will mitigate
4184 cache-timing and potential related attacks.
4186 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4187 and this is automatically used by BN_mod_exp_mont() if the new flag
4188 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4189 will use this BN flag for private exponents unless the flag
4190 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4191 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4193 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4195 *) Change the client implementation for SSLv23_method() and
4196 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4197 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4198 (Previously, the SSL 2.0 backwards compatible Client Hello
4199 message format would be used even with SSL_OP_NO_SSLv2.)
4202 *) Add support for smime-type MIME parameter in S/MIME messages which some
4206 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4207 a threadsafe manner. Modify rsa code to use new function and add calls
4208 to dsa and dh code (which had race conditions before).
4211 *) Include the fixed error library code in the C error file definitions
4212 instead of fixing them up at runtime. This keeps the error code
4213 structures constant.
4216 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4218 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4221 *) Fixes for newer kerberos headers. NB: the casts are needed because
4222 the 'length' field is signed on one version and unsigned on another
4223 with no (?) obvious way to tell the difference, without these VC++
4224 complains. Also the "definition" of FAR (blank) is no longer included
4225 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4226 some needed definitions.
4229 *) Undo Cygwin change.
4232 *) Added support for proxy certificates according to RFC 3820.
4233 Because they may be a security thread to unaware applications,
4234 they must be explicitely allowed in run-time. See
4235 docs/HOWTO/proxy_certificates.txt for further information.
4238 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4240 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4241 server and client random values. Previously
4242 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4243 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4245 This change has negligible security impact because:
4247 1. Server and client random values still have 24 bytes of pseudo random
4250 2. Server and client random values are sent in the clear in the initial
4253 3. The master secret is derived using the premaster secret (48 bytes in
4254 size for static RSA ciphersuites) as well as client server and random
4257 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4260 [Stephen Henson, reported by UK NISCC]
4262 *) Use Windows randomness collection on Cygwin.
4265 *) Fix hang in EGD/PRNGD query when communication socket is closed
4266 prematurely by EGD/PRNGD.
4267 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4269 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4272 *) Back-port of selected performance improvements from development
4273 branch, as well as improved support for PowerPC platforms.
4276 *) Add lots of checks for memory allocation failure, error codes to indicate
4277 failure and freeing up memory if a failure occurs.
4278 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4280 *) Add new -passin argument to dgst.
4283 *) Perform some character comparisons of different types in X509_NAME_cmp:
4284 this is needed for some certificates that reencode DNs into UTF8Strings
4285 (in violation of RFC3280) and can't or wont issue name rollover
4289 *) Make an explicit check during certificate validation to see that
4290 the CA setting in each certificate on the chain is correct. As a
4291 side effect always do the following basic checks on extensions,
4292 not just when there's an associated purpose to the check:
4294 - if there is an unhandled critical extension (unless the user
4295 has chosen to ignore this fault)
4296 - if the path length has been exceeded (if one is set at all)
4297 - that certain extensions fit the associated purpose (if one has
4301 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4303 *) Avoid a race condition when CRLs are checked in a multi threaded
4304 environment. This would happen due to the reordering of the revoked
4305 entries during signature checking and serial number lookup. Now the
4306 encoding is cached and the serial number sort performed under a lock.
4307 Add new STACK function sk_is_sorted().
4310 *) Add Delta CRL to the extension code.
4313 *) Various fixes to s3_pkt.c so alerts are sent properly.
4314 [David Holmes <d.holmes@f5.com>]
4316 *) Reduce the chances of duplicate issuer name and serial numbers (in
4317 violation of RFC3280) using the OpenSSL certificate creation utilities.
4318 This is done by creating a random 64 bit value for the initial serial
4319 number when a serial number file is created or when a self signed
4320 certificate is created using 'openssl req -x509'. The initial serial
4321 number file is created using 'openssl x509 -next_serial' in CA.pl
4322 rather than being initialized to 1.
4325 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4327 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4328 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4329 [Joe Orton, Steve Henson]
4331 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4333 [Joe Orton, Steve Henson]
4335 *) Make it possible to have multiple active certificates with the same
4336 subject in the CA index file. This is done only if the keyword
4337 'unique_subject' is set to 'no' in the main CA section (default
4338 if 'CA_default') of the configuration file. The value is saved
4339 with the database itself in a separate index attribute file,
4340 named like the index file with '.attr' appended to the name.
4343 *) X509 verify fixes. Disable broken certificate workarounds when
4344 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4345 keyUsage extension present. Don't accept CRLs with unhandled critical
4346 extensions: since verify currently doesn't process CRL extensions this
4347 rejects a CRL with *any* critical extensions. Add new verify error codes
4351 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4352 A clarification of RFC2560 will require the use of OCTET STRINGs and
4353 some implementations cannot handle the current raw format. Since OpenSSL
4354 copies and compares OCSP nonces as opaque blobs without any attempt at
4355 parsing them this should not create any compatibility issues.
4358 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4359 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4360 this HMAC (and other) operations are several times slower than OpenSSL
4364 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4365 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4367 *) Use the correct content when signing type "other".
4370 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4372 *) Fix various bugs revealed by running the NISCC test suite:
4374 Stop out of bounds reads in the ASN1 code when presented with
4375 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4377 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4379 If verify callback ignores invalid public key errors don't try to check
4380 certificate signature with the NULL public key.
4384 *) New -ignore_err option in ocsp application to stop the server
4385 exiting on the first error in a request.
4388 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4389 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4393 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4394 extra data after the compression methods not only for TLS 1.0
4395 but also for SSL 3.0 (as required by the specification).
4396 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4398 *) Change X509_certificate_type() to mark the key as exported/exportable
4399 when it's 512 *bits* long, not 512 bytes.
4402 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4403 blocks during encryption.
4406 *) Various fixes to base64 BIO and non blocking I/O. On write
4407 flushes were not handled properly if the BIO retried. On read
4408 data was not being buffered properly and had various logic bugs.
4409 This also affects blocking I/O when the data being decoded is a
4413 *) Various S/MIME bugfixes and compatibility changes:
4414 output correct application/pkcs7 MIME type if
4415 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4416 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4417 of files as .eml work). Correctly handle very long lines in MIME
4421 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4423 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4424 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4425 a protocol version number mismatch like a decryption error
4426 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4429 *) Turn on RSA blinding by default in the default implementation
4430 to avoid a timing attack. Applications that don't want it can call
4431 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4432 They would be ill-advised to do so in most cases.
4433 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4435 *) Change RSA blinding code so that it works when the PRNG is not
4436 seeded (in this case, the secret RSA exponent is abused as
4437 an unpredictable seed -- if it is not unpredictable, there
4438 is no point in blinding anyway). Make RSA blinding thread-safe
4439 by remembering the creator's thread ID in rsa->blinding and
4440 having all other threads use local one-time blinding factors
4441 (this requires more computation than sharing rsa->blinding, but
4442 avoids excessive locking; and if an RSA object is not shared
4443 between threads, blinding will still be very fast).
4446 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4447 ENGINE as defaults for all supported algorithms irrespective of
4448 the 'flags' parameter. 'flags' is now honoured, so applications
4449 should make sure they are passing it correctly.
4452 *) Target "mingw" now allows native Windows code to be generated in
4453 the Cygwin environment as well as with the MinGW compiler.
4456 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4458 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4459 via timing by performing a MAC computation even if incorrrect
4460 block cipher padding has been found. This is a countermeasure
4461 against active attacks where the attacker has to distinguish
4462 between bad padding and a MAC verification error. (CVE-2003-0078)
4464 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4465 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4466 Martin Vuagnoux (EPFL, Ilion)]
4468 *) Make the no-err option work as intended. The intention with no-err
4469 is not to have the whole error stack handling routines removed from
4470 libcrypto, it's only intended to remove all the function name and
4471 reason texts, thereby removing some of the footprint that may not
4472 be interesting if those errors aren't displayed anyway.
4474 NOTE: it's still possible for any application or module to have it's
4475 own set of error texts inserted. The routines are there, just not
4476 used by default when no-err is given.
4479 *) Add support for FreeBSD on IA64.
4480 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4482 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4483 Kerberos function mit_des_cbc_cksum(). Before this change,
4484 the value returned by DES_cbc_cksum() was like the one from
4485 mit_des_cbc_cksum(), except the bytes were swapped.
4486 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4488 *) Allow an application to disable the automatic SSL chain building.
4489 Before this a rather primitive chain build was always performed in
4490 ssl3_output_cert_chain(): an application had no way to send the
4491 correct chain if the automatic operation produced an incorrect result.
4493 Now the chain builder is disabled if either:
4495 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4497 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4499 The reasoning behind this is that an application would not want the
4500 auto chain building to take place if extra chain certificates are
4501 present and it might also want a means of sending no additional
4502 certificates (for example the chain has two certificates and the
4506 *) Add the possibility to build without the ENGINE framework.
4507 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4509 *) Under Win32 gmtime() can return NULL: check return value in
4510 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4513 *) DSA routines: under certain error conditions uninitialized BN objects
4514 could be freed. Solution: make sure initialization is performed early
4515 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4516 Nils Larsch <nla@trustcenter.de> via PR#459)
4519 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4520 checked on reconnect on the client side, therefore session resumption
4521 could still fail with a "ssl session id is different" error. This
4522 behaviour is masked when SSL_OP_ALL is used due to
4523 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4524 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4525 followup to PR #377.
4528 *) IA-32 assembler support enhancements: unified ELF targets, support
4529 for SCO/Caldera platforms, fix for Cygwin shared build.
4532 *) Add support for FreeBSD on sparc64. As a consequence, support for
4533 FreeBSD on non-x86 processors is separate from x86 processors on
4534 the config script, much like the NetBSD support.
4535 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4537 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4539 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4542 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4543 code (06) was taken as the first octet of the session ID and the last
4544 octet was ignored consequently. As a result SSLv2 client side session
4545 caching could not have worked due to the session ID mismatch between
4547 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4551 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4552 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4556 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4557 seems that in spite of existing for more than a year, many application
4558 author have done nothing to provide the necessary callbacks, which
4559 means that this particular engine will not work properly anywhere.
4560 This is a very unfortunate situation which forces us, in the name
4561 of usability, to give the hw_ncipher.c a static lock, which is part
4563 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4564 appear in 0.9.8 or later. We EXPECT application authors to have
4565 dealt properly with this when 0.9.8 is released (unless we actually
4566 make such changes in the libcrypto locking code that changes will
4567 have to be made anyway).
4570 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4571 octets have been read, EOF or an error occurs. Without this change
4572 some truncated ASN1 structures will not produce an error.
4575 *) Disable Heimdal support, since it hasn't been fully implemented.
4576 Still give the possibility to force the use of Heimdal, but with
4577 warnings and a request that patches get sent to openssl-dev.
4580 *) Add the VC-CE target, introduce the WINCE sysname, and add
4581 INSTALL.WCE and appropriate conditionals to make it build.
4582 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4584 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4585 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4586 edit numbers of the version.
4587 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4589 *) Introduce safe string copy and catenation functions
4590 (BUF_strlcpy() and BUF_strlcat()).
4591 [Ben Laurie (CHATS) and Richard Levitte]
4593 *) Avoid using fixed-size buffers for one-line DNs.
4594 [Ben Laurie (CHATS)]
4596 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4597 resizing buffers containing secrets, and use where appropriate.
4598 [Ben Laurie (CHATS)]
4600 *) Avoid using fixed size buffers for configuration file location.
4601 [Ben Laurie (CHATS)]
4603 *) Avoid filename truncation for various CA files.
4604 [Ben Laurie (CHATS)]
4606 *) Use sizeof in preference to magic numbers.
4607 [Ben Laurie (CHATS)]
4609 *) Avoid filename truncation in cert requests.
4610 [Ben Laurie (CHATS)]
4612 *) Add assertions to check for (supposedly impossible) buffer
4614 [Ben Laurie (CHATS)]
4616 *) Don't cache truncated DNS entries in the local cache (this could
4617 potentially lead to a spoofing attack).
4618 [Ben Laurie (CHATS)]
4620 *) Fix various buffers to be large enough for hex/decimal
4621 representations in a platform independent manner.
4622 [Ben Laurie (CHATS)]
4624 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4625 resizing buffers containing secrets, and use where appropriate.
4626 [Ben Laurie (CHATS)]
4628 *) Add BIO_indent() to avoid much slightly worrying code to do
4630 [Ben Laurie (CHATS)]
4632 *) Convert sprintf()/BIO_puts() to BIO_printf().
4633 [Ben Laurie (CHATS)]
4635 *) buffer_gets() could terminate with the buffer only half
4637 [Ben Laurie (CHATS)]
4639 *) Add assertions to prevent user-supplied crypto functions from
4640 overflowing internal buffers by having large block sizes, etc.
4641 [Ben Laurie (CHATS)]
4643 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4645 [Ben Laurie (CHATS)]
4647 *) Eliminate unused copy of key in RC4.
4648 [Ben Laurie (CHATS)]
4650 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4651 [Ben Laurie (CHATS)]
4653 *) Fix off-by-one error in EGD path.
4654 [Ben Laurie (CHATS)]
4656 *) If RANDFILE path is too long, ignore instead of truncating.
4657 [Ben Laurie (CHATS)]
4659 *) Eliminate unused and incorrectly sized X.509 structure
4661 [Ben Laurie (CHATS)]
4663 *) Eliminate unused and dangerous function knumber().
4664 [Ben Laurie (CHATS)]
4666 *) Eliminate unused and dangerous structure, KSSL_ERR.
4667 [Ben Laurie (CHATS)]
4669 *) Protect against overlong session ID context length in an encoded
4670 session object. Since these are local, this does not appear to be
4672 [Ben Laurie (CHATS)]
4674 *) Change from security patch (see 0.9.6e below) that did not affect
4675 the 0.9.6 release series:
4677 Remote buffer overflow in SSL3 protocol - an attacker could
4678 supply an oversized master key in Kerberos-enabled versions.
4680 [Ben Laurie (CHATS)]
4682 *) Change the SSL kerb5 codes to match RFC 2712.
4685 *) Make -nameopt work fully for req and add -reqopt switch.
4686 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4688 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4689 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4691 *) Make sure tests can be performed even if the corresponding algorithms
4692 have been removed entirely. This was also the last step to make
4693 OpenSSL compilable with DJGPP under all reasonable conditions.
4694 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4696 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4697 to allow version independent disabling of normally unselected ciphers,
4698 which may be activated as a side-effect of selecting a single cipher.
4700 (E.g., cipher list string "RSA" enables ciphersuites that are left
4701 out of "ALL" because they do not provide symmetric encryption.
4702 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4703 [Lutz Jaenicke, Bodo Moeller]
4705 *) Add appropriate support for separate platform-dependent build
4706 directories. The recommended way to make a platform-dependent
4707 build directory is the following (tested on Linux), maybe with
4710 # Place yourself outside of the OpenSSL source tree. In
4711 # this example, the environment variable OPENSSL_SOURCE
4712 # is assumed to contain the absolute OpenSSL source directory.
4713 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4714 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4715 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4716 mkdir -p `dirname $F`
4717 ln -s $OPENSSL_SOURCE/$F $F
4720 To be absolutely sure not to disturb the source tree, a "make clean"
4721 is a good thing. If it isn't successfull, don't worry about it,
4722 it probably means the source directory is very clean.
4725 *) Make sure any ENGINE control commands make local copies of string
4726 pointers passed to them whenever necessary. Otherwise it is possible
4727 the caller may have overwritten (or deallocated) the original string
4728 data when a later ENGINE operation tries to use the stored values.
4729 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4731 *) Improve diagnostics in file reading and command-line digests.
4732 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4734 *) Add AES modes CFB and OFB to the object database. Correct an
4735 error in AES-CFB decryption.
4738 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4739 allows existing EVP_CIPHER_CTX structures to be reused after
4740 calling EVP_*Final(). This behaviour is used by encryption
4741 BIOs and some applications. This has the side effect that
4742 applications must explicitly clean up cipher contexts with
4743 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4746 *) Check the values of dna and dnb in bn_mul_recursive before calling
4747 bn_mul_comba (a non zero value means the a or b arrays do not contain
4748 n2 elements) and fallback to bn_mul_normal if either is not zero.
4751 *) Fix escaping of non-ASCII characters when using the -subj option
4752 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4755 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4756 form for "surname", serialNumber has no short form.
4757 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4758 therefore remove "mail" short name for "internet 7".
4759 The OID for unique identifiers in X509 certificates is
4760 x500UniqueIdentifier, not uniqueIdentifier.
4761 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4764 *) Add an "init" command to the ENGINE config module and auto initialize
4765 ENGINEs. Without any "init" command the ENGINE will be initialized
4766 after all ctrl commands have been executed on it. If init=1 the
4767 ENGINE is initailized at that point (ctrls before that point are run
4768 on the uninitialized ENGINE and after on the initialized one). If
4769 init=0 then the ENGINE will not be iniatialized at all.
4772 *) Fix the 'app_verify_callback' interface so that the user-defined
4773 argument is actually passed to the callback: In the
4774 SSL_CTX_set_cert_verify_callback() prototype, the callback
4775 declaration has been changed from
4778 int (*cb)(X509_STORE_CTX *,void *);
4779 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4780 i=s->ctx->app_verify_callback(&ctx)
4781 has been changed into
4782 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4784 To update applications using SSL_CTX_set_cert_verify_callback(),
4785 a dummy argument can be added to their callback functions.
4786 [D. K. Smetters <smetters@parc.xerox.com>]
4788 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4789 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4791 *) Add and OPENSSL_LOAD_CONF define which will cause
4792 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4793 This allows older applications to transparently support certain
4794 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4795 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4796 load the config file and OPENSSL_add_all_algorithms_conf() which will
4797 always load it have also been added.
4800 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4801 Adjust NIDs and EVP layer.
4802 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4804 *) Config modules support in openssl utility.
4806 Most commands now load modules from the config file,
4807 though in a few (such as version) this isn't done
4808 because it couldn't be used for anything.
4810 In the case of ca and req the config file used is
4811 the same as the utility itself: that is the -config
4812 command line option can be used to specify an
4816 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4817 use "openssl_conf" if filename is NULL use default openssl config file.
4820 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4821 config section name. Add a new flag to tolerate a missing config file
4822 and move code to CONF_modules_load_file().
4825 *) Support for crypto accelerator cards from Accelerated Encryption
4826 Processing, www.aep.ie. (Use engine 'aep')
4827 The support was copied from 0.9.6c [engine] and adapted/corrected
4828 to work with the new engine framework.
4829 [AEP Inc. and Richard Levitte]
4831 *) Support for SureWare crypto accelerator cards from Baltimore
4832 Technologies. (Use engine 'sureware')
4833 The support was copied from 0.9.6c [engine] and adapted
4834 to work with the new engine framework.
4837 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4838 make the newer ENGINE framework commands for the CHIL engine work.
4839 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4841 *) Make it possible to produce shared libraries on ReliantUNIX.
4842 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4844 *) Add the configuration target debug-linux-ppro.
4845 Make 'openssl rsa' use the general key loading routines
4846 implemented in apps.c, and make those routines able to
4847 handle the key format FORMAT_NETSCAPE and the variant
4849 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4851 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4852 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4854 *) Add -keyform to rsautl, and document -engine.
4855 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4857 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4858 BIO_R_NO_SUCH_FILE error code rather than the generic
4859 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4862 *) Add new functions
4864 ERR_peek_last_error_line
4865 ERR_peek_last_error_line_data.
4866 These are similar to
4869 ERR_peek_error_line_data,
4870 but report on the latest error recorded rather than the first one
4871 still in the error queue.
4872 [Ben Laurie, Bodo Moeller]
4874 *) default_algorithms option in ENGINE config module. This allows things
4876 default_algorithms = ALL
4877 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4880 *) Prelminary ENGINE config module.
4883 *) New experimental application configuration code.
4886 *) Change the AES code to follow the same name structure as all other
4887 symmetric ciphers, and behave the same way. Move everything to
4888 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4889 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4891 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4892 [Ben Laurie and Theo de Raadt]
4894 *) Add option to output public keys in req command.
4895 [Massimiliano Pala madwolf@openca.org]
4897 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4898 (up to about 10% better than before for P-192 and P-224).
4901 *) New functions/macros
4903 SSL_CTX_set_msg_callback(ctx, cb)
4904 SSL_CTX_set_msg_callback_arg(ctx, arg)
4905 SSL_set_msg_callback(ssl, cb)
4906 SSL_set_msg_callback_arg(ssl, arg)
4908 to request calling a callback function
4910 void cb(int write_p, int version, int content_type,
4911 const void *buf, size_t len, SSL *ssl, void *arg)
4913 whenever a protocol message has been completely received
4914 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4915 protocol version according to which the SSL library interprets
4916 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4917 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4918 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4919 specification (change_cipher_spec(20), alert(21), handshake(22)).
4920 'buf' and 'len' point to the actual message, 'ssl' to the
4921 SSL object, and 'arg' is the application-defined value set by
4922 SSL[_CTX]_set_msg_callback_arg().
4924 'openssl s_client' and 'openssl s_server' have new '-msg' options
4925 to enable a callback that displays all protocol messages.
4928 *) Change the shared library support so shared libraries are built as
4929 soon as the corresponding static library is finished, and thereby get
4930 openssl and the test programs linked against the shared library.
4931 This still only happens when the keyword "shard" has been given to
4932 the configuration scripts.
4934 NOTE: shared library support is still an experimental thing, and
4935 backward binary compatibility is still not guaranteed.
4936 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4938 *) Add support for Subject Information Access extension.
4939 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4941 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4942 additional bytes when new memory had to be allocated, not just
4943 when reusing an existing buffer.
4946 *) New command line and configuration option 'utf8' for the req command.
4947 This allows field values to be specified as UTF8 strings.
4950 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4951 runs for the former and machine-readable output for the latter.
4954 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4955 of the e-mail address in the DN (i.e., it will go into a certificate
4956 extension only). The new configuration file option 'email_in_dn = no'
4957 has the same effect.
4958 [Massimiliano Pala madwolf@openca.org]
4960 *) Change all functions with names starting with des_ to be starting
4961 with DES_ instead. Add wrappers that are compatible with libdes,
4962 but are named _ossl_old_des_*. Finally, add macros that map the
4963 des_* symbols to the corresponding _ossl_old_des_* if libdes
4964 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4965 desired, the des_* symbols will be mapped to DES_*, with one
4968 Since we provide two compatibility mappings, the user needs to
4969 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4970 compatibility is desired. The default (i.e., when that macro
4971 isn't defined) is OpenSSL 0.9.6c compatibility.
4973 There are also macros that enable and disable the support of old
4974 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4975 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4976 are defined, the default will apply: to support the old des routines.
4978 In either case, one must include openssl/des.h to get the correct
4979 definitions. Do not try to just include openssl/des_old.h, that
4982 NOTE: This is a major break of an old API into a new one. Software
4983 authors are encouraged to switch to the DES_ style functions. Some
4984 time in the future, des_old.h and the libdes compatibility functions
4985 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4986 default), and then completely removed.
4989 *) Test for certificates which contain unsupported critical extensions.
4990 If such a certificate is found during a verify operation it is
4991 rejected by default: this behaviour can be overridden by either
4992 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4993 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4994 X509_supported_extension() has also been added which returns 1 if a
4995 particular extension is supported.
4998 *) Modify the behaviour of EVP cipher functions in similar way to digests
4999 to retain compatibility with existing code.
5002 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5003 compatibility with existing code. In particular the 'ctx' parameter does
5004 not have to be to be initialized before the call to EVP_DigestInit() and
5005 it is tidied up after a call to EVP_DigestFinal(). New function
5006 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5007 EVP_MD_CTX_copy() changed to not require the destination to be
5008 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5009 requires the destination to be valid.
5011 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5012 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5015 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5016 so that complete 'Handshake' protocol structures are kept in memory
5017 instead of overwriting 'msg_type' and 'length' with 'body' data.
5020 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5021 [Massimo Santin via Richard Levitte]
5023 *) Major restructuring to the underlying ENGINE code. This includes
5024 reduction of linker bloat, separation of pure "ENGINE" manipulation
5025 (initialisation, etc) from functionality dealing with implementations
5026 of specific crypto iterfaces. This change also introduces integrated
5027 support for symmetric ciphers and digest implementations - so ENGINEs
5028 can now accelerate these by providing EVP_CIPHER and EVP_MD
5029 implementations of their own. This is detailed in crypto/engine/README
5030 as it couldn't be adequately described here. However, there are a few
5031 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5032 were changed in the original introduction of ENGINE code have now
5033 reverted back - the hooking from this code to ENGINE is now a good
5034 deal more passive and at run-time, operations deal directly with
5035 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5036 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5037 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5038 they were not being used by the framework as there is no concept of a
5039 BIGNUM_METHOD and they could not be generalised to the new
5040 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5041 ENGINE_cpy() has been removed as it cannot be consistently defined in
5045 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5048 *) Change mkdef.pl to sort symbols that get the same entry number,
5049 and make sure the automatically generated functions ERR_load_*
5050 become part of libeay.num as well.
5053 *) New function SSL_renegotiate_pending(). This returns true once
5054 renegotiation has been requested (either SSL_renegotiate() call
5055 or HelloRequest/ClientHello receveived from the peer) and becomes
5056 false once a handshake has been completed.
5057 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5058 sends a HelloRequest, but does not ensure that a handshake takes
5059 place. SSL_renegotiate_pending() is useful for checking if the
5060 client has followed the request.)
5063 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5064 By default, clients may request session resumption even during
5065 renegotiation (if session ID contexts permit); with this option,
5066 session resumption is possible only in the first handshake.
5068 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5069 more bits available for options that should not be part of
5070 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5073 *) Add some demos for certificate and certificate request creation.
5076 *) Make maximum certificate chain size accepted from the peer application
5077 settable (SSL*_get/set_max_cert_list()), as proposed by
5078 "Douglas E. Engert" <deengert@anl.gov>.
5081 *) Add support for shared libraries for Unixware-7
5082 (Boyd Lynn Gerber <gerberb@zenez.com>).
5085 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5086 be done prior to destruction. Use this to unload error strings from
5087 ENGINEs that load their own error strings. NB: This adds two new API
5088 functions to "get" and "set" this destroy handler in an ENGINE.
5091 *) Alter all existing ENGINE implementations (except "openssl" and
5092 "openbsd") to dynamically instantiate their own error strings. This
5093 makes them more flexible to be built both as statically-linked ENGINEs
5094 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5095 Also, add stub code to each that makes building them as self-contained
5096 shared-libraries easier (see README.ENGINE).
5099 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5100 implementations into applications that are completely implemented in
5101 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5102 commands that can be used to configure what shared-library to load and
5103 to control aspects of the way it is handled. Also, made an update to
5104 the README.ENGINE file that brings its information up-to-date and
5105 provides some information and instructions on the "dynamic" ENGINE
5106 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5109 *) Make it possible to unload ranges of ERR strings with a new
5110 "ERR_unload_strings" function.
5113 *) Add a copy() function to EVP_MD.
5116 *) Make EVP_MD routines take a context pointer instead of just the
5117 md_data void pointer.
5120 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5121 that the digest can only process a single chunk of data
5122 (typically because it is provided by a piece of
5123 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5124 is only going to provide a single chunk of data, and hence the
5125 framework needn't accumulate the data for oneshot drivers.
5128 *) As with "ERR", make it possible to replace the underlying "ex_data"
5129 functions. This change also alters the storage and management of global
5130 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5131 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5132 index counters. The API functions that use this state have been changed
5133 to take a "class_index" rather than pointers to the class's local STACK
5134 and counter, and there is now an API function to dynamically create new
5135 classes. This centralisation allows us to (a) plug a lot of the
5136 thread-safety problems that existed, and (b) makes it possible to clean
5137 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5138 such data would previously have always leaked in application code and
5139 workarounds were in place to make the memory debugging turn a blind eye
5140 to it. Application code that doesn't use this new function will still
5141 leak as before, but their memory debugging output will announce it now
5142 rather than letting it slide.
5144 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5145 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5146 has a return value to indicate success or failure.
5149 *) Make it possible to replace the underlying "ERR" functions such that the
5150 global state (2 LHASH tables and 2 locks) is only used by the "default"
5151 implementation. This change also adds two functions to "get" and "set"
5152 the implementation prior to it being automatically set the first time
5153 any other ERR function takes place. Ie. an application can call "get",
5154 pass the return value to a module it has just loaded, and that module
5155 can call its own "set" function using that value. This means the
5156 module's "ERR" operations will use (and modify) the error state in the
5157 application and not in its own statically linked copy of OpenSSL code.
5160 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5161 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5162 the operation, and provides a more encapsulated way for external code
5163 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5164 to use these functions rather than manually incrementing the counts.
5166 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5169 *) Add EVP test program.
5172 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5175 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5176 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5177 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5178 These allow a CRL to be built without having to access X509_CRL fields
5179 directly. Modify 'ca' application to use new functions.
5182 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5183 bug workarounds. Rollback attack detection is a security feature.
5184 The problem will only arise on OpenSSL servers when TLSv1 is not
5185 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5186 Software authors not wanting to support TLSv1 will have special reasons
5187 for their choice and can explicitly enable this option.
5188 [Bodo Moeller, Lutz Jaenicke]
5190 *) Rationalise EVP so it can be extended: don't include a union of
5191 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5192 (similar to those existing for EVP_CIPHER_CTX).
5197 EVP_MD_CTX_init(&md); /* new function call */
5198 EVP_DigestInit(&md, EVP_sha1());
5199 EVP_DigestUpdate(&md, in, len);
5200 EVP_DigestFinal(&md, out, NULL);
5201 EVP_MD_CTX_cleanup(&md); /* new function call */
5205 *) Make DES key schedule conform to the usual scheme, as well as
5206 correcting its structure. This means that calls to DES functions
5207 now have to pass a pointer to a des_key_schedule instead of a
5208 plain des_key_schedule (which was actually always a pointer
5211 des_key_schedule ks;
5213 des_set_key_checked(..., &ks);
5214 des_ncbc_encrypt(..., &ks, ...);
5216 (Note that a later change renames 'des_...' into 'DES_...'.)
5219 *) Initial reduction of linker bloat: the use of some functions, such as
5220 PEM causes large amounts of unused functions to be linked in due to
5221 poor organisation. For example pem_all.c contains every PEM function
5222 which has a knock on effect of linking in large amounts of (unused)
5223 ASN1 code. Grouping together similar functions and splitting unrelated
5224 functions prevents this.
5227 *) Cleanup of EVP macros.
5230 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5231 correct _ecb suffix.
5234 *) Add initial OCSP responder support to ocsp application. The
5235 revocation information is handled using the text based index
5236 use by the ca application. The responder can either handle
5237 requests generated internally, supplied in files (for example
5238 via a CGI script) or using an internal minimal server.
5241 *) Add configuration choices to get zlib compression for TLS.
5244 *) Changes to Kerberos SSL for RFC 2712 compliance:
5245 1. Implemented real KerberosWrapper, instead of just using
5246 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5247 2. Implemented optional authenticator field of KerberosWrapper.
5249 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5250 and authenticator structs; see crypto/krb5/.
5252 Generalized Kerberos calls to support multiple Kerberos libraries.
5253 [Vern Staats <staatsvr@asc.hpc.mil>,
5254 Jeffrey Altman <jaltman@columbia.edu>
5255 via Richard Levitte]
5257 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5258 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5259 values for each of the key sizes rather than having just
5260 parameters (and 'speed' generating keys each time).
5263 *) Speed up EVP routines.
5266 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5267 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5268 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5269 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5271 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5272 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5273 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5276 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5278 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5281 *) Added the OS2-EMX target.
5282 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5284 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5285 to support NCONF routines in extension code. New function CONF_set_nconf()
5286 to allow functions which take an NCONF to also handle the old LHASH
5287 structure: this means that the old CONF compatible routines can be
5288 retained (in particular wrt extensions) without having to duplicate the
5289 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5292 *) Enhance the general user interface with mechanisms for inner control
5293 and with possibilities to have yes/no kind of prompts.
5296 *) Change all calls to low level digest routines in the library and
5297 applications to use EVP. Add missing calls to HMAC_cleanup() and
5298 don't assume HMAC_CTX can be copied using memcpy().
5299 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5301 *) Add the possibility to control engines through control names but with
5302 arbitrary arguments instead of just a string.
5303 Change the key loaders to take a UI_METHOD instead of a callback
5304 function pointer. NOTE: this breaks binary compatibility with earlier
5305 versions of OpenSSL [engine].
5306 Adapt the nCipher code for these new conditions and add a card insertion
5310 *) Enhance the general user interface with mechanisms to better support
5311 dialog box interfaces, application-defined prompts, the possibility
5312 to use defaults (for example default passwords from somewhere else)
5313 and interrupts/cancellations.
5316 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5317 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5320 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5321 tidy up some unnecessarily weird code in 'sk_new()').
5322 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5324 *) Change the key loading routines for ENGINEs to use the same kind
5325 callback (pem_password_cb) as all other routines that need this
5329 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5330 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5331 than this minimum value is recommended.
5334 *) New random seeder for OpenVMS, using the system process statistics
5335 that are easily reachable.
5338 *) Windows apparently can't transparently handle global
5339 variables defined in DLLs. Initialisations such as:
5341 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5343 wont compile. This is used by the any applications that need to
5344 declare their own ASN1 modules. This was fixed by adding the option
5345 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5346 needed for static libraries under Win32.
5349 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5350 setting of purpose and trust fields. New X509_STORE trust and
5351 purpose functions and tidy up setting in other SSL functions.
5354 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5355 structure. These are inherited by X509_STORE_CTX when it is
5356 initialised. This allows various defaults to be set in the
5357 X509_STORE structure (such as flags for CRL checking and custom
5358 purpose or trust settings) for functions which only use X509_STORE_CTX
5359 internally such as S/MIME.
5361 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5362 trust settings if they are not set in X509_STORE. This allows X509_STORE
5363 purposes and trust (in S/MIME for example) to override any set by default.
5365 Add command line options for CRL checking to smime, s_client and s_server
5369 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5370 are set then the CRL is looked up in the X509_STORE structure and
5371 its validity and signature checked, then if the certificate is found
5372 in the CRL the verify fails with a revoked error.
5374 Various new CRL related callbacks added to X509_STORE_CTX structure.
5376 Command line options added to 'verify' application to support this.
5378 This needs some additional work, such as being able to handle multiple
5379 CRLs with different times, extension based lookup (rather than just
5380 by subject name) and ultimately more complete V2 CRL extension
5384 *) Add a general user interface API (crypto/ui/). This is designed
5385 to replace things like des_read_password and friends (backward
5386 compatibility functions using this new API are provided).
5387 The purpose is to remove prompting functions from the DES code
5388 section as well as provide for prompting through dialog boxes in
5389 a window system and the like.
5392 *) Add "ex_data" support to ENGINE so implementations can add state at a
5393 per-structure level rather than having to store it globally.
5396 *) Make it possible for ENGINE structures to be copied when retrieved by
5397 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5398 This causes the "original" ENGINE structure to act like a template,
5399 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5400 operational state can be localised to each ENGINE structure, despite the
5401 fact they all share the same "methods". New ENGINE structures returned in
5402 this case have no functional references and the return value is the single
5403 structural reference. This matches the single structural reference returned
5404 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5408 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5409 needs to match any other type at all we need to manually clear the
5413 *) Changes to the "openssl engine" utility to include;
5414 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5415 about an ENGINE's available control commands.
5416 - executing control commands from command line arguments using the
5417 '-pre' and '-post' switches. '-post' is only used if '-t' is
5418 specified and the ENGINE is successfully initialised. The syntax for
5419 the individual commands are colon-separated, for example;
5420 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5423 *) New dynamic control command support for ENGINEs. ENGINEs can now
5424 declare their own commands (numbers), names (strings), descriptions,
5425 and input types for run-time discovery by calling applications. A
5426 subset of these commands are implicitly classed as "executable"
5427 depending on their input type, and only these can be invoked through
5428 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5429 can be based on user input, config files, etc). The distinction is
5430 that "executable" commands cannot return anything other than a boolean
5431 result and can only support numeric or string input, whereas some
5432 discoverable commands may only be for direct use through
5433 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5434 pointers, or other custom uses. The "executable" commands are to
5435 support parameterisations of ENGINE behaviour that can be
5436 unambiguously defined by ENGINEs and used consistently across any
5437 OpenSSL-based application. Commands have been added to all the
5438 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5439 control over shared-library paths without source code alterations.
5442 *) Changed all ENGINE implementations to dynamically allocate their
5443 ENGINEs rather than declaring them statically. Apart from this being
5444 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5445 this also allows the implementations to compile without using the
5446 internal engine_int.h header.
5449 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5450 'const' value. Any code that should be able to modify a RAND_METHOD
5451 should already have non-const pointers to it (ie. they should only
5452 modify their own ones).
5455 *) Made a variety of little tweaks to the ENGINE code.
5456 - "atalla" and "ubsec" string definitions were moved from header files
5457 to C code. "nuron" string definitions were placed in variables
5458 rather than hard-coded - allowing parameterisation of these values
5459 later on via ctrl() commands.
5460 - Removed unused "#if 0"'d code.
5461 - Fixed engine list iteration code so it uses ENGINE_free() to release
5462 structural references.
5463 - Constified the RAND_METHOD element of ENGINE structures.
5464 - Constified various get/set functions as appropriate and added
5465 missing functions (including a catch-all ENGINE_cpy that duplicates
5466 all ENGINE values onto a new ENGINE except reference counts/state).
5467 - Removed NULL parameter checks in get/set functions. Setting a method
5468 or function to NULL is a way of cancelling out a previously set
5469 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5470 and doesn't justify the extra error symbols and code.
5471 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5472 flags from engine_int.h to engine.h.
5473 - Changed prototypes for ENGINE handler functions (init(), finish(),
5474 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5477 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5478 to the algorithm using long division. The binary algorithm can be
5479 used only if the modulus is odd. On 32-bit systems, it is faster
5480 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5481 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5482 up to 450 bits. In 64-bit environments, the binary algorithm
5483 appears to be advantageous for much longer moduli; here we use it
5484 for moduli up to 2048 bits.
5487 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5488 could not support the combine flag in choice fields.
5491 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5492 extensions from a certificate request to the certificate.
5495 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5496 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5497 file: this allows the display of the certificate about to be
5498 signed to be customised, to allow certain fields to be included
5499 or excluded and extension details. The old system didn't display
5500 multicharacter strings properly, omitted fields not in the policy
5501 and couldn't display additional details such as extensions.
5504 *) Function EC_POINTs_mul for multiple scalar multiplication
5505 of an arbitrary number of elliptic curve points
5506 \sum scalars[i]*points[i],
5507 optionally including the generator defined for the EC_GROUP:
5508 scalar*generator + \sum scalars[i]*points[i].
5510 EC_POINT_mul is a simple wrapper function for the typical case
5511 that the point list has just one item (besides the optional
5515 *) First EC_METHODs for curves over GF(p):
5517 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5518 operations and provides various method functions that can also
5519 operate with faster implementations of modular arithmetic.
5521 EC_GFp_mont_method() reuses most functions that are part of
5522 EC_GFp_simple_method, but uses Montgomery arithmetic.
5524 [Bodo Moeller; point addition and point doubling
5525 implementation directly derived from source code provided by
5526 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5528 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5529 crypto/ec/ec_lib.c):
5531 Curves are EC_GROUP objects (with an optional group generator)
5532 based on EC_METHODs that are built into the library.
5534 Points are EC_POINT objects based on EC_GROUP objects.
5536 Most of the framework would be able to handle curves over arbitrary
5537 finite fields, but as there are no obvious types for fields other
5538 than GF(p), some functions are limited to that for now.
5541 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5542 that the file contains a complete HTTP response.
5545 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5546 change the def and num file printf format specifier from "%-40sXXX"
5547 to "%-39s XXX". The latter will always guarantee a space after the
5548 field while the former will cause them to run together if the field
5549 is 40 of more characters long.
5552 *) Constify the cipher and digest 'method' functions and structures
5553 and modify related functions to take constant EVP_MD and EVP_CIPHER
5557 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5558 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5561 *) Modify EVP_Digest*() routines so they now return values. Although the
5562 internal software routines can never fail additional hardware versions
5566 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5568 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5569 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5572 ERR_R_NESTED_ASN1_ERROR
5574 ERR_R_MISSING_ASN1_EOS
5575 were 4 .. 9, conflicting with
5576 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5578 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5579 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5581 Add new error code 'ERR_R_INTERNAL_ERROR'.
5584 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5588 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5589 sets the subject name for a new request or supersedes the
5590 subject name in a given request. Formats that can be parsed are
5591 'CN=Some Name, OU=myOU, C=IT'
5593 'CN=Some Name/OU=myOU/C=IT'.
5595 Add options '-batch' and '-verbose' to 'openssl req'.
5596 [Massimiliano Pala <madwolf@hackmasters.net>]
5598 *) Introduce the possibility to access global variables through
5599 functions on platform were that's the best way to handle exporting
5600 global variables in shared libraries. To enable this functionality,
5601 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5602 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5603 is normally done by Configure or something similar).
5605 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5606 in the source file (foo.c) like this:
5608 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5609 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5611 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5612 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5614 OPENSSL_DECLARE_GLOBAL(int,foo);
5615 #define foo OPENSSL_GLOBAL_REF(foo)
5616 OPENSSL_DECLARE_GLOBAL(double,bar);
5617 #define bar OPENSSL_GLOBAL_REF(bar)
5619 The #defines are very important, and therefore so is including the
5620 header file everywhere where the defined globals are used.
5622 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5623 of ASN.1 items, but that structure is a bit different.
5625 The largest change is in util/mkdef.pl which has been enhanced with
5626 better and easier to understand logic to choose which symbols should
5627 go into the Windows .def files as well as a number of fixes and code
5628 cleanup (among others, algorithm keywords are now sorted
5629 lexicographically to avoid constant rewrites).
5632 *) In BN_div() keep a copy of the sign of 'num' before writing the
5633 result to 'rm' because if rm==num the value will be overwritten
5634 and produce the wrong result if 'num' is negative: this caused
5635 problems with BN_mod() and BN_nnmod().
5638 *) Function OCSP_request_verify(). This checks the signature on an
5639 OCSP request and verifies the signer certificate. The signer
5640 certificate is just checked for a generic purpose and OCSP request
5644 *) Add OCSP_check_validity() function to check the validity of OCSP
5645 responses. OCSP responses are prepared in real time and may only
5646 be a few seconds old. Simply checking that the current time lies
5647 between thisUpdate and nextUpdate max reject otherwise valid responses
5648 caused by either OCSP responder or client clock inaccuracy. Instead
5649 we allow thisUpdate and nextUpdate to fall within a certain period of
5650 the current time. The age of the response can also optionally be
5651 checked. Two new options -validity_period and -status_age added to
5655 *) If signature or public key algorithm is unrecognized print out its
5656 OID rather that just UNKNOWN.
5659 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5660 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5661 ID to be generated from the issuer certificate alone which can then be
5662 passed to OCSP_id_issuer_cmp().
5665 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5666 ASN1 modules to export functions returning ASN1_ITEM pointers
5667 instead of the ASN1_ITEM structures themselves. This adds several
5668 new macros which allow the underlying ASN1 function/structure to
5669 be accessed transparently. As a result code should not use ASN1_ITEM
5670 references directly (such as &X509_it) but instead use the relevant
5671 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5672 use of the new ASN1 code on platforms where exporting structures
5673 is problematical (for example in shared libraries) but exporting
5674 functions returning pointers to structures is not.
5677 *) Add support for overriding the generation of SSL/TLS session IDs.
5678 These callbacks can be registered either in an SSL_CTX or per SSL.
5679 The purpose of this is to allow applications to control, if they wish,
5680 the arbitrary values chosen for use as session IDs, particularly as it
5681 can be useful for session caching in multiple-server environments. A
5682 command-line switch for testing this (and any client code that wishes
5683 to use such a feature) has been added to "s_server".
5684 [Geoff Thorpe, Lutz Jaenicke]
5686 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5687 of the form '#if defined(...) || defined(...) || ...' and
5688 '#if !defined(...) && !defined(...) && ...'. This also avoids
5689 the growing number of special cases it was previously handling.
5692 *) Make all configuration macros available for application by making
5693 sure they are available in opensslconf.h, by giving them names starting
5694 with "OPENSSL_" to avoid conflicts with other packages and by making
5695 sure e_os2.h will cover all platform-specific cases together with
5697 Additionally, it is now possible to define configuration/platform-
5698 specific names (called "system identities"). In the C code, these
5699 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5700 macro with the name beginning with "OPENSSL_SYS_", which is determined
5701 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5705 *) New option -set_serial to 'req' and 'x509' this allows the serial
5706 number to use to be specified on the command line. Previously self
5707 signed certificates were hard coded with serial number 0 and the
5708 CA options of 'x509' had to use a serial number in a file which was
5712 *) New options to 'ca' utility to support V2 CRL entry extensions.
5713 Currently CRL reason, invalidity date and hold instruction are
5714 supported. Add new CRL extensions to V3 code and some new objects.
5717 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5718 disable standard block padding (aka PKCS#5 padding) in the EVP
5719 API, which was previously mandatory. This means that the data is
5720 not padded in any way and so the total length much be a multiple
5721 of the block size, otherwise an error occurs.
5724 *) Initial (incomplete) OCSP SSL support.
5727 *) New function OCSP_parse_url(). This splits up a URL into its host,
5728 port and path components: primarily to parse OCSP URLs. New -url
5729 option to ocsp utility.
5732 *) New nonce behavior. The return value of OCSP_check_nonce() now
5733 reflects the various checks performed. Applications can decide
5734 whether to tolerate certain situations such as an absent nonce
5735 in a response when one was present in a request: the ocsp application
5736 just prints out a warning. New function OCSP_add1_basic_nonce()
5737 this is to allow responders to include a nonce in a response even if
5738 the request is nonce-less.
5741 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5742 skipped when using openssl x509 multiple times on a single input file,
5743 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5746 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5747 set string type: to handle setting ASN1_TIME structures. Fix ca
5748 utility to correctly initialize revocation date of CRLs.
5751 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5752 the clients preferred ciphersuites and rather use its own preferences.
5753 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5754 Internet Explorer by ensuring unchanged hash method during stepup.
5755 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5758 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5759 to aes and add a new 'exist' option to print out symbols that don't
5763 *) Additional options to ocsp utility to allow flags to be set and
5764 additional certificates supplied.
5767 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5768 OCSP client a number of certificate to only verify the response
5772 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5773 handle the new API. Currently only ECB, CBC modes supported. Add new
5776 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5777 Encryption Standard (AES) Ciphersuites for Transport Layer
5778 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5779 not enabled by default and were not part of the "ALL" ciphersuite
5780 alias because they were not yet official; they could be
5781 explicitly requested by specifying the "AESdraft" ciphersuite
5782 group alias. In the final release of OpenSSL 0.9.7, the group
5783 alias is called "AES" and is part of "ALL".)
5784 [Ben Laurie, Steve Henson, Bodo Moeller]
5786 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5787 request to response.
5790 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5791 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5792 extract information from a certificate request. OCSP_response_create()
5793 creates a response and optionally adds a basic response structure.
5794 OCSP_basic_add1_status() adds a complete single response to a basic
5795 response and returns the OCSP_SINGLERESP structure just added (to allow
5796 extensions to be included for example). OCSP_basic_add1_cert() adds a
5797 certificate to a basic response and OCSP_basic_sign() signs a basic
5798 response with various flags. New helper functions ASN1_TIME_check()
5799 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5800 (converts ASN1_TIME to GeneralizedTime).
5803 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5804 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5805 structure from a certificate. X509_pubkey_digest() digests the public_key
5806 contents: this is used in various key identifiers.
5809 *) Make sk_sort() tolerate a NULL argument.
5810 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5812 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5813 passed by the function are trusted implicitly. If any of them signed the
5814 response then it is assumed to be valid and is not verified.
5817 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5818 to data. This was previously part of the PKCS7 ASN1 code. This
5819 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5820 [Steve Henson, reported by Kenneth R. Robinette
5821 <support@securenetterm.com>]
5823 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5824 routines: without these tracing memory leaks is very painful.
5825 Fix leaks in PKCS12 and PKCS7 routines.
5828 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5829 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5830 effectively meant GeneralizedTime would never be used. Now it
5831 is initialised to -1 but X509_time_adj() now has to check the value
5832 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5833 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5834 [Steve Henson, reported by Kenneth R. Robinette
5835 <support@securenetterm.com>]
5837 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5838 result in a zero length in the ASN1_INTEGER structure which was
5839 not consistent with the structure when d2i_ASN1_INTEGER() was used
5840 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5841 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5842 where it did not print out a minus for negative ASN1_INTEGER.
5845 *) Add summary printout to ocsp utility. The various functions which
5846 convert status values to strings have been renamed to:
5847 OCSP_response_status_str(), OCSP_cert_status_str() and
5848 OCSP_crl_reason_str() and are no longer static. New options
5849 to verify nonce values and to disable verification. OCSP response
5850 printout format cleaned up.
5853 *) Add additional OCSP certificate checks. These are those specified
5854 in RFC2560. This consists of two separate checks: the CA of the
5855 certificate being checked must either be the OCSP signer certificate
5856 or the issuer of the OCSP signer certificate. In the latter case the
5857 OCSP signer certificate must contain the OCSP signing extended key
5858 usage. This check is performed by attempting to match the OCSP
5859 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5860 in the OCSP_CERTID structures of the response.
5863 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5864 and related routines. This uses the standard OpenSSL certificate
5865 verify routines to perform initial checks (just CA validity) and
5866 to obtain the certificate chain. Then additional checks will be
5867 performed on the chain. Currently the root CA is checked to see
5868 if it is explicitly trusted for OCSP signing. This is used to set
5869 a root CA as a global signing root: that is any certificate that
5870 chains to that CA is an acceptable OCSP signing certificate.
5873 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5874 extensions from a separate configuration file.
5875 As when reading extensions from the main configuration file,
5876 the '-extensions ...' option may be used for specifying the
5878 [Massimiliano Pala <madwolf@comune.modena.it>]
5880 *) New OCSP utility. Allows OCSP requests to be generated or
5881 read. The request can be sent to a responder and the output
5882 parsed, outputed or printed in text form. Not complete yet:
5883 still needs to check the OCSP response validity.
5886 *) New subcommands for 'openssl ca':
5887 'openssl ca -status <serial>' prints the status of the cert with
5888 the given serial number (according to the index file).
5889 'openssl ca -updatedb' updates the expiry status of certificates
5891 [Massimiliano Pala <madwolf@comune.modena.it>]
5893 *) New '-newreq-nodes' command option to CA.pl. This is like
5894 '-newreq', but calls 'openssl req' with the '-nodes' option
5895 so that the resulting key is not encrypted.
5896 [Damien Miller <djm@mindrot.org>]
5898 *) New configuration for the GNU Hurd.
5899 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5901 *) Initial code to implement OCSP basic response verify. This
5902 is currently incomplete. Currently just finds the signer's
5903 certificate and verifies the signature on the response.
5906 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5907 value of OPENSSLDIR. This is available via the new '-d' option
5908 to 'openssl version', and is also included in 'openssl version -a'.
5911 *) Allowing defining memory allocation callbacks that will be given
5912 file name and line number information in additional arguments
5913 (a const char* and an int). The basic functionality remains, as
5914 well as the original possibility to just replace malloc(),
5915 realloc() and free() by functions that do not know about these
5916 additional arguments. To register and find out the current
5917 settings for extended allocation functions, the following
5918 functions are provided:
5920 CRYPTO_set_mem_ex_functions
5921 CRYPTO_set_locked_mem_ex_functions
5922 CRYPTO_get_mem_ex_functions
5923 CRYPTO_get_locked_mem_ex_functions
5925 These work the same way as CRYPTO_set_mem_functions and friends.
5926 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5927 extended allocation function is enabled.
5928 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5929 a conventional allocation function is enabled.
5930 [Richard Levitte, Bodo Moeller]
5932 *) Finish off removing the remaining LHASH function pointer casts.
5933 There should no longer be any prototype-casting required when using
5934 the LHASH abstraction, and any casts that remain are "bugs". See
5935 the callback types and macros at the head of lhash.h for details
5936 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5939 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5940 If /dev/[u]random devices are not available or do not return enough
5941 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5943 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5944 /etc/entropy will be queried once each in this sequence, quering stops
5945 when enough entropy was collected without querying more sockets.
5948 *) Change the Unix RAND_poll() variant to be able to poll several
5949 random devices, as specified by DEVRANDOM, until a sufficient amount
5950 of data has been collected. We spend at most 10 ms on each file
5951 (select timeout) and read in non-blocking mode. DEVRANDOM now
5952 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5953 (previously it was just the string "/dev/urandom"), so on typical
5954 platforms the 10 ms delay will never occur.
5955 Also separate out the Unix variant to its own file, rand_unix.c.
5956 For VMS, there's a currently-empty rand_vms.c.
5959 *) Move OCSP client related routines to ocsp_cl.c. These
5960 provide utility functions which an application needing
5961 to issue a request to an OCSP responder and analyse the
5962 response will typically need: as opposed to those which an
5963 OCSP responder itself would need which will be added later.
5965 OCSP_request_sign() signs an OCSP request with an API similar
5966 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5967 response. OCSP_response_get1_basic() extracts basic response
5968 from response. OCSP_resp_find_status(): finds and extracts status
5969 information from an OCSP_CERTID structure (which will be created
5970 when the request structure is built). These are built from lower
5971 level functions which work on OCSP_SINGLERESP structures but
5972 wont normally be used unless the application wishes to examine
5973 extensions in the OCSP response for example.
5975 Replace nonce routines with a pair of functions.
5976 OCSP_request_add1_nonce() adds a nonce value and optionally
5977 generates a random value. OCSP_check_nonce() checks the
5978 validity of the nonce in an OCSP response.
5981 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5982 This doesn't copy the supplied OCSP_CERTID and avoids the
5983 need to free up the newly created id. Change return type
5984 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5985 This can then be used to add extensions to the request.
5986 Deleted OCSP_request_new(), since most of its functionality
5987 is now in OCSP_REQUEST_new() (and the case insensitive name
5988 clash) apart from the ability to set the request name which
5989 will be added elsewhere.
5992 *) Update OCSP API. Remove obsolete extensions argument from
5993 various functions. Extensions are now handled using the new
5994 OCSP extension code. New simple OCSP HTTP function which
5995 can be used to send requests and parse the response.
5998 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5999 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6000 uses the special reorder version of SET OF to sort the attributes
6001 and reorder them to match the encoded order. This resolves a long
6002 standing problem: a verify on a PKCS7 structure just after signing
6003 it used to fail because the attribute order did not match the
6004 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6005 it uses the received order. This is necessary to tolerate some broken
6006 software that does not order SET OF. This is handled by encoding
6007 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6008 to produce the required SET OF.
6011 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6012 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6013 files to get correct declarations of the ASN.1 item variables.
6016 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6017 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6018 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6019 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6020 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6021 ASN1_ITEM and no wrapper functions.
6024 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6025 replace the old function pointer based I/O routines. Change most of
6026 the *_d2i_bio() and *_d2i_fp() functions to use these.
6029 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6030 lines, recognice more "algorithms" that can be deselected, and make
6031 it complain about algorithm deselection that isn't recognised.
6034 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6035 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6036 to use new functions. Add NO_ASN1_OLD which can be set to remove
6037 some old style ASN1 functions: this can be used to determine if old
6038 code will still work when these eventually go away.
6041 *) New extension functions for OCSP structures, these follow the
6042 same conventions as certificates and CRLs.
6045 *) New function X509V3_add1_i2d(). This automatically encodes and
6046 adds an extension. Its behaviour can be customised with various
6047 flags to append, replace or delete. Various wrappers added for
6048 certifcates and CRLs.
6051 *) Fix to avoid calling the underlying ASN1 print routine when
6052 an extension cannot be parsed. Correct a typo in the
6053 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6056 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
6057 entries for variables.
6060 *) Add functionality to apps/openssl.c for detecting locking
6061 problems: As the program is single-threaded, all we have
6062 to do is register a locking callback using an array for
6063 storing which locks are currently held by the program.
6066 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6067 SSL_get_ex_data_X509_STORE_idx(), which is used in
6068 ssl_verify_cert_chain() and thus can be called at any time
6069 during TLS/SSL handshakes so that thread-safety is essential.
6070 Unfortunately, the ex_data design is not at all suited
6071 for multi-threaded use, so it probably should be abolished.
6074 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6075 [Broadcom, tweaked and integrated by Geoff Thorpe]
6077 *) Move common extension printing code to new function
6078 X509V3_print_extensions(). Reorganise OCSP print routines and
6079 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6082 *) New function X509_signature_print() to remove duplication in some
6086 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6087 set (this was treated exactly the same as SET OF previously). This
6088 is used to reorder the STACK representing the structure to match the
6089 encoding. This will be used to get round a problem where a PKCS7
6090 structure which was signed could not be verified because the STACK
6091 order did not reflect the encoded order.
6094 *) Reimplement the OCSP ASN1 module using the new code.
6097 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6098 for its ASN1 operations. The old style function pointers still exist
6099 for now but they will eventually go away.
6102 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6103 completely replaces the old ASN1 functionality with a table driven
6104 encoder and decoder which interprets an ASN1_ITEM structure describing
6105 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6106 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6107 has also been converted to the new form.
6110 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6111 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6112 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6113 for negative moduli.
6116 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6117 of not touching the result's sign bit.
6120 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6124 *) Changed the LHASH code to use prototypes for callbacks, and created
6125 macros to declare and implement thin (optionally static) functions
6126 that provide type-safety and avoid function pointer casting for the
6127 type-specific callbacks.
6130 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6132 [Veers Staats <staatsvr@asc.hpc.mil>,
6133 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6135 *) Reformat the FAQ so the different questions and answers can be divided
6136 in sections depending on the subject.
6139 *) Have the zlib compression code load ZLIB.DLL dynamically under
6143 *) New function BN_mod_sqrt for computing square roots modulo a prime
6144 (using the probabilistic Tonelli-Shanks algorithm unless
6145 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6146 be handled deterministically).
6147 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6149 *) Make BN_mod_inverse faster by explicitly handling small quotients
6150 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6151 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6154 *) New function BN_kronecker.
6157 *) Fix BN_gcd so that it works on negative inputs; the result is
6158 positive unless both parameters are zero.
6159 Previously something reasonably close to an infinite loop was
6160 possible because numbers could be growing instead of shrinking
6161 in the implementation of Euclid's algorithm.
6164 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6165 sign of the number in question.
6167 Fix BN_is_word(a,w) to work correctly for w == 0.
6169 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6170 because its test if the absolute value of 'a' equals 'w'.
6171 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6172 it exists mostly for use in the implementations of BN_is_zero(),
6173 BN_is_one(), and BN_is_word().
6176 *) New function BN_swap.
6179 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6180 the exponentiation functions are more likely to produce reasonable
6181 results on negative inputs.
6184 *) Change BN_mod_mul so that the result is always non-negative.
6185 Previously, it could be negative if one of the factors was negative;
6186 I don't think anyone really wanted that behaviour.
6189 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6190 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6191 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6192 and add new functions:
6201 BN_mod_lshift1_quick
6205 These functions always generate non-negative results.
6207 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6208 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6210 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6211 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6212 be reduced modulo m.
6213 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6216 The following entry accidentily appeared in the CHANGES file
6217 distributed with OpenSSL 0.9.7. The modifications described in
6218 it do *not* apply to OpenSSL 0.9.7.
6220 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6221 was actually never needed) and in BN_mul(). The removal in BN_mul()
6222 required a small change in bn_mul_part_recursive() and the addition
6223 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6224 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6225 bn_sub_words() and bn_add_words() except they take arrays with
6230 *) In 'openssl passwd', verify passwords read from the terminal
6231 unless the '-salt' option is used (which usually means that
6232 verification would just waste user's time since the resulting
6233 hash is going to be compared with some given password hash)
6234 or the new '-noverify' option is used.
6236 This is an incompatible change, but it does not affect
6237 non-interactive use of 'openssl passwd' (passwords on the command
6238 line, '-stdin' option, '-in ...' option) and thus should not
6242 *) Remove all references to RSAref, since there's no more need for it.
6245 *) Make DSO load along a path given through an environment variable
6246 (SHLIB_PATH) with shl_load().
6249 *) Constify the ENGINE code as a result of BIGNUM constification.
6250 Also constify the RSA code and most things related to it. In a
6251 few places, most notable in the depth of the ASN.1 code, ugly
6252 casts back to non-const were required (to be solved at a later
6256 *) Make it so the openssl application has all engines loaded by default.
6259 *) Constify the BIGNUM routines a little more.
6262 *) Add the following functions:
6264 ENGINE_load_cswift()
6266 ENGINE_load_atalla()
6268 ENGINE_load_builtin_engines()
6270 That way, an application can itself choose if external engines that
6271 are built-in in OpenSSL shall ever be used or not. The benefit is
6272 that applications won't have to be linked with libdl or other dso
6273 libraries unless it's really needed.
6275 Changed 'openssl engine' to load all engines on demand.
6276 Changed the engine header files to avoid the duplication of some
6277 declarations (they differed!).
6280 *) 'openssl engine' can now list capabilities.
6283 *) Better error reporting in 'openssl engine'.
6286 *) Never call load_dh_param(NULL) in s_server.
6289 *) Add engine application. It can currently list engines by name and
6290 identity, and test if they are actually available.
6293 *) Improve RPM specification file by forcing symbolic linking and making
6294 sure the installed documentation is also owned by root.root.
6295 [Damien Miller <djm@mindrot.org>]
6297 *) Give the OpenSSL applications more possibilities to make use of
6298 keys (public as well as private) handled by engines.
6301 *) Add OCSP code that comes from CertCo.
6304 *) Add VMS support for the Rijndael code.
6307 *) Added untested support for Nuron crypto accelerator.
6310 *) Add support for external cryptographic devices. This code was
6311 previously distributed separately as the "engine" branch.
6312 [Geoff Thorpe, Richard Levitte]
6314 *) Rework the filename-translation in the DSO code. It is now possible to
6315 have far greater control over how a "name" is turned into a filename
6316 depending on the operating environment and any oddities about the
6317 different shared library filenames on each system.
6320 *) Support threads on FreeBSD-elf in Configure.
6323 *) Fix for SHA1 assembly problem with MASM: it produces
6324 warnings about corrupt line number information when assembling
6325 with debugging information. This is caused by the overlapping
6327 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6330 NCONF_get_number() has no error checking at all. As a replacement,
6331 NCONF_get_number_e() is defined (_e for "error checking") and is
6332 promoted strongly. The old NCONF_get_number is kept around for
6333 binary backward compatibility.
6334 Make it possible for methods to load from something other than a BIO,
6335 by providing a function pointer that is given a name instead of a BIO.
6336 For example, this could be used to load configuration data from an
6340 *) Fix for non blocking accept BIOs. Added new I/O special reason
6341 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6342 with non blocking I/O was not possible because no retry code was
6343 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6347 *) Added the beginnings of Rijndael support.
6350 *) Fix for bug in DirectoryString mask setting. Add support for
6351 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6352 to allow certificate printing to more controllable, additional
6353 'certopt' option to 'x509' to allow new printing options to be
6357 *) Clean old EAY MD5 hack from e_os.h.
6360 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6362 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6363 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6364 [Joe Orton, Steve Henson]
6366 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6368 *) Fix additional bug revealed by the NISCC test suite:
6370 Stop bug triggering large recursion when presented with
6371 certain ASN.1 tags (CVE-2003-0851)
6374 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6376 *) Fix various bugs revealed by running the NISCC test suite:
6378 Stop out of bounds reads in the ASN1 code when presented with
6379 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6381 If verify callback ignores invalid public key errors don't try to check
6382 certificate signature with the NULL public key.
6386 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6387 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6391 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6392 extra data after the compression methods not only for TLS 1.0
6393 but also for SSL 3.0 (as required by the specification).
6394 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6396 *) Change X509_certificate_type() to mark the key as exported/exportable
6397 when it's 512 *bits* long, not 512 bytes.
6400 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6402 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6403 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6404 a protocol version number mismatch like a decryption error
6405 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6408 *) Turn on RSA blinding by default in the default implementation
6409 to avoid a timing attack. Applications that don't want it can call
6410 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6411 They would be ill-advised to do so in most cases.
6412 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6414 *) Change RSA blinding code so that it works when the PRNG is not
6415 seeded (in this case, the secret RSA exponent is abused as
6416 an unpredictable seed -- if it is not unpredictable, there
6417 is no point in blinding anyway). Make RSA blinding thread-safe
6418 by remembering the creator's thread ID in rsa->blinding and
6419 having all other threads use local one-time blinding factors
6420 (this requires more computation than sharing rsa->blinding, but
6421 avoids excessive locking; and if an RSA object is not shared
6422 between threads, blinding will still be very fast).
6425 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6427 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6428 via timing by performing a MAC computation even if incorrrect
6429 block cipher padding has been found. This is a countermeasure
6430 against active attacks where the attacker has to distinguish
6431 between bad padding and a MAC verification error. (CVE-2003-0078)
6433 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6434 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6435 Martin Vuagnoux (EPFL, Ilion)]
6437 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6439 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6440 memory from it's contents. This is done with a counter that will
6441 place alternating values in each byte. This can be used to solve
6442 two issues: 1) the removal of calls to memset() by highly optimizing
6443 compilers, and 2) cleansing with other values than 0, since those can
6444 be read through on certain media, for example a swap space on disk.
6447 *) Bugfix: client side session caching did not work with external caching,
6448 because the session->cipher setting was not restored when reloading
6449 from the external cache. This problem was masked, when
6450 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6451 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6454 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6455 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6456 [Zeev Lieber <zeev-l@yahoo.com>]
6458 *) Undo an undocumented change introduced in 0.9.6e which caused
6459 repeated calls to OpenSSL_add_all_ciphers() and
6460 OpenSSL_add_all_digests() to be ignored, even after calling
6464 *) Change the default configuration reader to deal with last line not
6465 being properly terminated.
6468 *) Change X509_NAME_cmp() so it applies the special rules on handling
6469 DN values that are of type PrintableString, as well as RDNs of type
6470 emailAddress where the value has the type ia5String.
6471 [stefank@valicert.com via Richard Levitte]
6473 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6474 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6475 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6476 the bitwise-OR of the two for use by the majority of applications
6477 wanting this behaviour, and update the docs. The documented
6478 behaviour and actual behaviour were inconsistent and had been
6479 changing anyway, so this is more a bug-fix than a behavioural
6481 [Geoff Thorpe, diagnosed by Nadav Har'El]
6483 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6484 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6487 *) Fix initialization code race conditions in
6488 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6489 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6490 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6491 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6492 ssl2_get_cipher_by_char(),
6493 ssl3_get_cipher_by_char().
6494 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6496 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6497 the cached sessions are flushed, as the remove_cb() might use ex_data
6498 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6499 (see [openssl.org #212]).
6500 [Geoff Thorpe, Lutz Jaenicke]
6502 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6503 length, instead of the encoding length to d2i_ASN1_OBJECT.
6506 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6508 *) [In 0.9.6g-engine release:]
6509 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6510 [Lynn Gazis <lgazis@rainbow.com>]
6512 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6514 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6515 and get fix the header length calculation.
6516 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6517 Alon Kantor <alonk@checkpoint.com> (and others),
6520 *) Use proper error handling instead of 'assertions' in buffer
6521 overflow checks added in 0.9.6e. This prevents DoS (the
6522 assertions could call abort()).
6523 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6525 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6527 *) Add various sanity checks to asn1_get_length() to reject
6528 the ASN1 length bytes if they exceed sizeof(long), will appear
6529 negative or the content length exceeds the length of the
6531 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6533 *) Fix cipher selection routines: ciphers without encryption had no flags
6534 for the cipher strength set and where therefore not handled correctly
6535 by the selection routines (PR #130).
6538 *) Fix EVP_dsa_sha macro.
6542 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6543 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6544 that was added in OpenSSL 0.9.6d.
6546 As the countermeasure turned out to be incompatible with some
6547 broken SSL implementations, the new option is part of SSL_OP_ALL.
6548 SSL_OP_ALL is usually employed when compatibility with weird SSL
6549 implementations is desired (e.g. '-bugs' option to 's_client' and
6550 's_server'), so the new option is automatically set in many
6554 *) Changes in security patch:
6556 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6557 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6558 Air Force Materiel Command, USAF, under agreement number
6561 *) Add various sanity checks to asn1_get_length() to reject
6562 the ASN1 length bytes if they exceed sizeof(long), will appear
6563 negative or the content length exceeds the length of the
6564 supplied buffer. (CVE-2002-0659)
6565 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6567 *) Assertions for various potential buffer overflows, not known to
6569 [Ben Laurie (CHATS)]
6571 *) Various temporary buffers to hold ASCII versions of integers were
6572 too small for 64 bit platforms. (CVE-2002-0655)
6573 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6575 *) Remote buffer overflow in SSL3 protocol - an attacker could
6576 supply an oversized session ID to a client. (CVE-2002-0656)
6577 [Ben Laurie (CHATS)]
6579 *) Remote buffer overflow in SSL2 protocol - an attacker could
6580 supply an oversized client master key. (CVE-2002-0656)
6581 [Ben Laurie (CHATS)]
6583 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6585 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6586 encoded as NULL) with id-dsa-with-sha1.
6587 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6589 *) Check various X509_...() return values in apps/req.c.
6590 [Nils Larsch <nla@trustcenter.de>]
6592 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6593 an end-of-file condition would erronously be flagged, when the CRLF
6594 was just at the end of a processed block. The bug was discovered when
6595 processing data through a buffering memory BIO handing the data to a
6596 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6597 <ptsekov@syntrex.com> and Nedelcho Stanev.
6600 *) Implement a countermeasure against a vulnerability recently found
6601 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6602 before application data chunks to avoid the use of known IVs
6603 with data potentially chosen by the attacker.
6606 *) Fix length checks in ssl3_get_client_hello().
6609 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6610 to prevent ssl3_read_internal() from incorrectly assuming that
6611 ssl3_read_bytes() found application data while handshake
6612 processing was enabled when in fact s->s3->in_read_app_data was
6613 merely automatically cleared during the initial handshake.
6614 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6616 *) Fix object definitions for Private and Enterprise: they were not
6617 recognized in their shortname (=lowercase) representation. Extend
6618 obj_dat.pl to issue an error when using undefined keywords instead
6619 of silently ignoring the problem (Svenning Sorensen
6620 <sss@sss.dnsalias.net>).
6623 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6624 generators, i.e. generators other than 2 and 5. (Previously, the
6625 code did not properly initialise the 'add' and 'rem' values to
6626 BN_generate_prime().)
6628 In the new general case, we do not insist that 'generator' is
6629 actually a primitive root: This requirement is rather pointless;
6630 a generator of the order-q subgroup is just as good, if not
6634 *) Map new X509 verification errors to alerts. Discovered and submitted by
6635 Tom Wu <tom@arcot.com>.
6638 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6639 returning non-zero before the data has been completely received
6640 when using non-blocking I/O.
6641 [Bodo Moeller; problem pointed out by John Hughes]
6643 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6644 [Ben Laurie, Lutz Jaenicke]
6646 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6647 Yoram Zahavi <YoramZ@gilian.com>).
6650 *) Add information about CygWin 1.3 and on, and preserve proper
6651 configuration for the versions before that.
6652 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6654 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6655 check whether we deal with a copy of a session and do not delete from
6656 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6657 <izhar@checkpoint.com>.
6660 *) Do not store session data into the internal session cache, if it
6661 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6662 flag is set). Proposed by Aslam <aslam@funk.com>.
6665 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6669 *) [In 0.9.6d-engine release:]
6670 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6671 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6673 *) Add the configuration target linux-s390x.
6674 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6676 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6677 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6678 variable as an indication that a ClientHello message has been
6679 received. As the flag value will be lost between multiple
6680 invocations of ssl3_accept when using non-blocking I/O, the
6681 function may not be aware that a handshake has actually taken
6682 place, thus preventing a new session from being added to the
6685 To avoid this problem, we now set s->new_session to 2 instead of
6686 using a local variable.
6687 [Lutz Jaenicke, Bodo Moeller]
6689 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6690 if the SSL_R_LENGTH_MISMATCH error is detected.
6691 [Geoff Thorpe, Bodo Moeller]
6693 *) New 'shared_ldflag' column in Configure platform table.
6696 *) Fix EVP_CIPHER_mode macro.
6697 ["Dan S. Camper" <dan@bti.net>]
6699 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6700 type, we must throw them away by setting rr->length to 0.
6701 [D P Chang <dpc@qualys.com>]
6703 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6705 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6706 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6707 worked incorrectly for those cases where range = 10..._2 and
6708 3*range is two bits longer than range.)
6711 *) Only add signing time to PKCS7 structures if it is not already
6715 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6716 OBJ_ld_ce should be OBJ_id_ce.
6717 Also some ip-pda OIDs in crypto/objects/objects.txt were
6718 incorrect (cf. RFC 3039).
6719 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6721 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6722 returns early because it has nothing to do.
6723 [Andy Schneider <andy.schneider@bjss.co.uk>]
6725 *) [In 0.9.6c-engine release:]
6726 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6727 [Andy Schneider <andy.schneider@bjss.co.uk>]
6729 *) [In 0.9.6c-engine release:]
6730 Add support for Cryptographic Appliance's keyserver technology.
6731 (Use engine 'keyclient')
6732 [Cryptographic Appliances and Geoff Thorpe]
6734 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6735 is called via tools/c89.sh because arguments have to be
6736 rearranged (all '-L' options must appear before the first object
6738 [Richard Shapiro <rshapiro@abinitio.com>]
6740 *) [In 0.9.6c-engine release:]
6741 Add support for Broadcom crypto accelerator cards, backported
6743 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6745 *) [In 0.9.6c-engine release:]
6746 Add support for SureWare crypto accelerator cards from
6747 Baltimore Technologies. (Use engine 'sureware')
6748 [Baltimore Technologies and Mark Cox]
6750 *) [In 0.9.6c-engine release:]
6751 Add support for crypto accelerator cards from Accelerated
6752 Encryption Processing, www.aep.ie. (Use engine 'aep')
6753 [AEP Inc. and Mark Cox]
6755 *) Add a configuration entry for gcc on UnixWare.
6756 [Gary Benson <gbenson@redhat.com>]
6758 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6759 messages are stored in a single piece (fixed-length part and
6760 variable-length part combined) and fix various bugs found on the way.
6763 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6764 instead. BIO_gethostbyname() does not know what timeouts are
6765 appropriate, so entries would stay in cache even when they have
6767 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6769 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6770 faced with a pathologically small ClientHello fragment that does
6771 not contain client_version: Instead of aborting with an error,
6772 simply choose the highest available protocol version (i.e.,
6773 TLS 1.0 unless it is disabled). In practice, ClientHello
6774 messages are never sent like this, but this change gives us
6775 strictly correct behaviour at least for TLS.
6778 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6779 never resets s->method to s->ctx->method when called from within
6780 one of the SSL handshake functions.
6781 [Bodo Moeller; problem pointed out by Niko Baric]
6783 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6784 (sent using the client's version number) if client_version is
6785 smaller than the protocol version in use. Also change
6786 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6787 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6788 the client will at least see that alert.
6791 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6795 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6796 client receives HelloRequest while in a handshake.
6797 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6799 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6800 should end in 'break', not 'goto end' which circuments various
6801 cleanups done in state SSL_ST_OK. But session related stuff
6802 must be disabled for SSL_ST_OK in the case that we just sent a
6805 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6806 before just sending a HelloRequest.
6807 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6809 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6810 reveal whether illegal block cipher padding was found or a MAC
6811 verification error occured. (Neither SSLerr() codes nor alerts
6812 are directly visible to potential attackers, but the information
6813 may leak via logfiles.)
6815 Similar changes are not required for the SSL 2.0 implementation
6816 because the number of padding bytes is sent in clear for SSL 2.0,
6817 and the extra bytes are just ignored. However ssl/s2_pkt.c
6818 failed to verify that the purported number of padding bytes is in
6822 *) Add OpenUNIX-8 support including shared libraries
6823 (Boyd Lynn Gerber <gerberb@zenez.com>).
6826 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6827 'wristwatch attack' using huge encoding parameters (cf.
6828 James H. Manger's CRYPTO 2001 paper). Note that the
6829 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6830 encoding parameters and hence was not vulnerable.
6833 *) BN_sqr() bug fix.
6834 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6836 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6837 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6838 followed by modular reduction.
6839 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6841 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6842 equivalent based on BN_pseudo_rand() instead of BN_rand().
6845 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6846 This function was broken, as the check for a new client hello message
6847 to handle SGC did not allow these large messages.
6848 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6851 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6854 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6855 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6858 *) Rework the configuration and shared library support for Tru64 Unix.
6859 The configuration part makes use of modern compiler features and
6860 still retains old compiler behavior for those that run older versions
6861 of the OS. The shared library support part includes a variant that
6862 uses the RPATH feature, and is available through the special
6863 configuration target "alpha-cc-rpath", which will never be selected
6865 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6867 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6868 with the same message size as in ssl3_get_certificate_request().
6869 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6870 messages might inadvertently be reject as too long.
6871 [Petr Lampa <lampa@fee.vutbr.cz>]
6873 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6876 *) Modified SSL library such that the verify_callback that has been set
6877 specificly for an SSL object with SSL_set_verify() is actually being
6878 used. Before the change, a verify_callback set with this function was
6879 ignored and the verify_callback() set in the SSL_CTX at the time of
6880 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6881 to allow the necessary settings.
6884 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6885 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6886 done automatically (in contradiction to the requirements of the C
6887 standard). This made problems when used from OpenSSH.
6890 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6891 dh->length and always used
6893 BN_rand_range(priv_key, dh->p).
6895 BN_rand_range() is not necessary for Diffie-Hellman, and this
6896 specific range makes Diffie-Hellman unnecessarily inefficient if
6897 dh->length (recommended exponent length) is much smaller than the
6898 length of dh->p. We could use BN_rand_range() if the order of
6899 the subgroup was stored in the DH structure, but we only have
6904 BN_rand(priv_key, l, ...)
6906 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6912 RSA_eay_public_encrypt
6913 RSA_eay_private_decrypt
6914 RSA_eay_private_encrypt (signing)
6915 RSA_eay_public_decrypt (signature verification)
6917 (default implementations for RSA_public_encrypt,
6918 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6919 always reject numbers >= n.
6922 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6923 to synchronize access to 'locking_thread'. This is necessary on
6924 systems where access to 'locking_thread' (an 'unsigned long'
6925 variable) is not atomic.
6928 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6929 *before* setting the 'crypto_lock_rand' flag. The previous code had
6930 a race condition if 0 is a valid thread ID.
6931 [Travis Vitek <vitek@roguewave.com>]
6933 *) Add support for shared libraries under Irix.
6934 [Albert Chin-A-Young <china@thewrittenword.com>]
6936 *) Add configuration option to build on Linux on both big-endian and
6938 [Ralf Baechle <ralf@uni-koblenz.de>]
6940 *) Add the possibility to create shared libraries on HP-UX.
6943 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6945 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6946 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6947 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6948 PRNG state recovery was possible based on the output of
6949 one PRNG request appropriately sized to gain knowledge on
6950 'md' followed by enough consecutive 1-byte PRNG requests
6951 to traverse all of 'state'.
6953 1. When updating 'md_local' (the current thread's copy of 'md')
6954 during PRNG output generation, hash all of the previous
6955 'md_local' value, not just the half used for PRNG output.
6957 2. Make the number of bytes from 'state' included into the hash
6958 independent from the number of PRNG bytes requested.
6960 The first measure alone would be sufficient to avoid
6961 Markku-Juhani's attack. (Actually it had never occurred
6962 to me that the half of 'md_local' used for chaining was the
6963 half from which PRNG output bytes were taken -- I had always
6964 assumed that the secret half would be used.) The second
6965 measure makes sure that additional data from 'state' is never
6966 mixed into 'md_local' in small portions; this heuristically
6967 further strengthens the PRNG.
6970 *) Fix crypto/bn/asm/mips3.s.
6973 *) When only the key is given to "enc", the IV is undefined. Print out
6974 an error message in this case.
6977 *) Handle special case when X509_NAME is empty in X509 printing routines.
6980 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6981 positive and less than q.
6984 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6985 used: it isn't thread safe and the add_lock_callback should handle
6987 [Paul Rose <Paul.Rose@bridge.com>]
6989 *) Verify that incoming data obeys the block size in
6990 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6994 [Ulf Möller, Bodo Möller]
6996 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6997 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6998 when fixing the server behaviour for backwards-compatible 'client
6999 hello' messages. (Note that the attack is impractical against
7000 SSL 3.0 and TLS 1.0 anyway because length and version checking
7001 means that the probability of guessing a valid ciphertext is
7002 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7005 Before 0.9.5, the countermeasure (hide the error by generating a
7006 random 'decryption result') did not work properly because
7007 ERR_clear_error() was missing, meaning that SSL_get_error() would
7008 detect the supposedly ignored error.
7010 Both problems are now fixed.
7013 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7014 (previously it was 1024).
7017 *) Fix for compatibility mode trust settings: ignore trust settings
7018 unless some valid trust or reject settings are present.
7021 *) Fix for blowfish EVP: its a variable length cipher.
7024 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7025 parameters in DSA public key structures and return an error in the
7026 DSA routines if parameters are absent.
7029 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7030 in the current directory if neither $RANDFILE nor $HOME was set.
7031 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7032 caused some confusion to Windows users who haven't defined $HOME.
7033 Thus RAND_file_name() is changed again: e_os.h can define a
7034 DEFAULT_HOME, which will be used if $HOME is not set.
7035 For Windows, we use "C:"; on other platforms, we still require
7036 environment variables.
7038 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7039 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7040 having multiple threads call RAND_poll() concurrently.
7043 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7044 combination of a flag and a thread ID variable.
7045 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7046 flag), *other* threads can enter ssleay_add_bytes without obeying
7047 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7048 that they do not hold after the first thread unsets add_do_not_lock).
7051 *) Change bctest again: '-x' expressions are not available in all
7055 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7057 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7058 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7060 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7061 the default extension for executables, if any. Also, make the perl
7062 scripts that use symlink() to test if it really exists and use "cp"
7063 if it doesn't. All this made OpenSSL compilable and installable in
7067 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7068 If SEQUENCE is length is indefinite just set c->slen to the total
7069 amount of data available.
7070 [Steve Henson, reported by shige@FreeBSD.org]
7071 [This change does not apply to 0.9.7.]
7073 *) Change bctest to avoid here-documents inside command substitution
7074 (workaround for FreeBSD /bin/sh bug).
7075 For compatibility with Ultrix, avoid shell functions (introduced
7076 in the bctest version that searches along $PATH).
7079 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7080 with des_encrypt() defined on some operating systems, like Solaris
7084 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7085 On the Importance of Eliminating Errors in Cryptographic
7086 Computations, J. Cryptology 14 (2001) 2, 101-119,
7087 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7090 *) MIPS assembler BIGNUM division bug fix.
7093 *) Disabled incorrect Alpha assembler code.
7096 *) Fix PKCS#7 decode routines so they correctly update the length
7097 after reading an EOC for the EXPLICIT tag.
7099 [This change does not apply to 0.9.7.]
7101 *) Fix bug in PKCS#12 key generation routines. This was triggered
7102 if a 3DES key was generated with a 0 initial byte. Include
7103 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7104 (but broken) behaviour.
7107 *) Enhance bctest to search for a working bc along $PATH and print
7109 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7111 *) Fix memory leaks in err.c: free err_data string if necessary;
7112 don't write to the wrong index in ERR_set_error_data.
7115 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7119 *) Replace rdtsc with _emit statements for VC++ version 5.
7120 [Jeremy Cooper <jeremy@baymoo.org>]
7122 *) Make it possible to reuse SSLv2 sessions.
7125 *) In copy_email() check for >= 0 as a return value for
7126 X509_NAME_get_index_by_NID() since 0 is a valid index.
7127 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7129 *) Avoid coredump with unsupported or invalid public keys by checking if
7130 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7131 PKCS7_verify() fails with non detached data.
7134 *) Don't use getenv in library functions when run as setuid/setgid.
7135 New function OPENSSL_issetugid().
7138 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7139 due to incorrect handling of multi-threading:
7141 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7143 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7145 3. Count how many times MemCheck_off() has been called so that
7146 nested use can be treated correctly. This also avoids
7147 inband-signalling in the previous code (which relied on the
7148 assumption that thread ID 0 is impossible).
7151 *) Add "-rand" option also to s_client and s_server.
7154 *) Fix CPU detection on Irix 6.x.
7155 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7156 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7158 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7161 [This change does not apply to 0.9.7.]
7163 *) Use the cached encoding of an X509_NAME structure rather than
7164 copying it. This is apparently the reason for the libsafe "errors"
7165 but the code is actually correct.
7168 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7169 Bleichenbacher's DSA attack.
7170 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7171 to be set and top=0 forces the highest bit to be set; top=-1 is new
7172 and leaves the highest bit random.
7173 [Ulf Moeller, Bodo Moeller]
7175 *) In the NCONF_...-based implementations for CONF_... queries
7176 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7177 a temporary CONF structure with the data component set to NULL
7178 (which gives segmentation faults in lh_retrieve).
7179 Instead, use NULL for the CONF pointer in CONF_get_string and
7180 CONF_get_number (which may use environment variables) and directly
7181 return NULL from CONF_get_section.
7184 *) Fix potential buffer overrun for EBCDIC.
7187 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7188 keyUsage if basicConstraints absent for a CA.
7191 *) Make SMIME_write_PKCS7() write mail header values with a format that
7192 is more generally accepted (no spaces before the semicolon), since
7193 some programs can't parse those values properly otherwise. Also make
7194 sure BIO's that break lines after each write do not create invalid
7198 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7199 macros previously used would not encode an empty SEQUENCE OF
7200 and break the signature.
7202 [This change does not apply to 0.9.7.]
7204 *) Zero the premaster secret after deriving the master secret in
7208 *) Add some EVP_add_digest_alias registrations (as found in
7209 OpenSSL_add_all_digests()) to SSL_library_init()
7210 aka OpenSSL_add_ssl_algorithms(). This provides improved
7211 compatibility with peers using X.509 certificates
7212 with unconventional AlgorithmIdentifier OIDs.
7215 *) Fix for Irix with NO_ASM.
7216 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7218 *) ./config script fixes.
7219 [Ulf Moeller, Richard Levitte]
7221 *) Fix 'openssl passwd -1'.
7224 *) Change PKCS12_key_gen_asc() so it can cope with non null
7225 terminated strings whose length is passed in the passlen
7226 parameter, for example from PEM callbacks. This was done
7227 by adding an extra length parameter to asc2uni().
7228 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7230 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7231 call failed, free the DSA structure.
7234 *) Fix to uni2asc() to cope with zero length Unicode strings.
7235 These are present in some PKCS#12 files.
7238 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7239 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7240 when writing a 32767 byte record.
7241 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7243 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7244 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7246 (RSA objects have a reference count access to which is protected
7247 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7248 so they are meant to be shared between threads.)
7249 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7250 "Reddie, Steven" <Steven.Reddie@ca.com>]
7252 *) Fix a deadlock in CRYPTO_mem_leaks().
7255 *) Use better test patterns in bntest.
7258 *) rand_win.c fix for Borland C.
7261 *) BN_rshift bugfix for n == 0.
7264 *) Add a 'bctest' script that checks for some known 'bc' bugs
7265 so that 'make test' does not abort just because 'bc' is broken.
7268 *) Store verify_result within SSL_SESSION also for client side to
7269 avoid potential security hole. (Re-used sessions on the client side
7270 always resulted in verify_result==X509_V_OK, not using the original
7271 result of the server certificate verification.)
7274 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7275 SSL3_RT_APPLICATION_DATA, return 0.
7276 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7280 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7281 releases, have been re-implemented by renaming the previous
7282 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7283 and ssl3_read_internal, respectively, and adding 'peek' parameters
7284 to them. The new ssl[23]_{read,peek} functions are calls to
7285 ssl[23]_read_internal with the 'peek' flag set appropriately.
7286 A 'peek' parameter has also been added to ssl3_read_bytes, which
7287 does the actual work for ssl3_read_internal.
7290 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7291 the method-specific "init()" handler. Also clean up ex_data after
7292 calling the method-specific "finish()" handler. Previously, this was
7293 happening the other way round.
7296 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7297 The previous value, 12, was not always sufficient for BN_mod_exp().
7300 *) Make sure that shared libraries get the internal name engine with
7301 the full version number and not just 0. This should mark the
7302 shared libraries as not backward compatible. Of course, this should
7303 be changed again when we can guarantee backward binary compatibility.
7306 *) Fix typo in get_cert_by_subject() in by_dir.c
7307 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7309 *) Rework the system to generate shared libraries:
7311 - Make note of the expected extension for the shared libraries and
7312 if there is a need for symbolic links from for example libcrypto.so.0
7313 to libcrypto.so.0.9.7. There is extended info in Configure for
7316 - Make as few rebuilds of the shared libraries as possible.
7318 - Still avoid linking the OpenSSL programs with the shared libraries.
7320 - When installing, install the shared libraries separately from the
7324 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7326 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7327 and not in SSL_clear because the latter is also used by the
7328 accept/connect functions; previously, the settings made by
7329 SSL_set_read_ahead would be lost during the handshake.
7330 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7332 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7333 Previously, it would create entries for disableed algorithms no
7337 *) Added several new manual pages for SSL_* function.
7340 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7342 *) In ssl23_get_client_hello, generate an error message when faced
7343 with an initial SSL 3.0/TLS record that is too small to contain the
7344 first two bytes of the ClientHello message, i.e. client_version.
7345 (Note that this is a pathologic case that probably has never happened
7346 in real life.) The previous approach was to use the version number
7347 from the record header as a substitute; but our protocol choice
7348 should not depend on that one because it is not authenticated
7349 by the Finished messages.
7352 *) More robust randomness gathering functions for Windows.
7353 [Jeffrey Altman <jaltman@columbia.edu>]
7355 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7356 not set then we don't setup the error code for issuer check errors
7357 to avoid possibly overwriting other errors which the callback does
7358 handle. If an application does set the flag then we assume it knows
7359 what it is doing and can handle the new informational codes
7363 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7364 a general "ANY" type, as such it should be able to decode anything
7365 including tagged types. However it didn't check the class so it would
7366 wrongly interpret tagged types in the same way as their universal
7367 counterpart and unknown types were just rejected. Changed so that the
7368 tagged and unknown types are handled in the same way as a SEQUENCE:
7369 that is the encoding is stored intact. There is also a new type
7370 "V_ASN1_OTHER" which is used when the class is not universal, in this
7371 case we have no idea what the actual type is so we just lump them all
7375 *) On VMS, stdout may very well lead to a file that is written to
7376 in a record-oriented fashion. That means that every write() will
7377 write a separate record, which will be read separately by the
7378 programs trying to read from it. This can be very confusing.
7380 The solution is to put a BIO filter in the way that will buffer
7381 text until a linefeed is reached, and then write everything a
7382 line at a time, so every record written will be an actual line,
7383 not chunks of lines and not (usually doesn't happen, but I've
7384 seen it once) several lines in one record. BIO_f_linebuffer() is
7387 Currently, it's a VMS-only method, because that's where it has
7388 been tested well enough.
7391 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7392 it can return incorrect results.
7393 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7394 but it was in 0.9.6-beta[12].)
7397 *) Disable the check for content being present when verifying detached
7398 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7399 include zero length content when signing messages.
7402 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7403 BIO_ctrl (for BIO pairs).
7406 *) Add DSO method for VMS.
7409 *) Bug fix: Montgomery multiplication could produce results with the
7413 *) Add RPM specification openssl.spec and modify it to build three
7414 packages. The default package contains applications, application
7415 documentation and run-time libraries. The devel package contains
7416 include files, static libraries and function documentation. The
7417 doc package contains the contents of the doc directory. The original
7418 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7421 *) Add a large number of documentation files for many SSL routines.
7422 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7424 *) Add a configuration entry for Sony News 4.
7425 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7427 *) Don't set the two most significant bits to one when generating a
7428 random number < q in the DSA library.
7431 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7432 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7433 the underlying transport is blocking) if a handshake took place.
7434 (The default behaviour is needed by applications such as s_client
7435 and s_server that use select() to determine when to use SSL_read;
7436 but for applications that know in advance when to expect data, it
7437 just makes things more complicated.)
7440 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7444 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7445 work better on such systems.
7446 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7448 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7449 Update PKCS12_parse() so it copies the friendlyName and the
7450 keyid to the certificates aux info.
7453 *) Fix bug in PKCS7_verify() which caused an infinite loop
7454 if there was more than one signature.
7455 [Sven Uszpelkat <su@celocom.de>]
7457 *) Major change in util/mkdef.pl to include extra information
7458 about each symbol, as well as presentig variables as well
7459 as functions. This change means that there's n more need
7460 to rebuild the .num files when some algorithms are excluded.
7463 *) Allow the verify time to be set by an application,
7464 rather than always using the current time.
7467 *) Phase 2 verify code reorganisation. The certificate
7468 verify code now looks up an issuer certificate by a
7469 number of criteria: subject name, authority key id
7470 and key usage. It also verifies self signed certificates
7471 by the same criteria. The main comparison function is
7472 X509_check_issued() which performs these checks.
7474 Lot of changes were necessary in order to support this
7475 without completely rewriting the lookup code.
7477 Authority and subject key identifier are now cached.
7479 The LHASH 'certs' is X509_STORE has now been replaced
7480 by a STACK_OF(X509_OBJECT). This is mainly because an
7481 LHASH can't store or retrieve multiple objects with
7482 the same hash value.
7484 As a result various functions (which were all internal
7485 use only) have changed to handle the new X509_STORE
7486 structure. This will break anything that messed round
7487 with X509_STORE internally.
7489 The functions X509_STORE_add_cert() now checks for an
7490 exact match, rather than just subject name.
7492 The X509_STORE API doesn't directly support the retrieval
7493 of multiple certificates matching a given criteria, however
7494 this can be worked round by performing a lookup first
7495 (which will fill the cache with candidate certificates)
7496 and then examining the cache for matches. This is probably
7497 the best we can do without throwing out X509_LOOKUP
7498 entirely (maybe later...).
7500 The X509_VERIFY_CTX structure has been enhanced considerably.
7502 All certificate lookup operations now go via a get_issuer()
7503 callback. Although this currently uses an X509_STORE it
7504 can be replaced by custom lookups. This is a simple way
7505 to bypass the X509_STORE hackery necessary to make this
7506 work and makes it possible to use more efficient techniques
7507 in future. A very simple version which uses a simple
7508 STACK for its trusted certificate store is also provided
7509 using X509_STORE_CTX_trusted_stack().
7511 The verify_cb() and verify() callbacks now have equivalents
7512 in the X509_STORE_CTX structure.
7514 X509_STORE_CTX also has a 'flags' field which can be used
7515 to customise the verify behaviour.
7518 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7519 excludes S/MIME capabilities.
7522 *) When a certificate request is read in keep a copy of the
7523 original encoding of the signed data and use it when outputing
7524 again. Signatures then use the original encoding rather than
7525 a decoded, encoded version which may cause problems if the
7526 request is improperly encoded.
7529 *) For consistency with other BIO_puts implementations, call
7530 buffer_write(b, ...) directly in buffer_puts instead of calling
7533 In BIO_puts, increment b->num_write as in BIO_write.
7534 [Peter.Sylvester@EdelWeb.fr]
7536 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7537 BN_zero, we may not return a BIGNUM with an array consisting of
7541 *) Avoid calling abort() from within the library when problems are
7542 detected, except if preprocessor symbols have been defined
7543 (such as REF_CHECK, BN_DEBUG etc.).
7546 *) New openssl application 'rsautl'. This utility can be
7547 used for low level RSA operations. DER public key
7548 BIO/fp routines also added.
7551 *) New Configure entry and patches for compiling on QNX 4.
7552 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7554 *) A demo state-machine implementation was sponsored by
7555 Nuron (http://www.nuron.com/) and is now available in
7556 demos/state_machine.
7559 *) New options added to the 'dgst' utility for signature
7560 generation and verification.
7563 *) Unrecognized PKCS#7 content types are now handled via a
7564 catch all ASN1_TYPE structure. This allows unsupported
7565 types to be stored as a "blob" and an application can
7566 encode and decode it manually.
7569 *) Fix various signed/unsigned issues to make a_strex.c
7571 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7573 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7574 length if passed a buffer. ASN1_INTEGER_to_BN failed
7575 if passed a NULL BN and its argument was negative.
7576 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7578 *) Modification to PKCS#7 encoding routines to output definite
7579 length encoding. Since currently the whole structures are in
7580 memory there's not real point in using indefinite length
7581 constructed encoding. However if OpenSSL is compiled with
7582 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7585 *) Added BIO_vprintf() and BIO_vsnprintf().
7588 *) Added more prefixes to parse for in the the strings written
7589 through a logging bio, to cover all the levels that are available
7590 through syslog. The prefixes are now:
7592 PANIC, EMERG, EMR => LOG_EMERG
7593 ALERT, ALR => LOG_ALERT
7594 CRIT, CRI => LOG_CRIT
7595 ERROR, ERR => LOG_ERR
7596 WARNING, WARN, WAR => LOG_WARNING
7597 NOTICE, NOTE, NOT => LOG_NOTICE
7598 INFO, INF => LOG_INFO
7599 DEBUG, DBG => LOG_DEBUG
7601 and as before, if none of those prefixes are present at the
7602 beginning of the string, LOG_ERR is chosen.
7604 On Win32, the LOG_* levels are mapped according to this:
7606 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7607 LOG_WARNING => EVENTLOG_WARNING_TYPE
7608 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7612 *) Made it possible to reconfigure with just the configuration
7613 argument "reconf" or "reconfigure". The command line arguments
7614 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7615 and are retrieved from there when reconfiguring.
7619 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7621 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7624 *) The obj_dat.pl script was messing up the sorting of object
7625 names. The reason was that it compared the quoted version
7626 of strings as a result "OCSP" > "OCSP Signing" because
7627 " > SPACE. Changed script to store unquoted versions of
7628 names and add quotes on output. It was also omitting some
7629 names from the lookup table if they were given a default
7630 value (that is if SN is missing it is given the same
7631 value as LN and vice versa), these are now added on the
7632 grounds that if an object has a name we should be able to
7633 look it up. Finally added warning output when duplicate
7634 short or long names are found.
7637 *) Changes needed for Tandem NSK.
7638 [Scott Uroff <scott@xypro.com>]
7640 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7641 RSA_padding_check_SSLv23(), special padding was never detected
7642 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7643 version rollback attacks was not effective.
7645 In s23_clnt.c, don't use special rollback-attack detection padding
7646 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7647 client; similarly, in s23_srvr.c, don't do the rollback check if
7648 SSL 2.0 is the only protocol enabled in the server.
7651 *) Make it possible to get hexdumps of unprintable data with 'openssl
7652 asn1parse'. By implication, the functions ASN1_parse_dump() and
7653 BIO_dump_indent() are added.
7656 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7657 these print out strings and name structures based on various
7658 flags including RFC2253 support and proper handling of
7659 multibyte characters. Added options to the 'x509' utility
7660 to allow the various flags to be set.
7663 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7664 Also change the functions X509_cmp_current_time() and
7665 X509_gmtime_adj() work with an ASN1_TIME structure,
7666 this will enable certificates using GeneralizedTime in validity
7667 dates to be checked.
7670 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7671 negative public key encodings) on by default,
7672 NO_NEG_PUBKEY_BUG can be set to disable it.
7675 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7676 content octets. An i2c_ASN1_OBJECT is unnecessary because
7677 the encoding can be trivially obtained from the structure.
7680 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7681 not read locks (CRYPTO_r_[un]lock).
7684 *) A first attempt at creating official support for shared
7685 libraries through configuration. I've kept it so the
7686 default is static libraries only, and the OpenSSL programs
7687 are always statically linked for now, but there are
7688 preparations for dynamic linking in place.
7689 This has been tested on Linux and Tru64.
7692 *) Randomness polling function for Win9x, as described in:
7693 Peter Gutmann, Software Generation of Practically Strong
7697 *) Fix so PRNG is seeded in req if using an already existing
7701 *) New options to smime application. -inform and -outform
7702 allow alternative formats for the S/MIME message including
7703 PEM and DER. The -content option allows the content to be
7704 specified separately. This should allow things like Netscape
7705 form signing output easier to verify.
7708 *) Fix the ASN1 encoding of tags using the 'long form'.
7711 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7712 STRING types. These convert content octets to and from the
7713 underlying type. The actual tag and length octets are
7714 already assumed to have been read in and checked. These
7715 are needed because all other string types have virtually
7716 identical handling apart from the tag. By having versions
7717 of the ASN1 functions that just operate on content octets
7718 IMPLICIT tagging can be handled properly. It also allows
7719 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7720 and ASN1_INTEGER are identical apart from the tag.
7723 *) Change the handling of OID objects as follows:
7725 - New object identifiers are inserted in objects.txt, following
7726 the syntax given in objects.README.
7727 - objects.pl is used to process obj_mac.num and create a new
7729 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7732 This is currently kind of a hack, and the perl code in objects.pl
7733 isn't very elegant, but it works as I intended. The simplest way
7734 to check that it worked correctly is to look in obj_dat.h and
7735 check the array nid_objs and make sure the objects haven't moved
7736 around (this is important!). Additions are OK, as well as
7737 consistent name changes.
7740 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7743 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7744 The given file adds to whatever has already been seeded into the
7745 random pool through the RANDFILE configuration file option or
7746 environment variable, or the default random state file.
7749 *) mkstack.pl now sorts each macro group into lexical order.
7750 Previously the output order depended on the order the files
7751 appeared in the directory, resulting in needless rewriting
7755 *) Patches to make OpenSSL compile under Win32 again. Mostly
7756 work arounds for the VC++ problem that it treats func() as
7757 func(void). Also stripped out the parts of mkdef.pl that
7758 added extra typesafe functions: these no longer exist.
7761 *) Reorganisation of the stack code. The macros are now all
7762 collected in safestack.h . Each macro is defined in terms of
7763 a "stack macro" of the form SKM_<name>(type, a, b). The
7764 DEBUG_SAFESTACK is now handled in terms of function casts,
7765 this has the advantage of retaining type safety without the
7766 use of additional functions. If DEBUG_SAFESTACK is not defined
7767 then the non typesafe macros are used instead. Also modified the
7768 mkstack.pl script to handle the new form. Needs testing to see
7769 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7770 the default if no major problems. Similar behaviour for ASN1_SET_OF
7771 and PKCS12_STACK_OF.
7774 *) When some versions of IIS use the 'NET' form of private key the
7775 key derivation algorithm is different. Normally MD5(password) is
7776 used as a 128 bit RC4 key. In the modified case
7777 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7778 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7779 as the old Netscape_RSA functions except they have an additional
7780 'sgckey' parameter which uses the modified algorithm. Also added
7781 an -sgckey command line option to the rsa utility. Thanks to
7782 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7783 algorithm to openssl-dev.
7786 *) The evp_local.h macros were using 'c.##kname' which resulted in
7787 invalid expansion on some systems (SCO 5.0.5 for example).
7788 Corrected to 'c.kname'.
7789 [Phillip Porch <root@theporch.com>]
7791 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7792 a STACK of email addresses from a certificate or request, these look
7793 in the subject name and the subject alternative name extensions and
7794 omit any duplicate addresses.
7797 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7798 This makes DSA verification about 2 % faster.
7801 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7802 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7803 plus overhead for 1024 bit moduli).
7804 This makes exponentiations about 0.5 % faster for 1024 bit
7805 exponents (as measured by "openssl speed rsa2048").
7808 *) Rename memory handling macros to avoid conflicts with other
7810 Malloc => OPENSSL_malloc
7811 Malloc_locked => OPENSSL_malloc_locked
7812 Realloc => OPENSSL_realloc
7813 Free => OPENSSL_free
7816 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7817 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7820 *) CygWin32 support.
7821 [John Jarvie <jjarvie@newsguy.com>]
7823 *) The type-safe stack code has been rejigged. It is now only compiled
7824 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7825 by default all type-specific stack functions are "#define"d back to
7826 standard stack functions. This results in more streamlined output
7827 but retains the type-safety checking possibilities of the original
7831 *) The STACK code has been cleaned up, and certain type declarations
7832 that didn't make a lot of sense have been brought in line. This has
7833 also involved a cleanup of sorts in safestack.h to more correctly
7834 map type-safe stack functions onto their plain stack counterparts.
7835 This work has also resulted in a variety of "const"ifications of
7836 lots of the code, especially "_cmp" operations which should normally
7837 be prototyped with "const" parameters anyway.
7840 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7841 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7842 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7843 where all of 'md' is used each time the PRNG is used, but 'state'
7844 is used only indexed by a cyclic counter. As entropy may not be
7845 well distributed from the beginning, 'md' is important as a
7846 chaining variable. However, the output function chains only half
7847 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7848 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7849 in all of 'state' being rewritten, with the new values depending
7850 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7853 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7854 the handshake is continued after ssl_verify_cert_chain();
7855 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7856 can lead to 'unexplainable' connection aborts later.
7857 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7859 *) Major EVP API cipher revision.
7860 Add hooks for extra EVP features. This allows various cipher
7861 parameters to be set in the EVP interface. Support added for variable
7862 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7863 setting of RC2 and RC5 parameters.
7865 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7868 Remove lots of duplicated code from the EVP library. For example *every*
7869 cipher init() function handles the 'iv' in the same way according to the
7870 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7871 for CFB and OFB modes they zero ctx->num.
7873 New functionality allows removal of S/MIME code RC2 hack.
7875 Most of the routines have the same form and so can be declared in terms
7878 By shifting this to the top level EVP_CipherInit() it can be removed from
7879 all individual ciphers. If the cipher wants to handle IVs or keys
7880 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7883 Change lots of functions like EVP_EncryptUpdate() to now return a
7884 value: although software versions of the algorithms cannot fail
7885 any installed hardware versions can.
7888 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7889 this option is set, tolerate broken clients that send the negotiated
7890 protocol version number instead of the requested protocol version
7894 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7895 i.e. non-zero for export ciphersuites, zero otherwise.
7896 Previous versions had this flag inverted, inconsistent with
7897 rsa_tmp_cb (..._TMP_RSA_CB).
7898 [Bodo Moeller; problem reported by Amit Chopra]
7900 *) Add missing DSA library text string. Work around for some IIS
7901 key files with invalid SEQUENCE encoding.
7904 *) Add a document (doc/standards.txt) that list all kinds of standards
7905 and so on that are implemented in OpenSSL.
7908 *) Enhance c_rehash script. Old version would mishandle certificates
7909 with the same subject name hash and wouldn't handle CRLs at all.
7910 Added -fingerprint option to crl utility, to support new c_rehash
7914 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7917 *) Fix for SSL server purpose checking. Server checking was
7918 rejecting certificates which had extended key usage present
7919 but no ssl client purpose.
7920 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7922 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7923 is a little unclear about how a blank password is handled.
7924 Since the password in encoded as a BMPString with terminating
7925 double NULL a zero length password would end up as just the
7926 double NULL. However no password at all is different and is
7927 handled differently in the PKCS#12 key generation code. NS
7928 treats a blank password as zero length. MSIE treats it as no
7929 password on export: but it will try both on import. We now do
7930 the same: PKCS12_parse() tries zero length and no password if
7931 the password is set to "" or NULL (NULL is now a valid password:
7932 it wasn't before) as does the pkcs12 application.
7935 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7936 perror when PEM_read_bio_X509_REQ fails, the error message must
7937 be obtained from the error queue.
7940 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7941 it in ERR_remove_state if appropriate, and change ERR_get_state
7942 accordingly to avoid race conditions (this is necessary because
7943 thread_hash is no longer constant once set).
7946 *) Bugfix for linux-elf makefile.one.
7949 *) RSA_get_default_method() will now cause a default
7950 RSA_METHOD to be chosen if one doesn't exist already.
7951 Previously this was only set during a call to RSA_new()
7952 or RSA_new_method(NULL) meaning it was possible for
7953 RSA_get_default_method() to return NULL.
7956 *) Added native name translation to the existing DSO code
7957 that will convert (if the flag to do so is set) filenames
7958 that are sufficiently small and have no path information
7959 into a canonical native form. Eg. "blah" converted to
7960 "libblah.so" or "blah.dll" etc.
7963 *) New function ERR_error_string_n(e, buf, len) which is like
7964 ERR_error_string(e, buf), but writes at most 'len' bytes
7965 including the 0 terminator. For ERR_error_string_n, 'buf'
7967 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7969 *) CONF library reworked to become more general. A new CONF
7970 configuration file reader "class" is implemented as well as a
7971 new functions (NCONF_*, for "New CONF") to handle it. The now
7972 old CONF_* functions are still there, but are reimplemented to
7973 work in terms of the new functions. Also, a set of functions
7974 to handle the internal storage of the configuration data is
7975 provided to make it easier to write new configuration file
7976 reader "classes" (I can definitely see something reading a
7977 configuration file in XML format, for example), called _CONF_*,
7978 or "the configuration storage API"...
7980 The new configuration file reading functions are:
7982 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7983 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7985 NCONF_default, NCONF_WIN32
7987 NCONF_dump_fp, NCONF_dump_bio
7989 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7990 NCONF_new creates a new CONF object. This works in the same way
7991 as other interfaces in OpenSSL, like the BIO interface.
7992 NCONF_dump_* dump the internal storage of the configuration file,
7993 which is useful for debugging. All other functions take the same
7994 arguments as the old CONF_* functions wth the exception of the
7995 first that must be a `CONF *' instead of a `LHASH *'.
7997 To make it easer to use the new classes with the old CONF_* functions,
7998 the function CONF_set_default_method is provided.
8001 *) Add '-tls1' option to 'openssl ciphers', which was already
8002 mentioned in the documentation but had not been implemented.
8003 (This option is not yet really useful because even the additional
8004 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8007 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8008 OpenSSL-based applications) load shared libraries and bind to
8009 them in a portable way.
8010 [Geoff Thorpe, with contributions from Richard Levitte]
8012 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8014 *) Make sure _lrotl and _lrotr are only used with MSVC.
8016 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8017 (the default implementation of RAND_status).
8019 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8020 to '-clrext' (= clear extensions), as intended and documented.
8021 [Bodo Moeller; inconsistency pointed out by Michael Attili
8024 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8025 was larger than the MD block size.
8026 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8028 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8029 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8030 using the passed key: if the passed key was a private key the result
8031 of X509_print(), for example, would be to print out all the private key
8035 *) des_quad_cksum() byte order bug fix.
8036 [Ulf Möller, using the problem description in krb4-0.9.7, where
8037 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8039 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8041 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8043 *) For easily testing in shell scripts whether some command
8044 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8045 returns with exit code 0 iff no command of the given name is available.
8046 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8047 the output goes to stdout and nothing is printed to stderr.
8048 Additional arguments are always ignored.
8050 Since for each cipher there is a command of the same name,
8051 the 'no-cipher' compilation switches can be tested this way.
8053 ('openssl no-XXX' is not able to detect pseudo-commands such
8054 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8057 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8060 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8061 is set; it will be thrown away anyway because each handshake creates
8063 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8064 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8065 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
8066 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8069 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8070 'Q' and 'R' lose their special meanings (quit/renegotiate).
8071 This is part of what -quiet does; unlike -quiet, -ign_eof
8072 does not suppress any output.
8075 *) Add compatibility options to the purpose and trust code. The
8076 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8077 accepts a certificate or CA, this was the previous behaviour,
8078 with all the associated security issues.
8080 X509_TRUST_COMPAT is the old trust behaviour: only and
8081 automatically trust self signed roots in certificate store. A
8082 new trust setting X509_TRUST_DEFAULT is used to specify that
8083 a purpose has no associated trust setting and it should instead
8084 use the value in the default purpose.
8087 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8088 and fix a memory leak.
8091 *) In util/mkerr.pl (which implements 'make errors'), preserve
8092 reason strings from the previous version of the .c file, as
8093 the default to have only downcase letters (and digits) in
8094 automatically generated reasons codes is not always appropriate.
8097 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8098 using strerror. Previously, ERR_reason_error_string() returned
8099 library names as reason strings for SYSerr; but SYSerr is a special
8100 case where small numbers are errno values, not library numbers.
8103 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8104 converts DSA parameters into DH parameters. (When creating parameters,
8105 DSA_generate_parameters is used.)
8108 *) Include 'length' (recommended exponent length) in C code generated
8109 by 'openssl dhparam -C'.
8112 *) The second argument to set_label in perlasm was already being used
8113 so couldn't be used as a "file scope" flag. Moved to third argument
8117 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8118 instead of RAND_bytes for encryption IVs and salts.
8121 *) Include RAND_status() into RAND_METHOD instead of implementing
8122 it only for md_rand.c Otherwise replacing the PRNG by calling
8123 RAND_set_rand_method would be impossible.
8126 *) Don't let DSA_generate_key() enter an infinite loop if the random
8127 number generation fails.
8130 *) New 'rand' application for creating pseudo-random output.
8133 *) Added configuration support for Linux/IA64
8134 [Rolf Haberrecker <rolf@suse.de>]
8136 *) Assembler module support for Mingw32.
8139 *) Shared library support for HPUX (in shlib/).
8140 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8142 *) Shared library support for Solaris gcc.
8143 [Lutz Behnke <behnke@trustcenter.de>]
8145 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8147 *) PKCS7_encrypt() was adding text MIME headers twice because they
8148 were added manually and by SMIME_crlf_copy().
8151 *) In bntest.c don't call BN_rand with zero bits argument.
8152 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8154 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8155 case was implemented. This caused BN_div_recp() to fail occasionally.
8158 *) Add an optional second argument to the set_label() in the perl
8159 assembly language builder. If this argument exists and is set
8160 to 1 it signals that the assembler should use a symbol whose
8161 scope is the entire file, not just the current function. This
8162 is needed with MASM which uses the format label:: for this scope.
8163 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8165 *) Change the ASN1 types so they are typedefs by default. Before
8166 almost all types were #define'd to ASN1_STRING which was causing
8167 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8171 *) Change names of new functions to the new get1/get0 naming
8172 convention: After 'get1', the caller owns a reference count
8173 and has to call ..._free; 'get0' returns a pointer to some
8174 data structure without incrementing reference counters.
8175 (Some of the existing 'get' functions increment a reference
8176 counter, some don't.)
8177 Similarly, 'set1' and 'add1' functions increase reference
8178 counters or duplicate objects.
8181 *) Allow for the possibility of temp RSA key generation failure:
8182 the code used to assume it always worked and crashed on failure.
8185 *) Fix potential buffer overrun problem in BIO_printf().
8186 [Ulf Möller, using public domain code by Patrick Powell; problem
8187 pointed out by David Sacerdote <das33@cornell.edu>]
8189 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8190 RAND_egd() and RAND_status(). In the command line application,
8191 the EGD socket can be specified like a seed file using RANDFILE
8195 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8196 Some CAs (e.g. Verisign) distribute certificates in this form.
8199 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8200 list to exclude them. This means that no special compilation option
8201 is needed to use anonymous DH: it just needs to be included in the
8205 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8206 EVP_MD_type. The old functionality is available in a new macro called
8207 EVP_MD_md(). Change code that uses it and update docs.
8210 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8211 where the 'void *' argument is replaced by a function pointer argument.
8212 Previously 'void *' was abused to point to functions, which works on
8213 many platforms, but is not correct. As these functions are usually
8214 called by macros defined in OpenSSL header files, most source code
8215 should work without changes.
8218 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8219 sections with information on -D... compiler switches used for
8220 compiling the library so that applications can see them. To enable
8221 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8222 must be defined. E.g.,
8223 #define OPENSSL_ALGORITHM_DEFINES
8224 #include <openssl/opensslconf.h>
8225 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8226 [Richard Levitte, Ulf and Bodo Möller]
8228 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8232 *) Change the 'other' type in certificate aux info to a STACK_OF
8233 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8234 the required ASN1 format: arbitrary types determined by an OID.
8237 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8238 argument to 'req'. This is not because the function is newer or
8239 better than others it just uses the work 'NEW' in the certificate
8240 request header lines. Some software needs this.
8243 *) Reorganise password command line arguments: now passwords can be
8244 obtained from various sources. Delete the PEM_cb function and make
8245 it the default behaviour: i.e. if the callback is NULL and the
8246 usrdata argument is not NULL interpret it as a null terminated pass
8247 phrase. If usrdata and the callback are NULL then the pass phrase
8248 is prompted for as usual.
8251 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8252 the support is automatically enabled. The resulting binaries will
8253 autodetect the card and use it if present.
8254 [Ben Laurie and Compaq Inc.]
8256 *) Work around for Netscape hang bug. This sends certificate request
8257 and server done in one record. Since this is perfectly legal in the
8258 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8259 the bugs/SSLv3 entry for more info.
8262 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8265 *) Add -rand argument to smime and pkcs12 applications and read/write
8269 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8272 *) Add command line password options to the remaining applications.
8275 *) Bug fix for BN_div_recp() for numerators with an even number of
8279 *) More tests in bntest.c, and changed test_bn output.
8282 *) ./config recognizes MacOS X now.
8285 *) Bug fix for BN_div() when the first words of num and divsor are
8286 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8289 *) Add support for various broken PKCS#8 formats, and command line
8290 options to produce them.
8293 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8294 get temporary BIGNUMs from a BN_CTX.
8297 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8301 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8302 include a #define from the old name to the new. The original intent
8303 was that statically linked binaries could for example just call
8304 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8305 link with digests. This never worked becayse SSLeay_add_all_digests()
8306 and SSLeay_add_all_ciphers() were in the same source file so calling
8307 one would link with the other. They are now in separate source files.
8310 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8313 *) Use a less unusual form of the Miller-Rabin primality test (it used
8314 a binary algorithm for exponentiation integrated into the Miller-Rabin
8315 loop, our standard modexp algorithms are faster).
8318 *) Support for the EBCDIC character set completed.
8319 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8321 *) Source code cleanups: use const where appropriate, eliminate casts,
8322 use void * instead of char * in lhash.
8325 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8326 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8327 this the server could overwrite ephemeral keys that the client
8331 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8332 using 50 iterations of the Rabin-Miller test.
8334 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8335 iterations of the Rabin-Miller test as required by the appendix
8336 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8337 As BN_is_prime_fasttest includes trial division, DSA parameter
8338 generation becomes much faster.
8340 This implies a change for the callback functions in DSA_is_prime
8341 and DSA_generate_parameters: The callback function is called once
8342 for each positive witness in the Rabin-Miller test, not just
8343 occasionally in the inner loop; and the parameters to the
8344 callback function now provide an iteration count for the outer
8345 loop rather than for the current invocation of the inner loop.
8346 DSA_generate_parameters additionally can call the callback
8347 function with an 'iteration count' of -1, meaning that a
8348 candidate has passed the trial division test (when q is generated
8349 from an application-provided seed, trial division is skipped).
8352 *) New function BN_is_prime_fasttest that optionally does trial
8353 division before starting the Rabin-Miller test and has
8354 an additional BN_CTX * argument (whereas BN_is_prime always
8355 has to allocate at least one BN_CTX).
8356 'callback(1, -1, cb_arg)' is called when a number has passed the
8357 trial division stage.
8360 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8364 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8367 *) New function BN_pseudo_rand().
8370 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8371 bignum version of BN_from_montgomery() with the working code from
8372 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8376 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8377 made it impossible to use the same SSL_SESSION data structure in
8378 SSL2 clients in multiple threads.
8381 *) The return value of RAND_load_file() no longer counts bytes obtained
8382 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8383 to seed the PRNG (previously an explicit byte count was required).
8384 [Ulf Möller, Bodo Möller]
8386 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8387 used (char *) instead of (void *) and had casts all over the place.
8390 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8393 *) Retain source code compatibility for BN_prime_checks macro:
8394 BN_is_prime(..., BN_prime_checks, ...) now uses
8395 BN_prime_checks_for_size to determine the appropriate number of
8396 Rabin-Miller iterations.
8399 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8400 DH_CHECK_P_NOT_SAFE_PRIME.
8401 (Check if this is true? OpenPGP calls them "strong".)
8404 *) Merge the functionality of "dh" and "gendh" programs into a new program
8405 "dhparam". The old programs are retained for now but will handle DH keys
8406 (instead of parameters) in future.
8409 *) Make the ciphers, s_server and s_client programs check the return values
8410 when a new cipher list is set.
8413 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8414 ciphers. Before when the 56bit ciphers were enabled the sorting was
8417 The syntax for the cipher sorting has been extended to support sorting by
8418 cipher-strength (using the strength_bits hard coded in the tables).
8419 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8421 Fix a bug in the cipher-command parser: when supplying a cipher command
8422 string with an "undefined" symbol (neither command nor alphanumeric
8423 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8424 an error is flagged.
8426 Due to the strength-sorting extension, the code of the
8427 ssl_create_cipher_list() function was completely rearranged. I hope that
8428 the readability was also increased :-)
8429 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8431 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8432 for the first serial number and places 2 in the serial number file. This
8433 avoids problems when the root CA is created with serial number zero and
8434 the first user certificate has the same issuer name and serial number
8438 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8439 the new code. Add documentation for this stuff.
8442 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8443 X509_*() to X509at_*() on the grounds that they don't handle X509
8444 structures and behave in an analagous way to the X509v3 functions:
8445 they shouldn't be called directly but wrapper functions should be used
8448 So we also now have some wrapper functions that call the X509at functions
8449 when passed certificate requests. (TO DO: similar things can be done with
8450 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8451 things. Some of these need some d2i or i2d and print functionality
8452 because they handle more complex structures.)
8455 *) Add missing #ifndefs that caused missing symbols when building libssl
8456 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8457 NO_RSA in ssl/s2*.c.
8458 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8460 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8461 has a return value which indicates the quality of the random data
8462 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8463 error queue. New function RAND_pseudo_bytes() generates output that is
8464 guaranteed to be unique but not unpredictable. RAND_add is like
8465 RAND_seed, but takes an extra argument for an entropy estimate
8466 (RAND_seed always assumes full entropy).
8469 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8470 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8471 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8472 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8473 false-positive rate of at most 2^-80 for random input.
8476 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8479 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8480 in the 0.9.5 release), this returns the chain
8481 from an X509_CTX structure with a dup of the stack and all
8482 the X509 reference counts upped: so the stack will exist
8483 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8486 Also make SSL_SESSION_print() print out the verify return
8490 *) Add manpage for the pkcs12 command. Also change the default
8491 behaviour so MAC iteration counts are used unless the new
8492 -nomaciter option is used. This improves file security and
8493 only older versions of MSIE (4.0 for example) need it.
8496 *) Honor the no-xxx Configure options when creating .DEF files.
8499 *) Add PKCS#10 attributes to field table: challengePassword,
8500 unstructuredName and unstructuredAddress. These are taken from
8501 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8502 international characters are used.
8504 More changes to X509_ATTRIBUTE code: allow the setting of types
8505 based on strings. Remove the 'loc' parameter when adding
8506 attributes because these will be a SET OF encoding which is sorted
8510 *) Initial changes to the 'req' utility to allow request generation
8511 automation. This will allow an application to just generate a template
8512 file containing all the field values and have req construct the
8515 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8516 used all over the place including certificate requests and PKCS#7
8517 structures. They are currently handled manually where necessary with
8518 some primitive wrappers for PKCS#7. The new functions behave in a
8519 manner analogous to the X509 extension functions: they allow
8520 attributes to be looked up by NID and added.
8522 Later something similar to the X509V3 code would be desirable to
8523 automatically handle the encoding, decoding and printing of the
8524 more complex types. The string types like challengePassword can
8525 be handled by the string table functions.
8527 Also modified the multi byte string table handling. Now there is
8528 a 'global mask' which masks out certain types. The table itself
8529 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8530 is useful when for example there is only one permissible type
8531 (as in countryName) and using the mask might result in no valid
8535 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8536 SSL_get_peer_finished to allow applications to obtain the latest
8537 Finished messages sent to the peer or expected from the peer,
8538 respectively. (SSL_get_peer_finished is usually the Finished message
8539 actually received from the peer, otherwise the protocol will be aborted.)
8541 As the Finished message are message digests of the complete handshake
8542 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8543 be used for external authentication procedures when the authentication
8544 provided by SSL/TLS is not desired or is not enough.
8547 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8548 the host supports BWX extension and if Compaq C is present on the
8549 $PATH. Just exploiting of the BWX extension results in 20-30%
8550 performance kick for some algorithms, e.g. DES and RC4 to mention
8551 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8555 *) Add support for MS "fast SGC". This is arguably a violation of the
8556 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8557 weak crypto and after checking the certificate is SGC a second one
8558 with strong crypto. MS SGC stops the first handshake after receiving
8559 the server certificate message and sends a second client hello. Since
8560 a server will typically do all the time consuming operations before
8561 expecting any further messages from the client (server key exchange
8562 is the most expensive) there is little difference between the two.
8564 To get OpenSSL to support MS SGC we have to permit a second client
8565 hello message after we have sent server done. In addition we have to
8566 reset the MAC if we do get this second client hello.
8569 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8570 if a DER encoded private key is RSA or DSA traditional format. Changed
8571 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8572 format DER encoded private key. Newer code should use PKCS#8 format which
8573 has the key type encoded in the ASN1 structure. Added DER private key
8574 support to pkcs8 application.
8577 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8578 ciphersuites has been selected (as required by the SSL 3/TLS 1
8579 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8580 is set, we interpret this as a request to violate the specification
8581 (the worst that can happen is a handshake failure, and 'correct'
8582 behaviour would result in a handshake failure anyway).
8585 *) In SSL_CTX_add_session, take into account that there might be multiple
8586 SSL_SESSION structures with the same session ID (e.g. when two threads
8587 concurrently obtain them from an external cache).
8588 The internal cache can handle only one SSL_SESSION with a given ID,
8589 so if there's a conflict, we now throw out the old one to achieve
8593 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8594 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8595 some routines that use cipher OIDs: some ciphers do not have OIDs
8596 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8600 *) Simplify the trust setting structure and code. Now we just have
8601 two sequences of OIDs for trusted and rejected settings. These will
8602 typically have values the same as the extended key usage extension
8603 and any application specific purposes.
8605 The trust checking code now has a default behaviour: it will just
8606 check for an object with the same NID as the passed id. Functions can
8607 be provided to override either the default behaviour or the behaviour
8608 for a given id. SSL client, server and email already have functions
8609 in place for compatibility: they check the NID and also return "trusted"
8610 if the certificate is self signed.
8613 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8614 traditional format into an EVP_PKEY structure.
8617 *) Add a password callback function PEM_cb() which either prompts for
8618 a password if usr_data is NULL or otherwise assumes it is a null
8619 terminated password. Allow passwords to be passed on command line
8620 environment or config files in a few more utilities.
8623 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8624 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8625 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8626 Update documentation.
8629 *) Support for ASN1 "NULL" type. This could be handled before by using
8630 ASN1_TYPE but there wasn't any function that would try to read a NULL
8631 and produce an error if it couldn't. For compatibility we also have
8632 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8633 don't allocate anything because they don't need to.
8636 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8638 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8640 *) Rebuild of the memory allocation routines used by OpenSSL code and
8641 possibly others as well. The purpose is to make an interface that
8642 provide hooks so anyone can build a separate set of allocation and
8643 deallocation routines to be used by OpenSSL, for example memory
8644 pool implementations, or something else, which was previously hard
8645 since Malloc(), Realloc() and Free() were defined as macros having
8646 the values malloc, realloc and free, respectively (except for Win32
8647 compilations). The same is provided for memory debugging code.
8648 OpenSSL already comes with functionality to find memory leaks, but
8649 this gives people a chance to debug other memory problems.
8651 With these changes, a new set of functions and macros have appeared:
8653 CRYPTO_set_mem_debug_functions() [F]
8654 CRYPTO_get_mem_debug_functions() [F]
8655 CRYPTO_dbg_set_options() [F]
8656 CRYPTO_dbg_get_options() [F]
8657 CRYPTO_malloc_debug_init() [M]
8659 The memory debug functions are NULL by default, unless the library
8660 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8661 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8662 gives the standard debugging functions that come with OpenSSL) or
8663 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8664 provided by the library user) must be used. When the standard
8665 debugging functions are used, CRYPTO_dbg_set_options can be used to
8666 request additional information:
8667 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8668 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8670 Also, things like CRYPTO_set_mem_functions will always give the
8671 expected result (the new set of functions is used for allocation
8672 and deallocation) at all times, regardless of platform and compiler
8675 To finish it up, some functions that were never use in any other
8676 way than through macros have a new API and new semantic:
8679 CRYPTO_dbg_realloc()
8682 All macros of value have retained their old syntax.
8683 [Richard Levitte and Bodo Moeller]
8685 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8686 ordering of SMIMECapabilities wasn't in "strength order" and there
8687 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8691 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8692 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8693 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8695 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8696 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8697 functionality to handle multipart/signed properly) and a utility
8698 called 'smime' to call all this stuff. This is based on code I
8699 originally wrote for Celo who have kindly allowed it to be
8700 included in OpenSSL.
8703 *) Add variants des_set_key_checked and des_set_key_unchecked of
8704 des_set_key (aka des_key_sched). Global variable des_check_key
8705 decides which of these is called by des_set_key; this way
8706 des_check_key behaves as it always did, but applications and
8707 the library itself, which was buggy for des_check_key == 1,
8708 have a cleaner way to pick the version they need.
8711 *) New function PKCS12_newpass() which changes the password of a
8715 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8716 dynamic mix. In both cases the ids can be used as an index into the
8717 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8718 functions so they accept a list of the field values and the
8719 application doesn't need to directly manipulate the X509_TRUST
8723 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8727 *) Modify the way the V3 extension code looks up extensions. This now
8728 works in a similar way to the object code: we have some "standard"
8729 extensions in a static table which is searched with OBJ_bsearch()
8730 and the application can add dynamic ones if needed. The file
8731 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8732 updated whenever a new extension is added to the core code and kept
8733 in ext_nid order. There is a simple program 'tabtest.c' which checks
8734 this. New extensions are not added too often so this file can readily
8735 be maintained manually.
8737 There are two big advantages in doing things this way. The extensions
8738 can be looked up immediately and no longer need to be "added" using
8739 X509V3_add_standard_extensions(): this function now does nothing.
8740 [Side note: I get *lots* of email saying the extension code doesn't
8741 work because people forget to call this function]
8742 Also no dynamic allocation is done unless new extensions are added:
8743 so if we don't add custom extensions there is no need to call
8744 X509V3_EXT_cleanup().
8747 *) Modify enc utility's salting as follows: make salting the default. Add a
8748 magic header, so unsalted files fail gracefully instead of just decrypting
8749 to garbage. This is because not salting is a big security hole, so people
8750 should be discouraged from doing it.
8753 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8754 digest to be passed on the command line but it only used this
8755 parameter when signing a certificate. Modified so all relevant
8756 operations are affected by the digest parameter including the
8757 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8758 DSA key was used because it didn't fix the digest.
8761 *) Initial certificate chain verify code. Currently tests the untrusted
8762 certificates for consistency with the verify purpose (which is set
8763 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8765 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8766 this is because it will reject chains with invalid extensions whereas
8767 every previous version of OpenSSL and SSLeay made no checks at all.
8769 Trust code: checks the root CA for the relevant trust settings. Trust
8770 settings have an initial value consistent with the verify purpose: e.g.
8771 if the verify purpose is for SSL client use it expects the CA to be
8772 trusted for SSL client use. However the default value can be changed to
8773 permit custom trust settings: one example of this would be to only trust
8774 certificates from a specific "secure" set of CAs.
8776 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8777 which should be used for version portability: especially since the
8778 verify structure is likely to change more often now.
8780 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8781 to set them. If not set then assume SSL clients will verify SSL servers
8784 Two new options to the verify program: -untrusted allows a set of
8785 untrusted certificates to be passed in and -purpose which sets the
8786 intended purpose of the certificate. If a purpose is set then the
8787 new chain verify code is used to check extension consistency.
8790 *) Support for the authority information access extension.
8793 *) Modify RSA and DSA PEM read routines to transparently handle
8794 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8795 public keys in a format compatible with certificate
8796 SubjectPublicKeyInfo structures. Unfortunately there were already
8797 functions called *_PublicKey_* which used various odd formats so
8798 these are retained for compatibility: however the DSA variants were
8799 never in a public release so they have been deleted. Changed dsa/rsa
8800 utilities to handle the new format: note no releases ever handled public
8801 keys so we should be OK.
8803 The primary motivation for this change is to avoid the same fiasco
8804 that dogs private keys: there are several incompatible private key
8805 formats some of which are standard and some OpenSSL specific and
8806 require various evil hacks to allow partial transparent handling and
8807 even then it doesn't work with DER formats. Given the option anything
8808 other than PKCS#8 should be dumped: but the other formats have to
8809 stay in the name of compatibility.
8811 With public keys and the benefit of hindsight one standard format
8812 is used which works with EVP_PKEY, RSA or DSA structures: though
8813 it clearly returns an error if you try to read the wrong kind of key.
8815 Added a -pubkey option to the 'x509' utility to output the public key.
8816 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8817 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8818 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8819 that do the same as the EVP_PKEY_assign_*() except they up the
8820 reference count of the added key (they don't "swallow" the
8824 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8825 CRLs would fail if the file contained no certificates or no CRLs:
8826 added a new function to read in both types and return the number
8827 read: this means that if none are read it will be an error. The
8828 DER versions of the certificate and CRL reader would always fail
8829 because it isn't possible to mix certificates and CRLs in DER format
8830 without choking one or the other routine. Changed this to just read
8831 a certificate: this is the best we can do. Also modified the code
8832 in apps/verify.c to take notice of return codes: it was previously
8833 attempting to read in certificates from NULL pointers and ignoring
8834 any errors: this is one reason why the cert and CRL reader seemed
8835 to work. It doesn't check return codes from the default certificate
8836 routines: these may well fail if the certificates aren't installed.
8839 *) Code to support otherName option in GeneralName.
8842 *) First update to verify code. Change the verify utility
8843 so it warns if it is passed a self signed certificate:
8844 for consistency with the normal behaviour. X509_verify
8845 has been modified to it will now verify a self signed
8846 certificate if *exactly* the same certificate appears
8847 in the store: it was previously impossible to trust a
8848 single self signed certificate. This means that:
8849 openssl verify ss.pem
8850 now gives a warning about a self signed certificate but
8851 openssl verify -CAfile ss.pem ss.pem
8855 *) For servers, store verify_result in SSL_SESSION data structure
8856 (and add it to external session representation).
8857 This is needed when client certificate verifications fails,
8858 but an application-provided verification callback (set by
8859 SSL_CTX_set_cert_verify_callback) allows accepting the session
8860 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8861 but returns 1): When the session is reused, we have to set
8862 ssl->verify_result to the appropriate error code to avoid
8864 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8866 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8867 case in PKCS7_dataInit() where the signed PKCS7 structure
8868 didn't contain any existing data because it was being created.
8869 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8871 *) Add a salt to the key derivation routines in enc.c. This
8872 forms the first 8 bytes of the encrypted file. Also add a
8873 -S option to allow a salt to be input on the command line.
8876 *) New function X509_cmp(). Oddly enough there wasn't a function
8877 to compare two certificates. We do this by working out the SHA1
8878 hash and comparing that. X509_cmp() will be needed by the trust
8882 *) SSL_get1_session() is like SSL_get_session(), but increments
8883 the reference count in the SSL_SESSION returned.
8884 [Geoff Thorpe <geoff@eu.c2.net>]
8886 *) Fix for 'req': it was adding a null to request attributes.
8887 Also change the X509_LOOKUP and X509_INFO code to handle
8888 certificate auxiliary information.
8891 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8895 *) Add the possibility to add extra information to the memory leak
8896 detecting output, to form tracebacks, showing from where each
8897 allocation was originated: CRYPTO_push_info("constant string") adds
8898 the string plus current file name and line number to a per-thread
8899 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8900 is like calling CYRPTO_pop_info() until the stack is empty.
8901 Also updated memory leak detection code to be multi-thread-safe.
8904 *) Add options -text and -noout to pkcs7 utility and delete the
8905 encryption options which never did anything. Update docs.
8908 *) Add options to some of the utilities to allow the pass phrase
8909 to be included on either the command line (not recommended on
8910 OSes like Unix) or read from the environment. Update the
8911 manpages and fix a few bugs.
8914 *) Add a few manpages for some of the openssl commands.
8917 *) Fix the -revoke option in ca. It was freeing up memory twice,
8918 leaking and not finding already revoked certificates.
8921 *) Extensive changes to support certificate auxiliary information.
8922 This involves the use of X509_CERT_AUX structure and X509_AUX
8923 functions. An X509_AUX function such as PEM_read_X509_AUX()
8924 can still read in a certificate file in the usual way but it
8925 will also read in any additional "auxiliary information". By
8926 doing things this way a fair degree of compatibility can be
8927 retained: existing certificates can have this information added
8928 using the new 'x509' options.
8930 Current auxiliary information includes an "alias" and some trust
8931 settings. The trust settings will ultimately be used in enhanced
8932 certificate chain verification routines: currently a certificate
8933 can only be trusted if it is self signed and then it is trusted
8937 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8938 The problem was that one of the replacement routines had not been working
8939 since SSLeay releases. For now the offending routine has been replaced
8940 with non-optimised assembler. Even so, this now gives around 95%
8941 performance improvement for 1024 bit RSA signs.
8944 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8945 handling. Most clients have the effective key size in bits equal to
8946 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8947 A few however don't do this and instead use the size of the decrypted key
8948 to determine the RC2 key length and the AlgorithmIdentifier to determine
8949 the effective key length. In this case the effective key length can still
8950 be 40 bits but the key length can be 168 bits for example. This is fixed
8951 by manually forcing an RC2 key into the EVP_PKEY structure because the
8952 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8953 the key length and effective key length are equal.
8956 *) Add a bunch of functions that should simplify the creation of
8957 X509_NAME structures. Now you should be able to do:
8958 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8959 and have it automatically work out the correct field type and fill in
8960 the structures. The more adventurous can try:
8961 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8962 and it will (hopefully) work out the correct multibyte encoding.
8965 *) Change the 'req' utility to use the new field handling and multibyte
8966 copy routines. Before the DN field creation was handled in an ad hoc
8967 way in req, ca, and x509 which was rather broken and didn't support
8968 BMPStrings or UTF8Strings. Since some software doesn't implement
8969 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8970 using the dirstring_type option. See the new comment in the default
8971 openssl.cnf for more info.
8974 *) Make crypto/rand/md_rand.c more robust:
8975 - Assure unique random numbers after fork().
8976 - Make sure that concurrent threads access the global counter and
8977 md serializably so that we never lose entropy in them
8978 or use exactly the same state in multiple threads.
8979 Access to the large state is not always serializable because
8980 the additional locking could be a performance killer, and
8981 md should be large enough anyway.
8984 *) New file apps/app_rand.c with commonly needed functionality
8985 for handling the random seed file.
8987 Use the random seed file in some applications that previously did not:
8989 dsaparam -genkey (which also ignored its '-rand' option),
8992 x509 (when signing).
8993 Except on systems with /dev/urandom, it is crucial to have a random
8994 seed file at least for key creation, DSA signing, and for DH exchanges;
8995 for RSA signatures we could do without one.
8997 gendh and gendsa (unlike genrsa) used to read only the first byte
8998 of each file listed in the '-rand' option. The function as previously
8999 found in genrsa is now in app_rand.c and is used by all programs
9000 that support '-rand'.
9003 *) In RAND_write_file, use mode 0600 for creating files;
9004 don't just chmod when it may be too late.
9007 *) Report an error from X509_STORE_load_locations
9008 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9011 *) New function ASN1_mbstring_copy() this copies a string in either
9012 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9013 into an ASN1_STRING type. A mask of permissible types is passed
9014 and it chooses the "minimal" type to use or an error if not type
9018 *) Add function equivalents to the various macros in asn1.h. The old
9019 macros are retained with an M_ prefix. Code inside the library can
9020 use the M_ macros. External code (including the openssl utility)
9021 should *NOT* in order to be "shared library friendly".
9024 *) Add various functions that can check a certificate's extensions
9025 to see if it usable for various purposes such as SSL client,
9026 server or S/MIME and CAs of these types. This is currently
9027 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9028 verification. Also added a -purpose flag to x509 utility to
9029 print out all the purposes.
9032 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9036 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9037 for, obtain and decode and extension and obtain its critical flag.
9038 This allows all the necessary extension code to be handled in a
9039 single function call.
9042 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9043 platforms. See crypto/rc4/rc4_enc.c for further details.
9046 *) New -noout option to asn1parse. This causes no output to be produced
9047 its main use is when combined with -strparse and -out to extract data
9048 from a file (which may not be in ASN.1 format).
9051 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9052 when producing the local key id.
9053 [Richard Levitte <levitte@stacken.kth.se>]
9055 *) New option -dhparam in s_server. This allows a DH parameter file to be
9056 stated explicitly. If it is not stated then it tries the first server
9057 certificate file. The previous behaviour hard coded the filename
9061 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9062 a public key to be input or output. For example:
9063 openssl rsa -in key.pem -pubout -out pubkey.pem
9064 Also added necessary DSA public key functions to handle this.
9067 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9068 in the message. This was handled by allowing
9069 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9070 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9072 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9073 to the end of the strings whereas this didn't. This would cause problems
9074 if strings read with d2i_ASN1_bytes() were later modified.
9075 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9077 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9078 data and it contains EOF it will end up returning an error. This is
9079 caused by input 46 bytes long. The cause is due to the way base64
9080 BIOs find the start of base64 encoded data. They do this by trying a
9081 trial decode on each line until they find one that works. When they
9082 do a flag is set and it starts again knowing it can pass all the
9083 data directly through the decoder. Unfortunately it doesn't reset
9084 the context it uses. This means that if EOF is reached an attempt
9085 is made to pass two EOFs through the context and this causes the
9086 resulting error. This can also cause other problems as well. As is
9087 usual with these problems it takes *ages* to find and the fix is
9088 trivial: move one line.
9089 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9091 *) Ugly workaround to get s_client and s_server working under Windows. The
9092 old code wouldn't work because it needed to select() on sockets and the
9093 tty (for keypresses and to see if data could be written). Win32 only
9094 supports select() on sockets so we select() with a 1s timeout on the
9095 sockets and then see if any characters are waiting to be read, if none
9096 are present then we retry, we also assume we can always write data to
9097 the tty. This isn't nice because the code then blocks until we've
9098 received a complete line of data and it is effectively polling the
9099 keyboard at 1s intervals: however it's quite a bit better than not
9100 working at all :-) A dedicated Windows application might handle this
9101 with an event loop for example.
9104 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9105 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9106 will be called when RSA_sign() and RSA_verify() are used. This is useful
9107 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9108 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9109 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9110 This necessitated the support of an extra signature type NID_md5_sha1
9111 for SSL signatures and modifications to the SSL library to use it instead
9112 of calling RSA_public_decrypt() and RSA_private_encrypt().
9115 *) Add new -verify -CAfile and -CApath options to the crl program, these
9116 will lookup a CRL issuers certificate and verify the signature in a
9117 similar way to the verify program. Tidy up the crl program so it
9118 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9119 less strict. It will now permit CRL extensions even if it is not
9120 a V2 CRL: this will allow it to tolerate some broken CRLs.
9123 *) Initialize all non-automatic variables each time one of the openssl
9124 sub-programs is started (this is necessary as they may be started
9125 multiple times from the "OpenSSL>" prompt).
9126 [Lennart Bang, Bodo Moeller]
9128 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9129 removing all other RSA functionality (this is what NO_RSA does). This
9130 is so (for example) those in the US can disable those operations covered
9131 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9135 *) Non-copying interface to BIO pairs.
9136 (still largely untested)
9139 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9140 ASCII string. This was handled independently in various places before.
9143 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9144 UTF8 strings a character at a time.
9147 *) Use client_version from client hello to select the protocol
9148 (s23_srvr.c) and for RSA client key exchange verification
9149 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9152 *) Add various utility functions to handle SPKACs, these were previously
9153 handled by poking round in the structure internals. Added new function
9154 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9155 print, verify and generate SPKACs. Based on an original idea from
9156 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9159 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9162 *) Allow the config file extension section to be overwritten on the
9163 command line. Based on an original idea from Massimiliano Pala
9164 <madwolf@comune.modena.it>. The new option is called -extensions
9165 and can be applied to ca, req and x509. Also -reqexts to override
9166 the request extensions in req and -crlexts to override the crl extensions
9170 *) Add new feature to the SPKAC handling in ca. Now you can include
9171 the same field multiple times by preceding it by "XXXX." for example:
9174 this is the same syntax as used in the req config file.
9177 *) Allow certificate extensions to be added to certificate requests. These
9178 are specified in a 'req_extensions' option of the req section of the
9179 config file. They can be printed out with the -text option to req but
9180 are otherwise ignored at present.
9183 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9184 data read consists of only the final block it would not decrypted because
9185 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9186 A misplaced 'break' also meant the decrypted final block might not be
9187 copied until the next read.
9190 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9191 a few extra parameters to the DH structure: these will be useful if
9192 for example we want the value of 'q' or implement X9.42 DH.
9195 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9196 provides hooks that allow the default DSA functions or functions on a
9197 "per key" basis to be replaced. This allows hardware acceleration and
9198 hardware key storage to be handled without major modification to the
9199 library. Also added low level modexp hooks and CRYPTO_EX structure and
9200 associated functions.
9203 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9204 as "read only": it can't be written to and the buffer it points to will
9205 not be freed. Reading from a read only BIO is much more efficient than
9206 a normal memory BIO. This was added because there are several times when
9207 an area of memory needs to be read from a BIO. The previous method was
9208 to create a memory BIO and write the data to it, this results in two
9209 copies of the data and an O(n^2) reading algorithm. There is a new
9210 function BIO_new_mem_buf() which creates a read only memory BIO from
9211 an area of memory. Also modified the PKCS#7 routines to use read only
9215 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9216 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9217 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9218 but a retry condition occured while trying to read the rest.
9221 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9222 NID_pkcs7_encrypted by default: this was wrong since this should almost
9223 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9224 the encrypted data type: this is a more sensible place to put it and it
9225 allows the PKCS#12 code to be tidied up that duplicated this
9229 *) Changed obj_dat.pl script so it takes its input and output files on
9230 the command line. This should avoid shell escape redirection problems
9234 *) Initial support for certificate extension requests, these are included
9235 in things like Xenroll certificate requests. Included functions to allow
9236 extensions to be obtained and added.
9239 *) -crlf option to s_client and s_server for sending newlines as
9240 CRLF (as required by many protocols).
9243 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9245 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9246 [Ralf S. Engelschall]
9248 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9249 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9251 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9255 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9256 DH parameters/keys (q is lost during that conversion, but the resulting
9257 DH parameters contain its length).
9259 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9260 much faster than DH_generate_parameters (which creates parameters
9261 where p = 2*q + 1), and also the smaller q makes DH computations
9262 much more efficient (160-bit exponentiation instead of 1024-bit
9263 exponentiation); so this provides a convenient way to support DHE
9264 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9265 utter importance to use
9266 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9268 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9269 when such DH parameters are used, because otherwise small subgroup
9270 attacks may become possible!
9273 *) Avoid memory leak in i2d_DHparams.
9276 *) Allow the -k option to be used more than once in the enc program:
9277 this allows the same encrypted message to be read by multiple recipients.
9280 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9281 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9282 it will always use the numerical form of the OID, even if it has a short
9286 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9287 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9288 otherwise bn_mod_exp was called. In the case of hardware keys for example
9289 no private key components need be present and it might store extra data
9290 in the RSA structure, which cannot be accessed from bn_mod_exp.
9291 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9292 private key operations.
9295 *) Added support for SPARC Linux.
9298 *) pem_password_cb function type incompatibly changed from
9299 typedef int pem_password_cb(char *buf, int size, int rwflag);
9301 ....(char *buf, int size, int rwflag, void *userdata);
9302 so that applications can pass data to their callbacks:
9303 The PEM[_ASN1]_{read,write}... functions and macros now take an
9304 additional void * argument, which is just handed through whenever
9305 the password callback is called.
9306 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9308 New function SSL_CTX_set_default_passwd_cb_userdata.
9310 Compatibility note: As many C implementations push function arguments
9311 onto the stack in reverse order, the new library version is likely to
9312 interoperate with programs that have been compiled with the old
9313 pem_password_cb definition (PEM_whatever takes some data that
9314 happens to be on the stack as its last argument, and the callback
9315 just ignores this garbage); but there is no guarantee whatsoever that
9318 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9319 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9320 problems not only on Windows, but also on some Unix platforms.
9321 To avoid problematic command lines, these definitions are now in an
9322 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9323 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9326 *) MIPS III/IV assembler module is reimplemented.
9329 *) More DES library cleanups: remove references to srand/rand and
9330 delete an unused file.
9333 *) Add support for the the free Netwide assembler (NASM) under Win32,
9334 since not many people have MASM (ml) and it can be hard to obtain.
9335 This is currently experimental but it seems to work OK and pass all
9336 the tests. Check out INSTALL.W32 for info.
9339 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9340 without temporary keys kept an extra copy of the server key,
9341 and connections with temporary keys did not free everything in case
9345 *) New function RSA_check_key and new openssl rsa option -check
9346 for verifying the consistency of RSA keys.
9347 [Ulf Moeller, Bodo Moeller]
9349 *) Various changes to make Win32 compile work:
9350 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9351 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9352 comparison" warnings.
9353 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9356 *) Add a debugging option to PKCS#5 v2 key generation function: when
9357 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9358 derived keys are printed to stderr.
9361 *) Copy the flags in ASN1_STRING_dup().
9362 [Roman E. Pavlov <pre@mo.msk.ru>]
9364 *) The x509 application mishandled signing requests containing DSA
9365 keys when the signing key was also DSA and the parameters didn't match.
9367 It was supposed to omit the parameters when they matched the signing key:
9368 the verifying software was then supposed to automatically use the CA's
9369 parameters if they were absent from the end user certificate.
9371 Omitting parameters is no longer recommended. The test was also
9372 the wrong way round! This was probably due to unusual behaviour in
9373 EVP_cmp_parameters() which returns 1 if the parameters match.
9374 This meant that parameters were omitted when they *didn't* match and
9375 the certificate was useless. Certificates signed with 'ca' didn't have
9377 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9379 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9380 The interface is as follows:
9381 Applications can use
9382 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9383 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9384 "off" is now the default.
9385 The library internally uses
9386 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9387 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9388 to disable memory-checking temporarily.
9390 Some inconsistent states that previously were possible (and were
9391 even the default) are now avoided.
9393 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9394 with each memory chunk allocated; this is occasionally more helpful
9395 than just having a counter.
9397 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9399 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9403 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9404 which largely parallels "options", but is for changing API behaviour,
9405 whereas "options" are about protocol behaviour.
9406 Initial "mode" flags are:
9408 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9409 a single record has been written.
9410 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9411 retries use the same buffer location.
9412 (But all of the contents must be
9416 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9419 *) Fix problems with no-hmac etc.
9420 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9422 *) New functions RSA_get_default_method(), RSA_set_method() and
9423 RSA_get_method(). These allows replacement of RSA_METHODs without having
9424 to mess around with the internals of an RSA structure.
9427 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9428 Also really enable memory leak checks in openssl.c and in some
9430 [Chad C. Mulligan, Bodo Moeller]
9432 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9433 up the length of negative integers. This has now been simplified to just
9434 store the length when it is first determined and use it later, rather
9435 than trying to keep track of where data is copied and updating it to
9437 [Steve Henson, reported by Brien Wheeler
9438 <bwheeler@authentica-security.com>]
9440 *) Add a new function PKCS7_signatureVerify. This allows the verification
9441 of a PKCS#7 signature but with the signing certificate passed to the
9442 function itself. This contrasts with PKCS7_dataVerify which assumes the
9443 certificate is present in the PKCS#7 structure. This isn't always the
9444 case: certificates can be omitted from a PKCS#7 structure and be
9445 distributed by "out of band" means (such as a certificate database).
9448 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9449 function prototypes in pem.h, also change util/mkdef.pl to add the
9450 necessary function names.
9453 *) mk1mf.pl (used by Windows builds) did not properly read the
9454 options set by Configure in the top level Makefile, and Configure
9455 was not even able to write more than one option correctly.
9456 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9459 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9460 file to be loaded from a BIO or FILE pointer. The BIO version will
9461 for example allow memory BIOs to contain config info.
9464 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9465 Whoever hopes to achieve shared-library compatibility across versions
9466 must use this, not the compile-time macro.
9467 (Exercise 0.9.4: Which is the minimum library version required by
9469 Note: All this applies only to multi-threaded programs, others don't
9473 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9474 through a BIO pair triggered the default case, i.e.
9475 SSLerr(...,SSL_R_UNKNOWN_STATE).
9478 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9479 can use the SSL library even if none of the specific BIOs is
9483 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9484 for the encoded length.
9485 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9487 *) Add initial documentation of the X509V3 functions.
9490 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9491 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9492 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9493 secure PKCS#8 private key format with a high iteration count.
9496 *) Fix determination of Perl interpreter: A perl or perl5
9497 _directory_ in $PATH was also accepted as the interpreter.
9498 [Ralf S. Engelschall]
9500 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9501 wrong with it but it was very old and did things like calling
9502 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9506 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9507 to use the new extension code.
9510 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9511 with macros. This should make it easier to change their form, add extra
9512 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9516 *) Add to configuration table a new entry that can specify an alternative
9517 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9518 according to Mark Crispin <MRC@Panda.COM>.
9522 *) DES CBC did not update the IV. Weird.
9525 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9526 Changing the behaviour of the former might break existing programs --
9527 where IV updating is needed, des_ncbc_encrypt can be used.
9530 *) When bntest is run from "make test" it drives bc to check its
9531 calculations, as well as internally checking them. If an internal check
9532 fails, it needs to cause bc to give a non-zero result or make test carries
9533 on without noticing the failure. Fixed.
9536 *) DES library cleanups.
9539 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9540 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9541 ciphers. NOTE: although the key derivation function has been verified
9542 against some published test vectors it has not been extensively tested
9543 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9547 *) Instead of "mkdir -p", which is not fully portable, use new
9548 Perl script "util/mkdir-p.pl".
9551 *) Rewrite the way password based encryption (PBE) is handled. It used to
9552 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9553 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9554 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9555 the 'parameter' field of the AlgorithmIdentifier is passed to the
9556 underlying key generation function so it must do its own ASN1 parsing.
9557 This has also changed the EVP_PBE_CipherInit() function which now has a
9558 'parameter' argument instead of literal salt and iteration count values
9559 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9562 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9563 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9564 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9565 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9566 value was just used as a "magic string" and not used directly its
9567 value doesn't matter.
9570 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9574 *) "linux-sparc64" configuration (ultrapenguin).
9575 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9576 "linux-sparc" configuration.
9577 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9579 *) config now generates no-xxx options for missing ciphers.
9582 *) Support the EBCDIC character set (work in progress).
9583 File ebcdic.c not yet included because it has a different license.
9584 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9586 *) Support BS2000/OSD-POSIX.
9587 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9589 *) Make callbacks for key generation use void * instead of char *.
9592 *) Make S/MIME samples compile (not yet tested).
9595 *) Additional typesafe stacks.
9598 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9602 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9604 *) New configuration variant "sco5-gcc".
9606 *) Updated some demos.
9607 [Sean O Riordain, Wade Scholine]
9609 *) Add missing BIO_free at exit of pkcs12 application.
9612 *) Fix memory leak in conf.c.
9615 *) Updates for Win32 to assembler version of MD5.
9618 *) Set #! path to perl in apps/der_chop to where we found it
9619 instead of using a fixed path.
9622 *) SHA library changes for irix64-mips4-cc.
9625 *) Improvements for VMS support.
9629 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9631 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9632 This also avoids the problems with SC4.2 and unpatched SC5.
9633 [Andy Polyakov <appro@fy.chalmers.se>]
9635 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9636 These are required because of the typesafe stack would otherwise break
9637 existing code. If old code used a structure member which used to be STACK
9638 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9639 sk_num or sk_value it would produce an error because the num, data members
9640 are not present in STACK_OF. Now it just produces a warning. sk_set
9641 replaces the old method of assigning a value to sk_value
9642 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9643 that does this will no longer work (and should use sk_set instead) but
9644 this could be regarded as a "questionable" behaviour anyway.
9647 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9648 correctly handle encrypted S/MIME data.
9651 *) Change type of various DES function arguments from des_cblock
9652 (which means, in function argument declarations, pointer to char)
9653 to des_cblock * (meaning pointer to array with 8 char elements),
9654 which allows the compiler to do more typechecking; it was like
9655 that back in SSLeay, but with lots of ugly casts.
9657 Introduce new type const_des_cblock.
9660 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9661 problems: find RecipientInfo structure that matches recipient certificate
9662 and initialise the ASN1 structures properly based on passed cipher.
9665 *) Belatedly make the BN tests actually check the results.
9668 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9669 to and from BNs: it was completely broken. New compilation option
9670 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9671 key elements as negative integers.
9674 *) Reorganize and speed up MD5.
9675 [Andy Polyakov <appro@fy.chalmers.se>]
9678 [Richard Levitte <richard@levitte.org>]
9680 *) New option -out to asn1parse to allow the parsed structure to be
9681 output to a file. This is most useful when combined with the -strparse
9682 option to examine the output of things like OCTET STRINGS.
9685 *) Make SSL library a little more fool-proof by not requiring any longer
9686 that SSL_set_{accept,connect}_state be called before
9687 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9688 in many applications because usually everything *appeared* to work as
9689 intended anyway -- now it really works as intended).
9692 *) Move openssl.cnf out of lib/.
9695 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9696 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9697 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9698 [Ralf S. Engelschall]
9700 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9701 handle PKCS#7 enveloped data properly.
9702 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9704 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9705 copying pointers. The cert_st handling is changed by this in
9706 various ways (and thus what used to be known as ctx->default_cert
9707 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9708 any longer when s->cert does not give us what we need).
9709 ssl_cert_instantiate becomes obsolete by this change.
9710 As soon as we've got the new code right (possibly it already is?),
9711 we have solved a couple of bugs of the earlier code where s->cert
9712 was used as if it could not have been shared with other SSL structures.
9714 Note that using the SSL API in certain dirty ways now will result
9715 in different behaviour than observed with earlier library versions:
9716 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9717 does not influence s as it used to.
9719 In order to clean up things more thoroughly, inside SSL_SESSION
9720 we don't use CERT any longer, but a new structure SESS_CERT
9721 that holds per-session data (if available); currently, this is
9722 the peer's certificate chain and, for clients, the server's certificate
9723 and temporary key. CERT holds only those values that can have
9724 meaningful defaults in an SSL_CTX.
9727 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9728 from the internal representation. Various PKCS#7 fixes: remove some
9729 evil casts and set the enc_dig_alg field properly based on the signing
9733 *) Allow PKCS#12 password to be set from the command line or the
9734 environment. Let 'ca' get its config file name from the environment
9735 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9739 *) Allow certificate policies extension to use an IA5STRING for the
9740 organization field. This is contrary to the PKIX definition but
9741 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9745 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9746 without disallowing inline assembler and the like for non-pedantic builds.
9749 *) Support Borland C++ builder.
9750 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9755 *) SHA-1 cleanups and performance enhancements.
9756 [Andy Polyakov <appro@fy.chalmers.se>]
9758 *) Sparc v8plus assembler for the bignum library.
9759 [Andy Polyakov <appro@fy.chalmers.se>]
9761 *) Accept any -xxx and +xxx compiler options in Configure.
9764 *) Update HPUX configuration.
9767 *) Add missing sk_<type>_unshift() function to safestack.h
9768 [Ralf S. Engelschall]
9770 *) New function SSL_CTX_use_certificate_chain_file that sets the
9771 "extra_cert"s in addition to the certificate. (This makes sense
9772 only for "PEM" format files, as chains as a whole are not
9776 *) Support verify_depth from the SSL API.
9777 x509_vfy.c had what can be considered an off-by-one-error:
9778 Its depth (which was not part of the external interface)
9779 was actually counting the number of certificates in a chain;
9780 now it really counts the depth.
9783 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9784 instead of X509err, which often resulted in confusing error
9785 messages since the error codes are not globally unique
9786 (e.g. an alleged error in ssl3_accept when a certificate
9787 didn't match the private key).
9789 *) New function SSL_CTX_set_session_id_context that allows to set a default
9790 value (so that you don't need SSL_set_session_id_context for each
9791 connection using the SSL_CTX).
9794 *) OAEP decoding bug fix.
9797 *) Support INSTALL_PREFIX for package builders, as proposed by
9801 *) New Configure options "threads" and "no-threads". For systems
9802 where the proper compiler options are known (currently Solaris
9803 and Linux), "threads" is the default.
9806 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9809 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9810 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9811 such as /usr/local/bin.
9814 *) "make linux-shared" to build shared libraries.
9815 [Niels Poppe <niels@netbox.org>]
9817 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9820 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9821 extension adding in x509 utility.
9824 *) Remove NOPROTO sections and error code comments.
9827 *) Partial rewrite of the DEF file generator to now parse the ANSI
9831 *) New Configure options --prefix=DIR and --openssldir=DIR.
9834 *) Complete rewrite of the error code script(s). It is all now handled
9835 by one script at the top level which handles error code gathering,
9836 header rewriting and C source file generation. It should be much better
9837 than the old method: it now uses a modified version of Ulf's parser to
9838 read the ANSI prototypes in all header files (thus the old K&R definitions
9839 aren't needed for error creation any more) and do a better job of
9840 translating function codes into names. The old 'ASN1 error code imbedded
9841 in a comment' is no longer necessary and it doesn't use .err files which
9842 have now been deleted. Also the error code call doesn't have to appear all
9843 on one line (which resulted in some large lines...).
9846 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9849 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9850 0 (which usually indicates a closed connection), but continue reading.
9853 *) Fix some race conditions.
9856 *) Add support for CRL distribution points extension. Add Certificate
9857 Policies and CRL distribution points documentation.
9860 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9863 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9864 8 of keying material. Merlin has also confirmed interop with this fix
9865 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9866 [Merlin Hughes <merlin@baltimore.ie>]
9868 *) Fix lots of warnings.
9869 [Richard Levitte <levitte@stacken.kth.se>]
9871 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9872 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9873 [Richard Levitte <levitte@stacken.kth.se>]
9875 *) Fix problems with sizeof(long) == 8.
9876 [Andy Polyakov <appro@fy.chalmers.se>]
9878 *) Change functions to ANSI C.
9881 *) Fix typos in error codes.
9882 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9884 *) Remove defunct assembler files from Configure.
9887 *) SPARC v8 assembler BIGNUM implementation.
9888 [Andy Polyakov <appro@fy.chalmers.se>]
9890 *) Support for Certificate Policies extension: both print and set.
9891 Various additions to support the r2i method this uses.
9894 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9895 return a const string when you are expecting an allocated buffer.
9898 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9899 types DirectoryString and DisplayText.
9902 *) Add code to allow r2i extensions to access the configuration database,
9903 add an LHASH database driver and add several ctx helper functions.
9906 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9907 fail when they extended the size of a BIGNUM.
9910 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9911 support typesafe stack.
9914 *) Fix typo in SSL_[gs]et_options().
9915 [Nils Frostberg <nils@medcom.se>]
9917 *) Delete various functions and files that belonged to the (now obsolete)
9918 old X509V3 handling code.
9921 *) New Configure option "rsaref".
9924 *) Don't auto-generate pem.h.
9927 *) Introduce type-safe ASN.1 SETs.
9930 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9931 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9933 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9934 that links with OpenSSL (well at least cause lots of warnings), but fear
9935 not: the conversion is trivial, and it eliminates loads of evil casts. A
9936 few STACKed things have been converted already. Feel free to convert more.
9937 In the fullness of time, I'll do away with the STACK type altogether.
9940 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9941 specified in <certfile> by updating the entry in the index.txt file.
9942 This way one no longer has to edit the index.txt file manually for
9943 revoking a certificate. The -revoke option does the gory details now.
9944 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9946 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9947 `-text' option at all and this way the `-noout -text' combination was
9948 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9949 [Ralf S. Engelschall]
9951 *) Make sure a corresponding plain text error message exists for the
9952 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9953 verify callback function determined that a certificate was revoked.
9954 [Ralf S. Engelschall]
9956 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9957 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9958 all available cipers including rc5, which was forgotten until now.
9959 In order to let the testing shell script know which algorithms
9960 are available, a new (up to now undocumented) command
9961 "openssl list-cipher-commands" is used.
9964 *) Bugfix: s_client occasionally would sleep in select() when
9965 it should have checked SSL_pending() first.
9968 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9969 the raw DSA values prior to ASN.1 encoding.
9972 *) Tweaks to Configure
9973 [Niels Poppe <niels@netbox.org>]
9975 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9979 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9982 *) New config option to avoid instructions that are illegal on the 80386.
9983 The default code is faster, but requires at least a 486.
9986 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9987 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9988 same as SSL2_VERSION anyway.
9991 *) New "-showcerts" option for s_client.
9994 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9995 application. Various cleanups and fixes.
9998 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9999 modify error routines to work internally. Add error codes and PBE init
10000 to library startup routines.
10003 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10004 packing functions to asn1 and evp. Changed function names and error
10005 codes along the way.
10008 *) PKCS12 integration: and so it begins... First of several patches to
10009 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10010 objects to objects.h
10013 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10014 and display support for Thawte strong extranet extension.
10017 *) Add LinuxPPC support.
10018 [Jeff Dubrule <igor@pobox.org>]
10020 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10021 bn_div_words in alpha.s.
10022 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10024 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10025 OAEP isn't supported when OpenSSL is built with RSAref.
10026 [Ulf Moeller <ulf@fitug.de>]
10028 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10029 so they no longer are missing under -DNOPROTO.
10030 [Soren S. Jorvang <soren@t.dk>]
10033 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10035 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10036 doesn't work when the session is reused. Coming soon!
10039 *) Fix a security hole, that allows sessions to be reused in the wrong
10040 context thus bypassing client cert protection! All software that uses
10041 client certs and session caches in multiple contexts NEEDS PATCHING to
10042 allow session reuse! A fuller solution is in the works.
10043 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10045 *) Some more source tree cleanups (removed obsolete files
10046 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10047 permission on "config" script to be executable) and a fix for the INSTALL
10049 [Ulf Moeller <ulf@fitug.de>]
10051 *) Remove some legacy and erroneous uses of malloc, free instead of
10053 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10055 *) Make rsa_oaep_test return non-zero on error.
10056 [Ulf Moeller <ulf@fitug.de>]
10058 *) Add support for native Solaris shared libraries. Configure
10059 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10060 if someone would make that last step automatic.
10061 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10063 *) ctx_size was not built with the right compiler during "make links". Fixed.
10066 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10067 except NULL ciphers". This means the default cipher list will no longer
10068 enable NULL ciphers. They need to be specifically enabled e.g. with
10069 the string "DEFAULT:eNULL".
10072 *) Fix to RSA private encryption routines: if p < q then it would
10073 occasionally produce an invalid result. This will only happen with
10074 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10077 *) Be less restrictive and allow also `perl util/perlpath.pl
10078 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10079 because this way one can also use an interpreter named `perl5' (which is
10080 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10081 installed as `perl').
10082 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10084 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10085 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10087 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10088 advapi32.lib to Win32 build and change the pem test comparision
10089 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10090 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10091 and crypto/des/ede_cbcm_enc.c.
10094 *) DES quad checksum was broken on big-endian architectures. Fixed.
10097 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10098 Win32 test batch file so it (might) work again. The Win32 test batch file
10099 is horrible: I feel ill....
10102 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10103 in e_os.h. Audit of header files to check ANSI and non ANSI
10104 sections: 10 functions were absent from non ANSI section and not exported
10105 from Windows DLLs. Fixed up libeay.num for new functions.
10108 *) Make `openssl version' output lines consistent.
10109 [Ralf S. Engelschall]
10111 *) Fix Win32 symbol export lists for BIO functions: Added
10112 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10113 to ms/libeay{16,32}.def.
10114 [Ralf S. Engelschall]
10116 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10117 fine under Unix and passes some trivial tests I've now added. But the
10118 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10119 added to make sure no one expects that this stuff really works in the
10120 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10121 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10123 [Ralf S. Engelschall]
10125 *) Fix the generation of two part addresses in perl.
10126 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10128 *) Add config entry for Linux on MIPS.
10129 [John Tobey <jtobey@channel1.com>]
10131 *) Make links whenever Configure is run, unless we are on Windoze.
10134 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10135 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10139 *) Add a useful kludge to allow package maintainers to specify compiler and
10140 other platforms details on the command line without having to patch the
10141 Configure script everytime: One now can use ``perl Configure
10142 <id>:<details>'', i.e. platform ids are allowed to have details appended
10143 to them (seperated by colons). This is treated as there would be a static
10144 pre-configured entry in Configure's %table under key <id> with value
10145 <details> and ``perl Configure <id>'' is called. So, when you want to
10146 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10147 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10148 now, which overrides the FreeBSD-elf entry on-the-fly.
10149 [Ralf S. Engelschall]
10151 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10154 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10155 on the `perl Configure ...' command line. This way one can compile
10156 OpenSSL libraries with Position Independent Code (PIC) which is needed
10157 for linking it into DSOs.
10158 [Ralf S. Engelschall]
10160 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10164 *) Cleaned up the LICENSE document: The official contact for any license
10165 questions now is the OpenSSL core team under openssl-core@openssl.org.
10166 And add a paragraph about the dual-license situation to make sure people
10167 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10168 to the OpenSSL toolkit.
10169 [Ralf S. Engelschall]
10171 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10172 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10173 Additonally cleaned up the `make links' target: Remove unnecessary
10174 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10175 to speed processing and no longer clutter the display with confusing
10176 stuff. Instead only the actually done links are displayed.
10177 [Ralf S. Engelschall]
10179 *) Permit null encryption ciphersuites, used for authentication only. It used
10180 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10181 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10185 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10186 signed attributes when verifying signatures (this would break them),
10187 the detached data encoding was wrong and public keys obtained using
10188 X509_get_pubkey() weren't freed.
10191 *) Add text documentation for the BUFFER functions. Also added a work around
10192 to a Win95 console bug. This was triggered by the password read stuff: the
10193 last character typed gets carried over to the next fread(). If you were
10194 generating a new cert request using 'req' for example then the last
10195 character of the passphrase would be CR which would then enter the first
10199 *) Added the new `Includes OpenSSL Cryptography Software' button as
10200 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10201 button and can be used by applications based on OpenSSL to show the
10202 relationship to the OpenSSL project.
10203 [Ralf S. Engelschall]
10205 *) Remove confusing variables in function signatures in files
10206 ssl/ssl_lib.c and ssl/ssl.h.
10207 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10209 *) Don't install bss_file.c under PREFIX/include/
10210 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10212 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10213 functions that return function pointers and has support for NT specific
10214 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10215 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10216 unsigned to signed types: this was killing the Win32 compile.
10219 *) Add new certificate file to stack functions,
10220 SSL_add_dir_cert_subjects_to_stack() and
10221 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10222 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10223 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10224 This means that Apache-SSL and similar packages don't have to mess around
10225 to add as many CAs as they want to the preferred list.
10228 *) Experiment with doxygen documentation. Currently only partially applied to
10230 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10231 openssl.doxy as the configuration file.
10234 *) Get rid of remaining C++-style comments which strict C compilers hate.
10235 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10237 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10238 compiled in by default: it has problems with large keys.
10241 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10242 DH private keys and/or callback functions which directly correspond to
10243 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10244 is needed for applications which have to configure certificates on a
10245 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10247 For the RSA certificate situation is makes no difference, but
10248 for the DSA certificate situation this fixes the "no shared cipher"
10249 problem where the OpenSSL cipher selection procedure failed because the
10250 temporary keys were not overtaken from the context and the API provided
10251 no way to reconfigure them.
10252 The new functions now let applications reconfigure the stuff and they
10253 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10254 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10255 non-public-API function ssl_cert_instantiate() is used as a helper
10256 function and also to reduce code redundancy inside ssl_rsa.c.
10257 [Ralf S. Engelschall]
10259 *) Move s_server -dcert and -dkey options out of the undocumented feature
10260 area because they are useful for the DSA situation and should be
10261 recognized by the users.
10262 [Ralf S. Engelschall]
10264 *) Fix the cipher decision scheme for export ciphers: the export bits are
10265 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10266 SSL_EXP_MASK. So, the original variable has to be used instead of the
10267 already masked variable.
10268 [Richard Levitte <levitte@stacken.kth.se>]
10270 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10271 [Richard Levitte <levitte@stacken.kth.se>]
10273 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10274 from `int' to `unsigned int' because it's a length and initialized by
10275 EVP_DigestFinal() which expects an `unsigned int *'.
10276 [Richard Levitte <levitte@stacken.kth.se>]
10278 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10279 script. Instead use the usual Shell->Perl transition trick.
10280 [Ralf S. Engelschall]
10282 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10283 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10284 -noout -modulus' as it's already the case for `openssl rsa -noout
10285 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10286 currently the public key is printed (a decision which was already done by
10287 `openssl dsa -modulus' in the past) which serves a similar purpose.
10288 Additionally the NO_RSA no longer completely removes the whole -modulus
10289 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10291 [Ralf S. Engelschall]
10293 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10294 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10295 [Arne Ansper <arne@ats.cyber.ee>]
10297 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10298 to be added. Now both 'req' and 'ca' can use new objects defined in the
10302 *) Add cool BIO that does syslog (or event log on NT).
10303 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10305 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10306 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10307 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10308 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10311 *) Add preliminary config info for new extension code.
10314 *) Make RSA_NO_PADDING really use no padding.
10315 [Ulf Moeller <ulf@fitug.de>]
10317 *) Generate errors when private/public key check is done.
10320 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10321 for some CRL extensions and new objects added.
10324 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10325 key usage extension and fuller support for authority key id.
10328 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10329 padding method for RSA, which is recommended for new applications in PKCS
10330 #1 v2.0 (RFC 2437, October 1998).
10331 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10332 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10333 against Bleichbacher's attack on RSA.
10334 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10337 *) Updates to the new SSL compression code
10338 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10340 *) Fix so that the version number in the master secret, when passed
10341 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10342 (because the server will not accept higher), that the version number
10343 is 0x03,0x01, not 0x03,0x00
10344 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10346 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10347 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10348 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10351 *) Support for RAW extensions where an arbitrary extension can be
10352 created by including its DER encoding. See apps/openssl.cnf for
10356 *) Make sure latest Perl versions don't interpret some generated C array
10357 code as Perl array code in the crypto/err/err_genc.pl script.
10358 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10360 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10361 not many people have the assembler. Various Win32 compilation fixes and
10362 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10363 build instructions.
10366 *) Modify configure script 'Configure' to automatically create crypto/date.h
10367 file under Win32 and also build pem.h from pem.org. New script
10368 util/mkfiles.pl to create the MINFO file on environments that can't do a
10369 'make files': perl util/mkfiles.pl >MINFO should work.
10372 *) Major rework of DES function declarations, in the pursuit of correctness
10373 and purity. As a result, many evil casts evaporated, and some weirdness,
10374 too. You may find this causes warnings in your code. Zapping your evil
10375 casts will probably fix them. Mostly.
10378 *) Fix for a typo in asn1.h. Bug fix to object creation script
10379 obj_dat.pl. It considered a zero in an object definition to mean
10380 "end of object": none of the objects in objects.h have any zeros
10381 so it wasn't spotted.
10382 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10384 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10385 Masking (CBCM). In the absence of test vectors, the best I have been able
10386 to do is check that the decrypt undoes the encrypt, so far. Send me test
10387 vectors if you have them.
10390 *) Correct calculation of key length for export ciphers (too much space was
10391 allocated for null ciphers). This has not been tested!
10394 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10395 message is now correct (it understands "crypto" and "ssl" on its
10396 command line). There is also now an "update" option. This will update
10397 the util/ssleay.num and util/libeay.num files with any new functions.
10399 perl util/mkdef.pl crypto ssl update
10400 it will update them.
10403 *) Overhauled the Perl interface (perl/*):
10404 - ported BN stuff to OpenSSL's different BN library
10405 - made the perl/ source tree CVS-aware
10406 - renamed the package from SSLeay to OpenSSL (the files still contain
10407 their history because I've copied them in the repository)
10408 - removed obsolete files (the test scripts will be replaced
10409 by better Test::Harness variants in the future)
10410 [Ralf S. Engelschall]
10412 *) First cut for a very conservative source tree cleanup:
10413 1. merge various obsolete readme texts into doc/ssleay.txt
10414 where we collect the old documents and readme texts.
10415 2. remove the first part of files where I'm already sure that we no
10416 longer need them because of three reasons: either they are just temporary
10417 files which were left by Eric or they are preserved original files where
10418 I've verified that the diff is also available in the CVS via "cvs diff
10419 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10420 the crypto/md/ stuff).
10421 [Ralf S. Engelschall]
10423 *) More extension code. Incomplete support for subject and issuer alt
10424 name, issuer and authority key id. Change the i2v function parameters
10425 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10426 what that's for :-) Fix to ASN1 macro which messed up
10427 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10430 *) Preliminary support for ENUMERATED type. This is largely copied from the
10434 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10435 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10437 *) Make sure `make rehash' target really finds the `openssl' program.
10438 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10440 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10441 like to hear about it if this slows down other processors.
10444 *) Add CygWin32 platform information to Configure script.
10445 [Alan Batie <batie@aahz.jf.intel.com>]
10447 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10448 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10450 *) New program nseq to manipulate netscape certificate sequences
10453 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10457 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10458 but the BN code had some problems that would cause failures when
10459 doing certificate verification and some other functions.
10460 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10462 *) Add ASN1 and PEM code to support netscape certificate sequences.
10465 *) Add ASN1 and PEM code to support netscape certificate sequences.
10468 *) Add several PKIX and private extended key usage OIDs.
10471 *) Modify the 'ca' program to handle the new extension code. Modify
10472 openssl.cnf for new extension format, add comments.
10475 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10476 and add a sample to openssl.cnf so req -x509 now adds appropriate
10480 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10481 error code, add initial support to X509_print() and x509 application.
10484 *) Takes a deep breath and start addding X509 V3 extension support code. Add
10485 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10486 stuff is currently isolated and isn't even compiled yet.
10489 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10490 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10491 Removed the versions check from X509 routines when loading extensions:
10492 this allows certain broken certificates that don't set the version
10493 properly to be processed.
10496 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10497 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10498 can still be regenerated with "make depend".
10501 *) Spelling mistake in C version of CAST-128.
10502 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10504 *) Changes to the error generation code. The perl script err-code.pl
10505 now reads in the old error codes and retains the old numbers, only
10506 adding new ones if necessary. It also only changes the .err files if new
10507 codes are added. The makefiles have been modified to only insert errors
10508 when needed (to avoid needlessly modifying header files). This is done
10509 by only inserting errors if the .err file is newer than the auto generated
10510 C file. To rebuild all the error codes from scratch (the old behaviour)
10511 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10512 or delete all the .err files.
10515 *) CAST-128 was incorrectly implemented for short keys. The C version has
10516 been fixed, but is untested. The assembler versions are also fixed, but
10517 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10518 to regenerate it if needed.
10519 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10520 Hagino <itojun@kame.net>]
10522 *) File was opened incorrectly in randfile.c.
10523 [Ulf Möller <ulf@fitug.de>]
10525 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10526 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10527 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10528 al: it's just almost always a UTCTime. Note this patch adds new error
10529 codes so do a "make errors" if there are problems.
10532 *) Correct Linux 1 recognition in config.
10533 [Ulf Möller <ulf@fitug.de>]
10535 *) Remove pointless MD5 hash when using DSA keys in ca.
10536 [Anonymous <nobody@replay.com>]
10538 *) Generate an error if given an empty string as a cert directory. Also
10539 generate an error if handed NULL (previously returned 0 to indicate an
10540 error, but didn't set one).
10541 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10543 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10546 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10547 parameters. This was causing a warning which killed off the Win32 compile.
10550 *) Remove C++ style comments from crypto/bn/bn_local.h.
10551 [Neil Costigan <neil.costigan@celocom.com>]
10553 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10554 based on a text string, looking up short and long names and finally
10555 "dot" format. The "dot" format stuff didn't work. Added new function
10556 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10557 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10558 OID is not part of the table.
10561 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10562 X509_LOOKUP_by_alias().
10565 *) Sort openssl functions by name.
10568 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10569 encryption from sample DSA keys (in case anyone is interested the password
10573 *) Make _all_ *_free functions accept a NULL pointer.
10574 [Frans Heymans <fheymans@isaserver.be>]
10576 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10578 [Anonymous <nobody@replay.com>]
10580 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10581 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10583 *) Don't blow it for numeric -newkey arguments to apps/req.
10584 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10586 *) Temp key "for export" tests were wrong in s3_srvr.c.
10587 [Anonymous <nobody@replay.com>]
10589 *) Add prototype for temp key callback functions
10590 SSL_CTX_set_tmp_{rsa,dh}_callback().
10593 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10594 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10597 *) X509_name_add_entry() freed the wrong thing after an error.
10598 [Arne Ansper <arne@ats.cyber.ee>]
10600 *) rsa_eay.c would attempt to free a NULL context.
10601 [Arne Ansper <arne@ats.cyber.ee>]
10603 *) BIO_s_socket() had a broken should_retry() on Windoze.
10604 [Arne Ansper <arne@ats.cyber.ee>]
10606 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10607 [Arne Ansper <arne@ats.cyber.ee>]
10609 *) Make sure the already existing X509_STORE->depth variable is initialized
10610 in X509_STORE_new(), but document the fact that this variable is still
10611 unused in the certificate verification process.
10612 [Ralf S. Engelschall]
10614 *) Fix the various library and apps files to free up pkeys obtained from
10615 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10618 *) Fix reference counting in X509_PUBKEY_get(). This makes
10619 demos/maurice/example2.c work, amongst others, probably.
10620 [Steve Henson and Ben Laurie]
10622 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10623 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10624 are no longer created. This way we have a single and consistent command
10625 line interface `openssl <command>', similar to `cvs <command>'.
10626 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10628 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10629 BIT STRING wrapper always have zero unused bits.
10632 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10635 *) Make the top-level INSTALL documentation easier to understand.
10638 *) Makefiles updated to exit if an error occurs in a sub-directory
10639 make (including if user presses ^C) [Paul Sutton]
10641 *) Make Montgomery context stuff explicit in RSA data structure.
10644 *) Fix build order of pem and err to allow for generated pem.h.
10647 *) Fix renumbering bug in X509_NAME_delete_entry().
10650 *) Enhanced the err-ins.pl script so it makes the error library number
10651 global and can add a library name. This is needed for external ASN1 and
10652 other error libraries.
10655 *) Fixed sk_insert which never worked properly.
10658 *) Fix ASN1 macros so they can handle indefinite length construted
10659 EXPLICIT tags. Some non standard certificates use these: they can now
10663 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10664 into a single doc/ssleay.txt bundle. This way the information is still
10665 preserved but no longer messes up this directory. Now it's new room for
10666 the new set of documenation files.
10667 [Ralf S. Engelschall]
10669 *) SETs were incorrectly DER encoded. This was a major pain, because they
10670 shared code with SEQUENCEs, which aren't coded the same. This means that
10671 almost everything to do with SETs or SEQUENCEs has either changed name or
10672 number of arguments.
10673 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10675 *) Fix test data to work with the above.
10678 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10679 was already fixed by Eric for 0.9.1 it seems.
10680 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10682 *) Autodetect FreeBSD3.
10685 *) Fix various bugs in Configure. This affects the following platforms:
10689 unixware-2.0-pentium
10693 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10694 before they are needed.
10697 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10701 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10703 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10704 changed SSLeay to OpenSSL in version strings.
10705 [Ralf S. Engelschall]
10707 *) Some fixups to the top-level documents.
10710 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10711 because the symlink to include/ was missing.
10712 [Ralf S. Engelschall]
10714 *) Incorporated the popular no-RSA/DSA-only patches
10715 which allow to compile a RSA-free SSLeay.
10716 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10718 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10719 when "ssleay" is still not found.
10720 [Ralf S. Engelschall]
10722 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10723 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10725 *) Updated the README file.
10726 [Ralf S. Engelschall]
10728 *) Added various .cvsignore files in the CVS repository subdirs
10729 to make a "cvs update" really silent.
10730 [Ralf S. Engelschall]
10732 *) Recompiled the error-definition header files and added
10733 missing symbols to the Win32 linker tables.
10734 [Ralf S. Engelschall]
10736 *) Cleaned up the top-level documents;
10737 o new files: CHANGES and LICENSE
10738 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10739 o merged COPYRIGHT into LICENSE
10740 o removed obsolete TODO file
10741 o renamed MICROSOFT to INSTALL.W32
10742 [Ralf S. Engelschall]
10744 *) Removed dummy files from the 0.9.1b source tree:
10745 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10746 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10747 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10748 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10749 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10750 [Ralf S. Engelschall]
10752 *) Added various platform portability fixes.
10755 *) The Genesis of the OpenSSL rpject:
10756 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10757 Young and Tim J. Hudson created while they were working for C2Net until
10759 [The OpenSSL Project]
10762 Changes between 0.9.0b and 0.9.1b [not released]
10764 *) Updated a few CA certificates under certs/
10767 *) Changed some BIGNUM api stuff.
10770 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10771 DGUX x86, Linux Alpha, etc.
10774 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10775 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10779 *) Add -strparse option to asn1pars program which parses nested
10781 [Dr Stephen Henson <shenson@bigfoot.com>]
10783 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10786 *) DSA fix for "ca" program.
10789 *) Added "-genkey" option to "dsaparam" program.
10792 *) Added RIPE MD160 (rmd160) message digest.
10795 *) Added -a (all) option to "ssleay version" command.
10798 *) Added PLATFORM define which is the id given to Configure.
10801 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10804 *) Extended the ASN.1 parser routines.
10807 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10810 *) Added a BN_CTX to the BN library.
10813 *) Fixed the weak key values in DES library
10816 *) Changed API in EVP library for cipher aliases.
10819 *) Added support for RC2/64bit cipher.
10822 *) Converted the lhash library to the crypto/mem.c functions.
10825 *) Added more recognized ASN.1 object ids.
10828 *) Added more RSA padding checks for SSL/TLS.
10831 *) Added BIO proxy/filter functionality.
10834 *) Added extra_certs to SSL_CTX which can be used
10835 send extra CA certificates to the client in the CA cert chain sending
10836 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10839 *) Now Fortezza is denied in the authentication phase because
10840 this is key exchange mechanism is not supported by SSLeay at all.
10843 *) Additional PKCS1 checks.
10846 *) Support the string "TLSv1" for all TLS v1 ciphers.
10849 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10850 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10853 *) Fixed a few memory leaks.
10856 *) Fixed various code and comment typos.
10859 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10860 bytes sent in the client random.
10861 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob