5 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
7 *) ClientHello sigalgs DoS fix
9 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
10 invalid signature algorithms extension a NULL pointer dereference will
11 occur. This can be exploited in a DoS attack against the server.
13 This issue was was reported to OpenSSL by David Ramos of Stanford
16 [Stephen Henson and Matt Caswell]
18 *) Multiblock corrupted pointer fix
20 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
21 feature only applies on 64 bit x86 architecture platforms that support AES
22 NI instructions. A defect in the implementation of "multiblock" can cause
23 OpenSSL's internal write buffer to become incorrectly set to NULL when
24 using non-blocking IO. Typically, when the user application is using a
25 socket BIO for writing, this will only result in a failed connection.
26 However if some other BIO is used then it is likely that a segmentation
27 fault will be triggered, thus enabling a potential DoS attack.
29 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
33 *) Segmentation fault in DTLSv1_listen fix
35 The DTLSv1_listen function is intended to be stateless and processes the
36 initial ClientHello from many peers. It is common for user code to loop
37 over the call to DTLSv1_listen until a valid ClientHello is received with
38 an associated cookie. A defect in the implementation of DTLSv1_listen means
39 that state is preserved in the SSL object from one invocation to the next
40 that can lead to a segmentation fault. Errors processing the initial
41 ClientHello can trigger this scenario. An example of such an error could be
42 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
45 This issue was reported to OpenSSL by Per Allansson.
49 *) Segmentation fault in ASN1_TYPE_cmp fix
51 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
52 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
53 certificate signature algorithm consistency this can be used to crash any
54 certificate verification operation and exploited in a DoS attack. Any
55 application which performs certificate verification is vulnerable including
56 OpenSSL clients and servers which enable client authentication.
60 *) Segmentation fault for invalid PSS parameters fix
62 The signature verification routines will crash with a NULL pointer
63 dereference if presented with an ASN.1 signature using the RSA PSS
64 algorithm and invalid parameters. Since these routines are used to verify
65 certificate signature algorithms this can be used to crash any
66 certificate verification operation and exploited in a DoS attack. Any
67 application which performs certificate verification is vulnerable including
68 OpenSSL clients and servers which enable client authentication.
70 This issue was was reported to OpenSSL by Brian Carpenter.
74 *) ASN.1 structure reuse memory corruption fix
76 Reusing a structure in ASN.1 parsing may allow an attacker to cause
77 memory corruption via an invalid write. Such reuse is and has been
78 strongly discouraged and is believed to be rare.
80 Applications that parse structures containing CHOICE or ANY DEFINED BY
81 components may be affected. Certificate parsing (d2i_X509 and related
82 functions) are however not affected. OpenSSL clients and servers are
87 *) PKCS7 NULL pointer dereferences fix
89 The PKCS#7 parsing code does not handle missing outer ContentInfo
90 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
91 missing content and trigger a NULL pointer dereference on parsing.
93 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
94 otherwise parse PKCS#7 structures from untrusted sources are
95 affected. OpenSSL clients and servers are not affected.
97 This issue was reported to OpenSSL by Michal Zalewski (Google).
101 *) DoS via reachable assert in SSLv2 servers fix
103 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
104 servers that both support SSLv2 and enable export cipher suites by sending
105 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
107 This issue was discovered by Sean Burford (Google) and Emilia Käsper
108 (OpenSSL development team).
112 *) Empty CKE with client auth and DHE fix
114 If client auth is used then a server can seg fault in the event of a DHE
115 ciphersuite being selected and a zero length ClientKeyExchange message
116 being sent by the client. This could be exploited in a DoS attack.
120 *) Handshake with unseeded PRNG fix
122 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
123 with an unseeded PRNG. The conditions are:
124 - The client is on a platform where the PRNG has not been seeded
125 automatically, and the user has not seeded manually
126 - A protocol specific client method version has been used (i.e. not
127 SSL_client_methodv23)
128 - A ciphersuite is used that does not require additional random data from
129 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
131 If the handshake succeeds then the client random that has been used will
132 have been generated from a PRNG with insufficient entropy and therefore the
133 output may be predictable.
135 For example using the following command with an unseeded openssl will
136 succeed on an unpatched platform:
138 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
142 *) Use After Free following d2i_ECPrivatekey error fix
144 A malformed EC private key file consumed via the d2i_ECPrivateKey function
145 could cause a use after free condition. This, in turn, could cause a double
146 free in several private key parsing functions (such as d2i_PrivateKey
147 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
148 for applications that receive EC private keys from untrusted
149 sources. This scenario is considered rare.
151 This issue was discovered by the BoringSSL project and fixed in their
156 *) X509_to_X509_REQ NULL pointer deref fix
158 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
159 the certificate key is invalid. This function is rarely used in practice.
161 This issue was discovered by Brian Carpenter.
165 *) Removed the export ciphers from the DEFAULT ciphers
168 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
170 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
171 ARMv5 through ARMv8, as opposite to "locking" it to single one.
172 So far those who have to target multiple plaforms would compromise
173 and argue that binary targeting say ARMv5 would still execute on
174 ARMv8. "Universal" build resolves this compromise by providing
175 near-optimal performance even on newer platforms.
178 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
179 (other platforms pending).
180 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
182 *) Add support for the SignedCertificateTimestampList certificate and
183 OCSP response extensions from RFC6962.
186 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
187 for corner cases. (Certain input points at infinity could lead to
188 bogus results, with non-infinity inputs mapped to infinity too.)
191 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
192 This covers AES, SHA256/512 and GHASH. "Initial" means that most
193 common cases are optimized and there still is room for further
194 improvements. Vector Permutation AES for Altivec is also added.
197 *) Add support for little-endian ppc64 Linux target.
198 [Marcelo Cerri (IBM)]
200 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
201 SHA1, SHA256 and GHASH. "Initial" means that most common cases
202 are optimized and there still is room for further improvements.
203 Both 32- and 64-bit modes are supported.
204 [Andy Polyakov, Ard Biesheuvel (Linaro)]
206 *) Improved ARMv7 NEON support.
209 *) Support for SPARC Architecture 2011 crypto extensions, first
210 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
211 SHA256/512, MD5, GHASH and modular exponentiation.
212 [Andy Polyakov, David Miller]
214 *) Accelerated modular exponentiation for Intel processors, a.k.a.
216 [Shay Gueron & Vlad Krasnov (Intel Corp)]
218 *) Support for new and upcoming Intel processors, including AVX2,
219 BMI and SHA ISA extensions. This includes additional "stitched"
220 implementations, AESNI-SHA256 and GCM, and multi-buffer support
223 This work was sponsored by Intel Corp.
226 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
227 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
228 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
231 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
232 this fixes a limiation in previous versions of OpenSSL.
235 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
236 MGF1 digest and OAEP label.
239 *) Add EVP support for key wrapping algorithms, to avoid problems with
240 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
241 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
242 algorithms and include tests cases.
245 *) Add functions to allocate and set the fields of an ECDSA_METHOD
247 [Douglas E. Engert, Steve Henson]
249 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
250 difference in days and seconds between two tm or ASN1_TIME structures.
253 *) Add -rev test option to s_server to just reverse order of characters
254 received by client and send back to server. Also prints an abbreviated
255 summary of the connection parameters.
258 *) New option -brief for s_client and s_server to print out a brief summary
259 of connection parameters.
262 *) Add callbacks for arbitrary TLS extensions.
263 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
265 *) New option -crl_download in several openssl utilities to download CRLs
266 from CRLDP extension in certificates.
269 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
272 *) New function X509_CRL_diff to generate a delta CRL from the difference
273 of two full CRLs. Add support to "crl" utility.
276 *) New functions to set lookup_crls function and to retrieve
277 X509_STORE from X509_STORE_CTX.
280 *) Print out deprecated issuer and subject unique ID fields in
284 *) Extend OCSP I/O functions so they can be used for simple general purpose
285 HTTP as well as OCSP. New wrapper function which can be used to download
286 CRLs using the OCSP API.
289 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
292 *) SSL_CONF* functions. These provide a common framework for application
293 configuration using configuration files or command lines.
296 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
297 message callback and prints the results. Needs compile time option
298 "enable-ssl-trace". New options to s_client and s_server to enable
302 *) New ctrl and macro to retrieve supported points extensions.
303 Print out extension in s_server and s_client.
306 *) New functions to retrieve certificate signature and signature
310 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
314 *) New Suite B modes for TLS code. These use and enforce the requirements
315 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
316 only use Suite B curves. The Suite B modes can be set by using the
317 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
320 *) New chain verification flags for Suite B levels of security. Check
321 algorithms are acceptable when flags are set in X509_verify_cert.
324 *) Make tls1_check_chain return a set of flags indicating checks passed
325 by a certificate chain. Add additional tests to handle client
326 certificates: checks for matching certificate type and issuer name
330 *) If an attempt is made to use a signature algorithm not in the peer
331 preference list abort the handshake. If client has no suitable
332 signature algorithms in response to a certificate request do not
336 *) If server EC tmp key is not in client preference list abort handshake.
339 *) Add support for certificate stores in CERT structure. This makes it
340 possible to have different stores per SSL structure or one store in
341 the parent SSL_CTX. Include distint stores for certificate chain
342 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
343 to build and store a certificate chain in CERT structure: returing
344 an error if the chain cannot be built: this will allow applications
345 to test if a chain is correctly configured.
347 Note: if the CERT based stores are not set then the parent SSL_CTX
348 store is used to retain compatibility with existing behaviour.
352 *) New function ssl_set_client_disabled to set a ciphersuite disabled
353 mask based on the current session, check mask when sending client
354 hello and checking the requested ciphersuite.
357 *) New ctrls to retrieve and set certificate types in a certificate
358 request message. Print out received values in s_client. If certificate
359 types is not set with custom values set sensible values based on
360 supported signature algorithms.
363 *) Support for distinct client and server supported signature algorithms.
366 *) Add certificate callback. If set this is called whenever a certificate
367 is required by client or server. An application can decide which
368 certificate chain to present based on arbitrary criteria: for example
369 supported signature algorithms. Add very simple example to s_server.
370 This fixes many of the problems and restrictions of the existing client
371 certificate callback: for example you can now clear an existing
372 certificate and specify the whole chain.
375 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
376 the certificate can be used for (if anything). Set valid_flags field
377 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
378 to have similar checks in it.
380 Add new "cert_flags" field to CERT structure and include a "strict mode".
381 This enforces some TLS certificate requirements (such as only permitting
382 certificate signature algorithms contained in the supported algorithms
383 extension) which some implementations ignore: this option should be used
384 with caution as it could cause interoperability issues.
387 *) Update and tidy signature algorithm extension processing. Work out
388 shared signature algorithms based on preferences and peer algorithms
389 and print them out in s_client and s_server. Abort handshake if no
390 shared signature algorithms.
393 *) Add new functions to allow customised supported signature algorithms
394 for SSL and SSL_CTX structures. Add options to s_client and s_server
398 *) New function SSL_certs_clear() to delete all references to certificates
399 from an SSL structure. Before this once a certificate had been added
400 it couldn't be removed.
403 *) Integrate hostname, email address and IP address checking with certificate
404 verification. New verify options supporting checking in opensl utility.
407 *) Fixes and wildcard matching support to hostname and email checking
408 functions. Add manual page.
409 [Florian Weimer (Red Hat Product Security Team)]
411 *) New functions to check a hostname email or IP address against a
412 certificate. Add options x509 utility to print results of checks against
416 *) Fix OCSP checking.
417 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
419 *) Initial experimental support for explicitly trusted non-root CAs.
420 OpenSSL still tries to build a complete chain to a root but if an
421 intermediate CA has a trust setting included that is used. The first
422 setting is used: whether to trust (e.g., -addtrust option to the x509
426 *) Add -trusted_first option which attempts to find certificates in the
427 trusted store even if an untrusted chain is also supplied.
430 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
431 platform support for Linux and Android.
434 *) Support for linux-x32, ILP32 environment in x86_64 framework.
437 *) Experimental multi-implementation support for FIPS capable OpenSSL.
438 When in FIPS mode the approved implementations are used as normal,
439 when not in FIPS mode the internal unapproved versions are used instead.
440 This means that the FIPS capable OpenSSL isn't forced to use the
441 (often lower perfomance) FIPS implementations outside FIPS mode.
444 *) Transparently support X9.42 DH parameters when calling
445 PEM_read_bio_DHparameters. This means existing applications can handle
446 the new parameter format automatically.
449 *) Initial experimental support for X9.42 DH parameter format: mainly
450 to support use of 'q' parameter for RFC5114 parameters.
453 *) Add DH parameters from RFC5114 including test data to dhtest.
456 *) Support for automatic EC temporary key parameter selection. If enabled
457 the most preferred EC parameters are automatically used instead of
458 hardcoded fixed parameters. Now a server just has to call:
459 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
460 support ECDH and use the most appropriate parameters.
463 *) Enhance and tidy EC curve and point format TLS extension code. Use
464 static structures instead of allocation if default values are used.
465 New ctrls to set curves we wish to support and to retrieve shared curves.
466 Print out shared curves in s_server. New options to s_server and s_client
467 to set list of supported curves.
470 *) New ctrls to retrieve supported signature algorithms and
471 supported curve values as an array of NIDs. Extend openssl utility
472 to print out received values.
475 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
476 between NIDs and the more common NIST names such as "P-256". Enhance
477 ecparam utility and ECC method to recognise the NIST names for curves.
480 *) Enhance SSL/TLS certificate chain handling to support different
481 chains for each certificate instead of one chain in the parent SSL_CTX.
484 *) Support for fixed DH ciphersuite client authentication: where both
485 server and client use DH certificates with common parameters.
488 *) Support for fixed DH ciphersuites: those requiring DH server
492 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
494 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
495 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
496 X509_CINF_get_signature were reverted post internal team review.
498 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
500 *) Build fixes for the Windows and OpenVMS platforms
501 [Matt Caswell and Richard Levitte]
503 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
505 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
506 message can cause a segmentation fault in OpenSSL due to a NULL pointer
507 dereference. This could lead to a Denial Of Service attack. Thanks to
508 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
512 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
513 dtls1_buffer_record function under certain conditions. In particular this
514 could occur if an attacker sent repeated DTLS records with the same
515 sequence number but for the next epoch. The memory leak could be exploited
516 by an attacker in a Denial of Service attack through memory exhaustion.
517 Thanks to Chris Mueller for reporting this issue.
521 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
522 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
523 method would be set to NULL which could later result in a NULL pointer
524 dereference. Thanks to Frank Schmirler for reporting this issue.
528 *) Abort handshake if server key exchange message is omitted for ephemeral
531 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
532 reporting this issue.
536 *) Remove non-export ephemeral RSA code on client and server. This code
537 violated the TLS standard by allowing the use of temporary RSA keys in
538 non-export ciphersuites and could be used by a server to effectively
539 downgrade the RSA key length used to a value smaller than the server
540 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
541 INRIA or reporting this issue.
545 *) Fixed issue where DH client certificates are accepted without verification.
546 An OpenSSL server will accept a DH certificate for client authentication
547 without the certificate verify message. This effectively allows a client to
548 authenticate without the use of a private key. This only affects servers
549 which trust a client certificate authority which issues certificates
550 containing DH keys: these are extremely rare and hardly ever encountered.
551 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
556 *) Ensure that the session ID context of an SSL is updated when its
557 SSL_CTX is updated via SSL_set_SSL_CTX.
559 The session ID context is typically set from the parent SSL_CTX,
560 and can vary with the CTX.
563 *) Fix various certificate fingerprint issues.
565 By using non-DER or invalid encodings outside the signed portion of a
566 certificate the fingerprint can be changed without breaking the signature.
567 Although no details of the signed portion of the certificate can be changed
568 this can cause problems with some applications: e.g. those using the
569 certificate fingerprint for blacklists.
571 1. Reject signatures with non zero unused bits.
573 If the BIT STRING containing the signature has non zero unused bits reject
574 the signature. All current signature algorithms require zero unused bits.
576 2. Check certificate algorithm consistency.
578 Check the AlgorithmIdentifier inside TBS matches the one in the
579 certificate signature. NB: this will result in signature failure
580 errors for some broken certificates.
582 Thanks to Konrad Kraszewski from Google for reporting this issue.
584 3. Check DSA/ECDSA signatures use DER.
586 Reencode DSA/ECDSA signatures and compare with the original received
587 signature. Return an error if there is a mismatch.
589 This will reject various cases including garbage after signature
590 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
591 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
592 (negative or with leading zeroes).
594 Further analysis was conducted and fixes were developed by Stephen Henson
595 of the OpenSSL core team.
600 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
601 results on some platforms, including x86_64. This bug occurs at random
602 with a very low probability, and is not known to be exploitable in any
603 way, though its exact impact is difficult to determine. Thanks to Pieter
604 Wuille (Blockstream) who reported this issue and also suggested an initial
605 fix. Further analysis was conducted by the OpenSSL development team and
606 Adam Langley of Google. The final fix was developed by Andy Polyakov of
607 the OpenSSL core team.
611 *) Do not resume sessions on the server if the negotiated protocol
612 version does not match the session's version. Resuming with a different
613 version, while not strictly forbidden by the RFC, is of questionable
614 sanity and breaks all known clients.
615 [David Benjamin, Emilia Käsper]
617 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
618 early CCS messages during renegotiation. (Note that because
619 renegotiation is encrypted, this early CCS was not exploitable.)
622 *) Tighten client-side session ticket handling during renegotiation:
623 ensure that the client only accepts a session ticket if the server sends
624 the extension anew in the ServerHello. Previously, a TLS client would
625 reuse the old extension state and thus accept a session ticket if one was
626 announced in the initial ServerHello.
628 Similarly, ensure that the client requires a session ticket if one
629 was advertised in the ServerHello. Previously, a TLS client would
630 ignore a missing NewSessionTicket message.
633 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
637 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
638 sends a carefully crafted handshake message, to cause OpenSSL to fail
639 to free up to 64k of memory causing a memory leak. This could be
640 exploited in a Denial Of Service attack. This issue affects OpenSSL
641 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
642 whether SRTP is used or configured. Implementations of OpenSSL that
643 have been compiled with OPENSSL_NO_SRTP defined are not affected.
645 The fix was developed by the OpenSSL team.
649 *) Session Ticket Memory Leak.
651 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
652 integrity of that ticket is first verified. In the event of a session
653 ticket integrity check failing, OpenSSL will fail to free memory
654 causing a memory leak. By sending a large number of invalid session
655 tickets an attacker could exploit this issue in a Denial Of Service
660 *) Build option no-ssl3 is incomplete.
662 When OpenSSL is configured with "no-ssl3" as a build option, servers
663 could accept and complete a SSL 3.0 handshake, and clients could be
664 configured to send them.
666 [Akamai and the OpenSSL team]
668 *) Add support for TLS_FALLBACK_SCSV.
669 Client applications doing fallback retries should call
670 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
672 [Adam Langley, Bodo Moeller]
674 *) Add additional DigestInfo checks.
676 Reencode DigestInto in DER and check against the original when
677 verifying RSA signature: this will reject any improperly encoded
678 DigestInfo structures.
680 Note: this is a precautionary measure and no attacks are currently known.
684 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
686 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
687 SRP code can be overrun an internal buffer. Add sanity check that
688 g, A, B < N to SRP code.
690 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
691 Group for discovering this issue.
695 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
696 TLS 1.0 instead of higher protocol versions when the ClientHello message
697 is badly fragmented. This allows a man-in-the-middle attacker to force a
698 downgrade to TLS 1.0 even if both the server and the client support a
699 higher protocol version, by modifying the client's TLS records.
701 Thanks to David Benjamin and Adam Langley (Google) for discovering and
702 researching this issue.
706 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
707 to a denial of service attack. A malicious server can crash the client
708 with a null pointer dereference (read) by specifying an anonymous (EC)DH
709 ciphersuite and sending carefully crafted handshake messages.
711 Thanks to Felix Gröbert (Google) for discovering and researching this
716 *) By sending carefully crafted DTLS packets an attacker could cause openssl
717 to leak memory. This can be exploited through a Denial of Service attack.
718 Thanks to Adam Langley for discovering and researching this issue.
722 *) An attacker can force openssl to consume large amounts of memory whilst
723 processing DTLS handshake messages. This can be exploited through a
724 Denial of Service attack.
725 Thanks to Adam Langley for discovering and researching this issue.
729 *) An attacker can force an error condition which causes openssl to crash
730 whilst processing DTLS packets due to memory being freed twice. This
731 can be exploited through a Denial of Service attack.
732 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
737 *) If a multithreaded client connects to a malicious server using a resumed
738 session and the server sends an ec point format extension it could write
739 up to 255 bytes to freed memory.
741 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
746 *) A malicious server can crash an OpenSSL client with a null pointer
747 dereference (read) by specifying an SRP ciphersuite even though it was not
748 properly negotiated with the client. This can be exploited through a
749 Denial of Service attack.
751 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
752 discovering and researching this issue.
756 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
757 X509_name_oneline, X509_name_print_ex et al. to leak some information
758 from the stack. Applications may be affected if they echo pretty printing
759 output to the attacker.
761 Thanks to Ivan Fratric (Google) for discovering this issue.
763 [Emilia Käsper, and Steve Henson]
765 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
766 for corner cases. (Certain input points at infinity could lead to
767 bogus results, with non-infinity inputs mapped to infinity too.)
770 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
772 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
773 handshake can force the use of weak keying material in OpenSSL
774 SSL/TLS clients and servers.
776 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
777 researching this issue. (CVE-2014-0224)
778 [KIKUCHI Masashi, Steve Henson]
780 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
781 OpenSSL DTLS client the code can be made to recurse eventually crashing
784 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
786 [Imre Rad, Steve Henson]
788 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
789 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
790 client or server. This is potentially exploitable to run arbitrary
791 code on a vulnerable client or server.
793 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
794 [Jüri Aedla, Steve Henson]
796 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
797 are subject to a denial of service attack.
799 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
800 this issue. (CVE-2014-3470)
801 [Felix Gröbert, Ivan Fratric, Steve Henson]
803 *) Harmonize version and its documentation. -f flag is used to display
805 [mancha <mancha1@zoho.com>]
807 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
809 [mancha <mancha1@zoho.com>]
811 *) Fix some double frees. These are not thought to be exploitable.
812 [mancha <mancha1@zoho.com>]
814 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
816 *) A missing bounds check in the handling of the TLS heartbeat extension
817 can be used to reveal up to 64k of memory to a connected client or
820 Thanks for Neel Mehta of Google Security for discovering this bug and to
821 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
822 preparing the fix (CVE-2014-0160)
823 [Adam Langley, Bodo Moeller]
825 *) Fix for the attack described in the paper "Recovering OpenSSL
826 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
827 by Yuval Yarom and Naomi Benger. Details can be obtained from:
828 http://eprint.iacr.org/2014/140
830 Thanks to Yuval Yarom and Naomi Benger for discovering this
831 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
832 [Yuval Yarom and Naomi Benger]
834 *) TLS pad extension: draft-agl-tls-padding-03
836 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
837 TLS client Hello record length value would otherwise be > 255 and
838 less that 512 pad with a dummy extension containing zeroes so it
839 is at least 512 bytes long.
841 [Adam Langley, Steve Henson]
843 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
845 *) Fix for TLS record tampering bug. A carefully crafted invalid
846 handshake could crash OpenSSL with a NULL pointer exception.
847 Thanks to Anton Johansson for reporting this issues.
850 *) Keep original DTLS digest and encryption contexts in retransmission
851 structures so we can use the previous session parameters if they need
852 to be resent. (CVE-2013-6450)
855 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
856 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
857 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
858 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
859 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
860 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
861 [Rob Stradling, Adam Langley]
863 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
865 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
866 supporting platforms or when small records were transferred.
867 [Andy Polyakov, Steve Henson]
869 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
871 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
873 This addresses the flaw in CBC record processing discovered by
874 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
875 at: http://www.isg.rhul.ac.uk/tls/
877 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
878 Security Group at Royal Holloway, University of London
879 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
880 Emilia Käsper for the initial patch.
882 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
884 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
885 ciphersuites which can be exploited in a denial of service attack.
886 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
887 and detecting this bug and to Wolfgang Ettlinger
888 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
892 *) Return an error when checking OCSP signatures when key is NULL.
893 This fixes a DoS attack. (CVE-2013-0166)
896 *) Make openssl verify return errors.
897 [Chris Palmer <palmer@google.com> and Ben Laurie]
899 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
900 the right response is stapled. Also change SSL_get_certificate()
901 so it returns the certificate actually sent.
902 See http://rt.openssl.org/Ticket/Display.html?id=2836.
903 [Rob Stradling <rob.stradling@comodo.com>]
905 *) Fix possible deadlock when decoding public keys.
908 *) Don't use TLS 1.0 record version number in initial client hello
912 Changes between 1.0.1b and 1.0.1c [10 May 2012]
914 *) Sanity check record length before skipping explicit IV in TLS
915 1.2, 1.1 and DTLS to fix DoS attack.
917 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
918 fuzzing as a service testing platform.
922 *) Initialise tkeylen properly when encrypting CMS messages.
923 Thanks to Solar Designer of Openwall for reporting this issue.
926 *) In FIPS mode don't try to use composite ciphers as they are not
930 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
932 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
933 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
934 mean any application compiled against OpenSSL 1.0.0 headers setting
935 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
936 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
937 0x10000000L Any application which was previously compiled against
938 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
939 will need to be recompiled as a result. Letting be results in
940 inability to disable specifically TLS 1.1 and in client context,
941 in unlike event, limit maximum offered version to TLS 1.0 [see below].
944 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
945 disable just protocol X, but all protocols above X *if* there are
946 protocols *below* X still enabled. In more practical terms it means
947 that if application wants to disable TLS1.0 in favor of TLS1.1 and
948 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
949 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
953 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
955 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
956 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
957 in CRYPTO_realloc_clean.
959 Thanks to Tavis Ormandy, Google Security Team, for discovering this
960 issue and to Adam Langley <agl@chromium.org> for fixing it.
962 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
964 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
967 *) Workarounds for some broken servers that "hang" if a client hello
968 record length exceeds 255 bytes.
970 1. Do not use record version number > TLS 1.0 in initial client
971 hello: some (but not all) hanging servers will now work.
972 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
973 the number of ciphers sent in the client hello. This should be
974 set to an even number, such as 50, for example by passing:
975 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
976 Most broken servers should now work.
977 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
978 TLS 1.2 client support entirely.
981 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
984 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
986 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
987 STRING form instead of a DigestInfo.
990 *) The format used for MDC2 RSA signatures is inconsistent between EVP
991 and the RSA_sign/RSA_verify functions. This was made more apparent when
992 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
993 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
994 the correct format in RSA_verify so both forms transparently work.
997 *) Some servers which support TLS 1.0 can choke if we initially indicate
998 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
999 encrypted premaster secret. As a workaround use the maximum pemitted
1000 client version in client hello, this should keep such servers happy
1001 and still work with previous versions of OpenSSL.
1004 *) Add support for TLS/DTLS heartbeats.
1005 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1007 *) Add support for SCTP.
1008 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1010 *) Improved PRNG seeding for VOS.
1011 [Paul Green <Paul.Green@stratus.com>]
1013 *) Extensive assembler packs updates, most notably:
1015 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1016 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1017 - x86_64: bit-sliced AES implementation;
1018 - ARM: NEON support, contemporary platforms optimizations;
1019 - s390x: z196 support;
1020 - *: GHASH and GF(2^m) multiplication implementations;
1024 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1025 (removal of unnecessary code)
1026 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1028 *) Add TLS key material exporter from RFC 5705.
1031 *) Add DTLS-SRTP negotiation from RFC 5764.
1034 *) Add Next Protocol Negotiation,
1035 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1036 disabled with a no-npn flag to config or Configure. Code donated
1038 [Adam Langley <agl@google.com> and Ben Laurie]
1040 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1041 NIST-P256, NIST-P521, with constant-time single point multiplication on
1042 typical inputs. Compiler support for the nonstandard type __uint128_t is
1043 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1044 Code made available under Apache License version 2.0.
1046 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1047 line to include this in your build of OpenSSL, and run "make depend" (or
1048 "make update"). This enables the following EC_METHODs:
1050 EC_GFp_nistp224_method()
1051 EC_GFp_nistp256_method()
1052 EC_GFp_nistp521_method()
1054 EC_GROUP_new_by_curve_name() will automatically use these (while
1055 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1057 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1059 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1060 all platforms. Move ssize_t definition from e_os.h to the public
1061 header file e_os2.h as it now appears in public header file cms.h
1064 *) New -sigopt option to the ca, req and x509 utilities. Additional
1065 signature parameters can be passed using this option and in
1069 *) Add RSA PSS signing function. This will generate and set the
1070 appropriate AlgorithmIdentifiers for PSS based on those in the
1071 corresponding EVP_MD_CTX structure. No application support yet.
1074 *) Support for companion algorithm specific ASN1 signing routines.
1075 New function ASN1_item_sign_ctx() signs a pre-initialised
1076 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1077 the appropriate parameters.
1080 *) Add new algorithm specific ASN1 verification initialisation function
1081 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1082 handling will be the same no matter what EVP_PKEY_METHOD is used.
1083 Add a PSS handler to support verification of PSS signatures: checked
1084 against a number of sample certificates.
1087 *) Add signature printing for PSS. Add PSS OIDs.
1088 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1090 *) Add algorithm specific signature printing. An individual ASN1 method
1091 can now print out signatures instead of the standard hex dump.
1093 More complex signatures (e.g. PSS) can print out more meaningful
1094 information. Include DSA version that prints out the signature
1098 *) Password based recipient info support for CMS library: implementing
1102 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1103 neatly separates the code into cipher and PBE sections and is required
1104 for some algorithms that split PBES2 into separate pieces (such as
1105 password based CMS).
1108 *) Session-handling fixes:
1109 - Fix handling of connections that are resuming with a session ID,
1110 but also support Session Tickets.
1111 - Fix a bug that suppressed issuing of a new ticket if the client
1112 presented a ticket with an expired session.
1113 - Try to set the ticket lifetime hint to something reasonable.
1114 - Make tickets shorter by excluding irrelevant information.
1115 - On the client side, don't ignore renewed tickets.
1116 [Adam Langley, Bodo Moeller (Google)]
1118 *) Fix PSK session representation.
1121 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1123 This work was sponsored by Intel.
1126 *) Add GCM support to TLS library. Some custom code is needed to split
1127 the IV between the fixed (from PRF) and explicit (from TLS record)
1128 portions. This adds all GCM ciphersuites supported by RFC5288 and
1129 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
1130 add a special AESGCM string for GCM only.
1133 *) Expand range of ctrls for AES GCM. Permit setting invocation
1134 field on decrypt and retrieval of invocation field only on encrypt.
1137 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1138 As required by RFC5289 these ciphersuites cannot be used if for
1139 versions of TLS earlier than 1.2.
1142 *) For FIPS capable OpenSSL interpret a NULL default public key method
1143 as unset and return the appopriate default but do *not* set the default.
1144 This means we can return the appopriate method in applications that
1145 swicth between FIPS and non-FIPS modes.
1148 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1149 ENGINE is used then we cannot handle that in the FIPS module so we
1150 keep original code iff non-FIPS operations are allowed.
1153 *) Add -attime option to openssl utilities.
1154 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1156 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1159 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1160 FIPS EC methods unconditionally for now.
1163 *) New build option no-ec2m to disable characteristic 2 code.
1166 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1167 all cases can be covered as some introduce binary incompatibilities.
1170 *) Redirect RSA operations to FIPS module including keygen,
1171 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1174 *) Add similar low level API blocking to ciphers.
1177 *) Low level digest APIs are not approved in FIPS mode: any attempt
1178 to use these will cause a fatal error. Applications that *really* want
1179 to use them can use the private_* version instead.
1182 *) Redirect cipher operations to FIPS module for FIPS builds.
1185 *) Redirect digest operations to FIPS module for FIPS builds.
1188 *) Update build system to add "fips" flag which will link in fipscanister.o
1189 for static and shared library builds embedding a signature if needed.
1192 *) Output TLS supported curves in preference order instead of numerical
1193 order. This is currently hardcoded for the highest order curves first.
1194 This should be configurable so applications can judge speed vs strength.
1197 *) Add TLS v1.2 server support for client authentication.
1200 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1204 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1205 FIPS modules versions.
1208 *) Add TLS v1.2 client side support for client authentication. Keep cache
1209 of handshake records longer as we don't know the hash algorithm to use
1210 until after the certificate request message is received.
1213 *) Initial TLS v1.2 client support. Add a default signature algorithms
1214 extension including all the algorithms we support. Parse new signature
1215 format in client key exchange. Relax some ECC signing restrictions for
1216 TLS v1.2 as indicated in RFC5246.
1219 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1220 to new signature format when needed using client digest preference.
1221 All server ciphersuites should now work correctly in TLS v1.2. No client
1222 support yet and no support for client certificates.
1225 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1226 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1227 ciphersuites. At present only RSA key exchange ciphersuites work with
1228 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1229 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1230 and version checking.
1233 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1234 with this defined it will not be affected by any changes to ssl internal
1235 structures. Add several utility functions to allow openssl application
1236 to work with OPENSSL_NO_SSL_INTERN defined.
1240 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1242 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1245 *) Permit abbreviated handshakes when renegotiating using the function
1246 SSL_renegotiate_abbreviated().
1247 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1249 *) Add call to ENGINE_register_all_complete() to
1250 ENGINE_load_builtin_engines(), so some implementations get used
1251 automatically instead of needing explicit application support.
1254 *) Add support for TLS key exporter as described in RFC5705.
1255 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1257 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1258 a few changes are required:
1260 Add SSL_OP_NO_TLSv1_1 flag.
1261 Add TLSv1_1 methods.
1262 Update version checking logic to handle version 1.1.
1263 Add explicit IV handling (ported from DTLS code).
1264 Add command line options to s_client/s_server.
1267 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1269 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1270 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1271 content decryption and always return the same error. Note: this attack
1272 needs on average 2^20 messages so it only affects automated senders. The
1273 old behaviour can be reenabled in the CMS code by setting the
1274 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1275 an MMA defence is not necessary.
1276 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1277 this issue. (CVE-2012-0884)
1280 *) Fix CVE-2011-4619: make sure we really are receiving a
1281 client hello before rejecting multiple SGC restarts. Thanks to
1282 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1285 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1287 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1288 Thanks to Antonio Martin, Enterprise Secure Access Research and
1289 Development, Cisco Systems, Inc. for discovering this bug and
1290 preparing a fix. (CVE-2012-0050)
1293 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1295 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1296 of the Vaudenay padding oracle attack on CBC mode encryption
1297 which enables an efficient plaintext recovery attack against
1298 the OpenSSL implementation of DTLS. Their attack exploits timing
1299 differences arising during decryption processing. A research
1300 paper describing this attack can be found at:
1301 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1302 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1303 Security Group at Royal Holloway, University of London
1304 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1305 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1306 for preparing the fix. (CVE-2011-4108)
1307 [Robin Seggelmann, Michael Tuexen]
1309 *) Clear bytes used for block padding of SSL 3.0 records.
1311 [Adam Langley (Google)]
1313 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1314 Kadianakis <desnacked@gmail.com> for discovering this issue and
1315 Adam Langley for preparing the fix. (CVE-2011-4619)
1316 [Adam Langley (Google)]
1318 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1319 [Andrey Kulikov <amdeich@gmail.com>]
1321 *) Prevent malformed RFC3779 data triggering an assertion failure.
1322 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1323 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1324 [Rob Austein <sra@hactrn.net>]
1326 *) Improved PRNG seeding for VOS.
1327 [Paul Green <Paul.Green@stratus.com>]
1329 *) Fix ssl_ciph.c set-up race.
1330 [Adam Langley (Google)]
1332 *) Fix spurious failures in ecdsatest.c.
1333 [Emilia Käsper (Google)]
1335 *) Fix the BIO_f_buffer() implementation (which was mixing different
1336 interpretations of the '..._len' fields).
1337 [Adam Langley (Google)]
1339 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1340 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1341 threads won't reuse the same blinding coefficients.
1343 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1344 lock to call BN_BLINDING_invert_ex, and avoids one use of
1345 BN_BLINDING_update for each BN_BLINDING structure (previously,
1346 the last update always remained unused).
1347 [Emilia Käsper (Google)]
1349 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1350 [Bob Buckholz (Google)]
1352 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1354 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1355 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1356 [Kaspar Brand <ossl@velox.ch>]
1358 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1359 for multi-threaded use of ECDH. (CVE-2011-3210)
1360 [Adam Langley (Google)]
1362 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1365 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1366 signature public key algorithm by using OID xref utilities instead.
1367 Before this you could only use some ECC ciphersuites with SHA1 only.
1370 *) Add protection against ECDSA timing attacks as mentioned in the paper
1371 by Billy Bob Brumley and Nicola Tuveri, see:
1373 http://eprint.iacr.org/2011/232.pdf
1375 [Billy Bob Brumley and Nicola Tuveri]
1377 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1379 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1380 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1382 *) Fix bug in string printing code: if *any* escaping is enabled we must
1383 escape the escape character (backslash) or the resulting string is
1387 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1389 *) Disable code workaround for ancient and obsolete Netscape browsers
1390 and servers: an attacker can use it in a ciphersuite downgrade attack.
1391 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1394 *) Fixed J-PAKE implementation error, originally discovered by
1395 Sebastien Martini, further info and confirmation from Stefan
1396 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1399 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1401 *) Fix extension code to avoid race conditions which can result in a buffer
1402 overrun vulnerability: resumed sessions must not be modified as they can
1403 be shared by multiple threads. CVE-2010-3864
1406 *) Fix WIN32 build system to correctly link an ENGINE directory into
1410 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1412 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1414 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1416 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1418 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1419 context. The operation can be customised via the ctrl mechanism in
1420 case ENGINEs want to include additional functionality.
1423 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1426 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1427 output hashes compatible with older versions of OpenSSL.
1428 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1430 *) Fix compression algorithm handling: if resuming a session use the
1431 compression algorithm of the resumed session instead of determining
1432 it from client hello again. Don't allow server to change algorithm.
1435 *) Add load_crls() function to apps tidying load_certs() too. Add option
1436 to verify utility to allow additional CRLs to be included.
1439 *) Update OCSP request code to permit adding custom headers to the request:
1440 some responders need this.
1443 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1445 [Julia Lawall <julia@diku.dk>]
1447 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1448 needlessly dereferenced structures, used obsolete functions and
1449 didn't handle all updated verify codes correctly.
1452 *) Disable MD2 in the default configuration.
1455 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1456 indicate the initial BIO being pushed or popped. This makes it possible
1457 to determine whether the BIO is the one explicitly called or as a result
1458 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1459 it handles reference counts correctly and doesn't zero out the I/O bio
1460 when it is not being explicitly popped. WARNING: applications which
1461 included workarounds for the old buggy behaviour will need to be modified
1462 or they could free up already freed BIOs.
1465 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1466 renaming to all platforms (within the 0.9.8 branch, this was
1467 done conditionally on Netware platforms to avoid a name clash).
1468 [Guenter <lists@gknw.net>]
1470 *) Add ECDHE and PSK support to DTLS.
1471 [Michael Tuexen <tuexen@fh-muenster.de>]
1473 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1477 *) Add "missing" function EVP_MD_flags() (without this the only way to
1478 retrieve a digest flags is by accessing the structure directly. Update
1479 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1480 or cipher is registered as in the "from" argument. Print out all
1481 registered digests in the dgst usage message instead of manually
1482 attempting to work them out.
1485 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1486 this allows the use of compression and extensions. Change default cipher
1487 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1488 by default unless an application cipher string requests it.
1491 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1492 key ids to find matching certificates and keys but some PKCS#12 files
1493 don't follow the (somewhat unwritten) rules and this strategy fails.
1494 Now just gather all certificates together and the first private key
1495 then look for the first certificate that matches the key.
1498 *) Support use of registered digest and cipher names for dgst and cipher
1499 commands instead of having to add each one as a special case. So now
1506 openssl dgst -sha256 foo
1508 and this works for ENGINE based algorithms too.
1512 *) Update Gost ENGINE to support parameter files.
1513 [Victor B. Wagner <vitus@cryptocom.ru>]
1515 *) Support GeneralizedTime in ca utility.
1516 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1518 *) Enhance the hash format used for certificate directory links. The new
1519 form uses the canonical encoding (meaning equivalent names will work
1520 even if they aren't identical) and uses SHA1 instead of MD5. This form
1521 is incompatible with the older format and as a result c_rehash should
1522 be used to rebuild symbolic links.
1525 *) Make PKCS#8 the default write format for private keys, replacing the
1526 traditional format. This form is standardised, more secure and doesn't
1527 include an implicit MD5 dependency.
1530 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1531 committed to OpenSSL should pass this lot as a minimum.
1534 *) Add session ticket override functionality for use by EAP-FAST.
1535 [Jouni Malinen <j@w1.fi>]
1537 *) Modify HMAC functions to return a value. Since these can be implemented
1538 in an ENGINE errors can occur.
1541 *) Type-checked OBJ_bsearch_ex.
1544 *) Type-checked OBJ_bsearch. Also some constification necessitated
1545 by type-checking. Still to come: TXT_DB, bsearch(?),
1546 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1550 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1551 seconds to a tm structure directly, instead of going through OS
1552 specific date routines. This avoids any issues with OS routines such
1553 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1554 and X509_time_adj_ex() to cover the extended range. The existing
1555 X509_time_adj() is still usable and will no longer have any date issues.
1558 *) Delta CRL support. New use deltas option which will attempt to locate
1559 and search any appropriate delta CRLs available.
1561 This work was sponsored by Google.
1564 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1565 code and add additional score elements. Validate alternate CRL paths
1566 as part of the CRL checking and indicate a new error "CRL path validation
1567 error" in this case. Applications wanting additional details can use
1568 the verify callback and check the new "parent" field. If this is not
1569 NULL CRL path validation is taking place. Existing applications wont
1570 see this because it requires extended CRL support which is off by
1573 This work was sponsored by Google.
1576 *) Support for freshest CRL extension.
1578 This work was sponsored by Google.
1581 *) Initial indirect CRL support. Currently only supported in the CRLs
1582 passed directly and not via lookup. Process certificate issuer
1583 CRL entry extension and lookup CRL entries by bother issuer name
1584 and serial number. Check and process CRL issuer entry in IDP extension.
1586 This work was sponsored by Google.
1589 *) Add support for distinct certificate and CRL paths. The CRL issuer
1590 certificate is validated separately in this case. Only enabled if
1591 an extended CRL support flag is set: this flag will enable additional
1592 CRL functionality in future.
1594 This work was sponsored by Google.
1597 *) Add support for policy mappings extension.
1599 This work was sponsored by Google.
1602 *) Fixes to pathlength constraint, self issued certificate handling,
1603 policy processing to align with RFC3280 and PKITS tests.
1605 This work was sponsored by Google.
1608 *) Support for name constraints certificate extension. DN, email, DNS
1609 and URI types are currently supported.
1611 This work was sponsored by Google.
1614 *) To cater for systems that provide a pointer-based thread ID rather
1615 than numeric, deprecate the current numeric thread ID mechanism and
1616 replace it with a structure and associated callback type. This
1617 mechanism allows a numeric "hash" to be extracted from a thread ID in
1618 either case, and on platforms where pointers are larger than 'long',
1619 mixing is done to help ensure the numeric 'hash' is usable even if it
1620 can't be guaranteed unique. The default mechanism is to use "&errno"
1621 as a pointer-based thread ID to distinguish between threads.
1623 Applications that want to provide their own thread IDs should now use
1624 CRYPTO_THREADID_set_callback() to register a callback that will call
1625 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1627 Note that ERR_remove_state() is now deprecated, because it is tied
1628 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1629 to free the current thread's error state should be replaced by
1630 ERR_remove_thread_state(NULL).
1632 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1633 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1634 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1635 application was previously providing a numeric thread callback that
1636 was inappropriate for distinguishing threads, then uniqueness might
1637 have been obtained with &errno that happened immediately in the
1638 intermediate development versions of OpenSSL; this is no longer the
1639 case, the numeric thread callback will now override the automatic use
1641 [Geoff Thorpe, with help from Bodo Moeller]
1643 *) Initial support for different CRL issuing certificates. This covers a
1644 simple case where the self issued certificates in the chain exist and
1645 the real CRL issuer is higher in the existing chain.
1647 This work was sponsored by Google.
1650 *) Removed effectively defunct crypto/store from the build.
1653 *) Revamp of STACK to provide stronger type-checking. Still to come:
1654 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1655 ASN1_STRING, CONF_VALUE.
1658 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1659 RAM on SSL connections. This option can save about 34k per idle SSL.
1662 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1663 STACK, TXT_DB, bsearch, qsort.
1666 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1667 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1668 support for data, signedData, compressedData, digestedData and
1669 encryptedData, envelopedData types included. Scripts to check against
1670 RFC4134 examples draft and interop and consistency checks of many
1671 content types and variants.
1674 *) Add options to enc utility to support use of zlib compression BIO.
1677 *) Extend mk1mf to support importing of options and assembly language
1678 files from Configure script, currently only included in VC-WIN32.
1679 The assembly language rules can now optionally generate the source
1680 files from the associated perl scripts.
1683 *) Implement remaining functionality needed to support GOST ciphersuites.
1684 Interop testing has been performed using CryptoPro implementations.
1685 [Victor B. Wagner <vitus@cryptocom.ru>]
1687 *) s390x assembler pack.
1690 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1694 *) Implement Opaque PRF Input TLS extension as specified in
1695 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1696 official specification yet and no extension type assignment by
1697 IANA exists, this extension (for now) will have to be explicitly
1698 enabled when building OpenSSL by providing the extension number
1699 to use. For example, specify an option
1701 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1703 to the "config" or "Configure" script to enable the extension,
1704 assuming extension number 0x9527 (which is a completely arbitrary
1705 and unofficial assignment based on the MD5 hash of the Internet
1706 Draft). Note that by doing so, you potentially lose
1707 interoperability with other TLS implementations since these might
1708 be using the same extension number for other purposes.
1710 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1711 opaque PRF input value to use in the handshake. This will create
1712 an interal copy of the length-'len' string at 'src', and will
1713 return non-zero for success.
1715 To get more control and flexibility, provide a callback function
1718 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1719 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1723 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1726 Callback function 'cb' will be called in handshakes, and is
1727 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1728 Argument 'arg' is for application purposes (the value as given to
1729 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1730 be provided to the callback function). The callback function
1731 has to return non-zero to report success: usually 1 to use opaque
1732 PRF input just if possible, or 2 to enforce use of the opaque PRF
1733 input. In the latter case, the library will abort the handshake
1734 if opaque PRF input is not successfully negotiated.
1736 Arguments 'peerinput' and 'len' given to the callback function
1737 will always be NULL and 0 in the case of a client. A server will
1738 see the client's opaque PRF input through these variables if
1739 available (NULL and 0 otherwise). Note that if the server
1740 provides an opaque PRF input, the length must be the same as the
1741 length of the client's opaque PRF input.
1743 Note that the callback function will only be called when creating
1744 a new session (session resumption can resume whatever was
1745 previously negotiated), and will not be called in SSL 2.0
1746 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1747 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1748 for applications that need to enforce opaque PRF input.
1752 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1755 [Victor B. Wagner <vitus@cryptocom.ru>]
1757 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1758 RFC4507bis. The encrypted ticket format is an encrypted encoded
1759 SSL_SESSION structure, that way new session features are automatically
1762 If a client application caches session in an SSL_SESSION structure
1763 support is transparent because tickets are now stored in the encoded
1766 The SSL_CTX structure automatically generates keys for ticket
1767 protection in servers so again support should be possible
1768 with no application modification.
1770 If a client or server wishes to disable RFC4507 support then the option
1771 SSL_OP_NO_TICKET can be set.
1773 Add a TLS extension debugging callback to allow the contents of any client
1774 or server extensions to be examined.
1776 This work was sponsored by Google.
1779 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1780 OpenSSL should now compile cleanly on gcc 4.2
1781 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1783 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1784 support including streaming MAC support: this is required for GOST
1785 ciphersuite support.
1786 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1788 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1789 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1790 to output in BER and PEM format.
1793 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1794 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1795 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1796 ENGINE support for HMAC keys which are unextractable. New -mac and
1797 -macopt options to dgst utility.
1800 *) New option -sigopt to dgst utility. Update dgst to use
1801 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1802 alternative signing paramaters such as X9.31 or PSS in the dgst
1806 *) Change ssl_cipher_apply_rule(), the internal function that does
1807 the work each time a ciphersuite string requests enabling
1808 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1809 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1810 the order of disabled ciphersuites such that those ciphersuites
1811 that most recently went from enabled to disabled not only stay
1812 in order with respect to each other, but also have higher priority
1813 than other disabled ciphersuites the next time ciphersuites are
1816 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1817 the same ciphersuites as with "HIGH" alone, but in a specific
1818 order where the PSK ciphersuites come first (since they are the
1819 most recently disabled ciphersuites when "HIGH" is parsed).
1821 Also, change ssl_create_cipher_list() (using this new
1822 funcionality) such that between otherwise identical
1823 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1827 *) Change ssl_create_cipher_list() so that it automatically
1828 arranges the ciphersuites in reasonable order before starting
1829 to process the rule string. Thus, the definition for "DEFAULT"
1830 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1831 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1832 This makes it much easier to arrive at a reasonable default order
1833 in applications for which anonymous ciphers are OK (meaning
1834 that you can't actually use DEFAULT).
1835 [Bodo Moeller; suggested by Victor Duchovni]
1837 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1838 processing) into multiple integers instead of setting
1839 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1840 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1841 (These masks as well as the individual bit definitions are hidden
1842 away into the non-exported interface ssl/ssl_locl.h, so this
1843 change to the definition of the SSL_CIPHER structure shouldn't
1844 affect applications.) This give us more bits for each of these
1845 categories, so there is no longer a need to coagulate AES128 and
1846 AES256 into a single algorithm bit, and to coagulate Camellia128
1847 and Camellia256 into a single algorithm bit, which has led to all
1850 Thus, among other things, the kludge introduced in 0.9.7m and
1851 0.9.8e for masking out AES256 independently of AES128 or masking
1852 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1854 With the change, we also introduce new ciphersuite aliases that
1855 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1859 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1860 Use the leftmost N bytes of the signature input if the input is
1861 larger than the prime q (with N being the size in bytes of q).
1864 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1865 it yet and it is largely untested.
1868 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1871 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1872 some compilers (gcc 4.2 and later) reject their use. Safestack is
1873 reimplemented. Update ASN1 to avoid use of legacy functions.
1876 *) Win32/64 targets are linked with Winsock2.
1879 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1880 to external functions. This can be used to increase CRL handling
1881 efficiency especially when CRLs are very large by (for example) storing
1882 the CRL revoked certificates in a database.
1885 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1886 new CRLs added to a directory can be used. New command line option
1887 -verify_return_error to s_client and s_server. This causes real errors
1888 to be returned by the verify callback instead of carrying on no matter
1889 what. This reflects the way a "real world" verify callback would behave.
1892 *) GOST engine, supporting several GOST algorithms and public key formats.
1893 Kindly donated by Cryptocom.
1896 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1897 partitioned by DP are handled but no indirect CRL or reason partitioning
1898 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1899 selected via a scoring technique which handles IDP and AKID in CRLs.
1902 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1903 will ultimately be used for all verify operations: this will remove the
1904 X509_STORE dependency on certificate verification and allow alternative
1905 lookup methods. X509_STORE based implementations of these two callbacks.
1908 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1909 Modify get_crl() to find a valid (unexpired) CRL if possible.
1912 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1913 this would be called X509_CRL_cmp() but that name is already used by
1914 a function that just compares CRL issuer names. Cache several CRL
1915 extensions in X509_CRL structure and cache CRLDP in X509.
1918 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1919 this maps equivalent X509_NAME structures into a consistent structure.
1920 Name comparison can then be performed rapidly using memcmp().
1923 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1927 *) Allow digests to supply their own micalg string for S/MIME type using
1928 the ctrl EVP_MD_CTRL_MICALG.
1931 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1932 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1933 ctrl. It can then customise the structure before and/or after signing
1937 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1938 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1939 to free up any added signature OIDs.
1942 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1943 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1944 digest and cipher tables. New options added to openssl utility:
1945 list-message-digest-algorithms and list-cipher-algorithms.
1948 *) Change the array representation of binary polynomials: the list
1949 of degrees of non-zero coefficients is now terminated with -1.
1950 Previously it was terminated with 0, which was also part of the
1951 value; thus, the array representation was not applicable to
1952 polynomials where t^0 has coefficient zero. This change makes
1953 the array representation useful in a more general context.
1956 *) Various modifications and fixes to SSL/TLS cipher string
1957 handling. For ECC, the code now distinguishes between fixed ECDH
1958 with RSA certificates on the one hand and with ECDSA certificates
1959 on the other hand, since these are separate ciphersuites. The
1960 unused code for Fortezza ciphersuites has been removed.
1962 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1963 (not "ECDHE"). For consistency with the code for DH
1964 certificates, use of ECDH certificates is now considered ECDH
1965 authentication, not RSA or ECDSA authentication (the latter is
1966 merely the CA's signing algorithm and not actively used in the
1969 The temporary ciphersuite alias "ECCdraft" is no longer
1970 available, and ECC ciphersuites are no longer excluded from "ALL"
1971 and "DEFAULT". The following aliases now exist for RFC 4492
1972 ciphersuites, most of these by analogy with the DH case:
1974 kECDHr - ECDH cert, signed with RSA
1975 kECDHe - ECDH cert, signed with ECDSA
1976 kECDH - ECDH cert (signed with either RSA or ECDSA)
1977 kEECDH - ephemeral ECDH
1978 ECDH - ECDH cert or ephemeral ECDH
1984 AECDH - anonymous ECDH
1985 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1989 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1990 Use correct micalg parameters depending on digest(s) in signed message.
1993 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1994 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1997 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1998 an engine to register a method. Add ENGINE lookups for methods and
1999 functional reference processing.
2002 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2003 EVP_{Sign,Verify}* which allow an application to customise the signature
2007 *) New -resign option to smime utility. This adds one or more signers
2008 to an existing PKCS#7 signedData structure. Also -md option to use an
2009 alternative message digest algorithm for signing.
2012 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2013 create PKCS7 structures containing multiple signers. Update smime
2014 application to support multiple signers.
2017 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2021 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2022 Reorganize PBE internals to lookup from a static table using NIDs,
2023 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2024 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2025 PRF which will be automatically used with PBES2.
2028 *) Replace the algorithm specific calls to generate keys in "req" with the
2032 *) Update PKCS#7 enveloped data routines to use new API. This is now
2033 supported by any public key method supporting the encrypt operation. A
2034 ctrl is added to allow the public key algorithm to examine or modify
2035 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2039 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2040 a default digest type to use. In most cases this will be SHA1 but some
2041 algorithms (such as GOST) need to specify an alternative digest. The
2042 return value indicates how strong the prefernce is 1 means optional and
2043 2 is mandatory (that is it is the only supported type). Modify
2044 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2045 use the default md. Update openssl utilities to use the default digest
2046 type for signing if it is not explicitly indicated.
2049 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2050 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2051 signing method from the key type. This effectively removes the link
2052 between digests and public key types.
2055 *) Add an OID cross reference table and utility functions. Its purpose is to
2056 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2057 rsaEncryption. This will allow some of the algorithm specific hackery
2058 needed to use the correct OID to be removed.
2061 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2062 structures for PKCS7_sign(). They are now set up by the relevant public
2066 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2069 *) Add support for key derivation (agreement) in the API, DH method and
2073 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2074 public and private key formats. As a side effect these add additional
2075 command line functionality not previously available: DSA signatures can be
2076 generated and verified using pkeyutl and DH key support and generation in
2081 [Oliver Tappe <zooey@hirschkaefer.de>]
2083 *) New make target "install_html_docs" installs HTML renditions of the
2085 [Oliver Tappe <zooey@hirschkaefer.de>]
2087 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
2088 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2089 support key and parameter generation and add initial key generation
2090 functionality for RSA.
2093 *) Add functions for main EVP_PKEY_method operations. The undocumented
2094 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2095 EVP_PKEY_{encrypt,decrypt}_old.
2098 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2099 key API, doesn't do much yet.
2102 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2103 public key algorithms. New option to openssl utility:
2104 "list-public-key-algorithms" to print out info.
2107 *) Implement the Supported Elliptic Curves Extension for
2108 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2111 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2112 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2115 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2116 utilities such as rsa, dsa, dsaparam etc except they process any key
2120 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2121 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2122 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2126 *) Initial support for pluggable public key ASN1.
2127 De-spaghettify the public key ASN1 handling. Move public and private
2128 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2129 algorithm specific handling to a single module within the relevant
2130 algorithm directory. Add functions to allow (near) opaque processing
2131 of public and private key structures.
2134 *) Implement the Supported Point Formats Extension for
2135 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2138 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2139 for the psk identity [hint] and the psk callback functions to the
2140 SSL_SESSION, SSL and SSL_CTX structure.
2143 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2147 SSL_CTX_use_psk_identity_hint
2148 SSL_get_psk_identity_hint
2149 SSL_get_psk_identity
2150 SSL_use_psk_identity_hint
2152 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2154 *) Add RFC 3161 compliant time stamp request creation, response generation
2155 and response verification functionality.
2156 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2158 *) Add initial support for TLS extensions, specifically for the server_name
2159 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2160 have new members for a host name. The SSL data structure has an
2161 additional member SSL_CTX *initial_ctx so that new sessions can be
2162 stored in that context to allow for session resumption, even after the
2163 SSL has been switched to a new SSL_CTX in reaction to a client's
2164 server_name extension.
2166 New functions (subject to change):
2168 SSL_get_servername()
2169 SSL_get_servername_type()
2172 New CTRL codes and macros (subject to change):
2174 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2175 - SSL_CTX_set_tlsext_servername_callback()
2176 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2177 - SSL_CTX_set_tlsext_servername_arg()
2178 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2180 openssl s_client has a new '-servername ...' option.
2182 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2183 '-key2 ...', '-servername_fatal' (subject to change). This allows
2184 testing the HostName extension for a specific single host name ('-cert'
2185 and '-key' remain fallbacks for handshakes without HostName
2186 negotiation). If the unrecogninzed_name alert has to be sent, this by
2187 default is a warning; it becomes fatal with the '-servername_fatal'
2190 [Peter Sylvester, Remy Allais, Christophe Renou]
2192 *) Whirlpool hash implementation is added.
2195 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2196 bn(64,32). Because of instruction set limitations it doesn't have
2197 any negative impact on performance. This was done mostly in order
2198 to make it possible to share assembler modules, such as bn_mul_mont
2199 implementations, between 32- and 64-bit builds without hassle.
2202 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2203 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2207 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2208 dedicated Montgomery multiplication procedure, is introduced.
2209 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2210 "64-bit" performance on certain 32-bit targets.
2213 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2214 in SSL structures. New SSL ctrl to set maximum send fragment size.
2215 Save memory by seeting the I/O buffer sizes dynamically instead of
2216 using the maximum available value.
2219 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2220 in addition to the text details.
2223 *) Very, very preliminary EXPERIMENTAL support for printing of general
2224 ASN1 structures. This currently produces rather ugly output and doesn't
2225 handle several customised structures at all.
2228 *) Integrated support for PVK file format and some related formats such
2229 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2230 these in the 'rsa' and 'dsa' utilities.
2233 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2236 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2237 place for the (very old) "NETSCAPE" format certificates which are now
2238 handled using new ASN1 code equivalents.
2241 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2242 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2243 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2246 *) Modify CRL distribution points extension code to print out previously
2247 unsupported fields. Enhance extension setting code to allow setting of
2251 *) Add print and set support for Issuing Distribution Point CRL extension.
2254 *) Change 'Configure' script to enable Camellia by default.
2257 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2259 *) When rejecting SSL/TLS records due to an incorrect version number, never
2260 update s->server with a new major version number. As of
2261 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2262 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2263 the previous behavior could result in a read attempt at NULL when
2264 receiving specific incorrect SSL/TLS records once record payload
2265 protection is active. (CVE-2010-0740)
2266 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2268 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2269 could be crashed if the relevant tables were not present (e.g. chrooted).
2270 [Tomas Hoger <thoger@redhat.com>]
2272 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2274 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2275 [Martin Olsson, Neel Mehta]
2277 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2278 accommodate for stack sorting, always a write lock!).
2281 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2282 excessive delays in the RAND_poll(): over a minute. As a workaround
2283 include a time check in the inner Heap32Next loop too.
2286 *) The code that handled flushing of data in SSL/TLS originally used the
2287 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2288 the problem outlined in PR#1949. The fix suggested there however can
2289 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2290 of Apache). So instead simplify the code to flush unconditionally.
2291 This should be fine since flushing with no data to flush is a no op.
2294 *) Handle TLS versions 2.0 and later properly and correctly use the
2295 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2296 off ancient servers have a habit of sticking around for a while...
2299 *) Modify compression code so it frees up structures without using the
2300 ex_data callbacks. This works around a problem where some applications
2301 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2302 restarting) then use compression (e.g. SSL with compression) later.
2303 This results in significant per-connection memory leaks and
2304 has caused some security issues including CVE-2008-1678 and
2308 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2309 change when encrypting or decrypting.
2312 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2313 connect and renegotiate with servers which do not support RI.
2314 Until RI is more widely deployed this option is enabled by default.
2317 *) Add "missing" ssl ctrls to clear options and mode.
2320 *) If client attempts to renegotiate and doesn't support RI respond with
2321 a no_renegotiation alert as required by RFC5746. Some renegotiating
2322 TLS clients will continue a connection gracefully when they receive
2323 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2324 waiting for a server hello which it will never receive. Now we treat a
2325 received no_renegotiation alert as a fatal error. This is because
2326 applications requesting a renegotiation might well expect it to succeed
2327 and would have no code in place to handle the server denying it so the
2328 only safe thing to do is to terminate the connection.
2331 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2332 peer supports secure renegotiation and 0 otherwise. Print out peer
2333 renegotiation support in s_client/s_server.
2336 *) Replace the highly broken and deprecated SPKAC certification method with
2337 the updated NID creation version. This should correctly handle UTF8.
2340 *) Implement RFC5746. Re-enable renegotiation but require the extension
2341 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2342 turns out to be a bad idea. It has been replaced by
2343 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2344 SSL_CTX_set_options(). This is really not recommended unless you
2345 know what you are doing.
2346 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2348 *) Fixes to stateless session resumption handling. Use initial_ctx when
2349 issuing and attempting to decrypt tickets in case it has changed during
2350 servername handling. Use a non-zero length session ID when attempting
2351 stateless session resumption: this makes it possible to determine if
2352 a resumption has occurred immediately after receiving server hello
2353 (several places in OpenSSL subtly assume this) instead of later in
2357 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2358 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2359 fixes for a few places where the return code is not checked
2361 [Julia Lawall <julia@diku.dk>]
2363 *) Add --strict-warnings option to Configure script to include devteam
2364 warnings in other configurations.
2367 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2368 makes it possible to install openssl libraries in locations which
2369 have names other than "lib", for example "/usr/lib64" which some
2371 [Steve Henson, based on patch from Jeremy Utley]
2373 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2374 X690 8.9.12 and can produce some misleading textual output of OIDs.
2375 [Steve Henson, reported by Dan Kaminsky]
2377 *) Delete MD2 from algorithm tables. This follows the recommendation in
2378 several standards that it is not used in new applications due to
2379 several cryptographic weaknesses. For binary compatibility reasons
2380 the MD2 API is still compiled in by default.
2383 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2387 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2388 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2390 [Guenter <lists@gknw.net>]
2392 *) Fix the server certificate chain building code to use X509_verify_cert(),
2393 it used to have an ad-hoc builder which was unable to cope with anything
2394 other than a simple chain.
2395 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2397 *) Don't check self signed certificate signatures in X509_verify_cert()
2398 by default (a flag can override this): it just wastes time without
2399 adding any security. As a useful side effect self signed root CAs
2400 with non-FIPS digests are now usable in FIPS mode.
2403 *) In dtls1_process_out_of_seq_message() the check if the current message
2404 is already buffered was missing. For every new message was memory
2405 allocated, allowing an attacker to perform an denial of service attack
2406 with sending out of seq handshake messages until there is no memory
2407 left. Additionally every future messege was buffered, even if the
2408 sequence number made no sense and would be part of another handshake.
2409 So only messages with sequence numbers less than 10 in advance will be
2410 buffered. (CVE-2009-1378)
2411 [Robin Seggelmann, discovered by Daniel Mentz]
2413 *) Records are buffered if they arrive with a future epoch to be
2414 processed after finishing the corresponding handshake. There is
2415 currently no limitation to this buffer allowing an attacker to perform
2416 a DOS attack with sending records with future epochs until there is no
2417 memory left. This patch adds the pqueue_size() function to detemine
2418 the size of a buffer and limits the record buffer to 100 entries.
2420 [Robin Seggelmann, discovered by Daniel Mentz]
2422 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2423 parent structure is freed. (CVE-2009-1379)
2426 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2427 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2430 [Ilya O. <vrghost@gmail.com>]
2432 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2434 *) Disable renegotiation completely - this fixes a severe security
2435 problem (CVE-2009-3555) at the cost of breaking all
2436 renegotiation. Renegotiation can be re-enabled by setting
2437 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2438 run-time. This is really not recommended unless you know what
2442 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2444 *) Don't set val to NULL when freeing up structures, it is freed up by
2445 underlying code. If sizeof(void *) > sizeof(long) this can result in
2446 zeroing past the valid field. (CVE-2009-0789)
2447 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2449 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2450 checked correctly. This would allow some invalid signed attributes to
2451 appear to verify correctly. (CVE-2009-0591)
2452 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2454 *) Reject UniversalString and BMPString types with invalid lengths. This
2455 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2456 a legal length. (CVE-2009-0590)
2459 *) Set S/MIME signing as the default purpose rather than setting it
2460 unconditionally. This allows applications to override it at the store
2464 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2465 to handle some structures.
2468 *) Improve efficiency of mem_gets: don't search whole buffer each time
2470 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2472 *) New -hex option for openssl rand.
2475 *) Print out UTF8String and NumericString when parsing ASN1.
2478 *) Support NumericString type for name components.
2481 *) Allow CC in the environment to override the automatically chosen
2482 compiler. Note that nothing is done to ensure flags work with the
2486 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2488 *) Properly check EVP_VerifyFinal() and similar return values
2490 [Ben Laurie, Bodo Moeller, Google Security Team]
2492 *) Enable TLS extensions by default.
2495 *) Allow the CHIL engine to be loaded, whether the application is
2496 multithreaded or not. (This does not release the developer from the
2497 obligation to set up the dynamic locking callbacks.)
2498 [Sander Temme <sander@temme.net>]
2500 *) Use correct exit code if there is an error in dgst command.
2501 [Steve Henson; problem pointed out by Roland Dirlewanger]
2503 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2504 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2507 *) Add experimental JPAKE support, including demo authentication in
2508 s_client and s_server.
2511 *) Set the comparison function in v3_addr_canonize().
2512 [Rob Austein <sra@hactrn.net>]
2514 *) Add support for XMPP STARTTLS in s_client.
2515 [Philip Paeps <philip@freebsd.org>]
2517 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2518 to ensure that even with this option, only ciphersuites in the
2519 server's preference list will be accepted. (Note that the option
2520 applies only when resuming a session, so the earlier behavior was
2521 just about the algorithm choice for symmetric cryptography.)
2524 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2526 *) Fix NULL pointer dereference if a DTLS server received
2527 ChangeCipherSpec as first record (CVE-2009-1386).
2530 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2531 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2534 *) The fix in 0.9.8c that supposedly got rid of unsafe
2535 double-checked locking was incomplete for RSA blinding,
2536 addressing just one layer of what turns out to have been
2537 doubly unsafe triple-checked locking.
2539 So now fix this for real by retiring the MONT_HELPER macro
2540 in crypto/rsa/rsa_eay.c.
2542 [Bodo Moeller; problem pointed out by Marius Schilder]
2544 *) Various precautionary measures:
2546 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2548 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2549 (NB: This would require knowledge of the secret session ticket key
2550 to exploit, in which case you'd be SOL either way.)
2552 - Change bn_nist.c so that it will properly handle input BIGNUMs
2553 outside the expected range.
2555 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2558 [Neel Mehta, Bodo Moeller]
2560 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2561 the load fails. Useful for distros.
2562 [Ben Laurie and the FreeBSD team]
2564 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2567 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2570 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2572 This work was sponsored by Logica.
2575 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2576 keystores. Support for SSL/TLS client authentication too.
2577 Not compiled unless enable-capieng specified to Configure.
2579 This work was sponsored by Logica.
2582 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2583 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2584 attribute creation routines such as certifcate requests and PKCS#12
2588 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2590 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2591 handshake which could lead to a cilent crash as found using the
2592 Codenomicon TLS test suite (CVE-2008-1672)
2593 [Steve Henson, Mark Cox]
2595 *) Fix double free in TLS server name extensions which could lead to
2596 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2599 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2601 Clear the error queue to ensure that error entries left from
2602 older function calls do not interfere with the correct operation.
2603 [Lutz Jaenicke, Erik de Castro Lopo]
2605 *) Remove root CA certificates of commercial CAs:
2607 The OpenSSL project does not recommend any specific CA and does not
2608 have any policy with respect to including or excluding any CA.
2609 Therefore it does not make any sense to ship an arbitrary selection
2610 of root CA certificates with the OpenSSL software.
2613 *) RSA OAEP patches to fix two separate invalid memory reads.
2614 The first one involves inputs when 'lzero' is greater than
2615 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2616 before the beginning of from). The second one involves inputs where
2617 the 'db' section contains nothing but zeroes (there is a one-byte
2618 invalid read after the end of 'db').
2619 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2621 *) Partial backport from 0.9.9-dev:
2623 Introduce bn_mul_mont (dedicated Montgomery multiplication
2624 procedure) as a candidate for BIGNUM assembler implementation.
2625 While 0.9.9-dev uses assembler for various architectures, only
2626 x86_64 is available by default here in the 0.9.8 branch, and
2627 32-bit x86 is available through a compile-time setting.
2629 To try the 32-bit x86 assembler implementation, use Configure
2630 option "enable-montasm" (which exists only for this backport).
2632 As "enable-montasm" for 32-bit x86 disclaims code stability
2633 anyway, in this constellation we activate additional code
2634 backported from 0.9.9-dev for further performance improvements,
2635 namely BN_from_montgomery_word. (To enable this otherwise,
2636 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2638 [Andy Polyakov (backport partially by Bodo Moeller)]
2640 *) Add TLS session ticket callback. This allows an application to set
2641 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2642 values. This is useful for key rollover for example where several key
2643 sets may exist with different names.
2646 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2647 This was broken until now in 0.9.8 releases, such that the only way
2648 a registered ENGINE could be used (assuming it initialises
2649 successfully on the host) was to explicitly set it as the default
2650 for the relevant algorithms. This is in contradiction with 0.9.7
2651 behaviour and the documentation. With this fix, when an ENGINE is
2652 registered into a given algorithm's table of implementations, the
2653 'uptodate' flag is reset so that auto-discovery will be used next
2654 time a new context for that algorithm attempts to select an
2656 [Ian Lister (tweaked by Geoff Thorpe)]
2658 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2659 implemention in the following ways:
2661 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2664 Lack of BER streaming support means one pass streaming processing is
2665 only supported if data is detached: setting the streaming flag is
2666 ignored for embedded content.
2668 CMS support is disabled by default and must be explicitly enabled
2669 with the enable-cms configuration option.
2672 *) Update the GMP engine glue to do direct copies between BIGNUM and
2673 mpz_t when openssl and GMP use the same limb size. Otherwise the
2674 existing "conversion via a text string export" trick is still used.
2675 [Paul Sheer <paulsheer@gmail.com>]
2677 *) Zlib compression BIO. This is a filter BIO which compressed and
2678 uncompresses any data passed through it.
2681 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2682 RFC3394 compatible AES key wrapping.
2685 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2686 sets string data without copying. X509_ALGOR_set0() and
2687 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2688 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2689 from an X509_ATTRIBUTE structure optionally checking it occurs only
2690 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2694 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2695 to get the expected BN_FLG_CONSTTIME behavior.
2696 [Bodo Moeller (Google)]
2700 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2701 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2702 - added some more tests to do_tests.pl
2703 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2704 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2705 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2706 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2707 - various changes to netware.pl to enable gcc-cross builds on Win32
2709 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2710 - various changes to fix missing prototype warnings
2711 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2712 - added AES, WHIRLPOOL and CPUID assembler code to build files
2713 - added missing AES assembler make rules to mk1mf.pl
2714 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2715 [Guenter Knauf <eflash@gmx.net>]
2717 *) Implement certificate status request TLS extension defined in RFC3546.
2718 A client can set the appropriate parameters and receive the encoded
2719 OCSP response via a callback. A server can query the supplied parameters
2720 and set the encoded OCSP response in the callback. Add simplified examples
2721 to s_client and s_server.
2724 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2726 *) Fix various bugs:
2727 + Binary incompatibility of ssl_ctx_st structure
2728 + DTLS interoperation with non-compliant servers
2729 + Don't call get_session_cb() without proposed session
2730 + Fix ia64 assembler code
2731 [Andy Polyakov, Steve Henson]
2733 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2735 *) DTLS Handshake overhaul. There were longstanding issues with
2736 OpenSSL DTLS implementation, which were making it impossible for
2737 RFC 4347 compliant client to communicate with OpenSSL server.
2738 Unfortunately just fixing these incompatibilities would "cut off"
2739 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2740 server keeps tolerating non RFC compliant syntax. The opposite is
2741 not true, 0.9.8f client can not communicate with earlier server.
2742 This update even addresses CVE-2007-4995.
2745 *) Changes to avoid need for function casts in OpenSSL: some compilers
2746 (gcc 4.2 and later) reject their use.
2747 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2750 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2751 RFC4507bis. The encrypted ticket format is an encrypted encoded
2752 SSL_SESSION structure, that way new session features are automatically
2755 If a client application caches session in an SSL_SESSION structure
2756 support is transparent because tickets are now stored in the encoded
2759 The SSL_CTX structure automatically generates keys for ticket
2760 protection in servers so again support should be possible
2761 with no application modification.
2763 If a client or server wishes to disable RFC4507 support then the option
2764 SSL_OP_NO_TICKET can be set.
2766 Add a TLS extension debugging callback to allow the contents of any client
2767 or server extensions to be examined.
2769 This work was sponsored by Google.
2772 *) Add initial support for TLS extensions, specifically for the server_name
2773 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2774 have new members for a host name. The SSL data structure has an
2775 additional member SSL_CTX *initial_ctx so that new sessions can be
2776 stored in that context to allow for session resumption, even after the
2777 SSL has been switched to a new SSL_CTX in reaction to a client's
2778 server_name extension.
2780 New functions (subject to change):
2782 SSL_get_servername()
2783 SSL_get_servername_type()
2786 New CTRL codes and macros (subject to change):
2788 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2789 - SSL_CTX_set_tlsext_servername_callback()
2790 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2791 - SSL_CTX_set_tlsext_servername_arg()
2792 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2794 openssl s_client has a new '-servername ...' option.
2796 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2797 '-key2 ...', '-servername_fatal' (subject to change). This allows
2798 testing the HostName extension for a specific single host name ('-cert'
2799 and '-key' remain fallbacks for handshakes without HostName
2800 negotiation). If the unrecogninzed_name alert has to be sent, this by
2801 default is a warning; it becomes fatal with the '-servername_fatal'
2804 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2806 *) Add AES and SSE2 assembly language support to VC++ build.
2809 *) Mitigate attack on final subtraction in Montgomery reduction.
2812 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2813 (which previously caused an internal error).
2816 *) Squeeze another 10% out of IGE mode when in != out.
2819 *) AES IGE mode speedup.
2820 [Dean Gaudet (Google)]
2822 *) Add the Korean symmetric 128-bit cipher SEED (see
2823 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2824 add SEED ciphersuites from RFC 4162:
2826 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2827 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2828 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2829 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2831 To minimize changes between patchlevels in the OpenSSL 0.9.8
2832 series, SEED remains excluded from compilation unless OpenSSL
2833 is configured with 'enable-seed'.
2834 [KISA, Bodo Moeller]
2836 *) Mitigate branch prediction attacks, which can be practical if a
2837 single processor is shared, allowing a spy process to extract
2838 information. For detailed background information, see
2839 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2840 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2841 and Necessary Software Countermeasures"). The core of the change
2842 are new versions BN_div_no_branch() and
2843 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2844 respectively, which are slower, but avoid the security-relevant
2845 conditional branches. These are automatically called by BN_div()
2846 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2847 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2848 remove a conditional branch.
2850 BN_FLG_CONSTTIME is the new name for the previous
2851 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2852 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2853 in the exponent causes BN_mod_exp_mont() to use the alternative
2854 implementation in BN_mod_exp_mont_consttime().) The old name
2855 remains as a deprecated alias.
2857 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2858 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2859 constant-time implementations for more than just exponentiation.
2860 Here too the old name is kept as a deprecated alias.
2862 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2863 the BN_BLINDING structure gets an independent copy of the
2864 modulus. This means that the previous "BIGNUM *m" argument to
2865 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2866 essentially becomes "const BIGNUM *m", although we can't actually
2867 change this in the header file before 0.9.9. It allows
2868 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2869 enable BN_FLG_CONSTTIME.
2871 [Matthew D Wood (Intel Corp)]
2873 *) In the SSL/TLS server implementation, be strict about session ID
2874 context matching (which matters if an application uses a single
2875 external cache for different purposes). Previously,
2876 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2877 set. This did ensure strict client verification, but meant that,
2878 with applications using a single external cache for quite
2879 different requirements, clients could circumvent ciphersuite
2880 restrictions for a given session ID context by starting a session
2881 in a different context.
2884 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2885 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2886 authentication-only ciphersuites.
2889 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2890 not complete and could lead to a possible single byte overflow
2891 (CVE-2007-5135) [Ben Laurie]
2893 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2895 *) Since AES128 and AES256 (and similarly Camellia128 and
2896 Camellia256) share a single mask bit in the logic of
2897 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2898 kludge to work properly if AES128 is available and AES256 isn't
2899 (or if Camellia128 is available and Camellia256 isn't).
2902 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2903 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2904 When a point or a seed is encoded in a BIT STRING, we need to
2905 prevent the removal of trailing zero bits to get the proper DER
2906 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2907 of a NamedBitList, for which trailing 0 bits need to be removed.)
2910 *) Have SSL/TLS server implementation tolerate "mismatched" record
2911 protocol version while receiving ClientHello even if the
2912 ClientHello is fragmented. (The server can't insist on the
2913 particular protocol version it has chosen before the ServerHello
2914 message has informed the client about his choice.)
2917 *) Add RFC 3779 support.
2918 [Rob Austein for ARIN, Ben Laurie]
2920 *) Load error codes if they are not already present instead of using a
2921 static variable. This allows them to be cleanly unloaded and reloaded.
2922 Improve header file function name parsing.
2925 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2926 or CAPABILITY handshake as required by RFCs.
2929 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2931 *) Introduce limits to prevent malicious keys being able to
2932 cause a denial of service. (CVE-2006-2940)
2933 [Steve Henson, Bodo Moeller]
2935 *) Fix ASN.1 parsing of certain invalid structures that can result
2936 in a denial of service. (CVE-2006-2937) [Steve Henson]
2938 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2939 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2941 *) Fix SSL client code which could crash if connecting to a
2942 malicious SSLv2 server. (CVE-2006-4343)
2943 [Tavis Ormandy and Will Drewry, Google Security Team]
2945 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2946 match only those. Before that, "AES256-SHA" would be interpreted
2947 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2948 the same strength classification in 0.9.7h) as we currently only
2949 have a single AES bit in the ciphersuite description bitmap.
2950 That change, however, also applied to ciphersuite strings such as
2951 "RC4-MD5" that intentionally matched multiple ciphersuites --
2952 namely, SSL 2.0 ciphersuites in addition to the more common ones
2953 from SSL 3.0/TLS 1.0.
2955 So we change the selection algorithm again: Naming an explicit
2956 ciphersuite selects this one ciphersuite, and any other similar
2957 ciphersuite (same bitmap) from *other* protocol versions.
2958 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2959 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2961 Since SSL 2.0 does not have any ciphersuites for which the
2962 128/256 bit distinction would be relevant, this works for now.
2963 The proper fix will be to use different bits for AES128 and
2964 AES256, which would have avoided the problems from the beginning;
2965 however, bits are scarce, so we can only do this in a new release
2966 (not just a patchlevel) when we can change the SSL_CIPHER
2967 definition to split the single 'unsigned long mask' bitmap into
2968 multiple values to extend the available space.
2972 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2974 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2975 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2977 *) Add AES IGE and biIGE modes.
2980 *) Change the Unix randomness entropy gathering to use poll() when
2981 possible instead of select(), since the latter has some
2982 undesirable limitations.
2983 [Darryl Miles via Richard Levitte and Bodo Moeller]
2985 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2986 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2987 cannot be implicitly activated as part of, e.g., the "AES" alias.
2988 However, please upgrade to OpenSSL 0.9.9[-dev] for
2989 non-experimental use of the ECC ciphersuites to get TLS extension
2990 support, which is required for curve and point format negotiation
2991 to avoid potential handshake problems.
2994 *) Disable rogue ciphersuites:
2996 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2997 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2998 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3000 The latter two were purportedly from
3001 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3004 Also deactivate the remaining ciphersuites from
3005 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3006 unofficial, and the ID has long expired.
3009 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3010 dual-core machines) and other potential thread-safety issues.
3013 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3014 versions), which is now available for royalty-free use
3015 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3016 Also, add Camellia TLS ciphersuites from RFC 4132.
3018 To minimize changes between patchlevels in the OpenSSL 0.9.8
3019 series, Camellia remains excluded from compilation unless OpenSSL
3020 is configured with 'enable-camellia'.
3023 *) Disable the padding bug check when compression is in use. The padding
3024 bug check assumes the first packet is of even length, this is not
3025 necessarily true if compresssion is enabled and can result in false
3026 positives causing handshake failure. The actual bug test is ancient
3027 code so it is hoped that implementations will either have fixed it by
3028 now or any which still have the bug do not support compression.
3031 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3033 *) When applying a cipher rule check to see if string match is an explicit
3034 cipher suite and only match that one cipher suite if it is.
3037 *) Link in manifests for VC++ if needed.
3038 [Austin Ziegler <halostatue@gmail.com>]
3040 *) Update support for ECC-based TLS ciphersuites according to
3041 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3042 TLS extensions, which are supported starting with the 0.9.9
3043 branch, not in the OpenSSL 0.9.8 branch).
3046 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3047 opaque EVP_CIPHER_CTX handling.
3050 *) Fixes and enhancements to zlib compression code. We now only use
3051 "zlib1.dll" and use the default __cdecl calling convention on Win32
3052 to conform with the standards mentioned here:
3053 http://www.zlib.net/DLL_FAQ.txt
3054 Static zlib linking now works on Windows and the new --with-zlib-include
3055 --with-zlib-lib options to Configure can be used to supply the location
3056 of the headers and library. Gracefully handle case where zlib library
3060 *) Several fixes and enhancements to the OID generation code. The old code
3061 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3062 handle numbers larger than ULONG_MAX, truncated printing and had a
3063 non standard OBJ_obj2txt() behaviour.
3066 *) Add support for building of engines under engine/ as shared libraries
3067 under VC++ build system.
3070 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3071 Hopefully, we will not see any false combination of paths any more.
3074 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3076 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3077 (part of SSL_OP_ALL). This option used to disable the
3078 countermeasure against man-in-the-middle protocol-version
3079 rollback in the SSL 2.0 server implementation, which is a bad
3080 idea. (CVE-2005-2969)
3082 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3083 for Information Security, National Institute of Advanced Industrial
3084 Science and Technology [AIST], Japan)]
3086 *) Add two function to clear and return the verify parameter flags.
3089 *) Keep cipherlists sorted in the source instead of sorting them at
3090 runtime, thus removing the need for a lock.
3093 *) Avoid some small subgroup attacks in Diffie-Hellman.
3094 [Nick Mathewson and Ben Laurie]
3096 *) Add functions for well-known primes.
3099 *) Extended Windows CE support.
3100 [Satoshi Nakamura and Andy Polyakov]
3102 *) Initialize SSL_METHOD structures at compile time instead of during
3103 runtime, thus removing the need for a lock.
3106 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3107 attempting to decrypt each encrypted key in turn. Add support to
3111 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3113 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3116 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3119 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3120 key into the same file any more.
3123 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3126 *) Add -utf8 command line and config file option to 'ca'.
3127 [Stefan <stf@udoma.org]
3129 *) Removed the macro des_crypt(), as it seems to conflict with some
3130 libraries. Use DES_crypt().
3133 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3134 involves renaming the source and generated shared-libs for
3135 both. The engines will accept the corrected or legacy ids
3136 ('ncipher' and '4758_cca' respectively) when binding. NB,
3137 this only applies when building 'shared'.
3138 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3140 *) Add attribute functions to EVP_PKEY structure. Modify
3141 PKCS12_create() to recognize a CSP name attribute and
3142 use it. Make -CSP option work again in pkcs12 utility.
3145 *) Add new functionality to the bn blinding code:
3146 - automatic re-creation of the BN_BLINDING parameters after
3147 a fixed number of uses (currently 32)
3148 - add new function for parameter creation
3149 - introduce flags to control the update behaviour of the
3150 BN_BLINDING parameters
3151 - hide BN_BLINDING structure
3152 Add a second BN_BLINDING slot to the RSA structure to improve
3153 performance when a single RSA object is shared among several
3157 *) Add support for DTLS.
3158 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3160 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3161 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3164 *) Remove buggy and incompletet DH cert support from
3165 ssl/ssl_rsa.c and ssl/s3_both.c
3168 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3169 the apps/openssl applications.
3172 *) Compile clean with "-Wall -Wmissing-prototypes
3173 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3174 DEBUG_SAFESTACK must also be set.
3177 *) Change ./Configure so that certain algorithms can be disabled by default.
3178 The new counterpiece to "no-xxx" is "enable-xxx".
3180 The patented RC5 and MDC2 algorithms will now be disabled unless
3181 "enable-rc5" and "enable-mdc2", respectively, are specified.
3183 (IDEA remains enabled despite being patented. This is because IDEA
3184 is frequently required for interoperability, and there is no license
3185 fee for non-commercial use. As before, "no-idea" can be used to
3186 avoid this algorithm.)
3190 *) Add processing of proxy certificates (see RFC 3820). This work was
3191 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3192 EGEE (Enabling Grids for E-science in Europe).
3195 *) RC4 performance overhaul on modern architectures/implementations, such
3196 as Intel P4, IA-64 and AMD64.
3199 *) New utility extract-section.pl. This can be used specify an alternative
3200 section number in a pod file instead of having to treat each file as
3201 a separate case in Makefile. This can be done by adding two lines to the
3204 =for comment openssl_section:XXX
3206 The blank line is mandatory.
3210 *) New arguments -certform, -keyform and -pass for s_client and s_server
3211 to allow alternative format key and certificate files and passphrase
3215 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3216 update associated structures and add various utility functions.
3218 Add new policy related verify parameters, include policy checking in
3219 standard verify code. Enhance 'smime' application with extra parameters
3220 to support policy checking and print out.
3223 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3224 Nehemiah processors. These extensions support AES encryption in hardware
3225 as well as RNG (though RNG support is currently disabled).
3226 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3228 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3231 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3232 [Andy Polyakov and a number of other people]
3234 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3235 implementation contributed by IBM.
3236 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3238 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3239 exponent rather than 'unsigned long'. There is a corresponding change to
3240 the new 'rsa_keygen' element of the RSA_METHOD structure.
3241 [Jelte Jansen, Geoff Thorpe]
3243 *) Functionality for creating the initial serial number file is now
3244 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3246 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3247 number file to 1, which is bound to cause problems. To avoid
3248 the problems while respecting compatibility between different 0.9.7
3249 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3250 CA.pl for serial number initialization. With the new release 0.9.8,
3251 we can fix the problem directly in the 'ca' utility.)
3254 *) Reduced header interdepencies by declaring more opaque objects in
3255 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3256 give fewer recursive includes, which could break lazy source code - so
3257 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3258 developers should define this symbol when building and using openssl to
3259 ensure they track the recommended behaviour, interfaces, [etc], but
3260 backwards-compatible behaviour prevails when this isn't defined.
3263 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3266 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3267 This will generate a random key of the appropriate length based on the
3268 cipher context. The EVP_CIPHER can provide its own random key generation
3269 routine to support keys of a specific form. This is used in the des and
3270 3des routines to generate a key of the correct parity. Update S/MIME
3271 code to use new functions and hence generate correct parity DES keys.
3272 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3273 valid (weak or incorrect parity).
3276 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3277 as looking them up. This is useful when the verified structure may contain
3278 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3279 present unless the new PKCS7_NO_CRL flag is asserted.
3282 *) Extend ASN1 oid configuration module. It now additionally accepts the
3285 shortName = some long name, 1.2.3.4
3288 *) Reimplemented the BN_CTX implementation. There is now no more static
3289 limitation on the number of variables it can handle nor the depth of the
3290 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3291 information can now expand as required, and rather than having a single
3292 static array of bignums, BN_CTX now uses a linked-list of such arrays
3293 allowing it to expand on demand whilst maintaining the usefulness of
3294 BN_CTX's "bundling".
3297 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3298 to allow all RSA operations to function using a single BN_CTX.
3301 *) Preliminary support for certificate policy evaluation and checking. This
3302 is initially intended to pass the tests outlined in "Conformance Testing
3303 of Relying Party Client Certificate Path Processing Logic" v1.07.
3306 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3307 remained unused and not that useful. A variety of other little bignum
3308 tweaks and fixes have also been made continuing on from the audit (see
3312 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3313 associated ASN1, EVP and SSL functions and old ASN1 macros.
3316 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3317 and this should never fail. So the return value from the use of
3318 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3319 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3322 *) BN_CTX_get() should return zero-valued bignums, providing the same
3323 initialised value as BN_new().
3324 [Geoff Thorpe, suggested by Ulf Möller]
3326 *) Support for inhibitAnyPolicy certificate extension.
3329 *) An audit of the BIGNUM code is underway, for which debugging code is
3330 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3331 is considered valid when processing BIGNUMs, and causes execution to
3332 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3333 further steps are taken to deliberately pollute unused data in BIGNUM
3334 structures to try and expose faulty code further on. For now, openssl will
3335 (in its default mode of operation) continue to tolerate the inconsistent
3336 forms that it has tolerated in the past, but authors and packagers should
3337 consider trying openssl and their own applications when compiled with
3338 these debugging symbols defined. It will help highlight potential bugs in
3339 their own code, and will improve the test coverage for OpenSSL itself. At
3340 some point, these tighter rules will become openssl's default to improve
3341 maintainability, though the assert()s and other overheads will remain only
3342 in debugging configurations. See bn.h for more details.
3343 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3345 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3346 that can only be obtained through BN_CTX_new() (which implicitly
3347 initialises it). The presence of this function only made it possible
3348 to overwrite an existing structure (and cause memory leaks).
3351 *) Because of the callback-based approach for implementing LHASH as a
3352 template type, lh_insert() adds opaque objects to hash-tables and
3353 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3354 to clean up those corresponding objects before destroying the hash table
3355 (and losing the object pointers). So some over-zealous constifications in
3356 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3357 objects as "const" and the lh_doall[_arg] callback wrappers are not
3358 prototyped to have "const" restrictions on the object pointers they are
3359 given (and so aren't required to cast them away any more).
3362 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3363 (speed) prefers to use its own implementation. The two implementations
3364 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3365 its object type properly exposed (MS_TM) instead of casting to/from "char
3366 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3367 aren't necessarily the greatest nomenclatures - but this is what was used
3368 internally to the implementation so I've used that for now.
3371 *) Ensure that deprecated functions do not get compiled when
3372 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3373 the self-tests were still using deprecated key-generation functions so
3374 these have been updated also.
3377 *) Reorganise PKCS#7 code to separate the digest location functionality
3378 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
3379 New function PKCS7_set_digest() to set the digest type for PKCS#7
3380 digestedData type. Add additional code to correctly generate the
3381 digestedData type and add support for this type in PKCS7 initialization
3385 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3386 structure of type "other".
3389 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3390 sure the loop does correctly stop and breaking ("division by zero")
3391 modulus operations are not performed. The (pre-generated) prime
3392 table crypto/bn/bn_prime.h was already correct, but it could not be
3393 re-generated on some platforms because of the "division by zero"
3394 situation in the script.
3395 [Ralf S. Engelschall]
3397 *) Update support for ECC-based TLS ciphersuites according to
3398 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3399 SHA-1 now is only used for "small" curves (where the
3400 representation of a field element takes up to 24 bytes); for
3401 larger curves, the field element resulting from ECDH is directly
3402 used as premaster secret.
3403 [Douglas Stebila (Sun Microsystems Laboratories)]
3405 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3406 curve secp160r1 to the tests.
3407 [Douglas Stebila (Sun Microsystems Laboratories)]
3409 *) Add the possibility to load symbols globally with DSO.
3410 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3412 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3413 control of the error stack.
3416 *) Add support for STORE in ENGINE.
3419 *) Add the STORE type. The intention is to provide a common interface
3420 to certificate and key stores, be they simple file-based stores, or
3421 HSM-type store, or LDAP stores, or...
3422 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3425 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3426 pass a list of arguments to any function as well as provide a way
3427 for a function to pass data back to the caller.
3430 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3431 works like BUF_strdup() but can be used to duplicate a portion of
3432 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3436 *) Add the function sk_find_ex() which works like sk_find(), but will
3437 return an index to an element even if an exact match couldn't be
3438 found. The index is guaranteed to point at the element where the
3439 searched-for key would be inserted to preserve sorting order.
3442 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3443 takes an extra flags argument for optional functionality. Currently,
3444 the following flags are defined:
3446 OBJ_BSEARCH_VALUE_ON_NOMATCH
3447 This one gets OBJ_bsearch_ex() to return a pointer to the first
3448 element where the comparing function returns a negative or zero
3451 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3452 This one gets OBJ_bsearch_ex() to return a pointer to the first
3453 element where the comparing function returns zero. This is useful
3454 if there are more than one element where the comparing function
3458 *) Make it possible to create self-signed certificates with 'openssl ca'
3459 in such a way that the self-signed certificate becomes part of the
3460 CA database and uses the same mechanisms for serial number generation
3461 as all other certificate signing. The new flag '-selfsign' enables
3462 this functionality. Adapt CA.sh and CA.pl.in.
3465 *) Add functionality to check the public key of a certificate request
3466 against a given private. This is useful to check that a certificate
3467 request can be signed by that key (self-signing).
3470 *) Make it possible to have multiple active certificates with the same
3471 subject in the CA index file. This is done only if the keyword
3472 'unique_subject' is set to 'no' in the main CA section (default
3473 if 'CA_default') of the configuration file. The value is saved
3474 with the database itself in a separate index attribute file,
3475 named like the index file with '.attr' appended to the name.
3478 *) Generate muti valued AVAs using '+' notation in config files for
3482 *) Support for nameConstraints certificate extension.
3485 *) Support for policyConstraints certificate extension.
3488 *) Support for policyMappings certificate extension.
3491 *) Make sure the default DSA_METHOD implementation only uses its
3492 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3493 and change its own handlers to be NULL so as to remove unnecessary
3494 indirection. This lets alternative implementations fallback to the
3495 default implementation more easily.
3498 *) Support for directoryName in GeneralName related extensions
3502 *) Make it possible to link applications using Makefile.shared.
3503 Make that possible even when linking against static libraries!
3506 *) Support for single pass processing for S/MIME signing. This now
3507 means that S/MIME signing can be done from a pipe, in addition
3508 cleartext signing (multipart/signed type) is effectively streaming
3509 and the signed data does not need to be all held in memory.
3511 This is done with a new flag PKCS7_STREAM. When this flag is set
3512 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3513 is done after the data is output (and digests calculated) in
3514 SMIME_write_PKCS7().
3517 *) Add full support for -rpath/-R, both in shared libraries and
3518 applications, at least on the platforms where it's known how
3522 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3523 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3524 will now compute a table of multiples of the generator that
3525 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3526 faster (notably in the case of a single point multiplication,
3527 scalar * generator).
3528 [Nils Larsch, Bodo Moeller]
3530 *) IPv6 support for certificate extensions. The various extensions
3531 which use the IP:a.b.c.d can now take IPv6 addresses using the
3532 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3536 *) Added an ENGINE that implements RSA by performing private key
3537 exponentiations with the GMP library. The conversions to and from
3538 GMP's mpz_t format aren't optimised nor are any montgomery forms
3539 cached, and on x86 it appears OpenSSL's own performance has caught up.
3540 However there are likely to be other architectures where GMP could
3541 provide a boost. This ENGINE is not built in by default, but it can be
3542 specified at Configure time and should be accompanied by the necessary
3543 linker additions, eg;
3544 ./config -DOPENSSL_USE_GMP -lgmp
3547 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3548 testing availability of engines with "-t" - the old behaviour is
3549 produced by increasing the feature's verbosity with "-tt".
3552 *) ECDSA routines: under certain error conditions uninitialized BN objects
3553 could be freed. Solution: make sure initialization is performed early
3554 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3558 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3559 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3560 software implementations. For DSA and DH, parameter generation can
3561 also be overriden by providing the appropriate method callbacks.
3564 *) Change the "progress" mechanism used in key-generation and
3565 primality testing to functions that take a new BN_GENCB pointer in
3566 place of callback/argument pairs. The new API functions have "_ex"
3567 postfixes and the older functions are reimplemented as wrappers for
3568 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3569 declarations of the old functions to help (graceful) attempts to
3570 migrate to the new functions. Also, the new key-generation API
3571 functions operate on a caller-supplied key-structure and return
3572 success/failure rather than returning a key or NULL - this is to
3573 help make "keygen" another member function of RSA_METHOD etc.
3575 Example for using the new callback interface:
3577 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3581 BN_GENCB_set(&my_cb, my_callback, my_arg);
3583 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3584 /* For the meaning of a, b in calls to my_callback(), see the
3585 * documentation of the function that calls the callback.
3586 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3587 * my_callback should return 1 if it wants BN_is_prime_ex()
3588 * to continue, or 0 to stop.
3593 *) Change the ZLIB compression method to be stateful, and make it
3594 available to TLS with the number defined in
3595 draft-ietf-tls-compression-04.txt.
3598 *) Add the ASN.1 structures and functions for CertificatePair, which
3599 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3601 CertificatePair ::= SEQUENCE {
3602 forward [0] Certificate OPTIONAL,
3603 reverse [1] Certificate OPTIONAL,
3604 -- at least one of the pair shall be present -- }
3606 Also implement the PEM functions to read and write certificate
3607 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3609 This needed to be defined, mostly for the sake of the LDAP
3610 attribute crossCertificatePair, but may prove useful elsewhere as
3614 *) Make it possible to inhibit symlinking of shared libraries in
3615 Makefile.shared, for Cygwin's sake.
3618 *) Extend the BIGNUM API by creating a function
3619 void BN_set_negative(BIGNUM *a, int neg);
3620 and a macro that behave like
3621 int BN_is_negative(const BIGNUM *a);
3623 to avoid the need to access 'a->neg' directly in applications.
3626 *) Implement fast modular reduction for pseudo-Mersenne primes
3627 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3628 EC_GROUP_new_curve_GFp() will now automatically use this
3630 [Nils Larsch <nla@trustcenter.de>]
3632 *) Add new lock type (CRYPTO_LOCK_BN).
3635 *) Change the ENGINE framework to automatically load engines
3636 dynamically from specific directories unless they could be
3637 found to already be built in or loaded. Move all the
3638 current engines except for the cryptodev one to a new
3640 The engines in engines/ are built as shared libraries if
3641 the "shared" options was given to ./Configure or ./config.
3642 Otherwise, they are inserted in libcrypto.a.
3643 /usr/local/ssl/engines is the default directory for dynamic
3644 engines, but that can be overriden at configure time through
3645 the usual use of --prefix and/or --openssldir, and at run
3646 time with the environment variable OPENSSL_ENGINES.
3647 [Geoff Thorpe and Richard Levitte]
3649 *) Add Makefile.shared, a helper makefile to build shared
3650 libraries. Addapt Makefile.org.
3653 *) Add version info to Win32 DLLs.
3654 [Peter 'Luna' Runestig" <peter@runestig.com>]
3656 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3657 can be added using this API to created arbitrary PKCS#12
3658 files while avoiding the low level API.
3660 New options to PKCS12_create(), key or cert can be NULL and
3661 will then be omitted from the output file. The encryption
3662 algorithm NIDs can be set to -1 for no encryption, the mac
3663 iteration count can be set to 0 to omit the mac.
3665 Enhance pkcs12 utility by making the -nokeys and -nocerts
3666 options work when creating a PKCS#12 file. New option -nomac
3667 to omit the mac, NONE can be set for an encryption algorithm.
3668 New code is modified to use the enhanced PKCS12_create()
3669 instead of the low level API.
3672 *) Extend ASN1 encoder to support indefinite length constructed
3673 encoding. This can output sequences tags and octet strings in
3674 this form. Modify pk7_asn1.c to support indefinite length
3675 encoding. This is experimental and needs additional code to
3676 be useful, such as an ASN1 bio and some enhanced streaming
3679 Extend template encode functionality so that tagging is passed
3680 down to the template encoder.
3683 *) Let 'openssl req' fail if an argument to '-newkey' is not
3684 recognized instead of using RSA as a default.
3687 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3688 As these are not official, they are not included in "ALL";
3689 the "ECCdraft" ciphersuite group alias can be used to select them.
3690 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3692 *) Add ECDH engine support.
3693 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3695 *) Add ECDH in new directory crypto/ecdh/.
3696 [Douglas Stebila (Sun Microsystems Laboratories)]
3698 *) Let BN_rand_range() abort with an error after 100 iterations
3699 without success (which indicates a broken PRNG).
3702 *) Change BN_mod_sqrt() so that it verifies that the input value
3703 is really the square of the return value. (Previously,
3704 BN_mod_sqrt would show GIGO behaviour.)
3707 *) Add named elliptic curves over binary fields from X9.62, SECG,
3708 and WAP/WTLS; add OIDs that were still missing.
3710 [Sheueling Chang Shantz and Douglas Stebila
3711 (Sun Microsystems Laboratories)]
3713 *) Extend the EC library for elliptic curves over binary fields
3714 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3717 EC_GF2m_simple_method
3721 EC_GROUP_new_curve_GF2m
3722 EC_GROUP_set_curve_GF2m
3723 EC_GROUP_get_curve_GF2m
3724 EC_POINT_set_affine_coordinates_GF2m
3725 EC_POINT_get_affine_coordinates_GF2m
3726 EC_POINT_set_compressed_coordinates_GF2m
3728 Point compression for binary fields is disabled by default for
3729 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3732 As binary polynomials are represented as BIGNUMs, various members
3733 of the EC_GROUP and EC_POINT data structures can be shared
3734 between the implementations for prime fields and binary fields;
3735 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3736 are essentially identical to their ..._GFp counterparts.
3737 (For simplicity, the '..._GFp' prefix has been dropped from
3738 various internal method names.)
3740 An internal 'field_div' method (similar to 'field_mul' and
3741 'field_sqr') has been added; this is used only for binary fields.
3743 [Sheueling Chang Shantz and Douglas Stebila
3744 (Sun Microsystems Laboratories)]
3746 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3747 through methods ('mul', 'precompute_mult').
3749 The generic implementations (now internally called 'ec_wNAF_mul'
3750 and 'ec_wNAF_precomputed_mult') remain the default if these
3751 methods are undefined.
3753 [Sheueling Chang Shantz and Douglas Stebila
3754 (Sun Microsystems Laboratories)]
3756 *) New function EC_GROUP_get_degree, which is defined through
3757 EC_METHOD. For curves over prime fields, this returns the bit
3758 length of the modulus.
3760 [Sheueling Chang Shantz and Douglas Stebila
3761 (Sun Microsystems Laboratories)]
3763 *) New functions EC_GROUP_dup, EC_POINT_dup.
3764 (These simply call ..._new and ..._copy).
3766 [Sheueling Chang Shantz and Douglas Stebila
3767 (Sun Microsystems Laboratories)]
3769 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3770 Polynomials are represented as BIGNUMs (where the sign bit is not
3771 used) in the following functions [macros]:
3774 BN_GF2m_sub [= BN_GF2m_add]
3775 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3776 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3777 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3779 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3780 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3781 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3782 BN_GF2m_cmp [= BN_ucmp]
3784 (Note that only the 'mod' functions are actually for fields GF(2^m).
3785 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3787 For some functions, an the irreducible polynomial defining a
3788 field can be given as an 'unsigned int[]' with strictly
3789 decreasing elements giving the indices of those bits that are set;
3790 i.e., p[] represents the polynomial
3791 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3793 p[0] > p[1] > ... > p[k] = 0.
3794 This applies to the following functions:
3799 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3800 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3802 BN_GF2m_mod_sqrt_arr
3803 BN_GF2m_mod_solve_quad_arr
3807 Conversion can be performed by the following functions:
3812 bntest.c has additional tests for binary polynomial arithmetic.
3814 Two implementations for BN_GF2m_mod_div() are available.
3815 The default algorithm simply uses BN_GF2m_mod_inv() and
3816 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3817 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3818 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3820 [Sheueling Chang Shantz and Douglas Stebila
3821 (Sun Microsystems Laboratories)]
3823 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3824 functionality is disabled at compile-time.
3825 [Douglas Stebila <douglas.stebila@sun.com>]
3827 *) Change default behaviour of 'openssl asn1parse' so that more
3828 information is visible when viewing, e.g., a certificate:
3830 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3831 mode the content of non-printable OCTET STRINGs is output in a
3832 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3833 avoid the appearance of a printable string.
3834 [Nils Larsch <nla@trustcenter.de>]
3836 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3838 EC_GROUP_set_asn1_flag()
3839 EC_GROUP_get_asn1_flag()
3840 EC_GROUP_set_point_conversion_form()
3841 EC_GROUP_get_point_conversion_form()
3842 These control ASN1 encoding details:
3843 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3844 has been set to OPENSSL_EC_NAMED_CURVE.
3845 - Points are encoded in uncompressed form by default; options for
3846 asn1_for are as for point2oct, namely
3847 POINT_CONVERSION_COMPRESSED
3848 POINT_CONVERSION_UNCOMPRESSED
3849 POINT_CONVERSION_HYBRID
3851 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3854 EC_GROUP_get0_seed()
3855 EC_GROUP_get_seed_len()
3856 This is used only for ASN1 purposes (so far).
3857 [Nils Larsch <nla@trustcenter.de>]
3859 *) Add 'field_type' member to EC_METHOD, which holds the NID
3860 of the appropriate field type OID. The new function
3861 EC_METHOD_get_field_type() returns this value.
3862 [Nils Larsch <nla@trustcenter.de>]
3867 EC_POINT_point2hex()
3868 EC_POINT_hex2point()
3869 providing useful interfaces to EC_POINT_point2oct() and
3870 EC_POINT_oct2point().
3871 [Nils Larsch <nla@trustcenter.de>]
3873 *) Change internals of the EC library so that the functions
3874 EC_GROUP_set_generator()
3875 EC_GROUP_get_generator()
3876 EC_GROUP_get_order()
3877 EC_GROUP_get_cofactor()
3878 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3879 to methods, which would lead to unnecessary code duplication when
3880 adding different types of curves.
3881 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3883 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3884 arithmetic, and such that modified wNAFs are generated
3885 (which avoid length expansion in many cases).
3888 *) Add a function EC_GROUP_check_discriminant() (defined via
3889 EC_METHOD) that verifies that the curve discriminant is non-zero.
3891 Add a function EC_GROUP_check() that makes some sanity tests
3892 on a EC_GROUP, its generator and order. This includes
3893 EC_GROUP_check_discriminant().
3894 [Nils Larsch <nla@trustcenter.de>]
3896 *) Add ECDSA in new directory crypto/ecdsa/.
3898 Add applications 'openssl ecparam' and 'openssl ecdsa'
3899 (these are based on 'openssl dsaparam' and 'openssl dsa').
3901 ECDSA support is also included in various other files across the
3902 library. Most notably,
3903 - 'openssl req' now has a '-newkey ecdsa:file' option;
3904 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3905 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3906 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3907 them suitable for ECDSA where domain parameters must be
3908 extracted before the specific public key;
3909 - ECDSA engine support has been added.
3910 [Nils Larsch <nla@trustcenter.de>]
3912 *) Include some named elliptic curves, and add OIDs from X9.62,
3913 SECG, and WAP/WTLS. Each curve can be obtained from the new
3915 EC_GROUP_new_by_curve_name(),
3916 and the list of available named curves can be obtained with
3917 EC_get_builtin_curves().
3918 Also add a 'curve_name' member to EC_GROUP objects, which can be
3920 EC_GROUP_set_curve_name()
3921 EC_GROUP_get_curve_name()
3922 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3924 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3925 was actually never needed) and in BN_mul(). The removal in BN_mul()
3926 required a small change in bn_mul_part_recursive() and the addition
3927 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3928 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3929 bn_sub_words() and bn_add_words() except they take arrays with
3933 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3935 *) Cleanse PEM buffers before freeing them since they may contain
3937 [Benjamin Bennett <ben@psc.edu>]
3939 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3940 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3941 authentication-only ciphersuites.
3944 *) Since AES128 and AES256 share a single mask bit in the logic of
3945 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3946 kludge to work properly if AES128 is available and AES256 isn't.
3949 *) Expand security boundary to match 1.1.1 module.
3952 *) Remove redundant features: hash file source, editing of test vectors
3953 modify fipsld to use external fips_premain.c signature.
3956 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3957 run algorithm test programs.
3960 *) Make algorithm test programs more tolerant of whitespace.
3963 *) Have SSL/TLS server implementation tolerate "mismatched" record
3964 protocol version while receiving ClientHello even if the
3965 ClientHello is fragmented. (The server can't insist on the
3966 particular protocol version it has chosen before the ServerHello
3967 message has informed the client about his choice.)
3970 *) Load error codes if they are not already present instead of using a
3971 static variable. This allows them to be cleanly unloaded and reloaded.
3974 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3976 *) Introduce limits to prevent malicious keys being able to
3977 cause a denial of service. (CVE-2006-2940)
3978 [Steve Henson, Bodo Moeller]
3980 *) Fix ASN.1 parsing of certain invalid structures that can result
3981 in a denial of service. (CVE-2006-2937) [Steve Henson]
3983 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3984 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3986 *) Fix SSL client code which could crash if connecting to a
3987 malicious SSLv2 server. (CVE-2006-4343)
3988 [Tavis Ormandy and Will Drewry, Google Security Team]
3990 *) Change ciphersuite string processing so that an explicit
3991 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3992 will no longer include "AES128-SHA"), and any other similar
3993 ciphersuite (same bitmap) from *other* protocol versions (so that
3994 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3995 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3996 changes from 0.9.8b and 0.9.8d.
3999 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4001 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4002 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4004 *) Change the Unix randomness entropy gathering to use poll() when
4005 possible instead of select(), since the latter has some
4006 undesirable limitations.
4007 [Darryl Miles via Richard Levitte and Bodo Moeller]
4009 *) Disable rogue ciphersuites:
4011 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4012 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4013 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4015 The latter two were purportedly from
4016 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4019 Also deactive the remaining ciphersuites from
4020 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4021 unofficial, and the ID has long expired.
4024 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4025 dual-core machines) and other potential thread-safety issues.
4028 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4030 *) Adapt fipsld and the build system to link against the validated FIPS
4031 module in FIPS mode.
4034 *) Fixes for VC++ 2005 build under Windows.
4037 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4038 from a Windows bash shell such as MSYS. It is autodetected from the
4039 "config" script when run from a VC++ environment. Modify standard VC++
4040 build to use fipscanister.o from the GNU make build.
4043 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4045 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4046 The value now differs depending on if you build for FIPS or not.
4047 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4048 safely run with a non-FIPSed libcrypto, as it may crash because of
4049 the difference induced by this change.
4052 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4054 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4055 (part of SSL_OP_ALL). This option used to disable the
4056 countermeasure against man-in-the-middle protocol-version
4057 rollback in the SSL 2.0 server implementation, which is a bad
4058 idea. (CVE-2005-2969)
4060 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4061 for Information Security, National Institute of Advanced Industrial
4062 Science and Technology [AIST], Japan)]
4064 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4065 mainly for FIPS compliance and not fully integrated at this stage.
4068 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4069 the exponentiation using a fixed-length exponent. (Otherwise,
4070 the information leaked through timing could expose the secret key
4071 after many signatures; cf. Bleichenbacher's attack on DSA with
4075 *) Make a new fixed-window mod_exp implementation the default for
4076 RSA, DSA, and DH private-key operations so that the sequence of
4077 squares and multiplies and the memory access pattern are
4078 independent of the particular secret key. This will mitigate
4079 cache-timing and potential related attacks.
4081 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4082 and this is automatically used by BN_mod_exp_mont() if the new flag
4083 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4084 will use this BN flag for private exponents unless the flag
4085 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4086 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4088 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4090 *) Change the client implementation for SSLv23_method() and
4091 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4092 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4093 (Previously, the SSL 2.0 backwards compatible Client Hello
4094 message format would be used even with SSL_OP_NO_SSLv2.)
4097 *) Add support for smime-type MIME parameter in S/MIME messages which some
4101 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4102 a threadsafe manner. Modify rsa code to use new function and add calls
4103 to dsa and dh code (which had race conditions before).
4106 *) Include the fixed error library code in the C error file definitions
4107 instead of fixing them up at runtime. This keeps the error code
4108 structures constant.
4111 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4113 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4116 *) Fixes for newer kerberos headers. NB: the casts are needed because
4117 the 'length' field is signed on one version and unsigned on another
4118 with no (?) obvious way to tell the difference, without these VC++
4119 complains. Also the "definition" of FAR (blank) is no longer included
4120 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4121 some needed definitions.
4124 *) Undo Cygwin change.
4127 *) Added support for proxy certificates according to RFC 3820.
4128 Because they may be a security thread to unaware applications,
4129 they must be explicitely allowed in run-time. See
4130 docs/HOWTO/proxy_certificates.txt for further information.
4133 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4135 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4136 server and client random values. Previously
4137 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4138 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4140 This change has negligible security impact because:
4142 1. Server and client random values still have 24 bytes of pseudo random
4145 2. Server and client random values are sent in the clear in the initial
4148 3. The master secret is derived using the premaster secret (48 bytes in
4149 size for static RSA ciphersuites) as well as client server and random
4152 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4155 [Stephen Henson, reported by UK NISCC]
4157 *) Use Windows randomness collection on Cygwin.
4160 *) Fix hang in EGD/PRNGD query when communication socket is closed
4161 prematurely by EGD/PRNGD.
4162 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4164 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4167 *) Back-port of selected performance improvements from development
4168 branch, as well as improved support for PowerPC platforms.
4171 *) Add lots of checks for memory allocation failure, error codes to indicate
4172 failure and freeing up memory if a failure occurs.
4173 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4175 *) Add new -passin argument to dgst.
4178 *) Perform some character comparisons of different types in X509_NAME_cmp:
4179 this is needed for some certificates that reencode DNs into UTF8Strings
4180 (in violation of RFC3280) and can't or wont issue name rollover
4184 *) Make an explicit check during certificate validation to see that
4185 the CA setting in each certificate on the chain is correct. As a
4186 side effect always do the following basic checks on extensions,
4187 not just when there's an associated purpose to the check:
4189 - if there is an unhandled critical extension (unless the user
4190 has chosen to ignore this fault)
4191 - if the path length has been exceeded (if one is set at all)
4192 - that certain extensions fit the associated purpose (if one has
4196 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4198 *) Avoid a race condition when CRLs are checked in a multi threaded
4199 environment. This would happen due to the reordering of the revoked
4200 entries during signature checking and serial number lookup. Now the
4201 encoding is cached and the serial number sort performed under a lock.
4202 Add new STACK function sk_is_sorted().
4205 *) Add Delta CRL to the extension code.
4208 *) Various fixes to s3_pkt.c so alerts are sent properly.
4209 [David Holmes <d.holmes@f5.com>]
4211 *) Reduce the chances of duplicate issuer name and serial numbers (in
4212 violation of RFC3280) using the OpenSSL certificate creation utilities.
4213 This is done by creating a random 64 bit value for the initial serial
4214 number when a serial number file is created or when a self signed
4215 certificate is created using 'openssl req -x509'. The initial serial
4216 number file is created using 'openssl x509 -next_serial' in CA.pl
4217 rather than being initialized to 1.
4220 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4222 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4223 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4224 [Joe Orton, Steve Henson]
4226 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4228 [Joe Orton, Steve Henson]
4230 *) Make it possible to have multiple active certificates with the same
4231 subject in the CA index file. This is done only if the keyword
4232 'unique_subject' is set to 'no' in the main CA section (default
4233 if 'CA_default') of the configuration file. The value is saved
4234 with the database itself in a separate index attribute file,
4235 named like the index file with '.attr' appended to the name.
4238 *) X509 verify fixes. Disable broken certificate workarounds when
4239 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4240 keyUsage extension present. Don't accept CRLs with unhandled critical
4241 extensions: since verify currently doesn't process CRL extensions this
4242 rejects a CRL with *any* critical extensions. Add new verify error codes
4246 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4247 A clarification of RFC2560 will require the use of OCTET STRINGs and
4248 some implementations cannot handle the current raw format. Since OpenSSL
4249 copies and compares OCSP nonces as opaque blobs without any attempt at
4250 parsing them this should not create any compatibility issues.
4253 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4254 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4255 this HMAC (and other) operations are several times slower than OpenSSL
4259 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4260 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4262 *) Use the correct content when signing type "other".
4265 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4267 *) Fix various bugs revealed by running the NISCC test suite:
4269 Stop out of bounds reads in the ASN1 code when presented with
4270 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4272 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4274 If verify callback ignores invalid public key errors don't try to check
4275 certificate signature with the NULL public key.
4279 *) New -ignore_err option in ocsp application to stop the server
4280 exiting on the first error in a request.
4283 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4284 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4288 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4289 extra data after the compression methods not only for TLS 1.0
4290 but also for SSL 3.0 (as required by the specification).
4291 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4293 *) Change X509_certificate_type() to mark the key as exported/exportable
4294 when it's 512 *bits* long, not 512 bytes.
4297 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4298 blocks during encryption.
4301 *) Various fixes to base64 BIO and non blocking I/O. On write
4302 flushes were not handled properly if the BIO retried. On read
4303 data was not being buffered properly and had various logic bugs.
4304 This also affects blocking I/O when the data being decoded is a
4308 *) Various S/MIME bugfixes and compatibility changes:
4309 output correct application/pkcs7 MIME type if
4310 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4311 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4312 of files as .eml work). Correctly handle very long lines in MIME
4316 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4318 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4319 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4320 a protocol version number mismatch like a decryption error
4321 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4324 *) Turn on RSA blinding by default in the default implementation
4325 to avoid a timing attack. Applications that don't want it can call
4326 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4327 They would be ill-advised to do so in most cases.
4328 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4330 *) Change RSA blinding code so that it works when the PRNG is not
4331 seeded (in this case, the secret RSA exponent is abused as
4332 an unpredictable seed -- if it is not unpredictable, there
4333 is no point in blinding anyway). Make RSA blinding thread-safe
4334 by remembering the creator's thread ID in rsa->blinding and
4335 having all other threads use local one-time blinding factors
4336 (this requires more computation than sharing rsa->blinding, but
4337 avoids excessive locking; and if an RSA object is not shared
4338 between threads, blinding will still be very fast).
4341 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4342 ENGINE as defaults for all supported algorithms irrespective of
4343 the 'flags' parameter. 'flags' is now honoured, so applications
4344 should make sure they are passing it correctly.
4347 *) Target "mingw" now allows native Windows code to be generated in
4348 the Cygwin environment as well as with the MinGW compiler.
4351 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4353 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4354 via timing by performing a MAC computation even if incorrrect
4355 block cipher padding has been found. This is a countermeasure
4356 against active attacks where the attacker has to distinguish
4357 between bad padding and a MAC verification error. (CVE-2003-0078)
4359 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4360 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4361 Martin Vuagnoux (EPFL, Ilion)]
4363 *) Make the no-err option work as intended. The intention with no-err
4364 is not to have the whole error stack handling routines removed from
4365 libcrypto, it's only intended to remove all the function name and
4366 reason texts, thereby removing some of the footprint that may not
4367 be interesting if those errors aren't displayed anyway.
4369 NOTE: it's still possible for any application or module to have it's
4370 own set of error texts inserted. The routines are there, just not
4371 used by default when no-err is given.
4374 *) Add support for FreeBSD on IA64.
4375 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4377 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4378 Kerberos function mit_des_cbc_cksum(). Before this change,
4379 the value returned by DES_cbc_cksum() was like the one from
4380 mit_des_cbc_cksum(), except the bytes were swapped.
4381 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4383 *) Allow an application to disable the automatic SSL chain building.
4384 Before this a rather primitive chain build was always performed in
4385 ssl3_output_cert_chain(): an application had no way to send the
4386 correct chain if the automatic operation produced an incorrect result.
4388 Now the chain builder is disabled if either:
4390 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4392 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4394 The reasoning behind this is that an application would not want the
4395 auto chain building to take place if extra chain certificates are
4396 present and it might also want a means of sending no additional
4397 certificates (for example the chain has two certificates and the
4401 *) Add the possibility to build without the ENGINE framework.
4402 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4404 *) Under Win32 gmtime() can return NULL: check return value in
4405 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4408 *) DSA routines: under certain error conditions uninitialized BN objects
4409 could be freed. Solution: make sure initialization is performed early
4410 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4411 Nils Larsch <nla@trustcenter.de> via PR#459)
4414 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4415 checked on reconnect on the client side, therefore session resumption
4416 could still fail with a "ssl session id is different" error. This
4417 behaviour is masked when SSL_OP_ALL is used due to
4418 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4419 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4420 followup to PR #377.
4423 *) IA-32 assembler support enhancements: unified ELF targets, support
4424 for SCO/Caldera platforms, fix for Cygwin shared build.
4427 *) Add support for FreeBSD on sparc64. As a consequence, support for
4428 FreeBSD on non-x86 processors is separate from x86 processors on
4429 the config script, much like the NetBSD support.
4430 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4432 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4434 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4437 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4438 code (06) was taken as the first octet of the session ID and the last
4439 octet was ignored consequently. As a result SSLv2 client side session
4440 caching could not have worked due to the session ID mismatch between
4442 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4446 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4447 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4451 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4452 seems that in spite of existing for more than a year, many application
4453 author have done nothing to provide the necessary callbacks, which
4454 means that this particular engine will not work properly anywhere.
4455 This is a very unfortunate situation which forces us, in the name
4456 of usability, to give the hw_ncipher.c a static lock, which is part
4458 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4459 appear in 0.9.8 or later. We EXPECT application authors to have
4460 dealt properly with this when 0.9.8 is released (unless we actually
4461 make such changes in the libcrypto locking code that changes will
4462 have to be made anyway).
4465 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4466 octets have been read, EOF or an error occurs. Without this change
4467 some truncated ASN1 structures will not produce an error.
4470 *) Disable Heimdal support, since it hasn't been fully implemented.
4471 Still give the possibility to force the use of Heimdal, but with
4472 warnings and a request that patches get sent to openssl-dev.
4475 *) Add the VC-CE target, introduce the WINCE sysname, and add
4476 INSTALL.WCE and appropriate conditionals to make it build.
4477 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4479 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4480 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4481 edit numbers of the version.
4482 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4484 *) Introduce safe string copy and catenation functions
4485 (BUF_strlcpy() and BUF_strlcat()).
4486 [Ben Laurie (CHATS) and Richard Levitte]
4488 *) Avoid using fixed-size buffers for one-line DNs.
4489 [Ben Laurie (CHATS)]
4491 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4492 resizing buffers containing secrets, and use where appropriate.
4493 [Ben Laurie (CHATS)]
4495 *) Avoid using fixed size buffers for configuration file location.
4496 [Ben Laurie (CHATS)]
4498 *) Avoid filename truncation for various CA files.
4499 [Ben Laurie (CHATS)]
4501 *) Use sizeof in preference to magic numbers.
4502 [Ben Laurie (CHATS)]
4504 *) Avoid filename truncation in cert requests.
4505 [Ben Laurie (CHATS)]
4507 *) Add assertions to check for (supposedly impossible) buffer
4509 [Ben Laurie (CHATS)]
4511 *) Don't cache truncated DNS entries in the local cache (this could
4512 potentially lead to a spoofing attack).
4513 [Ben Laurie (CHATS)]
4515 *) Fix various buffers to be large enough for hex/decimal
4516 representations in a platform independent manner.
4517 [Ben Laurie (CHATS)]
4519 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4520 resizing buffers containing secrets, and use where appropriate.
4521 [Ben Laurie (CHATS)]
4523 *) Add BIO_indent() to avoid much slightly worrying code to do
4525 [Ben Laurie (CHATS)]
4527 *) Convert sprintf()/BIO_puts() to BIO_printf().
4528 [Ben Laurie (CHATS)]
4530 *) buffer_gets() could terminate with the buffer only half
4532 [Ben Laurie (CHATS)]
4534 *) Add assertions to prevent user-supplied crypto functions from
4535 overflowing internal buffers by having large block sizes, etc.
4536 [Ben Laurie (CHATS)]
4538 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4540 [Ben Laurie (CHATS)]
4542 *) Eliminate unused copy of key in RC4.
4543 [Ben Laurie (CHATS)]
4545 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4546 [Ben Laurie (CHATS)]
4548 *) Fix off-by-one error in EGD path.
4549 [Ben Laurie (CHATS)]
4551 *) If RANDFILE path is too long, ignore instead of truncating.
4552 [Ben Laurie (CHATS)]
4554 *) Eliminate unused and incorrectly sized X.509 structure
4556 [Ben Laurie (CHATS)]
4558 *) Eliminate unused and dangerous function knumber().
4559 [Ben Laurie (CHATS)]
4561 *) Eliminate unused and dangerous structure, KSSL_ERR.
4562 [Ben Laurie (CHATS)]
4564 *) Protect against overlong session ID context length in an encoded
4565 session object. Since these are local, this does not appear to be
4567 [Ben Laurie (CHATS)]
4569 *) Change from security patch (see 0.9.6e below) that did not affect
4570 the 0.9.6 release series:
4572 Remote buffer overflow in SSL3 protocol - an attacker could
4573 supply an oversized master key in Kerberos-enabled versions.
4575 [Ben Laurie (CHATS)]
4577 *) Change the SSL kerb5 codes to match RFC 2712.
4580 *) Make -nameopt work fully for req and add -reqopt switch.
4581 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4583 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4584 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4586 *) Make sure tests can be performed even if the corresponding algorithms
4587 have been removed entirely. This was also the last step to make
4588 OpenSSL compilable with DJGPP under all reasonable conditions.
4589 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4591 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4592 to allow version independent disabling of normally unselected ciphers,
4593 which may be activated as a side-effect of selecting a single cipher.
4595 (E.g., cipher list string "RSA" enables ciphersuites that are left
4596 out of "ALL" because they do not provide symmetric encryption.
4597 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4598 [Lutz Jaenicke, Bodo Moeller]
4600 *) Add appropriate support for separate platform-dependent build
4601 directories. The recommended way to make a platform-dependent
4602 build directory is the following (tested on Linux), maybe with
4605 # Place yourself outside of the OpenSSL source tree. In
4606 # this example, the environment variable OPENSSL_SOURCE
4607 # is assumed to contain the absolute OpenSSL source directory.
4608 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4609 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4610 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4611 mkdir -p `dirname $F`
4612 ln -s $OPENSSL_SOURCE/$F $F
4615 To be absolutely sure not to disturb the source tree, a "make clean"
4616 is a good thing. If it isn't successfull, don't worry about it,
4617 it probably means the source directory is very clean.
4620 *) Make sure any ENGINE control commands make local copies of string
4621 pointers passed to them whenever necessary. Otherwise it is possible
4622 the caller may have overwritten (or deallocated) the original string
4623 data when a later ENGINE operation tries to use the stored values.
4624 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4626 *) Improve diagnostics in file reading and command-line digests.
4627 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4629 *) Add AES modes CFB and OFB to the object database. Correct an
4630 error in AES-CFB decryption.
4633 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4634 allows existing EVP_CIPHER_CTX structures to be reused after
4635 calling EVP_*Final(). This behaviour is used by encryption
4636 BIOs and some applications. This has the side effect that
4637 applications must explicitly clean up cipher contexts with
4638 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4641 *) Check the values of dna and dnb in bn_mul_recursive before calling
4642 bn_mul_comba (a non zero value means the a or b arrays do not contain
4643 n2 elements) and fallback to bn_mul_normal if either is not zero.
4646 *) Fix escaping of non-ASCII characters when using the -subj option
4647 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4650 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4651 form for "surname", serialNumber has no short form.
4652 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4653 therefore remove "mail" short name for "internet 7".
4654 The OID for unique identifiers in X509 certificates is
4655 x500UniqueIdentifier, not uniqueIdentifier.
4656 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4659 *) Add an "init" command to the ENGINE config module and auto initialize
4660 ENGINEs. Without any "init" command the ENGINE will be initialized
4661 after all ctrl commands have been executed on it. If init=1 the
4662 ENGINE is initailized at that point (ctrls before that point are run
4663 on the uninitialized ENGINE and after on the initialized one). If
4664 init=0 then the ENGINE will not be iniatialized at all.
4667 *) Fix the 'app_verify_callback' interface so that the user-defined
4668 argument is actually passed to the callback: In the
4669 SSL_CTX_set_cert_verify_callback() prototype, the callback
4670 declaration has been changed from
4673 int (*cb)(X509_STORE_CTX *,void *);
4674 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4675 i=s->ctx->app_verify_callback(&ctx)
4676 has been changed into
4677 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4679 To update applications using SSL_CTX_set_cert_verify_callback(),
4680 a dummy argument can be added to their callback functions.
4681 [D. K. Smetters <smetters@parc.xerox.com>]
4683 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4684 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4686 *) Add and OPENSSL_LOAD_CONF define which will cause
4687 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4688 This allows older applications to transparently support certain
4689 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4690 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4691 load the config file and OPENSSL_add_all_algorithms_conf() which will
4692 always load it have also been added.
4695 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4696 Adjust NIDs and EVP layer.
4697 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4699 *) Config modules support in openssl utility.
4701 Most commands now load modules from the config file,
4702 though in a few (such as version) this isn't done
4703 because it couldn't be used for anything.
4705 In the case of ca and req the config file used is
4706 the same as the utility itself: that is the -config
4707 command line option can be used to specify an
4711 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4712 use "openssl_conf" if filename is NULL use default openssl config file.
4715 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4716 config section name. Add a new flag to tolerate a missing config file
4717 and move code to CONF_modules_load_file().
4720 *) Support for crypto accelerator cards from Accelerated Encryption
4721 Processing, www.aep.ie. (Use engine 'aep')
4722 The support was copied from 0.9.6c [engine] and adapted/corrected
4723 to work with the new engine framework.
4724 [AEP Inc. and Richard Levitte]
4726 *) Support for SureWare crypto accelerator cards from Baltimore
4727 Technologies. (Use engine 'sureware')
4728 The support was copied from 0.9.6c [engine] and adapted
4729 to work with the new engine framework.
4732 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4733 make the newer ENGINE framework commands for the CHIL engine work.
4734 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4736 *) Make it possible to produce shared libraries on ReliantUNIX.
4737 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4739 *) Add the configuration target debug-linux-ppro.
4740 Make 'openssl rsa' use the general key loading routines
4741 implemented in apps.c, and make those routines able to
4742 handle the key format FORMAT_NETSCAPE and the variant
4744 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4746 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4747 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4749 *) Add -keyform to rsautl, and document -engine.
4750 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4752 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4753 BIO_R_NO_SUCH_FILE error code rather than the generic
4754 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4757 *) Add new functions
4759 ERR_peek_last_error_line
4760 ERR_peek_last_error_line_data.
4761 These are similar to
4764 ERR_peek_error_line_data,
4765 but report on the latest error recorded rather than the first one
4766 still in the error queue.
4767 [Ben Laurie, Bodo Moeller]
4769 *) default_algorithms option in ENGINE config module. This allows things
4771 default_algorithms = ALL
4772 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4775 *) Prelminary ENGINE config module.
4778 *) New experimental application configuration code.
4781 *) Change the AES code to follow the same name structure as all other
4782 symmetric ciphers, and behave the same way. Move everything to
4783 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4784 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4786 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4787 [Ben Laurie and Theo de Raadt]
4789 *) Add option to output public keys in req command.
4790 [Massimiliano Pala madwolf@openca.org]
4792 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4793 (up to about 10% better than before for P-192 and P-224).
4796 *) New functions/macros
4798 SSL_CTX_set_msg_callback(ctx, cb)
4799 SSL_CTX_set_msg_callback_arg(ctx, arg)
4800 SSL_set_msg_callback(ssl, cb)
4801 SSL_set_msg_callback_arg(ssl, arg)
4803 to request calling a callback function
4805 void cb(int write_p, int version, int content_type,
4806 const void *buf, size_t len, SSL *ssl, void *arg)
4808 whenever a protocol message has been completely received
4809 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4810 protocol version according to which the SSL library interprets
4811 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4812 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4813 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4814 specification (change_cipher_spec(20), alert(21), handshake(22)).
4815 'buf' and 'len' point to the actual message, 'ssl' to the
4816 SSL object, and 'arg' is the application-defined value set by
4817 SSL[_CTX]_set_msg_callback_arg().
4819 'openssl s_client' and 'openssl s_server' have new '-msg' options
4820 to enable a callback that displays all protocol messages.
4823 *) Change the shared library support so shared libraries are built as
4824 soon as the corresponding static library is finished, and thereby get
4825 openssl and the test programs linked against the shared library.
4826 This still only happens when the keyword "shard" has been given to
4827 the configuration scripts.
4829 NOTE: shared library support is still an experimental thing, and
4830 backward binary compatibility is still not guaranteed.
4831 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4833 *) Add support for Subject Information Access extension.
4834 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4836 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4837 additional bytes when new memory had to be allocated, not just
4838 when reusing an existing buffer.
4841 *) New command line and configuration option 'utf8' for the req command.
4842 This allows field values to be specified as UTF8 strings.
4845 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4846 runs for the former and machine-readable output for the latter.
4849 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4850 of the e-mail address in the DN (i.e., it will go into a certificate
4851 extension only). The new configuration file option 'email_in_dn = no'
4852 has the same effect.
4853 [Massimiliano Pala madwolf@openca.org]
4855 *) Change all functions with names starting with des_ to be starting
4856 with DES_ instead. Add wrappers that are compatible with libdes,
4857 but are named _ossl_old_des_*. Finally, add macros that map the
4858 des_* symbols to the corresponding _ossl_old_des_* if libdes
4859 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4860 desired, the des_* symbols will be mapped to DES_*, with one
4863 Since we provide two compatibility mappings, the user needs to
4864 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4865 compatibility is desired. The default (i.e., when that macro
4866 isn't defined) is OpenSSL 0.9.6c compatibility.
4868 There are also macros that enable and disable the support of old
4869 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4870 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4871 are defined, the default will apply: to support the old des routines.
4873 In either case, one must include openssl/des.h to get the correct
4874 definitions. Do not try to just include openssl/des_old.h, that
4877 NOTE: This is a major break of an old API into a new one. Software
4878 authors are encouraged to switch to the DES_ style functions. Some
4879 time in the future, des_old.h and the libdes compatibility functions
4880 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4881 default), and then completely removed.
4884 *) Test for certificates which contain unsupported critical extensions.
4885 If such a certificate is found during a verify operation it is
4886 rejected by default: this behaviour can be overridden by either
4887 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4888 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4889 X509_supported_extension() has also been added which returns 1 if a
4890 particular extension is supported.
4893 *) Modify the behaviour of EVP cipher functions in similar way to digests
4894 to retain compatibility with existing code.
4897 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4898 compatibility with existing code. In particular the 'ctx' parameter does
4899 not have to be to be initialized before the call to EVP_DigestInit() and
4900 it is tidied up after a call to EVP_DigestFinal(). New function
4901 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4902 EVP_MD_CTX_copy() changed to not require the destination to be
4903 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4904 requires the destination to be valid.
4906 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4907 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4910 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4911 so that complete 'Handshake' protocol structures are kept in memory
4912 instead of overwriting 'msg_type' and 'length' with 'body' data.
4915 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4916 [Massimo Santin via Richard Levitte]
4918 *) Major restructuring to the underlying ENGINE code. This includes
4919 reduction of linker bloat, separation of pure "ENGINE" manipulation
4920 (initialisation, etc) from functionality dealing with implementations
4921 of specific crypto iterfaces. This change also introduces integrated
4922 support for symmetric ciphers and digest implementations - so ENGINEs
4923 can now accelerate these by providing EVP_CIPHER and EVP_MD
4924 implementations of their own. This is detailed in crypto/engine/README
4925 as it couldn't be adequately described here. However, there are a few
4926 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4927 were changed in the original introduction of ENGINE code have now
4928 reverted back - the hooking from this code to ENGINE is now a good
4929 deal more passive and at run-time, operations deal directly with
4930 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4931 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4932 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4933 they were not being used by the framework as there is no concept of a
4934 BIGNUM_METHOD and they could not be generalised to the new
4935 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4936 ENGINE_cpy() has been removed as it cannot be consistently defined in
4940 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4943 *) Change mkdef.pl to sort symbols that get the same entry number,
4944 and make sure the automatically generated functions ERR_load_*
4945 become part of libeay.num as well.
4948 *) New function SSL_renegotiate_pending(). This returns true once
4949 renegotiation has been requested (either SSL_renegotiate() call
4950 or HelloRequest/ClientHello receveived from the peer) and becomes
4951 false once a handshake has been completed.
4952 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4953 sends a HelloRequest, but does not ensure that a handshake takes
4954 place. SSL_renegotiate_pending() is useful for checking if the
4955 client has followed the request.)
4958 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4959 By default, clients may request session resumption even during
4960 renegotiation (if session ID contexts permit); with this option,
4961 session resumption is possible only in the first handshake.
4963 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4964 more bits available for options that should not be part of
4965 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4968 *) Add some demos for certificate and certificate request creation.
4971 *) Make maximum certificate chain size accepted from the peer application
4972 settable (SSL*_get/set_max_cert_list()), as proposed by
4973 "Douglas E. Engert" <deengert@anl.gov>.
4976 *) Add support for shared libraries for Unixware-7
4977 (Boyd Lynn Gerber <gerberb@zenez.com>).
4980 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4981 be done prior to destruction. Use this to unload error strings from
4982 ENGINEs that load their own error strings. NB: This adds two new API
4983 functions to "get" and "set" this destroy handler in an ENGINE.
4986 *) Alter all existing ENGINE implementations (except "openssl" and
4987 "openbsd") to dynamically instantiate their own error strings. This
4988 makes them more flexible to be built both as statically-linked ENGINEs
4989 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4990 Also, add stub code to each that makes building them as self-contained
4991 shared-libraries easier (see README.ENGINE).
4994 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4995 implementations into applications that are completely implemented in
4996 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4997 commands that can be used to configure what shared-library to load and
4998 to control aspects of the way it is handled. Also, made an update to
4999 the README.ENGINE file that brings its information up-to-date and
5000 provides some information and instructions on the "dynamic" ENGINE
5001 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5004 *) Make it possible to unload ranges of ERR strings with a new
5005 "ERR_unload_strings" function.
5008 *) Add a copy() function to EVP_MD.
5011 *) Make EVP_MD routines take a context pointer instead of just the
5012 md_data void pointer.
5015 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5016 that the digest can only process a single chunk of data
5017 (typically because it is provided by a piece of
5018 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5019 is only going to provide a single chunk of data, and hence the
5020 framework needn't accumulate the data for oneshot drivers.
5023 *) As with "ERR", make it possible to replace the underlying "ex_data"
5024 functions. This change also alters the storage and management of global
5025 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5026 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5027 index counters. The API functions that use this state have been changed
5028 to take a "class_index" rather than pointers to the class's local STACK
5029 and counter, and there is now an API function to dynamically create new
5030 classes. This centralisation allows us to (a) plug a lot of the
5031 thread-safety problems that existed, and (b) makes it possible to clean
5032 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5033 such data would previously have always leaked in application code and
5034 workarounds were in place to make the memory debugging turn a blind eye
5035 to it. Application code that doesn't use this new function will still
5036 leak as before, but their memory debugging output will announce it now
5037 rather than letting it slide.
5039 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5040 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5041 has a return value to indicate success or failure.
5044 *) Make it possible to replace the underlying "ERR" functions such that the
5045 global state (2 LHASH tables and 2 locks) is only used by the "default"
5046 implementation. This change also adds two functions to "get" and "set"
5047 the implementation prior to it being automatically set the first time
5048 any other ERR function takes place. Ie. an application can call "get",
5049 pass the return value to a module it has just loaded, and that module
5050 can call its own "set" function using that value. This means the
5051 module's "ERR" operations will use (and modify) the error state in the
5052 application and not in its own statically linked copy of OpenSSL code.
5055 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5056 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5057 the operation, and provides a more encapsulated way for external code
5058 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5059 to use these functions rather than manually incrementing the counts.
5061 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5064 *) Add EVP test program.
5067 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5070 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5071 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5072 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5073 These allow a CRL to be built without having to access X509_CRL fields
5074 directly. Modify 'ca' application to use new functions.
5077 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5078 bug workarounds. Rollback attack detection is a security feature.
5079 The problem will only arise on OpenSSL servers when TLSv1 is not
5080 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5081 Software authors not wanting to support TLSv1 will have special reasons
5082 for their choice and can explicitly enable this option.
5083 [Bodo Moeller, Lutz Jaenicke]
5085 *) Rationalise EVP so it can be extended: don't include a union of
5086 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5087 (similar to those existing for EVP_CIPHER_CTX).
5092 EVP_MD_CTX_init(&md); /* new function call */
5093 EVP_DigestInit(&md, EVP_sha1());
5094 EVP_DigestUpdate(&md, in, len);
5095 EVP_DigestFinal(&md, out, NULL);
5096 EVP_MD_CTX_cleanup(&md); /* new function call */
5100 *) Make DES key schedule conform to the usual scheme, as well as
5101 correcting its structure. This means that calls to DES functions
5102 now have to pass a pointer to a des_key_schedule instead of a
5103 plain des_key_schedule (which was actually always a pointer
5106 des_key_schedule ks;
5108 des_set_key_checked(..., &ks);
5109 des_ncbc_encrypt(..., &ks, ...);
5111 (Note that a later change renames 'des_...' into 'DES_...'.)
5114 *) Initial reduction of linker bloat: the use of some functions, such as
5115 PEM causes large amounts of unused functions to be linked in due to
5116 poor organisation. For example pem_all.c contains every PEM function
5117 which has a knock on effect of linking in large amounts of (unused)
5118 ASN1 code. Grouping together similar functions and splitting unrelated
5119 functions prevents this.
5122 *) Cleanup of EVP macros.
5125 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5126 correct _ecb suffix.
5129 *) Add initial OCSP responder support to ocsp application. The
5130 revocation information is handled using the text based index
5131 use by the ca application. The responder can either handle
5132 requests generated internally, supplied in files (for example
5133 via a CGI script) or using an internal minimal server.
5136 *) Add configuration choices to get zlib compression for TLS.
5139 *) Changes to Kerberos SSL for RFC 2712 compliance:
5140 1. Implemented real KerberosWrapper, instead of just using
5141 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5142 2. Implemented optional authenticator field of KerberosWrapper.
5144 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5145 and authenticator structs; see crypto/krb5/.
5147 Generalized Kerberos calls to support multiple Kerberos libraries.
5148 [Vern Staats <staatsvr@asc.hpc.mil>,
5149 Jeffrey Altman <jaltman@columbia.edu>
5150 via Richard Levitte]
5152 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5153 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5154 values for each of the key sizes rather than having just
5155 parameters (and 'speed' generating keys each time).
5158 *) Speed up EVP routines.
5161 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5162 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5163 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5164 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5166 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5167 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5168 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5171 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5173 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5176 *) Added the OS2-EMX target.
5177 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5179 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5180 to support NCONF routines in extension code. New function CONF_set_nconf()
5181 to allow functions which take an NCONF to also handle the old LHASH
5182 structure: this means that the old CONF compatible routines can be
5183 retained (in particular wrt extensions) without having to duplicate the
5184 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5187 *) Enhance the general user interface with mechanisms for inner control
5188 and with possibilities to have yes/no kind of prompts.
5191 *) Change all calls to low level digest routines in the library and
5192 applications to use EVP. Add missing calls to HMAC_cleanup() and
5193 don't assume HMAC_CTX can be copied using memcpy().
5194 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5196 *) Add the possibility to control engines through control names but with
5197 arbitrary arguments instead of just a string.
5198 Change the key loaders to take a UI_METHOD instead of a callback
5199 function pointer. NOTE: this breaks binary compatibility with earlier
5200 versions of OpenSSL [engine].
5201 Adapt the nCipher code for these new conditions and add a card insertion
5205 *) Enhance the general user interface with mechanisms to better support
5206 dialog box interfaces, application-defined prompts, the possibility
5207 to use defaults (for example default passwords from somewhere else)
5208 and interrupts/cancellations.
5211 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5212 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5215 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5216 tidy up some unnecessarily weird code in 'sk_new()').
5217 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5219 *) Change the key loading routines for ENGINEs to use the same kind
5220 callback (pem_password_cb) as all other routines that need this
5224 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5225 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5226 than this minimum value is recommended.
5229 *) New random seeder for OpenVMS, using the system process statistics
5230 that are easily reachable.
5233 *) Windows apparently can't transparently handle global
5234 variables defined in DLLs. Initialisations such as:
5236 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5238 wont compile. This is used by the any applications that need to
5239 declare their own ASN1 modules. This was fixed by adding the option
5240 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5241 needed for static libraries under Win32.
5244 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5245 setting of purpose and trust fields. New X509_STORE trust and
5246 purpose functions and tidy up setting in other SSL functions.
5249 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5250 structure. These are inherited by X509_STORE_CTX when it is
5251 initialised. This allows various defaults to be set in the
5252 X509_STORE structure (such as flags for CRL checking and custom
5253 purpose or trust settings) for functions which only use X509_STORE_CTX
5254 internally such as S/MIME.
5256 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5257 trust settings if they are not set in X509_STORE. This allows X509_STORE
5258 purposes and trust (in S/MIME for example) to override any set by default.
5260 Add command line options for CRL checking to smime, s_client and s_server
5264 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5265 are set then the CRL is looked up in the X509_STORE structure and
5266 its validity and signature checked, then if the certificate is found
5267 in the CRL the verify fails with a revoked error.
5269 Various new CRL related callbacks added to X509_STORE_CTX structure.
5271 Command line options added to 'verify' application to support this.
5273 This needs some additional work, such as being able to handle multiple
5274 CRLs with different times, extension based lookup (rather than just
5275 by subject name) and ultimately more complete V2 CRL extension
5279 *) Add a general user interface API (crypto/ui/). This is designed
5280 to replace things like des_read_password and friends (backward
5281 compatibility functions using this new API are provided).
5282 The purpose is to remove prompting functions from the DES code
5283 section as well as provide for prompting through dialog boxes in
5284 a window system and the like.
5287 *) Add "ex_data" support to ENGINE so implementations can add state at a
5288 per-structure level rather than having to store it globally.
5291 *) Make it possible for ENGINE structures to be copied when retrieved by
5292 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5293 This causes the "original" ENGINE structure to act like a template,
5294 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5295 operational state can be localised to each ENGINE structure, despite the
5296 fact they all share the same "methods". New ENGINE structures returned in
5297 this case have no functional references and the return value is the single
5298 structural reference. This matches the single structural reference returned
5299 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5303 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5304 needs to match any other type at all we need to manually clear the
5308 *) Changes to the "openssl engine" utility to include;
5309 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5310 about an ENGINE's available control commands.
5311 - executing control commands from command line arguments using the
5312 '-pre' and '-post' switches. '-post' is only used if '-t' is
5313 specified and the ENGINE is successfully initialised. The syntax for
5314 the individual commands are colon-separated, for example;
5315 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5318 *) New dynamic control command support for ENGINEs. ENGINEs can now
5319 declare their own commands (numbers), names (strings), descriptions,
5320 and input types for run-time discovery by calling applications. A
5321 subset of these commands are implicitly classed as "executable"
5322 depending on their input type, and only these can be invoked through
5323 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5324 can be based on user input, config files, etc). The distinction is
5325 that "executable" commands cannot return anything other than a boolean
5326 result and can only support numeric or string input, whereas some
5327 discoverable commands may only be for direct use through
5328 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5329 pointers, or other custom uses. The "executable" commands are to
5330 support parameterisations of ENGINE behaviour that can be
5331 unambiguously defined by ENGINEs and used consistently across any
5332 OpenSSL-based application. Commands have been added to all the
5333 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5334 control over shared-library paths without source code alterations.
5337 *) Changed all ENGINE implementations to dynamically allocate their
5338 ENGINEs rather than declaring them statically. Apart from this being
5339 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5340 this also allows the implementations to compile without using the
5341 internal engine_int.h header.
5344 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5345 'const' value. Any code that should be able to modify a RAND_METHOD
5346 should already have non-const pointers to it (ie. they should only
5347 modify their own ones).
5350 *) Made a variety of little tweaks to the ENGINE code.
5351 - "atalla" and "ubsec" string definitions were moved from header files
5352 to C code. "nuron" string definitions were placed in variables
5353 rather than hard-coded - allowing parameterisation of these values
5354 later on via ctrl() commands.
5355 - Removed unused "#if 0"'d code.
5356 - Fixed engine list iteration code so it uses ENGINE_free() to release
5357 structural references.
5358 - Constified the RAND_METHOD element of ENGINE structures.
5359 - Constified various get/set functions as appropriate and added
5360 missing functions (including a catch-all ENGINE_cpy that duplicates
5361 all ENGINE values onto a new ENGINE except reference counts/state).
5362 - Removed NULL parameter checks in get/set functions. Setting a method
5363 or function to NULL is a way of cancelling out a previously set
5364 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5365 and doesn't justify the extra error symbols and code.
5366 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5367 flags from engine_int.h to engine.h.
5368 - Changed prototypes for ENGINE handler functions (init(), finish(),
5369 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5372 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5373 to the algorithm using long division. The binary algorithm can be
5374 used only if the modulus is odd. On 32-bit systems, it is faster
5375 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5376 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5377 up to 450 bits. In 64-bit environments, the binary algorithm
5378 appears to be advantageous for much longer moduli; here we use it
5379 for moduli up to 2048 bits.
5382 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5383 could not support the combine flag in choice fields.
5386 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5387 extensions from a certificate request to the certificate.
5390 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5391 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5392 file: this allows the display of the certificate about to be
5393 signed to be customised, to allow certain fields to be included
5394 or excluded and extension details. The old system didn't display
5395 multicharacter strings properly, omitted fields not in the policy
5396 and couldn't display additional details such as extensions.
5399 *) Function EC_POINTs_mul for multiple scalar multiplication
5400 of an arbitrary number of elliptic curve points
5401 \sum scalars[i]*points[i],
5402 optionally including the generator defined for the EC_GROUP:
5403 scalar*generator + \sum scalars[i]*points[i].
5405 EC_POINT_mul is a simple wrapper function for the typical case
5406 that the point list has just one item (besides the optional
5410 *) First EC_METHODs for curves over GF(p):
5412 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5413 operations and provides various method functions that can also
5414 operate with faster implementations of modular arithmetic.
5416 EC_GFp_mont_method() reuses most functions that are part of
5417 EC_GFp_simple_method, but uses Montgomery arithmetic.
5419 [Bodo Moeller; point addition and point doubling
5420 implementation directly derived from source code provided by
5421 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5423 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5424 crypto/ec/ec_lib.c):
5426 Curves are EC_GROUP objects (with an optional group generator)
5427 based on EC_METHODs that are built into the library.
5429 Points are EC_POINT objects based on EC_GROUP objects.
5431 Most of the framework would be able to handle curves over arbitrary
5432 finite fields, but as there are no obvious types for fields other
5433 than GF(p), some functions are limited to that for now.
5436 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5437 that the file contains a complete HTTP response.
5440 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5441 change the def and num file printf format specifier from "%-40sXXX"
5442 to "%-39s XXX". The latter will always guarantee a space after the
5443 field while the former will cause them to run together if the field
5444 is 40 of more characters long.
5447 *) Constify the cipher and digest 'method' functions and structures
5448 and modify related functions to take constant EVP_MD and EVP_CIPHER
5452 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5453 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5456 *) Modify EVP_Digest*() routines so they now return values. Although the
5457 internal software routines can never fail additional hardware versions
5461 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5463 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5464 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5467 ERR_R_NESTED_ASN1_ERROR
5469 ERR_R_MISSING_ASN1_EOS
5470 were 4 .. 9, conflicting with
5471 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5473 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5474 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5476 Add new error code 'ERR_R_INTERNAL_ERROR'.
5479 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5483 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5484 sets the subject name for a new request or supersedes the
5485 subject name in a given request. Formats that can be parsed are
5486 'CN=Some Name, OU=myOU, C=IT'
5488 'CN=Some Name/OU=myOU/C=IT'.
5490 Add options '-batch' and '-verbose' to 'openssl req'.
5491 [Massimiliano Pala <madwolf@hackmasters.net>]
5493 *) Introduce the possibility to access global variables through
5494 functions on platform were that's the best way to handle exporting
5495 global variables in shared libraries. To enable this functionality,
5496 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5497 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5498 is normally done by Configure or something similar).
5500 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5501 in the source file (foo.c) like this:
5503 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5504 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5506 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5507 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5509 OPENSSL_DECLARE_GLOBAL(int,foo);
5510 #define foo OPENSSL_GLOBAL_REF(foo)
5511 OPENSSL_DECLARE_GLOBAL(double,bar);
5512 #define bar OPENSSL_GLOBAL_REF(bar)
5514 The #defines are very important, and therefore so is including the
5515 header file everywhere where the defined globals are used.
5517 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5518 of ASN.1 items, but that structure is a bit different.
5520 The largest change is in util/mkdef.pl which has been enhanced with
5521 better and easier to understand logic to choose which symbols should
5522 go into the Windows .def files as well as a number of fixes and code
5523 cleanup (among others, algorithm keywords are now sorted
5524 lexicographically to avoid constant rewrites).
5527 *) In BN_div() keep a copy of the sign of 'num' before writing the
5528 result to 'rm' because if rm==num the value will be overwritten
5529 and produce the wrong result if 'num' is negative: this caused
5530 problems with BN_mod() and BN_nnmod().
5533 *) Function OCSP_request_verify(). This checks the signature on an
5534 OCSP request and verifies the signer certificate. The signer
5535 certificate is just checked for a generic purpose and OCSP request
5539 *) Add OCSP_check_validity() function to check the validity of OCSP
5540 responses. OCSP responses are prepared in real time and may only
5541 be a few seconds old. Simply checking that the current time lies
5542 between thisUpdate and nextUpdate max reject otherwise valid responses
5543 caused by either OCSP responder or client clock inaccuracy. Instead
5544 we allow thisUpdate and nextUpdate to fall within a certain period of
5545 the current time. The age of the response can also optionally be
5546 checked. Two new options -validity_period and -status_age added to
5550 *) If signature or public key algorithm is unrecognized print out its
5551 OID rather that just UNKNOWN.
5554 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5555 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5556 ID to be generated from the issuer certificate alone which can then be
5557 passed to OCSP_id_issuer_cmp().
5560 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5561 ASN1 modules to export functions returning ASN1_ITEM pointers
5562 instead of the ASN1_ITEM structures themselves. This adds several
5563 new macros which allow the underlying ASN1 function/structure to
5564 be accessed transparently. As a result code should not use ASN1_ITEM
5565 references directly (such as &X509_it) but instead use the relevant
5566 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5567 use of the new ASN1 code on platforms where exporting structures
5568 is problematical (for example in shared libraries) but exporting
5569 functions returning pointers to structures is not.
5572 *) Add support for overriding the generation of SSL/TLS session IDs.
5573 These callbacks can be registered either in an SSL_CTX or per SSL.
5574 The purpose of this is to allow applications to control, if they wish,
5575 the arbitrary values chosen for use as session IDs, particularly as it
5576 can be useful for session caching in multiple-server environments. A
5577 command-line switch for testing this (and any client code that wishes
5578 to use such a feature) has been added to "s_server".
5579 [Geoff Thorpe, Lutz Jaenicke]
5581 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5582 of the form '#if defined(...) || defined(...) || ...' and
5583 '#if !defined(...) && !defined(...) && ...'. This also avoids
5584 the growing number of special cases it was previously handling.
5587 *) Make all configuration macros available for application by making
5588 sure they are available in opensslconf.h, by giving them names starting
5589 with "OPENSSL_" to avoid conflicts with other packages and by making
5590 sure e_os2.h will cover all platform-specific cases together with
5592 Additionally, it is now possible to define configuration/platform-
5593 specific names (called "system identities"). In the C code, these
5594 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5595 macro with the name beginning with "OPENSSL_SYS_", which is determined
5596 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5600 *) New option -set_serial to 'req' and 'x509' this allows the serial
5601 number to use to be specified on the command line. Previously self
5602 signed certificates were hard coded with serial number 0 and the
5603 CA options of 'x509' had to use a serial number in a file which was
5607 *) New options to 'ca' utility to support V2 CRL entry extensions.
5608 Currently CRL reason, invalidity date and hold instruction are
5609 supported. Add new CRL extensions to V3 code and some new objects.
5612 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5613 disable standard block padding (aka PKCS#5 padding) in the EVP
5614 API, which was previously mandatory. This means that the data is
5615 not padded in any way and so the total length much be a multiple
5616 of the block size, otherwise an error occurs.
5619 *) Initial (incomplete) OCSP SSL support.
5622 *) New function OCSP_parse_url(). This splits up a URL into its host,
5623 port and path components: primarily to parse OCSP URLs. New -url
5624 option to ocsp utility.
5627 *) New nonce behavior. The return value of OCSP_check_nonce() now
5628 reflects the various checks performed. Applications can decide
5629 whether to tolerate certain situations such as an absent nonce
5630 in a response when one was present in a request: the ocsp application
5631 just prints out a warning. New function OCSP_add1_basic_nonce()
5632 this is to allow responders to include a nonce in a response even if
5633 the request is nonce-less.
5636 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5637 skipped when using openssl x509 multiple times on a single input file,
5638 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5641 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5642 set string type: to handle setting ASN1_TIME structures. Fix ca
5643 utility to correctly initialize revocation date of CRLs.
5646 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5647 the clients preferred ciphersuites and rather use its own preferences.
5648 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5649 Internet Explorer by ensuring unchanged hash method during stepup.
5650 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5653 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5654 to aes and add a new 'exist' option to print out symbols that don't
5658 *) Additional options to ocsp utility to allow flags to be set and
5659 additional certificates supplied.
5662 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5663 OCSP client a number of certificate to only verify the response
5667 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5668 handle the new API. Currently only ECB, CBC modes supported. Add new
5671 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5672 Encryption Standard (AES) Ciphersuites for Transport Layer
5673 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5674 not enabled by default and were not part of the "ALL" ciphersuite
5675 alias because they were not yet official; they could be
5676 explicitly requested by specifying the "AESdraft" ciphersuite
5677 group alias. In the final release of OpenSSL 0.9.7, the group
5678 alias is called "AES" and is part of "ALL".)
5679 [Ben Laurie, Steve Henson, Bodo Moeller]
5681 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5682 request to response.
5685 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5686 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5687 extract information from a certificate request. OCSP_response_create()
5688 creates a response and optionally adds a basic response structure.
5689 OCSP_basic_add1_status() adds a complete single response to a basic
5690 response and returns the OCSP_SINGLERESP structure just added (to allow
5691 extensions to be included for example). OCSP_basic_add1_cert() adds a
5692 certificate to a basic response and OCSP_basic_sign() signs a basic
5693 response with various flags. New helper functions ASN1_TIME_check()
5694 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5695 (converts ASN1_TIME to GeneralizedTime).
5698 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5699 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5700 structure from a certificate. X509_pubkey_digest() digests the public_key
5701 contents: this is used in various key identifiers.
5704 *) Make sk_sort() tolerate a NULL argument.
5705 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5707 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5708 passed by the function are trusted implicitly. If any of them signed the
5709 response then it is assumed to be valid and is not verified.
5712 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5713 to data. This was previously part of the PKCS7 ASN1 code. This
5714 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5715 [Steve Henson, reported by Kenneth R. Robinette
5716 <support@securenetterm.com>]
5718 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5719 routines: without these tracing memory leaks is very painful.
5720 Fix leaks in PKCS12 and PKCS7 routines.
5723 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5724 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5725 effectively meant GeneralizedTime would never be used. Now it
5726 is initialised to -1 but X509_time_adj() now has to check the value
5727 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5728 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5729 [Steve Henson, reported by Kenneth R. Robinette
5730 <support@securenetterm.com>]
5732 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5733 result in a zero length in the ASN1_INTEGER structure which was
5734 not consistent with the structure when d2i_ASN1_INTEGER() was used
5735 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5736 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5737 where it did not print out a minus for negative ASN1_INTEGER.
5740 *) Add summary printout to ocsp utility. The various functions which
5741 convert status values to strings have been renamed to:
5742 OCSP_response_status_str(), OCSP_cert_status_str() and
5743 OCSP_crl_reason_str() and are no longer static. New options
5744 to verify nonce values and to disable verification. OCSP response
5745 printout format cleaned up.
5748 *) Add additional OCSP certificate checks. These are those specified
5749 in RFC2560. This consists of two separate checks: the CA of the
5750 certificate being checked must either be the OCSP signer certificate
5751 or the issuer of the OCSP signer certificate. In the latter case the
5752 OCSP signer certificate must contain the OCSP signing extended key
5753 usage. This check is performed by attempting to match the OCSP
5754 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5755 in the OCSP_CERTID structures of the response.
5758 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5759 and related routines. This uses the standard OpenSSL certificate
5760 verify routines to perform initial checks (just CA validity) and
5761 to obtain the certificate chain. Then additional checks will be
5762 performed on the chain. Currently the root CA is checked to see
5763 if it is explicitly trusted for OCSP signing. This is used to set
5764 a root CA as a global signing root: that is any certificate that
5765 chains to that CA is an acceptable OCSP signing certificate.
5768 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5769 extensions from a separate configuration file.
5770 As when reading extensions from the main configuration file,
5771 the '-extensions ...' option may be used for specifying the
5773 [Massimiliano Pala <madwolf@comune.modena.it>]
5775 *) New OCSP utility. Allows OCSP requests to be generated or
5776 read. The request can be sent to a responder and the output
5777 parsed, outputed or printed in text form. Not complete yet:
5778 still needs to check the OCSP response validity.
5781 *) New subcommands for 'openssl ca':
5782 'openssl ca -status <serial>' prints the status of the cert with
5783 the given serial number (according to the index file).
5784 'openssl ca -updatedb' updates the expiry status of certificates
5786 [Massimiliano Pala <madwolf@comune.modena.it>]
5788 *) New '-newreq-nodes' command option to CA.pl. This is like
5789 '-newreq', but calls 'openssl req' with the '-nodes' option
5790 so that the resulting key is not encrypted.
5791 [Damien Miller <djm@mindrot.org>]
5793 *) New configuration for the GNU Hurd.
5794 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5796 *) Initial code to implement OCSP basic response verify. This
5797 is currently incomplete. Currently just finds the signer's
5798 certificate and verifies the signature on the response.
5801 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5802 value of OPENSSLDIR. This is available via the new '-d' option
5803 to 'openssl version', and is also included in 'openssl version -a'.
5806 *) Allowing defining memory allocation callbacks that will be given
5807 file name and line number information in additional arguments
5808 (a const char* and an int). The basic functionality remains, as
5809 well as the original possibility to just replace malloc(),
5810 realloc() and free() by functions that do not know about these
5811 additional arguments. To register and find out the current
5812 settings for extended allocation functions, the following
5813 functions are provided:
5815 CRYPTO_set_mem_ex_functions
5816 CRYPTO_set_locked_mem_ex_functions
5817 CRYPTO_get_mem_ex_functions
5818 CRYPTO_get_locked_mem_ex_functions
5820 These work the same way as CRYPTO_set_mem_functions and friends.
5821 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5822 extended allocation function is enabled.
5823 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5824 a conventional allocation function is enabled.
5825 [Richard Levitte, Bodo Moeller]
5827 *) Finish off removing the remaining LHASH function pointer casts.
5828 There should no longer be any prototype-casting required when using
5829 the LHASH abstraction, and any casts that remain are "bugs". See
5830 the callback types and macros at the head of lhash.h for details
5831 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5834 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5835 If /dev/[u]random devices are not available or do not return enough
5836 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5838 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5839 /etc/entropy will be queried once each in this sequence, quering stops
5840 when enough entropy was collected without querying more sockets.
5843 *) Change the Unix RAND_poll() variant to be able to poll several
5844 random devices, as specified by DEVRANDOM, until a sufficient amount
5845 of data has been collected. We spend at most 10 ms on each file
5846 (select timeout) and read in non-blocking mode. DEVRANDOM now
5847 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5848 (previously it was just the string "/dev/urandom"), so on typical
5849 platforms the 10 ms delay will never occur.
5850 Also separate out the Unix variant to its own file, rand_unix.c.
5851 For VMS, there's a currently-empty rand_vms.c.
5854 *) Move OCSP client related routines to ocsp_cl.c. These
5855 provide utility functions which an application needing
5856 to issue a request to an OCSP responder and analyse the
5857 response will typically need: as opposed to those which an
5858 OCSP responder itself would need which will be added later.
5860 OCSP_request_sign() signs an OCSP request with an API similar
5861 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5862 response. OCSP_response_get1_basic() extracts basic response
5863 from response. OCSP_resp_find_status(): finds and extracts status
5864 information from an OCSP_CERTID structure (which will be created
5865 when the request structure is built). These are built from lower
5866 level functions which work on OCSP_SINGLERESP structures but
5867 wont normally be used unless the application wishes to examine
5868 extensions in the OCSP response for example.
5870 Replace nonce routines with a pair of functions.
5871 OCSP_request_add1_nonce() adds a nonce value and optionally
5872 generates a random value. OCSP_check_nonce() checks the
5873 validity of the nonce in an OCSP response.
5876 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5877 This doesn't copy the supplied OCSP_CERTID and avoids the
5878 need to free up the newly created id. Change return type
5879 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5880 This can then be used to add extensions to the request.
5881 Deleted OCSP_request_new(), since most of its functionality
5882 is now in OCSP_REQUEST_new() (and the case insensitive name
5883 clash) apart from the ability to set the request name which
5884 will be added elsewhere.
5887 *) Update OCSP API. Remove obsolete extensions argument from
5888 various functions. Extensions are now handled using the new
5889 OCSP extension code. New simple OCSP HTTP function which
5890 can be used to send requests and parse the response.
5893 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5894 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5895 uses the special reorder version of SET OF to sort the attributes
5896 and reorder them to match the encoded order. This resolves a long
5897 standing problem: a verify on a PKCS7 structure just after signing
5898 it used to fail because the attribute order did not match the
5899 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5900 it uses the received order. This is necessary to tolerate some broken
5901 software that does not order SET OF. This is handled by encoding
5902 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5903 to produce the required SET OF.
5906 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5907 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5908 files to get correct declarations of the ASN.1 item variables.
5911 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5912 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5913 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5914 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5915 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5916 ASN1_ITEM and no wrapper functions.
5919 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5920 replace the old function pointer based I/O routines. Change most of
5921 the *_d2i_bio() and *_d2i_fp() functions to use these.
5924 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5925 lines, recognice more "algorithms" that can be deselected, and make
5926 it complain about algorithm deselection that isn't recognised.
5929 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5930 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5931 to use new functions. Add NO_ASN1_OLD which can be set to remove
5932 some old style ASN1 functions: this can be used to determine if old
5933 code will still work when these eventually go away.
5936 *) New extension functions for OCSP structures, these follow the
5937 same conventions as certificates and CRLs.
5940 *) New function X509V3_add1_i2d(). This automatically encodes and
5941 adds an extension. Its behaviour can be customised with various
5942 flags to append, replace or delete. Various wrappers added for
5943 certifcates and CRLs.
5946 *) Fix to avoid calling the underlying ASN1 print routine when
5947 an extension cannot be parsed. Correct a typo in the
5948 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5951 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5952 entries for variables.
5955 *) Add functionality to apps/openssl.c for detecting locking
5956 problems: As the program is single-threaded, all we have
5957 to do is register a locking callback using an array for
5958 storing which locks are currently held by the program.
5961 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5962 SSL_get_ex_data_X509_STORE_idx(), which is used in
5963 ssl_verify_cert_chain() and thus can be called at any time
5964 during TLS/SSL handshakes so that thread-safety is essential.
5965 Unfortunately, the ex_data design is not at all suited
5966 for multi-threaded use, so it probably should be abolished.
5969 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5970 [Broadcom, tweaked and integrated by Geoff Thorpe]
5972 *) Move common extension printing code to new function
5973 X509V3_print_extensions(). Reorganise OCSP print routines and
5974 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5977 *) New function X509_signature_print() to remove duplication in some
5981 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5982 set (this was treated exactly the same as SET OF previously). This
5983 is used to reorder the STACK representing the structure to match the
5984 encoding. This will be used to get round a problem where a PKCS7
5985 structure which was signed could not be verified because the STACK
5986 order did not reflect the encoded order.
5989 *) Reimplement the OCSP ASN1 module using the new code.
5992 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5993 for its ASN1 operations. The old style function pointers still exist
5994 for now but they will eventually go away.
5997 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5998 completely replaces the old ASN1 functionality with a table driven
5999 encoder and decoder which interprets an ASN1_ITEM structure describing
6000 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6001 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6002 has also been converted to the new form.
6005 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6006 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6007 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6008 for negative moduli.
6011 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6012 of not touching the result's sign bit.
6015 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6019 *) Changed the LHASH code to use prototypes for callbacks, and created
6020 macros to declare and implement thin (optionally static) functions
6021 that provide type-safety and avoid function pointer casting for the
6022 type-specific callbacks.
6025 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6027 [Veers Staats <staatsvr@asc.hpc.mil>,
6028 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6030 *) Reformat the FAQ so the different questions and answers can be divided
6031 in sections depending on the subject.
6034 *) Have the zlib compression code load ZLIB.DLL dynamically under
6038 *) New function BN_mod_sqrt for computing square roots modulo a prime
6039 (using the probabilistic Tonelli-Shanks algorithm unless
6040 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6041 be handled deterministically).
6042 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6044 *) Make BN_mod_inverse faster by explicitly handling small quotients
6045 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6046 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6049 *) New function BN_kronecker.
6052 *) Fix BN_gcd so that it works on negative inputs; the result is
6053 positive unless both parameters are zero.
6054 Previously something reasonably close to an infinite loop was
6055 possible because numbers could be growing instead of shrinking
6056 in the implementation of Euclid's algorithm.
6059 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6060 sign of the number in question.
6062 Fix BN_is_word(a,w) to work correctly for w == 0.
6064 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6065 because its test if the absolute value of 'a' equals 'w'.
6066 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6067 it exists mostly for use in the implementations of BN_is_zero(),
6068 BN_is_one(), and BN_is_word().
6071 *) New function BN_swap.
6074 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6075 the exponentiation functions are more likely to produce reasonable
6076 results on negative inputs.
6079 *) Change BN_mod_mul so that the result is always non-negative.
6080 Previously, it could be negative if one of the factors was negative;
6081 I don't think anyone really wanted that behaviour.
6084 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6085 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6086 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6087 and add new functions:
6096 BN_mod_lshift1_quick
6100 These functions always generate non-negative results.
6102 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6103 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6105 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6106 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6107 be reduced modulo m.
6108 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6111 The following entry accidentily appeared in the CHANGES file
6112 distributed with OpenSSL 0.9.7. The modifications described in
6113 it do *not* apply to OpenSSL 0.9.7.
6115 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6116 was actually never needed) and in BN_mul(). The removal in BN_mul()
6117 required a small change in bn_mul_part_recursive() and the addition
6118 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6119 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6120 bn_sub_words() and bn_add_words() except they take arrays with
6125 *) In 'openssl passwd', verify passwords read from the terminal
6126 unless the '-salt' option is used (which usually means that
6127 verification would just waste user's time since the resulting
6128 hash is going to be compared with some given password hash)
6129 or the new '-noverify' option is used.
6131 This is an incompatible change, but it does not affect
6132 non-interactive use of 'openssl passwd' (passwords on the command
6133 line, '-stdin' option, '-in ...' option) and thus should not
6137 *) Remove all references to RSAref, since there's no more need for it.
6140 *) Make DSO load along a path given through an environment variable
6141 (SHLIB_PATH) with shl_load().
6144 *) Constify the ENGINE code as a result of BIGNUM constification.
6145 Also constify the RSA code and most things related to it. In a
6146 few places, most notable in the depth of the ASN.1 code, ugly
6147 casts back to non-const were required (to be solved at a later
6151 *) Make it so the openssl application has all engines loaded by default.
6154 *) Constify the BIGNUM routines a little more.
6157 *) Add the following functions:
6159 ENGINE_load_cswift()
6161 ENGINE_load_atalla()
6163 ENGINE_load_builtin_engines()
6165 That way, an application can itself choose if external engines that
6166 are built-in in OpenSSL shall ever be used or not. The benefit is
6167 that applications won't have to be linked with libdl or other dso
6168 libraries unless it's really needed.
6170 Changed 'openssl engine' to load all engines on demand.
6171 Changed the engine header files to avoid the duplication of some
6172 declarations (they differed!).
6175 *) 'openssl engine' can now list capabilities.
6178 *) Better error reporting in 'openssl engine'.
6181 *) Never call load_dh_param(NULL) in s_server.
6184 *) Add engine application. It can currently list engines by name and
6185 identity, and test if they are actually available.
6188 *) Improve RPM specification file by forcing symbolic linking and making
6189 sure the installed documentation is also owned by root.root.
6190 [Damien Miller <djm@mindrot.org>]
6192 *) Give the OpenSSL applications more possibilities to make use of
6193 keys (public as well as private) handled by engines.
6196 *) Add OCSP code that comes from CertCo.
6199 *) Add VMS support for the Rijndael code.
6202 *) Added untested support for Nuron crypto accelerator.
6205 *) Add support for external cryptographic devices. This code was
6206 previously distributed separately as the "engine" branch.
6207 [Geoff Thorpe, Richard Levitte]
6209 *) Rework the filename-translation in the DSO code. It is now possible to
6210 have far greater control over how a "name" is turned into a filename
6211 depending on the operating environment and any oddities about the
6212 different shared library filenames on each system.
6215 *) Support threads on FreeBSD-elf in Configure.
6218 *) Fix for SHA1 assembly problem with MASM: it produces
6219 warnings about corrupt line number information when assembling
6220 with debugging information. This is caused by the overlapping
6222 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6225 NCONF_get_number() has no error checking at all. As a replacement,
6226 NCONF_get_number_e() is defined (_e for "error checking") and is
6227 promoted strongly. The old NCONF_get_number is kept around for
6228 binary backward compatibility.
6229 Make it possible for methods to load from something other than a BIO,
6230 by providing a function pointer that is given a name instead of a BIO.
6231 For example, this could be used to load configuration data from an
6235 *) Fix for non blocking accept BIOs. Added new I/O special reason
6236 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6237 with non blocking I/O was not possible because no retry code was
6238 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6242 *) Added the beginnings of Rijndael support.
6245 *) Fix for bug in DirectoryString mask setting. Add support for
6246 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6247 to allow certificate printing to more controllable, additional
6248 'certopt' option to 'x509' to allow new printing options to be
6252 *) Clean old EAY MD5 hack from e_os.h.
6255 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6257 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6258 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6259 [Joe Orton, Steve Henson]
6261 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6263 *) Fix additional bug revealed by the NISCC test suite:
6265 Stop bug triggering large recursion when presented with
6266 certain ASN.1 tags (CVE-2003-0851)
6269 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6271 *) Fix various bugs revealed by running the NISCC test suite:
6273 Stop out of bounds reads in the ASN1 code when presented with
6274 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6276 If verify callback ignores invalid public key errors don't try to check
6277 certificate signature with the NULL public key.
6281 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6282 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6286 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6287 extra data after the compression methods not only for TLS 1.0
6288 but also for SSL 3.0 (as required by the specification).
6289 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6291 *) Change X509_certificate_type() to mark the key as exported/exportable
6292 when it's 512 *bits* long, not 512 bytes.
6295 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6297 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6298 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6299 a protocol version number mismatch like a decryption error
6300 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6303 *) Turn on RSA blinding by default in the default implementation
6304 to avoid a timing attack. Applications that don't want it can call
6305 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6306 They would be ill-advised to do so in most cases.
6307 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6309 *) Change RSA blinding code so that it works when the PRNG is not
6310 seeded (in this case, the secret RSA exponent is abused as
6311 an unpredictable seed -- if it is not unpredictable, there
6312 is no point in blinding anyway). Make RSA blinding thread-safe
6313 by remembering the creator's thread ID in rsa->blinding and
6314 having all other threads use local one-time blinding factors
6315 (this requires more computation than sharing rsa->blinding, but
6316 avoids excessive locking; and if an RSA object is not shared
6317 between threads, blinding will still be very fast).
6320 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6322 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6323 via timing by performing a MAC computation even if incorrrect
6324 block cipher padding has been found. This is a countermeasure
6325 against active attacks where the attacker has to distinguish
6326 between bad padding and a MAC verification error. (CVE-2003-0078)
6328 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6329 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6330 Martin Vuagnoux (EPFL, Ilion)]
6332 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6334 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6335 memory from it's contents. This is done with a counter that will
6336 place alternating values in each byte. This can be used to solve
6337 two issues: 1) the removal of calls to memset() by highly optimizing
6338 compilers, and 2) cleansing with other values than 0, since those can
6339 be read through on certain media, for example a swap space on disk.
6342 *) Bugfix: client side session caching did not work with external caching,
6343 because the session->cipher setting was not restored when reloading
6344 from the external cache. This problem was masked, when
6345 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6346 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6349 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6350 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6351 [Zeev Lieber <zeev-l@yahoo.com>]
6353 *) Undo an undocumented change introduced in 0.9.6e which caused
6354 repeated calls to OpenSSL_add_all_ciphers() and
6355 OpenSSL_add_all_digests() to be ignored, even after calling
6359 *) Change the default configuration reader to deal with last line not
6360 being properly terminated.
6363 *) Change X509_NAME_cmp() so it applies the special rules on handling
6364 DN values that are of type PrintableString, as well as RDNs of type
6365 emailAddress where the value has the type ia5String.
6366 [stefank@valicert.com via Richard Levitte]
6368 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6369 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6370 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6371 the bitwise-OR of the two for use by the majority of applications
6372 wanting this behaviour, and update the docs. The documented
6373 behaviour and actual behaviour were inconsistent and had been
6374 changing anyway, so this is more a bug-fix than a behavioural
6376 [Geoff Thorpe, diagnosed by Nadav Har'El]
6378 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6379 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6382 *) Fix initialization code race conditions in
6383 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6384 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6385 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6386 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6387 ssl2_get_cipher_by_char(),
6388 ssl3_get_cipher_by_char().
6389 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6391 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6392 the cached sessions are flushed, as the remove_cb() might use ex_data
6393 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6394 (see [openssl.org #212]).
6395 [Geoff Thorpe, Lutz Jaenicke]
6397 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6398 length, instead of the encoding length to d2i_ASN1_OBJECT.
6401 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6403 *) [In 0.9.6g-engine release:]
6404 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6405 [Lynn Gazis <lgazis@rainbow.com>]
6407 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6409 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6410 and get fix the header length calculation.
6411 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6412 Alon Kantor <alonk@checkpoint.com> (and others),
6415 *) Use proper error handling instead of 'assertions' in buffer
6416 overflow checks added in 0.9.6e. This prevents DoS (the
6417 assertions could call abort()).
6418 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6420 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6422 *) Add various sanity checks to asn1_get_length() to reject
6423 the ASN1 length bytes if they exceed sizeof(long), will appear
6424 negative or the content length exceeds the length of the
6426 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6428 *) Fix cipher selection routines: ciphers without encryption had no flags
6429 for the cipher strength set and where therefore not handled correctly
6430 by the selection routines (PR #130).
6433 *) Fix EVP_dsa_sha macro.
6437 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6438 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6439 that was added in OpenSSL 0.9.6d.
6441 As the countermeasure turned out to be incompatible with some
6442 broken SSL implementations, the new option is part of SSL_OP_ALL.
6443 SSL_OP_ALL is usually employed when compatibility with weird SSL
6444 implementations is desired (e.g. '-bugs' option to 's_client' and
6445 's_server'), so the new option is automatically set in many
6449 *) Changes in security patch:
6451 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6452 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6453 Air Force Materiel Command, USAF, under agreement number
6456 *) Add various sanity checks to asn1_get_length() to reject
6457 the ASN1 length bytes if they exceed sizeof(long), will appear
6458 negative or the content length exceeds the length of the
6459 supplied buffer. (CVE-2002-0659)
6460 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6462 *) Assertions for various potential buffer overflows, not known to
6464 [Ben Laurie (CHATS)]
6466 *) Various temporary buffers to hold ASCII versions of integers were
6467 too small for 64 bit platforms. (CVE-2002-0655)
6468 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6470 *) Remote buffer overflow in SSL3 protocol - an attacker could
6471 supply an oversized session ID to a client. (CVE-2002-0656)
6472 [Ben Laurie (CHATS)]
6474 *) Remote buffer overflow in SSL2 protocol - an attacker could
6475 supply an oversized client master key. (CVE-2002-0656)
6476 [Ben Laurie (CHATS)]
6478 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6480 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6481 encoded as NULL) with id-dsa-with-sha1.
6482 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6484 *) Check various X509_...() return values in apps/req.c.
6485 [Nils Larsch <nla@trustcenter.de>]
6487 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6488 an end-of-file condition would erronously be flagged, when the CRLF
6489 was just at the end of a processed block. The bug was discovered when
6490 processing data through a buffering memory BIO handing the data to a
6491 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6492 <ptsekov@syntrex.com> and Nedelcho Stanev.
6495 *) Implement a countermeasure against a vulnerability recently found
6496 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6497 before application data chunks to avoid the use of known IVs
6498 with data potentially chosen by the attacker.
6501 *) Fix length checks in ssl3_get_client_hello().
6504 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6505 to prevent ssl3_read_internal() from incorrectly assuming that
6506 ssl3_read_bytes() found application data while handshake
6507 processing was enabled when in fact s->s3->in_read_app_data was
6508 merely automatically cleared during the initial handshake.
6509 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6511 *) Fix object definitions for Private and Enterprise: they were not
6512 recognized in their shortname (=lowercase) representation. Extend
6513 obj_dat.pl to issue an error when using undefined keywords instead
6514 of silently ignoring the problem (Svenning Sorensen
6515 <sss@sss.dnsalias.net>).
6518 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6519 generators, i.e. generators other than 2 and 5. (Previously, the
6520 code did not properly initialise the 'add' and 'rem' values to
6521 BN_generate_prime().)
6523 In the new general case, we do not insist that 'generator' is
6524 actually a primitive root: This requirement is rather pointless;
6525 a generator of the order-q subgroup is just as good, if not
6529 *) Map new X509 verification errors to alerts. Discovered and submitted by
6530 Tom Wu <tom@arcot.com>.
6533 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6534 returning non-zero before the data has been completely received
6535 when using non-blocking I/O.
6536 [Bodo Moeller; problem pointed out by John Hughes]
6538 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6539 [Ben Laurie, Lutz Jaenicke]
6541 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6542 Yoram Zahavi <YoramZ@gilian.com>).
6545 *) Add information about CygWin 1.3 and on, and preserve proper
6546 configuration for the versions before that.
6547 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6549 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6550 check whether we deal with a copy of a session and do not delete from
6551 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6552 <izhar@checkpoint.com>.
6555 *) Do not store session data into the internal session cache, if it
6556 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6557 flag is set). Proposed by Aslam <aslam@funk.com>.
6560 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6564 *) [In 0.9.6d-engine release:]
6565 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6566 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6568 *) Add the configuration target linux-s390x.
6569 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6571 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6572 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6573 variable as an indication that a ClientHello message has been
6574 received. As the flag value will be lost between multiple
6575 invocations of ssl3_accept when using non-blocking I/O, the
6576 function may not be aware that a handshake has actually taken
6577 place, thus preventing a new session from being added to the
6580 To avoid this problem, we now set s->new_session to 2 instead of
6581 using a local variable.
6582 [Lutz Jaenicke, Bodo Moeller]
6584 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6585 if the SSL_R_LENGTH_MISMATCH error is detected.
6586 [Geoff Thorpe, Bodo Moeller]
6588 *) New 'shared_ldflag' column in Configure platform table.
6591 *) Fix EVP_CIPHER_mode macro.
6592 ["Dan S. Camper" <dan@bti.net>]
6594 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6595 type, we must throw them away by setting rr->length to 0.
6596 [D P Chang <dpc@qualys.com>]
6598 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6600 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6601 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6602 worked incorrectly for those cases where range = 10..._2 and
6603 3*range is two bits longer than range.)
6606 *) Only add signing time to PKCS7 structures if it is not already
6610 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6611 OBJ_ld_ce should be OBJ_id_ce.
6612 Also some ip-pda OIDs in crypto/objects/objects.txt were
6613 incorrect (cf. RFC 3039).
6614 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6616 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6617 returns early because it has nothing to do.
6618 [Andy Schneider <andy.schneider@bjss.co.uk>]
6620 *) [In 0.9.6c-engine release:]
6621 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6622 [Andy Schneider <andy.schneider@bjss.co.uk>]
6624 *) [In 0.9.6c-engine release:]
6625 Add support for Cryptographic Appliance's keyserver technology.
6626 (Use engine 'keyclient')
6627 [Cryptographic Appliances and Geoff Thorpe]
6629 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6630 is called via tools/c89.sh because arguments have to be
6631 rearranged (all '-L' options must appear before the first object
6633 [Richard Shapiro <rshapiro@abinitio.com>]
6635 *) [In 0.9.6c-engine release:]
6636 Add support for Broadcom crypto accelerator cards, backported
6638 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6640 *) [In 0.9.6c-engine release:]
6641 Add support for SureWare crypto accelerator cards from
6642 Baltimore Technologies. (Use engine 'sureware')
6643 [Baltimore Technologies and Mark Cox]
6645 *) [In 0.9.6c-engine release:]
6646 Add support for crypto accelerator cards from Accelerated
6647 Encryption Processing, www.aep.ie. (Use engine 'aep')
6648 [AEP Inc. and Mark Cox]
6650 *) Add a configuration entry for gcc on UnixWare.
6651 [Gary Benson <gbenson@redhat.com>]
6653 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6654 messages are stored in a single piece (fixed-length part and
6655 variable-length part combined) and fix various bugs found on the way.
6658 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6659 instead. BIO_gethostbyname() does not know what timeouts are
6660 appropriate, so entries would stay in cache even when they have
6662 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6664 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6665 faced with a pathologically small ClientHello fragment that does
6666 not contain client_version: Instead of aborting with an error,
6667 simply choose the highest available protocol version (i.e.,
6668 TLS 1.0 unless it is disabled). In practice, ClientHello
6669 messages are never sent like this, but this change gives us
6670 strictly correct behaviour at least for TLS.
6673 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6674 never resets s->method to s->ctx->method when called from within
6675 one of the SSL handshake functions.
6676 [Bodo Moeller; problem pointed out by Niko Baric]
6678 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6679 (sent using the client's version number) if client_version is
6680 smaller than the protocol version in use. Also change
6681 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6682 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6683 the client will at least see that alert.
6686 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6690 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6691 client receives HelloRequest while in a handshake.
6692 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6694 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6695 should end in 'break', not 'goto end' which circuments various
6696 cleanups done in state SSL_ST_OK. But session related stuff
6697 must be disabled for SSL_ST_OK in the case that we just sent a
6700 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6701 before just sending a HelloRequest.
6702 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6704 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6705 reveal whether illegal block cipher padding was found or a MAC
6706 verification error occured. (Neither SSLerr() codes nor alerts
6707 are directly visible to potential attackers, but the information
6708 may leak via logfiles.)
6710 Similar changes are not required for the SSL 2.0 implementation
6711 because the number of padding bytes is sent in clear for SSL 2.0,
6712 and the extra bytes are just ignored. However ssl/s2_pkt.c
6713 failed to verify that the purported number of padding bytes is in
6717 *) Add OpenUNIX-8 support including shared libraries
6718 (Boyd Lynn Gerber <gerberb@zenez.com>).
6721 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6722 'wristwatch attack' using huge encoding parameters (cf.
6723 James H. Manger's CRYPTO 2001 paper). Note that the
6724 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6725 encoding parameters and hence was not vulnerable.
6728 *) BN_sqr() bug fix.
6729 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6731 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6732 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6733 followed by modular reduction.
6734 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6736 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6737 equivalent based on BN_pseudo_rand() instead of BN_rand().
6740 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6741 This function was broken, as the check for a new client hello message
6742 to handle SGC did not allow these large messages.
6743 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6746 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6749 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6750 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6753 *) Rework the configuration and shared library support for Tru64 Unix.
6754 The configuration part makes use of modern compiler features and
6755 still retains old compiler behavior for those that run older versions
6756 of the OS. The shared library support part includes a variant that
6757 uses the RPATH feature, and is available through the special
6758 configuration target "alpha-cc-rpath", which will never be selected
6760 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6762 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6763 with the same message size as in ssl3_get_certificate_request().
6764 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6765 messages might inadvertently be reject as too long.
6766 [Petr Lampa <lampa@fee.vutbr.cz>]
6768 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6771 *) Modified SSL library such that the verify_callback that has been set
6772 specificly for an SSL object with SSL_set_verify() is actually being
6773 used. Before the change, a verify_callback set with this function was
6774 ignored and the verify_callback() set in the SSL_CTX at the time of
6775 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6776 to allow the necessary settings.
6779 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6780 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6781 done automatically (in contradiction to the requirements of the C
6782 standard). This made problems when used from OpenSSH.
6785 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6786 dh->length and always used
6788 BN_rand_range(priv_key, dh->p).
6790 BN_rand_range() is not necessary for Diffie-Hellman, and this
6791 specific range makes Diffie-Hellman unnecessarily inefficient if
6792 dh->length (recommended exponent length) is much smaller than the
6793 length of dh->p. We could use BN_rand_range() if the order of
6794 the subgroup was stored in the DH structure, but we only have
6799 BN_rand(priv_key, l, ...)
6801 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6807 RSA_eay_public_encrypt
6808 RSA_eay_private_decrypt
6809 RSA_eay_private_encrypt (signing)
6810 RSA_eay_public_decrypt (signature verification)
6812 (default implementations for RSA_public_encrypt,
6813 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6814 always reject numbers >= n.
6817 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6818 to synchronize access to 'locking_thread'. This is necessary on
6819 systems where access to 'locking_thread' (an 'unsigned long'
6820 variable) is not atomic.
6823 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6824 *before* setting the 'crypto_lock_rand' flag. The previous code had
6825 a race condition if 0 is a valid thread ID.
6826 [Travis Vitek <vitek@roguewave.com>]
6828 *) Add support for shared libraries under Irix.
6829 [Albert Chin-A-Young <china@thewrittenword.com>]
6831 *) Add configuration option to build on Linux on both big-endian and
6833 [Ralf Baechle <ralf@uni-koblenz.de>]
6835 *) Add the possibility to create shared libraries on HP-UX.
6838 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6840 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6841 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6842 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6843 PRNG state recovery was possible based on the output of
6844 one PRNG request appropriately sized to gain knowledge on
6845 'md' followed by enough consecutive 1-byte PRNG requests
6846 to traverse all of 'state'.
6848 1. When updating 'md_local' (the current thread's copy of 'md')
6849 during PRNG output generation, hash all of the previous
6850 'md_local' value, not just the half used for PRNG output.
6852 2. Make the number of bytes from 'state' included into the hash
6853 independent from the number of PRNG bytes requested.
6855 The first measure alone would be sufficient to avoid
6856 Markku-Juhani's attack. (Actually it had never occurred
6857 to me that the half of 'md_local' used for chaining was the
6858 half from which PRNG output bytes were taken -- I had always
6859 assumed that the secret half would be used.) The second
6860 measure makes sure that additional data from 'state' is never
6861 mixed into 'md_local' in small portions; this heuristically
6862 further strengthens the PRNG.
6865 *) Fix crypto/bn/asm/mips3.s.
6868 *) When only the key is given to "enc", the IV is undefined. Print out
6869 an error message in this case.
6872 *) Handle special case when X509_NAME is empty in X509 printing routines.
6875 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6876 positive and less than q.
6879 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6880 used: it isn't thread safe and the add_lock_callback should handle
6882 [Paul Rose <Paul.Rose@bridge.com>]
6884 *) Verify that incoming data obeys the block size in
6885 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6889 [Ulf Möller, Bodo Möller]
6891 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6892 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6893 when fixing the server behaviour for backwards-compatible 'client
6894 hello' messages. (Note that the attack is impractical against
6895 SSL 3.0 and TLS 1.0 anyway because length and version checking
6896 means that the probability of guessing a valid ciphertext is
6897 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6900 Before 0.9.5, the countermeasure (hide the error by generating a
6901 random 'decryption result') did not work properly because
6902 ERR_clear_error() was missing, meaning that SSL_get_error() would
6903 detect the supposedly ignored error.
6905 Both problems are now fixed.
6908 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6909 (previously it was 1024).
6912 *) Fix for compatibility mode trust settings: ignore trust settings
6913 unless some valid trust or reject settings are present.
6916 *) Fix for blowfish EVP: its a variable length cipher.
6919 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6920 parameters in DSA public key structures and return an error in the
6921 DSA routines if parameters are absent.
6924 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6925 in the current directory if neither $RANDFILE nor $HOME was set.
6926 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6927 caused some confusion to Windows users who haven't defined $HOME.
6928 Thus RAND_file_name() is changed again: e_os.h can define a
6929 DEFAULT_HOME, which will be used if $HOME is not set.
6930 For Windows, we use "C:"; on other platforms, we still require
6931 environment variables.
6933 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6934 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6935 having multiple threads call RAND_poll() concurrently.
6938 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6939 combination of a flag and a thread ID variable.
6940 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6941 flag), *other* threads can enter ssleay_add_bytes without obeying
6942 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6943 that they do not hold after the first thread unsets add_do_not_lock).
6946 *) Change bctest again: '-x' expressions are not available in all
6950 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6952 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6953 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6955 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6956 the default extension for executables, if any. Also, make the perl
6957 scripts that use symlink() to test if it really exists and use "cp"
6958 if it doesn't. All this made OpenSSL compilable and installable in
6962 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6963 If SEQUENCE is length is indefinite just set c->slen to the total
6964 amount of data available.
6965 [Steve Henson, reported by shige@FreeBSD.org]
6966 [This change does not apply to 0.9.7.]
6968 *) Change bctest to avoid here-documents inside command substitution
6969 (workaround for FreeBSD /bin/sh bug).
6970 For compatibility with Ultrix, avoid shell functions (introduced
6971 in the bctest version that searches along $PATH).
6974 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6975 with des_encrypt() defined on some operating systems, like Solaris
6979 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6980 On the Importance of Eliminating Errors in Cryptographic
6981 Computations, J. Cryptology 14 (2001) 2, 101-119,
6982 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6985 *) MIPS assembler BIGNUM division bug fix.
6988 *) Disabled incorrect Alpha assembler code.
6991 *) Fix PKCS#7 decode routines so they correctly update the length
6992 after reading an EOC for the EXPLICIT tag.
6994 [This change does not apply to 0.9.7.]
6996 *) Fix bug in PKCS#12 key generation routines. This was triggered
6997 if a 3DES key was generated with a 0 initial byte. Include
6998 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6999 (but broken) behaviour.
7002 *) Enhance bctest to search for a working bc along $PATH and print
7004 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7006 *) Fix memory leaks in err.c: free err_data string if necessary;
7007 don't write to the wrong index in ERR_set_error_data.
7010 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7014 *) Replace rdtsc with _emit statements for VC++ version 5.
7015 [Jeremy Cooper <jeremy@baymoo.org>]
7017 *) Make it possible to reuse SSLv2 sessions.
7020 *) In copy_email() check for >= 0 as a return value for
7021 X509_NAME_get_index_by_NID() since 0 is a valid index.
7022 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7024 *) Avoid coredump with unsupported or invalid public keys by checking if
7025 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7026 PKCS7_verify() fails with non detached data.
7029 *) Don't use getenv in library functions when run as setuid/setgid.
7030 New function OPENSSL_issetugid().
7033 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7034 due to incorrect handling of multi-threading:
7036 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7038 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7040 3. Count how many times MemCheck_off() has been called so that
7041 nested use can be treated correctly. This also avoids
7042 inband-signalling in the previous code (which relied on the
7043 assumption that thread ID 0 is impossible).
7046 *) Add "-rand" option also to s_client and s_server.
7049 *) Fix CPU detection on Irix 6.x.
7050 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7051 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7053 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7056 [This change does not apply to 0.9.7.]
7058 *) Use the cached encoding of an X509_NAME structure rather than
7059 copying it. This is apparently the reason for the libsafe "errors"
7060 but the code is actually correct.
7063 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7064 Bleichenbacher's DSA attack.
7065 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7066 to be set and top=0 forces the highest bit to be set; top=-1 is new
7067 and leaves the highest bit random.
7068 [Ulf Moeller, Bodo Moeller]
7070 *) In the NCONF_...-based implementations for CONF_... queries
7071 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7072 a temporary CONF structure with the data component set to NULL
7073 (which gives segmentation faults in lh_retrieve).
7074 Instead, use NULL for the CONF pointer in CONF_get_string and
7075 CONF_get_number (which may use environment variables) and directly
7076 return NULL from CONF_get_section.
7079 *) Fix potential buffer overrun for EBCDIC.
7082 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7083 keyUsage if basicConstraints absent for a CA.
7086 *) Make SMIME_write_PKCS7() write mail header values with a format that
7087 is more generally accepted (no spaces before the semicolon), since
7088 some programs can't parse those values properly otherwise. Also make
7089 sure BIO's that break lines after each write do not create invalid
7093 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7094 macros previously used would not encode an empty SEQUENCE OF
7095 and break the signature.
7097 [This change does not apply to 0.9.7.]
7099 *) Zero the premaster secret after deriving the master secret in
7103 *) Add some EVP_add_digest_alias registrations (as found in
7104 OpenSSL_add_all_digests()) to SSL_library_init()
7105 aka OpenSSL_add_ssl_algorithms(). This provides improved
7106 compatibility with peers using X.509 certificates
7107 with unconventional AlgorithmIdentifier OIDs.
7110 *) Fix for Irix with NO_ASM.
7111 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7113 *) ./config script fixes.
7114 [Ulf Moeller, Richard Levitte]
7116 *) Fix 'openssl passwd -1'.
7119 *) Change PKCS12_key_gen_asc() so it can cope with non null
7120 terminated strings whose length is passed in the passlen
7121 parameter, for example from PEM callbacks. This was done
7122 by adding an extra length parameter to asc2uni().
7123 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7125 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7126 call failed, free the DSA structure.
7129 *) Fix to uni2asc() to cope with zero length Unicode strings.
7130 These are present in some PKCS#12 files.
7133 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7134 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7135 when writing a 32767 byte record.
7136 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7138 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7139 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7141 (RSA objects have a reference count access to which is protected
7142 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7143 so they are meant to be shared between threads.)
7144 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7145 "Reddie, Steven" <Steven.Reddie@ca.com>]
7147 *) Fix a deadlock in CRYPTO_mem_leaks().
7150 *) Use better test patterns in bntest.
7153 *) rand_win.c fix for Borland C.
7156 *) BN_rshift bugfix for n == 0.
7159 *) Add a 'bctest' script that checks for some known 'bc' bugs
7160 so that 'make test' does not abort just because 'bc' is broken.
7163 *) Store verify_result within SSL_SESSION also for client side to
7164 avoid potential security hole. (Re-used sessions on the client side
7165 always resulted in verify_result==X509_V_OK, not using the original
7166 result of the server certificate verification.)
7169 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7170 SSL3_RT_APPLICATION_DATA, return 0.
7171 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7175 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7176 releases, have been re-implemented by renaming the previous
7177 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7178 and ssl3_read_internal, respectively, and adding 'peek' parameters
7179 to them. The new ssl[23]_{read,peek} functions are calls to
7180 ssl[23]_read_internal with the 'peek' flag set appropriately.
7181 A 'peek' parameter has also been added to ssl3_read_bytes, which
7182 does the actual work for ssl3_read_internal.
7185 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7186 the method-specific "init()" handler. Also clean up ex_data after
7187 calling the method-specific "finish()" handler. Previously, this was
7188 happening the other way round.
7191 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7192 The previous value, 12, was not always sufficient for BN_mod_exp().
7195 *) Make sure that shared libraries get the internal name engine with
7196 the full version number and not just 0. This should mark the
7197 shared libraries as not backward compatible. Of course, this should
7198 be changed again when we can guarantee backward binary compatibility.
7201 *) Fix typo in get_cert_by_subject() in by_dir.c
7202 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7204 *) Rework the system to generate shared libraries:
7206 - Make note of the expected extension for the shared libraries and
7207 if there is a need for symbolic links from for example libcrypto.so.0
7208 to libcrypto.so.0.9.7. There is extended info in Configure for
7211 - Make as few rebuilds of the shared libraries as possible.
7213 - Still avoid linking the OpenSSL programs with the shared libraries.
7215 - When installing, install the shared libraries separately from the
7219 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7221 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7222 and not in SSL_clear because the latter is also used by the
7223 accept/connect functions; previously, the settings made by
7224 SSL_set_read_ahead would be lost during the handshake.
7225 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7227 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7228 Previously, it would create entries for disableed algorithms no
7232 *) Added several new manual pages for SSL_* function.
7235 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7237 *) In ssl23_get_client_hello, generate an error message when faced
7238 with an initial SSL 3.0/TLS record that is too small to contain the
7239 first two bytes of the ClientHello message, i.e. client_version.
7240 (Note that this is a pathologic case that probably has never happened
7241 in real life.) The previous approach was to use the version number
7242 from the record header as a substitute; but our protocol choice
7243 should not depend on that one because it is not authenticated
7244 by the Finished messages.
7247 *) More robust randomness gathering functions for Windows.
7248 [Jeffrey Altman <jaltman@columbia.edu>]
7250 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7251 not set then we don't setup the error code for issuer check errors
7252 to avoid possibly overwriting other errors which the callback does
7253 handle. If an application does set the flag then we assume it knows
7254 what it is doing and can handle the new informational codes
7258 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7259 a general "ANY" type, as such it should be able to decode anything
7260 including tagged types. However it didn't check the class so it would
7261 wrongly interpret tagged types in the same way as their universal
7262 counterpart and unknown types were just rejected. Changed so that the
7263 tagged and unknown types are handled in the same way as a SEQUENCE:
7264 that is the encoding is stored intact. There is also a new type
7265 "V_ASN1_OTHER" which is used when the class is not universal, in this
7266 case we have no idea what the actual type is so we just lump them all
7270 *) On VMS, stdout may very well lead to a file that is written to
7271 in a record-oriented fashion. That means that every write() will
7272 write a separate record, which will be read separately by the
7273 programs trying to read from it. This can be very confusing.
7275 The solution is to put a BIO filter in the way that will buffer
7276 text until a linefeed is reached, and then write everything a
7277 line at a time, so every record written will be an actual line,
7278 not chunks of lines and not (usually doesn't happen, but I've
7279 seen it once) several lines in one record. BIO_f_linebuffer() is
7282 Currently, it's a VMS-only method, because that's where it has
7283 been tested well enough.
7286 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7287 it can return incorrect results.
7288 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7289 but it was in 0.9.6-beta[12].)
7292 *) Disable the check for content being present when verifying detached
7293 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7294 include zero length content when signing messages.
7297 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7298 BIO_ctrl (for BIO pairs).
7301 *) Add DSO method for VMS.
7304 *) Bug fix: Montgomery multiplication could produce results with the
7308 *) Add RPM specification openssl.spec and modify it to build three
7309 packages. The default package contains applications, application
7310 documentation and run-time libraries. The devel package contains
7311 include files, static libraries and function documentation. The
7312 doc package contains the contents of the doc directory. The original
7313 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7316 *) Add a large number of documentation files for many SSL routines.
7317 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7319 *) Add a configuration entry for Sony News 4.
7320 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7322 *) Don't set the two most significant bits to one when generating a
7323 random number < q in the DSA library.
7326 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7327 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7328 the underlying transport is blocking) if a handshake took place.
7329 (The default behaviour is needed by applications such as s_client
7330 and s_server that use select() to determine when to use SSL_read;
7331 but for applications that know in advance when to expect data, it
7332 just makes things more complicated.)
7335 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7339 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7340 work better on such systems.
7341 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7343 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7344 Update PKCS12_parse() so it copies the friendlyName and the
7345 keyid to the certificates aux info.
7348 *) Fix bug in PKCS7_verify() which caused an infinite loop
7349 if there was more than one signature.
7350 [Sven Uszpelkat <su@celocom.de>]
7352 *) Major change in util/mkdef.pl to include extra information
7353 about each symbol, as well as presentig variables as well
7354 as functions. This change means that there's n more need
7355 to rebuild the .num files when some algorithms are excluded.
7358 *) Allow the verify time to be set by an application,
7359 rather than always using the current time.
7362 *) Phase 2 verify code reorganisation. The certificate
7363 verify code now looks up an issuer certificate by a
7364 number of criteria: subject name, authority key id
7365 and key usage. It also verifies self signed certificates
7366 by the same criteria. The main comparison function is
7367 X509_check_issued() which performs these checks.
7369 Lot of changes were necessary in order to support this
7370 without completely rewriting the lookup code.
7372 Authority and subject key identifier are now cached.
7374 The LHASH 'certs' is X509_STORE has now been replaced
7375 by a STACK_OF(X509_OBJECT). This is mainly because an
7376 LHASH can't store or retrieve multiple objects with
7377 the same hash value.
7379 As a result various functions (which were all internal
7380 use only) have changed to handle the new X509_STORE
7381 structure. This will break anything that messed round
7382 with X509_STORE internally.
7384 The functions X509_STORE_add_cert() now checks for an
7385 exact match, rather than just subject name.
7387 The X509_STORE API doesn't directly support the retrieval
7388 of multiple certificates matching a given criteria, however
7389 this can be worked round by performing a lookup first
7390 (which will fill the cache with candidate certificates)
7391 and then examining the cache for matches. This is probably
7392 the best we can do without throwing out X509_LOOKUP
7393 entirely (maybe later...).
7395 The X509_VERIFY_CTX structure has been enhanced considerably.
7397 All certificate lookup operations now go via a get_issuer()
7398 callback. Although this currently uses an X509_STORE it
7399 can be replaced by custom lookups. This is a simple way
7400 to bypass the X509_STORE hackery necessary to make this
7401 work and makes it possible to use more efficient techniques
7402 in future. A very simple version which uses a simple
7403 STACK for its trusted certificate store is also provided
7404 using X509_STORE_CTX_trusted_stack().
7406 The verify_cb() and verify() callbacks now have equivalents
7407 in the X509_STORE_CTX structure.
7409 X509_STORE_CTX also has a 'flags' field which can be used
7410 to customise the verify behaviour.
7413 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7414 excludes S/MIME capabilities.
7417 *) When a certificate request is read in keep a copy of the
7418 original encoding of the signed data and use it when outputing
7419 again. Signatures then use the original encoding rather than
7420 a decoded, encoded version which may cause problems if the
7421 request is improperly encoded.
7424 *) For consistency with other BIO_puts implementations, call
7425 buffer_write(b, ...) directly in buffer_puts instead of calling
7428 In BIO_puts, increment b->num_write as in BIO_write.
7429 [Peter.Sylvester@EdelWeb.fr]
7431 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7432 BN_zero, we may not return a BIGNUM with an array consisting of
7436 *) Avoid calling abort() from within the library when problems are
7437 detected, except if preprocessor symbols have been defined
7438 (such as REF_CHECK, BN_DEBUG etc.).
7441 *) New openssl application 'rsautl'. This utility can be
7442 used for low level RSA operations. DER public key
7443 BIO/fp routines also added.
7446 *) New Configure entry and patches for compiling on QNX 4.
7447 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7449 *) A demo state-machine implementation was sponsored by
7450 Nuron (http://www.nuron.com/) and is now available in
7451 demos/state_machine.
7454 *) New options added to the 'dgst' utility for signature
7455 generation and verification.
7458 *) Unrecognized PKCS#7 content types are now handled via a
7459 catch all ASN1_TYPE structure. This allows unsupported
7460 types to be stored as a "blob" and an application can
7461 encode and decode it manually.
7464 *) Fix various signed/unsigned issues to make a_strex.c
7466 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7468 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7469 length if passed a buffer. ASN1_INTEGER_to_BN failed
7470 if passed a NULL BN and its argument was negative.
7471 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7473 *) Modification to PKCS#7 encoding routines to output definite
7474 length encoding. Since currently the whole structures are in
7475 memory there's not real point in using indefinite length
7476 constructed encoding. However if OpenSSL is compiled with
7477 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7480 *) Added BIO_vprintf() and BIO_vsnprintf().
7483 *) Added more prefixes to parse for in the the strings written
7484 through a logging bio, to cover all the levels that are available
7485 through syslog. The prefixes are now:
7487 PANIC, EMERG, EMR => LOG_EMERG
7488 ALERT, ALR => LOG_ALERT
7489 CRIT, CRI => LOG_CRIT
7490 ERROR, ERR => LOG_ERR
7491 WARNING, WARN, WAR => LOG_WARNING
7492 NOTICE, NOTE, NOT => LOG_NOTICE
7493 INFO, INF => LOG_INFO
7494 DEBUG, DBG => LOG_DEBUG
7496 and as before, if none of those prefixes are present at the
7497 beginning of the string, LOG_ERR is chosen.
7499 On Win32, the LOG_* levels are mapped according to this:
7501 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7502 LOG_WARNING => EVENTLOG_WARNING_TYPE
7503 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7507 *) Made it possible to reconfigure with just the configuration
7508 argument "reconf" or "reconfigure". The command line arguments
7509 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7510 and are retrieved from there when reconfiguring.
7514 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7516 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7519 *) The obj_dat.pl script was messing up the sorting of object
7520 names. The reason was that it compared the quoted version
7521 of strings as a result "OCSP" > "OCSP Signing" because
7522 " > SPACE. Changed script to store unquoted versions of
7523 names and add quotes on output. It was also omitting some
7524 names from the lookup table if they were given a default
7525 value (that is if SN is missing it is given the same
7526 value as LN and vice versa), these are now added on the
7527 grounds that if an object has a name we should be able to
7528 look it up. Finally added warning output when duplicate
7529 short or long names are found.
7532 *) Changes needed for Tandem NSK.
7533 [Scott Uroff <scott@xypro.com>]
7535 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7536 RSA_padding_check_SSLv23(), special padding was never detected
7537 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7538 version rollback attacks was not effective.
7540 In s23_clnt.c, don't use special rollback-attack detection padding
7541 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7542 client; similarly, in s23_srvr.c, don't do the rollback check if
7543 SSL 2.0 is the only protocol enabled in the server.
7546 *) Make it possible to get hexdumps of unprintable data with 'openssl
7547 asn1parse'. By implication, the functions ASN1_parse_dump() and
7548 BIO_dump_indent() are added.
7551 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7552 these print out strings and name structures based on various
7553 flags including RFC2253 support and proper handling of
7554 multibyte characters. Added options to the 'x509' utility
7555 to allow the various flags to be set.
7558 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7559 Also change the functions X509_cmp_current_time() and
7560 X509_gmtime_adj() work with an ASN1_TIME structure,
7561 this will enable certificates using GeneralizedTime in validity
7562 dates to be checked.
7565 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7566 negative public key encodings) on by default,
7567 NO_NEG_PUBKEY_BUG can be set to disable it.
7570 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7571 content octets. An i2c_ASN1_OBJECT is unnecessary because
7572 the encoding can be trivially obtained from the structure.
7575 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7576 not read locks (CRYPTO_r_[un]lock).
7579 *) A first attempt at creating official support for shared
7580 libraries through configuration. I've kept it so the
7581 default is static libraries only, and the OpenSSL programs
7582 are always statically linked for now, but there are
7583 preparations for dynamic linking in place.
7584 This has been tested on Linux and Tru64.
7587 *) Randomness polling function for Win9x, as described in:
7588 Peter Gutmann, Software Generation of Practically Strong
7592 *) Fix so PRNG is seeded in req if using an already existing
7596 *) New options to smime application. -inform and -outform
7597 allow alternative formats for the S/MIME message including
7598 PEM and DER. The -content option allows the content to be
7599 specified separately. This should allow things like Netscape
7600 form signing output easier to verify.
7603 *) Fix the ASN1 encoding of tags using the 'long form'.
7606 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7607 STRING types. These convert content octets to and from the
7608 underlying type. The actual tag and length octets are
7609 already assumed to have been read in and checked. These
7610 are needed because all other string types have virtually
7611 identical handling apart from the tag. By having versions
7612 of the ASN1 functions that just operate on content octets
7613 IMPLICIT tagging can be handled properly. It also allows
7614 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7615 and ASN1_INTEGER are identical apart from the tag.
7618 *) Change the handling of OID objects as follows:
7620 - New object identifiers are inserted in objects.txt, following
7621 the syntax given in objects.README.
7622 - objects.pl is used to process obj_mac.num and create a new
7624 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7627 This is currently kind of a hack, and the perl code in objects.pl
7628 isn't very elegant, but it works as I intended. The simplest way
7629 to check that it worked correctly is to look in obj_dat.h and
7630 check the array nid_objs and make sure the objects haven't moved
7631 around (this is important!). Additions are OK, as well as
7632 consistent name changes.
7635 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7638 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7639 The given file adds to whatever has already been seeded into the
7640 random pool through the RANDFILE configuration file option or
7641 environment variable, or the default random state file.
7644 *) mkstack.pl now sorts each macro group into lexical order.
7645 Previously the output order depended on the order the files
7646 appeared in the directory, resulting in needless rewriting
7650 *) Patches to make OpenSSL compile under Win32 again. Mostly
7651 work arounds for the VC++ problem that it treats func() as
7652 func(void). Also stripped out the parts of mkdef.pl that
7653 added extra typesafe functions: these no longer exist.
7656 *) Reorganisation of the stack code. The macros are now all
7657 collected in safestack.h . Each macro is defined in terms of
7658 a "stack macro" of the form SKM_<name>(type, a, b). The
7659 DEBUG_SAFESTACK is now handled in terms of function casts,
7660 this has the advantage of retaining type safety without the
7661 use of additional functions. If DEBUG_SAFESTACK is not defined
7662 then the non typesafe macros are used instead. Also modified the
7663 mkstack.pl script to handle the new form. Needs testing to see
7664 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7665 the default if no major problems. Similar behaviour for ASN1_SET_OF
7666 and PKCS12_STACK_OF.
7669 *) When some versions of IIS use the 'NET' form of private key the
7670 key derivation algorithm is different. Normally MD5(password) is
7671 used as a 128 bit RC4 key. In the modified case
7672 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7673 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7674 as the old Netscape_RSA functions except they have an additional
7675 'sgckey' parameter which uses the modified algorithm. Also added
7676 an -sgckey command line option to the rsa utility. Thanks to
7677 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7678 algorithm to openssl-dev.
7681 *) The evp_local.h macros were using 'c.##kname' which resulted in
7682 invalid expansion on some systems (SCO 5.0.5 for example).
7683 Corrected to 'c.kname'.
7684 [Phillip Porch <root@theporch.com>]
7686 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7687 a STACK of email addresses from a certificate or request, these look
7688 in the subject name and the subject alternative name extensions and
7689 omit any duplicate addresses.
7692 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7693 This makes DSA verification about 2 % faster.
7696 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7697 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7698 plus overhead for 1024 bit moduli).
7699 This makes exponentiations about 0.5 % faster for 1024 bit
7700 exponents (as measured by "openssl speed rsa2048").
7703 *) Rename memory handling macros to avoid conflicts with other
7705 Malloc => OPENSSL_malloc
7706 Malloc_locked => OPENSSL_malloc_locked
7707 Realloc => OPENSSL_realloc
7708 Free => OPENSSL_free
7711 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7712 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7715 *) CygWin32 support.
7716 [John Jarvie <jjarvie@newsguy.com>]
7718 *) The type-safe stack code has been rejigged. It is now only compiled
7719 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7720 by default all type-specific stack functions are "#define"d back to
7721 standard stack functions. This results in more streamlined output
7722 but retains the type-safety checking possibilities of the original
7726 *) The STACK code has been cleaned up, and certain type declarations
7727 that didn't make a lot of sense have been brought in line. This has
7728 also involved a cleanup of sorts in safestack.h to more correctly
7729 map type-safe stack functions onto their plain stack counterparts.
7730 This work has also resulted in a variety of "const"ifications of
7731 lots of the code, especially "_cmp" operations which should normally
7732 be prototyped with "const" parameters anyway.
7735 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7736 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7737 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7738 where all of 'md' is used each time the PRNG is used, but 'state'
7739 is used only indexed by a cyclic counter. As entropy may not be
7740 well distributed from the beginning, 'md' is important as a
7741 chaining variable. However, the output function chains only half
7742 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7743 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7744 in all of 'state' being rewritten, with the new values depending
7745 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7748 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7749 the handshake is continued after ssl_verify_cert_chain();
7750 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7751 can lead to 'unexplainable' connection aborts later.
7752 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7754 *) Major EVP API cipher revision.
7755 Add hooks for extra EVP features. This allows various cipher
7756 parameters to be set in the EVP interface. Support added for variable
7757 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7758 setting of RC2 and RC5 parameters.
7760 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7763 Remove lots of duplicated code from the EVP library. For example *every*
7764 cipher init() function handles the 'iv' in the same way according to the
7765 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7766 for CFB and OFB modes they zero ctx->num.
7768 New functionality allows removal of S/MIME code RC2 hack.
7770 Most of the routines have the same form and so can be declared in terms
7773 By shifting this to the top level EVP_CipherInit() it can be removed from
7774 all individual ciphers. If the cipher wants to handle IVs or keys
7775 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7778 Change lots of functions like EVP_EncryptUpdate() to now return a
7779 value: although software versions of the algorithms cannot fail
7780 any installed hardware versions can.
7783 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7784 this option is set, tolerate broken clients that send the negotiated
7785 protocol version number instead of the requested protocol version
7789 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7790 i.e. non-zero for export ciphersuites, zero otherwise.
7791 Previous versions had this flag inverted, inconsistent with
7792 rsa_tmp_cb (..._TMP_RSA_CB).
7793 [Bodo Moeller; problem reported by Amit Chopra]
7795 *) Add missing DSA library text string. Work around for some IIS
7796 key files with invalid SEQUENCE encoding.
7799 *) Add a document (doc/standards.txt) that list all kinds of standards
7800 and so on that are implemented in OpenSSL.
7803 *) Enhance c_rehash script. Old version would mishandle certificates
7804 with the same subject name hash and wouldn't handle CRLs at all.
7805 Added -fingerprint option to crl utility, to support new c_rehash
7809 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7812 *) Fix for SSL server purpose checking. Server checking was
7813 rejecting certificates which had extended key usage present
7814 but no ssl client purpose.
7815 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7817 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7818 is a little unclear about how a blank password is handled.
7819 Since the password in encoded as a BMPString with terminating
7820 double NULL a zero length password would end up as just the
7821 double NULL. However no password at all is different and is
7822 handled differently in the PKCS#12 key generation code. NS
7823 treats a blank password as zero length. MSIE treats it as no
7824 password on export: but it will try both on import. We now do
7825 the same: PKCS12_parse() tries zero length and no password if
7826 the password is set to "" or NULL (NULL is now a valid password:
7827 it wasn't before) as does the pkcs12 application.
7830 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7831 perror when PEM_read_bio_X509_REQ fails, the error message must
7832 be obtained from the error queue.
7835 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7836 it in ERR_remove_state if appropriate, and change ERR_get_state
7837 accordingly to avoid race conditions (this is necessary because
7838 thread_hash is no longer constant once set).
7841 *) Bugfix for linux-elf makefile.one.
7844 *) RSA_get_default_method() will now cause a default
7845 RSA_METHOD to be chosen if one doesn't exist already.
7846 Previously this was only set during a call to RSA_new()
7847 or RSA_new_method(NULL) meaning it was possible for
7848 RSA_get_default_method() to return NULL.
7851 *) Added native name translation to the existing DSO code
7852 that will convert (if the flag to do so is set) filenames
7853 that are sufficiently small and have no path information
7854 into a canonical native form. Eg. "blah" converted to
7855 "libblah.so" or "blah.dll" etc.
7858 *) New function ERR_error_string_n(e, buf, len) which is like
7859 ERR_error_string(e, buf), but writes at most 'len' bytes
7860 including the 0 terminator. For ERR_error_string_n, 'buf'
7862 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7864 *) CONF library reworked to become more general. A new CONF
7865 configuration file reader "class" is implemented as well as a
7866 new functions (NCONF_*, for "New CONF") to handle it. The now
7867 old CONF_* functions are still there, but are reimplemented to
7868 work in terms of the new functions. Also, a set of functions
7869 to handle the internal storage of the configuration data is
7870 provided to make it easier to write new configuration file
7871 reader "classes" (I can definitely see something reading a
7872 configuration file in XML format, for example), called _CONF_*,
7873 or "the configuration storage API"...
7875 The new configuration file reading functions are:
7877 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7878 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7880 NCONF_default, NCONF_WIN32
7882 NCONF_dump_fp, NCONF_dump_bio
7884 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7885 NCONF_new creates a new CONF object. This works in the same way
7886 as other interfaces in OpenSSL, like the BIO interface.
7887 NCONF_dump_* dump the internal storage of the configuration file,
7888 which is useful for debugging. All other functions take the same
7889 arguments as the old CONF_* functions wth the exception of the
7890 first that must be a `CONF *' instead of a `LHASH *'.
7892 To make it easer to use the new classes with the old CONF_* functions,
7893 the function CONF_set_default_method is provided.
7896 *) Add '-tls1' option to 'openssl ciphers', which was already
7897 mentioned in the documentation but had not been implemented.
7898 (This option is not yet really useful because even the additional
7899 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7902 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7903 OpenSSL-based applications) load shared libraries and bind to
7904 them in a portable way.
7905 [Geoff Thorpe, with contributions from Richard Levitte]
7907 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7909 *) Make sure _lrotl and _lrotr are only used with MSVC.
7911 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7912 (the default implementation of RAND_status).
7914 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7915 to '-clrext' (= clear extensions), as intended and documented.
7916 [Bodo Moeller; inconsistency pointed out by Michael Attili
7919 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7920 was larger than the MD block size.
7921 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7923 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7924 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7925 using the passed key: if the passed key was a private key the result
7926 of X509_print(), for example, would be to print out all the private key
7930 *) des_quad_cksum() byte order bug fix.
7931 [Ulf Möller, using the problem description in krb4-0.9.7, where
7932 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7934 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7936 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7938 *) For easily testing in shell scripts whether some command
7939 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7940 returns with exit code 0 iff no command of the given name is available.
7941 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7942 the output goes to stdout and nothing is printed to stderr.
7943 Additional arguments are always ignored.
7945 Since for each cipher there is a command of the same name,
7946 the 'no-cipher' compilation switches can be tested this way.
7948 ('openssl no-XXX' is not able to detect pseudo-commands such
7949 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7952 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7955 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7956 is set; it will be thrown away anyway because each handshake creates
7958 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7959 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7960 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7961 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7964 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7965 'Q' and 'R' lose their special meanings (quit/renegotiate).
7966 This is part of what -quiet does; unlike -quiet, -ign_eof
7967 does not suppress any output.
7970 *) Add compatibility options to the purpose and trust code. The
7971 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7972 accepts a certificate or CA, this was the previous behaviour,
7973 with all the associated security issues.
7975 X509_TRUST_COMPAT is the old trust behaviour: only and
7976 automatically trust self signed roots in certificate store. A
7977 new trust setting X509_TRUST_DEFAULT is used to specify that
7978 a purpose has no associated trust setting and it should instead
7979 use the value in the default purpose.
7982 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7983 and fix a memory leak.
7986 *) In util/mkerr.pl (which implements 'make errors'), preserve
7987 reason strings from the previous version of the .c file, as
7988 the default to have only downcase letters (and digits) in
7989 automatically generated reasons codes is not always appropriate.
7992 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7993 using strerror. Previously, ERR_reason_error_string() returned
7994 library names as reason strings for SYSerr; but SYSerr is a special
7995 case where small numbers are errno values, not library numbers.
7998 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7999 converts DSA parameters into DH parameters. (When creating parameters,
8000 DSA_generate_parameters is used.)
8003 *) Include 'length' (recommended exponent length) in C code generated
8004 by 'openssl dhparam -C'.
8007 *) The second argument to set_label in perlasm was already being used
8008 so couldn't be used as a "file scope" flag. Moved to third argument
8012 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8013 instead of RAND_bytes for encryption IVs and salts.
8016 *) Include RAND_status() into RAND_METHOD instead of implementing
8017 it only for md_rand.c Otherwise replacing the PRNG by calling
8018 RAND_set_rand_method would be impossible.
8021 *) Don't let DSA_generate_key() enter an infinite loop if the random
8022 number generation fails.
8025 *) New 'rand' application for creating pseudo-random output.
8028 *) Added configuration support for Linux/IA64
8029 [Rolf Haberrecker <rolf@suse.de>]
8031 *) Assembler module support for Mingw32.
8034 *) Shared library support for HPUX (in shlib/).
8035 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8037 *) Shared library support for Solaris gcc.
8038 [Lutz Behnke <behnke@trustcenter.de>]
8040 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8042 *) PKCS7_encrypt() was adding text MIME headers twice because they
8043 were added manually and by SMIME_crlf_copy().
8046 *) In bntest.c don't call BN_rand with zero bits argument.
8047 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8049 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8050 case was implemented. This caused BN_div_recp() to fail occasionally.
8053 *) Add an optional second argument to the set_label() in the perl
8054 assembly language builder. If this argument exists and is set
8055 to 1 it signals that the assembler should use a symbol whose
8056 scope is the entire file, not just the current function. This
8057 is needed with MASM which uses the format label:: for this scope.
8058 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8060 *) Change the ASN1 types so they are typedefs by default. Before
8061 almost all types were #define'd to ASN1_STRING which was causing
8062 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8066 *) Change names of new functions to the new get1/get0 naming
8067 convention: After 'get1', the caller owns a reference count
8068 and has to call ..._free; 'get0' returns a pointer to some
8069 data structure without incrementing reference counters.
8070 (Some of the existing 'get' functions increment a reference
8071 counter, some don't.)
8072 Similarly, 'set1' and 'add1' functions increase reference
8073 counters or duplicate objects.
8076 *) Allow for the possibility of temp RSA key generation failure:
8077 the code used to assume it always worked and crashed on failure.
8080 *) Fix potential buffer overrun problem in BIO_printf().
8081 [Ulf Möller, using public domain code by Patrick Powell; problem
8082 pointed out by David Sacerdote <das33@cornell.edu>]
8084 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8085 RAND_egd() and RAND_status(). In the command line application,
8086 the EGD socket can be specified like a seed file using RANDFILE
8090 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8091 Some CAs (e.g. Verisign) distribute certificates in this form.
8094 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8095 list to exclude them. This means that no special compilation option
8096 is needed to use anonymous DH: it just needs to be included in the
8100 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8101 EVP_MD_type. The old functionality is available in a new macro called
8102 EVP_MD_md(). Change code that uses it and update docs.
8105 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8106 where the 'void *' argument is replaced by a function pointer argument.
8107 Previously 'void *' was abused to point to functions, which works on
8108 many platforms, but is not correct. As these functions are usually
8109 called by macros defined in OpenSSL header files, most source code
8110 should work without changes.
8113 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8114 sections with information on -D... compiler switches used for
8115 compiling the library so that applications can see them. To enable
8116 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8117 must be defined. E.g.,
8118 #define OPENSSL_ALGORITHM_DEFINES
8119 #include <openssl/opensslconf.h>
8120 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8121 [Richard Levitte, Ulf and Bodo Möller]
8123 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8127 *) Change the 'other' type in certificate aux info to a STACK_OF
8128 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8129 the required ASN1 format: arbitrary types determined by an OID.
8132 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8133 argument to 'req'. This is not because the function is newer or
8134 better than others it just uses the work 'NEW' in the certificate
8135 request header lines. Some software needs this.
8138 *) Reorganise password command line arguments: now passwords can be
8139 obtained from various sources. Delete the PEM_cb function and make
8140 it the default behaviour: i.e. if the callback is NULL and the
8141 usrdata argument is not NULL interpret it as a null terminated pass
8142 phrase. If usrdata and the callback are NULL then the pass phrase
8143 is prompted for as usual.
8146 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8147 the support is automatically enabled. The resulting binaries will
8148 autodetect the card and use it if present.
8149 [Ben Laurie and Compaq Inc.]
8151 *) Work around for Netscape hang bug. This sends certificate request
8152 and server done in one record. Since this is perfectly legal in the
8153 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8154 the bugs/SSLv3 entry for more info.
8157 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8160 *) Add -rand argument to smime and pkcs12 applications and read/write
8164 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8167 *) Add command line password options to the remaining applications.
8170 *) Bug fix for BN_div_recp() for numerators with an even number of
8174 *) More tests in bntest.c, and changed test_bn output.
8177 *) ./config recognizes MacOS X now.
8180 *) Bug fix for BN_div() when the first words of num and divsor are
8181 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8184 *) Add support for various broken PKCS#8 formats, and command line
8185 options to produce them.
8188 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8189 get temporary BIGNUMs from a BN_CTX.
8192 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8196 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8197 include a #define from the old name to the new. The original intent
8198 was that statically linked binaries could for example just call
8199 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8200 link with digests. This never worked becayse SSLeay_add_all_digests()
8201 and SSLeay_add_all_ciphers() were in the same source file so calling
8202 one would link with the other. They are now in separate source files.
8205 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8208 *) Use a less unusual form of the Miller-Rabin primality test (it used
8209 a binary algorithm for exponentiation integrated into the Miller-Rabin
8210 loop, our standard modexp algorithms are faster).
8213 *) Support for the EBCDIC character set completed.
8214 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8216 *) Source code cleanups: use const where appropriate, eliminate casts,
8217 use void * instead of char * in lhash.
8220 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8221 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8222 this the server could overwrite ephemeral keys that the client
8226 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8227 using 50 iterations of the Rabin-Miller test.
8229 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8230 iterations of the Rabin-Miller test as required by the appendix
8231 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8232 As BN_is_prime_fasttest includes trial division, DSA parameter
8233 generation becomes much faster.
8235 This implies a change for the callback functions in DSA_is_prime
8236 and DSA_generate_parameters: The callback function is called once
8237 for each positive witness in the Rabin-Miller test, not just
8238 occasionally in the inner loop; and the parameters to the
8239 callback function now provide an iteration count for the outer
8240 loop rather than for the current invocation of the inner loop.
8241 DSA_generate_parameters additionally can call the callback
8242 function with an 'iteration count' of -1, meaning that a
8243 candidate has passed the trial division test (when q is generated
8244 from an application-provided seed, trial division is skipped).
8247 *) New function BN_is_prime_fasttest that optionally does trial
8248 division before starting the Rabin-Miller test and has
8249 an additional BN_CTX * argument (whereas BN_is_prime always
8250 has to allocate at least one BN_CTX).
8251 'callback(1, -1, cb_arg)' is called when a number has passed the
8252 trial division stage.
8255 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8259 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8262 *) New function BN_pseudo_rand().
8265 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8266 bignum version of BN_from_montgomery() with the working code from
8267 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8271 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8272 made it impossible to use the same SSL_SESSION data structure in
8273 SSL2 clients in multiple threads.
8276 *) The return value of RAND_load_file() no longer counts bytes obtained
8277 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8278 to seed the PRNG (previously an explicit byte count was required).
8279 [Ulf Möller, Bodo Möller]
8281 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8282 used (char *) instead of (void *) and had casts all over the place.
8285 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8288 *) Retain source code compatibility for BN_prime_checks macro:
8289 BN_is_prime(..., BN_prime_checks, ...) now uses
8290 BN_prime_checks_for_size to determine the appropriate number of
8291 Rabin-Miller iterations.
8294 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8295 DH_CHECK_P_NOT_SAFE_PRIME.
8296 (Check if this is true? OpenPGP calls them "strong".)
8299 *) Merge the functionality of "dh" and "gendh" programs into a new program
8300 "dhparam". The old programs are retained for now but will handle DH keys
8301 (instead of parameters) in future.
8304 *) Make the ciphers, s_server and s_client programs check the return values
8305 when a new cipher list is set.
8308 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8309 ciphers. Before when the 56bit ciphers were enabled the sorting was
8312 The syntax for the cipher sorting has been extended to support sorting by
8313 cipher-strength (using the strength_bits hard coded in the tables).
8314 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8316 Fix a bug in the cipher-command parser: when supplying a cipher command
8317 string with an "undefined" symbol (neither command nor alphanumeric
8318 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8319 an error is flagged.
8321 Due to the strength-sorting extension, the code of the
8322 ssl_create_cipher_list() function was completely rearranged. I hope that
8323 the readability was also increased :-)
8324 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8326 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8327 for the first serial number and places 2 in the serial number file. This
8328 avoids problems when the root CA is created with serial number zero and
8329 the first user certificate has the same issuer name and serial number
8333 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8334 the new code. Add documentation for this stuff.
8337 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8338 X509_*() to X509at_*() on the grounds that they don't handle X509
8339 structures and behave in an analagous way to the X509v3 functions:
8340 they shouldn't be called directly but wrapper functions should be used
8343 So we also now have some wrapper functions that call the X509at functions
8344 when passed certificate requests. (TO DO: similar things can be done with
8345 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8346 things. Some of these need some d2i or i2d and print functionality
8347 because they handle more complex structures.)
8350 *) Add missing #ifndefs that caused missing symbols when building libssl
8351 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8352 NO_RSA in ssl/s2*.c.
8353 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8355 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8356 has a return value which indicates the quality of the random data
8357 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8358 error queue. New function RAND_pseudo_bytes() generates output that is
8359 guaranteed to be unique but not unpredictable. RAND_add is like
8360 RAND_seed, but takes an extra argument for an entropy estimate
8361 (RAND_seed always assumes full entropy).
8364 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8365 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8366 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8367 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8368 false-positive rate of at most 2^-80 for random input.
8371 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8374 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8375 in the 0.9.5 release), this returns the chain
8376 from an X509_CTX structure with a dup of the stack and all
8377 the X509 reference counts upped: so the stack will exist
8378 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8381 Also make SSL_SESSION_print() print out the verify return
8385 *) Add manpage for the pkcs12 command. Also change the default
8386 behaviour so MAC iteration counts are used unless the new
8387 -nomaciter option is used. This improves file security and
8388 only older versions of MSIE (4.0 for example) need it.
8391 *) Honor the no-xxx Configure options when creating .DEF files.
8394 *) Add PKCS#10 attributes to field table: challengePassword,
8395 unstructuredName and unstructuredAddress. These are taken from
8396 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8397 international characters are used.
8399 More changes to X509_ATTRIBUTE code: allow the setting of types
8400 based on strings. Remove the 'loc' parameter when adding
8401 attributes because these will be a SET OF encoding which is sorted
8405 *) Initial changes to the 'req' utility to allow request generation
8406 automation. This will allow an application to just generate a template
8407 file containing all the field values and have req construct the
8410 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8411 used all over the place including certificate requests and PKCS#7
8412 structures. They are currently handled manually where necessary with
8413 some primitive wrappers for PKCS#7. The new functions behave in a
8414 manner analogous to the X509 extension functions: they allow
8415 attributes to be looked up by NID and added.
8417 Later something similar to the X509V3 code would be desirable to
8418 automatically handle the encoding, decoding and printing of the
8419 more complex types. The string types like challengePassword can
8420 be handled by the string table functions.
8422 Also modified the multi byte string table handling. Now there is
8423 a 'global mask' which masks out certain types. The table itself
8424 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8425 is useful when for example there is only one permissible type
8426 (as in countryName) and using the mask might result in no valid
8430 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8431 SSL_get_peer_finished to allow applications to obtain the latest
8432 Finished messages sent to the peer or expected from the peer,
8433 respectively. (SSL_get_peer_finished is usually the Finished message
8434 actually received from the peer, otherwise the protocol will be aborted.)
8436 As the Finished message are message digests of the complete handshake
8437 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8438 be used for external authentication procedures when the authentication
8439 provided by SSL/TLS is not desired or is not enough.
8442 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8443 the host supports BWX extension and if Compaq C is present on the
8444 $PATH. Just exploiting of the BWX extension results in 20-30%
8445 performance kick for some algorithms, e.g. DES and RC4 to mention
8446 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8450 *) Add support for MS "fast SGC". This is arguably a violation of the
8451 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8452 weak crypto and after checking the certificate is SGC a second one
8453 with strong crypto. MS SGC stops the first handshake after receiving
8454 the server certificate message and sends a second client hello. Since
8455 a server will typically do all the time consuming operations before
8456 expecting any further messages from the client (server key exchange
8457 is the most expensive) there is little difference between the two.
8459 To get OpenSSL to support MS SGC we have to permit a second client
8460 hello message after we have sent server done. In addition we have to
8461 reset the MAC if we do get this second client hello.
8464 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8465 if a DER encoded private key is RSA or DSA traditional format. Changed
8466 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8467 format DER encoded private key. Newer code should use PKCS#8 format which
8468 has the key type encoded in the ASN1 structure. Added DER private key
8469 support to pkcs8 application.
8472 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8473 ciphersuites has been selected (as required by the SSL 3/TLS 1
8474 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8475 is set, we interpret this as a request to violate the specification
8476 (the worst that can happen is a handshake failure, and 'correct'
8477 behaviour would result in a handshake failure anyway).
8480 *) In SSL_CTX_add_session, take into account that there might be multiple
8481 SSL_SESSION structures with the same session ID (e.g. when two threads
8482 concurrently obtain them from an external cache).
8483 The internal cache can handle only one SSL_SESSION with a given ID,
8484 so if there's a conflict, we now throw out the old one to achieve
8488 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8489 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8490 some routines that use cipher OIDs: some ciphers do not have OIDs
8491 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8495 *) Simplify the trust setting structure and code. Now we just have
8496 two sequences of OIDs for trusted and rejected settings. These will
8497 typically have values the same as the extended key usage extension
8498 and any application specific purposes.
8500 The trust checking code now has a default behaviour: it will just
8501 check for an object with the same NID as the passed id. Functions can
8502 be provided to override either the default behaviour or the behaviour
8503 for a given id. SSL client, server and email already have functions
8504 in place for compatibility: they check the NID and also return "trusted"
8505 if the certificate is self signed.
8508 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8509 traditional format into an EVP_PKEY structure.
8512 *) Add a password callback function PEM_cb() which either prompts for
8513 a password if usr_data is NULL or otherwise assumes it is a null
8514 terminated password. Allow passwords to be passed on command line
8515 environment or config files in a few more utilities.
8518 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8519 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8520 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8521 Update documentation.
8524 *) Support for ASN1 "NULL" type. This could be handled before by using
8525 ASN1_TYPE but there wasn't any function that would try to read a NULL
8526 and produce an error if it couldn't. For compatibility we also have
8527 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8528 don't allocate anything because they don't need to.
8531 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8533 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8535 *) Rebuild of the memory allocation routines used by OpenSSL code and
8536 possibly others as well. The purpose is to make an interface that
8537 provide hooks so anyone can build a separate set of allocation and
8538 deallocation routines to be used by OpenSSL, for example memory
8539 pool implementations, or something else, which was previously hard
8540 since Malloc(), Realloc() and Free() were defined as macros having
8541 the values malloc, realloc and free, respectively (except for Win32
8542 compilations). The same is provided for memory debugging code.
8543 OpenSSL already comes with functionality to find memory leaks, but
8544 this gives people a chance to debug other memory problems.
8546 With these changes, a new set of functions and macros have appeared:
8548 CRYPTO_set_mem_debug_functions() [F]
8549 CRYPTO_get_mem_debug_functions() [F]
8550 CRYPTO_dbg_set_options() [F]
8551 CRYPTO_dbg_get_options() [F]
8552 CRYPTO_malloc_debug_init() [M]
8554 The memory debug functions are NULL by default, unless the library
8555 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8556 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8557 gives the standard debugging functions that come with OpenSSL) or
8558 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8559 provided by the library user) must be used. When the standard
8560 debugging functions are used, CRYPTO_dbg_set_options can be used to
8561 request additional information:
8562 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8563 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8565 Also, things like CRYPTO_set_mem_functions will always give the
8566 expected result (the new set of functions is used for allocation
8567 and deallocation) at all times, regardless of platform and compiler
8570 To finish it up, some functions that were never use in any other
8571 way than through macros have a new API and new semantic:
8574 CRYPTO_dbg_realloc()
8577 All macros of value have retained their old syntax.
8578 [Richard Levitte and Bodo Moeller]
8580 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8581 ordering of SMIMECapabilities wasn't in "strength order" and there
8582 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8586 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8587 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8588 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8590 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8591 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8592 functionality to handle multipart/signed properly) and a utility
8593 called 'smime' to call all this stuff. This is based on code I
8594 originally wrote for Celo who have kindly allowed it to be
8595 included in OpenSSL.
8598 *) Add variants des_set_key_checked and des_set_key_unchecked of
8599 des_set_key (aka des_key_sched). Global variable des_check_key
8600 decides which of these is called by des_set_key; this way
8601 des_check_key behaves as it always did, but applications and
8602 the library itself, which was buggy for des_check_key == 1,
8603 have a cleaner way to pick the version they need.
8606 *) New function PKCS12_newpass() which changes the password of a
8610 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8611 dynamic mix. In both cases the ids can be used as an index into the
8612 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8613 functions so they accept a list of the field values and the
8614 application doesn't need to directly manipulate the X509_TRUST
8618 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8622 *) Modify the way the V3 extension code looks up extensions. This now
8623 works in a similar way to the object code: we have some "standard"
8624 extensions in a static table which is searched with OBJ_bsearch()
8625 and the application can add dynamic ones if needed. The file
8626 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8627 updated whenever a new extension is added to the core code and kept
8628 in ext_nid order. There is a simple program 'tabtest.c' which checks
8629 this. New extensions are not added too often so this file can readily
8630 be maintained manually.
8632 There are two big advantages in doing things this way. The extensions
8633 can be looked up immediately and no longer need to be "added" using
8634 X509V3_add_standard_extensions(): this function now does nothing.
8635 [Side note: I get *lots* of email saying the extension code doesn't
8636 work because people forget to call this function]
8637 Also no dynamic allocation is done unless new extensions are added:
8638 so if we don't add custom extensions there is no need to call
8639 X509V3_EXT_cleanup().
8642 *) Modify enc utility's salting as follows: make salting the default. Add a
8643 magic header, so unsalted files fail gracefully instead of just decrypting
8644 to garbage. This is because not salting is a big security hole, so people
8645 should be discouraged from doing it.
8648 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8649 digest to be passed on the command line but it only used this
8650 parameter when signing a certificate. Modified so all relevant
8651 operations are affected by the digest parameter including the
8652 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8653 DSA key was used because it didn't fix the digest.
8656 *) Initial certificate chain verify code. Currently tests the untrusted
8657 certificates for consistency with the verify purpose (which is set
8658 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8660 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8661 this is because it will reject chains with invalid extensions whereas
8662 every previous version of OpenSSL and SSLeay made no checks at all.
8664 Trust code: checks the root CA for the relevant trust settings. Trust
8665 settings have an initial value consistent with the verify purpose: e.g.
8666 if the verify purpose is for SSL client use it expects the CA to be
8667 trusted for SSL client use. However the default value can be changed to
8668 permit custom trust settings: one example of this would be to only trust
8669 certificates from a specific "secure" set of CAs.
8671 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8672 which should be used for version portability: especially since the
8673 verify structure is likely to change more often now.
8675 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8676 to set them. If not set then assume SSL clients will verify SSL servers
8679 Two new options to the verify program: -untrusted allows a set of
8680 untrusted certificates to be passed in and -purpose which sets the
8681 intended purpose of the certificate. If a purpose is set then the
8682 new chain verify code is used to check extension consistency.
8685 *) Support for the authority information access extension.
8688 *) Modify RSA and DSA PEM read routines to transparently handle
8689 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8690 public keys in a format compatible with certificate
8691 SubjectPublicKeyInfo structures. Unfortunately there were already
8692 functions called *_PublicKey_* which used various odd formats so
8693 these are retained for compatibility: however the DSA variants were
8694 never in a public release so they have been deleted. Changed dsa/rsa
8695 utilities to handle the new format: note no releases ever handled public
8696 keys so we should be OK.
8698 The primary motivation for this change is to avoid the same fiasco
8699 that dogs private keys: there are several incompatible private key
8700 formats some of which are standard and some OpenSSL specific and
8701 require various evil hacks to allow partial transparent handling and
8702 even then it doesn't work with DER formats. Given the option anything
8703 other than PKCS#8 should be dumped: but the other formats have to
8704 stay in the name of compatibility.
8706 With public keys and the benefit of hindsight one standard format
8707 is used which works with EVP_PKEY, RSA or DSA structures: though
8708 it clearly returns an error if you try to read the wrong kind of key.
8710 Added a -pubkey option to the 'x509' utility to output the public key.
8711 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8712 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8713 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8714 that do the same as the EVP_PKEY_assign_*() except they up the
8715 reference count of the added key (they don't "swallow" the
8719 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8720 CRLs would fail if the file contained no certificates or no CRLs:
8721 added a new function to read in both types and return the number
8722 read: this means that if none are read it will be an error. The
8723 DER versions of the certificate and CRL reader would always fail
8724 because it isn't possible to mix certificates and CRLs in DER format
8725 without choking one or the other routine. Changed this to just read
8726 a certificate: this is the best we can do. Also modified the code
8727 in apps/verify.c to take notice of return codes: it was previously
8728 attempting to read in certificates from NULL pointers and ignoring
8729 any errors: this is one reason why the cert and CRL reader seemed
8730 to work. It doesn't check return codes from the default certificate
8731 routines: these may well fail if the certificates aren't installed.
8734 *) Code to support otherName option in GeneralName.
8737 *) First update to verify code. Change the verify utility
8738 so it warns if it is passed a self signed certificate:
8739 for consistency with the normal behaviour. X509_verify
8740 has been modified to it will now verify a self signed
8741 certificate if *exactly* the same certificate appears
8742 in the store: it was previously impossible to trust a
8743 single self signed certificate. This means that:
8744 openssl verify ss.pem
8745 now gives a warning about a self signed certificate but
8746 openssl verify -CAfile ss.pem ss.pem
8750 *) For servers, store verify_result in SSL_SESSION data structure
8751 (and add it to external session representation).
8752 This is needed when client certificate verifications fails,
8753 but an application-provided verification callback (set by
8754 SSL_CTX_set_cert_verify_callback) allows accepting the session
8755 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8756 but returns 1): When the session is reused, we have to set
8757 ssl->verify_result to the appropriate error code to avoid
8759 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8761 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8762 case in PKCS7_dataInit() where the signed PKCS7 structure
8763 didn't contain any existing data because it was being created.
8764 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8766 *) Add a salt to the key derivation routines in enc.c. This
8767 forms the first 8 bytes of the encrypted file. Also add a
8768 -S option to allow a salt to be input on the command line.
8771 *) New function X509_cmp(). Oddly enough there wasn't a function
8772 to compare two certificates. We do this by working out the SHA1
8773 hash and comparing that. X509_cmp() will be needed by the trust
8777 *) SSL_get1_session() is like SSL_get_session(), but increments
8778 the reference count in the SSL_SESSION returned.
8779 [Geoff Thorpe <geoff@eu.c2.net>]
8781 *) Fix for 'req': it was adding a null to request attributes.
8782 Also change the X509_LOOKUP and X509_INFO code to handle
8783 certificate auxiliary information.
8786 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8790 *) Add the possibility to add extra information to the memory leak
8791 detecting output, to form tracebacks, showing from where each
8792 allocation was originated: CRYPTO_push_info("constant string") adds
8793 the string plus current file name and line number to a per-thread
8794 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8795 is like calling CYRPTO_pop_info() until the stack is empty.
8796 Also updated memory leak detection code to be multi-thread-safe.
8799 *) Add options -text and -noout to pkcs7 utility and delete the
8800 encryption options which never did anything. Update docs.
8803 *) Add options to some of the utilities to allow the pass phrase
8804 to be included on either the command line (not recommended on
8805 OSes like Unix) or read from the environment. Update the
8806 manpages and fix a few bugs.
8809 *) Add a few manpages for some of the openssl commands.
8812 *) Fix the -revoke option in ca. It was freeing up memory twice,
8813 leaking and not finding already revoked certificates.
8816 *) Extensive changes to support certificate auxiliary information.
8817 This involves the use of X509_CERT_AUX structure and X509_AUX
8818 functions. An X509_AUX function such as PEM_read_X509_AUX()
8819 can still read in a certificate file in the usual way but it
8820 will also read in any additional "auxiliary information". By
8821 doing things this way a fair degree of compatibility can be
8822 retained: existing certificates can have this information added
8823 using the new 'x509' options.
8825 Current auxiliary information includes an "alias" and some trust
8826 settings. The trust settings will ultimately be used in enhanced
8827 certificate chain verification routines: currently a certificate
8828 can only be trusted if it is self signed and then it is trusted
8832 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8833 The problem was that one of the replacement routines had not been working
8834 since SSLeay releases. For now the offending routine has been replaced
8835 with non-optimised assembler. Even so, this now gives around 95%
8836 performance improvement for 1024 bit RSA signs.
8839 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8840 handling. Most clients have the effective key size in bits equal to
8841 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8842 A few however don't do this and instead use the size of the decrypted key
8843 to determine the RC2 key length and the AlgorithmIdentifier to determine
8844 the effective key length. In this case the effective key length can still
8845 be 40 bits but the key length can be 168 bits for example. This is fixed
8846 by manually forcing an RC2 key into the EVP_PKEY structure because the
8847 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8848 the key length and effective key length are equal.
8851 *) Add a bunch of functions that should simplify the creation of
8852 X509_NAME structures. Now you should be able to do:
8853 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8854 and have it automatically work out the correct field type and fill in
8855 the structures. The more adventurous can try:
8856 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8857 and it will (hopefully) work out the correct multibyte encoding.
8860 *) Change the 'req' utility to use the new field handling and multibyte
8861 copy routines. Before the DN field creation was handled in an ad hoc
8862 way in req, ca, and x509 which was rather broken and didn't support
8863 BMPStrings or UTF8Strings. Since some software doesn't implement
8864 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8865 using the dirstring_type option. See the new comment in the default
8866 openssl.cnf for more info.
8869 *) Make crypto/rand/md_rand.c more robust:
8870 - Assure unique random numbers after fork().
8871 - Make sure that concurrent threads access the global counter and
8872 md serializably so that we never lose entropy in them
8873 or use exactly the same state in multiple threads.
8874 Access to the large state is not always serializable because
8875 the additional locking could be a performance killer, and
8876 md should be large enough anyway.
8879 *) New file apps/app_rand.c with commonly needed functionality
8880 for handling the random seed file.
8882 Use the random seed file in some applications that previously did not:
8884 dsaparam -genkey (which also ignored its '-rand' option),
8887 x509 (when signing).
8888 Except on systems with /dev/urandom, it is crucial to have a random
8889 seed file at least for key creation, DSA signing, and for DH exchanges;
8890 for RSA signatures we could do without one.
8892 gendh and gendsa (unlike genrsa) used to read only the first byte
8893 of each file listed in the '-rand' option. The function as previously
8894 found in genrsa is now in app_rand.c and is used by all programs
8895 that support '-rand'.
8898 *) In RAND_write_file, use mode 0600 for creating files;
8899 don't just chmod when it may be too late.
8902 *) Report an error from X509_STORE_load_locations
8903 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8906 *) New function ASN1_mbstring_copy() this copies a string in either
8907 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8908 into an ASN1_STRING type. A mask of permissible types is passed
8909 and it chooses the "minimal" type to use or an error if not type
8913 *) Add function equivalents to the various macros in asn1.h. The old
8914 macros are retained with an M_ prefix. Code inside the library can
8915 use the M_ macros. External code (including the openssl utility)
8916 should *NOT* in order to be "shared library friendly".
8919 *) Add various functions that can check a certificate's extensions
8920 to see if it usable for various purposes such as SSL client,
8921 server or S/MIME and CAs of these types. This is currently
8922 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8923 verification. Also added a -purpose flag to x509 utility to
8924 print out all the purposes.
8927 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8931 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8932 for, obtain and decode and extension and obtain its critical flag.
8933 This allows all the necessary extension code to be handled in a
8934 single function call.
8937 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8938 platforms. See crypto/rc4/rc4_enc.c for further details.
8941 *) New -noout option to asn1parse. This causes no output to be produced
8942 its main use is when combined with -strparse and -out to extract data
8943 from a file (which may not be in ASN.1 format).
8946 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8947 when producing the local key id.
8948 [Richard Levitte <levitte@stacken.kth.se>]
8950 *) New option -dhparam in s_server. This allows a DH parameter file to be
8951 stated explicitly. If it is not stated then it tries the first server
8952 certificate file. The previous behaviour hard coded the filename
8956 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8957 a public key to be input or output. For example:
8958 openssl rsa -in key.pem -pubout -out pubkey.pem
8959 Also added necessary DSA public key functions to handle this.
8962 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8963 in the message. This was handled by allowing
8964 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8965 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8967 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8968 to the end of the strings whereas this didn't. This would cause problems
8969 if strings read with d2i_ASN1_bytes() were later modified.
8970 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8972 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8973 data and it contains EOF it will end up returning an error. This is
8974 caused by input 46 bytes long. The cause is due to the way base64
8975 BIOs find the start of base64 encoded data. They do this by trying a
8976 trial decode on each line until they find one that works. When they
8977 do a flag is set and it starts again knowing it can pass all the
8978 data directly through the decoder. Unfortunately it doesn't reset
8979 the context it uses. This means that if EOF is reached an attempt
8980 is made to pass two EOFs through the context and this causes the
8981 resulting error. This can also cause other problems as well. As is
8982 usual with these problems it takes *ages* to find and the fix is
8983 trivial: move one line.
8984 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8986 *) Ugly workaround to get s_client and s_server working under Windows. The
8987 old code wouldn't work because it needed to select() on sockets and the
8988 tty (for keypresses and to see if data could be written). Win32 only
8989 supports select() on sockets so we select() with a 1s timeout on the
8990 sockets and then see if any characters are waiting to be read, if none
8991 are present then we retry, we also assume we can always write data to
8992 the tty. This isn't nice because the code then blocks until we've
8993 received a complete line of data and it is effectively polling the
8994 keyboard at 1s intervals: however it's quite a bit better than not
8995 working at all :-) A dedicated Windows application might handle this
8996 with an event loop for example.
8999 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9000 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9001 will be called when RSA_sign() and RSA_verify() are used. This is useful
9002 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9003 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9004 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9005 This necessitated the support of an extra signature type NID_md5_sha1
9006 for SSL signatures and modifications to the SSL library to use it instead
9007 of calling RSA_public_decrypt() and RSA_private_encrypt().
9010 *) Add new -verify -CAfile and -CApath options to the crl program, these
9011 will lookup a CRL issuers certificate and verify the signature in a
9012 similar way to the verify program. Tidy up the crl program so it
9013 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9014 less strict. It will now permit CRL extensions even if it is not
9015 a V2 CRL: this will allow it to tolerate some broken CRLs.
9018 *) Initialize all non-automatic variables each time one of the openssl
9019 sub-programs is started (this is necessary as they may be started
9020 multiple times from the "OpenSSL>" prompt).
9021 [Lennart Bang, Bodo Moeller]
9023 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9024 removing all other RSA functionality (this is what NO_RSA does). This
9025 is so (for example) those in the US can disable those operations covered
9026 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9030 *) Non-copying interface to BIO pairs.
9031 (still largely untested)
9034 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9035 ASCII string. This was handled independently in various places before.
9038 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9039 UTF8 strings a character at a time.
9042 *) Use client_version from client hello to select the protocol
9043 (s23_srvr.c) and for RSA client key exchange verification
9044 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9047 *) Add various utility functions to handle SPKACs, these were previously
9048 handled by poking round in the structure internals. Added new function
9049 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9050 print, verify and generate SPKACs. Based on an original idea from
9051 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9054 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9057 *) Allow the config file extension section to be overwritten on the
9058 command line. Based on an original idea from Massimiliano Pala
9059 <madwolf@comune.modena.it>. The new option is called -extensions
9060 and can be applied to ca, req and x509. Also -reqexts to override
9061 the request extensions in req and -crlexts to override the crl extensions
9065 *) Add new feature to the SPKAC handling in ca. Now you can include
9066 the same field multiple times by preceding it by "XXXX." for example:
9069 this is the same syntax as used in the req config file.
9072 *) Allow certificate extensions to be added to certificate requests. These
9073 are specified in a 'req_extensions' option of the req section of the
9074 config file. They can be printed out with the -text option to req but
9075 are otherwise ignored at present.
9078 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9079 data read consists of only the final block it would not decrypted because
9080 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9081 A misplaced 'break' also meant the decrypted final block might not be
9082 copied until the next read.
9085 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9086 a few extra parameters to the DH structure: these will be useful if
9087 for example we want the value of 'q' or implement X9.42 DH.
9090 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9091 provides hooks that allow the default DSA functions or functions on a
9092 "per key" basis to be replaced. This allows hardware acceleration and
9093 hardware key storage to be handled without major modification to the
9094 library. Also added low level modexp hooks and CRYPTO_EX structure and
9095 associated functions.
9098 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9099 as "read only": it can't be written to and the buffer it points to will
9100 not be freed. Reading from a read only BIO is much more efficient than
9101 a normal memory BIO. This was added because there are several times when
9102 an area of memory needs to be read from a BIO. The previous method was
9103 to create a memory BIO and write the data to it, this results in two
9104 copies of the data and an O(n^2) reading algorithm. There is a new
9105 function BIO_new_mem_buf() which creates a read only memory BIO from
9106 an area of memory. Also modified the PKCS#7 routines to use read only
9110 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9111 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9112 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9113 but a retry condition occured while trying to read the rest.
9116 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9117 NID_pkcs7_encrypted by default: this was wrong since this should almost
9118 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9119 the encrypted data type: this is a more sensible place to put it and it
9120 allows the PKCS#12 code to be tidied up that duplicated this
9124 *) Changed obj_dat.pl script so it takes its input and output files on
9125 the command line. This should avoid shell escape redirection problems
9129 *) Initial support for certificate extension requests, these are included
9130 in things like Xenroll certificate requests. Included functions to allow
9131 extensions to be obtained and added.
9134 *) -crlf option to s_client and s_server for sending newlines as
9135 CRLF (as required by many protocols).
9138 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9140 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9141 [Ralf S. Engelschall]
9143 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9144 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9146 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9150 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9151 DH parameters/keys (q is lost during that conversion, but the resulting
9152 DH parameters contain its length).
9154 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9155 much faster than DH_generate_parameters (which creates parameters
9156 where p = 2*q + 1), and also the smaller q makes DH computations
9157 much more efficient (160-bit exponentiation instead of 1024-bit
9158 exponentiation); so this provides a convenient way to support DHE
9159 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9160 utter importance to use
9161 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9163 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9164 when such DH parameters are used, because otherwise small subgroup
9165 attacks may become possible!
9168 *) Avoid memory leak in i2d_DHparams.
9171 *) Allow the -k option to be used more than once in the enc program:
9172 this allows the same encrypted message to be read by multiple recipients.
9175 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9176 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9177 it will always use the numerical form of the OID, even if it has a short
9181 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9182 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9183 otherwise bn_mod_exp was called. In the case of hardware keys for example
9184 no private key components need be present and it might store extra data
9185 in the RSA structure, which cannot be accessed from bn_mod_exp.
9186 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9187 private key operations.
9190 *) Added support for SPARC Linux.
9193 *) pem_password_cb function type incompatibly changed from
9194 typedef int pem_password_cb(char *buf, int size, int rwflag);
9196 ....(char *buf, int size, int rwflag, void *userdata);
9197 so that applications can pass data to their callbacks:
9198 The PEM[_ASN1]_{read,write}... functions and macros now take an
9199 additional void * argument, which is just handed through whenever
9200 the password callback is called.
9201 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9203 New function SSL_CTX_set_default_passwd_cb_userdata.
9205 Compatibility note: As many C implementations push function arguments
9206 onto the stack in reverse order, the new library version is likely to
9207 interoperate with programs that have been compiled with the old
9208 pem_password_cb definition (PEM_whatever takes some data that
9209 happens to be on the stack as its last argument, and the callback
9210 just ignores this garbage); but there is no guarantee whatsoever that
9213 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9214 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9215 problems not only on Windows, but also on some Unix platforms.
9216 To avoid problematic command lines, these definitions are now in an
9217 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9218 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9221 *) MIPS III/IV assembler module is reimplemented.
9224 *) More DES library cleanups: remove references to srand/rand and
9225 delete an unused file.
9228 *) Add support for the the free Netwide assembler (NASM) under Win32,
9229 since not many people have MASM (ml) and it can be hard to obtain.
9230 This is currently experimental but it seems to work OK and pass all
9231 the tests. Check out INSTALL.W32 for info.
9234 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9235 without temporary keys kept an extra copy of the server key,
9236 and connections with temporary keys did not free everything in case
9240 *) New function RSA_check_key and new openssl rsa option -check
9241 for verifying the consistency of RSA keys.
9242 [Ulf Moeller, Bodo Moeller]
9244 *) Various changes to make Win32 compile work:
9245 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9246 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9247 comparison" warnings.
9248 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9251 *) Add a debugging option to PKCS#5 v2 key generation function: when
9252 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9253 derived keys are printed to stderr.
9256 *) Copy the flags in ASN1_STRING_dup().
9257 [Roman E. Pavlov <pre@mo.msk.ru>]
9259 *) The x509 application mishandled signing requests containing DSA
9260 keys when the signing key was also DSA and the parameters didn't match.
9262 It was supposed to omit the parameters when they matched the signing key:
9263 the verifying software was then supposed to automatically use the CA's
9264 parameters if they were absent from the end user certificate.
9266 Omitting parameters is no longer recommended. The test was also
9267 the wrong way round! This was probably due to unusual behaviour in
9268 EVP_cmp_parameters() which returns 1 if the parameters match.
9269 This meant that parameters were omitted when they *didn't* match and
9270 the certificate was useless. Certificates signed with 'ca' didn't have
9272 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9274 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9275 The interface is as follows:
9276 Applications can use
9277 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9278 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9279 "off" is now the default.
9280 The library internally uses
9281 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9282 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9283 to disable memory-checking temporarily.
9285 Some inconsistent states that previously were possible (and were
9286 even the default) are now avoided.
9288 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9289 with each memory chunk allocated; this is occasionally more helpful
9290 than just having a counter.
9292 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9294 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9298 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9299 which largely parallels "options", but is for changing API behaviour,
9300 whereas "options" are about protocol behaviour.
9301 Initial "mode" flags are:
9303 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9304 a single record has been written.
9305 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9306 retries use the same buffer location.
9307 (But all of the contents must be
9311 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9314 *) Fix problems with no-hmac etc.
9315 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9317 *) New functions RSA_get_default_method(), RSA_set_method() and
9318 RSA_get_method(). These allows replacement of RSA_METHODs without having
9319 to mess around with the internals of an RSA structure.
9322 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9323 Also really enable memory leak checks in openssl.c and in some
9325 [Chad C. Mulligan, Bodo Moeller]
9327 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9328 up the length of negative integers. This has now been simplified to just
9329 store the length when it is first determined and use it later, rather
9330 than trying to keep track of where data is copied and updating it to
9332 [Steve Henson, reported by Brien Wheeler
9333 <bwheeler@authentica-security.com>]
9335 *) Add a new function PKCS7_signatureVerify. This allows the verification
9336 of a PKCS#7 signature but with the signing certificate passed to the
9337 function itself. This contrasts with PKCS7_dataVerify which assumes the
9338 certificate is present in the PKCS#7 structure. This isn't always the
9339 case: certificates can be omitted from a PKCS#7 structure and be
9340 distributed by "out of band" means (such as a certificate database).
9343 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9344 function prototypes in pem.h, also change util/mkdef.pl to add the
9345 necessary function names.
9348 *) mk1mf.pl (used by Windows builds) did not properly read the
9349 options set by Configure in the top level Makefile, and Configure
9350 was not even able to write more than one option correctly.
9351 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9354 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9355 file to be loaded from a BIO or FILE pointer. The BIO version will
9356 for example allow memory BIOs to contain config info.
9359 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9360 Whoever hopes to achieve shared-library compatibility across versions
9361 must use this, not the compile-time macro.
9362 (Exercise 0.9.4: Which is the minimum library version required by
9364 Note: All this applies only to multi-threaded programs, others don't
9368 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9369 through a BIO pair triggered the default case, i.e.
9370 SSLerr(...,SSL_R_UNKNOWN_STATE).
9373 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9374 can use the SSL library even if none of the specific BIOs is
9378 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9379 for the encoded length.
9380 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9382 *) Add initial documentation of the X509V3 functions.
9385 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9386 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9387 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9388 secure PKCS#8 private key format with a high iteration count.
9391 *) Fix determination of Perl interpreter: A perl or perl5
9392 _directory_ in $PATH was also accepted as the interpreter.
9393 [Ralf S. Engelschall]
9395 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9396 wrong with it but it was very old and did things like calling
9397 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9401 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9402 to use the new extension code.
9405 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9406 with macros. This should make it easier to change their form, add extra
9407 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9411 *) Add to configuration table a new entry that can specify an alternative
9412 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9413 according to Mark Crispin <MRC@Panda.COM>.
9417 *) DES CBC did not update the IV. Weird.
9420 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9421 Changing the behaviour of the former might break existing programs --
9422 where IV updating is needed, des_ncbc_encrypt can be used.
9425 *) When bntest is run from "make test" it drives bc to check its
9426 calculations, as well as internally checking them. If an internal check
9427 fails, it needs to cause bc to give a non-zero result or make test carries
9428 on without noticing the failure. Fixed.
9431 *) DES library cleanups.
9434 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9435 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9436 ciphers. NOTE: although the key derivation function has been verified
9437 against some published test vectors it has not been extensively tested
9438 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9442 *) Instead of "mkdir -p", which is not fully portable, use new
9443 Perl script "util/mkdir-p.pl".
9446 *) Rewrite the way password based encryption (PBE) is handled. It used to
9447 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9448 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9449 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9450 the 'parameter' field of the AlgorithmIdentifier is passed to the
9451 underlying key generation function so it must do its own ASN1 parsing.
9452 This has also changed the EVP_PBE_CipherInit() function which now has a
9453 'parameter' argument instead of literal salt and iteration count values
9454 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9457 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9458 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9459 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9460 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9461 value was just used as a "magic string" and not used directly its
9462 value doesn't matter.
9465 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9469 *) "linux-sparc64" configuration (ultrapenguin).
9470 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9471 "linux-sparc" configuration.
9472 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9474 *) config now generates no-xxx options for missing ciphers.
9477 *) Support the EBCDIC character set (work in progress).
9478 File ebcdic.c not yet included because it has a different license.
9479 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9481 *) Support BS2000/OSD-POSIX.
9482 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9484 *) Make callbacks for key generation use void * instead of char *.
9487 *) Make S/MIME samples compile (not yet tested).
9490 *) Additional typesafe stacks.
9493 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9497 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9499 *) New configuration variant "sco5-gcc".
9501 *) Updated some demos.
9502 [Sean O Riordain, Wade Scholine]
9504 *) Add missing BIO_free at exit of pkcs12 application.
9507 *) Fix memory leak in conf.c.
9510 *) Updates for Win32 to assembler version of MD5.
9513 *) Set #! path to perl in apps/der_chop to where we found it
9514 instead of using a fixed path.
9517 *) SHA library changes for irix64-mips4-cc.
9520 *) Improvements for VMS support.
9524 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9526 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9527 This also avoids the problems with SC4.2 and unpatched SC5.
9528 [Andy Polyakov <appro@fy.chalmers.se>]
9530 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9531 These are required because of the typesafe stack would otherwise break
9532 existing code. If old code used a structure member which used to be STACK
9533 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9534 sk_num or sk_value it would produce an error because the num, data members
9535 are not present in STACK_OF. Now it just produces a warning. sk_set
9536 replaces the old method of assigning a value to sk_value
9537 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9538 that does this will no longer work (and should use sk_set instead) but
9539 this could be regarded as a "questionable" behaviour anyway.
9542 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9543 correctly handle encrypted S/MIME data.
9546 *) Change type of various DES function arguments from des_cblock
9547 (which means, in function argument declarations, pointer to char)
9548 to des_cblock * (meaning pointer to array with 8 char elements),
9549 which allows the compiler to do more typechecking; it was like
9550 that back in SSLeay, but with lots of ugly casts.
9552 Introduce new type const_des_cblock.
9555 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9556 problems: find RecipientInfo structure that matches recipient certificate
9557 and initialise the ASN1 structures properly based on passed cipher.
9560 *) Belatedly make the BN tests actually check the results.
9563 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9564 to and from BNs: it was completely broken. New compilation option
9565 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9566 key elements as negative integers.
9569 *) Reorganize and speed up MD5.
9570 [Andy Polyakov <appro@fy.chalmers.se>]
9573 [Richard Levitte <richard@levitte.org>]
9575 *) New option -out to asn1parse to allow the parsed structure to be
9576 output to a file. This is most useful when combined with the -strparse
9577 option to examine the output of things like OCTET STRINGS.
9580 *) Make SSL library a little more fool-proof by not requiring any longer
9581 that SSL_set_{accept,connect}_state be called before
9582 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9583 in many applications because usually everything *appeared* to work as
9584 intended anyway -- now it really works as intended).
9587 *) Move openssl.cnf out of lib/.
9590 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9591 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9592 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9593 [Ralf S. Engelschall]
9595 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9596 handle PKCS#7 enveloped data properly.
9597 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9599 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9600 copying pointers. The cert_st handling is changed by this in
9601 various ways (and thus what used to be known as ctx->default_cert
9602 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9603 any longer when s->cert does not give us what we need).
9604 ssl_cert_instantiate becomes obsolete by this change.
9605 As soon as we've got the new code right (possibly it already is?),
9606 we have solved a couple of bugs of the earlier code where s->cert
9607 was used as if it could not have been shared with other SSL structures.
9609 Note that using the SSL API in certain dirty ways now will result
9610 in different behaviour than observed with earlier library versions:
9611 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9612 does not influence s as it used to.
9614 In order to clean up things more thoroughly, inside SSL_SESSION
9615 we don't use CERT any longer, but a new structure SESS_CERT
9616 that holds per-session data (if available); currently, this is
9617 the peer's certificate chain and, for clients, the server's certificate
9618 and temporary key. CERT holds only those values that can have
9619 meaningful defaults in an SSL_CTX.
9622 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9623 from the internal representation. Various PKCS#7 fixes: remove some
9624 evil casts and set the enc_dig_alg field properly based on the signing
9628 *) Allow PKCS#12 password to be set from the command line or the
9629 environment. Let 'ca' get its config file name from the environment
9630 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9634 *) Allow certificate policies extension to use an IA5STRING for the
9635 organization field. This is contrary to the PKIX definition but
9636 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9640 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9641 without disallowing inline assembler and the like for non-pedantic builds.
9644 *) Support Borland C++ builder.
9645 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9650 *) SHA-1 cleanups and performance enhancements.
9651 [Andy Polyakov <appro@fy.chalmers.se>]
9653 *) Sparc v8plus assembler for the bignum library.
9654 [Andy Polyakov <appro@fy.chalmers.se>]
9656 *) Accept any -xxx and +xxx compiler options in Configure.
9659 *) Update HPUX configuration.
9662 *) Add missing sk_<type>_unshift() function to safestack.h
9663 [Ralf S. Engelschall]
9665 *) New function SSL_CTX_use_certificate_chain_file that sets the
9666 "extra_cert"s in addition to the certificate. (This makes sense
9667 only for "PEM" format files, as chains as a whole are not
9671 *) Support verify_depth from the SSL API.
9672 x509_vfy.c had what can be considered an off-by-one-error:
9673 Its depth (which was not part of the external interface)
9674 was actually counting the number of certificates in a chain;
9675 now it really counts the depth.
9678 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9679 instead of X509err, which often resulted in confusing error
9680 messages since the error codes are not globally unique
9681 (e.g. an alleged error in ssl3_accept when a certificate
9682 didn't match the private key).
9684 *) New function SSL_CTX_set_session_id_context that allows to set a default
9685 value (so that you don't need SSL_set_session_id_context for each
9686 connection using the SSL_CTX).
9689 *) OAEP decoding bug fix.
9692 *) Support INSTALL_PREFIX for package builders, as proposed by
9696 *) New Configure options "threads" and "no-threads". For systems
9697 where the proper compiler options are known (currently Solaris
9698 and Linux), "threads" is the default.
9701 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9704 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9705 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9706 such as /usr/local/bin.
9709 *) "make linux-shared" to build shared libraries.
9710 [Niels Poppe <niels@netbox.org>]
9712 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9715 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9716 extension adding in x509 utility.
9719 *) Remove NOPROTO sections and error code comments.
9722 *) Partial rewrite of the DEF file generator to now parse the ANSI
9726 *) New Configure options --prefix=DIR and --openssldir=DIR.
9729 *) Complete rewrite of the error code script(s). It is all now handled
9730 by one script at the top level which handles error code gathering,
9731 header rewriting and C source file generation. It should be much better
9732 than the old method: it now uses a modified version of Ulf's parser to
9733 read the ANSI prototypes in all header files (thus the old K&R definitions
9734 aren't needed for error creation any more) and do a better job of
9735 translating function codes into names. The old 'ASN1 error code imbedded
9736 in a comment' is no longer necessary and it doesn't use .err files which
9737 have now been deleted. Also the error code call doesn't have to appear all
9738 on one line (which resulted in some large lines...).
9741 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9744 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9745 0 (which usually indicates a closed connection), but continue reading.
9748 *) Fix some race conditions.
9751 *) Add support for CRL distribution points extension. Add Certificate
9752 Policies and CRL distribution points documentation.
9755 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9758 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9759 8 of keying material. Merlin has also confirmed interop with this fix
9760 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9761 [Merlin Hughes <merlin@baltimore.ie>]
9763 *) Fix lots of warnings.
9764 [Richard Levitte <levitte@stacken.kth.se>]
9766 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9767 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9768 [Richard Levitte <levitte@stacken.kth.se>]
9770 *) Fix problems with sizeof(long) == 8.
9771 [Andy Polyakov <appro@fy.chalmers.se>]
9773 *) Change functions to ANSI C.
9776 *) Fix typos in error codes.
9777 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9779 *) Remove defunct assembler files from Configure.
9782 *) SPARC v8 assembler BIGNUM implementation.
9783 [Andy Polyakov <appro@fy.chalmers.se>]
9785 *) Support for Certificate Policies extension: both print and set.
9786 Various additions to support the r2i method this uses.
9789 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9790 return a const string when you are expecting an allocated buffer.
9793 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9794 types DirectoryString and DisplayText.
9797 *) Add code to allow r2i extensions to access the configuration database,
9798 add an LHASH database driver and add several ctx helper functions.
9801 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9802 fail when they extended the size of a BIGNUM.
9805 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9806 support typesafe stack.
9809 *) Fix typo in SSL_[gs]et_options().
9810 [Nils Frostberg <nils@medcom.se>]
9812 *) Delete various functions and files that belonged to the (now obsolete)
9813 old X509V3 handling code.
9816 *) New Configure option "rsaref".
9819 *) Don't auto-generate pem.h.
9822 *) Introduce type-safe ASN.1 SETs.
9825 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9826 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9828 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9829 that links with OpenSSL (well at least cause lots of warnings), but fear
9830 not: the conversion is trivial, and it eliminates loads of evil casts. A
9831 few STACKed things have been converted already. Feel free to convert more.
9832 In the fullness of time, I'll do away with the STACK type altogether.
9835 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9836 specified in <certfile> by updating the entry in the index.txt file.
9837 This way one no longer has to edit the index.txt file manually for
9838 revoking a certificate. The -revoke option does the gory details now.
9839 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9841 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9842 `-text' option at all and this way the `-noout -text' combination was
9843 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9844 [Ralf S. Engelschall]
9846 *) Make sure a corresponding plain text error message exists for the
9847 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9848 verify callback function determined that a certificate was revoked.
9849 [Ralf S. Engelschall]
9851 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9852 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9853 all available cipers including rc5, which was forgotten until now.
9854 In order to let the testing shell script know which algorithms
9855 are available, a new (up to now undocumented) command
9856 "openssl list-cipher-commands" is used.
9859 *) Bugfix: s_client occasionally would sleep in select() when
9860 it should have checked SSL_pending() first.
9863 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9864 the raw DSA values prior to ASN.1 encoding.
9867 *) Tweaks to Configure
9868 [Niels Poppe <niels@netbox.org>]
9870 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9874 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9877 *) New config option to avoid instructions that are illegal on the 80386.
9878 The default code is faster, but requires at least a 486.
9881 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9882 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9883 same as SSL2_VERSION anyway.
9886 *) New "-showcerts" option for s_client.
9889 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9890 application. Various cleanups and fixes.
9893 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9894 modify error routines to work internally. Add error codes and PBE init
9895 to library startup routines.
9898 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9899 packing functions to asn1 and evp. Changed function names and error
9900 codes along the way.
9903 *) PKCS12 integration: and so it begins... First of several patches to
9904 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9905 objects to objects.h
9908 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9909 and display support for Thawte strong extranet extension.
9912 *) Add LinuxPPC support.
9913 [Jeff Dubrule <igor@pobox.org>]
9915 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9916 bn_div_words in alpha.s.
9917 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9919 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9920 OAEP isn't supported when OpenSSL is built with RSAref.
9921 [Ulf Moeller <ulf@fitug.de>]
9923 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9924 so they no longer are missing under -DNOPROTO.
9925 [Soren S. Jorvang <soren@t.dk>]
9928 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9930 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9931 doesn't work when the session is reused. Coming soon!
9934 *) Fix a security hole, that allows sessions to be reused in the wrong
9935 context thus bypassing client cert protection! All software that uses
9936 client certs and session caches in multiple contexts NEEDS PATCHING to
9937 allow session reuse! A fuller solution is in the works.
9938 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9940 *) Some more source tree cleanups (removed obsolete files
9941 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9942 permission on "config" script to be executable) and a fix for the INSTALL
9944 [Ulf Moeller <ulf@fitug.de>]
9946 *) Remove some legacy and erroneous uses of malloc, free instead of
9948 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9950 *) Make rsa_oaep_test return non-zero on error.
9951 [Ulf Moeller <ulf@fitug.de>]
9953 *) Add support for native Solaris shared libraries. Configure
9954 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9955 if someone would make that last step automatic.
9956 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9958 *) ctx_size was not built with the right compiler during "make links". Fixed.
9961 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9962 except NULL ciphers". This means the default cipher list will no longer
9963 enable NULL ciphers. They need to be specifically enabled e.g. with
9964 the string "DEFAULT:eNULL".
9967 *) Fix to RSA private encryption routines: if p < q then it would
9968 occasionally produce an invalid result. This will only happen with
9969 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9972 *) Be less restrictive and allow also `perl util/perlpath.pl
9973 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9974 because this way one can also use an interpreter named `perl5' (which is
9975 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9976 installed as `perl').
9977 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9979 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9980 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9982 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9983 advapi32.lib to Win32 build and change the pem test comparision
9984 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9985 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9986 and crypto/des/ede_cbcm_enc.c.
9989 *) DES quad checksum was broken on big-endian architectures. Fixed.
9992 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9993 Win32 test batch file so it (might) work again. The Win32 test batch file
9994 is horrible: I feel ill....
9997 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9998 in e_os.h. Audit of header files to check ANSI and non ANSI
9999 sections: 10 functions were absent from non ANSI section and not exported
10000 from Windows DLLs. Fixed up libeay.num for new functions.
10003 *) Make `openssl version' output lines consistent.
10004 [Ralf S. Engelschall]
10006 *) Fix Win32 symbol export lists for BIO functions: Added
10007 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10008 to ms/libeay{16,32}.def.
10009 [Ralf S. Engelschall]
10011 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10012 fine under Unix and passes some trivial tests I've now added. But the
10013 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10014 added to make sure no one expects that this stuff really works in the
10015 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10016 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10018 [Ralf S. Engelschall]
10020 *) Fix the generation of two part addresses in perl.
10021 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10023 *) Add config entry for Linux on MIPS.
10024 [John Tobey <jtobey@channel1.com>]
10026 *) Make links whenever Configure is run, unless we are on Windoze.
10029 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10030 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10034 *) Add a useful kludge to allow package maintainers to specify compiler and
10035 other platforms details on the command line without having to patch the
10036 Configure script everytime: One now can use ``perl Configure
10037 <id>:<details>'', i.e. platform ids are allowed to have details appended
10038 to them (seperated by colons). This is treated as there would be a static
10039 pre-configured entry in Configure's %table under key <id> with value
10040 <details> and ``perl Configure <id>'' is called. So, when you want to
10041 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10042 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10043 now, which overrides the FreeBSD-elf entry on-the-fly.
10044 [Ralf S. Engelschall]
10046 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10049 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10050 on the `perl Configure ...' command line. This way one can compile
10051 OpenSSL libraries with Position Independent Code (PIC) which is needed
10052 for linking it into DSOs.
10053 [Ralf S. Engelschall]
10055 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10059 *) Cleaned up the LICENSE document: The official contact for any license
10060 questions now is the OpenSSL core team under openssl-core@openssl.org.
10061 And add a paragraph about the dual-license situation to make sure people
10062 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10063 to the OpenSSL toolkit.
10064 [Ralf S. Engelschall]
10066 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10067 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10068 Additonally cleaned up the `make links' target: Remove unnecessary
10069 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10070 to speed processing and no longer clutter the display with confusing
10071 stuff. Instead only the actually done links are displayed.
10072 [Ralf S. Engelschall]
10074 *) Permit null encryption ciphersuites, used for authentication only. It used
10075 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10076 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10080 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10081 signed attributes when verifying signatures (this would break them),
10082 the detached data encoding was wrong and public keys obtained using
10083 X509_get_pubkey() weren't freed.
10086 *) Add text documentation for the BUFFER functions. Also added a work around
10087 to a Win95 console bug. This was triggered by the password read stuff: the
10088 last character typed gets carried over to the next fread(). If you were
10089 generating a new cert request using 'req' for example then the last
10090 character of the passphrase would be CR which would then enter the first
10094 *) Added the new `Includes OpenSSL Cryptography Software' button as
10095 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10096 button and can be used by applications based on OpenSSL to show the
10097 relationship to the OpenSSL project.
10098 [Ralf S. Engelschall]
10100 *) Remove confusing variables in function signatures in files
10101 ssl/ssl_lib.c and ssl/ssl.h.
10102 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10104 *) Don't install bss_file.c under PREFIX/include/
10105 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10107 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10108 functions that return function pointers and has support for NT specific
10109 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10110 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10111 unsigned to signed types: this was killing the Win32 compile.
10114 *) Add new certificate file to stack functions,
10115 SSL_add_dir_cert_subjects_to_stack() and
10116 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10117 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10118 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10119 This means that Apache-SSL and similar packages don't have to mess around
10120 to add as many CAs as they want to the preferred list.
10123 *) Experiment with doxygen documentation. Currently only partially applied to
10125 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10126 openssl.doxy as the configuration file.
10129 *) Get rid of remaining C++-style comments which strict C compilers hate.
10130 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10132 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10133 compiled in by default: it has problems with large keys.
10136 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10137 DH private keys and/or callback functions which directly correspond to
10138 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10139 is needed for applications which have to configure certificates on a
10140 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10142 For the RSA certificate situation is makes no difference, but
10143 for the DSA certificate situation this fixes the "no shared cipher"
10144 problem where the OpenSSL cipher selection procedure failed because the
10145 temporary keys were not overtaken from the context and the API provided
10146 no way to reconfigure them.
10147 The new functions now let applications reconfigure the stuff and they
10148 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10149 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10150 non-public-API function ssl_cert_instantiate() is used as a helper
10151 function and also to reduce code redundancy inside ssl_rsa.c.
10152 [Ralf S. Engelschall]
10154 *) Move s_server -dcert and -dkey options out of the undocumented feature
10155 area because they are useful for the DSA situation and should be
10156 recognized by the users.
10157 [Ralf S. Engelschall]
10159 *) Fix the cipher decision scheme for export ciphers: the export bits are
10160 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10161 SSL_EXP_MASK. So, the original variable has to be used instead of the
10162 already masked variable.
10163 [Richard Levitte <levitte@stacken.kth.se>]
10165 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10166 [Richard Levitte <levitte@stacken.kth.se>]
10168 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10169 from `int' to `unsigned int' because it's a length and initialized by
10170 EVP_DigestFinal() which expects an `unsigned int *'.
10171 [Richard Levitte <levitte@stacken.kth.se>]
10173 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10174 script. Instead use the usual Shell->Perl transition trick.
10175 [Ralf S. Engelschall]
10177 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10178 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10179 -noout -modulus' as it's already the case for `openssl rsa -noout
10180 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10181 currently the public key is printed (a decision which was already done by
10182 `openssl dsa -modulus' in the past) which serves a similar purpose.
10183 Additionally the NO_RSA no longer completely removes the whole -modulus
10184 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10186 [Ralf S. Engelschall]
10188 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10189 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10190 [Arne Ansper <arne@ats.cyber.ee>]
10192 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10193 to be added. Now both 'req' and 'ca' can use new objects defined in the
10197 *) Add cool BIO that does syslog (or event log on NT).
10198 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10200 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10201 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10202 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10203 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10206 *) Add preliminary config info for new extension code.
10209 *) Make RSA_NO_PADDING really use no padding.
10210 [Ulf Moeller <ulf@fitug.de>]
10212 *) Generate errors when private/public key check is done.
10215 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10216 for some CRL extensions and new objects added.
10219 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10220 key usage extension and fuller support for authority key id.
10223 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10224 padding method for RSA, which is recommended for new applications in PKCS
10225 #1 v2.0 (RFC 2437, October 1998).
10226 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10227 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10228 against Bleichbacher's attack on RSA.
10229 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10232 *) Updates to the new SSL compression code
10233 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10235 *) Fix so that the version number in the master secret, when passed
10236 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10237 (because the server will not accept higher), that the version number
10238 is 0x03,0x01, not 0x03,0x00
10239 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10241 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10242 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10243 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10246 *) Support for RAW extensions where an arbitrary extension can be
10247 created by including its DER encoding. See apps/openssl.cnf for
10251 *) Make sure latest Perl versions don't interpret some generated C array
10252 code as Perl array code in the crypto/err/err_genc.pl script.
10253 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10255 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10256 not many people have the assembler. Various Win32 compilation fixes and
10257 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10258 build instructions.
10261 *) Modify configure script 'Configure' to automatically create crypto/date.h
10262 file under Win32 and also build pem.h from pem.org. New script
10263 util/mkfiles.pl to create the MINFO file on environments that can't do a
10264 'make files': perl util/mkfiles.pl >MINFO should work.
10267 *) Major rework of DES function declarations, in the pursuit of correctness
10268 and purity. As a result, many evil casts evaporated, and some weirdness,
10269 too. You may find this causes warnings in your code. Zapping your evil
10270 casts will probably fix them. Mostly.
10273 *) Fix for a typo in asn1.h. Bug fix to object creation script
10274 obj_dat.pl. It considered a zero in an object definition to mean
10275 "end of object": none of the objects in objects.h have any zeros
10276 so it wasn't spotted.
10277 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10279 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10280 Masking (CBCM). In the absence of test vectors, the best I have been able
10281 to do is check that the decrypt undoes the encrypt, so far. Send me test
10282 vectors if you have them.
10285 *) Correct calculation of key length for export ciphers (too much space was
10286 allocated for null ciphers). This has not been tested!
10289 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10290 message is now correct (it understands "crypto" and "ssl" on its
10291 command line). There is also now an "update" option. This will update
10292 the util/ssleay.num and util/libeay.num files with any new functions.
10294 perl util/mkdef.pl crypto ssl update
10295 it will update them.
10298 *) Overhauled the Perl interface (perl/*):
10299 - ported BN stuff to OpenSSL's different BN library
10300 - made the perl/ source tree CVS-aware
10301 - renamed the package from SSLeay to OpenSSL (the files still contain
10302 their history because I've copied them in the repository)
10303 - removed obsolete files (the test scripts will be replaced
10304 by better Test::Harness variants in the future)
10305 [Ralf S. Engelschall]
10307 *) First cut for a very conservative source tree cleanup:
10308 1. merge various obsolete readme texts into doc/ssleay.txt
10309 where we collect the old documents and readme texts.
10310 2. remove the first part of files where I'm already sure that we no
10311 longer need them because of three reasons: either they are just temporary
10312 files which were left by Eric or they are preserved original files where
10313 I've verified that the diff is also available in the CVS via "cvs diff
10314 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10315 the crypto/md/ stuff).
10316 [Ralf S. Engelschall]
10318 *) More extension code. Incomplete support for subject and issuer alt
10319 name, issuer and authority key id. Change the i2v function parameters
10320 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10321 what that's for :-) Fix to ASN1 macro which messed up
10322 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10325 *) Preliminary support for ENUMERATED type. This is largely copied from the
10329 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10330 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10332 *) Make sure `make rehash' target really finds the `openssl' program.
10333 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10335 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10336 like to hear about it if this slows down other processors.
10339 *) Add CygWin32 platform information to Configure script.
10340 [Alan Batie <batie@aahz.jf.intel.com>]
10342 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10343 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10345 *) New program nseq to manipulate netscape certificate sequences
10348 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10352 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10353 but the BN code had some problems that would cause failures when
10354 doing certificate verification and some other functions.
10355 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10357 *) Add ASN1 and PEM code to support netscape certificate sequences.
10360 *) Add ASN1 and PEM code to support netscape certificate sequences.
10363 *) Add several PKIX and private extended key usage OIDs.
10366 *) Modify the 'ca' program to handle the new extension code. Modify
10367 openssl.cnf for new extension format, add comments.
10370 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10371 and add a sample to openssl.cnf so req -x509 now adds appropriate
10375 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10376 error code, add initial support to X509_print() and x509 application.
10379 *) Takes a deep breath and start addding X509 V3 extension support code. Add
10380 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10381 stuff is currently isolated and isn't even compiled yet.
10384 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10385 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10386 Removed the versions check from X509 routines when loading extensions:
10387 this allows certain broken certificates that don't set the version
10388 properly to be processed.
10391 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10392 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10393 can still be regenerated with "make depend".
10396 *) Spelling mistake in C version of CAST-128.
10397 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10399 *) Changes to the error generation code. The perl script err-code.pl
10400 now reads in the old error codes and retains the old numbers, only
10401 adding new ones if necessary. It also only changes the .err files if new
10402 codes are added. The makefiles have been modified to only insert errors
10403 when needed (to avoid needlessly modifying header files). This is done
10404 by only inserting errors if the .err file is newer than the auto generated
10405 C file. To rebuild all the error codes from scratch (the old behaviour)
10406 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10407 or delete all the .err files.
10410 *) CAST-128 was incorrectly implemented for short keys. The C version has
10411 been fixed, but is untested. The assembler versions are also fixed, but
10412 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10413 to regenerate it if needed.
10414 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10415 Hagino <itojun@kame.net>]
10417 *) File was opened incorrectly in randfile.c.
10418 [Ulf Möller <ulf@fitug.de>]
10420 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10421 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10422 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10423 al: it's just almost always a UTCTime. Note this patch adds new error
10424 codes so do a "make errors" if there are problems.
10427 *) Correct Linux 1 recognition in config.
10428 [Ulf Möller <ulf@fitug.de>]
10430 *) Remove pointless MD5 hash when using DSA keys in ca.
10431 [Anonymous <nobody@replay.com>]
10433 *) Generate an error if given an empty string as a cert directory. Also
10434 generate an error if handed NULL (previously returned 0 to indicate an
10435 error, but didn't set one).
10436 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10438 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10441 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10442 parameters. This was causing a warning which killed off the Win32 compile.
10445 *) Remove C++ style comments from crypto/bn/bn_local.h.
10446 [Neil Costigan <neil.costigan@celocom.com>]
10448 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10449 based on a text string, looking up short and long names and finally
10450 "dot" format. The "dot" format stuff didn't work. Added new function
10451 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10452 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10453 OID is not part of the table.
10456 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10457 X509_LOOKUP_by_alias().
10460 *) Sort openssl functions by name.
10463 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10464 encryption from sample DSA keys (in case anyone is interested the password
10468 *) Make _all_ *_free functions accept a NULL pointer.
10469 [Frans Heymans <fheymans@isaserver.be>]
10471 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10473 [Anonymous <nobody@replay.com>]
10475 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10476 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10478 *) Don't blow it for numeric -newkey arguments to apps/req.
10479 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10481 *) Temp key "for export" tests were wrong in s3_srvr.c.
10482 [Anonymous <nobody@replay.com>]
10484 *) Add prototype for temp key callback functions
10485 SSL_CTX_set_tmp_{rsa,dh}_callback().
10488 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10489 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10492 *) X509_name_add_entry() freed the wrong thing after an error.
10493 [Arne Ansper <arne@ats.cyber.ee>]
10495 *) rsa_eay.c would attempt to free a NULL context.
10496 [Arne Ansper <arne@ats.cyber.ee>]
10498 *) BIO_s_socket() had a broken should_retry() on Windoze.
10499 [Arne Ansper <arne@ats.cyber.ee>]
10501 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10502 [Arne Ansper <arne@ats.cyber.ee>]
10504 *) Make sure the already existing X509_STORE->depth variable is initialized
10505 in X509_STORE_new(), but document the fact that this variable is still
10506 unused in the certificate verification process.
10507 [Ralf S. Engelschall]
10509 *) Fix the various library and apps files to free up pkeys obtained from
10510 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10513 *) Fix reference counting in X509_PUBKEY_get(). This makes
10514 demos/maurice/example2.c work, amongst others, probably.
10515 [Steve Henson and Ben Laurie]
10517 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10518 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10519 are no longer created. This way we have a single and consistent command
10520 line interface `openssl <command>', similar to `cvs <command>'.
10521 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10523 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10524 BIT STRING wrapper always have zero unused bits.
10527 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10530 *) Make the top-level INSTALL documentation easier to understand.
10533 *) Makefiles updated to exit if an error occurs in a sub-directory
10534 make (including if user presses ^C) [Paul Sutton]
10536 *) Make Montgomery context stuff explicit in RSA data structure.
10539 *) Fix build order of pem and err to allow for generated pem.h.
10542 *) Fix renumbering bug in X509_NAME_delete_entry().
10545 *) Enhanced the err-ins.pl script so it makes the error library number
10546 global and can add a library name. This is needed for external ASN1 and
10547 other error libraries.
10550 *) Fixed sk_insert which never worked properly.
10553 *) Fix ASN1 macros so they can handle indefinite length construted
10554 EXPLICIT tags. Some non standard certificates use these: they can now
10558 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10559 into a single doc/ssleay.txt bundle. This way the information is still
10560 preserved but no longer messes up this directory. Now it's new room for
10561 the new set of documenation files.
10562 [Ralf S. Engelschall]
10564 *) SETs were incorrectly DER encoded. This was a major pain, because they
10565 shared code with SEQUENCEs, which aren't coded the same. This means that
10566 almost everything to do with SETs or SEQUENCEs has either changed name or
10567 number of arguments.
10568 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10570 *) Fix test data to work with the above.
10573 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10574 was already fixed by Eric for 0.9.1 it seems.
10575 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10577 *) Autodetect FreeBSD3.
10580 *) Fix various bugs in Configure. This affects the following platforms:
10584 unixware-2.0-pentium
10588 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10589 before they are needed.
10592 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10596 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10598 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10599 changed SSLeay to OpenSSL in version strings.
10600 [Ralf S. Engelschall]
10602 *) Some fixups to the top-level documents.
10605 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10606 because the symlink to include/ was missing.
10607 [Ralf S. Engelschall]
10609 *) Incorporated the popular no-RSA/DSA-only patches
10610 which allow to compile a RSA-free SSLeay.
10611 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10613 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10614 when "ssleay" is still not found.
10615 [Ralf S. Engelschall]
10617 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10618 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10620 *) Updated the README file.
10621 [Ralf S. Engelschall]
10623 *) Added various .cvsignore files in the CVS repository subdirs
10624 to make a "cvs update" really silent.
10625 [Ralf S. Engelschall]
10627 *) Recompiled the error-definition header files and added
10628 missing symbols to the Win32 linker tables.
10629 [Ralf S. Engelschall]
10631 *) Cleaned up the top-level documents;
10632 o new files: CHANGES and LICENSE
10633 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10634 o merged COPYRIGHT into LICENSE
10635 o removed obsolete TODO file
10636 o renamed MICROSOFT to INSTALL.W32
10637 [Ralf S. Engelschall]
10639 *) Removed dummy files from the 0.9.1b source tree:
10640 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10641 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10642 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10643 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10644 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10645 [Ralf S. Engelschall]
10647 *) Added various platform portability fixes.
10650 *) The Genesis of the OpenSSL rpject:
10651 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10652 Young and Tim J. Hudson created while they were working for C2Net until
10654 [The OpenSSL Project]
10657 Changes between 0.9.0b and 0.9.1b [not released]
10659 *) Updated a few CA certificates under certs/
10662 *) Changed some BIGNUM api stuff.
10665 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10666 DGUX x86, Linux Alpha, etc.
10669 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10670 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10674 *) Add -strparse option to asn1pars program which parses nested
10676 [Dr Stephen Henson <shenson@bigfoot.com>]
10678 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10681 *) DSA fix for "ca" program.
10684 *) Added "-genkey" option to "dsaparam" program.
10687 *) Added RIPE MD160 (rmd160) message digest.
10690 *) Added -a (all) option to "ssleay version" command.
10693 *) Added PLATFORM define which is the id given to Configure.
10696 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10699 *) Extended the ASN.1 parser routines.
10702 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10705 *) Added a BN_CTX to the BN library.
10708 *) Fixed the weak key values in DES library
10711 *) Changed API in EVP library for cipher aliases.
10714 *) Added support for RC2/64bit cipher.
10717 *) Converted the lhash library to the crypto/mem.c functions.
10720 *) Added more recognized ASN.1 object ids.
10723 *) Added more RSA padding checks for SSL/TLS.
10726 *) Added BIO proxy/filter functionality.
10729 *) Added extra_certs to SSL_CTX which can be used
10730 send extra CA certificates to the client in the CA cert chain sending
10731 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10734 *) Now Fortezza is denied in the authentication phase because
10735 this is key exchange mechanism is not supported by SSLeay at all.
10738 *) Additional PKCS1 checks.
10741 *) Support the string "TLSv1" for all TLS v1 ciphers.
10744 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10745 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10748 *) Fixed a few memory leaks.
10751 *) Fixed various code and comment typos.
10754 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10755 bytes sent in the client random.
10756 [Edward Bishop <ebishop@spyglass.com>]