5 Changes between 1.0.2 and 1.0.2a [xx XXX xxxx]
7 *) Removed the export ciphers from the DEFAULT ciphers
10 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
14 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
15 sends a carefully crafted handshake message, to cause OpenSSL to fail
16 to free up to 64k of memory causing a memory leak. This could be
17 exploited in a Denial Of Service attack. This issue affects OpenSSL
18 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
19 whether SRTP is used or configured. Implementations of OpenSSL that
20 have been compiled with OPENSSL_NO_SRTP defined are not affected.
22 The fix was developed by the OpenSSL team.
26 *) Session Ticket Memory Leak.
28 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
29 integrity of that ticket is first verified. In the event of a session
30 ticket integrity check failing, OpenSSL will fail to free memory
31 causing a memory leak. By sending a large number of invalid session
32 tickets an attacker could exploit this issue in a Denial Of Service
37 *) Build option no-ssl3 is incomplete.
39 When OpenSSL is configured with "no-ssl3" as a build option, servers
40 could accept and complete a SSL 3.0 handshake, and clients could be
41 configured to send them.
43 [Akamai and the OpenSSL team]
45 *) Add support for TLS_FALLBACK_SCSV.
46 Client applications doing fallback retries should call
47 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
49 [Adam Langley, Bodo Moeller]
51 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
52 ARMv5 through ARMv8, as opposite to "locking" it to single one.
53 So far those who have to target multiple plaforms would compromise
54 and argue that binary targeting say ARMv5 would still execute on
55 ARMv8. "Universal" build resolves this compromise by providing
56 near-optimal performance even on newer platforms.
59 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
60 (other platforms pending).
61 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
63 *) Add support for the SignedCertificateTimestampList certificate and
64 OCSP response extensions from RFC6962.
67 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
68 for corner cases. (Certain input points at infinity could lead to
69 bogus results, with non-infinity inputs mapped to infinity too.)
72 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
73 This covers AES, SHA256/512 and GHASH. "Initial" means that most
74 common cases are optimized and there still is room for further
75 improvements. Vector Permutation AES for Altivec is also added.
78 *) Add support for little-endian ppc64 Linux target.
81 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
82 SHA1, SHA256 and GHASH. "Initial" means that most common cases
83 are optimized and there still is room for further improvements.
84 Both 32- and 64-bit modes are supported.
85 [Andy Polyakov, Ard Biesheuvel (Linaro)]
87 *) Improved ARMv7 NEON support.
90 *) Support for SPARC Architecture 2011 crypto extensions, first
91 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
92 SHA256/512, MD5, GHASH and modular exponentiation.
93 [Andy Polyakov, David Miller]
95 *) Accelerated modular exponentiation for Intel processors, a.k.a.
97 [Shay Gueron & Vlad Krasnov (Intel Corp)]
99 *) Support for new and upcoming Intel processors, including AVX2,
100 BMI and SHA ISA extensions. This includes additional "stitched"
101 implementations, AESNI-SHA256 and GCM, and multi-buffer support
104 This work was sponsored by Intel Corp.
107 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
108 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
109 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
112 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
113 this fixes a limiation in previous versions of OpenSSL.
116 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
117 MGF1 digest and OAEP label.
120 *) Add EVP support for key wrapping algorithms, to avoid problems with
121 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
122 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
123 algorithms and include tests cases.
126 *) Add functions to allocate and set the fields of an ECDSA_METHOD
128 [Douglas E. Engert, Steve Henson]
130 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
131 difference in days and seconds between two tm or ASN1_TIME structures.
134 *) Add -rev test option to s_server to just reverse order of characters
135 received by client and send back to server. Also prints an abbreviated
136 summary of the connection parameters.
139 *) New option -brief for s_client and s_server to print out a brief summary
140 of connection parameters.
143 *) Add callbacks for arbitrary TLS extensions.
144 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
146 *) New option -crl_download in several openssl utilities to download CRLs
147 from CRLDP extension in certificates.
150 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
153 *) New function X509_CRL_diff to generate a delta CRL from the difference
154 of two full CRLs. Add support to "crl" utility.
157 *) New functions to set lookup_crls function and to retrieve
158 X509_STORE from X509_STORE_CTX.
161 *) Print out deprecated issuer and subject unique ID fields in
165 *) Extend OCSP I/O functions so they can be used for simple general purpose
166 HTTP as well as OCSP. New wrapper function which can be used to download
167 CRLs using the OCSP API.
170 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
173 *) SSL_CONF* functions. These provide a common framework for application
174 configuration using configuration files or command lines.
177 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
178 message callback and prints the results. Needs compile time option
179 "enable-ssl-trace". New options to s_client and s_server to enable
183 *) New ctrl and macro to retrieve supported points extensions.
184 Print out extension in s_server and s_client.
187 *) New functions to retrieve certificate signature and signature
191 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
195 *) New Suite B modes for TLS code. These use and enforce the requirements
196 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
197 only use Suite B curves. The Suite B modes can be set by using the
198 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
201 *) New chain verification flags for Suite B levels of security. Check
202 algorithms are acceptable when flags are set in X509_verify_cert.
205 *) Make tls1_check_chain return a set of flags indicating checks passed
206 by a certificate chain. Add additional tests to handle client
207 certificates: checks for matching certificate type and issuer name
211 *) If an attempt is made to use a signature algorithm not in the peer
212 preference list abort the handshake. If client has no suitable
213 signature algorithms in response to a certificate request do not
217 *) If server EC tmp key is not in client preference list abort handshake.
220 *) Add support for certificate stores in CERT structure. This makes it
221 possible to have different stores per SSL structure or one store in
222 the parent SSL_CTX. Include distint stores for certificate chain
223 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
224 to build and store a certificate chain in CERT structure: returing
225 an error if the chain cannot be built: this will allow applications
226 to test if a chain is correctly configured.
228 Note: if the CERT based stores are not set then the parent SSL_CTX
229 store is used to retain compatibility with existing behaviour.
233 *) New function ssl_set_client_disabled to set a ciphersuite disabled
234 mask based on the current session, check mask when sending client
235 hello and checking the requested ciphersuite.
238 *) New ctrls to retrieve and set certificate types in a certificate
239 request message. Print out received values in s_client. If certificate
240 types is not set with custom values set sensible values based on
241 supported signature algorithms.
244 *) Support for distinct client and server supported signature algorithms.
247 *) Add certificate callback. If set this is called whenever a certificate
248 is required by client or server. An application can decide which
249 certificate chain to present based on arbitrary criteria: for example
250 supported signature algorithms. Add very simple example to s_server.
251 This fixes many of the problems and restrictions of the existing client
252 certificate callback: for example you can now clear an existing
253 certificate and specify the whole chain.
256 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
257 the certificate can be used for (if anything). Set valid_flags field
258 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
259 to have similar checks in it.
261 Add new "cert_flags" field to CERT structure and include a "strict mode".
262 This enforces some TLS certificate requirements (such as only permitting
263 certificate signature algorithms contained in the supported algorithms
264 extension) which some implementations ignore: this option should be used
265 with caution as it could cause interoperability issues.
268 *) Update and tidy signature algorithm extension processing. Work out
269 shared signature algorithms based on preferences and peer algorithms
270 and print them out in s_client and s_server. Abort handshake if no
271 shared signature algorithms.
274 *) Add new functions to allow customised supported signature algorithms
275 for SSL and SSL_CTX structures. Add options to s_client and s_server
279 *) New function SSL_certs_clear() to delete all references to certificates
280 from an SSL structure. Before this once a certificate had been added
281 it couldn't be removed.
284 *) Integrate hostname, email address and IP address checking with certificate
285 verification. New verify options supporting checking in opensl utility.
288 *) Fixes and wildcard matching support to hostname and email checking
289 functions. Add manual page.
290 [Florian Weimer (Red Hat Product Security Team)]
292 *) New functions to check a hostname email or IP address against a
293 certificate. Add options x509 utility to print results of checks against
297 *) Fix OCSP checking.
298 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
300 *) Initial experimental support for explicitly trusted non-root CAs.
301 OpenSSL still tries to build a complete chain to a root but if an
302 intermediate CA has a trust setting included that is used. The first
303 setting is used: whether to trust (e.g., -addtrust option to the x509
307 *) Add -trusted_first option which attempts to find certificates in the
308 trusted store even if an untrusted chain is also supplied.
311 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
312 platform support for Linux and Android.
315 *) Support for linux-x32, ILP32 environment in x86_64 framework.
318 *) Experimental multi-implementation support for FIPS capable OpenSSL.
319 When in FIPS mode the approved implementations are used as normal,
320 when not in FIPS mode the internal unapproved versions are used instead.
321 This means that the FIPS capable OpenSSL isn't forced to use the
322 (often lower perfomance) FIPS implementations outside FIPS mode.
325 *) Transparently support X9.42 DH parameters when calling
326 PEM_read_bio_DHparameters. This means existing applications can handle
327 the new parameter format automatically.
330 *) Initial experimental support for X9.42 DH parameter format: mainly
331 to support use of 'q' parameter for RFC5114 parameters.
334 *) Add DH parameters from RFC5114 including test data to dhtest.
337 *) Support for automatic EC temporary key parameter selection. If enabled
338 the most preferred EC parameters are automatically used instead of
339 hardcoded fixed parameters. Now a server just has to call:
340 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
341 support ECDH and use the most appropriate parameters.
344 *) Enhance and tidy EC curve and point format TLS extension code. Use
345 static structures instead of allocation if default values are used.
346 New ctrls to set curves we wish to support and to retrieve shared curves.
347 Print out shared curves in s_server. New options to s_server and s_client
348 to set list of supported curves.
351 *) New ctrls to retrieve supported signature algorithms and
352 supported curve values as an array of NIDs. Extend openssl utility
353 to print out received values.
356 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
357 between NIDs and the more common NIST names such as "P-256". Enhance
358 ecparam utility and ECC method to recognise the NIST names for curves.
361 *) Enhance SSL/TLS certificate chain handling to support different
362 chains for each certificate instead of one chain in the parent SSL_CTX.
365 *) Support for fixed DH ciphersuite client authentication: where both
366 server and client use DH certificates with common parameters.
369 *) Support for fixed DH ciphersuites: those requiring DH server
373 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
375 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
376 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
377 X509_CINF_get_signature were reverted post internal team review.
379 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
381 *) Build fixes for the Windows and OpenVMS platforms
382 [Matt Caswell and Richard Levitte]
384 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
386 *) Abort handshake if server key exchange message is omitted for ephemeral
389 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
390 reporting this issue.
394 *) Remove non-export ephemeral RSA code on client and server. This code
395 violated the TLS standard by allowing the use of temporary RSA keys in
396 non-export ciphersuites and could be used by a server to effectively
397 downgrade the RSA key length used to a value smaller than the server
398 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
399 INRIA or reporting this issue.
403 *) Ensure that the session ID context of an SSL is updated when its
404 SSL_CTX is updated via SSL_set_SSL_CTX.
406 The session ID context is typically set from the parent SSL_CTX,
407 and can vary with the CTX.
410 *) Fix various certificate fingerprint issues.
412 By using non-DER or invalid encodings outside the signed portion of a
413 certificate the fingerprint can be changed without breaking the signature.
414 Although no details of the signed portion of the certificate can be changed
415 this can cause problems with some applications: e.g. those using the
416 certificate fingerprint for blacklists.
418 1. Reject signatures with non zero unused bits.
420 If the BIT STRING containing the signature has non zero unused bits reject
421 the signature. All current signature algorithms require zero unused bits.
423 2. Check certificate algorithm consistency.
425 Check the AlgorithmIdentifier inside TBS matches the one in the
426 certificate signature. NB: this will result in signature failure
427 errors for some broken certificates.
429 Thanks to Konrad Kraszewski from Google for reporting this issue.
431 3. Check DSA/ECDSA signatures use DER.
433 Reencode DSA/ECDSA signatures and compare with the original received
434 signature. Return an error if there is a mismatch.
436 This will reject various cases including garbage after signature
437 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
438 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
439 (negative or with leading zeroes).
441 Further analysis was conducted and fixes were developed by Stephen Henson
442 of the OpenSSL core team.
447 *) Do not resume sessions on the server if the negotiated protocol
448 version does not match the session's version. Resuming with a different
449 version, while not strictly forbidden by the RFC, is of questionable
450 sanity and breaks all known clients.
451 [David Benjamin, Emilia Käsper]
453 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
454 early CCS messages during renegotiation. (Note that because
455 renegotiation is encrypted, this early CCS was not exploitable.)
458 *) Tighten client-side session ticket handling during renegotiation:
459 ensure that the client only accepts a session ticket if the server sends
460 the extension anew in the ServerHello. Previously, a TLS client would
461 reuse the old extension state and thus accept a session ticket if one was
462 announced in the initial ServerHello.
464 Similarly, ensure that the client requires a session ticket if one
465 was advertised in the ServerHello. Previously, a TLS client would
466 ignore a missing NewSessionTicket message.
469 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
473 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
474 sends a carefully crafted handshake message, to cause OpenSSL to fail
475 to free up to 64k of memory causing a memory leak. This could be
476 exploited in a Denial Of Service attack. This issue affects OpenSSL
477 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
478 whether SRTP is used or configured. Implementations of OpenSSL that
479 have been compiled with OPENSSL_NO_SRTP defined are not affected.
481 The fix was developed by the OpenSSL team.
485 *) Session Ticket Memory Leak.
487 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
488 integrity of that ticket is first verified. In the event of a session
489 ticket integrity check failing, OpenSSL will fail to free memory
490 causing a memory leak. By sending a large number of invalid session
491 tickets an attacker could exploit this issue in a Denial Of Service
496 *) Build option no-ssl3 is incomplete.
498 When OpenSSL is configured with "no-ssl3" as a build option, servers
499 could accept and complete a SSL 3.0 handshake, and clients could be
500 configured to send them.
502 [Akamai and the OpenSSL team]
504 *) Add support for TLS_FALLBACK_SCSV.
505 Client applications doing fallback retries should call
506 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
508 [Adam Langley, Bodo Moeller]
510 *) Add additional DigestInfo checks.
512 Reencode DigestInto in DER and check against the original when
513 verifying RSA signature: this will reject any improperly encoded
514 DigestInfo structures.
516 Note: this is a precautionary measure and no attacks are currently known.
520 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
522 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
523 SRP code can be overrun an internal buffer. Add sanity check that
524 g, A, B < N to SRP code.
526 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
527 Group for discovering this issue.
531 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
532 TLS 1.0 instead of higher protocol versions when the ClientHello message
533 is badly fragmented. This allows a man-in-the-middle attacker to force a
534 downgrade to TLS 1.0 even if both the server and the client support a
535 higher protocol version, by modifying the client's TLS records.
537 Thanks to David Benjamin and Adam Langley (Google) for discovering and
538 researching this issue.
542 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
543 to a denial of service attack. A malicious server can crash the client
544 with a null pointer dereference (read) by specifying an anonymous (EC)DH
545 ciphersuite and sending carefully crafted handshake messages.
547 Thanks to Felix Gröbert (Google) for discovering and researching this
552 *) By sending carefully crafted DTLS packets an attacker could cause openssl
553 to leak memory. This can be exploited through a Denial of Service attack.
554 Thanks to Adam Langley for discovering and researching this issue.
558 *) An attacker can force openssl to consume large amounts of memory whilst
559 processing DTLS handshake messages. This can be exploited through a
560 Denial of Service attack.
561 Thanks to Adam Langley for discovering and researching this issue.
565 *) An attacker can force an error condition which causes openssl to crash
566 whilst processing DTLS packets due to memory being freed twice. This
567 can be exploited through a Denial of Service attack.
568 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
573 *) If a multithreaded client connects to a malicious server using a resumed
574 session and the server sends an ec point format extension it could write
575 up to 255 bytes to freed memory.
577 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
582 *) A malicious server can crash an OpenSSL client with a null pointer
583 dereference (read) by specifying an SRP ciphersuite even though it was not
584 properly negotiated with the client. This can be exploited through a
585 Denial of Service attack.
587 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
588 discovering and researching this issue.
592 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
593 X509_name_oneline, X509_name_print_ex et al. to leak some information
594 from the stack. Applications may be affected if they echo pretty printing
595 output to the attacker.
597 Thanks to Ivan Fratric (Google) for discovering this issue.
599 [Emilia Käsper, and Steve Henson]
601 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
602 for corner cases. (Certain input points at infinity could lead to
603 bogus results, with non-infinity inputs mapped to infinity too.)
606 Changes between 1.0.1i and 1.0.1j [xx XXX xxxx]
608 *) Add additional DigestInfo checks.
610 Reencode DigestInto in DER and check against the original when
611 verifying RSA signature: this will reject any improperly encoded
612 DigestInfo structures.
614 Note: this is a precautionary measure and no attacks are currently known.
618 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
620 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
621 handshake can force the use of weak keying material in OpenSSL
622 SSL/TLS clients and servers.
624 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
625 researching this issue. (CVE-2014-0224)
626 [KIKUCHI Masashi, Steve Henson]
628 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
629 OpenSSL DTLS client the code can be made to recurse eventually crashing
632 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
634 [Imre Rad, Steve Henson]
636 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
637 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
638 client or server. This is potentially exploitable to run arbitrary
639 code on a vulnerable client or server.
641 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
642 [Jüri Aedla, Steve Henson]
644 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
645 are subject to a denial of service attack.
647 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
648 this issue. (CVE-2014-3470)
649 [Felix Gröbert, Ivan Fratric, Steve Henson]
651 *) Harmonize version and its documentation. -f flag is used to display
653 [mancha <mancha1@zoho.com>]
655 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
657 [mancha <mancha1@zoho.com>]
659 *) Fix some double frees. These are not thought to be exploitable.
660 [mancha <mancha1@zoho.com>]
662 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
664 *) A missing bounds check in the handling of the TLS heartbeat extension
665 can be used to reveal up to 64k of memory to a connected client or
668 Thanks for Neel Mehta of Google Security for discovering this bug and to
669 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
670 preparing the fix (CVE-2014-0160)
671 [Adam Langley, Bodo Moeller]
673 *) Fix for the attack described in the paper "Recovering OpenSSL
674 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
675 by Yuval Yarom and Naomi Benger. Details can be obtained from:
676 http://eprint.iacr.org/2014/140
678 Thanks to Yuval Yarom and Naomi Benger for discovering this
679 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
680 [Yuval Yarom and Naomi Benger]
682 *) TLS pad extension: draft-agl-tls-padding-03
684 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
685 TLS client Hello record length value would otherwise be > 255 and
686 less that 512 pad with a dummy extension containing zeroes so it
687 is at least 512 bytes long.
689 [Adam Langley, Steve Henson]
691 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
693 *) Fix for TLS record tampering bug. A carefully crafted invalid
694 handshake could crash OpenSSL with a NULL pointer exception.
695 Thanks to Anton Johansson for reporting this issues.
698 *) Keep original DTLS digest and encryption contexts in retransmission
699 structures so we can use the previous session parameters if they need
700 to be resent. (CVE-2013-6450)
703 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
704 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
705 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
706 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
707 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
708 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
709 [Rob Stradling, Adam Langley]
711 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
713 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
714 supporting platforms or when small records were transferred.
715 [Andy Polyakov, Steve Henson]
717 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
719 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
721 This addresses the flaw in CBC record processing discovered by
722 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
723 at: http://www.isg.rhul.ac.uk/tls/
725 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
726 Security Group at Royal Holloway, University of London
727 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
728 Emilia Käsper for the initial patch.
730 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
732 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
733 ciphersuites which can be exploited in a denial of service attack.
734 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
735 and detecting this bug and to Wolfgang Ettlinger
736 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
740 *) Return an error when checking OCSP signatures when key is NULL.
741 This fixes a DoS attack. (CVE-2013-0166)
744 *) Make openssl verify return errors.
745 [Chris Palmer <palmer@google.com> and Ben Laurie]
747 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
748 the right response is stapled. Also change SSL_get_certificate()
749 so it returns the certificate actually sent.
750 See http://rt.openssl.org/Ticket/Display.html?id=2836.
751 [Rob Stradling <rob.stradling@comodo.com>]
753 *) Fix possible deadlock when decoding public keys.
756 *) Don't use TLS 1.0 record version number in initial client hello
760 Changes between 1.0.1b and 1.0.1c [10 May 2012]
762 *) Sanity check record length before skipping explicit IV in TLS
763 1.2, 1.1 and DTLS to fix DoS attack.
765 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
766 fuzzing as a service testing platform.
770 *) Initialise tkeylen properly when encrypting CMS messages.
771 Thanks to Solar Designer of Openwall for reporting this issue.
774 *) In FIPS mode don't try to use composite ciphers as they are not
778 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
780 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
781 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
782 mean any application compiled against OpenSSL 1.0.0 headers setting
783 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
784 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
785 0x10000000L Any application which was previously compiled against
786 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
787 will need to be recompiled as a result. Letting be results in
788 inability to disable specifically TLS 1.1 and in client context,
789 in unlike event, limit maximum offered version to TLS 1.0 [see below].
792 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
793 disable just protocol X, but all protocols above X *if* there are
794 protocols *below* X still enabled. In more practical terms it means
795 that if application wants to disable TLS1.0 in favor of TLS1.1 and
796 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
797 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
801 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
803 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
804 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
805 in CRYPTO_realloc_clean.
807 Thanks to Tavis Ormandy, Google Security Team, for discovering this
808 issue and to Adam Langley <agl@chromium.org> for fixing it.
810 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
812 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
815 *) Workarounds for some broken servers that "hang" if a client hello
816 record length exceeds 255 bytes.
818 1. Do not use record version number > TLS 1.0 in initial client
819 hello: some (but not all) hanging servers will now work.
820 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
821 the number of ciphers sent in the client hello. This should be
822 set to an even number, such as 50, for example by passing:
823 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
824 Most broken servers should now work.
825 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
826 TLS 1.2 client support entirely.
829 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
832 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
834 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
835 STRING form instead of a DigestInfo.
838 *) The format used for MDC2 RSA signatures is inconsistent between EVP
839 and the RSA_sign/RSA_verify functions. This was made more apparent when
840 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
841 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
842 the correct format in RSA_verify so both forms transparently work.
845 *) Some servers which support TLS 1.0 can choke if we initially indicate
846 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
847 encrypted premaster secret. As a workaround use the maximum pemitted
848 client version in client hello, this should keep such servers happy
849 and still work with previous versions of OpenSSL.
852 *) Add support for TLS/DTLS heartbeats.
853 [Robin Seggelmann <seggelmann@fh-muenster.de>]
855 *) Add support for SCTP.
856 [Robin Seggelmann <seggelmann@fh-muenster.de>]
858 *) Improved PRNG seeding for VOS.
859 [Paul Green <Paul.Green@stratus.com>]
861 *) Extensive assembler packs updates, most notably:
863 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
864 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
865 - x86_64: bit-sliced AES implementation;
866 - ARM: NEON support, contemporary platforms optimizations;
867 - s390x: z196 support;
868 - *: GHASH and GF(2^m) multiplication implementations;
872 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
873 (removal of unnecessary code)
874 [Peter Sylvester <peter.sylvester@edelweb.fr>]
876 *) Add TLS key material exporter from RFC 5705.
879 *) Add DTLS-SRTP negotiation from RFC 5764.
882 *) Add Next Protocol Negotiation,
883 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
884 disabled with a no-npn flag to config or Configure. Code donated
886 [Adam Langley <agl@google.com> and Ben Laurie]
888 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
889 NIST-P256, NIST-P521, with constant-time single point multiplication on
890 typical inputs. Compiler support for the nonstandard type __uint128_t is
891 required to use this (present in gcc 4.4 and later, for 64-bit builds).
892 Code made available under Apache License version 2.0.
894 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
895 line to include this in your build of OpenSSL, and run "make depend" (or
896 "make update"). This enables the following EC_METHODs:
898 EC_GFp_nistp224_method()
899 EC_GFp_nistp256_method()
900 EC_GFp_nistp521_method()
902 EC_GROUP_new_by_curve_name() will automatically use these (while
903 EC_GROUP_new_curve_GFp() currently prefers the more flexible
905 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
907 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
908 all platforms. Move ssize_t definition from e_os.h to the public
909 header file e_os2.h as it now appears in public header file cms.h
912 *) New -sigopt option to the ca, req and x509 utilities. Additional
913 signature parameters can be passed using this option and in
917 *) Add RSA PSS signing function. This will generate and set the
918 appropriate AlgorithmIdentifiers for PSS based on those in the
919 corresponding EVP_MD_CTX structure. No application support yet.
922 *) Support for companion algorithm specific ASN1 signing routines.
923 New function ASN1_item_sign_ctx() signs a pre-initialised
924 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
925 the appropriate parameters.
928 *) Add new algorithm specific ASN1 verification initialisation function
929 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
930 handling will be the same no matter what EVP_PKEY_METHOD is used.
931 Add a PSS handler to support verification of PSS signatures: checked
932 against a number of sample certificates.
935 *) Add signature printing for PSS. Add PSS OIDs.
936 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
938 *) Add algorithm specific signature printing. An individual ASN1 method
939 can now print out signatures instead of the standard hex dump.
941 More complex signatures (e.g. PSS) can print out more meaningful
942 information. Include DSA version that prints out the signature
946 *) Password based recipient info support for CMS library: implementing
950 *) Split password based encryption into PBES2 and PBKDF2 functions. This
951 neatly separates the code into cipher and PBE sections and is required
952 for some algorithms that split PBES2 into separate pieces (such as
956 *) Session-handling fixes:
957 - Fix handling of connections that are resuming with a session ID,
958 but also support Session Tickets.
959 - Fix a bug that suppressed issuing of a new ticket if the client
960 presented a ticket with an expired session.
961 - Try to set the ticket lifetime hint to something reasonable.
962 - Make tickets shorter by excluding irrelevant information.
963 - On the client side, don't ignore renewed tickets.
964 [Adam Langley, Bodo Moeller (Google)]
966 *) Fix PSK session representation.
969 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
971 This work was sponsored by Intel.
974 *) Add GCM support to TLS library. Some custom code is needed to split
975 the IV between the fixed (from PRF) and explicit (from TLS record)
976 portions. This adds all GCM ciphersuites supported by RFC5288 and
977 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
978 add a special AESGCM string for GCM only.
981 *) Expand range of ctrls for AES GCM. Permit setting invocation
982 field on decrypt and retrieval of invocation field only on encrypt.
985 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
986 As required by RFC5289 these ciphersuites cannot be used if for
987 versions of TLS earlier than 1.2.
990 *) For FIPS capable OpenSSL interpret a NULL default public key method
991 as unset and return the appopriate default but do *not* set the default.
992 This means we can return the appopriate method in applications that
993 swicth between FIPS and non-FIPS modes.
996 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
997 ENGINE is used then we cannot handle that in the FIPS module so we
998 keep original code iff non-FIPS operations are allowed.
1001 *) Add -attime option to openssl utilities.
1002 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1004 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1007 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1008 FIPS EC methods unconditionally for now.
1011 *) New build option no-ec2m to disable characteristic 2 code.
1014 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1015 all cases can be covered as some introduce binary incompatibilities.
1018 *) Redirect RSA operations to FIPS module including keygen,
1019 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1022 *) Add similar low level API blocking to ciphers.
1025 *) Low level digest APIs are not approved in FIPS mode: any attempt
1026 to use these will cause a fatal error. Applications that *really* want
1027 to use them can use the private_* version instead.
1030 *) Redirect cipher operations to FIPS module for FIPS builds.
1033 *) Redirect digest operations to FIPS module for FIPS builds.
1036 *) Update build system to add "fips" flag which will link in fipscanister.o
1037 for static and shared library builds embedding a signature if needed.
1040 *) Output TLS supported curves in preference order instead of numerical
1041 order. This is currently hardcoded for the highest order curves first.
1042 This should be configurable so applications can judge speed vs strength.
1045 *) Add TLS v1.2 server support for client authentication.
1048 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1052 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1053 FIPS modules versions.
1056 *) Add TLS v1.2 client side support for client authentication. Keep cache
1057 of handshake records longer as we don't know the hash algorithm to use
1058 until after the certificate request message is received.
1061 *) Initial TLS v1.2 client support. Add a default signature algorithms
1062 extension including all the algorithms we support. Parse new signature
1063 format in client key exchange. Relax some ECC signing restrictions for
1064 TLS v1.2 as indicated in RFC5246.
1067 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1068 to new signature format when needed using client digest preference.
1069 All server ciphersuites should now work correctly in TLS v1.2. No client
1070 support yet and no support for client certificates.
1073 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1074 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1075 ciphersuites. At present only RSA key exchange ciphersuites work with
1076 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1077 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1078 and version checking.
1081 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1082 with this defined it will not be affected by any changes to ssl internal
1083 structures. Add several utility functions to allow openssl application
1084 to work with OPENSSL_NO_SSL_INTERN defined.
1088 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1090 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1093 *) Permit abbreviated handshakes when renegotiating using the function
1094 SSL_renegotiate_abbreviated().
1095 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1097 *) Add call to ENGINE_register_all_complete() to
1098 ENGINE_load_builtin_engines(), so some implementations get used
1099 automatically instead of needing explicit application support.
1102 *) Add support for TLS key exporter as described in RFC5705.
1103 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1105 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1106 a few changes are required:
1108 Add SSL_OP_NO_TLSv1_1 flag.
1109 Add TLSv1_1 methods.
1110 Update version checking logic to handle version 1.1.
1111 Add explicit IV handling (ported from DTLS code).
1112 Add command line options to s_client/s_server.
1115 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
1117 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1119 This addresses the flaw in CBC record processing discovered by
1120 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1121 at: http://www.isg.rhul.ac.uk/tls/
1123 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1124 Security Group at Royal Holloway, University of London
1125 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1126 Emilia Käsper for the initial patch.
1128 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1130 *) Return an error when checking OCSP signatures when key is NULL.
1131 This fixes a DoS attack. (CVE-2013-0166)
1134 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1135 the right response is stapled. Also change SSL_get_certificate()
1136 so it returns the certificate actually sent.
1137 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1138 (This is a backport)
1139 [Rob Stradling <rob.stradling@comodo.com>]
1141 *) Fix possible deadlock when decoding public keys.
1144 Changes between 1.0.0i and 1.0.0j [10 May 2012]
1146 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
1149 *) Sanity check record length before skipping explicit IV in DTLS
1152 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1153 fuzzing as a service testing platform.
1157 *) Initialise tkeylen properly when encrypting CMS messages.
1158 Thanks to Solar Designer of Openwall for reporting this issue.
1161 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
1163 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1164 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1165 in CRYPTO_realloc_clean.
1167 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1168 issue and to Adam Langley <agl@chromium.org> for fixing it.
1170 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1172 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1174 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1175 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1176 content decryption and always return the same error. Note: this attack
1177 needs on average 2^20 messages so it only affects automated senders. The
1178 old behaviour can be reenabled in the CMS code by setting the
1179 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1180 an MMA defence is not necessary.
1181 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1182 this issue. (CVE-2012-0884)
1185 *) Fix CVE-2011-4619: make sure we really are receiving a
1186 client hello before rejecting multiple SGC restarts. Thanks to
1187 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1190 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1192 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1193 Thanks to Antonio Martin, Enterprise Secure Access Research and
1194 Development, Cisco Systems, Inc. for discovering this bug and
1195 preparing a fix. (CVE-2012-0050)
1198 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1200 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1201 of the Vaudenay padding oracle attack on CBC mode encryption
1202 which enables an efficient plaintext recovery attack against
1203 the OpenSSL implementation of DTLS. Their attack exploits timing
1204 differences arising during decryption processing. A research
1205 paper describing this attack can be found at:
1206 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1207 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1208 Security Group at Royal Holloway, University of London
1209 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1210 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1211 for preparing the fix. (CVE-2011-4108)
1212 [Robin Seggelmann, Michael Tuexen]
1214 *) Clear bytes used for block padding of SSL 3.0 records.
1216 [Adam Langley (Google)]
1218 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1219 Kadianakis <desnacked@gmail.com> for discovering this issue and
1220 Adam Langley for preparing the fix. (CVE-2011-4619)
1221 [Adam Langley (Google)]
1223 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1224 [Andrey Kulikov <amdeich@gmail.com>]
1226 *) Prevent malformed RFC3779 data triggering an assertion failure.
1227 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1228 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1229 [Rob Austein <sra@hactrn.net>]
1231 *) Improved PRNG seeding for VOS.
1232 [Paul Green <Paul.Green@stratus.com>]
1234 *) Fix ssl_ciph.c set-up race.
1235 [Adam Langley (Google)]
1237 *) Fix spurious failures in ecdsatest.c.
1238 [Emilia Käsper (Google)]
1240 *) Fix the BIO_f_buffer() implementation (which was mixing different
1241 interpretations of the '..._len' fields).
1242 [Adam Langley (Google)]
1244 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1245 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1246 threads won't reuse the same blinding coefficients.
1248 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1249 lock to call BN_BLINDING_invert_ex, and avoids one use of
1250 BN_BLINDING_update for each BN_BLINDING structure (previously,
1251 the last update always remained unused).
1252 [Emilia Käsper (Google)]
1254 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1255 [Bob Buckholz (Google)]
1257 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1259 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1260 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1261 [Kaspar Brand <ossl@velox.ch>]
1263 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1264 for multi-threaded use of ECDH. (CVE-2011-3210)
1265 [Adam Langley (Google)]
1267 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1270 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1271 signature public key algorithm by using OID xref utilities instead.
1272 Before this you could only use some ECC ciphersuites with SHA1 only.
1275 *) Add protection against ECDSA timing attacks as mentioned in the paper
1276 by Billy Bob Brumley and Nicola Tuveri, see:
1278 http://eprint.iacr.org/2011/232.pdf
1280 [Billy Bob Brumley and Nicola Tuveri]
1282 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1284 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1285 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1287 *) Fix bug in string printing code: if *any* escaping is enabled we must
1288 escape the escape character (backslash) or the resulting string is
1292 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1294 *) Disable code workaround for ancient and obsolete Netscape browsers
1295 and servers: an attacker can use it in a ciphersuite downgrade attack.
1296 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1299 *) Fixed J-PAKE implementation error, originally discovered by
1300 Sebastien Martini, further info and confirmation from Stefan
1301 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1304 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1306 *) Fix extension code to avoid race conditions which can result in a buffer
1307 overrun vulnerability: resumed sessions must not be modified as they can
1308 be shared by multiple threads. CVE-2010-3864
1311 *) Fix WIN32 build system to correctly link an ENGINE directory into
1315 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1317 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1319 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1321 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1323 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1324 context. The operation can be customised via the ctrl mechanism in
1325 case ENGINEs want to include additional functionality.
1328 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1331 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1332 output hashes compatible with older versions of OpenSSL.
1333 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1335 *) Fix compression algorithm handling: if resuming a session use the
1336 compression algorithm of the resumed session instead of determining
1337 it from client hello again. Don't allow server to change algorithm.
1340 *) Add load_crls() function to apps tidying load_certs() too. Add option
1341 to verify utility to allow additional CRLs to be included.
1344 *) Update OCSP request code to permit adding custom headers to the request:
1345 some responders need this.
1348 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1350 [Julia Lawall <julia@diku.dk>]
1352 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1353 needlessly dereferenced structures, used obsolete functions and
1354 didn't handle all updated verify codes correctly.
1357 *) Disable MD2 in the default configuration.
1360 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1361 indicate the initial BIO being pushed or popped. This makes it possible
1362 to determine whether the BIO is the one explicitly called or as a result
1363 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1364 it handles reference counts correctly and doesn't zero out the I/O bio
1365 when it is not being explicitly popped. WARNING: applications which
1366 included workarounds for the old buggy behaviour will need to be modified
1367 or they could free up already freed BIOs.
1370 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1371 renaming to all platforms (within the 0.9.8 branch, this was
1372 done conditionally on Netware platforms to avoid a name clash).
1373 [Guenter <lists@gknw.net>]
1375 *) Add ECDHE and PSK support to DTLS.
1376 [Michael Tuexen <tuexen@fh-muenster.de>]
1378 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1382 *) Add "missing" function EVP_MD_flags() (without this the only way to
1383 retrieve a digest flags is by accessing the structure directly. Update
1384 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1385 or cipher is registered as in the "from" argument. Print out all
1386 registered digests in the dgst usage message instead of manually
1387 attempting to work them out.
1390 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1391 this allows the use of compression and extensions. Change default cipher
1392 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1393 by default unless an application cipher string requests it.
1396 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1397 key ids to find matching certificates and keys but some PKCS#12 files
1398 don't follow the (somewhat unwritten) rules and this strategy fails.
1399 Now just gather all certificates together and the first private key
1400 then look for the first certificate that matches the key.
1403 *) Support use of registered digest and cipher names for dgst and cipher
1404 commands instead of having to add each one as a special case. So now
1411 openssl dgst -sha256 foo
1413 and this works for ENGINE based algorithms too.
1417 *) Update Gost ENGINE to support parameter files.
1418 [Victor B. Wagner <vitus@cryptocom.ru>]
1420 *) Support GeneralizedTime in ca utility.
1421 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1423 *) Enhance the hash format used for certificate directory links. The new
1424 form uses the canonical encoding (meaning equivalent names will work
1425 even if they aren't identical) and uses SHA1 instead of MD5. This form
1426 is incompatible with the older format and as a result c_rehash should
1427 be used to rebuild symbolic links.
1430 *) Make PKCS#8 the default write format for private keys, replacing the
1431 traditional format. This form is standardised, more secure and doesn't
1432 include an implicit MD5 dependency.
1435 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1436 committed to OpenSSL should pass this lot as a minimum.
1439 *) Add session ticket override functionality for use by EAP-FAST.
1440 [Jouni Malinen <j@w1.fi>]
1442 *) Modify HMAC functions to return a value. Since these can be implemented
1443 in an ENGINE errors can occur.
1446 *) Type-checked OBJ_bsearch_ex.
1449 *) Type-checked OBJ_bsearch. Also some constification necessitated
1450 by type-checking. Still to come: TXT_DB, bsearch(?),
1451 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1455 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1456 seconds to a tm structure directly, instead of going through OS
1457 specific date routines. This avoids any issues with OS routines such
1458 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1459 and X509_time_adj_ex() to cover the extended range. The existing
1460 X509_time_adj() is still usable and will no longer have any date issues.
1463 *) Delta CRL support. New use deltas option which will attempt to locate
1464 and search any appropriate delta CRLs available.
1466 This work was sponsored by Google.
1469 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1470 code and add additional score elements. Validate alternate CRL paths
1471 as part of the CRL checking and indicate a new error "CRL path validation
1472 error" in this case. Applications wanting additional details can use
1473 the verify callback and check the new "parent" field. If this is not
1474 NULL CRL path validation is taking place. Existing applications wont
1475 see this because it requires extended CRL support which is off by
1478 This work was sponsored by Google.
1481 *) Support for freshest CRL extension.
1483 This work was sponsored by Google.
1486 *) Initial indirect CRL support. Currently only supported in the CRLs
1487 passed directly and not via lookup. Process certificate issuer
1488 CRL entry extension and lookup CRL entries by bother issuer name
1489 and serial number. Check and process CRL issuer entry in IDP extension.
1491 This work was sponsored by Google.
1494 *) Add support for distinct certificate and CRL paths. The CRL issuer
1495 certificate is validated separately in this case. Only enabled if
1496 an extended CRL support flag is set: this flag will enable additional
1497 CRL functionality in future.
1499 This work was sponsored by Google.
1502 *) Add support for policy mappings extension.
1504 This work was sponsored by Google.
1507 *) Fixes to pathlength constraint, self issued certificate handling,
1508 policy processing to align with RFC3280 and PKITS tests.
1510 This work was sponsored by Google.
1513 *) Support for name constraints certificate extension. DN, email, DNS
1514 and URI types are currently supported.
1516 This work was sponsored by Google.
1519 *) To cater for systems that provide a pointer-based thread ID rather
1520 than numeric, deprecate the current numeric thread ID mechanism and
1521 replace it with a structure and associated callback type. This
1522 mechanism allows a numeric "hash" to be extracted from a thread ID in
1523 either case, and on platforms where pointers are larger than 'long',
1524 mixing is done to help ensure the numeric 'hash' is usable even if it
1525 can't be guaranteed unique. The default mechanism is to use "&errno"
1526 as a pointer-based thread ID to distinguish between threads.
1528 Applications that want to provide their own thread IDs should now use
1529 CRYPTO_THREADID_set_callback() to register a callback that will call
1530 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1532 Note that ERR_remove_state() is now deprecated, because it is tied
1533 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1534 to free the current thread's error state should be replaced by
1535 ERR_remove_thread_state(NULL).
1537 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1538 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1539 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1540 application was previously providing a numeric thread callback that
1541 was inappropriate for distinguishing threads, then uniqueness might
1542 have been obtained with &errno that happened immediately in the
1543 intermediate development versions of OpenSSL; this is no longer the
1544 case, the numeric thread callback will now override the automatic use
1546 [Geoff Thorpe, with help from Bodo Moeller]
1548 *) Initial support for different CRL issuing certificates. This covers a
1549 simple case where the self issued certificates in the chain exist and
1550 the real CRL issuer is higher in the existing chain.
1552 This work was sponsored by Google.
1555 *) Removed effectively defunct crypto/store from the build.
1558 *) Revamp of STACK to provide stronger type-checking. Still to come:
1559 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1560 ASN1_STRING, CONF_VALUE.
1563 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1564 RAM on SSL connections. This option can save about 34k per idle SSL.
1567 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1568 STACK, TXT_DB, bsearch, qsort.
1571 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1572 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1573 support for data, signedData, compressedData, digestedData and
1574 encryptedData, envelopedData types included. Scripts to check against
1575 RFC4134 examples draft and interop and consistency checks of many
1576 content types and variants.
1579 *) Add options to enc utility to support use of zlib compression BIO.
1582 *) Extend mk1mf to support importing of options and assembly language
1583 files from Configure script, currently only included in VC-WIN32.
1584 The assembly language rules can now optionally generate the source
1585 files from the associated perl scripts.
1588 *) Implement remaining functionality needed to support GOST ciphersuites.
1589 Interop testing has been performed using CryptoPro implementations.
1590 [Victor B. Wagner <vitus@cryptocom.ru>]
1592 *) s390x assembler pack.
1595 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1599 *) Implement Opaque PRF Input TLS extension as specified in
1600 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1601 official specification yet and no extension type assignment by
1602 IANA exists, this extension (for now) will have to be explicitly
1603 enabled when building OpenSSL by providing the extension number
1604 to use. For example, specify an option
1606 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1608 to the "config" or "Configure" script to enable the extension,
1609 assuming extension number 0x9527 (which is a completely arbitrary
1610 and unofficial assignment based on the MD5 hash of the Internet
1611 Draft). Note that by doing so, you potentially lose
1612 interoperability with other TLS implementations since these might
1613 be using the same extension number for other purposes.
1615 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1616 opaque PRF input value to use in the handshake. This will create
1617 an interal copy of the length-'len' string at 'src', and will
1618 return non-zero for success.
1620 To get more control and flexibility, provide a callback function
1623 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1624 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1628 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1631 Callback function 'cb' will be called in handshakes, and is
1632 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1633 Argument 'arg' is for application purposes (the value as given to
1634 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1635 be provided to the callback function). The callback function
1636 has to return non-zero to report success: usually 1 to use opaque
1637 PRF input just if possible, or 2 to enforce use of the opaque PRF
1638 input. In the latter case, the library will abort the handshake
1639 if opaque PRF input is not successfully negotiated.
1641 Arguments 'peerinput' and 'len' given to the callback function
1642 will always be NULL and 0 in the case of a client. A server will
1643 see the client's opaque PRF input through these variables if
1644 available (NULL and 0 otherwise). Note that if the server
1645 provides an opaque PRF input, the length must be the same as the
1646 length of the client's opaque PRF input.
1648 Note that the callback function will only be called when creating
1649 a new session (session resumption can resume whatever was
1650 previously negotiated), and will not be called in SSL 2.0
1651 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1652 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1653 for applications that need to enforce opaque PRF input.
1657 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1660 [Victor B. Wagner <vitus@cryptocom.ru>]
1662 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1663 RFC4507bis. The encrypted ticket format is an encrypted encoded
1664 SSL_SESSION structure, that way new session features are automatically
1667 If a client application caches session in an SSL_SESSION structure
1668 support is transparent because tickets are now stored in the encoded
1671 The SSL_CTX structure automatically generates keys for ticket
1672 protection in servers so again support should be possible
1673 with no application modification.
1675 If a client or server wishes to disable RFC4507 support then the option
1676 SSL_OP_NO_TICKET can be set.
1678 Add a TLS extension debugging callback to allow the contents of any client
1679 or server extensions to be examined.
1681 This work was sponsored by Google.
1684 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1685 OpenSSL should now compile cleanly on gcc 4.2
1686 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1688 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1689 support including streaming MAC support: this is required for GOST
1690 ciphersuite support.
1691 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1693 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1694 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1695 to output in BER and PEM format.
1698 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1699 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1700 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1701 ENGINE support for HMAC keys which are unextractable. New -mac and
1702 -macopt options to dgst utility.
1705 *) New option -sigopt to dgst utility. Update dgst to use
1706 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1707 alternative signing paramaters such as X9.31 or PSS in the dgst
1711 *) Change ssl_cipher_apply_rule(), the internal function that does
1712 the work each time a ciphersuite string requests enabling
1713 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1714 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1715 the order of disabled ciphersuites such that those ciphersuites
1716 that most recently went from enabled to disabled not only stay
1717 in order with respect to each other, but also have higher priority
1718 than other disabled ciphersuites the next time ciphersuites are
1721 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1722 the same ciphersuites as with "HIGH" alone, but in a specific
1723 order where the PSK ciphersuites come first (since they are the
1724 most recently disabled ciphersuites when "HIGH" is parsed).
1726 Also, change ssl_create_cipher_list() (using this new
1727 funcionality) such that between otherwise identical
1728 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1732 *) Change ssl_create_cipher_list() so that it automatically
1733 arranges the ciphersuites in reasonable order before starting
1734 to process the rule string. Thus, the definition for "DEFAULT"
1735 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1736 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1737 This makes it much easier to arrive at a reasonable default order
1738 in applications for which anonymous ciphers are OK (meaning
1739 that you can't actually use DEFAULT).
1740 [Bodo Moeller; suggested by Victor Duchovni]
1742 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1743 processing) into multiple integers instead of setting
1744 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1745 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1746 (These masks as well as the individual bit definitions are hidden
1747 away into the non-exported interface ssl/ssl_locl.h, so this
1748 change to the definition of the SSL_CIPHER structure shouldn't
1749 affect applications.) This give us more bits for each of these
1750 categories, so there is no longer a need to coagulate AES128 and
1751 AES256 into a single algorithm bit, and to coagulate Camellia128
1752 and Camellia256 into a single algorithm bit, which has led to all
1755 Thus, among other things, the kludge introduced in 0.9.7m and
1756 0.9.8e for masking out AES256 independently of AES128 or masking
1757 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1759 With the change, we also introduce new ciphersuite aliases that
1760 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1764 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1765 Use the leftmost N bytes of the signature input if the input is
1766 larger than the prime q (with N being the size in bytes of q).
1769 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1770 it yet and it is largely untested.
1773 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1776 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1777 some compilers (gcc 4.2 and later) reject their use. Safestack is
1778 reimplemented. Update ASN1 to avoid use of legacy functions.
1781 *) Win32/64 targets are linked with Winsock2.
1784 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1785 to external functions. This can be used to increase CRL handling
1786 efficiency especially when CRLs are very large by (for example) storing
1787 the CRL revoked certificates in a database.
1790 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1791 new CRLs added to a directory can be used. New command line option
1792 -verify_return_error to s_client and s_server. This causes real errors
1793 to be returned by the verify callback instead of carrying on no matter
1794 what. This reflects the way a "real world" verify callback would behave.
1797 *) GOST engine, supporting several GOST algorithms and public key formats.
1798 Kindly donated by Cryptocom.
1801 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1802 partitioned by DP are handled but no indirect CRL or reason partitioning
1803 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1804 selected via a scoring technique which handles IDP and AKID in CRLs.
1807 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1808 will ultimately be used for all verify operations: this will remove the
1809 X509_STORE dependency on certificate verification and allow alternative
1810 lookup methods. X509_STORE based implementations of these two callbacks.
1813 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1814 Modify get_crl() to find a valid (unexpired) CRL if possible.
1817 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1818 this would be called X509_CRL_cmp() but that name is already used by
1819 a function that just compares CRL issuer names. Cache several CRL
1820 extensions in X509_CRL structure and cache CRLDP in X509.
1823 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1824 this maps equivalent X509_NAME structures into a consistent structure.
1825 Name comparison can then be performed rapidly using memcmp().
1828 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1832 *) Allow digests to supply their own micalg string for S/MIME type using
1833 the ctrl EVP_MD_CTRL_MICALG.
1836 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1837 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1838 ctrl. It can then customise the structure before and/or after signing
1842 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1843 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1844 to free up any added signature OIDs.
1847 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1848 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1849 digest and cipher tables. New options added to openssl utility:
1850 list-message-digest-algorithms and list-cipher-algorithms.
1853 *) Change the array representation of binary polynomials: the list
1854 of degrees of non-zero coefficients is now terminated with -1.
1855 Previously it was terminated with 0, which was also part of the
1856 value; thus, the array representation was not applicable to
1857 polynomials where t^0 has coefficient zero. This change makes
1858 the array representation useful in a more general context.
1861 *) Various modifications and fixes to SSL/TLS cipher string
1862 handling. For ECC, the code now distinguishes between fixed ECDH
1863 with RSA certificates on the one hand and with ECDSA certificates
1864 on the other hand, since these are separate ciphersuites. The
1865 unused code for Fortezza ciphersuites has been removed.
1867 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1868 (not "ECDHE"). For consistency with the code for DH
1869 certificates, use of ECDH certificates is now considered ECDH
1870 authentication, not RSA or ECDSA authentication (the latter is
1871 merely the CA's signing algorithm and not actively used in the
1874 The temporary ciphersuite alias "ECCdraft" is no longer
1875 available, and ECC ciphersuites are no longer excluded from "ALL"
1876 and "DEFAULT". The following aliases now exist for RFC 4492
1877 ciphersuites, most of these by analogy with the DH case:
1879 kECDHr - ECDH cert, signed with RSA
1880 kECDHe - ECDH cert, signed with ECDSA
1881 kECDH - ECDH cert (signed with either RSA or ECDSA)
1882 kEECDH - ephemeral ECDH
1883 ECDH - ECDH cert or ephemeral ECDH
1889 AECDH - anonymous ECDH
1890 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1894 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1895 Use correct micalg parameters depending on digest(s) in signed message.
1898 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1899 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1902 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1903 an engine to register a method. Add ENGINE lookups for methods and
1904 functional reference processing.
1907 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1908 EVP_{Sign,Verify}* which allow an application to customise the signature
1912 *) New -resign option to smime utility. This adds one or more signers
1913 to an existing PKCS#7 signedData structure. Also -md option to use an
1914 alternative message digest algorithm for signing.
1917 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1918 create PKCS7 structures containing multiple signers. Update smime
1919 application to support multiple signers.
1922 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1926 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1927 Reorganize PBE internals to lookup from a static table using NIDs,
1928 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1929 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1930 PRF which will be automatically used with PBES2.
1933 *) Replace the algorithm specific calls to generate keys in "req" with the
1937 *) Update PKCS#7 enveloped data routines to use new API. This is now
1938 supported by any public key method supporting the encrypt operation. A
1939 ctrl is added to allow the public key algorithm to examine or modify
1940 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1944 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1945 a default digest type to use. In most cases this will be SHA1 but some
1946 algorithms (such as GOST) need to specify an alternative digest. The
1947 return value indicates how strong the prefernce is 1 means optional and
1948 2 is mandatory (that is it is the only supported type). Modify
1949 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1950 use the default md. Update openssl utilities to use the default digest
1951 type for signing if it is not explicitly indicated.
1954 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1955 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1956 signing method from the key type. This effectively removes the link
1957 between digests and public key types.
1960 *) Add an OID cross reference table and utility functions. Its purpose is to
1961 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1962 rsaEncryption. This will allow some of the algorithm specific hackery
1963 needed to use the correct OID to be removed.
1966 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1967 structures for PKCS7_sign(). They are now set up by the relevant public
1971 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1974 *) Add support for key derivation (agreement) in the API, DH method and
1978 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1979 public and private key formats. As a side effect these add additional
1980 command line functionality not previously available: DSA signatures can be
1981 generated and verified using pkeyutl and DH key support and generation in
1986 [Oliver Tappe <zooey@hirschkaefer.de>]
1988 *) New make target "install_html_docs" installs HTML renditions of the
1990 [Oliver Tappe <zooey@hirschkaefer.de>]
1992 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1993 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1994 support key and parameter generation and add initial key generation
1995 functionality for RSA.
1998 *) Add functions for main EVP_PKEY_method operations. The undocumented
1999 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2000 EVP_PKEY_{encrypt,decrypt}_old.
2003 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2004 key API, doesn't do much yet.
2007 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2008 public key algorithms. New option to openssl utility:
2009 "list-public-key-algorithms" to print out info.
2012 *) Implement the Supported Elliptic Curves Extension for
2013 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2016 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2017 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2020 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2021 utilities such as rsa, dsa, dsaparam etc except they process any key
2025 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2026 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2027 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2031 *) Initial support for pluggable public key ASN1.
2032 De-spaghettify the public key ASN1 handling. Move public and private
2033 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2034 algorithm specific handling to a single module within the relevant
2035 algorithm directory. Add functions to allow (near) opaque processing
2036 of public and private key structures.
2039 *) Implement the Supported Point Formats Extension for
2040 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2043 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2044 for the psk identity [hint] and the psk callback functions to the
2045 SSL_SESSION, SSL and SSL_CTX structure.
2048 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2052 SSL_CTX_use_psk_identity_hint
2053 SSL_get_psk_identity_hint
2054 SSL_get_psk_identity
2055 SSL_use_psk_identity_hint
2057 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2059 *) Add RFC 3161 compliant time stamp request creation, response generation
2060 and response verification functionality.
2061 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2063 *) Add initial support for TLS extensions, specifically for the server_name
2064 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2065 have new members for a host name. The SSL data structure has an
2066 additional member SSL_CTX *initial_ctx so that new sessions can be
2067 stored in that context to allow for session resumption, even after the
2068 SSL has been switched to a new SSL_CTX in reaction to a client's
2069 server_name extension.
2071 New functions (subject to change):
2073 SSL_get_servername()
2074 SSL_get_servername_type()
2077 New CTRL codes and macros (subject to change):
2079 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2080 - SSL_CTX_set_tlsext_servername_callback()
2081 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2082 - SSL_CTX_set_tlsext_servername_arg()
2083 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2085 openssl s_client has a new '-servername ...' option.
2087 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2088 '-key2 ...', '-servername_fatal' (subject to change). This allows
2089 testing the HostName extension for a specific single host name ('-cert'
2090 and '-key' remain fallbacks for handshakes without HostName
2091 negotiation). If the unrecogninzed_name alert has to be sent, this by
2092 default is a warning; it becomes fatal with the '-servername_fatal'
2095 [Peter Sylvester, Remy Allais, Christophe Renou]
2097 *) Whirlpool hash implementation is added.
2100 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2101 bn(64,32). Because of instruction set limitations it doesn't have
2102 any negative impact on performance. This was done mostly in order
2103 to make it possible to share assembler modules, such as bn_mul_mont
2104 implementations, between 32- and 64-bit builds without hassle.
2107 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2108 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2112 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2113 dedicated Montgomery multiplication procedure, is introduced.
2114 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2115 "64-bit" performance on certain 32-bit targets.
2118 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2119 in SSL structures. New SSL ctrl to set maximum send fragment size.
2120 Save memory by seeting the I/O buffer sizes dynamically instead of
2121 using the maximum available value.
2124 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2125 in addition to the text details.
2128 *) Very, very preliminary EXPERIMENTAL support for printing of general
2129 ASN1 structures. This currently produces rather ugly output and doesn't
2130 handle several customised structures at all.
2133 *) Integrated support for PVK file format and some related formats such
2134 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2135 these in the 'rsa' and 'dsa' utilities.
2138 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2141 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2142 place for the (very old) "NETSCAPE" format certificates which are now
2143 handled using new ASN1 code equivalents.
2146 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2147 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2148 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2151 *) Modify CRL distribution points extension code to print out previously
2152 unsupported fields. Enhance extension setting code to allow setting of
2156 *) Add print and set support for Issuing Distribution Point CRL extension.
2159 *) Change 'Configure' script to enable Camellia by default.
2162 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
2164 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2166 This addresses the flaw in CBC record processing discovered by
2167 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2168 at: http://www.isg.rhul.ac.uk/tls/
2170 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2171 Security Group at Royal Holloway, University of London
2172 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2173 Emilia Käsper for the initial patch.
2175 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2177 *) Return an error when checking OCSP signatures when key is NULL.
2178 This fixes a DoS attack. (CVE-2013-0166)
2181 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2182 the right response is stapled. Also change SSL_get_certificate()
2183 so it returns the certificate actually sent.
2184 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2185 (This is a backport)
2186 [Rob Stradling <rob.stradling@comodo.com>]
2188 *) Fix possible deadlock when decoding public keys.
2191 Changes between 0.9.8w and 0.9.8x [10 May 2012]
2193 *) Sanity check record length before skipping explicit IV in DTLS
2196 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2197 fuzzing as a service testing platform.
2201 *) Initialise tkeylen properly when encrypting CMS messages.
2202 Thanks to Solar Designer of Openwall for reporting this issue.
2205 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
2207 *) The fix for CVE-2012-2110 did not take into account that the
2208 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
2209 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
2210 rejecting negative len parameter. (CVE-2012-2131)
2211 [Tomas Hoger <thoger@redhat.com>]
2213 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
2215 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2216 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2217 in CRYPTO_realloc_clean.
2219 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2220 issue and to Adam Langley <agl@chromium.org> for fixing it.
2222 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2224 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
2226 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2227 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2228 content decryption and always return the same error. Note: this attack
2229 needs on average 2^20 messages so it only affects automated senders. The
2230 old behaviour can be reenabled in the CMS code by setting the
2231 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2232 an MMA defence is not necessary.
2233 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2234 this issue. (CVE-2012-0884)
2237 *) Fix CVE-2011-4619: make sure we really are receiving a
2238 client hello before rejecting multiple SGC restarts. Thanks to
2239 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2242 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
2244 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2245 Thanks to Antonio Martin, Enterprise Secure Access Research and
2246 Development, Cisco Systems, Inc. for discovering this bug and
2247 preparing a fix. (CVE-2012-0050)
2250 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
2252 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2253 of the Vaudenay padding oracle attack on CBC mode encryption
2254 which enables an efficient plaintext recovery attack against
2255 the OpenSSL implementation of DTLS. Their attack exploits timing
2256 differences arising during decryption processing. A research
2257 paper describing this attack can be found at:
2258 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2259 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2260 Security Group at Royal Holloway, University of London
2261 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2262 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2263 for preparing the fix. (CVE-2011-4108)
2264 [Robin Seggelmann, Michael Tuexen]
2266 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
2267 [Ben Laurie, Kasper <ekasper@google.com>]
2269 *) Clear bytes used for block padding of SSL 3.0 records.
2271 [Adam Langley (Google)]
2273 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2274 Kadianakis <desnacked@gmail.com> for discovering this issue and
2275 Adam Langley for preparing the fix. (CVE-2011-4619)
2276 [Adam Langley (Google)]
2278 *) Prevent malformed RFC3779 data triggering an assertion failure.
2279 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2280 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2281 [Rob Austein <sra@hactrn.net>]
2283 *) Fix ssl_ciph.c set-up race.
2284 [Adam Langley (Google)]
2286 *) Fix spurious failures in ecdsatest.c.
2287 [Emilia Käsper (Google)]
2289 *) Fix the BIO_f_buffer() implementation (which was mixing different
2290 interpretations of the '..._len' fields).
2291 [Adam Langley (Google)]
2293 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2294 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2295 threads won't reuse the same blinding coefficients.
2297 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2298 lock to call BN_BLINDING_invert_ex, and avoids one use of
2299 BN_BLINDING_update for each BN_BLINDING structure (previously,
2300 the last update always remained unused).
2301 [Emilia Käsper (Google)]
2303 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2304 for multi-threaded use of ECDH.
2305 [Adam Langley (Google)]
2307 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2310 *) Add protection against ECDSA timing attacks as mentioned in the paper
2311 by Billy Bob Brumley and Nicola Tuveri, see:
2313 http://eprint.iacr.org/2011/232.pdf
2315 [Billy Bob Brumley and Nicola Tuveri]
2317 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
2319 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2320 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2322 *) Fix bug in string printing code: if *any* escaping is enabled we must
2323 escape the escape character (backslash) or the resulting string is
2327 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
2329 *) Disable code workaround for ancient and obsolete Netscape browsers
2330 and servers: an attacker can use it in a ciphersuite downgrade attack.
2331 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2334 *) Fixed J-PAKE implementation error, originally discovered by
2335 Sebastien Martini, further info and confirmation from Stefan
2336 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2339 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
2341 *) Fix extension code to avoid race conditions which can result in a buffer
2342 overrun vulnerability: resumed sessions must not be modified as they can
2343 be shared by multiple threads. CVE-2010-3864
2346 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
2349 *) Don't reencode certificate when calculating signature: cache and use
2350 the original encoding instead. This makes signature verification of
2351 some broken encodings work correctly.
2354 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
2355 is also one of the inputs.
2356 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2358 *) Don't repeatedly append PBE algorithms to table if they already exist.
2359 Sort table on each new add. This effectively makes the table read only
2360 after all algorithms are added and subsequent calls to PKCS12_pbe_add
2364 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
2366 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
2369 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
2370 access or freeing data twice (CVE-2010-0742)
2371 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
2373 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
2374 common in certificates and some applications which only call
2375 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
2379 Reduce copying into .apps and .test in makevms.com
2380 Don't try to use blank CA certificate in CA.com
2381 Allow use of C files from original directories in maketests.com
2382 [Steven M. Schweda" <sms@antinode.info>]
2384 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2386 *) When rejecting SSL/TLS records due to an incorrect version number, never
2387 update s->server with a new major version number. As of
2388 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2389 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2390 the previous behavior could result in a read attempt at NULL when
2391 receiving specific incorrect SSL/TLS records once record payload
2392 protection is active. (CVE-2010-0740)
2393 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2395 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2396 could be crashed if the relevant tables were not present (e.g. chrooted).
2397 [Tomas Hoger <thoger@redhat.com>]
2399 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2401 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2402 [Martin Olsson, Neel Mehta]
2404 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2405 accommodate for stack sorting, always a write lock!).
2408 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2409 excessive delays in the RAND_poll(): over a minute. As a workaround
2410 include a time check in the inner Heap32Next loop too.
2413 *) The code that handled flushing of data in SSL/TLS originally used the
2414 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2415 the problem outlined in PR#1949. The fix suggested there however can
2416 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2417 of Apache). So instead simplify the code to flush unconditionally.
2418 This should be fine since flushing with no data to flush is a no op.
2421 *) Handle TLS versions 2.0 and later properly and correctly use the
2422 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2423 off ancient servers have a habit of sticking around for a while...
2426 *) Modify compression code so it frees up structures without using the
2427 ex_data callbacks. This works around a problem where some applications
2428 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2429 restarting) then use compression (e.g. SSL with compression) later.
2430 This results in significant per-connection memory leaks and
2431 has caused some security issues including CVE-2008-1678 and
2435 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2436 change when encrypting or decrypting.
2439 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2440 connect and renegotiate with servers which do not support RI.
2441 Until RI is more widely deployed this option is enabled by default.
2444 *) Add "missing" ssl ctrls to clear options and mode.
2447 *) If client attempts to renegotiate and doesn't support RI respond with
2448 a no_renegotiation alert as required by RFC5746. Some renegotiating
2449 TLS clients will continue a connection gracefully when they receive
2450 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2451 waiting for a server hello which it will never receive. Now we treat a
2452 received no_renegotiation alert as a fatal error. This is because
2453 applications requesting a renegotiation might well expect it to succeed
2454 and would have no code in place to handle the server denying it so the
2455 only safe thing to do is to terminate the connection.
2458 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2459 peer supports secure renegotiation and 0 otherwise. Print out peer
2460 renegotiation support in s_client/s_server.
2463 *) Replace the highly broken and deprecated SPKAC certification method with
2464 the updated NID creation version. This should correctly handle UTF8.
2467 *) Implement RFC5746. Re-enable renegotiation but require the extension
2468 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2469 turns out to be a bad idea. It has been replaced by
2470 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2471 SSL_CTX_set_options(). This is really not recommended unless you
2472 know what you are doing.
2473 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2475 *) Fixes to stateless session resumption handling. Use initial_ctx when
2476 issuing and attempting to decrypt tickets in case it has changed during
2477 servername handling. Use a non-zero length session ID when attempting
2478 stateless session resumption: this makes it possible to determine if
2479 a resumption has occurred immediately after receiving server hello
2480 (several places in OpenSSL subtly assume this) instead of later in
2484 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2485 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2486 fixes for a few places where the return code is not checked
2488 [Julia Lawall <julia@diku.dk>]
2490 *) Add --strict-warnings option to Configure script to include devteam
2491 warnings in other configurations.
2494 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2495 makes it possible to install openssl libraries in locations which
2496 have names other than "lib", for example "/usr/lib64" which some
2498 [Steve Henson, based on patch from Jeremy Utley]
2500 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2501 X690 8.9.12 and can produce some misleading textual output of OIDs.
2502 [Steve Henson, reported by Dan Kaminsky]
2504 *) Delete MD2 from algorithm tables. This follows the recommendation in
2505 several standards that it is not used in new applications due to
2506 several cryptographic weaknesses. For binary compatibility reasons
2507 the MD2 API is still compiled in by default.
2510 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2514 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2515 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2517 [Guenter <lists@gknw.net>]
2519 *) Fix the server certificate chain building code to use X509_verify_cert(),
2520 it used to have an ad-hoc builder which was unable to cope with anything
2521 other than a simple chain.
2522 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2524 *) Don't check self signed certificate signatures in X509_verify_cert()
2525 by default (a flag can override this): it just wastes time without
2526 adding any security. As a useful side effect self signed root CAs
2527 with non-FIPS digests are now usable in FIPS mode.
2530 *) In dtls1_process_out_of_seq_message() the check if the current message
2531 is already buffered was missing. For every new message was memory
2532 allocated, allowing an attacker to perform an denial of service attack
2533 with sending out of seq handshake messages until there is no memory
2534 left. Additionally every future messege was buffered, even if the
2535 sequence number made no sense and would be part of another handshake.
2536 So only messages with sequence numbers less than 10 in advance will be
2537 buffered. (CVE-2009-1378)
2538 [Robin Seggelmann, discovered by Daniel Mentz]
2540 *) Records are buffered if they arrive with a future epoch to be
2541 processed after finishing the corresponding handshake. There is
2542 currently no limitation to this buffer allowing an attacker to perform
2543 a DOS attack with sending records with future epochs until there is no
2544 memory left. This patch adds the pqueue_size() function to detemine
2545 the size of a buffer and limits the record buffer to 100 entries.
2547 [Robin Seggelmann, discovered by Daniel Mentz]
2549 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2550 parent structure is freed. (CVE-2009-1379)
2553 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2554 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2557 [Ilya O. <vrghost@gmail.com>]
2559 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2561 *) Disable renegotiation completely - this fixes a severe security
2562 problem (CVE-2009-3555) at the cost of breaking all
2563 renegotiation. Renegotiation can be re-enabled by setting
2564 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2565 run-time. This is really not recommended unless you know what
2569 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2571 *) Don't set val to NULL when freeing up structures, it is freed up by
2572 underlying code. If sizeof(void *) > sizeof(long) this can result in
2573 zeroing past the valid field. (CVE-2009-0789)
2574 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2576 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2577 checked correctly. This would allow some invalid signed attributes to
2578 appear to verify correctly. (CVE-2009-0591)
2579 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2581 *) Reject UniversalString and BMPString types with invalid lengths. This
2582 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2583 a legal length. (CVE-2009-0590)
2586 *) Set S/MIME signing as the default purpose rather than setting it
2587 unconditionally. This allows applications to override it at the store
2591 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2592 to handle some structures.
2595 *) Improve efficiency of mem_gets: don't search whole buffer each time
2597 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2599 *) New -hex option for openssl rand.
2602 *) Print out UTF8String and NumericString when parsing ASN1.
2605 *) Support NumericString type for name components.
2608 *) Allow CC in the environment to override the automatically chosen
2609 compiler. Note that nothing is done to ensure flags work with the
2613 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2615 *) Properly check EVP_VerifyFinal() and similar return values
2617 [Ben Laurie, Bodo Moeller, Google Security Team]
2619 *) Enable TLS extensions by default.
2622 *) Allow the CHIL engine to be loaded, whether the application is
2623 multithreaded or not. (This does not release the developer from the
2624 obligation to set up the dynamic locking callbacks.)
2625 [Sander Temme <sander@temme.net>]
2627 *) Use correct exit code if there is an error in dgst command.
2628 [Steve Henson; problem pointed out by Roland Dirlewanger]
2630 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2631 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2634 *) Add experimental JPAKE support, including demo authentication in
2635 s_client and s_server.
2638 *) Set the comparison function in v3_addr_canonize().
2639 [Rob Austein <sra@hactrn.net>]
2641 *) Add support for XMPP STARTTLS in s_client.
2642 [Philip Paeps <philip@freebsd.org>]
2644 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2645 to ensure that even with this option, only ciphersuites in the
2646 server's preference list will be accepted. (Note that the option
2647 applies only when resuming a session, so the earlier behavior was
2648 just about the algorithm choice for symmetric cryptography.)
2651 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2653 *) Fix NULL pointer dereference if a DTLS server received
2654 ChangeCipherSpec as first record (CVE-2009-1386).
2657 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2658 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2661 *) The fix in 0.9.8c that supposedly got rid of unsafe
2662 double-checked locking was incomplete for RSA blinding,
2663 addressing just one layer of what turns out to have been
2664 doubly unsafe triple-checked locking.
2666 So now fix this for real by retiring the MONT_HELPER macro
2667 in crypto/rsa/rsa_eay.c.
2669 [Bodo Moeller; problem pointed out by Marius Schilder]
2671 *) Various precautionary measures:
2673 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2675 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2676 (NB: This would require knowledge of the secret session ticket key
2677 to exploit, in which case you'd be SOL either way.)
2679 - Change bn_nist.c so that it will properly handle input BIGNUMs
2680 outside the expected range.
2682 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2685 [Neel Mehta, Bodo Moeller]
2687 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2688 the load fails. Useful for distros.
2689 [Ben Laurie and the FreeBSD team]
2691 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2694 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2697 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2699 This work was sponsored by Logica.
2702 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2703 keystores. Support for SSL/TLS client authentication too.
2704 Not compiled unless enable-capieng specified to Configure.
2706 This work was sponsored by Logica.
2709 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2710 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2711 attribute creation routines such as certifcate requests and PKCS#12
2715 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2717 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2718 handshake which could lead to a cilent crash as found using the
2719 Codenomicon TLS test suite (CVE-2008-1672)
2720 [Steve Henson, Mark Cox]
2722 *) Fix double free in TLS server name extensions which could lead to
2723 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2726 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2728 Clear the error queue to ensure that error entries left from
2729 older function calls do not interfere with the correct operation.
2730 [Lutz Jaenicke, Erik de Castro Lopo]
2732 *) Remove root CA certificates of commercial CAs:
2734 The OpenSSL project does not recommend any specific CA and does not
2735 have any policy with respect to including or excluding any CA.
2736 Therefore it does not make any sense to ship an arbitrary selection
2737 of root CA certificates with the OpenSSL software.
2740 *) RSA OAEP patches to fix two separate invalid memory reads.
2741 The first one involves inputs when 'lzero' is greater than
2742 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2743 before the beginning of from). The second one involves inputs where
2744 the 'db' section contains nothing but zeroes (there is a one-byte
2745 invalid read after the end of 'db').
2746 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2748 *) Partial backport from 0.9.9-dev:
2750 Introduce bn_mul_mont (dedicated Montgomery multiplication
2751 procedure) as a candidate for BIGNUM assembler implementation.
2752 While 0.9.9-dev uses assembler for various architectures, only
2753 x86_64 is available by default here in the 0.9.8 branch, and
2754 32-bit x86 is available through a compile-time setting.
2756 To try the 32-bit x86 assembler implementation, use Configure
2757 option "enable-montasm" (which exists only for this backport).
2759 As "enable-montasm" for 32-bit x86 disclaims code stability
2760 anyway, in this constellation we activate additional code
2761 backported from 0.9.9-dev for further performance improvements,
2762 namely BN_from_montgomery_word. (To enable this otherwise,
2763 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2765 [Andy Polyakov (backport partially by Bodo Moeller)]
2767 *) Add TLS session ticket callback. This allows an application to set
2768 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2769 values. This is useful for key rollover for example where several key
2770 sets may exist with different names.
2773 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2774 This was broken until now in 0.9.8 releases, such that the only way
2775 a registered ENGINE could be used (assuming it initialises
2776 successfully on the host) was to explicitly set it as the default
2777 for the relevant algorithms. This is in contradiction with 0.9.7
2778 behaviour and the documentation. With this fix, when an ENGINE is
2779 registered into a given algorithm's table of implementations, the
2780 'uptodate' flag is reset so that auto-discovery will be used next
2781 time a new context for that algorithm attempts to select an
2783 [Ian Lister (tweaked by Geoff Thorpe)]
2785 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2786 implemention in the following ways:
2788 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2791 Lack of BER streaming support means one pass streaming processing is
2792 only supported if data is detached: setting the streaming flag is
2793 ignored for embedded content.
2795 CMS support is disabled by default and must be explicitly enabled
2796 with the enable-cms configuration option.
2799 *) Update the GMP engine glue to do direct copies between BIGNUM and
2800 mpz_t when openssl and GMP use the same limb size. Otherwise the
2801 existing "conversion via a text string export" trick is still used.
2802 [Paul Sheer <paulsheer@gmail.com>]
2804 *) Zlib compression BIO. This is a filter BIO which compressed and
2805 uncompresses any data passed through it.
2808 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2809 RFC3394 compatible AES key wrapping.
2812 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2813 sets string data without copying. X509_ALGOR_set0() and
2814 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2815 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2816 from an X509_ATTRIBUTE structure optionally checking it occurs only
2817 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2821 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2822 to get the expected BN_FLG_CONSTTIME behavior.
2823 [Bodo Moeller (Google)]
2827 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2828 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2829 - added some more tests to do_tests.pl
2830 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2831 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2832 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2833 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2834 - various changes to netware.pl to enable gcc-cross builds on Win32
2836 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2837 - various changes to fix missing prototype warnings
2838 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2839 - added AES, WHIRLPOOL and CPUID assembler code to build files
2840 - added missing AES assembler make rules to mk1mf.pl
2841 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2842 [Guenter Knauf <eflash@gmx.net>]
2844 *) Implement certificate status request TLS extension defined in RFC3546.
2845 A client can set the appropriate parameters and receive the encoded
2846 OCSP response via a callback. A server can query the supplied parameters
2847 and set the encoded OCSP response in the callback. Add simplified examples
2848 to s_client and s_server.
2851 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2853 *) Fix various bugs:
2854 + Binary incompatibility of ssl_ctx_st structure
2855 + DTLS interoperation with non-compliant servers
2856 + Don't call get_session_cb() without proposed session
2857 + Fix ia64 assembler code
2858 [Andy Polyakov, Steve Henson]
2860 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2862 *) DTLS Handshake overhaul. There were longstanding issues with
2863 OpenSSL DTLS implementation, which were making it impossible for
2864 RFC 4347 compliant client to communicate with OpenSSL server.
2865 Unfortunately just fixing these incompatibilities would "cut off"
2866 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2867 server keeps tolerating non RFC compliant syntax. The opposite is
2868 not true, 0.9.8f client can not communicate with earlier server.
2869 This update even addresses CVE-2007-4995.
2872 *) Changes to avoid need for function casts in OpenSSL: some compilers
2873 (gcc 4.2 and later) reject their use.
2874 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2877 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2878 RFC4507bis. The encrypted ticket format is an encrypted encoded
2879 SSL_SESSION structure, that way new session features are automatically
2882 If a client application caches session in an SSL_SESSION structure
2883 support is transparent because tickets are now stored in the encoded
2886 The SSL_CTX structure automatically generates keys for ticket
2887 protection in servers so again support should be possible
2888 with no application modification.
2890 If a client or server wishes to disable RFC4507 support then the option
2891 SSL_OP_NO_TICKET can be set.
2893 Add a TLS extension debugging callback to allow the contents of any client
2894 or server extensions to be examined.
2896 This work was sponsored by Google.
2899 *) Add initial support for TLS extensions, specifically for the server_name
2900 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2901 have new members for a host name. The SSL data structure has an
2902 additional member SSL_CTX *initial_ctx so that new sessions can be
2903 stored in that context to allow for session resumption, even after the
2904 SSL has been switched to a new SSL_CTX in reaction to a client's
2905 server_name extension.
2907 New functions (subject to change):
2909 SSL_get_servername()
2910 SSL_get_servername_type()
2913 New CTRL codes and macros (subject to change):
2915 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2916 - SSL_CTX_set_tlsext_servername_callback()
2917 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2918 - SSL_CTX_set_tlsext_servername_arg()
2919 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2921 openssl s_client has a new '-servername ...' option.
2923 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2924 '-key2 ...', '-servername_fatal' (subject to change). This allows
2925 testing the HostName extension for a specific single host name ('-cert'
2926 and '-key' remain fallbacks for handshakes without HostName
2927 negotiation). If the unrecogninzed_name alert has to be sent, this by
2928 default is a warning; it becomes fatal with the '-servername_fatal'
2931 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2933 *) Add AES and SSE2 assembly language support to VC++ build.
2936 *) Mitigate attack on final subtraction in Montgomery reduction.
2939 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2940 (which previously caused an internal error).
2943 *) Squeeze another 10% out of IGE mode when in != out.
2946 *) AES IGE mode speedup.
2947 [Dean Gaudet (Google)]
2949 *) Add the Korean symmetric 128-bit cipher SEED (see
2950 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2951 add SEED ciphersuites from RFC 4162:
2953 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2954 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2955 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2956 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2958 To minimize changes between patchlevels in the OpenSSL 0.9.8
2959 series, SEED remains excluded from compilation unless OpenSSL
2960 is configured with 'enable-seed'.
2961 [KISA, Bodo Moeller]
2963 *) Mitigate branch prediction attacks, which can be practical if a
2964 single processor is shared, allowing a spy process to extract
2965 information. For detailed background information, see
2966 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2967 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2968 and Necessary Software Countermeasures"). The core of the change
2969 are new versions BN_div_no_branch() and
2970 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2971 respectively, which are slower, but avoid the security-relevant
2972 conditional branches. These are automatically called by BN_div()
2973 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2974 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2975 remove a conditional branch.
2977 BN_FLG_CONSTTIME is the new name for the previous
2978 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2979 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2980 in the exponent causes BN_mod_exp_mont() to use the alternative
2981 implementation in BN_mod_exp_mont_consttime().) The old name
2982 remains as a deprecated alias.
2984 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2985 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2986 constant-time implementations for more than just exponentiation.
2987 Here too the old name is kept as a deprecated alias.
2989 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2990 the BN_BLINDING structure gets an independent copy of the
2991 modulus. This means that the previous "BIGNUM *m" argument to
2992 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2993 essentially becomes "const BIGNUM *m", although we can't actually
2994 change this in the header file before 0.9.9. It allows
2995 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2996 enable BN_FLG_CONSTTIME.
2998 [Matthew D Wood (Intel Corp)]
3000 *) In the SSL/TLS server implementation, be strict about session ID
3001 context matching (which matters if an application uses a single
3002 external cache for different purposes). Previously,
3003 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3004 set. This did ensure strict client verification, but meant that,
3005 with applications using a single external cache for quite
3006 different requirements, clients could circumvent ciphersuite
3007 restrictions for a given session ID context by starting a session
3008 in a different context.
3011 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3012 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3013 authentication-only ciphersuites.
3016 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3017 not complete and could lead to a possible single byte overflow
3018 (CVE-2007-5135) [Ben Laurie]
3020 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3022 *) Since AES128 and AES256 (and similarly Camellia128 and
3023 Camellia256) share a single mask bit in the logic of
3024 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3025 kludge to work properly if AES128 is available and AES256 isn't
3026 (or if Camellia128 is available and Camellia256 isn't).
3029 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3030 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3031 When a point or a seed is encoded in a BIT STRING, we need to
3032 prevent the removal of trailing zero bits to get the proper DER
3033 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3034 of a NamedBitList, for which trailing 0 bits need to be removed.)
3037 *) Have SSL/TLS server implementation tolerate "mismatched" record
3038 protocol version while receiving ClientHello even if the
3039 ClientHello is fragmented. (The server can't insist on the
3040 particular protocol version it has chosen before the ServerHello
3041 message has informed the client about his choice.)
3044 *) Add RFC 3779 support.
3045 [Rob Austein for ARIN, Ben Laurie]
3047 *) Load error codes if they are not already present instead of using a
3048 static variable. This allows them to be cleanly unloaded and reloaded.
3049 Improve header file function name parsing.
3052 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3053 or CAPABILITY handshake as required by RFCs.
3056 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3058 *) Introduce limits to prevent malicious keys being able to
3059 cause a denial of service. (CVE-2006-2940)
3060 [Steve Henson, Bodo Moeller]
3062 *) Fix ASN.1 parsing of certain invalid structures that can result
3063 in a denial of service. (CVE-2006-2937) [Steve Henson]
3065 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3066 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3068 *) Fix SSL client code which could crash if connecting to a
3069 malicious SSLv2 server. (CVE-2006-4343)
3070 [Tavis Ormandy and Will Drewry, Google Security Team]
3072 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3073 match only those. Before that, "AES256-SHA" would be interpreted
3074 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3075 the same strength classification in 0.9.7h) as we currently only
3076 have a single AES bit in the ciphersuite description bitmap.
3077 That change, however, also applied to ciphersuite strings such as
3078 "RC4-MD5" that intentionally matched multiple ciphersuites --
3079 namely, SSL 2.0 ciphersuites in addition to the more common ones
3080 from SSL 3.0/TLS 1.0.
3082 So we change the selection algorithm again: Naming an explicit
3083 ciphersuite selects this one ciphersuite, and any other similar
3084 ciphersuite (same bitmap) from *other* protocol versions.
3085 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3086 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3088 Since SSL 2.0 does not have any ciphersuites for which the
3089 128/256 bit distinction would be relevant, this works for now.
3090 The proper fix will be to use different bits for AES128 and
3091 AES256, which would have avoided the problems from the beginning;
3092 however, bits are scarce, so we can only do this in a new release
3093 (not just a patchlevel) when we can change the SSL_CIPHER
3094 definition to split the single 'unsigned long mask' bitmap into
3095 multiple values to extend the available space.
3099 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3101 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3102 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3104 *) Add AES IGE and biIGE modes.
3107 *) Change the Unix randomness entropy gathering to use poll() when
3108 possible instead of select(), since the latter has some
3109 undesirable limitations.
3110 [Darryl Miles via Richard Levitte and Bodo Moeller]
3112 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3113 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3114 cannot be implicitly activated as part of, e.g., the "AES" alias.
3115 However, please upgrade to OpenSSL 0.9.9[-dev] for
3116 non-experimental use of the ECC ciphersuites to get TLS extension
3117 support, which is required for curve and point format negotiation
3118 to avoid potential handshake problems.
3121 *) Disable rogue ciphersuites:
3123 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3124 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3125 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3127 The latter two were purportedly from
3128 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3131 Also deactivate the remaining ciphersuites from
3132 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3133 unofficial, and the ID has long expired.
3136 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3137 dual-core machines) and other potential thread-safety issues.
3140 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3141 versions), which is now available for royalty-free use
3142 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3143 Also, add Camellia TLS ciphersuites from RFC 4132.
3145 To minimize changes between patchlevels in the OpenSSL 0.9.8
3146 series, Camellia remains excluded from compilation unless OpenSSL
3147 is configured with 'enable-camellia'.
3150 *) Disable the padding bug check when compression is in use. The padding
3151 bug check assumes the first packet is of even length, this is not
3152 necessarily true if compresssion is enabled and can result in false
3153 positives causing handshake failure. The actual bug test is ancient
3154 code so it is hoped that implementations will either have fixed it by
3155 now or any which still have the bug do not support compression.
3158 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3160 *) When applying a cipher rule check to see if string match is an explicit
3161 cipher suite and only match that one cipher suite if it is.
3164 *) Link in manifests for VC++ if needed.
3165 [Austin Ziegler <halostatue@gmail.com>]
3167 *) Update support for ECC-based TLS ciphersuites according to
3168 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3169 TLS extensions, which are supported starting with the 0.9.9
3170 branch, not in the OpenSSL 0.9.8 branch).
3173 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3174 opaque EVP_CIPHER_CTX handling.
3177 *) Fixes and enhancements to zlib compression code. We now only use
3178 "zlib1.dll" and use the default __cdecl calling convention on Win32
3179 to conform with the standards mentioned here:
3180 http://www.zlib.net/DLL_FAQ.txt
3181 Static zlib linking now works on Windows and the new --with-zlib-include
3182 --with-zlib-lib options to Configure can be used to supply the location
3183 of the headers and library. Gracefully handle case where zlib library
3187 *) Several fixes and enhancements to the OID generation code. The old code
3188 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3189 handle numbers larger than ULONG_MAX, truncated printing and had a
3190 non standard OBJ_obj2txt() behaviour.
3193 *) Add support for building of engines under engine/ as shared libraries
3194 under VC++ build system.
3197 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3198 Hopefully, we will not see any false combination of paths any more.
3201 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3203 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3204 (part of SSL_OP_ALL). This option used to disable the
3205 countermeasure against man-in-the-middle protocol-version
3206 rollback in the SSL 2.0 server implementation, which is a bad
3207 idea. (CVE-2005-2969)
3209 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3210 for Information Security, National Institute of Advanced Industrial
3211 Science and Technology [AIST], Japan)]
3213 *) Add two function to clear and return the verify parameter flags.
3216 *) Keep cipherlists sorted in the source instead of sorting them at
3217 runtime, thus removing the need for a lock.
3220 *) Avoid some small subgroup attacks in Diffie-Hellman.
3221 [Nick Mathewson and Ben Laurie]
3223 *) Add functions for well-known primes.
3226 *) Extended Windows CE support.
3227 [Satoshi Nakamura and Andy Polyakov]
3229 *) Initialize SSL_METHOD structures at compile time instead of during
3230 runtime, thus removing the need for a lock.
3233 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3234 attempting to decrypt each encrypted key in turn. Add support to
3238 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3240 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3243 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3246 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3247 key into the same file any more.
3250 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3253 *) Add -utf8 command line and config file option to 'ca'.
3254 [Stefan <stf@udoma.org]
3256 *) Removed the macro des_crypt(), as it seems to conflict with some
3257 libraries. Use DES_crypt().
3260 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3261 involves renaming the source and generated shared-libs for
3262 both. The engines will accept the corrected or legacy ids
3263 ('ncipher' and '4758_cca' respectively) when binding. NB,
3264 this only applies when building 'shared'.
3265 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3267 *) Add attribute functions to EVP_PKEY structure. Modify
3268 PKCS12_create() to recognize a CSP name attribute and
3269 use it. Make -CSP option work again in pkcs12 utility.
3272 *) Add new functionality to the bn blinding code:
3273 - automatic re-creation of the BN_BLINDING parameters after
3274 a fixed number of uses (currently 32)
3275 - add new function for parameter creation
3276 - introduce flags to control the update behaviour of the
3277 BN_BLINDING parameters
3278 - hide BN_BLINDING structure
3279 Add a second BN_BLINDING slot to the RSA structure to improve
3280 performance when a single RSA object is shared among several
3284 *) Add support for DTLS.
3285 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3287 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3288 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3291 *) Remove buggy and incompletet DH cert support from
3292 ssl/ssl_rsa.c and ssl/s3_both.c
3295 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3296 the apps/openssl applications.
3299 *) Compile clean with "-Wall -Wmissing-prototypes
3300 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3301 DEBUG_SAFESTACK must also be set.
3304 *) Change ./Configure so that certain algorithms can be disabled by default.
3305 The new counterpiece to "no-xxx" is "enable-xxx".
3307 The patented RC5 and MDC2 algorithms will now be disabled unless
3308 "enable-rc5" and "enable-mdc2", respectively, are specified.
3310 (IDEA remains enabled despite being patented. This is because IDEA
3311 is frequently required for interoperability, and there is no license
3312 fee for non-commercial use. As before, "no-idea" can be used to
3313 avoid this algorithm.)
3317 *) Add processing of proxy certificates (see RFC 3820). This work was
3318 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3319 EGEE (Enabling Grids for E-science in Europe).
3322 *) RC4 performance overhaul on modern architectures/implementations, such
3323 as Intel P4, IA-64 and AMD64.
3326 *) New utility extract-section.pl. This can be used specify an alternative
3327 section number in a pod file instead of having to treat each file as
3328 a separate case in Makefile. This can be done by adding two lines to the
3331 =for comment openssl_section:XXX
3333 The blank line is mandatory.
3337 *) New arguments -certform, -keyform and -pass for s_client and s_server
3338 to allow alternative format key and certificate files and passphrase
3342 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3343 update associated structures and add various utility functions.
3345 Add new policy related verify parameters, include policy checking in
3346 standard verify code. Enhance 'smime' application with extra parameters
3347 to support policy checking and print out.
3350 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3351 Nehemiah processors. These extensions support AES encryption in hardware
3352 as well as RNG (though RNG support is currently disabled).
3353 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3355 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3358 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3359 [Andy Polyakov and a number of other people]
3361 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3362 implementation contributed by IBM.
3363 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3365 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3366 exponent rather than 'unsigned long'. There is a corresponding change to
3367 the new 'rsa_keygen' element of the RSA_METHOD structure.
3368 [Jelte Jansen, Geoff Thorpe]
3370 *) Functionality for creating the initial serial number file is now
3371 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3373 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3374 number file to 1, which is bound to cause problems. To avoid
3375 the problems while respecting compatibility between different 0.9.7
3376 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3377 CA.pl for serial number initialization. With the new release 0.9.8,
3378 we can fix the problem directly in the 'ca' utility.)
3381 *) Reduced header interdepencies by declaring more opaque objects in
3382 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3383 give fewer recursive includes, which could break lazy source code - so
3384 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3385 developers should define this symbol when building and using openssl to
3386 ensure they track the recommended behaviour, interfaces, [etc], but
3387 backwards-compatible behaviour prevails when this isn't defined.
3390 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3393 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3394 This will generate a random key of the appropriate length based on the
3395 cipher context. The EVP_CIPHER can provide its own random key generation
3396 routine to support keys of a specific form. This is used in the des and
3397 3des routines to generate a key of the correct parity. Update S/MIME
3398 code to use new functions and hence generate correct parity DES keys.
3399 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3400 valid (weak or incorrect parity).
3403 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3404 as looking them up. This is useful when the verified structure may contain
3405 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3406 present unless the new PKCS7_NO_CRL flag is asserted.
3409 *) Extend ASN1 oid configuration module. It now additionally accepts the
3412 shortName = some long name, 1.2.3.4
3415 *) Reimplemented the BN_CTX implementation. There is now no more static
3416 limitation on the number of variables it can handle nor the depth of the
3417 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3418 information can now expand as required, and rather than having a single
3419 static array of bignums, BN_CTX now uses a linked-list of such arrays
3420 allowing it to expand on demand whilst maintaining the usefulness of
3421 BN_CTX's "bundling".
3424 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3425 to allow all RSA operations to function using a single BN_CTX.
3428 *) Preliminary support for certificate policy evaluation and checking. This
3429 is initially intended to pass the tests outlined in "Conformance Testing
3430 of Relying Party Client Certificate Path Processing Logic" v1.07.
3433 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3434 remained unused and not that useful. A variety of other little bignum
3435 tweaks and fixes have also been made continuing on from the audit (see
3439 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3440 associated ASN1, EVP and SSL functions and old ASN1 macros.
3443 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3444 and this should never fail. So the return value from the use of
3445 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3446 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3449 *) BN_CTX_get() should return zero-valued bignums, providing the same
3450 initialised value as BN_new().
3451 [Geoff Thorpe, suggested by Ulf Möller]
3453 *) Support for inhibitAnyPolicy certificate extension.
3456 *) An audit of the BIGNUM code is underway, for which debugging code is
3457 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3458 is considered valid when processing BIGNUMs, and causes execution to
3459 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3460 further steps are taken to deliberately pollute unused data in BIGNUM
3461 structures to try and expose faulty code further on. For now, openssl will
3462 (in its default mode of operation) continue to tolerate the inconsistent
3463 forms that it has tolerated in the past, but authors and packagers should
3464 consider trying openssl and their own applications when compiled with
3465 these debugging symbols defined. It will help highlight potential bugs in
3466 their own code, and will improve the test coverage for OpenSSL itself. At
3467 some point, these tighter rules will become openssl's default to improve
3468 maintainability, though the assert()s and other overheads will remain only
3469 in debugging configurations. See bn.h for more details.
3470 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3472 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3473 that can only be obtained through BN_CTX_new() (which implicitly
3474 initialises it). The presence of this function only made it possible
3475 to overwrite an existing structure (and cause memory leaks).
3478 *) Because of the callback-based approach for implementing LHASH as a
3479 template type, lh_insert() adds opaque objects to hash-tables and
3480 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3481 to clean up those corresponding objects before destroying the hash table
3482 (and losing the object pointers). So some over-zealous constifications in
3483 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3484 objects as "const" and the lh_doall[_arg] callback wrappers are not
3485 prototyped to have "const" restrictions on the object pointers they are
3486 given (and so aren't required to cast them away any more).
3489 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3490 (speed) prefers to use its own implementation. The two implementations
3491 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3492 its object type properly exposed (MS_TM) instead of casting to/from "char
3493 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3494 aren't necessarily the greatest nomenclatures - but this is what was used
3495 internally to the implementation so I've used that for now.
3498 *) Ensure that deprecated functions do not get compiled when
3499 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3500 the self-tests were still using deprecated key-generation functions so
3501 these have been updated also.
3504 *) Reorganise PKCS#7 code to separate the digest location functionality
3505 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
3506 New function PKCS7_set_digest() to set the digest type for PKCS#7
3507 digestedData type. Add additional code to correctly generate the
3508 digestedData type and add support for this type in PKCS7 initialization
3512 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3513 structure of type "other".
3516 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3517 sure the loop does correctly stop and breaking ("division by zero")
3518 modulus operations are not performed. The (pre-generated) prime
3519 table crypto/bn/bn_prime.h was already correct, but it could not be
3520 re-generated on some platforms because of the "division by zero"
3521 situation in the script.
3522 [Ralf S. Engelschall]
3524 *) Update support for ECC-based TLS ciphersuites according to
3525 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3526 SHA-1 now is only used for "small" curves (where the
3527 representation of a field element takes up to 24 bytes); for
3528 larger curves, the field element resulting from ECDH is directly
3529 used as premaster secret.
3530 [Douglas Stebila (Sun Microsystems Laboratories)]
3532 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3533 curve secp160r1 to the tests.
3534 [Douglas Stebila (Sun Microsystems Laboratories)]
3536 *) Add the possibility to load symbols globally with DSO.
3537 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3539 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3540 control of the error stack.
3543 *) Add support for STORE in ENGINE.
3546 *) Add the STORE type. The intention is to provide a common interface
3547 to certificate and key stores, be they simple file-based stores, or
3548 HSM-type store, or LDAP stores, or...
3549 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3552 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3553 pass a list of arguments to any function as well as provide a way
3554 for a function to pass data back to the caller.
3557 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3558 works like BUF_strdup() but can be used to duplicate a portion of
3559 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3563 *) Add the function sk_find_ex() which works like sk_find(), but will
3564 return an index to an element even if an exact match couldn't be
3565 found. The index is guaranteed to point at the element where the
3566 searched-for key would be inserted to preserve sorting order.
3569 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3570 takes an extra flags argument for optional functionality. Currently,
3571 the following flags are defined:
3573 OBJ_BSEARCH_VALUE_ON_NOMATCH
3574 This one gets OBJ_bsearch_ex() to return a pointer to the first
3575 element where the comparing function returns a negative or zero
3578 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3579 This one gets OBJ_bsearch_ex() to return a pointer to the first
3580 element where the comparing function returns zero. This is useful
3581 if there are more than one element where the comparing function
3585 *) Make it possible to create self-signed certificates with 'openssl ca'
3586 in such a way that the self-signed certificate becomes part of the
3587 CA database and uses the same mechanisms for serial number generation
3588 as all other certificate signing. The new flag '-selfsign' enables
3589 this functionality. Adapt CA.sh and CA.pl.in.
3592 *) Add functionality to check the public key of a certificate request
3593 against a given private. This is useful to check that a certificate
3594 request can be signed by that key (self-signing).
3597 *) Make it possible to have multiple active certificates with the same
3598 subject in the CA index file. This is done only if the keyword
3599 'unique_subject' is set to 'no' in the main CA section (default
3600 if 'CA_default') of the configuration file. The value is saved
3601 with the database itself in a separate index attribute file,
3602 named like the index file with '.attr' appended to the name.
3605 *) Generate muti valued AVAs using '+' notation in config files for
3609 *) Support for nameConstraints certificate extension.
3612 *) Support for policyConstraints certificate extension.
3615 *) Support for policyMappings certificate extension.
3618 *) Make sure the default DSA_METHOD implementation only uses its
3619 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3620 and change its own handlers to be NULL so as to remove unnecessary
3621 indirection. This lets alternative implementations fallback to the
3622 default implementation more easily.
3625 *) Support for directoryName in GeneralName related extensions
3629 *) Make it possible to link applications using Makefile.shared.
3630 Make that possible even when linking against static libraries!
3633 *) Support for single pass processing for S/MIME signing. This now
3634 means that S/MIME signing can be done from a pipe, in addition
3635 cleartext signing (multipart/signed type) is effectively streaming
3636 and the signed data does not need to be all held in memory.
3638 This is done with a new flag PKCS7_STREAM. When this flag is set
3639 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3640 is done after the data is output (and digests calculated) in
3641 SMIME_write_PKCS7().
3644 *) Add full support for -rpath/-R, both in shared libraries and
3645 applications, at least on the platforms where it's known how
3649 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3650 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3651 will now compute a table of multiples of the generator that
3652 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3653 faster (notably in the case of a single point multiplication,
3654 scalar * generator).
3655 [Nils Larsch, Bodo Moeller]
3657 *) IPv6 support for certificate extensions. The various extensions
3658 which use the IP:a.b.c.d can now take IPv6 addresses using the
3659 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3663 *) Added an ENGINE that implements RSA by performing private key
3664 exponentiations with the GMP library. The conversions to and from
3665 GMP's mpz_t format aren't optimised nor are any montgomery forms
3666 cached, and on x86 it appears OpenSSL's own performance has caught up.
3667 However there are likely to be other architectures where GMP could
3668 provide a boost. This ENGINE is not built in by default, but it can be
3669 specified at Configure time and should be accompanied by the necessary
3670 linker additions, eg;
3671 ./config -DOPENSSL_USE_GMP -lgmp
3674 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3675 testing availability of engines with "-t" - the old behaviour is
3676 produced by increasing the feature's verbosity with "-tt".
3679 *) ECDSA routines: under certain error conditions uninitialized BN objects
3680 could be freed. Solution: make sure initialization is performed early
3681 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3685 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3686 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3687 software implementations. For DSA and DH, parameter generation can
3688 also be overriden by providing the appropriate method callbacks.
3691 *) Change the "progress" mechanism used in key-generation and
3692 primality testing to functions that take a new BN_GENCB pointer in
3693 place of callback/argument pairs. The new API functions have "_ex"
3694 postfixes and the older functions are reimplemented as wrappers for
3695 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3696 declarations of the old functions to help (graceful) attempts to
3697 migrate to the new functions. Also, the new key-generation API
3698 functions operate on a caller-supplied key-structure and return
3699 success/failure rather than returning a key or NULL - this is to
3700 help make "keygen" another member function of RSA_METHOD etc.
3702 Example for using the new callback interface:
3704 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3708 BN_GENCB_set(&my_cb, my_callback, my_arg);
3710 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3711 /* For the meaning of a, b in calls to my_callback(), see the
3712 * documentation of the function that calls the callback.
3713 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3714 * my_callback should return 1 if it wants BN_is_prime_ex()
3715 * to continue, or 0 to stop.
3720 *) Change the ZLIB compression method to be stateful, and make it
3721 available to TLS with the number defined in
3722 draft-ietf-tls-compression-04.txt.
3725 *) Add the ASN.1 structures and functions for CertificatePair, which
3726 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3728 CertificatePair ::= SEQUENCE {
3729 forward [0] Certificate OPTIONAL,
3730 reverse [1] Certificate OPTIONAL,
3731 -- at least one of the pair shall be present -- }
3733 Also implement the PEM functions to read and write certificate
3734 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3736 This needed to be defined, mostly for the sake of the LDAP
3737 attribute crossCertificatePair, but may prove useful elsewhere as
3741 *) Make it possible to inhibit symlinking of shared libraries in
3742 Makefile.shared, for Cygwin's sake.
3745 *) Extend the BIGNUM API by creating a function
3746 void BN_set_negative(BIGNUM *a, int neg);
3747 and a macro that behave like
3748 int BN_is_negative(const BIGNUM *a);
3750 to avoid the need to access 'a->neg' directly in applications.
3753 *) Implement fast modular reduction for pseudo-Mersenne primes
3754 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3755 EC_GROUP_new_curve_GFp() will now automatically use this
3757 [Nils Larsch <nla@trustcenter.de>]
3759 *) Add new lock type (CRYPTO_LOCK_BN).
3762 *) Change the ENGINE framework to automatically load engines
3763 dynamically from specific directories unless they could be
3764 found to already be built in or loaded. Move all the
3765 current engines except for the cryptodev one to a new
3767 The engines in engines/ are built as shared libraries if
3768 the "shared" options was given to ./Configure or ./config.
3769 Otherwise, they are inserted in libcrypto.a.
3770 /usr/local/ssl/engines is the default directory for dynamic
3771 engines, but that can be overriden at configure time through
3772 the usual use of --prefix and/or --openssldir, and at run
3773 time with the environment variable OPENSSL_ENGINES.
3774 [Geoff Thorpe and Richard Levitte]
3776 *) Add Makefile.shared, a helper makefile to build shared
3777 libraries. Addapt Makefile.org.
3780 *) Add version info to Win32 DLLs.
3781 [Peter 'Luna' Runestig" <peter@runestig.com>]
3783 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3784 can be added using this API to created arbitrary PKCS#12
3785 files while avoiding the low level API.
3787 New options to PKCS12_create(), key or cert can be NULL and
3788 will then be omitted from the output file. The encryption
3789 algorithm NIDs can be set to -1 for no encryption, the mac
3790 iteration count can be set to 0 to omit the mac.
3792 Enhance pkcs12 utility by making the -nokeys and -nocerts
3793 options work when creating a PKCS#12 file. New option -nomac
3794 to omit the mac, NONE can be set for an encryption algorithm.
3795 New code is modified to use the enhanced PKCS12_create()
3796 instead of the low level API.
3799 *) Extend ASN1 encoder to support indefinite length constructed
3800 encoding. This can output sequences tags and octet strings in
3801 this form. Modify pk7_asn1.c to support indefinite length
3802 encoding. This is experimental and needs additional code to
3803 be useful, such as an ASN1 bio and some enhanced streaming
3806 Extend template encode functionality so that tagging is passed
3807 down to the template encoder.
3810 *) Let 'openssl req' fail if an argument to '-newkey' is not
3811 recognized instead of using RSA as a default.
3814 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3815 As these are not official, they are not included in "ALL";
3816 the "ECCdraft" ciphersuite group alias can be used to select them.
3817 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3819 *) Add ECDH engine support.
3820 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3822 *) Add ECDH in new directory crypto/ecdh/.
3823 [Douglas Stebila (Sun Microsystems Laboratories)]
3825 *) Let BN_rand_range() abort with an error after 100 iterations
3826 without success (which indicates a broken PRNG).
3829 *) Change BN_mod_sqrt() so that it verifies that the input value
3830 is really the square of the return value. (Previously,
3831 BN_mod_sqrt would show GIGO behaviour.)
3834 *) Add named elliptic curves over binary fields from X9.62, SECG,
3835 and WAP/WTLS; add OIDs that were still missing.
3837 [Sheueling Chang Shantz and Douglas Stebila
3838 (Sun Microsystems Laboratories)]
3840 *) Extend the EC library for elliptic curves over binary fields
3841 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3844 EC_GF2m_simple_method
3848 EC_GROUP_new_curve_GF2m
3849 EC_GROUP_set_curve_GF2m
3850 EC_GROUP_get_curve_GF2m
3851 EC_POINT_set_affine_coordinates_GF2m
3852 EC_POINT_get_affine_coordinates_GF2m
3853 EC_POINT_set_compressed_coordinates_GF2m
3855 Point compression for binary fields is disabled by default for
3856 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3859 As binary polynomials are represented as BIGNUMs, various members
3860 of the EC_GROUP and EC_POINT data structures can be shared
3861 between the implementations for prime fields and binary fields;
3862 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3863 are essentially identical to their ..._GFp counterparts.
3864 (For simplicity, the '..._GFp' prefix has been dropped from
3865 various internal method names.)
3867 An internal 'field_div' method (similar to 'field_mul' and
3868 'field_sqr') has been added; this is used only for binary fields.
3870 [Sheueling Chang Shantz and Douglas Stebila
3871 (Sun Microsystems Laboratories)]
3873 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3874 through methods ('mul', 'precompute_mult').
3876 The generic implementations (now internally called 'ec_wNAF_mul'
3877 and 'ec_wNAF_precomputed_mult') remain the default if these
3878 methods are undefined.
3880 [Sheueling Chang Shantz and Douglas Stebila
3881 (Sun Microsystems Laboratories)]
3883 *) New function EC_GROUP_get_degree, which is defined through
3884 EC_METHOD. For curves over prime fields, this returns the bit
3885 length of the modulus.
3887 [Sheueling Chang Shantz and Douglas Stebila
3888 (Sun Microsystems Laboratories)]
3890 *) New functions EC_GROUP_dup, EC_POINT_dup.
3891 (These simply call ..._new and ..._copy).
3893 [Sheueling Chang Shantz and Douglas Stebila
3894 (Sun Microsystems Laboratories)]
3896 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3897 Polynomials are represented as BIGNUMs (where the sign bit is not
3898 used) in the following functions [macros]:
3901 BN_GF2m_sub [= BN_GF2m_add]
3902 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3903 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3904 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3906 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3907 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3908 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3909 BN_GF2m_cmp [= BN_ucmp]
3911 (Note that only the 'mod' functions are actually for fields GF(2^m).
3912 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3914 For some functions, an the irreducible polynomial defining a
3915 field can be given as an 'unsigned int[]' with strictly
3916 decreasing elements giving the indices of those bits that are set;
3917 i.e., p[] represents the polynomial
3918 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3920 p[0] > p[1] > ... > p[k] = 0.
3921 This applies to the following functions:
3926 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3927 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3929 BN_GF2m_mod_sqrt_arr
3930 BN_GF2m_mod_solve_quad_arr
3934 Conversion can be performed by the following functions:
3939 bntest.c has additional tests for binary polynomial arithmetic.
3941 Two implementations for BN_GF2m_mod_div() are available.
3942 The default algorithm simply uses BN_GF2m_mod_inv() and
3943 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3944 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3945 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3947 [Sheueling Chang Shantz and Douglas Stebila
3948 (Sun Microsystems Laboratories)]
3950 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3951 functionality is disabled at compile-time.
3952 [Douglas Stebila <douglas.stebila@sun.com>]
3954 *) Change default behaviour of 'openssl asn1parse' so that more
3955 information is visible when viewing, e.g., a certificate:
3957 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3958 mode the content of non-printable OCTET STRINGs is output in a
3959 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3960 avoid the appearance of a printable string.
3961 [Nils Larsch <nla@trustcenter.de>]
3963 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3965 EC_GROUP_set_asn1_flag()
3966 EC_GROUP_get_asn1_flag()
3967 EC_GROUP_set_point_conversion_form()
3968 EC_GROUP_get_point_conversion_form()
3969 These control ASN1 encoding details:
3970 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3971 has been set to OPENSSL_EC_NAMED_CURVE.
3972 - Points are encoded in uncompressed form by default; options for
3973 asn1_for are as for point2oct, namely
3974 POINT_CONVERSION_COMPRESSED
3975 POINT_CONVERSION_UNCOMPRESSED
3976 POINT_CONVERSION_HYBRID
3978 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3981 EC_GROUP_get0_seed()
3982 EC_GROUP_get_seed_len()
3983 This is used only for ASN1 purposes (so far).
3984 [Nils Larsch <nla@trustcenter.de>]
3986 *) Add 'field_type' member to EC_METHOD, which holds the NID
3987 of the appropriate field type OID. The new function
3988 EC_METHOD_get_field_type() returns this value.
3989 [Nils Larsch <nla@trustcenter.de>]
3994 EC_POINT_point2hex()
3995 EC_POINT_hex2point()
3996 providing useful interfaces to EC_POINT_point2oct() and
3997 EC_POINT_oct2point().
3998 [Nils Larsch <nla@trustcenter.de>]
4000 *) Change internals of the EC library so that the functions
4001 EC_GROUP_set_generator()
4002 EC_GROUP_get_generator()
4003 EC_GROUP_get_order()
4004 EC_GROUP_get_cofactor()
4005 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4006 to methods, which would lead to unnecessary code duplication when
4007 adding different types of curves.
4008 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4010 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4011 arithmetic, and such that modified wNAFs are generated
4012 (which avoid length expansion in many cases).
4015 *) Add a function EC_GROUP_check_discriminant() (defined via
4016 EC_METHOD) that verifies that the curve discriminant is non-zero.
4018 Add a function EC_GROUP_check() that makes some sanity tests
4019 on a EC_GROUP, its generator and order. This includes
4020 EC_GROUP_check_discriminant().
4021 [Nils Larsch <nla@trustcenter.de>]
4023 *) Add ECDSA in new directory crypto/ecdsa/.
4025 Add applications 'openssl ecparam' and 'openssl ecdsa'
4026 (these are based on 'openssl dsaparam' and 'openssl dsa').
4028 ECDSA support is also included in various other files across the
4029 library. Most notably,
4030 - 'openssl req' now has a '-newkey ecdsa:file' option;
4031 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4032 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4033 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4034 them suitable for ECDSA where domain parameters must be
4035 extracted before the specific public key;
4036 - ECDSA engine support has been added.
4037 [Nils Larsch <nla@trustcenter.de>]
4039 *) Include some named elliptic curves, and add OIDs from X9.62,
4040 SECG, and WAP/WTLS. Each curve can be obtained from the new
4042 EC_GROUP_new_by_curve_name(),
4043 and the list of available named curves can be obtained with
4044 EC_get_builtin_curves().
4045 Also add a 'curve_name' member to EC_GROUP objects, which can be
4047 EC_GROUP_set_curve_name()
4048 EC_GROUP_get_curve_name()
4049 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4051 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4052 was actually never needed) and in BN_mul(). The removal in BN_mul()
4053 required a small change in bn_mul_part_recursive() and the addition
4054 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4055 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4056 bn_sub_words() and bn_add_words() except they take arrays with
4060 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4062 *) Cleanse PEM buffers before freeing them since they may contain
4064 [Benjamin Bennett <ben@psc.edu>]
4066 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4067 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4068 authentication-only ciphersuites.
4071 *) Since AES128 and AES256 share a single mask bit in the logic of
4072 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4073 kludge to work properly if AES128 is available and AES256 isn't.
4076 *) Expand security boundary to match 1.1.1 module.
4079 *) Remove redundant features: hash file source, editing of test vectors
4080 modify fipsld to use external fips_premain.c signature.
4083 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4084 run algorithm test programs.
4087 *) Make algorithm test programs more tolerant of whitespace.
4090 *) Have SSL/TLS server implementation tolerate "mismatched" record
4091 protocol version while receiving ClientHello even if the
4092 ClientHello is fragmented. (The server can't insist on the
4093 particular protocol version it has chosen before the ServerHello
4094 message has informed the client about his choice.)
4097 *) Load error codes if they are not already present instead of using a
4098 static variable. This allows them to be cleanly unloaded and reloaded.
4101 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4103 *) Introduce limits to prevent malicious keys being able to
4104 cause a denial of service. (CVE-2006-2940)
4105 [Steve Henson, Bodo Moeller]
4107 *) Fix ASN.1 parsing of certain invalid structures that can result
4108 in a denial of service. (CVE-2006-2937) [Steve Henson]
4110 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4111 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4113 *) Fix SSL client code which could crash if connecting to a
4114 malicious SSLv2 server. (CVE-2006-4343)
4115 [Tavis Ormandy and Will Drewry, Google Security Team]
4117 *) Change ciphersuite string processing so that an explicit
4118 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4119 will no longer include "AES128-SHA"), and any other similar
4120 ciphersuite (same bitmap) from *other* protocol versions (so that
4121 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4122 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4123 changes from 0.9.8b and 0.9.8d.
4126 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4128 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4129 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4131 *) Change the Unix randomness entropy gathering to use poll() when
4132 possible instead of select(), since the latter has some
4133 undesirable limitations.
4134 [Darryl Miles via Richard Levitte and Bodo Moeller]
4136 *) Disable rogue ciphersuites:
4138 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4139 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4140 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4142 The latter two were purportedly from
4143 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4146 Also deactive the remaining ciphersuites from
4147 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4148 unofficial, and the ID has long expired.
4151 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4152 dual-core machines) and other potential thread-safety issues.
4155 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4157 *) Adapt fipsld and the build system to link against the validated FIPS
4158 module in FIPS mode.
4161 *) Fixes for VC++ 2005 build under Windows.
4164 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4165 from a Windows bash shell such as MSYS. It is autodetected from the
4166 "config" script when run from a VC++ environment. Modify standard VC++
4167 build to use fipscanister.o from the GNU make build.
4170 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4172 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4173 The value now differs depending on if you build for FIPS or not.
4174 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4175 safely run with a non-FIPSed libcrypto, as it may crash because of
4176 the difference induced by this change.
4179 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4181 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4182 (part of SSL_OP_ALL). This option used to disable the
4183 countermeasure against man-in-the-middle protocol-version
4184 rollback in the SSL 2.0 server implementation, which is a bad
4185 idea. (CVE-2005-2969)
4187 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4188 for Information Security, National Institute of Advanced Industrial
4189 Science and Technology [AIST], Japan)]
4191 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4192 mainly for FIPS compliance and not fully integrated at this stage.
4195 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4196 the exponentiation using a fixed-length exponent. (Otherwise,
4197 the information leaked through timing could expose the secret key
4198 after many signatures; cf. Bleichenbacher's attack on DSA with
4202 *) Make a new fixed-window mod_exp implementation the default for
4203 RSA, DSA, and DH private-key operations so that the sequence of
4204 squares and multiplies and the memory access pattern are
4205 independent of the particular secret key. This will mitigate
4206 cache-timing and potential related attacks.
4208 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4209 and this is automatically used by BN_mod_exp_mont() if the new flag
4210 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4211 will use this BN flag for private exponents unless the flag
4212 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4213 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4215 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4217 *) Change the client implementation for SSLv23_method() and
4218 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4219 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4220 (Previously, the SSL 2.0 backwards compatible Client Hello
4221 message format would be used even with SSL_OP_NO_SSLv2.)
4224 *) Add support for smime-type MIME parameter in S/MIME messages which some
4228 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4229 a threadsafe manner. Modify rsa code to use new function and add calls
4230 to dsa and dh code (which had race conditions before).
4233 *) Include the fixed error library code in the C error file definitions
4234 instead of fixing them up at runtime. This keeps the error code
4235 structures constant.
4238 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4240 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4243 *) Fixes for newer kerberos headers. NB: the casts are needed because
4244 the 'length' field is signed on one version and unsigned on another
4245 with no (?) obvious way to tell the difference, without these VC++
4246 complains. Also the "definition" of FAR (blank) is no longer included
4247 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4248 some needed definitions.
4251 *) Undo Cygwin change.
4254 *) Added support for proxy certificates according to RFC 3820.
4255 Because they may be a security thread to unaware applications,
4256 they must be explicitely allowed in run-time. See
4257 docs/HOWTO/proxy_certificates.txt for further information.
4260 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4262 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4263 server and client random values. Previously
4264 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4265 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4267 This change has negligible security impact because:
4269 1. Server and client random values still have 24 bytes of pseudo random
4272 2. Server and client random values are sent in the clear in the initial
4275 3. The master secret is derived using the premaster secret (48 bytes in
4276 size for static RSA ciphersuites) as well as client server and random
4279 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4282 [Stephen Henson, reported by UK NISCC]
4284 *) Use Windows randomness collection on Cygwin.
4287 *) Fix hang in EGD/PRNGD query when communication socket is closed
4288 prematurely by EGD/PRNGD.
4289 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4291 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4294 *) Back-port of selected performance improvements from development
4295 branch, as well as improved support for PowerPC platforms.
4298 *) Add lots of checks for memory allocation failure, error codes to indicate
4299 failure and freeing up memory if a failure occurs.
4300 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4302 *) Add new -passin argument to dgst.
4305 *) Perform some character comparisons of different types in X509_NAME_cmp:
4306 this is needed for some certificates that reencode DNs into UTF8Strings
4307 (in violation of RFC3280) and can't or wont issue name rollover
4311 *) Make an explicit check during certificate validation to see that
4312 the CA setting in each certificate on the chain is correct. As a
4313 side effect always do the following basic checks on extensions,
4314 not just when there's an associated purpose to the check:
4316 - if there is an unhandled critical extension (unless the user
4317 has chosen to ignore this fault)
4318 - if the path length has been exceeded (if one is set at all)
4319 - that certain extensions fit the associated purpose (if one has
4323 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4325 *) Avoid a race condition when CRLs are checked in a multi threaded
4326 environment. This would happen due to the reordering of the revoked
4327 entries during signature checking and serial number lookup. Now the
4328 encoding is cached and the serial number sort performed under a lock.
4329 Add new STACK function sk_is_sorted().
4332 *) Add Delta CRL to the extension code.
4335 *) Various fixes to s3_pkt.c so alerts are sent properly.
4336 [David Holmes <d.holmes@f5.com>]
4338 *) Reduce the chances of duplicate issuer name and serial numbers (in
4339 violation of RFC3280) using the OpenSSL certificate creation utilities.
4340 This is done by creating a random 64 bit value for the initial serial
4341 number when a serial number file is created or when a self signed
4342 certificate is created using 'openssl req -x509'. The initial serial
4343 number file is created using 'openssl x509 -next_serial' in CA.pl
4344 rather than being initialized to 1.
4347 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4349 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4350 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4351 [Joe Orton, Steve Henson]
4353 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4355 [Joe Orton, Steve Henson]
4357 *) Make it possible to have multiple active certificates with the same
4358 subject in the CA index file. This is done only if the keyword
4359 'unique_subject' is set to 'no' in the main CA section (default
4360 if 'CA_default') of the configuration file. The value is saved
4361 with the database itself in a separate index attribute file,
4362 named like the index file with '.attr' appended to the name.
4365 *) X509 verify fixes. Disable broken certificate workarounds when
4366 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4367 keyUsage extension present. Don't accept CRLs with unhandled critical
4368 extensions: since verify currently doesn't process CRL extensions this
4369 rejects a CRL with *any* critical extensions. Add new verify error codes
4373 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4374 A clarification of RFC2560 will require the use of OCTET STRINGs and
4375 some implementations cannot handle the current raw format. Since OpenSSL
4376 copies and compares OCSP nonces as opaque blobs without any attempt at
4377 parsing them this should not create any compatibility issues.
4380 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4381 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4382 this HMAC (and other) operations are several times slower than OpenSSL
4386 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4387 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4389 *) Use the correct content when signing type "other".
4392 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4394 *) Fix various bugs revealed by running the NISCC test suite:
4396 Stop out of bounds reads in the ASN1 code when presented with
4397 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4399 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4401 If verify callback ignores invalid public key errors don't try to check
4402 certificate signature with the NULL public key.
4406 *) New -ignore_err option in ocsp application to stop the server
4407 exiting on the first error in a request.
4410 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4411 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4415 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4416 extra data after the compression methods not only for TLS 1.0
4417 but also for SSL 3.0 (as required by the specification).
4418 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4420 *) Change X509_certificate_type() to mark the key as exported/exportable
4421 when it's 512 *bits* long, not 512 bytes.
4424 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4425 blocks during encryption.
4428 *) Various fixes to base64 BIO and non blocking I/O. On write
4429 flushes were not handled properly if the BIO retried. On read
4430 data was not being buffered properly and had various logic bugs.
4431 This also affects blocking I/O when the data being decoded is a
4435 *) Various S/MIME bugfixes and compatibility changes:
4436 output correct application/pkcs7 MIME type if
4437 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4438 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4439 of files as .eml work). Correctly handle very long lines in MIME
4443 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4445 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4446 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4447 a protocol version number mismatch like a decryption error
4448 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4451 *) Turn on RSA blinding by default in the default implementation
4452 to avoid a timing attack. Applications that don't want it can call
4453 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4454 They would be ill-advised to do so in most cases.
4455 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4457 *) Change RSA blinding code so that it works when the PRNG is not
4458 seeded (in this case, the secret RSA exponent is abused as
4459 an unpredictable seed -- if it is not unpredictable, there
4460 is no point in blinding anyway). Make RSA blinding thread-safe
4461 by remembering the creator's thread ID in rsa->blinding and
4462 having all other threads use local one-time blinding factors
4463 (this requires more computation than sharing rsa->blinding, but
4464 avoids excessive locking; and if an RSA object is not shared
4465 between threads, blinding will still be very fast).
4468 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4469 ENGINE as defaults for all supported algorithms irrespective of
4470 the 'flags' parameter. 'flags' is now honoured, so applications
4471 should make sure they are passing it correctly.
4474 *) Target "mingw" now allows native Windows code to be generated in
4475 the Cygwin environment as well as with the MinGW compiler.
4478 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4480 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4481 via timing by performing a MAC computation even if incorrrect
4482 block cipher padding has been found. This is a countermeasure
4483 against active attacks where the attacker has to distinguish
4484 between bad padding and a MAC verification error. (CVE-2003-0078)
4486 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4487 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4488 Martin Vuagnoux (EPFL, Ilion)]
4490 *) Make the no-err option work as intended. The intention with no-err
4491 is not to have the whole error stack handling routines removed from
4492 libcrypto, it's only intended to remove all the function name and
4493 reason texts, thereby removing some of the footprint that may not
4494 be interesting if those errors aren't displayed anyway.
4496 NOTE: it's still possible for any application or module to have it's
4497 own set of error texts inserted. The routines are there, just not
4498 used by default when no-err is given.
4501 *) Add support for FreeBSD on IA64.
4502 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4504 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4505 Kerberos function mit_des_cbc_cksum(). Before this change,
4506 the value returned by DES_cbc_cksum() was like the one from
4507 mit_des_cbc_cksum(), except the bytes were swapped.
4508 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4510 *) Allow an application to disable the automatic SSL chain building.
4511 Before this a rather primitive chain build was always performed in
4512 ssl3_output_cert_chain(): an application had no way to send the
4513 correct chain if the automatic operation produced an incorrect result.
4515 Now the chain builder is disabled if either:
4517 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4519 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4521 The reasoning behind this is that an application would not want the
4522 auto chain building to take place if extra chain certificates are
4523 present and it might also want a means of sending no additional
4524 certificates (for example the chain has two certificates and the
4528 *) Add the possibility to build without the ENGINE framework.
4529 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4531 *) Under Win32 gmtime() can return NULL: check return value in
4532 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4535 *) DSA routines: under certain error conditions uninitialized BN objects
4536 could be freed. Solution: make sure initialization is performed early
4537 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4538 Nils Larsch <nla@trustcenter.de> via PR#459)
4541 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4542 checked on reconnect on the client side, therefore session resumption
4543 could still fail with a "ssl session id is different" error. This
4544 behaviour is masked when SSL_OP_ALL is used due to
4545 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4546 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4547 followup to PR #377.
4550 *) IA-32 assembler support enhancements: unified ELF targets, support
4551 for SCO/Caldera platforms, fix for Cygwin shared build.
4554 *) Add support for FreeBSD on sparc64. As a consequence, support for
4555 FreeBSD on non-x86 processors is separate from x86 processors on
4556 the config script, much like the NetBSD support.
4557 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4559 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4561 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4564 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4565 code (06) was taken as the first octet of the session ID and the last
4566 octet was ignored consequently. As a result SSLv2 client side session
4567 caching could not have worked due to the session ID mismatch between
4569 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4573 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4574 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4578 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4579 seems that in spite of existing for more than a year, many application
4580 author have done nothing to provide the necessary callbacks, which
4581 means that this particular engine will not work properly anywhere.
4582 This is a very unfortunate situation which forces us, in the name
4583 of usability, to give the hw_ncipher.c a static lock, which is part
4585 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4586 appear in 0.9.8 or later. We EXPECT application authors to have
4587 dealt properly with this when 0.9.8 is released (unless we actually
4588 make such changes in the libcrypto locking code that changes will
4589 have to be made anyway).
4592 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4593 octets have been read, EOF or an error occurs. Without this change
4594 some truncated ASN1 structures will not produce an error.
4597 *) Disable Heimdal support, since it hasn't been fully implemented.
4598 Still give the possibility to force the use of Heimdal, but with
4599 warnings and a request that patches get sent to openssl-dev.
4602 *) Add the VC-CE target, introduce the WINCE sysname, and add
4603 INSTALL.WCE and appropriate conditionals to make it build.
4604 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4606 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4607 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4608 edit numbers of the version.
4609 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4611 *) Introduce safe string copy and catenation functions
4612 (BUF_strlcpy() and BUF_strlcat()).
4613 [Ben Laurie (CHATS) and Richard Levitte]
4615 *) Avoid using fixed-size buffers for one-line DNs.
4616 [Ben Laurie (CHATS)]
4618 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4619 resizing buffers containing secrets, and use where appropriate.
4620 [Ben Laurie (CHATS)]
4622 *) Avoid using fixed size buffers for configuration file location.
4623 [Ben Laurie (CHATS)]
4625 *) Avoid filename truncation for various CA files.
4626 [Ben Laurie (CHATS)]
4628 *) Use sizeof in preference to magic numbers.
4629 [Ben Laurie (CHATS)]
4631 *) Avoid filename truncation in cert requests.
4632 [Ben Laurie (CHATS)]
4634 *) Add assertions to check for (supposedly impossible) buffer
4636 [Ben Laurie (CHATS)]
4638 *) Don't cache truncated DNS entries in the local cache (this could
4639 potentially lead to a spoofing attack).
4640 [Ben Laurie (CHATS)]
4642 *) Fix various buffers to be large enough for hex/decimal
4643 representations in a platform independent manner.
4644 [Ben Laurie (CHATS)]
4646 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4647 resizing buffers containing secrets, and use where appropriate.
4648 [Ben Laurie (CHATS)]
4650 *) Add BIO_indent() to avoid much slightly worrying code to do
4652 [Ben Laurie (CHATS)]
4654 *) Convert sprintf()/BIO_puts() to BIO_printf().
4655 [Ben Laurie (CHATS)]
4657 *) buffer_gets() could terminate with the buffer only half
4659 [Ben Laurie (CHATS)]
4661 *) Add assertions to prevent user-supplied crypto functions from
4662 overflowing internal buffers by having large block sizes, etc.
4663 [Ben Laurie (CHATS)]
4665 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4667 [Ben Laurie (CHATS)]
4669 *) Eliminate unused copy of key in RC4.
4670 [Ben Laurie (CHATS)]
4672 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4673 [Ben Laurie (CHATS)]
4675 *) Fix off-by-one error in EGD path.
4676 [Ben Laurie (CHATS)]
4678 *) If RANDFILE path is too long, ignore instead of truncating.
4679 [Ben Laurie (CHATS)]
4681 *) Eliminate unused and incorrectly sized X.509 structure
4683 [Ben Laurie (CHATS)]
4685 *) Eliminate unused and dangerous function knumber().
4686 [Ben Laurie (CHATS)]
4688 *) Eliminate unused and dangerous structure, KSSL_ERR.
4689 [Ben Laurie (CHATS)]
4691 *) Protect against overlong session ID context length in an encoded
4692 session object. Since these are local, this does not appear to be
4694 [Ben Laurie (CHATS)]
4696 *) Change from security patch (see 0.9.6e below) that did not affect
4697 the 0.9.6 release series:
4699 Remote buffer overflow in SSL3 protocol - an attacker could
4700 supply an oversized master key in Kerberos-enabled versions.
4702 [Ben Laurie (CHATS)]
4704 *) Change the SSL kerb5 codes to match RFC 2712.
4707 *) Make -nameopt work fully for req and add -reqopt switch.
4708 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4710 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4711 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4713 *) Make sure tests can be performed even if the corresponding algorithms
4714 have been removed entirely. This was also the last step to make
4715 OpenSSL compilable with DJGPP under all reasonable conditions.
4716 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4718 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4719 to allow version independent disabling of normally unselected ciphers,
4720 which may be activated as a side-effect of selecting a single cipher.
4722 (E.g., cipher list string "RSA" enables ciphersuites that are left
4723 out of "ALL" because they do not provide symmetric encryption.
4724 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4725 [Lutz Jaenicke, Bodo Moeller]
4727 *) Add appropriate support for separate platform-dependent build
4728 directories. The recommended way to make a platform-dependent
4729 build directory is the following (tested on Linux), maybe with
4732 # Place yourself outside of the OpenSSL source tree. In
4733 # this example, the environment variable OPENSSL_SOURCE
4734 # is assumed to contain the absolute OpenSSL source directory.
4735 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4736 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4737 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4738 mkdir -p `dirname $F`
4739 ln -s $OPENSSL_SOURCE/$F $F
4742 To be absolutely sure not to disturb the source tree, a "make clean"
4743 is a good thing. If it isn't successfull, don't worry about it,
4744 it probably means the source directory is very clean.
4747 *) Make sure any ENGINE control commands make local copies of string
4748 pointers passed to them whenever necessary. Otherwise it is possible
4749 the caller may have overwritten (or deallocated) the original string
4750 data when a later ENGINE operation tries to use the stored values.
4751 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4753 *) Improve diagnostics in file reading and command-line digests.
4754 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4756 *) Add AES modes CFB and OFB to the object database. Correct an
4757 error in AES-CFB decryption.
4760 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4761 allows existing EVP_CIPHER_CTX structures to be reused after
4762 calling EVP_*Final(). This behaviour is used by encryption
4763 BIOs and some applications. This has the side effect that
4764 applications must explicitly clean up cipher contexts with
4765 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4768 *) Check the values of dna and dnb in bn_mul_recursive before calling
4769 bn_mul_comba (a non zero value means the a or b arrays do not contain
4770 n2 elements) and fallback to bn_mul_normal if either is not zero.
4773 *) Fix escaping of non-ASCII characters when using the -subj option
4774 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4777 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4778 form for "surname", serialNumber has no short form.
4779 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4780 therefore remove "mail" short name for "internet 7".
4781 The OID for unique identifiers in X509 certificates is
4782 x500UniqueIdentifier, not uniqueIdentifier.
4783 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4786 *) Add an "init" command to the ENGINE config module and auto initialize
4787 ENGINEs. Without any "init" command the ENGINE will be initialized
4788 after all ctrl commands have been executed on it. If init=1 the
4789 ENGINE is initailized at that point (ctrls before that point are run
4790 on the uninitialized ENGINE and after on the initialized one). If
4791 init=0 then the ENGINE will not be iniatialized at all.
4794 *) Fix the 'app_verify_callback' interface so that the user-defined
4795 argument is actually passed to the callback: In the
4796 SSL_CTX_set_cert_verify_callback() prototype, the callback
4797 declaration has been changed from
4800 int (*cb)(X509_STORE_CTX *,void *);
4801 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4802 i=s->ctx->app_verify_callback(&ctx)
4803 has been changed into
4804 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4806 To update applications using SSL_CTX_set_cert_verify_callback(),
4807 a dummy argument can be added to their callback functions.
4808 [D. K. Smetters <smetters@parc.xerox.com>]
4810 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4811 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4813 *) Add and OPENSSL_LOAD_CONF define which will cause
4814 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4815 This allows older applications to transparently support certain
4816 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4817 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4818 load the config file and OPENSSL_add_all_algorithms_conf() which will
4819 always load it have also been added.
4822 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4823 Adjust NIDs and EVP layer.
4824 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4826 *) Config modules support in openssl utility.
4828 Most commands now load modules from the config file,
4829 though in a few (such as version) this isn't done
4830 because it couldn't be used for anything.
4832 In the case of ca and req the config file used is
4833 the same as the utility itself: that is the -config
4834 command line option can be used to specify an
4838 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4839 use "openssl_conf" if filename is NULL use default openssl config file.
4842 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4843 config section name. Add a new flag to tolerate a missing config file
4844 and move code to CONF_modules_load_file().
4847 *) Support for crypto accelerator cards from Accelerated Encryption
4848 Processing, www.aep.ie. (Use engine 'aep')
4849 The support was copied from 0.9.6c [engine] and adapted/corrected
4850 to work with the new engine framework.
4851 [AEP Inc. and Richard Levitte]
4853 *) Support for SureWare crypto accelerator cards from Baltimore
4854 Technologies. (Use engine 'sureware')
4855 The support was copied from 0.9.6c [engine] and adapted
4856 to work with the new engine framework.
4859 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4860 make the newer ENGINE framework commands for the CHIL engine work.
4861 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4863 *) Make it possible to produce shared libraries on ReliantUNIX.
4864 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4866 *) Add the configuration target debug-linux-ppro.
4867 Make 'openssl rsa' use the general key loading routines
4868 implemented in apps.c, and make those routines able to
4869 handle the key format FORMAT_NETSCAPE and the variant
4871 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4873 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4874 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4876 *) Add -keyform to rsautl, and document -engine.
4877 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4879 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4880 BIO_R_NO_SUCH_FILE error code rather than the generic
4881 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4884 *) Add new functions
4886 ERR_peek_last_error_line
4887 ERR_peek_last_error_line_data.
4888 These are similar to
4891 ERR_peek_error_line_data,
4892 but report on the latest error recorded rather than the first one
4893 still in the error queue.
4894 [Ben Laurie, Bodo Moeller]
4896 *) default_algorithms option in ENGINE config module. This allows things
4898 default_algorithms = ALL
4899 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4902 *) Prelminary ENGINE config module.
4905 *) New experimental application configuration code.
4908 *) Change the AES code to follow the same name structure as all other
4909 symmetric ciphers, and behave the same way. Move everything to
4910 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4911 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4913 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4914 [Ben Laurie and Theo de Raadt]
4916 *) Add option to output public keys in req command.
4917 [Massimiliano Pala madwolf@openca.org]
4919 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4920 (up to about 10% better than before for P-192 and P-224).
4923 *) New functions/macros
4925 SSL_CTX_set_msg_callback(ctx, cb)
4926 SSL_CTX_set_msg_callback_arg(ctx, arg)
4927 SSL_set_msg_callback(ssl, cb)
4928 SSL_set_msg_callback_arg(ssl, arg)
4930 to request calling a callback function
4932 void cb(int write_p, int version, int content_type,
4933 const void *buf, size_t len, SSL *ssl, void *arg)
4935 whenever a protocol message has been completely received
4936 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4937 protocol version according to which the SSL library interprets
4938 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4939 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4940 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4941 specification (change_cipher_spec(20), alert(21), handshake(22)).
4942 'buf' and 'len' point to the actual message, 'ssl' to the
4943 SSL object, and 'arg' is the application-defined value set by
4944 SSL[_CTX]_set_msg_callback_arg().
4946 'openssl s_client' and 'openssl s_server' have new '-msg' options
4947 to enable a callback that displays all protocol messages.
4950 *) Change the shared library support so shared libraries are built as
4951 soon as the corresponding static library is finished, and thereby get
4952 openssl and the test programs linked against the shared library.
4953 This still only happens when the keyword "shard" has been given to
4954 the configuration scripts.
4956 NOTE: shared library support is still an experimental thing, and
4957 backward binary compatibility is still not guaranteed.
4958 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4960 *) Add support for Subject Information Access extension.
4961 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4963 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4964 additional bytes when new memory had to be allocated, not just
4965 when reusing an existing buffer.
4968 *) New command line and configuration option 'utf8' for the req command.
4969 This allows field values to be specified as UTF8 strings.
4972 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4973 runs for the former and machine-readable output for the latter.
4976 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4977 of the e-mail address in the DN (i.e., it will go into a certificate
4978 extension only). The new configuration file option 'email_in_dn = no'
4979 has the same effect.
4980 [Massimiliano Pala madwolf@openca.org]
4982 *) Change all functions with names starting with des_ to be starting
4983 with DES_ instead. Add wrappers that are compatible with libdes,
4984 but are named _ossl_old_des_*. Finally, add macros that map the
4985 des_* symbols to the corresponding _ossl_old_des_* if libdes
4986 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4987 desired, the des_* symbols will be mapped to DES_*, with one
4990 Since we provide two compatibility mappings, the user needs to
4991 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4992 compatibility is desired. The default (i.e., when that macro
4993 isn't defined) is OpenSSL 0.9.6c compatibility.
4995 There are also macros that enable and disable the support of old
4996 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4997 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4998 are defined, the default will apply: to support the old des routines.
5000 In either case, one must include openssl/des.h to get the correct
5001 definitions. Do not try to just include openssl/des_old.h, that
5004 NOTE: This is a major break of an old API into a new one. Software
5005 authors are encouraged to switch to the DES_ style functions. Some
5006 time in the future, des_old.h and the libdes compatibility functions
5007 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5008 default), and then completely removed.
5011 *) Test for certificates which contain unsupported critical extensions.
5012 If such a certificate is found during a verify operation it is
5013 rejected by default: this behaviour can be overridden by either
5014 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5015 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5016 X509_supported_extension() has also been added which returns 1 if a
5017 particular extension is supported.
5020 *) Modify the behaviour of EVP cipher functions in similar way to digests
5021 to retain compatibility with existing code.
5024 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5025 compatibility with existing code. In particular the 'ctx' parameter does
5026 not have to be to be initialized before the call to EVP_DigestInit() and
5027 it is tidied up after a call to EVP_DigestFinal(). New function
5028 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5029 EVP_MD_CTX_copy() changed to not require the destination to be
5030 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5031 requires the destination to be valid.
5033 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5034 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5037 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5038 so that complete 'Handshake' protocol structures are kept in memory
5039 instead of overwriting 'msg_type' and 'length' with 'body' data.
5042 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5043 [Massimo Santin via Richard Levitte]
5045 *) Major restructuring to the underlying ENGINE code. This includes
5046 reduction of linker bloat, separation of pure "ENGINE" manipulation
5047 (initialisation, etc) from functionality dealing with implementations
5048 of specific crypto iterfaces. This change also introduces integrated
5049 support for symmetric ciphers and digest implementations - so ENGINEs
5050 can now accelerate these by providing EVP_CIPHER and EVP_MD
5051 implementations of their own. This is detailed in crypto/engine/README
5052 as it couldn't be adequately described here. However, there are a few
5053 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5054 were changed in the original introduction of ENGINE code have now
5055 reverted back - the hooking from this code to ENGINE is now a good
5056 deal more passive and at run-time, operations deal directly with
5057 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5058 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5059 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5060 they were not being used by the framework as there is no concept of a
5061 BIGNUM_METHOD and they could not be generalised to the new
5062 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5063 ENGINE_cpy() has been removed as it cannot be consistently defined in
5067 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5070 *) Change mkdef.pl to sort symbols that get the same entry number,
5071 and make sure the automatically generated functions ERR_load_*
5072 become part of libeay.num as well.
5075 *) New function SSL_renegotiate_pending(). This returns true once
5076 renegotiation has been requested (either SSL_renegotiate() call
5077 or HelloRequest/ClientHello receveived from the peer) and becomes
5078 false once a handshake has been completed.
5079 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5080 sends a HelloRequest, but does not ensure that a handshake takes
5081 place. SSL_renegotiate_pending() is useful for checking if the
5082 client has followed the request.)
5085 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5086 By default, clients may request session resumption even during
5087 renegotiation (if session ID contexts permit); with this option,
5088 session resumption is possible only in the first handshake.
5090 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5091 more bits available for options that should not be part of
5092 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5095 *) Add some demos for certificate and certificate request creation.
5098 *) Make maximum certificate chain size accepted from the peer application
5099 settable (SSL*_get/set_max_cert_list()), as proposed by
5100 "Douglas E. Engert" <deengert@anl.gov>.
5103 *) Add support for shared libraries for Unixware-7
5104 (Boyd Lynn Gerber <gerberb@zenez.com>).
5107 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5108 be done prior to destruction. Use this to unload error strings from
5109 ENGINEs that load their own error strings. NB: This adds two new API
5110 functions to "get" and "set" this destroy handler in an ENGINE.
5113 *) Alter all existing ENGINE implementations (except "openssl" and
5114 "openbsd") to dynamically instantiate their own error strings. This
5115 makes them more flexible to be built both as statically-linked ENGINEs
5116 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5117 Also, add stub code to each that makes building them as self-contained
5118 shared-libraries easier (see README.ENGINE).
5121 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5122 implementations into applications that are completely implemented in
5123 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5124 commands that can be used to configure what shared-library to load and
5125 to control aspects of the way it is handled. Also, made an update to
5126 the README.ENGINE file that brings its information up-to-date and
5127 provides some information and instructions on the "dynamic" ENGINE
5128 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5131 *) Make it possible to unload ranges of ERR strings with a new
5132 "ERR_unload_strings" function.
5135 *) Add a copy() function to EVP_MD.
5138 *) Make EVP_MD routines take a context pointer instead of just the
5139 md_data void pointer.
5142 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5143 that the digest can only process a single chunk of data
5144 (typically because it is provided by a piece of
5145 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5146 is only going to provide a single chunk of data, and hence the
5147 framework needn't accumulate the data for oneshot drivers.
5150 *) As with "ERR", make it possible to replace the underlying "ex_data"
5151 functions. This change also alters the storage and management of global
5152 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5153 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5154 index counters. The API functions that use this state have been changed
5155 to take a "class_index" rather than pointers to the class's local STACK
5156 and counter, and there is now an API function to dynamically create new
5157 classes. This centralisation allows us to (a) plug a lot of the
5158 thread-safety problems that existed, and (b) makes it possible to clean
5159 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5160 such data would previously have always leaked in application code and
5161 workarounds were in place to make the memory debugging turn a blind eye
5162 to it. Application code that doesn't use this new function will still
5163 leak as before, but their memory debugging output will announce it now
5164 rather than letting it slide.
5166 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5167 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5168 has a return value to indicate success or failure.
5171 *) Make it possible to replace the underlying "ERR" functions such that the
5172 global state (2 LHASH tables and 2 locks) is only used by the "default"
5173 implementation. This change also adds two functions to "get" and "set"
5174 the implementation prior to it being automatically set the first time
5175 any other ERR function takes place. Ie. an application can call "get",
5176 pass the return value to a module it has just loaded, and that module
5177 can call its own "set" function using that value. This means the
5178 module's "ERR" operations will use (and modify) the error state in the
5179 application and not in its own statically linked copy of OpenSSL code.
5182 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5183 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5184 the operation, and provides a more encapsulated way for external code
5185 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5186 to use these functions rather than manually incrementing the counts.
5188 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5191 *) Add EVP test program.
5194 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5197 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5198 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5199 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5200 These allow a CRL to be built without having to access X509_CRL fields
5201 directly. Modify 'ca' application to use new functions.
5204 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5205 bug workarounds. Rollback attack detection is a security feature.
5206 The problem will only arise on OpenSSL servers when TLSv1 is not
5207 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5208 Software authors not wanting to support TLSv1 will have special reasons
5209 for their choice and can explicitly enable this option.
5210 [Bodo Moeller, Lutz Jaenicke]
5212 *) Rationalise EVP so it can be extended: don't include a union of
5213 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5214 (similar to those existing for EVP_CIPHER_CTX).
5219 EVP_MD_CTX_init(&md); /* new function call */
5220 EVP_DigestInit(&md, EVP_sha1());
5221 EVP_DigestUpdate(&md, in, len);
5222 EVP_DigestFinal(&md, out, NULL);
5223 EVP_MD_CTX_cleanup(&md); /* new function call */
5227 *) Make DES key schedule conform to the usual scheme, as well as
5228 correcting its structure. This means that calls to DES functions
5229 now have to pass a pointer to a des_key_schedule instead of a
5230 plain des_key_schedule (which was actually always a pointer
5233 des_key_schedule ks;
5235 des_set_key_checked(..., &ks);
5236 des_ncbc_encrypt(..., &ks, ...);
5238 (Note that a later change renames 'des_...' into 'DES_...'.)
5241 *) Initial reduction of linker bloat: the use of some functions, such as
5242 PEM causes large amounts of unused functions to be linked in due to
5243 poor organisation. For example pem_all.c contains every PEM function
5244 which has a knock on effect of linking in large amounts of (unused)
5245 ASN1 code. Grouping together similar functions and splitting unrelated
5246 functions prevents this.
5249 *) Cleanup of EVP macros.
5252 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5253 correct _ecb suffix.
5256 *) Add initial OCSP responder support to ocsp application. The
5257 revocation information is handled using the text based index
5258 use by the ca application. The responder can either handle
5259 requests generated internally, supplied in files (for example
5260 via a CGI script) or using an internal minimal server.
5263 *) Add configuration choices to get zlib compression for TLS.
5266 *) Changes to Kerberos SSL for RFC 2712 compliance:
5267 1. Implemented real KerberosWrapper, instead of just using
5268 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5269 2. Implemented optional authenticator field of KerberosWrapper.
5271 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5272 and authenticator structs; see crypto/krb5/.
5274 Generalized Kerberos calls to support multiple Kerberos libraries.
5275 [Vern Staats <staatsvr@asc.hpc.mil>,
5276 Jeffrey Altman <jaltman@columbia.edu>
5277 via Richard Levitte]
5279 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5280 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5281 values for each of the key sizes rather than having just
5282 parameters (and 'speed' generating keys each time).
5285 *) Speed up EVP routines.
5288 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5289 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5290 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5291 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5293 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5294 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5295 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5298 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5300 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5303 *) Added the OS2-EMX target.
5304 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5306 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5307 to support NCONF routines in extension code. New function CONF_set_nconf()
5308 to allow functions which take an NCONF to also handle the old LHASH
5309 structure: this means that the old CONF compatible routines can be
5310 retained (in particular wrt extensions) without having to duplicate the
5311 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5314 *) Enhance the general user interface with mechanisms for inner control
5315 and with possibilities to have yes/no kind of prompts.
5318 *) Change all calls to low level digest routines in the library and
5319 applications to use EVP. Add missing calls to HMAC_cleanup() and
5320 don't assume HMAC_CTX can be copied using memcpy().
5321 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5323 *) Add the possibility to control engines through control names but with
5324 arbitrary arguments instead of just a string.
5325 Change the key loaders to take a UI_METHOD instead of a callback
5326 function pointer. NOTE: this breaks binary compatibility with earlier
5327 versions of OpenSSL [engine].
5328 Adapt the nCipher code for these new conditions and add a card insertion
5332 *) Enhance the general user interface with mechanisms to better support
5333 dialog box interfaces, application-defined prompts, the possibility
5334 to use defaults (for example default passwords from somewhere else)
5335 and interrupts/cancellations.
5338 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5339 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5342 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5343 tidy up some unnecessarily weird code in 'sk_new()').
5344 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5346 *) Change the key loading routines for ENGINEs to use the same kind
5347 callback (pem_password_cb) as all other routines that need this
5351 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5352 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5353 than this minimum value is recommended.
5356 *) New random seeder for OpenVMS, using the system process statistics
5357 that are easily reachable.
5360 *) Windows apparently can't transparently handle global
5361 variables defined in DLLs. Initialisations such as:
5363 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5365 wont compile. This is used by the any applications that need to
5366 declare their own ASN1 modules. This was fixed by adding the option
5367 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5368 needed for static libraries under Win32.
5371 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5372 setting of purpose and trust fields. New X509_STORE trust and
5373 purpose functions and tidy up setting in other SSL functions.
5376 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5377 structure. These are inherited by X509_STORE_CTX when it is
5378 initialised. This allows various defaults to be set in the
5379 X509_STORE structure (such as flags for CRL checking and custom
5380 purpose or trust settings) for functions which only use X509_STORE_CTX
5381 internally such as S/MIME.
5383 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5384 trust settings if they are not set in X509_STORE. This allows X509_STORE
5385 purposes and trust (in S/MIME for example) to override any set by default.
5387 Add command line options for CRL checking to smime, s_client and s_server
5391 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5392 are set then the CRL is looked up in the X509_STORE structure and
5393 its validity and signature checked, then if the certificate is found
5394 in the CRL the verify fails with a revoked error.
5396 Various new CRL related callbacks added to X509_STORE_CTX structure.
5398 Command line options added to 'verify' application to support this.
5400 This needs some additional work, such as being able to handle multiple
5401 CRLs with different times, extension based lookup (rather than just
5402 by subject name) and ultimately more complete V2 CRL extension
5406 *) Add a general user interface API (crypto/ui/). This is designed
5407 to replace things like des_read_password and friends (backward
5408 compatibility functions using this new API are provided).
5409 The purpose is to remove prompting functions from the DES code
5410 section as well as provide for prompting through dialog boxes in
5411 a window system and the like.
5414 *) Add "ex_data" support to ENGINE so implementations can add state at a
5415 per-structure level rather than having to store it globally.
5418 *) Make it possible for ENGINE structures to be copied when retrieved by
5419 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5420 This causes the "original" ENGINE structure to act like a template,
5421 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5422 operational state can be localised to each ENGINE structure, despite the
5423 fact they all share the same "methods". New ENGINE structures returned in
5424 this case have no functional references and the return value is the single
5425 structural reference. This matches the single structural reference returned
5426 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5430 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5431 needs to match any other type at all we need to manually clear the
5435 *) Changes to the "openssl engine" utility to include;
5436 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5437 about an ENGINE's available control commands.
5438 - executing control commands from command line arguments using the
5439 '-pre' and '-post' switches. '-post' is only used if '-t' is
5440 specified and the ENGINE is successfully initialised. The syntax for
5441 the individual commands are colon-separated, for example;
5442 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5445 *) New dynamic control command support for ENGINEs. ENGINEs can now
5446 declare their own commands (numbers), names (strings), descriptions,
5447 and input types for run-time discovery by calling applications. A
5448 subset of these commands are implicitly classed as "executable"
5449 depending on their input type, and only these can be invoked through
5450 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5451 can be based on user input, config files, etc). The distinction is
5452 that "executable" commands cannot return anything other than a boolean
5453 result and can only support numeric or string input, whereas some
5454 discoverable commands may only be for direct use through
5455 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5456 pointers, or other custom uses. The "executable" commands are to
5457 support parameterisations of ENGINE behaviour that can be
5458 unambiguously defined by ENGINEs and used consistently across any
5459 OpenSSL-based application. Commands have been added to all the
5460 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5461 control over shared-library paths without source code alterations.
5464 *) Changed all ENGINE implementations to dynamically allocate their
5465 ENGINEs rather than declaring them statically. Apart from this being
5466 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5467 this also allows the implementations to compile without using the
5468 internal engine_int.h header.
5471 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5472 'const' value. Any code that should be able to modify a RAND_METHOD
5473 should already have non-const pointers to it (ie. they should only
5474 modify their own ones).
5477 *) Made a variety of little tweaks to the ENGINE code.
5478 - "atalla" and "ubsec" string definitions were moved from header files
5479 to C code. "nuron" string definitions were placed in variables
5480 rather than hard-coded - allowing parameterisation of these values
5481 later on via ctrl() commands.
5482 - Removed unused "#if 0"'d code.
5483 - Fixed engine list iteration code so it uses ENGINE_free() to release
5484 structural references.
5485 - Constified the RAND_METHOD element of ENGINE structures.
5486 - Constified various get/set functions as appropriate and added
5487 missing functions (including a catch-all ENGINE_cpy that duplicates
5488 all ENGINE values onto a new ENGINE except reference counts/state).
5489 - Removed NULL parameter checks in get/set functions. Setting a method
5490 or function to NULL is a way of cancelling out a previously set
5491 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5492 and doesn't justify the extra error symbols and code.
5493 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5494 flags from engine_int.h to engine.h.
5495 - Changed prototypes for ENGINE handler functions (init(), finish(),
5496 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5499 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5500 to the algorithm using long division. The binary algorithm can be
5501 used only if the modulus is odd. On 32-bit systems, it is faster
5502 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5503 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5504 up to 450 bits. In 64-bit environments, the binary algorithm
5505 appears to be advantageous for much longer moduli; here we use it
5506 for moduli up to 2048 bits.
5509 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5510 could not support the combine flag in choice fields.
5513 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5514 extensions from a certificate request to the certificate.
5517 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5518 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5519 file: this allows the display of the certificate about to be
5520 signed to be customised, to allow certain fields to be included
5521 or excluded and extension details. The old system didn't display
5522 multicharacter strings properly, omitted fields not in the policy
5523 and couldn't display additional details such as extensions.
5526 *) Function EC_POINTs_mul for multiple scalar multiplication
5527 of an arbitrary number of elliptic curve points
5528 \sum scalars[i]*points[i],
5529 optionally including the generator defined for the EC_GROUP:
5530 scalar*generator + \sum scalars[i]*points[i].
5532 EC_POINT_mul is a simple wrapper function for the typical case
5533 that the point list has just one item (besides the optional
5537 *) First EC_METHODs for curves over GF(p):
5539 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5540 operations and provides various method functions that can also
5541 operate with faster implementations of modular arithmetic.
5543 EC_GFp_mont_method() reuses most functions that are part of
5544 EC_GFp_simple_method, but uses Montgomery arithmetic.
5546 [Bodo Moeller; point addition and point doubling
5547 implementation directly derived from source code provided by
5548 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5550 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5551 crypto/ec/ec_lib.c):
5553 Curves are EC_GROUP objects (with an optional group generator)
5554 based on EC_METHODs that are built into the library.
5556 Points are EC_POINT objects based on EC_GROUP objects.
5558 Most of the framework would be able to handle curves over arbitrary
5559 finite fields, but as there are no obvious types for fields other
5560 than GF(p), some functions are limited to that for now.
5563 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5564 that the file contains a complete HTTP response.
5567 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5568 change the def and num file printf format specifier from "%-40sXXX"
5569 to "%-39s XXX". The latter will always guarantee a space after the
5570 field while the former will cause them to run together if the field
5571 is 40 of more characters long.
5574 *) Constify the cipher and digest 'method' functions and structures
5575 and modify related functions to take constant EVP_MD and EVP_CIPHER
5579 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5580 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5583 *) Modify EVP_Digest*() routines so they now return values. Although the
5584 internal software routines can never fail additional hardware versions
5588 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5590 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5591 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5594 ERR_R_NESTED_ASN1_ERROR
5596 ERR_R_MISSING_ASN1_EOS
5597 were 4 .. 9, conflicting with
5598 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5600 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5601 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5603 Add new error code 'ERR_R_INTERNAL_ERROR'.
5606 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5610 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5611 sets the subject name for a new request or supersedes the
5612 subject name in a given request. Formats that can be parsed are
5613 'CN=Some Name, OU=myOU, C=IT'
5615 'CN=Some Name/OU=myOU/C=IT'.
5617 Add options '-batch' and '-verbose' to 'openssl req'.
5618 [Massimiliano Pala <madwolf@hackmasters.net>]
5620 *) Introduce the possibility to access global variables through
5621 functions on platform were that's the best way to handle exporting
5622 global variables in shared libraries. To enable this functionality,
5623 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5624 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5625 is normally done by Configure or something similar).
5627 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5628 in the source file (foo.c) like this:
5630 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5631 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5633 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5634 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5636 OPENSSL_DECLARE_GLOBAL(int,foo);
5637 #define foo OPENSSL_GLOBAL_REF(foo)
5638 OPENSSL_DECLARE_GLOBAL(double,bar);
5639 #define bar OPENSSL_GLOBAL_REF(bar)
5641 The #defines are very important, and therefore so is including the
5642 header file everywhere where the defined globals are used.
5644 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5645 of ASN.1 items, but that structure is a bit different.
5647 The largest change is in util/mkdef.pl which has been enhanced with
5648 better and easier to understand logic to choose which symbols should
5649 go into the Windows .def files as well as a number of fixes and code
5650 cleanup (among others, algorithm keywords are now sorted
5651 lexicographically to avoid constant rewrites).
5654 *) In BN_div() keep a copy of the sign of 'num' before writing the
5655 result to 'rm' because if rm==num the value will be overwritten
5656 and produce the wrong result if 'num' is negative: this caused
5657 problems with BN_mod() and BN_nnmod().
5660 *) Function OCSP_request_verify(). This checks the signature on an
5661 OCSP request and verifies the signer certificate. The signer
5662 certificate is just checked for a generic purpose and OCSP request
5666 *) Add OCSP_check_validity() function to check the validity of OCSP
5667 responses. OCSP responses are prepared in real time and may only
5668 be a few seconds old. Simply checking that the current time lies
5669 between thisUpdate and nextUpdate max reject otherwise valid responses
5670 caused by either OCSP responder or client clock inaccuracy. Instead
5671 we allow thisUpdate and nextUpdate to fall within a certain period of
5672 the current time. The age of the response can also optionally be
5673 checked. Two new options -validity_period and -status_age added to
5677 *) If signature or public key algorithm is unrecognized print out its
5678 OID rather that just UNKNOWN.
5681 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5682 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5683 ID to be generated from the issuer certificate alone which can then be
5684 passed to OCSP_id_issuer_cmp().
5687 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5688 ASN1 modules to export functions returning ASN1_ITEM pointers
5689 instead of the ASN1_ITEM structures themselves. This adds several
5690 new macros which allow the underlying ASN1 function/structure to
5691 be accessed transparently. As a result code should not use ASN1_ITEM
5692 references directly (such as &X509_it) but instead use the relevant
5693 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5694 use of the new ASN1 code on platforms where exporting structures
5695 is problematical (for example in shared libraries) but exporting
5696 functions returning pointers to structures is not.
5699 *) Add support for overriding the generation of SSL/TLS session IDs.
5700 These callbacks can be registered either in an SSL_CTX or per SSL.
5701 The purpose of this is to allow applications to control, if they wish,
5702 the arbitrary values chosen for use as session IDs, particularly as it
5703 can be useful for session caching in multiple-server environments. A
5704 command-line switch for testing this (and any client code that wishes
5705 to use such a feature) has been added to "s_server".
5706 [Geoff Thorpe, Lutz Jaenicke]
5708 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5709 of the form '#if defined(...) || defined(...) || ...' and
5710 '#if !defined(...) && !defined(...) && ...'. This also avoids
5711 the growing number of special cases it was previously handling.
5714 *) Make all configuration macros available for application by making
5715 sure they are available in opensslconf.h, by giving them names starting
5716 with "OPENSSL_" to avoid conflicts with other packages and by making
5717 sure e_os2.h will cover all platform-specific cases together with
5719 Additionally, it is now possible to define configuration/platform-
5720 specific names (called "system identities"). In the C code, these
5721 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5722 macro with the name beginning with "OPENSSL_SYS_", which is determined
5723 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5727 *) New option -set_serial to 'req' and 'x509' this allows the serial
5728 number to use to be specified on the command line. Previously self
5729 signed certificates were hard coded with serial number 0 and the
5730 CA options of 'x509' had to use a serial number in a file which was
5734 *) New options to 'ca' utility to support V2 CRL entry extensions.
5735 Currently CRL reason, invalidity date and hold instruction are
5736 supported. Add new CRL extensions to V3 code and some new objects.
5739 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5740 disable standard block padding (aka PKCS#5 padding) in the EVP
5741 API, which was previously mandatory. This means that the data is
5742 not padded in any way and so the total length much be a multiple
5743 of the block size, otherwise an error occurs.
5746 *) Initial (incomplete) OCSP SSL support.
5749 *) New function OCSP_parse_url(). This splits up a URL into its host,
5750 port and path components: primarily to parse OCSP URLs. New -url
5751 option to ocsp utility.
5754 *) New nonce behavior. The return value of OCSP_check_nonce() now
5755 reflects the various checks performed. Applications can decide
5756 whether to tolerate certain situations such as an absent nonce
5757 in a response when one was present in a request: the ocsp application
5758 just prints out a warning. New function OCSP_add1_basic_nonce()
5759 this is to allow responders to include a nonce in a response even if
5760 the request is nonce-less.
5763 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5764 skipped when using openssl x509 multiple times on a single input file,
5765 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5768 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5769 set string type: to handle setting ASN1_TIME structures. Fix ca
5770 utility to correctly initialize revocation date of CRLs.
5773 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5774 the clients preferred ciphersuites and rather use its own preferences.
5775 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5776 Internet Explorer by ensuring unchanged hash method during stepup.
5777 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5780 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5781 to aes and add a new 'exist' option to print out symbols that don't
5785 *) Additional options to ocsp utility to allow flags to be set and
5786 additional certificates supplied.
5789 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5790 OCSP client a number of certificate to only verify the response
5794 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5795 handle the new API. Currently only ECB, CBC modes supported. Add new
5798 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5799 Encryption Standard (AES) Ciphersuites for Transport Layer
5800 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5801 not enabled by default and were not part of the "ALL" ciphersuite
5802 alias because they were not yet official; they could be
5803 explicitly requested by specifying the "AESdraft" ciphersuite
5804 group alias. In the final release of OpenSSL 0.9.7, the group
5805 alias is called "AES" and is part of "ALL".)
5806 [Ben Laurie, Steve Henson, Bodo Moeller]
5808 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5809 request to response.
5812 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5813 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5814 extract information from a certificate request. OCSP_response_create()
5815 creates a response and optionally adds a basic response structure.
5816 OCSP_basic_add1_status() adds a complete single response to a basic
5817 response and returns the OCSP_SINGLERESP structure just added (to allow
5818 extensions to be included for example). OCSP_basic_add1_cert() adds a
5819 certificate to a basic response and OCSP_basic_sign() signs a basic
5820 response with various flags. New helper functions ASN1_TIME_check()
5821 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5822 (converts ASN1_TIME to GeneralizedTime).
5825 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5826 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5827 structure from a certificate. X509_pubkey_digest() digests the public_key
5828 contents: this is used in various key identifiers.
5831 *) Make sk_sort() tolerate a NULL argument.
5832 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5834 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5835 passed by the function are trusted implicitly. If any of them signed the
5836 response then it is assumed to be valid and is not verified.
5839 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5840 to data. This was previously part of the PKCS7 ASN1 code. This
5841 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5842 [Steve Henson, reported by Kenneth R. Robinette
5843 <support@securenetterm.com>]
5845 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5846 routines: without these tracing memory leaks is very painful.
5847 Fix leaks in PKCS12 and PKCS7 routines.
5850 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5851 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5852 effectively meant GeneralizedTime would never be used. Now it
5853 is initialised to -1 but X509_time_adj() now has to check the value
5854 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5855 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5856 [Steve Henson, reported by Kenneth R. Robinette
5857 <support@securenetterm.com>]
5859 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5860 result in a zero length in the ASN1_INTEGER structure which was
5861 not consistent with the structure when d2i_ASN1_INTEGER() was used
5862 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5863 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5864 where it did not print out a minus for negative ASN1_INTEGER.
5867 *) Add summary printout to ocsp utility. The various functions which
5868 convert status values to strings have been renamed to:
5869 OCSP_response_status_str(), OCSP_cert_status_str() and
5870 OCSP_crl_reason_str() and are no longer static. New options
5871 to verify nonce values and to disable verification. OCSP response
5872 printout format cleaned up.
5875 *) Add additional OCSP certificate checks. These are those specified
5876 in RFC2560. This consists of two separate checks: the CA of the
5877 certificate being checked must either be the OCSP signer certificate
5878 or the issuer of the OCSP signer certificate. In the latter case the
5879 OCSP signer certificate must contain the OCSP signing extended key
5880 usage. This check is performed by attempting to match the OCSP
5881 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5882 in the OCSP_CERTID structures of the response.
5885 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5886 and related routines. This uses the standard OpenSSL certificate
5887 verify routines to perform initial checks (just CA validity) and
5888 to obtain the certificate chain. Then additional checks will be
5889 performed on the chain. Currently the root CA is checked to see
5890 if it is explicitly trusted for OCSP signing. This is used to set
5891 a root CA as a global signing root: that is any certificate that
5892 chains to that CA is an acceptable OCSP signing certificate.
5895 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5896 extensions from a separate configuration file.
5897 As when reading extensions from the main configuration file,
5898 the '-extensions ...' option may be used for specifying the
5900 [Massimiliano Pala <madwolf@comune.modena.it>]
5902 *) New OCSP utility. Allows OCSP requests to be generated or
5903 read. The request can be sent to a responder and the output
5904 parsed, outputed or printed in text form. Not complete yet:
5905 still needs to check the OCSP response validity.
5908 *) New subcommands for 'openssl ca':
5909 'openssl ca -status <serial>' prints the status of the cert with
5910 the given serial number (according to the index file).
5911 'openssl ca -updatedb' updates the expiry status of certificates
5913 [Massimiliano Pala <madwolf@comune.modena.it>]
5915 *) New '-newreq-nodes' command option to CA.pl. This is like
5916 '-newreq', but calls 'openssl req' with the '-nodes' option
5917 so that the resulting key is not encrypted.
5918 [Damien Miller <djm@mindrot.org>]
5920 *) New configuration for the GNU Hurd.
5921 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5923 *) Initial code to implement OCSP basic response verify. This
5924 is currently incomplete. Currently just finds the signer's
5925 certificate and verifies the signature on the response.
5928 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5929 value of OPENSSLDIR. This is available via the new '-d' option
5930 to 'openssl version', and is also included in 'openssl version -a'.
5933 *) Allowing defining memory allocation callbacks that will be given
5934 file name and line number information in additional arguments
5935 (a const char* and an int). The basic functionality remains, as
5936 well as the original possibility to just replace malloc(),
5937 realloc() and free() by functions that do not know about these
5938 additional arguments. To register and find out the current
5939 settings for extended allocation functions, the following
5940 functions are provided:
5942 CRYPTO_set_mem_ex_functions
5943 CRYPTO_set_locked_mem_ex_functions
5944 CRYPTO_get_mem_ex_functions
5945 CRYPTO_get_locked_mem_ex_functions
5947 These work the same way as CRYPTO_set_mem_functions and friends.
5948 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5949 extended allocation function is enabled.
5950 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5951 a conventional allocation function is enabled.
5952 [Richard Levitte, Bodo Moeller]
5954 *) Finish off removing the remaining LHASH function pointer casts.
5955 There should no longer be any prototype-casting required when using
5956 the LHASH abstraction, and any casts that remain are "bugs". See
5957 the callback types and macros at the head of lhash.h for details
5958 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5961 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5962 If /dev/[u]random devices are not available or do not return enough
5963 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5965 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5966 /etc/entropy will be queried once each in this sequence, quering stops
5967 when enough entropy was collected without querying more sockets.
5970 *) Change the Unix RAND_poll() variant to be able to poll several
5971 random devices, as specified by DEVRANDOM, until a sufficient amount
5972 of data has been collected. We spend at most 10 ms on each file
5973 (select timeout) and read in non-blocking mode. DEVRANDOM now
5974 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5975 (previously it was just the string "/dev/urandom"), so on typical
5976 platforms the 10 ms delay will never occur.
5977 Also separate out the Unix variant to its own file, rand_unix.c.
5978 For VMS, there's a currently-empty rand_vms.c.
5981 *) Move OCSP client related routines to ocsp_cl.c. These
5982 provide utility functions which an application needing
5983 to issue a request to an OCSP responder and analyse the
5984 response will typically need: as opposed to those which an
5985 OCSP responder itself would need which will be added later.
5987 OCSP_request_sign() signs an OCSP request with an API similar
5988 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5989 response. OCSP_response_get1_basic() extracts basic response
5990 from response. OCSP_resp_find_status(): finds and extracts status
5991 information from an OCSP_CERTID structure (which will be created
5992 when the request structure is built). These are built from lower
5993 level functions which work on OCSP_SINGLERESP structures but
5994 wont normally be used unless the application wishes to examine
5995 extensions in the OCSP response for example.
5997 Replace nonce routines with a pair of functions.
5998 OCSP_request_add1_nonce() adds a nonce value and optionally
5999 generates a random value. OCSP_check_nonce() checks the
6000 validity of the nonce in an OCSP response.
6003 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6004 This doesn't copy the supplied OCSP_CERTID and avoids the
6005 need to free up the newly created id. Change return type
6006 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6007 This can then be used to add extensions to the request.
6008 Deleted OCSP_request_new(), since most of its functionality
6009 is now in OCSP_REQUEST_new() (and the case insensitive name
6010 clash) apart from the ability to set the request name which
6011 will be added elsewhere.
6014 *) Update OCSP API. Remove obsolete extensions argument from
6015 various functions. Extensions are now handled using the new
6016 OCSP extension code. New simple OCSP HTTP function which
6017 can be used to send requests and parse the response.
6020 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6021 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6022 uses the special reorder version of SET OF to sort the attributes
6023 and reorder them to match the encoded order. This resolves a long
6024 standing problem: a verify on a PKCS7 structure just after signing
6025 it used to fail because the attribute order did not match the
6026 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6027 it uses the received order. This is necessary to tolerate some broken
6028 software that does not order SET OF. This is handled by encoding
6029 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6030 to produce the required SET OF.
6033 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6034 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6035 files to get correct declarations of the ASN.1 item variables.
6038 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6039 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6040 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6041 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6042 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6043 ASN1_ITEM and no wrapper functions.
6046 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6047 replace the old function pointer based I/O routines. Change most of
6048 the *_d2i_bio() and *_d2i_fp() functions to use these.
6051 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6052 lines, recognice more "algorithms" that can be deselected, and make
6053 it complain about algorithm deselection that isn't recognised.
6056 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6057 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6058 to use new functions. Add NO_ASN1_OLD which can be set to remove
6059 some old style ASN1 functions: this can be used to determine if old
6060 code will still work when these eventually go away.
6063 *) New extension functions for OCSP structures, these follow the
6064 same conventions as certificates and CRLs.
6067 *) New function X509V3_add1_i2d(). This automatically encodes and
6068 adds an extension. Its behaviour can be customised with various
6069 flags to append, replace or delete. Various wrappers added for
6070 certifcates and CRLs.
6073 *) Fix to avoid calling the underlying ASN1 print routine when
6074 an extension cannot be parsed. Correct a typo in the
6075 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6078 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
6079 entries for variables.
6082 *) Add functionality to apps/openssl.c for detecting locking
6083 problems: As the program is single-threaded, all we have
6084 to do is register a locking callback using an array for
6085 storing which locks are currently held by the program.
6088 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6089 SSL_get_ex_data_X509_STORE_idx(), which is used in
6090 ssl_verify_cert_chain() and thus can be called at any time
6091 during TLS/SSL handshakes so that thread-safety is essential.
6092 Unfortunately, the ex_data design is not at all suited
6093 for multi-threaded use, so it probably should be abolished.
6096 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6097 [Broadcom, tweaked and integrated by Geoff Thorpe]
6099 *) Move common extension printing code to new function
6100 X509V3_print_extensions(). Reorganise OCSP print routines and
6101 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6104 *) New function X509_signature_print() to remove duplication in some
6108 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6109 set (this was treated exactly the same as SET OF previously). This
6110 is used to reorder the STACK representing the structure to match the
6111 encoding. This will be used to get round a problem where a PKCS7
6112 structure which was signed could not be verified because the STACK
6113 order did not reflect the encoded order.
6116 *) Reimplement the OCSP ASN1 module using the new code.
6119 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6120 for its ASN1 operations. The old style function pointers still exist
6121 for now but they will eventually go away.
6124 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6125 completely replaces the old ASN1 functionality with a table driven
6126 encoder and decoder which interprets an ASN1_ITEM structure describing
6127 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6128 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6129 has also been converted to the new form.
6132 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6133 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6134 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6135 for negative moduli.
6138 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6139 of not touching the result's sign bit.
6142 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6146 *) Changed the LHASH code to use prototypes for callbacks, and created
6147 macros to declare and implement thin (optionally static) functions
6148 that provide type-safety and avoid function pointer casting for the
6149 type-specific callbacks.
6152 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6154 [Veers Staats <staatsvr@asc.hpc.mil>,
6155 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6157 *) Reformat the FAQ so the different questions and answers can be divided
6158 in sections depending on the subject.
6161 *) Have the zlib compression code load ZLIB.DLL dynamically under
6165 *) New function BN_mod_sqrt for computing square roots modulo a prime
6166 (using the probabilistic Tonelli-Shanks algorithm unless
6167 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6168 be handled deterministically).
6169 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6171 *) Make BN_mod_inverse faster by explicitly handling small quotients
6172 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6173 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6176 *) New function BN_kronecker.
6179 *) Fix BN_gcd so that it works on negative inputs; the result is
6180 positive unless both parameters are zero.
6181 Previously something reasonably close to an infinite loop was
6182 possible because numbers could be growing instead of shrinking
6183 in the implementation of Euclid's algorithm.
6186 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6187 sign of the number in question.
6189 Fix BN_is_word(a,w) to work correctly for w == 0.
6191 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6192 because its test if the absolute value of 'a' equals 'w'.
6193 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6194 it exists mostly for use in the implementations of BN_is_zero(),
6195 BN_is_one(), and BN_is_word().
6198 *) New function BN_swap.
6201 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6202 the exponentiation functions are more likely to produce reasonable
6203 results on negative inputs.
6206 *) Change BN_mod_mul so that the result is always non-negative.
6207 Previously, it could be negative if one of the factors was negative;
6208 I don't think anyone really wanted that behaviour.
6211 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6212 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6213 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6214 and add new functions:
6223 BN_mod_lshift1_quick
6227 These functions always generate non-negative results.
6229 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6230 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6232 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6233 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6234 be reduced modulo m.
6235 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6238 The following entry accidentily appeared in the CHANGES file
6239 distributed with OpenSSL 0.9.7. The modifications described in
6240 it do *not* apply to OpenSSL 0.9.7.
6242 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6243 was actually never needed) and in BN_mul(). The removal in BN_mul()
6244 required a small change in bn_mul_part_recursive() and the addition
6245 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6246 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6247 bn_sub_words() and bn_add_words() except they take arrays with
6252 *) In 'openssl passwd', verify passwords read from the terminal
6253 unless the '-salt' option is used (which usually means that
6254 verification would just waste user's time since the resulting
6255 hash is going to be compared with some given password hash)
6256 or the new '-noverify' option is used.
6258 This is an incompatible change, but it does not affect
6259 non-interactive use of 'openssl passwd' (passwords on the command
6260 line, '-stdin' option, '-in ...' option) and thus should not
6264 *) Remove all references to RSAref, since there's no more need for it.
6267 *) Make DSO load along a path given through an environment variable
6268 (SHLIB_PATH) with shl_load().
6271 *) Constify the ENGINE code as a result of BIGNUM constification.
6272 Also constify the RSA code and most things related to it. In a
6273 few places, most notable in the depth of the ASN.1 code, ugly
6274 casts back to non-const were required (to be solved at a later
6278 *) Make it so the openssl application has all engines loaded by default.
6281 *) Constify the BIGNUM routines a little more.
6284 *) Add the following functions:
6286 ENGINE_load_cswift()
6288 ENGINE_load_atalla()
6290 ENGINE_load_builtin_engines()
6292 That way, an application can itself choose if external engines that
6293 are built-in in OpenSSL shall ever be used or not. The benefit is
6294 that applications won't have to be linked with libdl or other dso
6295 libraries unless it's really needed.
6297 Changed 'openssl engine' to load all engines on demand.
6298 Changed the engine header files to avoid the duplication of some
6299 declarations (they differed!).
6302 *) 'openssl engine' can now list capabilities.
6305 *) Better error reporting in 'openssl engine'.
6308 *) Never call load_dh_param(NULL) in s_server.
6311 *) Add engine application. It can currently list engines by name and
6312 identity, and test if they are actually available.
6315 *) Improve RPM specification file by forcing symbolic linking and making
6316 sure the installed documentation is also owned by root.root.
6317 [Damien Miller <djm@mindrot.org>]
6319 *) Give the OpenSSL applications more possibilities to make use of
6320 keys (public as well as private) handled by engines.
6323 *) Add OCSP code that comes from CertCo.
6326 *) Add VMS support for the Rijndael code.
6329 *) Added untested support for Nuron crypto accelerator.
6332 *) Add support for external cryptographic devices. This code was
6333 previously distributed separately as the "engine" branch.
6334 [Geoff Thorpe, Richard Levitte]
6336 *) Rework the filename-translation in the DSO code. It is now possible to
6337 have far greater control over how a "name" is turned into a filename
6338 depending on the operating environment and any oddities about the
6339 different shared library filenames on each system.
6342 *) Support threads on FreeBSD-elf in Configure.
6345 *) Fix for SHA1 assembly problem with MASM: it produces
6346 warnings about corrupt line number information when assembling
6347 with debugging information. This is caused by the overlapping
6349 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6352 NCONF_get_number() has no error checking at all. As a replacement,
6353 NCONF_get_number_e() is defined (_e for "error checking") and is
6354 promoted strongly. The old NCONF_get_number is kept around for
6355 binary backward compatibility.
6356 Make it possible for methods to load from something other than a BIO,
6357 by providing a function pointer that is given a name instead of a BIO.
6358 For example, this could be used to load configuration data from an
6362 *) Fix for non blocking accept BIOs. Added new I/O special reason
6363 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6364 with non blocking I/O was not possible because no retry code was
6365 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6369 *) Added the beginnings of Rijndael support.
6372 *) Fix for bug in DirectoryString mask setting. Add support for
6373 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6374 to allow certificate printing to more controllable, additional
6375 'certopt' option to 'x509' to allow new printing options to be
6379 *) Clean old EAY MD5 hack from e_os.h.
6382 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6384 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6385 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6386 [Joe Orton, Steve Henson]
6388 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6390 *) Fix additional bug revealed by the NISCC test suite:
6392 Stop bug triggering large recursion when presented with
6393 certain ASN.1 tags (CVE-2003-0851)
6396 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6398 *) Fix various bugs revealed by running the NISCC test suite:
6400 Stop out of bounds reads in the ASN1 code when presented with
6401 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6403 If verify callback ignores invalid public key errors don't try to check
6404 certificate signature with the NULL public key.
6408 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6409 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6413 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6414 extra data after the compression methods not only for TLS 1.0
6415 but also for SSL 3.0 (as required by the specification).
6416 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6418 *) Change X509_certificate_type() to mark the key as exported/exportable
6419 when it's 512 *bits* long, not 512 bytes.
6422 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6424 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6425 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6426 a protocol version number mismatch like a decryption error
6427 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6430 *) Turn on RSA blinding by default in the default implementation
6431 to avoid a timing attack. Applications that don't want it can call
6432 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6433 They would be ill-advised to do so in most cases.
6434 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6436 *) Change RSA blinding code so that it works when the PRNG is not
6437 seeded (in this case, the secret RSA exponent is abused as
6438 an unpredictable seed -- if it is not unpredictable, there
6439 is no point in blinding anyway). Make RSA blinding thread-safe
6440 by remembering the creator's thread ID in rsa->blinding and
6441 having all other threads use local one-time blinding factors
6442 (this requires more computation than sharing rsa->blinding, but
6443 avoids excessive locking; and if an RSA object is not shared
6444 between threads, blinding will still be very fast).
6447 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6449 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6450 via timing by performing a MAC computation even if incorrrect
6451 block cipher padding has been found. This is a countermeasure
6452 against active attacks where the attacker has to distinguish
6453 between bad padding and a MAC verification error. (CVE-2003-0078)
6455 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6456 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6457 Martin Vuagnoux (EPFL, Ilion)]
6459 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6461 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6462 memory from it's contents. This is done with a counter that will
6463 place alternating values in each byte. This can be used to solve
6464 two issues: 1) the removal of calls to memset() by highly optimizing
6465 compilers, and 2) cleansing with other values than 0, since those can
6466 be read through on certain media, for example a swap space on disk.
6469 *) Bugfix: client side session caching did not work with external caching,
6470 because the session->cipher setting was not restored when reloading
6471 from the external cache. This problem was masked, when
6472 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6473 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6476 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6477 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6478 [Zeev Lieber <zeev-l@yahoo.com>]
6480 *) Undo an undocumented change introduced in 0.9.6e which caused
6481 repeated calls to OpenSSL_add_all_ciphers() and
6482 OpenSSL_add_all_digests() to be ignored, even after calling
6486 *) Change the default configuration reader to deal with last line not
6487 being properly terminated.
6490 *) Change X509_NAME_cmp() so it applies the special rules on handling
6491 DN values that are of type PrintableString, as well as RDNs of type
6492 emailAddress where the value has the type ia5String.
6493 [stefank@valicert.com via Richard Levitte]
6495 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6496 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6497 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6498 the bitwise-OR of the two for use by the majority of applications
6499 wanting this behaviour, and update the docs. The documented
6500 behaviour and actual behaviour were inconsistent and had been
6501 changing anyway, so this is more a bug-fix than a behavioural
6503 [Geoff Thorpe, diagnosed by Nadav Har'El]
6505 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6506 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6509 *) Fix initialization code race conditions in
6510 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6511 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6512 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6513 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6514 ssl2_get_cipher_by_char(),
6515 ssl3_get_cipher_by_char().
6516 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6518 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6519 the cached sessions are flushed, as the remove_cb() might use ex_data
6520 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6521 (see [openssl.org #212]).
6522 [Geoff Thorpe, Lutz Jaenicke]
6524 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6525 length, instead of the encoding length to d2i_ASN1_OBJECT.
6528 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6530 *) [In 0.9.6g-engine release:]
6531 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6532 [Lynn Gazis <lgazis@rainbow.com>]
6534 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6536 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6537 and get fix the header length calculation.
6538 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6539 Alon Kantor <alonk@checkpoint.com> (and others),
6542 *) Use proper error handling instead of 'assertions' in buffer
6543 overflow checks added in 0.9.6e. This prevents DoS (the
6544 assertions could call abort()).
6545 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6547 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6549 *) Add various sanity checks to asn1_get_length() to reject
6550 the ASN1 length bytes if they exceed sizeof(long), will appear
6551 negative or the content length exceeds the length of the
6553 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6555 *) Fix cipher selection routines: ciphers without encryption had no flags
6556 for the cipher strength set and where therefore not handled correctly
6557 by the selection routines (PR #130).
6560 *) Fix EVP_dsa_sha macro.
6564 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6565 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6566 that was added in OpenSSL 0.9.6d.
6568 As the countermeasure turned out to be incompatible with some
6569 broken SSL implementations, the new option is part of SSL_OP_ALL.
6570 SSL_OP_ALL is usually employed when compatibility with weird SSL
6571 implementations is desired (e.g. '-bugs' option to 's_client' and
6572 's_server'), so the new option is automatically set in many
6576 *) Changes in security patch:
6578 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6579 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6580 Air Force Materiel Command, USAF, under agreement number
6583 *) Add various sanity checks to asn1_get_length() to reject
6584 the ASN1 length bytes if they exceed sizeof(long), will appear
6585 negative or the content length exceeds the length of the
6586 supplied buffer. (CVE-2002-0659)
6587 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6589 *) Assertions for various potential buffer overflows, not known to
6591 [Ben Laurie (CHATS)]
6593 *) Various temporary buffers to hold ASCII versions of integers were
6594 too small for 64 bit platforms. (CVE-2002-0655)
6595 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6597 *) Remote buffer overflow in SSL3 protocol - an attacker could
6598 supply an oversized session ID to a client. (CVE-2002-0656)
6599 [Ben Laurie (CHATS)]
6601 *) Remote buffer overflow in SSL2 protocol - an attacker could
6602 supply an oversized client master key. (CVE-2002-0656)
6603 [Ben Laurie (CHATS)]
6605 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6607 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6608 encoded as NULL) with id-dsa-with-sha1.
6609 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6611 *) Check various X509_...() return values in apps/req.c.
6612 [Nils Larsch <nla@trustcenter.de>]
6614 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6615 an end-of-file condition would erronously be flagged, when the CRLF
6616 was just at the end of a processed block. The bug was discovered when
6617 processing data through a buffering memory BIO handing the data to a
6618 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6619 <ptsekov@syntrex.com> and Nedelcho Stanev.
6622 *) Implement a countermeasure against a vulnerability recently found
6623 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6624 before application data chunks to avoid the use of known IVs
6625 with data potentially chosen by the attacker.
6628 *) Fix length checks in ssl3_get_client_hello().
6631 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6632 to prevent ssl3_read_internal() from incorrectly assuming that
6633 ssl3_read_bytes() found application data while handshake
6634 processing was enabled when in fact s->s3->in_read_app_data was
6635 merely automatically cleared during the initial handshake.
6636 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6638 *) Fix object definitions for Private and Enterprise: they were not
6639 recognized in their shortname (=lowercase) representation. Extend
6640 obj_dat.pl to issue an error when using undefined keywords instead
6641 of silently ignoring the problem (Svenning Sorensen
6642 <sss@sss.dnsalias.net>).
6645 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6646 generators, i.e. generators other than 2 and 5. (Previously, the
6647 code did not properly initialise the 'add' and 'rem' values to
6648 BN_generate_prime().)
6650 In the new general case, we do not insist that 'generator' is
6651 actually a primitive root: This requirement is rather pointless;
6652 a generator of the order-q subgroup is just as good, if not
6656 *) Map new X509 verification errors to alerts. Discovered and submitted by
6657 Tom Wu <tom@arcot.com>.
6660 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6661 returning non-zero before the data has been completely received
6662 when using non-blocking I/O.
6663 [Bodo Moeller; problem pointed out by John Hughes]
6665 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6666 [Ben Laurie, Lutz Jaenicke]
6668 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6669 Yoram Zahavi <YoramZ@gilian.com>).
6672 *) Add information about CygWin 1.3 and on, and preserve proper
6673 configuration for the versions before that.
6674 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6676 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6677 check whether we deal with a copy of a session and do not delete from
6678 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6679 <izhar@checkpoint.com>.
6682 *) Do not store session data into the internal session cache, if it
6683 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6684 flag is set). Proposed by Aslam <aslam@funk.com>.
6687 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6691 *) [In 0.9.6d-engine release:]
6692 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6693 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6695 *) Add the configuration target linux-s390x.
6696 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6698 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6699 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6700 variable as an indication that a ClientHello message has been
6701 received. As the flag value will be lost between multiple
6702 invocations of ssl3_accept when using non-blocking I/O, the
6703 function may not be aware that a handshake has actually taken
6704 place, thus preventing a new session from being added to the
6707 To avoid this problem, we now set s->new_session to 2 instead of
6708 using a local variable.
6709 [Lutz Jaenicke, Bodo Moeller]
6711 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6712 if the SSL_R_LENGTH_MISMATCH error is detected.
6713 [Geoff Thorpe, Bodo Moeller]
6715 *) New 'shared_ldflag' column in Configure platform table.
6718 *) Fix EVP_CIPHER_mode macro.
6719 ["Dan S. Camper" <dan@bti.net>]
6721 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6722 type, we must throw them away by setting rr->length to 0.
6723 [D P Chang <dpc@qualys.com>]
6725 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6727 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6728 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6729 worked incorrectly for those cases where range = 10..._2 and
6730 3*range is two bits longer than range.)
6733 *) Only add signing time to PKCS7 structures if it is not already
6737 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6738 OBJ_ld_ce should be OBJ_id_ce.
6739 Also some ip-pda OIDs in crypto/objects/objects.txt were
6740 incorrect (cf. RFC 3039).
6741 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6743 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6744 returns early because it has nothing to do.
6745 [Andy Schneider <andy.schneider@bjss.co.uk>]
6747 *) [In 0.9.6c-engine release:]
6748 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6749 [Andy Schneider <andy.schneider@bjss.co.uk>]
6751 *) [In 0.9.6c-engine release:]
6752 Add support for Cryptographic Appliance's keyserver technology.
6753 (Use engine 'keyclient')
6754 [Cryptographic Appliances and Geoff Thorpe]
6756 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6757 is called via tools/c89.sh because arguments have to be
6758 rearranged (all '-L' options must appear before the first object
6760 [Richard Shapiro <rshapiro@abinitio.com>]
6762 *) [In 0.9.6c-engine release:]
6763 Add support for Broadcom crypto accelerator cards, backported
6765 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6767 *) [In 0.9.6c-engine release:]
6768 Add support for SureWare crypto accelerator cards from
6769 Baltimore Technologies. (Use engine 'sureware')
6770 [Baltimore Technologies and Mark Cox]
6772 *) [In 0.9.6c-engine release:]
6773 Add support for crypto accelerator cards from Accelerated
6774 Encryption Processing, www.aep.ie. (Use engine 'aep')
6775 [AEP Inc. and Mark Cox]
6777 *) Add a configuration entry for gcc on UnixWare.
6778 [Gary Benson <gbenson@redhat.com>]
6780 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6781 messages are stored in a single piece (fixed-length part and
6782 variable-length part combined) and fix various bugs found on the way.
6785 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6786 instead. BIO_gethostbyname() does not know what timeouts are
6787 appropriate, so entries would stay in cache even when they have
6789 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6791 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6792 faced with a pathologically small ClientHello fragment that does
6793 not contain client_version: Instead of aborting with an error,
6794 simply choose the highest available protocol version (i.e.,
6795 TLS 1.0 unless it is disabled). In practice, ClientHello
6796 messages are never sent like this, but this change gives us
6797 strictly correct behaviour at least for TLS.
6800 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6801 never resets s->method to s->ctx->method when called from within
6802 one of the SSL handshake functions.
6803 [Bodo Moeller; problem pointed out by Niko Baric]
6805 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6806 (sent using the client's version number) if client_version is
6807 smaller than the protocol version in use. Also change
6808 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6809 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6810 the client will at least see that alert.
6813 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6817 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6818 client receives HelloRequest while in a handshake.
6819 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6821 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6822 should end in 'break', not 'goto end' which circuments various
6823 cleanups done in state SSL_ST_OK. But session related stuff
6824 must be disabled for SSL_ST_OK in the case that we just sent a
6827 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6828 before just sending a HelloRequest.
6829 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6831 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6832 reveal whether illegal block cipher padding was found or a MAC
6833 verification error occured. (Neither SSLerr() codes nor alerts
6834 are directly visible to potential attackers, but the information
6835 may leak via logfiles.)
6837 Similar changes are not required for the SSL 2.0 implementation
6838 because the number of padding bytes is sent in clear for SSL 2.0,
6839 and the extra bytes are just ignored. However ssl/s2_pkt.c
6840 failed to verify that the purported number of padding bytes is in
6844 *) Add OpenUNIX-8 support including shared libraries
6845 (Boyd Lynn Gerber <gerberb@zenez.com>).
6848 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6849 'wristwatch attack' using huge encoding parameters (cf.
6850 James H. Manger's CRYPTO 2001 paper). Note that the
6851 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6852 encoding parameters and hence was not vulnerable.
6855 *) BN_sqr() bug fix.
6856 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6858 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6859 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6860 followed by modular reduction.
6861 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6863 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6864 equivalent based on BN_pseudo_rand() instead of BN_rand().
6867 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6868 This function was broken, as the check for a new client hello message
6869 to handle SGC did not allow these large messages.
6870 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6873 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6876 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6877 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6880 *) Rework the configuration and shared library support for Tru64 Unix.
6881 The configuration part makes use of modern compiler features and
6882 still retains old compiler behavior for those that run older versions
6883 of the OS. The shared library support part includes a variant that
6884 uses the RPATH feature, and is available through the special
6885 configuration target "alpha-cc-rpath", which will never be selected
6887 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6889 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6890 with the same message size as in ssl3_get_certificate_request().
6891 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6892 messages might inadvertently be reject as too long.
6893 [Petr Lampa <lampa@fee.vutbr.cz>]
6895 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6898 *) Modified SSL library such that the verify_callback that has been set
6899 specificly for an SSL object with SSL_set_verify() is actually being
6900 used. Before the change, a verify_callback set with this function was
6901 ignored and the verify_callback() set in the SSL_CTX at the time of
6902 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6903 to allow the necessary settings.
6906 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6907 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6908 done automatically (in contradiction to the requirements of the C
6909 standard). This made problems when used from OpenSSH.
6912 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6913 dh->length and always used
6915 BN_rand_range(priv_key, dh->p).
6917 BN_rand_range() is not necessary for Diffie-Hellman, and this
6918 specific range makes Diffie-Hellman unnecessarily inefficient if
6919 dh->length (recommended exponent length) is much smaller than the
6920 length of dh->p. We could use BN_rand_range() if the order of
6921 the subgroup was stored in the DH structure, but we only have
6926 BN_rand(priv_key, l, ...)
6928 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6934 RSA_eay_public_encrypt
6935 RSA_eay_private_decrypt
6936 RSA_eay_private_encrypt (signing)
6937 RSA_eay_public_decrypt (signature verification)
6939 (default implementations for RSA_public_encrypt,
6940 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6941 always reject numbers >= n.
6944 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6945 to synchronize access to 'locking_thread'. This is necessary on
6946 systems where access to 'locking_thread' (an 'unsigned long'
6947 variable) is not atomic.
6950 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6951 *before* setting the 'crypto_lock_rand' flag. The previous code had
6952 a race condition if 0 is a valid thread ID.
6953 [Travis Vitek <vitek@roguewave.com>]
6955 *) Add support for shared libraries under Irix.
6956 [Albert Chin-A-Young <china@thewrittenword.com>]
6958 *) Add configuration option to build on Linux on both big-endian and
6960 [Ralf Baechle <ralf@uni-koblenz.de>]
6962 *) Add the possibility to create shared libraries on HP-UX.
6965 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6967 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6968 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6969 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6970 PRNG state recovery was possible based on the output of
6971 one PRNG request appropriately sized to gain knowledge on
6972 'md' followed by enough consecutive 1-byte PRNG requests
6973 to traverse all of 'state'.
6975 1. When updating 'md_local' (the current thread's copy of 'md')
6976 during PRNG output generation, hash all of the previous
6977 'md_local' value, not just the half used for PRNG output.
6979 2. Make the number of bytes from 'state' included into the hash
6980 independent from the number of PRNG bytes requested.
6982 The first measure alone would be sufficient to avoid
6983 Markku-Juhani's attack. (Actually it had never occurred
6984 to me that the half of 'md_local' used for chaining was the
6985 half from which PRNG output bytes were taken -- I had always
6986 assumed that the secret half would be used.) The second
6987 measure makes sure that additional data from 'state' is never
6988 mixed into 'md_local' in small portions; this heuristically
6989 further strengthens the PRNG.
6992 *) Fix crypto/bn/asm/mips3.s.
6995 *) When only the key is given to "enc", the IV is undefined. Print out
6996 an error message in this case.
6999 *) Handle special case when X509_NAME is empty in X509 printing routines.
7002 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7003 positive and less than q.
7006 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7007 used: it isn't thread safe and the add_lock_callback should handle
7009 [Paul Rose <Paul.Rose@bridge.com>]
7011 *) Verify that incoming data obeys the block size in
7012 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7016 [Ulf Möller, Bodo Möller]
7018 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7019 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7020 when fixing the server behaviour for backwards-compatible 'client
7021 hello' messages. (Note that the attack is impractical against
7022 SSL 3.0 and TLS 1.0 anyway because length and version checking
7023 means that the probability of guessing a valid ciphertext is
7024 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7027 Before 0.9.5, the countermeasure (hide the error by generating a
7028 random 'decryption result') did not work properly because
7029 ERR_clear_error() was missing, meaning that SSL_get_error() would
7030 detect the supposedly ignored error.
7032 Both problems are now fixed.
7035 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7036 (previously it was 1024).
7039 *) Fix for compatibility mode trust settings: ignore trust settings
7040 unless some valid trust or reject settings are present.
7043 *) Fix for blowfish EVP: its a variable length cipher.
7046 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7047 parameters in DSA public key structures and return an error in the
7048 DSA routines if parameters are absent.
7051 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7052 in the current directory if neither $RANDFILE nor $HOME was set.
7053 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7054 caused some confusion to Windows users who haven't defined $HOME.
7055 Thus RAND_file_name() is changed again: e_os.h can define a
7056 DEFAULT_HOME, which will be used if $HOME is not set.
7057 For Windows, we use "C:"; on other platforms, we still require
7058 environment variables.
7060 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7061 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7062 having multiple threads call RAND_poll() concurrently.
7065 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7066 combination of a flag and a thread ID variable.
7067 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7068 flag), *other* threads can enter ssleay_add_bytes without obeying
7069 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7070 that they do not hold after the first thread unsets add_do_not_lock).
7073 *) Change bctest again: '-x' expressions are not available in all
7077 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7079 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7080 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7082 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7083 the default extension for executables, if any. Also, make the perl
7084 scripts that use symlink() to test if it really exists and use "cp"
7085 if it doesn't. All this made OpenSSL compilable and installable in
7089 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7090 If SEQUENCE is length is indefinite just set c->slen to the total
7091 amount of data available.
7092 [Steve Henson, reported by shige@FreeBSD.org]
7093 [This change does not apply to 0.9.7.]
7095 *) Change bctest to avoid here-documents inside command substitution
7096 (workaround for FreeBSD /bin/sh bug).
7097 For compatibility with Ultrix, avoid shell functions (introduced
7098 in the bctest version that searches along $PATH).
7101 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7102 with des_encrypt() defined on some operating systems, like Solaris
7106 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7107 On the Importance of Eliminating Errors in Cryptographic
7108 Computations, J. Cryptology 14 (2001) 2, 101-119,
7109 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7112 *) MIPS assembler BIGNUM division bug fix.
7115 *) Disabled incorrect Alpha assembler code.
7118 *) Fix PKCS#7 decode routines so they correctly update the length
7119 after reading an EOC for the EXPLICIT tag.
7121 [This change does not apply to 0.9.7.]
7123 *) Fix bug in PKCS#12 key generation routines. This was triggered
7124 if a 3DES key was generated with a 0 initial byte. Include
7125 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7126 (but broken) behaviour.
7129 *) Enhance bctest to search for a working bc along $PATH and print
7131 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7133 *) Fix memory leaks in err.c: free err_data string if necessary;
7134 don't write to the wrong index in ERR_set_error_data.
7137 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7141 *) Replace rdtsc with _emit statements for VC++ version 5.
7142 [Jeremy Cooper <jeremy@baymoo.org>]
7144 *) Make it possible to reuse SSLv2 sessions.
7147 *) In copy_email() check for >= 0 as a return value for
7148 X509_NAME_get_index_by_NID() since 0 is a valid index.
7149 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7151 *) Avoid coredump with unsupported or invalid public keys by checking if
7152 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7153 PKCS7_verify() fails with non detached data.
7156 *) Don't use getenv in library functions when run as setuid/setgid.
7157 New function OPENSSL_issetugid().
7160 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7161 due to incorrect handling of multi-threading:
7163 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7165 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7167 3. Count how many times MemCheck_off() has been called so that
7168 nested use can be treated correctly. This also avoids
7169 inband-signalling in the previous code (which relied on the
7170 assumption that thread ID 0 is impossible).
7173 *) Add "-rand" option also to s_client and s_server.
7176 *) Fix CPU detection on Irix 6.x.
7177 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7178 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7180 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7183 [This change does not apply to 0.9.7.]
7185 *) Use the cached encoding of an X509_NAME structure rather than
7186 copying it. This is apparently the reason for the libsafe "errors"
7187 but the code is actually correct.
7190 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7191 Bleichenbacher's DSA attack.
7192 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7193 to be set and top=0 forces the highest bit to be set; top=-1 is new
7194 and leaves the highest bit random.
7195 [Ulf Moeller, Bodo Moeller]
7197 *) In the NCONF_...-based implementations for CONF_... queries
7198 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7199 a temporary CONF structure with the data component set to NULL
7200 (which gives segmentation faults in lh_retrieve).
7201 Instead, use NULL for the CONF pointer in CONF_get_string and
7202 CONF_get_number (which may use environment variables) and directly
7203 return NULL from CONF_get_section.
7206 *) Fix potential buffer overrun for EBCDIC.
7209 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7210 keyUsage if basicConstraints absent for a CA.
7213 *) Make SMIME_write_PKCS7() write mail header values with a format that
7214 is more generally accepted (no spaces before the semicolon), since
7215 some programs can't parse those values properly otherwise. Also make
7216 sure BIO's that break lines after each write do not create invalid
7220 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7221 macros previously used would not encode an empty SEQUENCE OF
7222 and break the signature.
7224 [This change does not apply to 0.9.7.]
7226 *) Zero the premaster secret after deriving the master secret in
7230 *) Add some EVP_add_digest_alias registrations (as found in
7231 OpenSSL_add_all_digests()) to SSL_library_init()
7232 aka OpenSSL_add_ssl_algorithms(). This provides improved
7233 compatibility with peers using X.509 certificates
7234 with unconventional AlgorithmIdentifier OIDs.
7237 *) Fix for Irix with NO_ASM.
7238 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7240 *) ./config script fixes.
7241 [Ulf Moeller, Richard Levitte]
7243 *) Fix 'openssl passwd -1'.
7246 *) Change PKCS12_key_gen_asc() so it can cope with non null
7247 terminated strings whose length is passed in the passlen
7248 parameter, for example from PEM callbacks. This was done
7249 by adding an extra length parameter to asc2uni().
7250 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7252 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7253 call failed, free the DSA structure.
7256 *) Fix to uni2asc() to cope with zero length Unicode strings.
7257 These are present in some PKCS#12 files.
7260 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7261 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7262 when writing a 32767 byte record.
7263 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7265 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7266 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7268 (RSA objects have a reference count access to which is protected
7269 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7270 so they are meant to be shared between threads.)
7271 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7272 "Reddie, Steven" <Steven.Reddie@ca.com>]
7274 *) Fix a deadlock in CRYPTO_mem_leaks().
7277 *) Use better test patterns in bntest.
7280 *) rand_win.c fix for Borland C.
7283 *) BN_rshift bugfix for n == 0.
7286 *) Add a 'bctest' script that checks for some known 'bc' bugs
7287 so that 'make test' does not abort just because 'bc' is broken.
7290 *) Store verify_result within SSL_SESSION also for client side to
7291 avoid potential security hole. (Re-used sessions on the client side
7292 always resulted in verify_result==X509_V_OK, not using the original
7293 result of the server certificate verification.)
7296 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7297 SSL3_RT_APPLICATION_DATA, return 0.
7298 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7302 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7303 releases, have been re-implemented by renaming the previous
7304 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7305 and ssl3_read_internal, respectively, and adding 'peek' parameters
7306 to them. The new ssl[23]_{read,peek} functions are calls to
7307 ssl[23]_read_internal with the 'peek' flag set appropriately.
7308 A 'peek' parameter has also been added to ssl3_read_bytes, which
7309 does the actual work for ssl3_read_internal.
7312 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7313 the method-specific "init()" handler. Also clean up ex_data after
7314 calling the method-specific "finish()" handler. Previously, this was
7315 happening the other way round.
7318 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7319 The previous value, 12, was not always sufficient for BN_mod_exp().
7322 *) Make sure that shared libraries get the internal name engine with
7323 the full version number and not just 0. This should mark the
7324 shared libraries as not backward compatible. Of course, this should
7325 be changed again when we can guarantee backward binary compatibility.
7328 *) Fix typo in get_cert_by_subject() in by_dir.c
7329 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7331 *) Rework the system to generate shared libraries:
7333 - Make note of the expected extension for the shared libraries and
7334 if there is a need for symbolic links from for example libcrypto.so.0
7335 to libcrypto.so.0.9.7. There is extended info in Configure for
7338 - Make as few rebuilds of the shared libraries as possible.
7340 - Still avoid linking the OpenSSL programs with the shared libraries.
7342 - When installing, install the shared libraries separately from the
7346 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7348 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7349 and not in SSL_clear because the latter is also used by the
7350 accept/connect functions; previously, the settings made by
7351 SSL_set_read_ahead would be lost during the handshake.
7352 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7354 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7355 Previously, it would create entries for disableed algorithms no
7359 *) Added several new manual pages for SSL_* function.
7362 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7364 *) In ssl23_get_client_hello, generate an error message when faced
7365 with an initial SSL 3.0/TLS record that is too small to contain the
7366 first two bytes of the ClientHello message, i.e. client_version.
7367 (Note that this is a pathologic case that probably has never happened
7368 in real life.) The previous approach was to use the version number
7369 from the record header as a substitute; but our protocol choice
7370 should not depend on that one because it is not authenticated
7371 by the Finished messages.
7374 *) More robust randomness gathering functions for Windows.
7375 [Jeffrey Altman <jaltman@columbia.edu>]
7377 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7378 not set then we don't setup the error code for issuer check errors
7379 to avoid possibly overwriting other errors which the callback does
7380 handle. If an application does set the flag then we assume it knows
7381 what it is doing and can handle the new informational codes
7385 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7386 a general "ANY" type, as such it should be able to decode anything
7387 including tagged types. However it didn't check the class so it would
7388 wrongly interpret tagged types in the same way as their universal
7389 counterpart and unknown types were just rejected. Changed so that the
7390 tagged and unknown types are handled in the same way as a SEQUENCE:
7391 that is the encoding is stored intact. There is also a new type
7392 "V_ASN1_OTHER" which is used when the class is not universal, in this
7393 case we have no idea what the actual type is so we just lump them all
7397 *) On VMS, stdout may very well lead to a file that is written to
7398 in a record-oriented fashion. That means that every write() will
7399 write a separate record, which will be read separately by the
7400 programs trying to read from it. This can be very confusing.
7402 The solution is to put a BIO filter in the way that will buffer
7403 text until a linefeed is reached, and then write everything a
7404 line at a time, so every record written will be an actual line,
7405 not chunks of lines and not (usually doesn't happen, but I've
7406 seen it once) several lines in one record. BIO_f_linebuffer() is
7409 Currently, it's a VMS-only method, because that's where it has
7410 been tested well enough.
7413 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7414 it can return incorrect results.
7415 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7416 but it was in 0.9.6-beta[12].)
7419 *) Disable the check for content being present when verifying detached
7420 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7421 include zero length content when signing messages.
7424 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7425 BIO_ctrl (for BIO pairs).
7428 *) Add DSO method for VMS.
7431 *) Bug fix: Montgomery multiplication could produce results with the
7435 *) Add RPM specification openssl.spec and modify it to build three
7436 packages. The default package contains applications, application
7437 documentation and run-time libraries. The devel package contains
7438 include files, static libraries and function documentation. The
7439 doc package contains the contents of the doc directory. The original
7440 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7443 *) Add a large number of documentation files for many SSL routines.
7444 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7446 *) Add a configuration entry for Sony News 4.
7447 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7449 *) Don't set the two most significant bits to one when generating a
7450 random number < q in the DSA library.
7453 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7454 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7455 the underlying transport is blocking) if a handshake took place.
7456 (The default behaviour is needed by applications such as s_client
7457 and s_server that use select() to determine when to use SSL_read;
7458 but for applications that know in advance when to expect data, it
7459 just makes things more complicated.)
7462 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7466 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7467 work better on such systems.
7468 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7470 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7471 Update PKCS12_parse() so it copies the friendlyName and the
7472 keyid to the certificates aux info.
7475 *) Fix bug in PKCS7_verify() which caused an infinite loop
7476 if there was more than one signature.
7477 [Sven Uszpelkat <su@celocom.de>]
7479 *) Major change in util/mkdef.pl to include extra information
7480 about each symbol, as well as presentig variables as well
7481 as functions. This change means that there's n more need
7482 to rebuild the .num files when some algorithms are excluded.
7485 *) Allow the verify time to be set by an application,
7486 rather than always using the current time.
7489 *) Phase 2 verify code reorganisation. The certificate
7490 verify code now looks up an issuer certificate by a
7491 number of criteria: subject name, authority key id
7492 and key usage. It also verifies self signed certificates
7493 by the same criteria. The main comparison function is
7494 X509_check_issued() which performs these checks.
7496 Lot of changes were necessary in order to support this
7497 without completely rewriting the lookup code.
7499 Authority and subject key identifier are now cached.
7501 The LHASH 'certs' is X509_STORE has now been replaced
7502 by a STACK_OF(X509_OBJECT). This is mainly because an
7503 LHASH can't store or retrieve multiple objects with
7504 the same hash value.
7506 As a result various functions (which were all internal
7507 use only) have changed to handle the new X509_STORE
7508 structure. This will break anything that messed round
7509 with X509_STORE internally.
7511 The functions X509_STORE_add_cert() now checks for an
7512 exact match, rather than just subject name.
7514 The X509_STORE API doesn't directly support the retrieval
7515 of multiple certificates matching a given criteria, however
7516 this can be worked round by performing a lookup first
7517 (which will fill the cache with candidate certificates)
7518 and then examining the cache for matches. This is probably
7519 the best we can do without throwing out X509_LOOKUP
7520 entirely (maybe later...).
7522 The X509_VERIFY_CTX structure has been enhanced considerably.
7524 All certificate lookup operations now go via a get_issuer()
7525 callback. Although this currently uses an X509_STORE it
7526 can be replaced by custom lookups. This is a simple way
7527 to bypass the X509_STORE hackery necessary to make this
7528 work and makes it possible to use more efficient techniques
7529 in future. A very simple version which uses a simple
7530 STACK for its trusted certificate store is also provided
7531 using X509_STORE_CTX_trusted_stack().
7533 The verify_cb() and verify() callbacks now have equivalents
7534 in the X509_STORE_CTX structure.
7536 X509_STORE_CTX also has a 'flags' field which can be used
7537 to customise the verify behaviour.
7540 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7541 excludes S/MIME capabilities.
7544 *) When a certificate request is read in keep a copy of the
7545 original encoding of the signed data and use it when outputing
7546 again. Signatures then use the original encoding rather than
7547 a decoded, encoded version which may cause problems if the
7548 request is improperly encoded.
7551 *) For consistency with other BIO_puts implementations, call
7552 buffer_write(b, ...) directly in buffer_puts instead of calling
7555 In BIO_puts, increment b->num_write as in BIO_write.
7556 [Peter.Sylvester@EdelWeb.fr]
7558 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7559 BN_zero, we may not return a BIGNUM with an array consisting of
7563 *) Avoid calling abort() from within the library when problems are
7564 detected, except if preprocessor symbols have been defined
7565 (such as REF_CHECK, BN_DEBUG etc.).
7568 *) New openssl application 'rsautl'. This utility can be
7569 used for low level RSA operations. DER public key
7570 BIO/fp routines also added.
7573 *) New Configure entry and patches for compiling on QNX 4.
7574 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7576 *) A demo state-machine implementation was sponsored by
7577 Nuron (http://www.nuron.com/) and is now available in
7578 demos/state_machine.
7581 *) New options added to the 'dgst' utility for signature
7582 generation and verification.
7585 *) Unrecognized PKCS#7 content types are now handled via a
7586 catch all ASN1_TYPE structure. This allows unsupported
7587 types to be stored as a "blob" and an application can
7588 encode and decode it manually.
7591 *) Fix various signed/unsigned issues to make a_strex.c
7593 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7595 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7596 length if passed a buffer. ASN1_INTEGER_to_BN failed
7597 if passed a NULL BN and its argument was negative.
7598 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7600 *) Modification to PKCS#7 encoding routines to output definite
7601 length encoding. Since currently the whole structures are in
7602 memory there's not real point in using indefinite length
7603 constructed encoding. However if OpenSSL is compiled with
7604 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7607 *) Added BIO_vprintf() and BIO_vsnprintf().
7610 *) Added more prefixes to parse for in the the strings written
7611 through a logging bio, to cover all the levels that are available
7612 through syslog. The prefixes are now:
7614 PANIC, EMERG, EMR => LOG_EMERG
7615 ALERT, ALR => LOG_ALERT
7616 CRIT, CRI => LOG_CRIT
7617 ERROR, ERR => LOG_ERR
7618 WARNING, WARN, WAR => LOG_WARNING
7619 NOTICE, NOTE, NOT => LOG_NOTICE
7620 INFO, INF => LOG_INFO
7621 DEBUG, DBG => LOG_DEBUG
7623 and as before, if none of those prefixes are present at the
7624 beginning of the string, LOG_ERR is chosen.
7626 On Win32, the LOG_* levels are mapped according to this:
7628 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7629 LOG_WARNING => EVENTLOG_WARNING_TYPE
7630 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7634 *) Made it possible to reconfigure with just the configuration
7635 argument "reconf" or "reconfigure". The command line arguments
7636 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7637 and are retrieved from there when reconfiguring.
7641 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7643 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7646 *) The obj_dat.pl script was messing up the sorting of object
7647 names. The reason was that it compared the quoted version
7648 of strings as a result "OCSP" > "OCSP Signing" because
7649 " > SPACE. Changed script to store unquoted versions of
7650 names and add quotes on output. It was also omitting some
7651 names from the lookup table if they were given a default
7652 value (that is if SN is missing it is given the same
7653 value as LN and vice versa), these are now added on the
7654 grounds that if an object has a name we should be able to
7655 look it up. Finally added warning output when duplicate
7656 short or long names are found.
7659 *) Changes needed for Tandem NSK.
7660 [Scott Uroff <scott@xypro.com>]
7662 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7663 RSA_padding_check_SSLv23(), special padding was never detected
7664 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7665 version rollback attacks was not effective.
7667 In s23_clnt.c, don't use special rollback-attack detection padding
7668 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7669 client; similarly, in s23_srvr.c, don't do the rollback check if
7670 SSL 2.0 is the only protocol enabled in the server.
7673 *) Make it possible to get hexdumps of unprintable data with 'openssl
7674 asn1parse'. By implication, the functions ASN1_parse_dump() and
7675 BIO_dump_indent() are added.
7678 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7679 these print out strings and name structures based on various
7680 flags including RFC2253 support and proper handling of
7681 multibyte characters. Added options to the 'x509' utility
7682 to allow the various flags to be set.
7685 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7686 Also change the functions X509_cmp_current_time() and
7687 X509_gmtime_adj() work with an ASN1_TIME structure,
7688 this will enable certificates using GeneralizedTime in validity
7689 dates to be checked.
7692 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7693 negative public key encodings) on by default,
7694 NO_NEG_PUBKEY_BUG can be set to disable it.
7697 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7698 content octets. An i2c_ASN1_OBJECT is unnecessary because
7699 the encoding can be trivially obtained from the structure.
7702 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7703 not read locks (CRYPTO_r_[un]lock).
7706 *) A first attempt at creating official support for shared
7707 libraries through configuration. I've kept it so the
7708 default is static libraries only, and the OpenSSL programs
7709 are always statically linked for now, but there are
7710 preparations for dynamic linking in place.
7711 This has been tested on Linux and Tru64.
7714 *) Randomness polling function for Win9x, as described in:
7715 Peter Gutmann, Software Generation of Practically Strong
7719 *) Fix so PRNG is seeded in req if using an already existing
7723 *) New options to smime application. -inform and -outform
7724 allow alternative formats for the S/MIME message including
7725 PEM and DER. The -content option allows the content to be
7726 specified separately. This should allow things like Netscape
7727 form signing output easier to verify.
7730 *) Fix the ASN1 encoding of tags using the 'long form'.
7733 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7734 STRING types. These convert content octets to and from the
7735 underlying type. The actual tag and length octets are
7736 already assumed to have been read in and checked. These
7737 are needed because all other string types have virtually
7738 identical handling apart from the tag. By having versions
7739 of the ASN1 functions that just operate on content octets
7740 IMPLICIT tagging can be handled properly. It also allows
7741 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7742 and ASN1_INTEGER are identical apart from the tag.
7745 *) Change the handling of OID objects as follows:
7747 - New object identifiers are inserted in objects.txt, following
7748 the syntax given in objects.README.
7749 - objects.pl is used to process obj_mac.num and create a new
7751 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7754 This is currently kind of a hack, and the perl code in objects.pl
7755 isn't very elegant, but it works as I intended. The simplest way
7756 to check that it worked correctly is to look in obj_dat.h and
7757 check the array nid_objs and make sure the objects haven't moved
7758 around (this is important!). Additions are OK, as well as
7759 consistent name changes.
7762 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7765 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7766 The given file adds to whatever has already been seeded into the
7767 random pool through the RANDFILE configuration file option or
7768 environment variable, or the default random state file.
7771 *) mkstack.pl now sorts each macro group into lexical order.
7772 Previously the output order depended on the order the files
7773 appeared in the directory, resulting in needless rewriting
7777 *) Patches to make OpenSSL compile under Win32 again. Mostly
7778 work arounds for the VC++ problem that it treats func() as
7779 func(void). Also stripped out the parts of mkdef.pl that
7780 added extra typesafe functions: these no longer exist.
7783 *) Reorganisation of the stack code. The macros are now all
7784 collected in safestack.h . Each macro is defined in terms of
7785 a "stack macro" of the form SKM_<name>(type, a, b). The
7786 DEBUG_SAFESTACK is now handled in terms of function casts,
7787 this has the advantage of retaining type safety without the
7788 use of additional functions. If DEBUG_SAFESTACK is not defined
7789 then the non typesafe macros are used instead. Also modified the
7790 mkstack.pl script to handle the new form. Needs testing to see
7791 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7792 the default if no major problems. Similar behaviour for ASN1_SET_OF
7793 and PKCS12_STACK_OF.
7796 *) When some versions of IIS use the 'NET' form of private key the
7797 key derivation algorithm is different. Normally MD5(password) is
7798 used as a 128 bit RC4 key. In the modified case
7799 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7800 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7801 as the old Netscape_RSA functions except they have an additional
7802 'sgckey' parameter which uses the modified algorithm. Also added
7803 an -sgckey command line option to the rsa utility. Thanks to
7804 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7805 algorithm to openssl-dev.
7808 *) The evp_local.h macros were using 'c.##kname' which resulted in
7809 invalid expansion on some systems (SCO 5.0.5 for example).
7810 Corrected to 'c.kname'.
7811 [Phillip Porch <root@theporch.com>]
7813 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7814 a STACK of email addresses from a certificate or request, these look
7815 in the subject name and the subject alternative name extensions and
7816 omit any duplicate addresses.
7819 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7820 This makes DSA verification about 2 % faster.
7823 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7824 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7825 plus overhead for 1024 bit moduli).
7826 This makes exponentiations about 0.5 % faster for 1024 bit
7827 exponents (as measured by "openssl speed rsa2048").
7830 *) Rename memory handling macros to avoid conflicts with other
7832 Malloc => OPENSSL_malloc
7833 Malloc_locked => OPENSSL_malloc_locked
7834 Realloc => OPENSSL_realloc
7835 Free => OPENSSL_free
7838 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7839 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7842 *) CygWin32 support.
7843 [John Jarvie <jjarvie@newsguy.com>]
7845 *) The type-safe stack code has been rejigged. It is now only compiled
7846 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7847 by default all type-specific stack functions are "#define"d back to
7848 standard stack functions. This results in more streamlined output
7849 but retains the type-safety checking possibilities of the original
7853 *) The STACK code has been cleaned up, and certain type declarations
7854 that didn't make a lot of sense have been brought in line. This has
7855 also involved a cleanup of sorts in safestack.h to more correctly
7856 map type-safe stack functions onto their plain stack counterparts.
7857 This work has also resulted in a variety of "const"ifications of
7858 lots of the code, especially "_cmp" operations which should normally
7859 be prototyped with "const" parameters anyway.
7862 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7863 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7864 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7865 where all of 'md' is used each time the PRNG is used, but 'state'
7866 is used only indexed by a cyclic counter. As entropy may not be
7867 well distributed from the beginning, 'md' is important as a
7868 chaining variable. However, the output function chains only half
7869 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7870 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7871 in all of 'state' being rewritten, with the new values depending
7872 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7875 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7876 the handshake is continued after ssl_verify_cert_chain();
7877 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7878 can lead to 'unexplainable' connection aborts later.
7879 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7881 *) Major EVP API cipher revision.
7882 Add hooks for extra EVP features. This allows various cipher
7883 parameters to be set in the EVP interface. Support added for variable
7884 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7885 setting of RC2 and RC5 parameters.
7887 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7890 Remove lots of duplicated code from the EVP library. For example *every*
7891 cipher init() function handles the 'iv' in the same way according to the
7892 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7893 for CFB and OFB modes they zero ctx->num.
7895 New functionality allows removal of S/MIME code RC2 hack.
7897 Most of the routines have the same form and so can be declared in terms
7900 By shifting this to the top level EVP_CipherInit() it can be removed from
7901 all individual ciphers. If the cipher wants to handle IVs or keys
7902 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7905 Change lots of functions like EVP_EncryptUpdate() to now return a
7906 value: although software versions of the algorithms cannot fail
7907 any installed hardware versions can.
7910 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7911 this option is set, tolerate broken clients that send the negotiated
7912 protocol version number instead of the requested protocol version
7916 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7917 i.e. non-zero for export ciphersuites, zero otherwise.
7918 Previous versions had this flag inverted, inconsistent with
7919 rsa_tmp_cb (..._TMP_RSA_CB).
7920 [Bodo Moeller; problem reported by Amit Chopra]
7922 *) Add missing DSA library text string. Work around for some IIS
7923 key files with invalid SEQUENCE encoding.
7926 *) Add a document (doc/standards.txt) that list all kinds of standards
7927 and so on that are implemented in OpenSSL.
7930 *) Enhance c_rehash script. Old version would mishandle certificates
7931 with the same subject name hash and wouldn't handle CRLs at all.
7932 Added -fingerprint option to crl utility, to support new c_rehash
7936 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7939 *) Fix for SSL server purpose checking. Server checking was
7940 rejecting certificates which had extended key usage present
7941 but no ssl client purpose.
7942 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7944 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7945 is a little unclear about how a blank password is handled.
7946 Since the password in encoded as a BMPString with terminating
7947 double NULL a zero length password would end up as just the
7948 double NULL. However no password at all is different and is
7949 handled differently in the PKCS#12 key generation code. NS
7950 treats a blank password as zero length. MSIE treats it as no
7951 password on export: but it will try both on import. We now do
7952 the same: PKCS12_parse() tries zero length and no password if
7953 the password is set to "" or NULL (NULL is now a valid password:
7954 it wasn't before) as does the pkcs12 application.
7957 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7958 perror when PEM_read_bio_X509_REQ fails, the error message must
7959 be obtained from the error queue.
7962 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7963 it in ERR_remove_state if appropriate, and change ERR_get_state
7964 accordingly to avoid race conditions (this is necessary because
7965 thread_hash is no longer constant once set).
7968 *) Bugfix for linux-elf makefile.one.
7971 *) RSA_get_default_method() will now cause a default
7972 RSA_METHOD to be chosen if one doesn't exist already.
7973 Previously this was only set during a call to RSA_new()
7974 or RSA_new_method(NULL) meaning it was possible for
7975 RSA_get_default_method() to return NULL.
7978 *) Added native name translation to the existing DSO code
7979 that will convert (if the flag to do so is set) filenames
7980 that are sufficiently small and have no path information
7981 into a canonical native form. Eg. "blah" converted to
7982 "libblah.so" or "blah.dll" etc.
7985 *) New function ERR_error_string_n(e, buf, len) which is like
7986 ERR_error_string(e, buf), but writes at most 'len' bytes
7987 including the 0 terminator. For ERR_error_string_n, 'buf'
7989 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7991 *) CONF library reworked to become more general. A new CONF
7992 configuration file reader "class" is implemented as well as a
7993 new functions (NCONF_*, for "New CONF") to handle it. The now
7994 old CONF_* functions are still there, but are reimplemented to
7995 work in terms of the new functions. Also, a set of functions
7996 to handle the internal storage of the configuration data is
7997 provided to make it easier to write new configuration file
7998 reader "classes" (I can definitely see something reading a
7999 configuration file in XML format, for example), called _CONF_*,
8000 or "the configuration storage API"...
8002 The new configuration file reading functions are:
8004 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8005 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8007 NCONF_default, NCONF_WIN32
8009 NCONF_dump_fp, NCONF_dump_bio
8011 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8012 NCONF_new creates a new CONF object. This works in the same way
8013 as other interfaces in OpenSSL, like the BIO interface.
8014 NCONF_dump_* dump the internal storage of the configuration file,
8015 which is useful for debugging. All other functions take the same
8016 arguments as the old CONF_* functions wth the exception of the
8017 first that must be a `CONF *' instead of a `LHASH *'.
8019 To make it easer to use the new classes with the old CONF_* functions,
8020 the function CONF_set_default_method is provided.
8023 *) Add '-tls1' option to 'openssl ciphers', which was already
8024 mentioned in the documentation but had not been implemented.
8025 (This option is not yet really useful because even the additional
8026 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8029 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8030 OpenSSL-based applications) load shared libraries and bind to
8031 them in a portable way.
8032 [Geoff Thorpe, with contributions from Richard Levitte]
8034 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8036 *) Make sure _lrotl and _lrotr are only used with MSVC.
8038 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8039 (the default implementation of RAND_status).
8041 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8042 to '-clrext' (= clear extensions), as intended and documented.
8043 [Bodo Moeller; inconsistency pointed out by Michael Attili
8046 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8047 was larger than the MD block size.
8048 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8050 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8051 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8052 using the passed key: if the passed key was a private key the result
8053 of X509_print(), for example, would be to print out all the private key
8057 *) des_quad_cksum() byte order bug fix.
8058 [Ulf Möller, using the problem description in krb4-0.9.7, where
8059 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8061 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8063 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8065 *) For easily testing in shell scripts whether some command
8066 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8067 returns with exit code 0 iff no command of the given name is available.
8068 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8069 the output goes to stdout and nothing is printed to stderr.
8070 Additional arguments are always ignored.
8072 Since for each cipher there is a command of the same name,
8073 the 'no-cipher' compilation switches can be tested this way.
8075 ('openssl no-XXX' is not able to detect pseudo-commands such
8076 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8079 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8082 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8083 is set; it will be thrown away anyway because each handshake creates
8085 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8086 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8087 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
8088 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8091 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8092 'Q' and 'R' lose their special meanings (quit/renegotiate).
8093 This is part of what -quiet does; unlike -quiet, -ign_eof
8094 does not suppress any output.
8097 *) Add compatibility options to the purpose and trust code. The
8098 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8099 accepts a certificate or CA, this was the previous behaviour,
8100 with all the associated security issues.
8102 X509_TRUST_COMPAT is the old trust behaviour: only and
8103 automatically trust self signed roots in certificate store. A
8104 new trust setting X509_TRUST_DEFAULT is used to specify that
8105 a purpose has no associated trust setting and it should instead
8106 use the value in the default purpose.
8109 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8110 and fix a memory leak.
8113 *) In util/mkerr.pl (which implements 'make errors'), preserve
8114 reason strings from the previous version of the .c file, as
8115 the default to have only downcase letters (and digits) in
8116 automatically generated reasons codes is not always appropriate.
8119 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8120 using strerror. Previously, ERR_reason_error_string() returned
8121 library names as reason strings for SYSerr; but SYSerr is a special
8122 case where small numbers are errno values, not library numbers.
8125 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8126 converts DSA parameters into DH parameters. (When creating parameters,
8127 DSA_generate_parameters is used.)
8130 *) Include 'length' (recommended exponent length) in C code generated
8131 by 'openssl dhparam -C'.
8134 *) The second argument to set_label in perlasm was already being used
8135 so couldn't be used as a "file scope" flag. Moved to third argument
8139 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8140 instead of RAND_bytes for encryption IVs and salts.
8143 *) Include RAND_status() into RAND_METHOD instead of implementing
8144 it only for md_rand.c Otherwise replacing the PRNG by calling
8145 RAND_set_rand_method would be impossible.
8148 *) Don't let DSA_generate_key() enter an infinite loop if the random
8149 number generation fails.
8152 *) New 'rand' application for creating pseudo-random output.
8155 *) Added configuration support for Linux/IA64
8156 [Rolf Haberrecker <rolf@suse.de>]
8158 *) Assembler module support for Mingw32.
8161 *) Shared library support for HPUX (in shlib/).
8162 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8164 *) Shared library support for Solaris gcc.
8165 [Lutz Behnke <behnke@trustcenter.de>]
8167 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8169 *) PKCS7_encrypt() was adding text MIME headers twice because they
8170 were added manually and by SMIME_crlf_copy().
8173 *) In bntest.c don't call BN_rand with zero bits argument.
8174 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8176 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8177 case was implemented. This caused BN_div_recp() to fail occasionally.
8180 *) Add an optional second argument to the set_label() in the perl
8181 assembly language builder. If this argument exists and is set
8182 to 1 it signals that the assembler should use a symbol whose
8183 scope is the entire file, not just the current function. This
8184 is needed with MASM which uses the format label:: for this scope.
8185 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8187 *) Change the ASN1 types so they are typedefs by default. Before
8188 almost all types were #define'd to ASN1_STRING which was causing
8189 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8193 *) Change names of new functions to the new get1/get0 naming
8194 convention: After 'get1', the caller owns a reference count
8195 and has to call ..._free; 'get0' returns a pointer to some
8196 data structure without incrementing reference counters.
8197 (Some of the existing 'get' functions increment a reference
8198 counter, some don't.)
8199 Similarly, 'set1' and 'add1' functions increase reference
8200 counters or duplicate objects.
8203 *) Allow for the possibility of temp RSA key generation failure:
8204 the code used to assume it always worked and crashed on failure.
8207 *) Fix potential buffer overrun problem in BIO_printf().
8208 [Ulf Möller, using public domain code by Patrick Powell; problem
8209 pointed out by David Sacerdote <das33@cornell.edu>]
8211 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8212 RAND_egd() and RAND_status(). In the command line application,
8213 the EGD socket can be specified like a seed file using RANDFILE
8217 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8218 Some CAs (e.g. Verisign) distribute certificates in this form.
8221 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8222 list to exclude them. This means that no special compilation option
8223 is needed to use anonymous DH: it just needs to be included in the
8227 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8228 EVP_MD_type. The old functionality is available in a new macro called
8229 EVP_MD_md(). Change code that uses it and update docs.
8232 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8233 where the 'void *' argument is replaced by a function pointer argument.
8234 Previously 'void *' was abused to point to functions, which works on
8235 many platforms, but is not correct. As these functions are usually
8236 called by macros defined in OpenSSL header files, most source code
8237 should work without changes.
8240 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8241 sections with information on -D... compiler switches used for
8242 compiling the library so that applications can see them. To enable
8243 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8244 must be defined. E.g.,
8245 #define OPENSSL_ALGORITHM_DEFINES
8246 #include <openssl/opensslconf.h>
8247 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8248 [Richard Levitte, Ulf and Bodo Möller]
8250 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8254 *) Change the 'other' type in certificate aux info to a STACK_OF
8255 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8256 the required ASN1 format: arbitrary types determined by an OID.
8259 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8260 argument to 'req'. This is not because the function is newer or
8261 better than others it just uses the work 'NEW' in the certificate
8262 request header lines. Some software needs this.
8265 *) Reorganise password command line arguments: now passwords can be
8266 obtained from various sources. Delete the PEM_cb function and make
8267 it the default behaviour: i.e. if the callback is NULL and the
8268 usrdata argument is not NULL interpret it as a null terminated pass
8269 phrase. If usrdata and the callback are NULL then the pass phrase
8270 is prompted for as usual.
8273 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8274 the support is automatically enabled. The resulting binaries will
8275 autodetect the card and use it if present.
8276 [Ben Laurie and Compaq Inc.]
8278 *) Work around for Netscape hang bug. This sends certificate request
8279 and server done in one record. Since this is perfectly legal in the
8280 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8281 the bugs/SSLv3 entry for more info.
8284 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8287 *) Add -rand argument to smime and pkcs12 applications and read/write
8291 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8294 *) Add command line password options to the remaining applications.
8297 *) Bug fix for BN_div_recp() for numerators with an even number of
8301 *) More tests in bntest.c, and changed test_bn output.
8304 *) ./config recognizes MacOS X now.
8307 *) Bug fix for BN_div() when the first words of num and divsor are
8308 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8311 *) Add support for various broken PKCS#8 formats, and command line
8312 options to produce them.
8315 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8316 get temporary BIGNUMs from a BN_CTX.
8319 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8323 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8324 include a #define from the old name to the new. The original intent
8325 was that statically linked binaries could for example just call
8326 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8327 link with digests. This never worked becayse SSLeay_add_all_digests()
8328 and SSLeay_add_all_ciphers() were in the same source file so calling
8329 one would link with the other. They are now in separate source files.
8332 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8335 *) Use a less unusual form of the Miller-Rabin primality test (it used
8336 a binary algorithm for exponentiation integrated into the Miller-Rabin
8337 loop, our standard modexp algorithms are faster).
8340 *) Support for the EBCDIC character set completed.
8341 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8343 *) Source code cleanups: use const where appropriate, eliminate casts,
8344 use void * instead of char * in lhash.
8347 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8348 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8349 this the server could overwrite ephemeral keys that the client
8353 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8354 using 50 iterations of the Rabin-Miller test.
8356 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8357 iterations of the Rabin-Miller test as required by the appendix
8358 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8359 As BN_is_prime_fasttest includes trial division, DSA parameter
8360 generation becomes much faster.
8362 This implies a change for the callback functions in DSA_is_prime
8363 and DSA_generate_parameters: The callback function is called once
8364 for each positive witness in the Rabin-Miller test, not just
8365 occasionally in the inner loop; and the parameters to the
8366 callback function now provide an iteration count for the outer
8367 loop rather than for the current invocation of the inner loop.
8368 DSA_generate_parameters additionally can call the callback
8369 function with an 'iteration count' of -1, meaning that a
8370 candidate has passed the trial division test (when q is generated
8371 from an application-provided seed, trial division is skipped).
8374 *) New function BN_is_prime_fasttest that optionally does trial
8375 division before starting the Rabin-Miller test and has
8376 an additional BN_CTX * argument (whereas BN_is_prime always
8377 has to allocate at least one BN_CTX).
8378 'callback(1, -1, cb_arg)' is called when a number has passed the
8379 trial division stage.
8382 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8386 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8389 *) New function BN_pseudo_rand().
8392 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8393 bignum version of BN_from_montgomery() with the working code from
8394 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8398 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8399 made it impossible to use the same SSL_SESSION data structure in
8400 SSL2 clients in multiple threads.
8403 *) The return value of RAND_load_file() no longer counts bytes obtained
8404 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8405 to seed the PRNG (previously an explicit byte count was required).
8406 [Ulf Möller, Bodo Möller]
8408 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8409 used (char *) instead of (void *) and had casts all over the place.
8412 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8415 *) Retain source code compatibility for BN_prime_checks macro:
8416 BN_is_prime(..., BN_prime_checks, ...) now uses
8417 BN_prime_checks_for_size to determine the appropriate number of
8418 Rabin-Miller iterations.
8421 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8422 DH_CHECK_P_NOT_SAFE_PRIME.
8423 (Check if this is true? OpenPGP calls them "strong".)
8426 *) Merge the functionality of "dh" and "gendh" programs into a new program
8427 "dhparam". The old programs are retained for now but will handle DH keys
8428 (instead of parameters) in future.
8431 *) Make the ciphers, s_server and s_client programs check the return values
8432 when a new cipher list is set.
8435 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8436 ciphers. Before when the 56bit ciphers were enabled the sorting was
8439 The syntax for the cipher sorting has been extended to support sorting by
8440 cipher-strength (using the strength_bits hard coded in the tables).
8441 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8443 Fix a bug in the cipher-command parser: when supplying a cipher command
8444 string with an "undefined" symbol (neither command nor alphanumeric
8445 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8446 an error is flagged.
8448 Due to the strength-sorting extension, the code of the
8449 ssl_create_cipher_list() function was completely rearranged. I hope that
8450 the readability was also increased :-)
8451 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8453 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8454 for the first serial number and places 2 in the serial number file. This
8455 avoids problems when the root CA is created with serial number zero and
8456 the first user certificate has the same issuer name and serial number
8460 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8461 the new code. Add documentation for this stuff.
8464 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8465 X509_*() to X509at_*() on the grounds that they don't handle X509
8466 structures and behave in an analagous way to the X509v3 functions:
8467 they shouldn't be called directly but wrapper functions should be used
8470 So we also now have some wrapper functions that call the X509at functions
8471 when passed certificate requests. (TO DO: similar things can be done with
8472 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8473 things. Some of these need some d2i or i2d and print functionality
8474 because they handle more complex structures.)
8477 *) Add missing #ifndefs that caused missing symbols when building libssl
8478 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8479 NO_RSA in ssl/s2*.c.
8480 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8482 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8483 has a return value which indicates the quality of the random data
8484 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8485 error queue. New function RAND_pseudo_bytes() generates output that is
8486 guaranteed to be unique but not unpredictable. RAND_add is like
8487 RAND_seed, but takes an extra argument for an entropy estimate
8488 (RAND_seed always assumes full entropy).
8491 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8492 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8493 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8494 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8495 false-positive rate of at most 2^-80 for random input.
8498 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8501 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8502 in the 0.9.5 release), this returns the chain
8503 from an X509_CTX structure with a dup of the stack and all
8504 the X509 reference counts upped: so the stack will exist
8505 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8508 Also make SSL_SESSION_print() print out the verify return
8512 *) Add manpage for the pkcs12 command. Also change the default
8513 behaviour so MAC iteration counts are used unless the new
8514 -nomaciter option is used. This improves file security and
8515 only older versions of MSIE (4.0 for example) need it.
8518 *) Honor the no-xxx Configure options when creating .DEF files.
8521 *) Add PKCS#10 attributes to field table: challengePassword,
8522 unstructuredName and unstructuredAddress. These are taken from
8523 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8524 international characters are used.
8526 More changes to X509_ATTRIBUTE code: allow the setting of types
8527 based on strings. Remove the 'loc' parameter when adding
8528 attributes because these will be a SET OF encoding which is sorted
8532 *) Initial changes to the 'req' utility to allow request generation
8533 automation. This will allow an application to just generate a template
8534 file containing all the field values and have req construct the
8537 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8538 used all over the place including certificate requests and PKCS#7
8539 structures. They are currently handled manually where necessary with
8540 some primitive wrappers for PKCS#7. The new functions behave in a
8541 manner analogous to the X509 extension functions: they allow
8542 attributes to be looked up by NID and added.
8544 Later something similar to the X509V3 code would be desirable to
8545 automatically handle the encoding, decoding and printing of the
8546 more complex types. The string types like challengePassword can
8547 be handled by the string table functions.
8549 Also modified the multi byte string table handling. Now there is
8550 a 'global mask' which masks out certain types. The table itself
8551 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8552 is useful when for example there is only one permissible type
8553 (as in countryName) and using the mask might result in no valid
8557 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8558 SSL_get_peer_finished to allow applications to obtain the latest
8559 Finished messages sent to the peer or expected from the peer,
8560 respectively. (SSL_get_peer_finished is usually the Finished message
8561 actually received from the peer, otherwise the protocol will be aborted.)
8563 As the Finished message are message digests of the complete handshake
8564 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8565 be used for external authentication procedures when the authentication
8566 provided by SSL/TLS is not desired or is not enough.
8569 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8570 the host supports BWX extension and if Compaq C is present on the
8571 $PATH. Just exploiting of the BWX extension results in 20-30%
8572 performance kick for some algorithms, e.g. DES and RC4 to mention
8573 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8577 *) Add support for MS "fast SGC". This is arguably a violation of the
8578 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8579 weak crypto and after checking the certificate is SGC a second one
8580 with strong crypto. MS SGC stops the first handshake after receiving
8581 the server certificate message and sends a second client hello. Since
8582 a server will typically do all the time consuming operations before
8583 expecting any further messages from the client (server key exchange
8584 is the most expensive) there is little difference between the two.
8586 To get OpenSSL to support MS SGC we have to permit a second client
8587 hello message after we have sent server done. In addition we have to
8588 reset the MAC if we do get this second client hello.
8591 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8592 if a DER encoded private key is RSA or DSA traditional format. Changed
8593 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8594 format DER encoded private key. Newer code should use PKCS#8 format which
8595 has the key type encoded in the ASN1 structure. Added DER private key
8596 support to pkcs8 application.
8599 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8600 ciphersuites has been selected (as required by the SSL 3/TLS 1
8601 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8602 is set, we interpret this as a request to violate the specification
8603 (the worst that can happen is a handshake failure, and 'correct'
8604 behaviour would result in a handshake failure anyway).
8607 *) In SSL_CTX_add_session, take into account that there might be multiple
8608 SSL_SESSION structures with the same session ID (e.g. when two threads
8609 concurrently obtain them from an external cache).
8610 The internal cache can handle only one SSL_SESSION with a given ID,
8611 so if there's a conflict, we now throw out the old one to achieve
8615 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8616 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8617 some routines that use cipher OIDs: some ciphers do not have OIDs
8618 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8622 *) Simplify the trust setting structure and code. Now we just have
8623 two sequences of OIDs for trusted and rejected settings. These will
8624 typically have values the same as the extended key usage extension
8625 and any application specific purposes.
8627 The trust checking code now has a default behaviour: it will just
8628 check for an object with the same NID as the passed id. Functions can
8629 be provided to override either the default behaviour or the behaviour
8630 for a given id. SSL client, server and email already have functions
8631 in place for compatibility: they check the NID and also return "trusted"
8632 if the certificate is self signed.
8635 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8636 traditional format into an EVP_PKEY structure.
8639 *) Add a password callback function PEM_cb() which either prompts for
8640 a password if usr_data is NULL or otherwise assumes it is a null
8641 terminated password. Allow passwords to be passed on command line
8642 environment or config files in a few more utilities.
8645 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8646 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8647 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8648 Update documentation.
8651 *) Support for ASN1 "NULL" type. This could be handled before by using
8652 ASN1_TYPE but there wasn't any function that would try to read a NULL
8653 and produce an error if it couldn't. For compatibility we also have
8654 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8655 don't allocate anything because they don't need to.
8658 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8660 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8662 *) Rebuild of the memory allocation routines used by OpenSSL code and
8663 possibly others as well. The purpose is to make an interface that
8664 provide hooks so anyone can build a separate set of allocation and
8665 deallocation routines to be used by OpenSSL, for example memory
8666 pool implementations, or something else, which was previously hard
8667 since Malloc(), Realloc() and Free() were defined as macros having
8668 the values malloc, realloc and free, respectively (except for Win32
8669 compilations). The same is provided for memory debugging code.
8670 OpenSSL already comes with functionality to find memory leaks, but
8671 this gives people a chance to debug other memory problems.
8673 With these changes, a new set of functions and macros have appeared:
8675 CRYPTO_set_mem_debug_functions() [F]
8676 CRYPTO_get_mem_debug_functions() [F]
8677 CRYPTO_dbg_set_options() [F]
8678 CRYPTO_dbg_get_options() [F]
8679 CRYPTO_malloc_debug_init() [M]
8681 The memory debug functions are NULL by default, unless the library
8682 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8683 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8684 gives the standard debugging functions that come with OpenSSL) or
8685 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8686 provided by the library user) must be used. When the standard
8687 debugging functions are used, CRYPTO_dbg_set_options can be used to
8688 request additional information:
8689 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8690 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8692 Also, things like CRYPTO_set_mem_functions will always give the
8693 expected result (the new set of functions is used for allocation
8694 and deallocation) at all times, regardless of platform and compiler
8697 To finish it up, some functions that were never use in any other
8698 way than through macros have a new API and new semantic:
8701 CRYPTO_dbg_realloc()
8704 All macros of value have retained their old syntax.
8705 [Richard Levitte and Bodo Moeller]
8707 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8708 ordering of SMIMECapabilities wasn't in "strength order" and there
8709 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8713 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8714 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8715 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8717 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8718 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8719 functionality to handle multipart/signed properly) and a utility
8720 called 'smime' to call all this stuff. This is based on code I
8721 originally wrote for Celo who have kindly allowed it to be
8722 included in OpenSSL.
8725 *) Add variants des_set_key_checked and des_set_key_unchecked of
8726 des_set_key (aka des_key_sched). Global variable des_check_key
8727 decides which of these is called by des_set_key; this way
8728 des_check_key behaves as it always did, but applications and
8729 the library itself, which was buggy for des_check_key == 1,
8730 have a cleaner way to pick the version they need.
8733 *) New function PKCS12_newpass() which changes the password of a
8737 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8738 dynamic mix. In both cases the ids can be used as an index into the
8739 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8740 functions so they accept a list of the field values and the
8741 application doesn't need to directly manipulate the X509_TRUST
8745 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8749 *) Modify the way the V3 extension code looks up extensions. This now
8750 works in a similar way to the object code: we have some "standard"
8751 extensions in a static table which is searched with OBJ_bsearch()
8752 and the application can add dynamic ones if needed. The file
8753 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8754 updated whenever a new extension is added to the core code and kept
8755 in ext_nid order. There is a simple program 'tabtest.c' which checks
8756 this. New extensions are not added too often so this file can readily
8757 be maintained manually.
8759 There are two big advantages in doing things this way. The extensions
8760 can be looked up immediately and no longer need to be "added" using
8761 X509V3_add_standard_extensions(): this function now does nothing.
8762 [Side note: I get *lots* of email saying the extension code doesn't
8763 work because people forget to call this function]
8764 Also no dynamic allocation is done unless new extensions are added:
8765 so if we don't add custom extensions there is no need to call
8766 X509V3_EXT_cleanup().
8769 *) Modify enc utility's salting as follows: make salting the default. Add a
8770 magic header, so unsalted files fail gracefully instead of just decrypting
8771 to garbage. This is because not salting is a big security hole, so people
8772 should be discouraged from doing it.
8775 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8776 digest to be passed on the command line but it only used this
8777 parameter when signing a certificate. Modified so all relevant
8778 operations are affected by the digest parameter including the
8779 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8780 DSA key was used because it didn't fix the digest.
8783 *) Initial certificate chain verify code. Currently tests the untrusted
8784 certificates for consistency with the verify purpose (which is set
8785 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8787 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8788 this is because it will reject chains with invalid extensions whereas
8789 every previous version of OpenSSL and SSLeay made no checks at all.
8791 Trust code: checks the root CA for the relevant trust settings. Trust
8792 settings have an initial value consistent with the verify purpose: e.g.
8793 if the verify purpose is for SSL client use it expects the CA to be
8794 trusted for SSL client use. However the default value can be changed to
8795 permit custom trust settings: one example of this would be to only trust
8796 certificates from a specific "secure" set of CAs.
8798 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8799 which should be used for version portability: especially since the
8800 verify structure is likely to change more often now.
8802 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8803 to set them. If not set then assume SSL clients will verify SSL servers
8806 Two new options to the verify program: -untrusted allows a set of
8807 untrusted certificates to be passed in and -purpose which sets the
8808 intended purpose of the certificate. If a purpose is set then the
8809 new chain verify code is used to check extension consistency.
8812 *) Support for the authority information access extension.
8815 *) Modify RSA and DSA PEM read routines to transparently handle
8816 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8817 public keys in a format compatible with certificate
8818 SubjectPublicKeyInfo structures. Unfortunately there were already
8819 functions called *_PublicKey_* which used various odd formats so
8820 these are retained for compatibility: however the DSA variants were
8821 never in a public release so they have been deleted. Changed dsa/rsa
8822 utilities to handle the new format: note no releases ever handled public
8823 keys so we should be OK.
8825 The primary motivation for this change is to avoid the same fiasco
8826 that dogs private keys: there are several incompatible private key
8827 formats some of which are standard and some OpenSSL specific and
8828 require various evil hacks to allow partial transparent handling and
8829 even then it doesn't work with DER formats. Given the option anything
8830 other than PKCS#8 should be dumped: but the other formats have to
8831 stay in the name of compatibility.
8833 With public keys and the benefit of hindsight one standard format
8834 is used which works with EVP_PKEY, RSA or DSA structures: though
8835 it clearly returns an error if you try to read the wrong kind of key.
8837 Added a -pubkey option to the 'x509' utility to output the public key.
8838 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8839 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8840 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8841 that do the same as the EVP_PKEY_assign_*() except they up the
8842 reference count of the added key (they don't "swallow" the
8846 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8847 CRLs would fail if the file contained no certificates or no CRLs:
8848 added a new function to read in both types and return the number
8849 read: this means that if none are read it will be an error. The
8850 DER versions of the certificate and CRL reader would always fail
8851 because it isn't possible to mix certificates and CRLs in DER format
8852 without choking one or the other routine. Changed this to just read
8853 a certificate: this is the best we can do. Also modified the code
8854 in apps/verify.c to take notice of return codes: it was previously
8855 attempting to read in certificates from NULL pointers and ignoring
8856 any errors: this is one reason why the cert and CRL reader seemed
8857 to work. It doesn't check return codes from the default certificate
8858 routines: these may well fail if the certificates aren't installed.
8861 *) Code to support otherName option in GeneralName.
8864 *) First update to verify code. Change the verify utility
8865 so it warns if it is passed a self signed certificate:
8866 for consistency with the normal behaviour. X509_verify
8867 has been modified to it will now verify a self signed
8868 certificate if *exactly* the same certificate appears
8869 in the store: it was previously impossible to trust a
8870 single self signed certificate. This means that:
8871 openssl verify ss.pem
8872 now gives a warning about a self signed certificate but
8873 openssl verify -CAfile ss.pem ss.pem
8877 *) For servers, store verify_result in SSL_SESSION data structure
8878 (and add it to external session representation).
8879 This is needed when client certificate verifications fails,
8880 but an application-provided verification callback (set by
8881 SSL_CTX_set_cert_verify_callback) allows accepting the session
8882 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8883 but returns 1): When the session is reused, we have to set
8884 ssl->verify_result to the appropriate error code to avoid
8886 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8888 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8889 case in PKCS7_dataInit() where the signed PKCS7 structure
8890 didn't contain any existing data because it was being created.
8891 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8893 *) Add a salt to the key derivation routines in enc.c. This
8894 forms the first 8 bytes of the encrypted file. Also add a
8895 -S option to allow a salt to be input on the command line.
8898 *) New function X509_cmp(). Oddly enough there wasn't a function
8899 to compare two certificates. We do this by working out the SHA1
8900 hash and comparing that. X509_cmp() will be needed by the trust
8904 *) SSL_get1_session() is like SSL_get_session(), but increments
8905 the reference count in the SSL_SESSION returned.
8906 [Geoff Thorpe <geoff@eu.c2.net>]
8908 *) Fix for 'req': it was adding a null to request attributes.
8909 Also change the X509_LOOKUP and X509_INFO code to handle
8910 certificate auxiliary information.
8913 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8917 *) Add the possibility to add extra information to the memory leak
8918 detecting output, to form tracebacks, showing from where each
8919 allocation was originated: CRYPTO_push_info("constant string") adds
8920 the string plus current file name and line number to a per-thread
8921 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8922 is like calling CYRPTO_pop_info() until the stack is empty.
8923 Also updated memory leak detection code to be multi-thread-safe.
8926 *) Add options -text and -noout to pkcs7 utility and delete the
8927 encryption options which never did anything. Update docs.
8930 *) Add options to some of the utilities to allow the pass phrase
8931 to be included on either the command line (not recommended on
8932 OSes like Unix) or read from the environment. Update the
8933 manpages and fix a few bugs.
8936 *) Add a few manpages for some of the openssl commands.
8939 *) Fix the -revoke option in ca. It was freeing up memory twice,
8940 leaking and not finding already revoked certificates.
8943 *) Extensive changes to support certificate auxiliary information.
8944 This involves the use of X509_CERT_AUX structure and X509_AUX
8945 functions. An X509_AUX function such as PEM_read_X509_AUX()
8946 can still read in a certificate file in the usual way but it
8947 will also read in any additional "auxiliary information". By
8948 doing things this way a fair degree of compatibility can be
8949 retained: existing certificates can have this information added
8950 using the new 'x509' options.
8952 Current auxiliary information includes an "alias" and some trust
8953 settings. The trust settings will ultimately be used in enhanced
8954 certificate chain verification routines: currently a certificate
8955 can only be trusted if it is self signed and then it is trusted
8959 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8960 The problem was that one of the replacement routines had not been working
8961 since SSLeay releases. For now the offending routine has been replaced
8962 with non-optimised assembler. Even so, this now gives around 95%
8963 performance improvement for 1024 bit RSA signs.
8966 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8967 handling. Most clients have the effective key size in bits equal to
8968 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8969 A few however don't do this and instead use the size of the decrypted key
8970 to determine the RC2 key length and the AlgorithmIdentifier to determine
8971 the effective key length. In this case the effective key length can still
8972 be 40 bits but the key length can be 168 bits for example. This is fixed
8973 by manually forcing an RC2 key into the EVP_PKEY structure because the
8974 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8975 the key length and effective key length are equal.
8978 *) Add a bunch of functions that should simplify the creation of
8979 X509_NAME structures. Now you should be able to do:
8980 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8981 and have it automatically work out the correct field type and fill in
8982 the structures. The more adventurous can try:
8983 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8984 and it will (hopefully) work out the correct multibyte encoding.
8987 *) Change the 'req' utility to use the new field handling and multibyte
8988 copy routines. Before the DN field creation was handled in an ad hoc
8989 way in req, ca, and x509 which was rather broken and didn't support
8990 BMPStrings or UTF8Strings. Since some software doesn't implement
8991 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8992 using the dirstring_type option. See the new comment in the default
8993 openssl.cnf for more info.
8996 *) Make crypto/rand/md_rand.c more robust:
8997 - Assure unique random numbers after fork().
8998 - Make sure that concurrent threads access the global counter and
8999 md serializably so that we never lose entropy in them
9000 or use exactly the same state in multiple threads.
9001 Access to the large state is not always serializable because
9002 the additional locking could be a performance killer, and
9003 md should be large enough anyway.
9006 *) New file apps/app_rand.c with commonly needed functionality
9007 for handling the random seed file.
9009 Use the random seed file in some applications that previously did not:
9011 dsaparam -genkey (which also ignored its '-rand' option),
9014 x509 (when signing).
9015 Except on systems with /dev/urandom, it is crucial to have a random
9016 seed file at least for key creation, DSA signing, and for DH exchanges;
9017 for RSA signatures we could do without one.
9019 gendh and gendsa (unlike genrsa) used to read only the first byte
9020 of each file listed in the '-rand' option. The function as previously
9021 found in genrsa is now in app_rand.c and is used by all programs
9022 that support '-rand'.
9025 *) In RAND_write_file, use mode 0600 for creating files;
9026 don't just chmod when it may be too late.
9029 *) Report an error from X509_STORE_load_locations
9030 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9033 *) New function ASN1_mbstring_copy() this copies a string in either
9034 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9035 into an ASN1_STRING type. A mask of permissible types is passed
9036 and it chooses the "minimal" type to use or an error if not type
9040 *) Add function equivalents to the various macros in asn1.h. The old
9041 macros are retained with an M_ prefix. Code inside the library can
9042 use the M_ macros. External code (including the openssl utility)
9043 should *NOT* in order to be "shared library friendly".
9046 *) Add various functions that can check a certificate's extensions
9047 to see if it usable for various purposes such as SSL client,
9048 server or S/MIME and CAs of these types. This is currently
9049 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9050 verification. Also added a -purpose flag to x509 utility to
9051 print out all the purposes.
9054 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9058 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9059 for, obtain and decode and extension and obtain its critical flag.
9060 This allows all the necessary extension code to be handled in a
9061 single function call.
9064 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9065 platforms. See crypto/rc4/rc4_enc.c for further details.
9068 *) New -noout option to asn1parse. This causes no output to be produced
9069 its main use is when combined with -strparse and -out to extract data
9070 from a file (which may not be in ASN.1 format).
9073 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9074 when producing the local key id.
9075 [Richard Levitte <levitte@stacken.kth.se>]
9077 *) New option -dhparam in s_server. This allows a DH parameter file to be
9078 stated explicitly. If it is not stated then it tries the first server
9079 certificate file. The previous behaviour hard coded the filename
9083 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9084 a public key to be input or output. For example:
9085 openssl rsa -in key.pem -pubout -out pubkey.pem
9086 Also added necessary DSA public key functions to handle this.
9089 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9090 in the message. This was handled by allowing
9091 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9092 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9094 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9095 to the end of the strings whereas this didn't. This would cause problems
9096 if strings read with d2i_ASN1_bytes() were later modified.
9097 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9099 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9100 data and it contains EOF it will end up returning an error. This is
9101 caused by input 46 bytes long. The cause is due to the way base64
9102 BIOs find the start of base64 encoded data. They do this by trying a
9103 trial decode on each line until they find one that works. When they
9104 do a flag is set and it starts again knowing it can pass all the
9105 data directly through the decoder. Unfortunately it doesn't reset
9106 the context it uses. This means that if EOF is reached an attempt
9107 is made to pass two EOFs through the context and this causes the
9108 resulting error. This can also cause other problems as well. As is
9109 usual with these problems it takes *ages* to find and the fix is
9110 trivial: move one line.
9111 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9113 *) Ugly workaround to get s_client and s_server working under Windows. The
9114 old code wouldn't work because it needed to select() on sockets and the
9115 tty (for keypresses and to see if data could be written). Win32 only
9116 supports select() on sockets so we select() with a 1s timeout on the
9117 sockets and then see if any characters are waiting to be read, if none
9118 are present then we retry, we also assume we can always write data to
9119 the tty. This isn't nice because the code then blocks until we've
9120 received a complete line of data and it is effectively polling the
9121 keyboard at 1s intervals: however it's quite a bit better than not
9122 working at all :-) A dedicated Windows application might handle this
9123 with an event loop for example.
9126 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9127 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9128 will be called when RSA_sign() and RSA_verify() are used. This is useful
9129 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9130 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9131 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9132 This necessitated the support of an extra signature type NID_md5_sha1
9133 for SSL signatures and modifications to the SSL library to use it instead
9134 of calling RSA_public_decrypt() and RSA_private_encrypt().
9137 *) Add new -verify -CAfile and -CApath options to the crl program, these
9138 will lookup a CRL issuers certificate and verify the signature in a
9139 similar way to the verify program. Tidy up the crl program so it
9140 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9141 less strict. It will now permit CRL extensions even if it is not
9142 a V2 CRL: this will allow it to tolerate some broken CRLs.
9145 *) Initialize all non-automatic variables each time one of the openssl
9146 sub-programs is started (this is necessary as they may be started
9147 multiple times from the "OpenSSL>" prompt).
9148 [Lennart Bang, Bodo Moeller]
9150 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9151 removing all other RSA functionality (this is what NO_RSA does). This
9152 is so (for example) those in the US can disable those operations covered
9153 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9157 *) Non-copying interface to BIO pairs.
9158 (still largely untested)
9161 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9162 ASCII string. This was handled independently in various places before.
9165 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9166 UTF8 strings a character at a time.
9169 *) Use client_version from client hello to select the protocol
9170 (s23_srvr.c) and for RSA client key exchange verification
9171 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9174 *) Add various utility functions to handle SPKACs, these were previously
9175 handled by poking round in the structure internals. Added new function
9176 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9177 print, verify and generate SPKACs. Based on an original idea from
9178 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9181 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9184 *) Allow the config file extension section to be overwritten on the
9185 command line. Based on an original idea from Massimiliano Pala
9186 <madwolf@comune.modena.it>. The new option is called -extensions
9187 and can be applied to ca, req and x509. Also -reqexts to override
9188 the request extensions in req and -crlexts to override the crl extensions
9192 *) Add new feature to the SPKAC handling in ca. Now you can include
9193 the same field multiple times by preceding it by "XXXX." for example:
9196 this is the same syntax as used in the req config file.
9199 *) Allow certificate extensions to be added to certificate requests. These
9200 are specified in a 'req_extensions' option of the req section of the
9201 config file. They can be printed out with the -text option to req but
9202 are otherwise ignored at present.
9205 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9206 data read consists of only the final block it would not decrypted because
9207 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9208 A misplaced 'break' also meant the decrypted final block might not be
9209 copied until the next read.
9212 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9213 a few extra parameters to the DH structure: these will be useful if
9214 for example we want the value of 'q' or implement X9.42 DH.
9217 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9218 provides hooks that allow the default DSA functions or functions on a
9219 "per key" basis to be replaced. This allows hardware acceleration and
9220 hardware key storage to be handled without major modification to the
9221 library. Also added low level modexp hooks and CRYPTO_EX structure and
9222 associated functions.
9225 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9226 as "read only": it can't be written to and the buffer it points to will
9227 not be freed. Reading from a read only BIO is much more efficient than
9228 a normal memory BIO. This was added because there are several times when
9229 an area of memory needs to be read from a BIO. The previous method was
9230 to create a memory BIO and write the data to it, this results in two
9231 copies of the data and an O(n^2) reading algorithm. There is a new
9232 function BIO_new_mem_buf() which creates a read only memory BIO from
9233 an area of memory. Also modified the PKCS#7 routines to use read only
9237 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9238 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9239 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9240 but a retry condition occured while trying to read the rest.
9243 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9244 NID_pkcs7_encrypted by default: this was wrong since this should almost
9245 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9246 the encrypted data type: this is a more sensible place to put it and it
9247 allows the PKCS#12 code to be tidied up that duplicated this
9251 *) Changed obj_dat.pl script so it takes its input and output files on
9252 the command line. This should avoid shell escape redirection problems
9256 *) Initial support for certificate extension requests, these are included
9257 in things like Xenroll certificate requests. Included functions to allow
9258 extensions to be obtained and added.
9261 *) -crlf option to s_client and s_server for sending newlines as
9262 CRLF (as required by many protocols).
9265 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9267 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9268 [Ralf S. Engelschall]
9270 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9271 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9273 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9277 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9278 DH parameters/keys (q is lost during that conversion, but the resulting
9279 DH parameters contain its length).
9281 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9282 much faster than DH_generate_parameters (which creates parameters
9283 where p = 2*q + 1), and also the smaller q makes DH computations
9284 much more efficient (160-bit exponentiation instead of 1024-bit
9285 exponentiation); so this provides a convenient way to support DHE
9286 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9287 utter importance to use
9288 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9290 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9291 when such DH parameters are used, because otherwise small subgroup
9292 attacks may become possible!
9295 *) Avoid memory leak in i2d_DHparams.
9298 *) Allow the -k option to be used more than once in the enc program:
9299 this allows the same encrypted message to be read by multiple recipients.
9302 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9303 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9304 it will always use the numerical form of the OID, even if it has a short
9308 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9309 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9310 otherwise bn_mod_exp was called. In the case of hardware keys for example
9311 no private key components need be present and it might store extra data
9312 in the RSA structure, which cannot be accessed from bn_mod_exp.
9313 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9314 private key operations.
9317 *) Added support for SPARC Linux.
9320 *) pem_password_cb function type incompatibly changed from
9321 typedef int pem_password_cb(char *buf, int size, int rwflag);
9323 ....(char *buf, int size, int rwflag, void *userdata);
9324 so that applications can pass data to their callbacks:
9325 The PEM[_ASN1]_{read,write}... functions and macros now take an
9326 additional void * argument, which is just handed through whenever
9327 the password callback is called.
9328 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9330 New function SSL_CTX_set_default_passwd_cb_userdata.
9332 Compatibility note: As many C implementations push function arguments
9333 onto the stack in reverse order, the new library version is likely to
9334 interoperate with programs that have been compiled with the old
9335 pem_password_cb definition (PEM_whatever takes some data that
9336 happens to be on the stack as its last argument, and the callback
9337 just ignores this garbage); but there is no guarantee whatsoever that
9340 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9341 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9342 problems not only on Windows, but also on some Unix platforms.
9343 To avoid problematic command lines, these definitions are now in an
9344 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9345 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9348 *) MIPS III/IV assembler module is reimplemented.
9351 *) More DES library cleanups: remove references to srand/rand and
9352 delete an unused file.
9355 *) Add support for the the free Netwide assembler (NASM) under Win32,
9356 since not many people have MASM (ml) and it can be hard to obtain.
9357 This is currently experimental but it seems to work OK and pass all
9358 the tests. Check out INSTALL.W32 for info.
9361 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9362 without temporary keys kept an extra copy of the server key,
9363 and connections with temporary keys did not free everything in case
9367 *) New function RSA_check_key and new openssl rsa option -check
9368 for verifying the consistency of RSA keys.
9369 [Ulf Moeller, Bodo Moeller]
9371 *) Various changes to make Win32 compile work:
9372 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9373 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9374 comparison" warnings.
9375 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9378 *) Add a debugging option to PKCS#5 v2 key generation function: when
9379 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9380 derived keys are printed to stderr.
9383 *) Copy the flags in ASN1_STRING_dup().
9384 [Roman E. Pavlov <pre@mo.msk.ru>]
9386 *) The x509 application mishandled signing requests containing DSA
9387 keys when the signing key was also DSA and the parameters didn't match.
9389 It was supposed to omit the parameters when they matched the signing key:
9390 the verifying software was then supposed to automatically use the CA's
9391 parameters if they were absent from the end user certificate.
9393 Omitting parameters is no longer recommended. The test was also
9394 the wrong way round! This was probably due to unusual behaviour in
9395 EVP_cmp_parameters() which returns 1 if the parameters match.
9396 This meant that parameters were omitted when they *didn't* match and
9397 the certificate was useless. Certificates signed with 'ca' didn't have
9399 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9401 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9402 The interface is as follows:
9403 Applications can use
9404 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9405 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9406 "off" is now the default.
9407 The library internally uses
9408 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9409 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9410 to disable memory-checking temporarily.
9412 Some inconsistent states that previously were possible (and were
9413 even the default) are now avoided.
9415 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9416 with each memory chunk allocated; this is occasionally more helpful
9417 than just having a counter.
9419 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9421 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9425 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9426 which largely parallels "options", but is for changing API behaviour,
9427 whereas "options" are about protocol behaviour.
9428 Initial "mode" flags are:
9430 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9431 a single record has been written.
9432 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9433 retries use the same buffer location.
9434 (But all of the contents must be
9438 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9441 *) Fix problems with no-hmac etc.
9442 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9444 *) New functions RSA_get_default_method(), RSA_set_method() and
9445 RSA_get_method(). These allows replacement of RSA_METHODs without having
9446 to mess around with the internals of an RSA structure.
9449 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9450 Also really enable memory leak checks in openssl.c and in some
9452 [Chad C. Mulligan, Bodo Moeller]
9454 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9455 up the length of negative integers. This has now been simplified to just
9456 store the length when it is first determined and use it later, rather
9457 than trying to keep track of where data is copied and updating it to
9459 [Steve Henson, reported by Brien Wheeler
9460 <bwheeler@authentica-security.com>]
9462 *) Add a new function PKCS7_signatureVerify. This allows the verification
9463 of a PKCS#7 signature but with the signing certificate passed to the
9464 function itself. This contrasts with PKCS7_dataVerify which assumes the
9465 certificate is present in the PKCS#7 structure. This isn't always the
9466 case: certificates can be omitted from a PKCS#7 structure and be
9467 distributed by "out of band" means (such as a certificate database).
9470 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9471 function prototypes in pem.h, also change util/mkdef.pl to add the
9472 necessary function names.
9475 *) mk1mf.pl (used by Windows builds) did not properly read the
9476 options set by Configure in the top level Makefile, and Configure
9477 was not even able to write more than one option correctly.
9478 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9481 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9482 file to be loaded from a BIO or FILE pointer. The BIO version will
9483 for example allow memory BIOs to contain config info.
9486 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9487 Whoever hopes to achieve shared-library compatibility across versions
9488 must use this, not the compile-time macro.
9489 (Exercise 0.9.4: Which is the minimum library version required by
9491 Note: All this applies only to multi-threaded programs, others don't
9495 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9496 through a BIO pair triggered the default case, i.e.
9497 SSLerr(...,SSL_R_UNKNOWN_STATE).
9500 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9501 can use the SSL library even if none of the specific BIOs is
9505 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9506 for the encoded length.
9507 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9509 *) Add initial documentation of the X509V3 functions.
9512 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9513 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9514 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9515 secure PKCS#8 private key format with a high iteration count.
9518 *) Fix determination of Perl interpreter: A perl or perl5
9519 _directory_ in $PATH was also accepted as the interpreter.
9520 [Ralf S. Engelschall]
9522 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9523 wrong with it but it was very old and did things like calling
9524 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9528 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9529 to use the new extension code.
9532 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9533 with macros. This should make it easier to change their form, add extra
9534 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9538 *) Add to configuration table a new entry that can specify an alternative
9539 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9540 according to Mark Crispin <MRC@Panda.COM>.
9544 *) DES CBC did not update the IV. Weird.
9547 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9548 Changing the behaviour of the former might break existing programs --
9549 where IV updating is needed, des_ncbc_encrypt can be used.
9552 *) When bntest is run from "make test" it drives bc to check its
9553 calculations, as well as internally checking them. If an internal check
9554 fails, it needs to cause bc to give a non-zero result or make test carries
9555 on without noticing the failure. Fixed.
9558 *) DES library cleanups.
9561 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9562 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9563 ciphers. NOTE: although the key derivation function has been verified
9564 against some published test vectors it has not been extensively tested
9565 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9569 *) Instead of "mkdir -p", which is not fully portable, use new
9570 Perl script "util/mkdir-p.pl".
9573 *) Rewrite the way password based encryption (PBE) is handled. It used to
9574 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9575 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9576 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9577 the 'parameter' field of the AlgorithmIdentifier is passed to the
9578 underlying key generation function so it must do its own ASN1 parsing.
9579 This has also changed the EVP_PBE_CipherInit() function which now has a
9580 'parameter' argument instead of literal salt and iteration count values
9581 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9584 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9585 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9586 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9587 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9588 value was just used as a "magic string" and not used directly its
9589 value doesn't matter.
9592 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9596 *) "linux-sparc64" configuration (ultrapenguin).
9597 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9598 "linux-sparc" configuration.
9599 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9601 *) config now generates no-xxx options for missing ciphers.
9604 *) Support the EBCDIC character set (work in progress).
9605 File ebcdic.c not yet included because it has a different license.
9606 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9608 *) Support BS2000/OSD-POSIX.
9609 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9611 *) Make callbacks for key generation use void * instead of char *.
9614 *) Make S/MIME samples compile (not yet tested).
9617 *) Additional typesafe stacks.
9620 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9624 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9626 *) New configuration variant "sco5-gcc".
9628 *) Updated some demos.
9629 [Sean O Riordain, Wade Scholine]
9631 *) Add missing BIO_free at exit of pkcs12 application.
9634 *) Fix memory leak in conf.c.
9637 *) Updates for Win32 to assembler version of MD5.
9640 *) Set #! path to perl in apps/der_chop to where we found it
9641 instead of using a fixed path.
9644 *) SHA library changes for irix64-mips4-cc.
9647 *) Improvements for VMS support.
9651 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9653 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9654 This also avoids the problems with SC4.2 and unpatched SC5.
9655 [Andy Polyakov <appro@fy.chalmers.se>]
9657 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9658 These are required because of the typesafe stack would otherwise break
9659 existing code. If old code used a structure member which used to be STACK
9660 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9661 sk_num or sk_value it would produce an error because the num, data members
9662 are not present in STACK_OF. Now it just produces a warning. sk_set
9663 replaces the old method of assigning a value to sk_value
9664 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9665 that does this will no longer work (and should use sk_set instead) but
9666 this could be regarded as a "questionable" behaviour anyway.
9669 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9670 correctly handle encrypted S/MIME data.
9673 *) Change type of various DES function arguments from des_cblock
9674 (which means, in function argument declarations, pointer to char)
9675 to des_cblock * (meaning pointer to array with 8 char elements),
9676 which allows the compiler to do more typechecking; it was like
9677 that back in SSLeay, but with lots of ugly casts.
9679 Introduce new type const_des_cblock.
9682 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9683 problems: find RecipientInfo structure that matches recipient certificate
9684 and initialise the ASN1 structures properly based on passed cipher.
9687 *) Belatedly make the BN tests actually check the results.
9690 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9691 to and from BNs: it was completely broken. New compilation option
9692 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9693 key elements as negative integers.
9696 *) Reorganize and speed up MD5.
9697 [Andy Polyakov <appro@fy.chalmers.se>]
9700 [Richard Levitte <richard@levitte.org>]
9702 *) New option -out to asn1parse to allow the parsed structure to be
9703 output to a file. This is most useful when combined with the -strparse
9704 option to examine the output of things like OCTET STRINGS.
9707 *) Make SSL library a little more fool-proof by not requiring any longer
9708 that SSL_set_{accept,connect}_state be called before
9709 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9710 in many applications because usually everything *appeared* to work as
9711 intended anyway -- now it really works as intended).
9714 *) Move openssl.cnf out of lib/.
9717 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9718 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9719 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9720 [Ralf S. Engelschall]
9722 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9723 handle PKCS#7 enveloped data properly.
9724 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9726 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9727 copying pointers. The cert_st handling is changed by this in
9728 various ways (and thus what used to be known as ctx->default_cert
9729 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9730 any longer when s->cert does not give us what we need).
9731 ssl_cert_instantiate becomes obsolete by this change.
9732 As soon as we've got the new code right (possibly it already is?),
9733 we have solved a couple of bugs of the earlier code where s->cert
9734 was used as if it could not have been shared with other SSL structures.
9736 Note that using the SSL API in certain dirty ways now will result
9737 in different behaviour than observed with earlier library versions:
9738 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9739 does not influence s as it used to.
9741 In order to clean up things more thoroughly, inside SSL_SESSION
9742 we don't use CERT any longer, but a new structure SESS_CERT
9743 that holds per-session data (if available); currently, this is
9744 the peer's certificate chain and, for clients, the server's certificate
9745 and temporary key. CERT holds only those values that can have
9746 meaningful defaults in an SSL_CTX.
9749 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9750 from the internal representation. Various PKCS#7 fixes: remove some
9751 evil casts and set the enc_dig_alg field properly based on the signing
9755 *) Allow PKCS#12 password to be set from the command line or the
9756 environment. Let 'ca' get its config file name from the environment
9757 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9761 *) Allow certificate policies extension to use an IA5STRING for the
9762 organization field. This is contrary to the PKIX definition but
9763 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9767 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9768 without disallowing inline assembler and the like for non-pedantic builds.
9771 *) Support Borland C++ builder.
9772 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9777 *) SHA-1 cleanups and performance enhancements.
9778 [Andy Polyakov <appro@fy.chalmers.se>]
9780 *) Sparc v8plus assembler for the bignum library.
9781 [Andy Polyakov <appro@fy.chalmers.se>]
9783 *) Accept any -xxx and +xxx compiler options in Configure.
9786 *) Update HPUX configuration.
9789 *) Add missing sk_<type>_unshift() function to safestack.h
9790 [Ralf S. Engelschall]
9792 *) New function SSL_CTX_use_certificate_chain_file that sets the
9793 "extra_cert"s in addition to the certificate. (This makes sense
9794 only for "PEM" format files, as chains as a whole are not
9798 *) Support verify_depth from the SSL API.
9799 x509_vfy.c had what can be considered an off-by-one-error:
9800 Its depth (which was not part of the external interface)
9801 was actually counting the number of certificates in a chain;
9802 now it really counts the depth.
9805 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9806 instead of X509err, which often resulted in confusing error
9807 messages since the error codes are not globally unique
9808 (e.g. an alleged error in ssl3_accept when a certificate
9809 didn't match the private key).
9811 *) New function SSL_CTX_set_session_id_context that allows to set a default
9812 value (so that you don't need SSL_set_session_id_context for each
9813 connection using the SSL_CTX).
9816 *) OAEP decoding bug fix.
9819 *) Support INSTALL_PREFIX for package builders, as proposed by
9823 *) New Configure options "threads" and "no-threads". For systems
9824 where the proper compiler options are known (currently Solaris
9825 and Linux), "threads" is the default.
9828 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9831 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9832 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9833 such as /usr/local/bin.
9836 *) "make linux-shared" to build shared libraries.
9837 [Niels Poppe <niels@netbox.org>]
9839 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9842 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9843 extension adding in x509 utility.
9846 *) Remove NOPROTO sections and error code comments.
9849 *) Partial rewrite of the DEF file generator to now parse the ANSI
9853 *) New Configure options --prefix=DIR and --openssldir=DIR.
9856 *) Complete rewrite of the error code script(s). It is all now handled
9857 by one script at the top level which handles error code gathering,
9858 header rewriting and C source file generation. It should be much better
9859 than the old method: it now uses a modified version of Ulf's parser to
9860 read the ANSI prototypes in all header files (thus the old K&R definitions
9861 aren't needed for error creation any more) and do a better job of
9862 translating function codes into names. The old 'ASN1 error code imbedded
9863 in a comment' is no longer necessary and it doesn't use .err files which
9864 have now been deleted. Also the error code call doesn't have to appear all
9865 on one line (which resulted in some large lines...).
9868 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9871 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9872 0 (which usually indicates a closed connection), but continue reading.
9875 *) Fix some race conditions.
9878 *) Add support for CRL distribution points extension. Add Certificate
9879 Policies and CRL distribution points documentation.
9882 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9885 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9886 8 of keying material. Merlin has also confirmed interop with this fix
9887 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9888 [Merlin Hughes <merlin@baltimore.ie>]
9890 *) Fix lots of warnings.
9891 [Richard Levitte <levitte@stacken.kth.se>]
9893 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9894 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9895 [Richard Levitte <levitte@stacken.kth.se>]
9897 *) Fix problems with sizeof(long) == 8.
9898 [Andy Polyakov <appro@fy.chalmers.se>]
9900 *) Change functions to ANSI C.
9903 *) Fix typos in error codes.
9904 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9906 *) Remove defunct assembler files from Configure.
9909 *) SPARC v8 assembler BIGNUM implementation.
9910 [Andy Polyakov <appro@fy.chalmers.se>]
9912 *) Support for Certificate Policies extension: both print and set.
9913 Various additions to support the r2i method this uses.
9916 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9917 return a const string when you are expecting an allocated buffer.
9920 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9921 types DirectoryString and DisplayText.
9924 *) Add code to allow r2i extensions to access the configuration database,
9925 add an LHASH database driver and add several ctx helper functions.
9928 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9929 fail when they extended the size of a BIGNUM.
9932 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9933 support typesafe stack.
9936 *) Fix typo in SSL_[gs]et_options().
9937 [Nils Frostberg <nils@medcom.se>]
9939 *) Delete various functions and files that belonged to the (now obsolete)
9940 old X509V3 handling code.
9943 *) New Configure option "rsaref".
9946 *) Don't auto-generate pem.h.
9949 *) Introduce type-safe ASN.1 SETs.
9952 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9953 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9955 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9956 that links with OpenSSL (well at least cause lots of warnings), but fear
9957 not: the conversion is trivial, and it eliminates loads of evil casts. A
9958 few STACKed things have been converted already. Feel free to convert more.
9959 In the fullness of time, I'll do away with the STACK type altogether.
9962 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9963 specified in <certfile> by updating the entry in the index.txt file.
9964 This way one no longer has to edit the index.txt file manually for
9965 revoking a certificate. The -revoke option does the gory details now.
9966 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9968 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9969 `-text' option at all and this way the `-noout -text' combination was
9970 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9971 [Ralf S. Engelschall]
9973 *) Make sure a corresponding plain text error message exists for the
9974 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9975 verify callback function determined that a certificate was revoked.
9976 [Ralf S. Engelschall]
9978 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9979 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9980 all available cipers including rc5, which was forgotten until now.
9981 In order to let the testing shell script know which algorithms
9982 are available, a new (up to now undocumented) command
9983 "openssl list-cipher-commands" is used.
9986 *) Bugfix: s_client occasionally would sleep in select() when
9987 it should have checked SSL_pending() first.
9990 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9991 the raw DSA values prior to ASN.1 encoding.
9994 *) Tweaks to Configure
9995 [Niels Poppe <niels@netbox.org>]
9997 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10001 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10004 *) New config option to avoid instructions that are illegal on the 80386.
10005 The default code is faster, but requires at least a 486.
10008 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10009 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10010 same as SSL2_VERSION anyway.
10013 *) New "-showcerts" option for s_client.
10016 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10017 application. Various cleanups and fixes.
10020 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10021 modify error routines to work internally. Add error codes and PBE init
10022 to library startup routines.
10025 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10026 packing functions to asn1 and evp. Changed function names and error
10027 codes along the way.
10030 *) PKCS12 integration: and so it begins... First of several patches to
10031 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10032 objects to objects.h
10035 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10036 and display support for Thawte strong extranet extension.
10039 *) Add LinuxPPC support.
10040 [Jeff Dubrule <igor@pobox.org>]
10042 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10043 bn_div_words in alpha.s.
10044 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10046 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10047 OAEP isn't supported when OpenSSL is built with RSAref.
10048 [Ulf Moeller <ulf@fitug.de>]
10050 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10051 so they no longer are missing under -DNOPROTO.
10052 [Soren S. Jorvang <soren@t.dk>]
10055 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10057 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10058 doesn't work when the session is reused. Coming soon!
10061 *) Fix a security hole, that allows sessions to be reused in the wrong
10062 context thus bypassing client cert protection! All software that uses
10063 client certs and session caches in multiple contexts NEEDS PATCHING to
10064 allow session reuse! A fuller solution is in the works.
10065 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10067 *) Some more source tree cleanups (removed obsolete files
10068 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10069 permission on "config" script to be executable) and a fix for the INSTALL
10071 [Ulf Moeller <ulf@fitug.de>]
10073 *) Remove some legacy and erroneous uses of malloc, free instead of
10075 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10077 *) Make rsa_oaep_test return non-zero on error.
10078 [Ulf Moeller <ulf@fitug.de>]
10080 *) Add support for native Solaris shared libraries. Configure
10081 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10082 if someone would make that last step automatic.
10083 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10085 *) ctx_size was not built with the right compiler during "make links". Fixed.
10088 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10089 except NULL ciphers". This means the default cipher list will no longer
10090 enable NULL ciphers. They need to be specifically enabled e.g. with
10091 the string "DEFAULT:eNULL".
10094 *) Fix to RSA private encryption routines: if p < q then it would
10095 occasionally produce an invalid result. This will only happen with
10096 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10099 *) Be less restrictive and allow also `perl util/perlpath.pl
10100 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10101 because this way one can also use an interpreter named `perl5' (which is
10102 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10103 installed as `perl').
10104 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10106 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10107 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10109 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10110 advapi32.lib to Win32 build and change the pem test comparision
10111 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10112 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10113 and crypto/des/ede_cbcm_enc.c.
10116 *) DES quad checksum was broken on big-endian architectures. Fixed.
10119 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10120 Win32 test batch file so it (might) work again. The Win32 test batch file
10121 is horrible: I feel ill....
10124 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10125 in e_os.h. Audit of header files to check ANSI and non ANSI
10126 sections: 10 functions were absent from non ANSI section and not exported
10127 from Windows DLLs. Fixed up libeay.num for new functions.
10130 *) Make `openssl version' output lines consistent.
10131 [Ralf S. Engelschall]
10133 *) Fix Win32 symbol export lists for BIO functions: Added
10134 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10135 to ms/libeay{16,32}.def.
10136 [Ralf S. Engelschall]
10138 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10139 fine under Unix and passes some trivial tests I've now added. But the
10140 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10141 added to make sure no one expects that this stuff really works in the
10142 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10143 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10145 [Ralf S. Engelschall]
10147 *) Fix the generation of two part addresses in perl.
10148 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10150 *) Add config entry for Linux on MIPS.
10151 [John Tobey <jtobey@channel1.com>]
10153 *) Make links whenever Configure is run, unless we are on Windoze.
10156 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10157 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10161 *) Add a useful kludge to allow package maintainers to specify compiler and
10162 other platforms details on the command line without having to patch the
10163 Configure script everytime: One now can use ``perl Configure
10164 <id>:<details>'', i.e. platform ids are allowed to have details appended
10165 to them (seperated by colons). This is treated as there would be a static
10166 pre-configured entry in Configure's %table under key <id> with value
10167 <details> and ``perl Configure <id>'' is called. So, when you want to
10168 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10169 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10170 now, which overrides the FreeBSD-elf entry on-the-fly.
10171 [Ralf S. Engelschall]
10173 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10176 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10177 on the `perl Configure ...' command line. This way one can compile
10178 OpenSSL libraries with Position Independent Code (PIC) which is needed
10179 for linking it into DSOs.
10180 [Ralf S. Engelschall]
10182 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10186 *) Cleaned up the LICENSE document: The official contact for any license
10187 questions now is the OpenSSL core team under openssl-core@openssl.org.
10188 And add a paragraph about the dual-license situation to make sure people
10189 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10190 to the OpenSSL toolkit.
10191 [Ralf S. Engelschall]
10193 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10194 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10195 Additonally cleaned up the `make links' target: Remove unnecessary
10196 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10197 to speed processing and no longer clutter the display with confusing
10198 stuff. Instead only the actually done links are displayed.
10199 [Ralf S. Engelschall]
10201 *) Permit null encryption ciphersuites, used for authentication only. It used
10202 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10203 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10207 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10208 signed attributes when verifying signatures (this would break them),
10209 the detached data encoding was wrong and public keys obtained using
10210 X509_get_pubkey() weren't freed.
10213 *) Add text documentation for the BUFFER functions. Also added a work around
10214 to a Win95 console bug. This was triggered by the password read stuff: the
10215 last character typed gets carried over to the next fread(). If you were
10216 generating a new cert request using 'req' for example then the last
10217 character of the passphrase would be CR which would then enter the first
10221 *) Added the new `Includes OpenSSL Cryptography Software' button as
10222 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10223 button and can be used by applications based on OpenSSL to show the
10224 relationship to the OpenSSL project.
10225 [Ralf S. Engelschall]
10227 *) Remove confusing variables in function signatures in files
10228 ssl/ssl_lib.c and ssl/ssl.h.
10229 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10231 *) Don't install bss_file.c under PREFIX/include/
10232 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10234 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10235 functions that return function pointers and has support for NT specific
10236 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10237 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10238 unsigned to signed types: this was killing the Win32 compile.
10241 *) Add new certificate file to stack functions,
10242 SSL_add_dir_cert_subjects_to_stack() and
10243 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10244 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10245 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10246 This means that Apache-SSL and similar packages don't have to mess around
10247 to add as many CAs as they want to the preferred list.
10250 *) Experiment with doxygen documentation. Currently only partially applied to
10252 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10253 openssl.doxy as the configuration file.
10256 *) Get rid of remaining C++-style comments which strict C compilers hate.
10257 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10259 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10260 compiled in by default: it has problems with large keys.
10263 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10264 DH private keys and/or callback functions which directly correspond to
10265 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10266 is needed for applications which have to configure certificates on a
10267 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10269 For the RSA certificate situation is makes no difference, but
10270 for the DSA certificate situation this fixes the "no shared cipher"
10271 problem where the OpenSSL cipher selection procedure failed because the
10272 temporary keys were not overtaken from the context and the API provided
10273 no way to reconfigure them.
10274 The new functions now let applications reconfigure the stuff and they
10275 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10276 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10277 non-public-API function ssl_cert_instantiate() is used as a helper
10278 function and also to reduce code redundancy inside ssl_rsa.c.
10279 [Ralf S. Engelschall]
10281 *) Move s_server -dcert and -dkey options out of the undocumented feature
10282 area because they are useful for the DSA situation and should be
10283 recognized by the users.
10284 [Ralf S. Engelschall]
10286 *) Fix the cipher decision scheme for export ciphers: the export bits are
10287 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10288 SSL_EXP_MASK. So, the original variable has to be used instead of the
10289 already masked variable.
10290 [Richard Levitte <levitte@stacken.kth.se>]
10292 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10293 [Richard Levitte <levitte@stacken.kth.se>]
10295 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10296 from `int' to `unsigned int' because it's a length and initialized by
10297 EVP_DigestFinal() which expects an `unsigned int *'.
10298 [Richard Levitte <levitte@stacken.kth.se>]
10300 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10301 script. Instead use the usual Shell->Perl transition trick.
10302 [Ralf S. Engelschall]
10304 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10305 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10306 -noout -modulus' as it's already the case for `openssl rsa -noout
10307 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10308 currently the public key is printed (a decision which was already done by
10309 `openssl dsa -modulus' in the past) which serves a similar purpose.
10310 Additionally the NO_RSA no longer completely removes the whole -modulus
10311 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10313 [Ralf S. Engelschall]
10315 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10316 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10317 [Arne Ansper <arne@ats.cyber.ee>]
10319 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10320 to be added. Now both 'req' and 'ca' can use new objects defined in the
10324 *) Add cool BIO that does syslog (or event log on NT).
10325 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10327 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10328 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10329 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10330 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10333 *) Add preliminary config info for new extension code.
10336 *) Make RSA_NO_PADDING really use no padding.
10337 [Ulf Moeller <ulf@fitug.de>]
10339 *) Generate errors when private/public key check is done.
10342 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10343 for some CRL extensions and new objects added.
10346 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10347 key usage extension and fuller support for authority key id.
10350 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10351 padding method for RSA, which is recommended for new applications in PKCS
10352 #1 v2.0 (RFC 2437, October 1998).
10353 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10354 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10355 against Bleichbacher's attack on RSA.
10356 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10359 *) Updates to the new SSL compression code
10360 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10362 *) Fix so that the version number in the master secret, when passed
10363 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10364 (because the server will not accept higher), that the version number
10365 is 0x03,0x01, not 0x03,0x00
10366 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10368 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10369 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10370 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10373 *) Support for RAW extensions where an arbitrary extension can be
10374 created by including its DER encoding. See apps/openssl.cnf for
10378 *) Make sure latest Perl versions don't interpret some generated C array
10379 code as Perl array code in the crypto/err/err_genc.pl script.
10380 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10382 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10383 not many people have the assembler. Various Win32 compilation fixes and
10384 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10385 build instructions.
10388 *) Modify configure script 'Configure' to automatically create crypto/date.h
10389 file under Win32 and also build pem.h from pem.org. New script
10390 util/mkfiles.pl to create the MINFO file on environments that can't do a
10391 'make files': perl util/mkfiles.pl >MINFO should work.
10394 *) Major rework of DES function declarations, in the pursuit of correctness
10395 and purity. As a result, many evil casts evaporated, and some weirdness,
10396 too. You may find this causes warnings in your code. Zapping your evil
10397 casts will probably fix them. Mostly.
10400 *) Fix for a typo in asn1.h. Bug fix to object creation script
10401 obj_dat.pl. It considered a zero in an object definition to mean
10402 "end of object": none of the objects in objects.h have any zeros
10403 so it wasn't spotted.
10404 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10406 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10407 Masking (CBCM). In the absence of test vectors, the best I have been able
10408 to do is check that the decrypt undoes the encrypt, so far. Send me test
10409 vectors if you have them.
10412 *) Correct calculation of key length for export ciphers (too much space was
10413 allocated for null ciphers). This has not been tested!
10416 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10417 message is now correct (it understands "crypto" and "ssl" on its
10418 command line). There is also now an "update" option. This will update
10419 the util/ssleay.num and util/libeay.num files with any new functions.
10421 perl util/mkdef.pl crypto ssl update
10422 it will update them.
10425 *) Overhauled the Perl interface (perl/*):
10426 - ported BN stuff to OpenSSL's different BN library
10427 - made the perl/ source tree CVS-aware
10428 - renamed the package from SSLeay to OpenSSL (the files still contain
10429 their history because I've copied them in the repository)
10430 - removed obsolete files (the test scripts will be replaced
10431 by better Test::Harness variants in the future)
10432 [Ralf S. Engelschall]
10434 *) First cut for a very conservative source tree cleanup:
10435 1. merge various obsolete readme texts into doc/ssleay.txt
10436 where we collect the old documents and readme texts.
10437 2. remove the first part of files where I'm already sure that we no
10438 longer need them because of three reasons: either they are just temporary
10439 files which were left by Eric or they are preserved original files where
10440 I've verified that the diff is also available in the CVS via "cvs diff
10441 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10442 the crypto/md/ stuff).
10443 [Ralf S. Engelschall]
10445 *) More extension code. Incomplete support for subject and issuer alt
10446 name, issuer and authority key id. Change the i2v function parameters
10447 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10448 what that's for :-) Fix to ASN1 macro which messed up
10449 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10452 *) Preliminary support for ENUMERATED type. This is largely copied from the
10456 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10457 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10459 *) Make sure `make rehash' target really finds the `openssl' program.
10460 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10462 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10463 like to hear about it if this slows down other processors.
10466 *) Add CygWin32 platform information to Configure script.
10467 [Alan Batie <batie@aahz.jf.intel.com>]
10469 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10470 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10472 *) New program nseq to manipulate netscape certificate sequences
10475 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10479 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10480 but the BN code had some problems that would cause failures when
10481 doing certificate verification and some other functions.
10482 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10484 *) Add ASN1 and PEM code to support netscape certificate sequences.
10487 *) Add ASN1 and PEM code to support netscape certificate sequences.
10490 *) Add several PKIX and private extended key usage OIDs.
10493 *) Modify the 'ca' program to handle the new extension code. Modify
10494 openssl.cnf for new extension format, add comments.
10497 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10498 and add a sample to openssl.cnf so req -x509 now adds appropriate
10502 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10503 error code, add initial support to X509_print() and x509 application.
10506 *) Takes a deep breath and start addding X509 V3 extension support code. Add
10507 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10508 stuff is currently isolated and isn't even compiled yet.
10511 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10512 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10513 Removed the versions check from X509 routines when loading extensions:
10514 this allows certain broken certificates that don't set the version
10515 properly to be processed.
10518 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10519 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10520 can still be regenerated with "make depend".
10523 *) Spelling mistake in C version of CAST-128.
10524 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10526 *) Changes to the error generation code. The perl script err-code.pl
10527 now reads in the old error codes and retains the old numbers, only
10528 adding new ones if necessary. It also only changes the .err files if new
10529 codes are added. The makefiles have been modified to only insert errors
10530 when needed (to avoid needlessly modifying header files). This is done
10531 by only inserting errors if the .err file is newer than the auto generated
10532 C file. To rebuild all the error codes from scratch (the old behaviour)
10533 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10534 or delete all the .err files.
10537 *) CAST-128 was incorrectly implemented for short keys. The C version has
10538 been fixed, but is untested. The assembler versions are also fixed, but
10539 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10540 to regenerate it if needed.
10541 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10542 Hagino <itojun@kame.net>]
10544 *) File was opened incorrectly in randfile.c.
10545 [Ulf Möller <ulf@fitug.de>]
10547 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10548 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10549 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10550 al: it's just almost always a UTCTime. Note this patch adds new error
10551 codes so do a "make errors" if there are problems.
10554 *) Correct Linux 1 recognition in config.
10555 [Ulf Möller <ulf@fitug.de>]
10557 *) Remove pointless MD5 hash when using DSA keys in ca.
10558 [Anonymous <nobody@replay.com>]
10560 *) Generate an error if given an empty string as a cert directory. Also
10561 generate an error if handed NULL (previously returned 0 to indicate an
10562 error, but didn't set one).
10563 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10565 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10568 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10569 parameters. This was causing a warning which killed off the Win32 compile.
10572 *) Remove C++ style comments from crypto/bn/bn_local.h.
10573 [Neil Costigan <neil.costigan@celocom.com>]
10575 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10576 based on a text string, looking up short and long names and finally
10577 "dot" format. The "dot" format stuff didn't work. Added new function
10578 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10579 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10580 OID is not part of the table.
10583 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10584 X509_LOOKUP_by_alias().
10587 *) Sort openssl functions by name.
10590 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10591 encryption from sample DSA keys (in case anyone is interested the password
10595 *) Make _all_ *_free functions accept a NULL pointer.
10596 [Frans Heymans <fheymans@isaserver.be>]
10598 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10600 [Anonymous <nobody@replay.com>]
10602 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10603 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10605 *) Don't blow it for numeric -newkey arguments to apps/req.
10606 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10608 *) Temp key "for export" tests were wrong in s3_srvr.c.
10609 [Anonymous <nobody@replay.com>]
10611 *) Add prototype for temp key callback functions
10612 SSL_CTX_set_tmp_{rsa,dh}_callback().
10615 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10616 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10619 *) X509_name_add_entry() freed the wrong thing after an error.
10620 [Arne Ansper <arne@ats.cyber.ee>]
10622 *) rsa_eay.c would attempt to free a NULL context.
10623 [Arne Ansper <arne@ats.cyber.ee>]
10625 *) BIO_s_socket() had a broken should_retry() on Windoze.
10626 [Arne Ansper <arne@ats.cyber.ee>]
10628 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10629 [Arne Ansper <arne@ats.cyber.ee>]
10631 *) Make sure the already existing X509_STORE->depth variable is initialized
10632 in X509_STORE_new(), but document the fact that this variable is still
10633 unused in the certificate verification process.
10634 [Ralf S. Engelschall]
10636 *) Fix the various library and apps files to free up pkeys obtained from
10637 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10640 *) Fix reference counting in X509_PUBKEY_get(). This makes
10641 demos/maurice/example2.c work, amongst others, probably.
10642 [Steve Henson and Ben Laurie]
10644 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10645 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10646 are no longer created. This way we have a single and consistent command
10647 line interface `openssl <command>', similar to `cvs <command>'.
10648 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10650 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10651 BIT STRING wrapper always have zero unused bits.
10654 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10657 *) Make the top-level INSTALL documentation easier to understand.
10660 *) Makefiles updated to exit if an error occurs in a sub-directory
10661 make (including if user presses ^C) [Paul Sutton]
10663 *) Make Montgomery context stuff explicit in RSA data structure.
10666 *) Fix build order of pem and err to allow for generated pem.h.
10669 *) Fix renumbering bug in X509_NAME_delete_entry().
10672 *) Enhanced the err-ins.pl script so it makes the error library number
10673 global and can add a library name. This is needed for external ASN1 and
10674 other error libraries.
10677 *) Fixed sk_insert which never worked properly.
10680 *) Fix ASN1 macros so they can handle indefinite length construted
10681 EXPLICIT tags. Some non standard certificates use these: they can now
10685 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10686 into a single doc/ssleay.txt bundle. This way the information is still
10687 preserved but no longer messes up this directory. Now it's new room for
10688 the new set of documenation files.
10689 [Ralf S. Engelschall]
10691 *) SETs were incorrectly DER encoded. This was a major pain, because they
10692 shared code with SEQUENCEs, which aren't coded the same. This means that
10693 almost everything to do with SETs or SEQUENCEs has either changed name or
10694 number of arguments.
10695 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10697 *) Fix test data to work with the above.
10700 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10701 was already fixed by Eric for 0.9.1 it seems.
10702 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10704 *) Autodetect FreeBSD3.
10707 *) Fix various bugs in Configure. This affects the following platforms:
10711 unixware-2.0-pentium
10715 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10716 before they are needed.
10719 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10723 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10725 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10726 changed SSLeay to OpenSSL in version strings.
10727 [Ralf S. Engelschall]
10729 *) Some fixups to the top-level documents.
10732 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10733 because the symlink to include/ was missing.
10734 [Ralf S. Engelschall]
10736 *) Incorporated the popular no-RSA/DSA-only patches
10737 which allow to compile a RSA-free SSLeay.
10738 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10740 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10741 when "ssleay" is still not found.
10742 [Ralf S. Engelschall]
10744 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10745 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10747 *) Updated the README file.
10748 [Ralf S. Engelschall]
10750 *) Added various .cvsignore files in the CVS repository subdirs
10751 to make a "cvs update" really silent.
10752 [Ralf S. Engelschall]
10754 *) Recompiled the error-definition header files and added
10755 missing symbols to the Win32 linker tables.
10756 [Ralf S. Engelschall]
10758 *) Cleaned up the top-level documents;
10759 o new files: CHANGES and LICENSE
10760 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10761 o merged COPYRIGHT into LICENSE
10762 o removed obsolete TODO file
10763 o renamed MICROSOFT to INSTALL.W32
10764 [Ralf S. Engelschall]
10766 *) Removed dummy files from the 0.9.1b source tree:
10767 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10768 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10769 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10770 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10771 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10772 [Ralf S. Engelschall]
10774 *) Added various platform portability fixes.
10777 *) The Genesis of the OpenSSL rpject:
10778 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10779 Young and Tim J. Hudson created while they were working for C2Net until
10781 [The OpenSSL Project]
10784 Changes between 0.9.0b and 0.9.1b [not released]
10786 *) Updated a few CA certificates under certs/
10789 *) Changed some BIGNUM api stuff.
10792 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10793 DGUX x86, Linux Alpha, etc.
10796 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10797 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10801 *) Add -strparse option to asn1pars program which parses nested
10803 [Dr Stephen Henson <shenson@bigfoot.com>]
10805 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10808 *) DSA fix for "ca" program.
10811 *) Added "-genkey" option to "dsaparam" program.
10814 *) Added RIPE MD160 (rmd160) message digest.
10817 *) Added -a (all) option to "ssleay version" command.
10820 *) Added PLATFORM define which is the id given to Configure.
10823 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10826 *) Extended the ASN.1 parser routines.
10829 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10832 *) Added a BN_CTX to the BN library.
10835 *) Fixed the weak key values in DES library
10838 *) Changed API in EVP library for cipher aliases.
10841 *) Added support for RC2/64bit cipher.
10844 *) Converted the lhash library to the crypto/mem.c functions.
10847 *) Added more recognized ASN.1 object ids.
10850 *) Added more RSA padding checks for SSL/TLS.
10853 *) Added BIO proxy/filter functionality.
10856 *) Added extra_certs to SSL_CTX which can be used
10857 send extra CA certificates to the client in the CA cert chain sending
10858 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10861 *) Now Fortezza is denied in the authentication phase because
10862 this is key exchange mechanism is not supported by SSLeay at all.
10865 *) Additional PKCS1 checks.
10868 *) Support the string "TLSv1" for all TLS v1 ciphers.
10871 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10872 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10875 *) Fixed a few memory leaks.
10878 *) Fixed various code and comment typos.
10881 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10882 bytes sent in the client random.
10883 [Edward Bishop <ebishop@spyglass.com>]