5 Changes between 1.0.1k and 1.0.2 [xx XXX xxxx]
9 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
10 sends a carefully crafted handshake message, to cause OpenSSL to fail
11 to free up to 64k of memory causing a memory leak. This could be
12 exploited in a Denial Of Service attack. This issue affects OpenSSL
13 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
14 whether SRTP is used or configured. Implementations of OpenSSL that
15 have been compiled with OPENSSL_NO_SRTP defined are not affected.
17 The fix was developed by the OpenSSL team.
21 *) Session Ticket Memory Leak.
23 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
24 integrity of that ticket is first verified. In the event of a session
25 ticket integrity check failing, OpenSSL will fail to free memory
26 causing a memory leak. By sending a large number of invalid session
27 tickets an attacker could exploit this issue in a Denial Of Service
32 *) Build option no-ssl3 is incomplete.
34 When OpenSSL is configured with "no-ssl3" as a build option, servers
35 could accept and complete a SSL 3.0 handshake, and clients could be
36 configured to send them.
38 [Akamai and the OpenSSL team]
40 *) Add support for TLS_FALLBACK_SCSV.
41 Client applications doing fallback retries should call
42 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
44 [Adam Langley, Bodo Moeller]
46 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
47 (other platforms pending).
48 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
50 *) Add support for the SignedCertificateTimestampList certificate and
51 OCSP response extensions from RFC6962.
54 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
55 for corner cases. (Certain input points at infinity could lead to
56 bogus results, with non-infinity inputs mapped to infinity too.)
59 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
60 This covers AES, SHA256/512 and GHASH. "Initial" means that most
61 common cases are optimized and there still is room for further
62 improvements. Vector Permutation AES for Altivec is also added.
65 *) Add support for little-endian ppc64 Linux target.
68 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
69 SHA1, SHA256 and GHASH. "Initial" means that most common cases
70 are optimized and there still is room for further improvements.
71 Both 32- and 64-bit modes are supported.
72 [Andy Polyakov, Ard Biesheuvel (Linaro)]
74 *) Improved ARMv7 NEON support.
77 *) Support for SPARC Architecture 2011 crypto extensions, first
78 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
79 SHA256/512, MD5, GHASH and modular exponentiation.
80 [Andy Polyakov, David Miller]
82 *) Accelerated modular exponentiation for Intel processors, a.k.a.
84 [Shay Gueron & Vlad Krasnov (Intel Corp)]
86 *) Support for new and upcoming Intel processors, including AVX2,
87 BMI and SHA ISA extensions. This includes additional "stitched"
88 implementations, AESNI-SHA256 and GCM, and multi-buffer support
91 This work was sponsored by Intel Corp.
94 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
95 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
96 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
99 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
100 this fixes a limiation in previous versions of OpenSSL.
103 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
104 MGF1 digest and OAEP label.
107 *) Add EVP support for key wrapping algorithms, to avoid problems with
108 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
109 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
110 algorithms and include tests cases.
113 *) Add functions to allocate and set the fields of an ECDSA_METHOD
115 [Douglas E. Engert, Steve Henson]
117 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
118 difference in days and seconds between two tm or ASN1_TIME structures.
121 *) Add -rev test option to s_server to just reverse order of characters
122 received by client and send back to server. Also prints an abbreviated
123 summary of the connection parameters.
126 *) New option -brief for s_client and s_server to print out a brief summary
127 of connection parameters.
130 *) Add callbacks for arbitrary TLS extensions.
131 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
133 *) New option -crl_download in several openssl utilities to download CRLs
134 from CRLDP extension in certificates.
137 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
140 *) New function X509_CRL_diff to generate a delta CRL from the difference
141 of two full CRLs. Add support to "crl" utility.
144 *) New functions to set lookup_crls function and to retrieve
145 X509_STORE from X509_STORE_CTX.
148 *) Print out deprecated issuer and subject unique ID fields in
152 *) Extend OCSP I/O functions so they can be used for simple general purpose
153 HTTP as well as OCSP. New wrapper function which can be used to download
154 CRLs using the OCSP API.
157 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
160 *) SSL_CONF* functions. These provide a common framework for application
161 configuration using configuration files or command lines.
164 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
165 message callback and prints the results. Needs compile time option
166 "enable-ssl-trace". New options to s_client and s_server to enable
170 *) New ctrl and macro to retrieve supported points extensions.
171 Print out extension in s_server and s_client.
174 *) New functions to retrieve certificate signature and signature
178 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
182 *) New Suite B modes for TLS code. These use and enforce the requirements
183 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
184 only use Suite B curves. The Suite B modes can be set by using the
185 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
188 *) New chain verification flags for Suite B levels of security. Check
189 algorithms are acceptable when flags are set in X509_verify_cert.
192 *) Make tls1_check_chain return a set of flags indicating checks passed
193 by a certificate chain. Add additional tests to handle client
194 certificates: checks for matching certificate type and issuer name
198 *) If an attempt is made to use a signature algorithm not in the peer
199 preference list abort the handshake. If client has no suitable
200 signature algorithms in response to a certificate request do not
204 *) If server EC tmp key is not in client preference list abort handshake.
207 *) Add support for certificate stores in CERT structure. This makes it
208 possible to have different stores per SSL structure or one store in
209 the parent SSL_CTX. Include distint stores for certificate chain
210 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
211 to build and store a certificate chain in CERT structure: returing
212 an error if the chain cannot be built: this will allow applications
213 to test if a chain is correctly configured.
215 Note: if the CERT based stores are not set then the parent SSL_CTX
216 store is used to retain compatibility with existing behaviour.
220 *) New function ssl_set_client_disabled to set a ciphersuite disabled
221 mask based on the current session, check mask when sending client
222 hello and checking the requested ciphersuite.
225 *) New ctrls to retrieve and set certificate types in a certificate
226 request message. Print out received values in s_client. If certificate
227 types is not set with custom values set sensible values based on
228 supported signature algorithms.
231 *) Support for distinct client and server supported signature algorithms.
234 *) Add certificate callback. If set this is called whenever a certificate
235 is required by client or server. An application can decide which
236 certificate chain to present based on arbitrary criteria: for example
237 supported signature algorithms. Add very simple example to s_server.
238 This fixes many of the problems and restrictions of the existing client
239 certificate callback: for example you can now clear an existing
240 certificate and specify the whole chain.
243 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
244 the certificate can be used for (if anything). Set valid_flags field
245 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
246 to have similar checks in it.
248 Add new "cert_flags" field to CERT structure and include a "strict mode".
249 This enforces some TLS certificate requirements (such as only permitting
250 certificate signature algorithms contained in the supported algorithms
251 extension) which some implementations ignore: this option should be used
252 with caution as it could cause interoperability issues.
255 *) Update and tidy signature algorithm extension processing. Work out
256 shared signature algorithms based on preferences and peer algorithms
257 and print them out in s_client and s_server. Abort handshake if no
258 shared signature algorithms.
261 *) Add new functions to allow customised supported signature algorithms
262 for SSL and SSL_CTX structures. Add options to s_client and s_server
266 *) New function SSL_certs_clear() to delete all references to certificates
267 from an SSL structure. Before this once a certificate had been added
268 it couldn't be removed.
271 *) Integrate hostname, email address and IP address checking with certificate
272 verification. New verify options supporting checking in opensl utility.
275 *) Fixes and wildcard matching support to hostname and email checking
276 functions. Add manual page.
277 [Florian Weimer (Red Hat Product Security Team)]
279 *) New functions to check a hostname email or IP address against a
280 certificate. Add options x509 utility to print results of checks against
284 *) Fix OCSP checking.
285 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
287 *) Initial experimental support for explicitly trusted non-root CAs.
288 OpenSSL still tries to build a complete chain to a root but if an
289 intermediate CA has a trust setting included that is used. The first
290 setting is used: whether to trust (e.g., -addtrust option to the x509
294 *) Add -trusted_first option which attempts to find certificates in the
295 trusted store even if an untrusted chain is also supplied.
298 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
299 platform support for Linux and Android.
302 *) Support for linux-x32, ILP32 environment in x86_64 framework.
305 *) Experimental multi-implementation support for FIPS capable OpenSSL.
306 When in FIPS mode the approved implementations are used as normal,
307 when not in FIPS mode the internal unapproved versions are used instead.
308 This means that the FIPS capable OpenSSL isn't forced to use the
309 (often lower perfomance) FIPS implementations outside FIPS mode.
312 *) Transparently support X9.42 DH parameters when calling
313 PEM_read_bio_DHparameters. This means existing applications can handle
314 the new parameter format automatically.
317 *) Initial experimental support for X9.42 DH parameter format: mainly
318 to support use of 'q' parameter for RFC5114 parameters.
321 *) Add DH parameters from RFC5114 including test data to dhtest.
324 *) Support for automatic EC temporary key parameter selection. If enabled
325 the most preferred EC parameters are automatically used instead of
326 hardcoded fixed parameters. Now a server just has to call:
327 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
328 support ECDH and use the most appropriate parameters.
331 *) Enhance and tidy EC curve and point format TLS extension code. Use
332 static structures instead of allocation if default values are used.
333 New ctrls to set curves we wish to support and to retrieve shared curves.
334 Print out shared curves in s_server. New options to s_server and s_client
335 to set list of supported curves.
338 *) New ctrls to retrieve supported signature algorithms and
339 supported curve values as an array of NIDs. Extend openssl utility
340 to print out received values.
343 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
344 between NIDs and the more common NIST names such as "P-256". Enhance
345 ecparam utility and ECC method to recognise the NIST names for curves.
348 *) Enhance SSL/TLS certificate chain handling to support different
349 chains for each certificate instead of one chain in the parent SSL_CTX.
352 *) Support for fixed DH ciphersuite client authentication: where both
353 server and client use DH certificates with common parameters.
356 *) Support for fixed DH ciphersuites: those requiring DH server
360 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
362 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
363 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
364 X509_CINF_get_signature were reverted post internal team review.
366 Changes between 1.0.1j and 1.0.1k [xx XXX xxxx]
368 *) Abort handshake if server key exchange message is omitted for ephemeral
371 Thanks to Karthikeyan Bhargavan for reporting this issue.
375 *) Ensure that the session ID context of an SSL is updated when its
376 SSL_CTX is updated via SSL_set_SSL_CTX.
378 The session ID context is typically set from the parent SSL_CTX,
379 and can vary with the CTX.
382 *) Fix various certificate fingerprint issues.
384 By using non-DER or invalid encodings outside the signed portion of a
385 certificate the fingerprint can be changed without breaking the signature.
386 Although no details of the signed portion of the certificate can be changed
387 this can cause problems with some applications: e.g. those using the
388 certificate fingerprint for blacklists.
390 1. Reject signatures with non zero unused bits.
392 If the BIT STRING containing the signature has non zero unused bits reject
393 the signature. All current signature algorithms require zero unused bits.
395 2. Check certificate algorithm consistency.
397 Check the AlgorithmIdentifier inside TBS matches the one in the
398 certificate signature. NB: this will result in signature failure
399 errors for some broken certificates.
401 Thanks to Konrad Kraszewski from Google for reporting this issue.
403 3. Check DSA/ECDSA signatures use DER.
405 Reencode DSA/ECDSA signatures and compare with the original received
406 signature. Return an error if there is a mismatch.
408 This will reject various cases including garbage after signature
409 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
410 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
411 (negative or with leading zeroes).
413 Further analysis was conducted and fixes were developed by Stephen Henson
414 of the OpenSSL core team.
419 *) Do not resume sessions on the server if the negotiated protocol
420 version does not match the session's version. Resuming with a different
421 version, while not strictly forbidden by the RFC, is of questionable
422 sanity and breaks all known clients.
423 [David Benjamin, Emilia Käsper]
425 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
426 early CCS messages during renegotiation. (Note that because
427 renegotiation is encrypted, this early CCS was not exploitable.)
430 *) Tighten client-side session ticket handling during renegotiation:
431 ensure that the client only accepts a session ticket if the server sends
432 the extension anew in the ServerHello. Previously, a TLS client would
433 reuse the old extension state and thus accept a session ticket if one was
434 announced in the initial ServerHello.
436 Similarly, ensure that the client requires a session ticket if one
437 was advertised in the ServerHello. Previously, a TLS client would
438 ignore a missing NewSessionTicket message.
441 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
445 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
446 sends a carefully crafted handshake message, to cause OpenSSL to fail
447 to free up to 64k of memory causing a memory leak. This could be
448 exploited in a Denial Of Service attack. This issue affects OpenSSL
449 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
450 whether SRTP is used or configured. Implementations of OpenSSL that
451 have been compiled with OPENSSL_NO_SRTP defined are not affected.
453 The fix was developed by the OpenSSL team.
457 *) Session Ticket Memory Leak.
459 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
460 integrity of that ticket is first verified. In the event of a session
461 ticket integrity check failing, OpenSSL will fail to free memory
462 causing a memory leak. By sending a large number of invalid session
463 tickets an attacker could exploit this issue in a Denial Of Service
468 *) Build option no-ssl3 is incomplete.
470 When OpenSSL is configured with "no-ssl3" as a build option, servers
471 could accept and complete a SSL 3.0 handshake, and clients could be
472 configured to send them.
474 [Akamai and the OpenSSL team]
476 *) Add support for TLS_FALLBACK_SCSV.
477 Client applications doing fallback retries should call
478 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
480 [Adam Langley, Bodo Moeller]
482 *) Add additional DigestInfo checks.
484 Reencode DigestInto in DER and check against the original when
485 verifying RSA signature: this will reject any improperly encoded
486 DigestInfo structures.
488 Note: this is a precautionary measure and no attacks are currently known.
492 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
494 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
495 SRP code can be overrun an internal buffer. Add sanity check that
496 g, A, B < N to SRP code.
498 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
499 Group for discovering this issue.
503 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
504 TLS 1.0 instead of higher protocol versions when the ClientHello message
505 is badly fragmented. This allows a man-in-the-middle attacker to force a
506 downgrade to TLS 1.0 even if both the server and the client support a
507 higher protocol version, by modifying the client's TLS records.
509 Thanks to David Benjamin and Adam Langley (Google) for discovering and
510 researching this issue.
514 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
515 to a denial of service attack. A malicious server can crash the client
516 with a null pointer dereference (read) by specifying an anonymous (EC)DH
517 ciphersuite and sending carefully crafted handshake messages.
519 Thanks to Felix Gröbert (Google) for discovering and researching this
524 *) By sending carefully crafted DTLS packets an attacker could cause openssl
525 to leak memory. This can be exploited through a Denial of Service attack.
526 Thanks to Adam Langley for discovering and researching this issue.
530 *) An attacker can force openssl to consume large amounts of memory whilst
531 processing DTLS handshake messages. This can be exploited through a
532 Denial of Service attack.
533 Thanks to Adam Langley for discovering and researching this issue.
537 *) An attacker can force an error condition which causes openssl to crash
538 whilst processing DTLS packets due to memory being freed twice. This
539 can be exploited through a Denial of Service attack.
540 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
545 *) If a multithreaded client connects to a malicious server using a resumed
546 session and the server sends an ec point format extension it could write
547 up to 255 bytes to freed memory.
549 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
554 *) A malicious server can crash an OpenSSL client with a null pointer
555 dereference (read) by specifying an SRP ciphersuite even though it was not
556 properly negotiated with the client. This can be exploited through a
557 Denial of Service attack.
559 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
560 discovering and researching this issue.
564 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
565 X509_name_oneline, X509_name_print_ex et al. to leak some information
566 from the stack. Applications may be affected if they echo pretty printing
567 output to the attacker.
569 Thanks to Ivan Fratric (Google) for discovering this issue.
571 [Emilia Käsper, and Steve Henson]
573 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
574 for corner cases. (Certain input points at infinity could lead to
575 bogus results, with non-infinity inputs mapped to infinity too.)
578 Changes between 1.0.1i and 1.0.1j [xx XXX xxxx]
580 *) Add additional DigestInfo checks.
582 Reencode DigestInto in DER and check against the original when
583 verifying RSA signature: this will reject any improperly encoded
584 DigestInfo structures.
586 Note: this is a precautionary measure and no attacks are currently known.
590 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
592 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
593 handshake can force the use of weak keying material in OpenSSL
594 SSL/TLS clients and servers.
596 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
597 researching this issue. (CVE-2014-0224)
598 [KIKUCHI Masashi, Steve Henson]
600 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
601 OpenSSL DTLS client the code can be made to recurse eventually crashing
604 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
606 [Imre Rad, Steve Henson]
608 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
609 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
610 client or server. This is potentially exploitable to run arbitrary
611 code on a vulnerable client or server.
613 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
614 [Jüri Aedla, Steve Henson]
616 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
617 are subject to a denial of service attack.
619 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
620 this issue. (CVE-2014-3470)
621 [Felix Gröbert, Ivan Fratric, Steve Henson]
623 *) Harmonize version and its documentation. -f flag is used to display
625 [mancha <mancha1@zoho.com>]
627 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
629 [mancha <mancha1@zoho.com>]
631 *) Fix some double frees. These are not thought to be exploitable.
632 [mancha <mancha1@zoho.com>]
634 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
636 *) A missing bounds check in the handling of the TLS heartbeat extension
637 can be used to reveal up to 64k of memory to a connected client or
640 Thanks for Neel Mehta of Google Security for discovering this bug and to
641 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
642 preparing the fix (CVE-2014-0160)
643 [Adam Langley, Bodo Moeller]
645 *) Fix for the attack described in the paper "Recovering OpenSSL
646 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
647 by Yuval Yarom and Naomi Benger. Details can be obtained from:
648 http://eprint.iacr.org/2014/140
650 Thanks to Yuval Yarom and Naomi Benger for discovering this
651 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
652 [Yuval Yarom and Naomi Benger]
654 *) TLS pad extension: draft-agl-tls-padding-03
656 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
657 TLS client Hello record length value would otherwise be > 255 and
658 less that 512 pad with a dummy extension containing zeroes so it
659 is at least 512 bytes long.
661 [Adam Langley, Steve Henson]
663 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
665 *) Fix for TLS record tampering bug. A carefully crafted invalid
666 handshake could crash OpenSSL with a NULL pointer exception.
667 Thanks to Anton Johansson for reporting this issues.
670 *) Keep original DTLS digest and encryption contexts in retransmission
671 structures so we can use the previous session parameters if they need
672 to be resent. (CVE-2013-6450)
675 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
676 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
677 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
678 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
679 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
680 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
681 [Rob Stradling, Adam Langley]
683 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
685 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
686 supporting platforms or when small records were transferred.
687 [Andy Polyakov, Steve Henson]
689 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
691 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
693 This addresses the flaw in CBC record processing discovered by
694 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
695 at: http://www.isg.rhul.ac.uk/tls/
697 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
698 Security Group at Royal Holloway, University of London
699 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
700 Emilia Käsper for the initial patch.
702 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
704 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
705 ciphersuites which can be exploited in a denial of service attack.
706 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
707 and detecting this bug and to Wolfgang Ettlinger
708 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
712 *) Return an error when checking OCSP signatures when key is NULL.
713 This fixes a DoS attack. (CVE-2013-0166)
716 *) Make openssl verify return errors.
717 [Chris Palmer <palmer@google.com> and Ben Laurie]
719 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
720 the right response is stapled. Also change SSL_get_certificate()
721 so it returns the certificate actually sent.
722 See http://rt.openssl.org/Ticket/Display.html?id=2836.
723 [Rob Stradling <rob.stradling@comodo.com>]
725 *) Fix possible deadlock when decoding public keys.
728 *) Don't use TLS 1.0 record version number in initial client hello
732 Changes between 1.0.1b and 1.0.1c [10 May 2012]
734 *) Sanity check record length before skipping explicit IV in TLS
735 1.2, 1.1 and DTLS to fix DoS attack.
737 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
738 fuzzing as a service testing platform.
742 *) Initialise tkeylen properly when encrypting CMS messages.
743 Thanks to Solar Designer of Openwall for reporting this issue.
746 *) In FIPS mode don't try to use composite ciphers as they are not
750 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
752 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
753 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
754 mean any application compiled against OpenSSL 1.0.0 headers setting
755 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
756 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
757 0x10000000L Any application which was previously compiled against
758 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
759 will need to be recompiled as a result. Letting be results in
760 inability to disable specifically TLS 1.1 and in client context,
761 in unlike event, limit maximum offered version to TLS 1.0 [see below].
764 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
765 disable just protocol X, but all protocols above X *if* there are
766 protocols *below* X still enabled. In more practical terms it means
767 that if application wants to disable TLS1.0 in favor of TLS1.1 and
768 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
769 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
773 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
775 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
776 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
777 in CRYPTO_realloc_clean.
779 Thanks to Tavis Ormandy, Google Security Team, for discovering this
780 issue and to Adam Langley <agl@chromium.org> for fixing it.
782 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
784 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
787 *) Workarounds for some broken servers that "hang" if a client hello
788 record length exceeds 255 bytes.
790 1. Do not use record version number > TLS 1.0 in initial client
791 hello: some (but not all) hanging servers will now work.
792 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
793 the number of ciphers sent in the client hello. This should be
794 set to an even number, such as 50, for example by passing:
795 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
796 Most broken servers should now work.
797 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
798 TLS 1.2 client support entirely.
801 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
804 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
806 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
807 STRING form instead of a DigestInfo.
810 *) The format used for MDC2 RSA signatures is inconsistent between EVP
811 and the RSA_sign/RSA_verify functions. This was made more apparent when
812 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
813 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
814 the correct format in RSA_verify so both forms transparently work.
817 *) Some servers which support TLS 1.0 can choke if we initially indicate
818 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
819 encrypted premaster secret. As a workaround use the maximum pemitted
820 client version in client hello, this should keep such servers happy
821 and still work with previous versions of OpenSSL.
824 *) Add support for TLS/DTLS heartbeats.
825 [Robin Seggelmann <seggelmann@fh-muenster.de>]
827 *) Add support for SCTP.
828 [Robin Seggelmann <seggelmann@fh-muenster.de>]
830 *) Improved PRNG seeding for VOS.
831 [Paul Green <Paul.Green@stratus.com>]
833 *) Extensive assembler packs updates, most notably:
835 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
836 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
837 - x86_64: bit-sliced AES implementation;
838 - ARM: NEON support, contemporary platforms optimizations;
839 - s390x: z196 support;
840 - *: GHASH and GF(2^m) multiplication implementations;
844 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
845 (removal of unnecessary code)
846 [Peter Sylvester <peter.sylvester@edelweb.fr>]
848 *) Add TLS key material exporter from RFC 5705.
851 *) Add DTLS-SRTP negotiation from RFC 5764.
854 *) Add Next Protocol Negotiation,
855 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
856 disabled with a no-npn flag to config or Configure. Code donated
858 [Adam Langley <agl@google.com> and Ben Laurie]
860 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
861 NIST-P256, NIST-P521, with constant-time single point multiplication on
862 typical inputs. Compiler support for the nonstandard type __uint128_t is
863 required to use this (present in gcc 4.4 and later, for 64-bit builds).
864 Code made available under Apache License version 2.0.
866 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
867 line to include this in your build of OpenSSL, and run "make depend" (or
868 "make update"). This enables the following EC_METHODs:
870 EC_GFp_nistp224_method()
871 EC_GFp_nistp256_method()
872 EC_GFp_nistp521_method()
874 EC_GROUP_new_by_curve_name() will automatically use these (while
875 EC_GROUP_new_curve_GFp() currently prefers the more flexible
877 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
879 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
880 all platforms. Move ssize_t definition from e_os.h to the public
881 header file e_os2.h as it now appears in public header file cms.h
884 *) New -sigopt option to the ca, req and x509 utilities. Additional
885 signature parameters can be passed using this option and in
889 *) Add RSA PSS signing function. This will generate and set the
890 appropriate AlgorithmIdentifiers for PSS based on those in the
891 corresponding EVP_MD_CTX structure. No application support yet.
894 *) Support for companion algorithm specific ASN1 signing routines.
895 New function ASN1_item_sign_ctx() signs a pre-initialised
896 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
897 the appropriate parameters.
900 *) Add new algorithm specific ASN1 verification initialisation function
901 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
902 handling will be the same no matter what EVP_PKEY_METHOD is used.
903 Add a PSS handler to support verification of PSS signatures: checked
904 against a number of sample certificates.
907 *) Add signature printing for PSS. Add PSS OIDs.
908 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
910 *) Add algorithm specific signature printing. An individual ASN1 method
911 can now print out signatures instead of the standard hex dump.
913 More complex signatures (e.g. PSS) can print out more meaningful
914 information. Include DSA version that prints out the signature
918 *) Password based recipient info support for CMS library: implementing
922 *) Split password based encryption into PBES2 and PBKDF2 functions. This
923 neatly separates the code into cipher and PBE sections and is required
924 for some algorithms that split PBES2 into separate pieces (such as
928 *) Session-handling fixes:
929 - Fix handling of connections that are resuming with a session ID,
930 but also support Session Tickets.
931 - Fix a bug that suppressed issuing of a new ticket if the client
932 presented a ticket with an expired session.
933 - Try to set the ticket lifetime hint to something reasonable.
934 - Make tickets shorter by excluding irrelevant information.
935 - On the client side, don't ignore renewed tickets.
936 [Adam Langley, Bodo Moeller (Google)]
938 *) Fix PSK session representation.
941 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
943 This work was sponsored by Intel.
946 *) Add GCM support to TLS library. Some custom code is needed to split
947 the IV between the fixed (from PRF) and explicit (from TLS record)
948 portions. This adds all GCM ciphersuites supported by RFC5288 and
949 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
950 add a special AESGCM string for GCM only.
953 *) Expand range of ctrls for AES GCM. Permit setting invocation
954 field on decrypt and retrieval of invocation field only on encrypt.
957 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
958 As required by RFC5289 these ciphersuites cannot be used if for
959 versions of TLS earlier than 1.2.
962 *) For FIPS capable OpenSSL interpret a NULL default public key method
963 as unset and return the appopriate default but do *not* set the default.
964 This means we can return the appopriate method in applications that
965 swicth between FIPS and non-FIPS modes.
968 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
969 ENGINE is used then we cannot handle that in the FIPS module so we
970 keep original code iff non-FIPS operations are allowed.
973 *) Add -attime option to openssl utilities.
974 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
976 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
979 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
980 FIPS EC methods unconditionally for now.
983 *) New build option no-ec2m to disable characteristic 2 code.
986 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
987 all cases can be covered as some introduce binary incompatibilities.
990 *) Redirect RSA operations to FIPS module including keygen,
991 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
994 *) Add similar low level API blocking to ciphers.
997 *) Low level digest APIs are not approved in FIPS mode: any attempt
998 to use these will cause a fatal error. Applications that *really* want
999 to use them can use the private_* version instead.
1002 *) Redirect cipher operations to FIPS module for FIPS builds.
1005 *) Redirect digest operations to FIPS module for FIPS builds.
1008 *) Update build system to add "fips" flag which will link in fipscanister.o
1009 for static and shared library builds embedding a signature if needed.
1012 *) Output TLS supported curves in preference order instead of numerical
1013 order. This is currently hardcoded for the highest order curves first.
1014 This should be configurable so applications can judge speed vs strength.
1017 *) Add TLS v1.2 server support for client authentication.
1020 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1024 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1025 FIPS modules versions.
1028 *) Add TLS v1.2 client side support for client authentication. Keep cache
1029 of handshake records longer as we don't know the hash algorithm to use
1030 until after the certificate request message is received.
1033 *) Initial TLS v1.2 client support. Add a default signature algorithms
1034 extension including all the algorithms we support. Parse new signature
1035 format in client key exchange. Relax some ECC signing restrictions for
1036 TLS v1.2 as indicated in RFC5246.
1039 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1040 to new signature format when needed using client digest preference.
1041 All server ciphersuites should now work correctly in TLS v1.2. No client
1042 support yet and no support for client certificates.
1045 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1046 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1047 ciphersuites. At present only RSA key exchange ciphersuites work with
1048 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1049 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1050 and version checking.
1053 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1054 with this defined it will not be affected by any changes to ssl internal
1055 structures. Add several utility functions to allow openssl application
1056 to work with OPENSSL_NO_SSL_INTERN defined.
1060 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1062 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1065 *) Permit abbreviated handshakes when renegotiating using the function
1066 SSL_renegotiate_abbreviated().
1067 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1069 *) Add call to ENGINE_register_all_complete() to
1070 ENGINE_load_builtin_engines(), so some implementations get used
1071 automatically instead of needing explicit application support.
1074 *) Add support for TLS key exporter as described in RFC5705.
1075 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1077 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1078 a few changes are required:
1080 Add SSL_OP_NO_TLSv1_1 flag.
1081 Add TLSv1_1 methods.
1082 Update version checking logic to handle version 1.1.
1083 Add explicit IV handling (ported from DTLS code).
1084 Add command line options to s_client/s_server.
1087 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
1089 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1091 This addresses the flaw in CBC record processing discovered by
1092 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1093 at: http://www.isg.rhul.ac.uk/tls/
1095 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1096 Security Group at Royal Holloway, University of London
1097 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1098 Emilia Käsper for the initial patch.
1100 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1102 *) Return an error when checking OCSP signatures when key is NULL.
1103 This fixes a DoS attack. (CVE-2013-0166)
1106 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1107 the right response is stapled. Also change SSL_get_certificate()
1108 so it returns the certificate actually sent.
1109 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1110 (This is a backport)
1111 [Rob Stradling <rob.stradling@comodo.com>]
1113 *) Fix possible deadlock when decoding public keys.
1116 Changes between 1.0.0i and 1.0.0j [10 May 2012]
1118 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
1121 *) Sanity check record length before skipping explicit IV in DTLS
1124 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1125 fuzzing as a service testing platform.
1129 *) Initialise tkeylen properly when encrypting CMS messages.
1130 Thanks to Solar Designer of Openwall for reporting this issue.
1133 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
1135 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1136 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1137 in CRYPTO_realloc_clean.
1139 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1140 issue and to Adam Langley <agl@chromium.org> for fixing it.
1142 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1144 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1146 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1147 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1148 content decryption and always return the same error. Note: this attack
1149 needs on average 2^20 messages so it only affects automated senders. The
1150 old behaviour can be reenabled in the CMS code by setting the
1151 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1152 an MMA defence is not necessary.
1153 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1154 this issue. (CVE-2012-0884)
1157 *) Fix CVE-2011-4619: make sure we really are receiving a
1158 client hello before rejecting multiple SGC restarts. Thanks to
1159 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1162 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1164 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1165 Thanks to Antonio Martin, Enterprise Secure Access Research and
1166 Development, Cisco Systems, Inc. for discovering this bug and
1167 preparing a fix. (CVE-2012-0050)
1170 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1172 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1173 of the Vaudenay padding oracle attack on CBC mode encryption
1174 which enables an efficient plaintext recovery attack against
1175 the OpenSSL implementation of DTLS. Their attack exploits timing
1176 differences arising during decryption processing. A research
1177 paper describing this attack can be found at:
1178 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1179 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1180 Security Group at Royal Holloway, University of London
1181 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1182 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1183 for preparing the fix. (CVE-2011-4108)
1184 [Robin Seggelmann, Michael Tuexen]
1186 *) Clear bytes used for block padding of SSL 3.0 records.
1188 [Adam Langley (Google)]
1190 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1191 Kadianakis <desnacked@gmail.com> for discovering this issue and
1192 Adam Langley for preparing the fix. (CVE-2011-4619)
1193 [Adam Langley (Google)]
1195 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1196 [Andrey Kulikov <amdeich@gmail.com>]
1198 *) Prevent malformed RFC3779 data triggering an assertion failure.
1199 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1200 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1201 [Rob Austein <sra@hactrn.net>]
1203 *) Improved PRNG seeding for VOS.
1204 [Paul Green <Paul.Green@stratus.com>]
1206 *) Fix ssl_ciph.c set-up race.
1207 [Adam Langley (Google)]
1209 *) Fix spurious failures in ecdsatest.c.
1210 [Emilia Käsper (Google)]
1212 *) Fix the BIO_f_buffer() implementation (which was mixing different
1213 interpretations of the '..._len' fields).
1214 [Adam Langley (Google)]
1216 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1217 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1218 threads won't reuse the same blinding coefficients.
1220 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1221 lock to call BN_BLINDING_invert_ex, and avoids one use of
1222 BN_BLINDING_update for each BN_BLINDING structure (previously,
1223 the last update always remained unused).
1224 [Emilia Käsper (Google)]
1226 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1227 [Bob Buckholz (Google)]
1229 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1231 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1232 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1233 [Kaspar Brand <ossl@velox.ch>]
1235 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1236 for multi-threaded use of ECDH. (CVE-2011-3210)
1237 [Adam Langley (Google)]
1239 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1242 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1243 signature public key algorithm by using OID xref utilities instead.
1244 Before this you could only use some ECC ciphersuites with SHA1 only.
1247 *) Add protection against ECDSA timing attacks as mentioned in the paper
1248 by Billy Bob Brumley and Nicola Tuveri, see:
1250 http://eprint.iacr.org/2011/232.pdf
1252 [Billy Bob Brumley and Nicola Tuveri]
1254 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1256 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1257 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1259 *) Fix bug in string printing code: if *any* escaping is enabled we must
1260 escape the escape character (backslash) or the resulting string is
1264 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1266 *) Disable code workaround for ancient and obsolete Netscape browsers
1267 and servers: an attacker can use it in a ciphersuite downgrade attack.
1268 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1271 *) Fixed J-PAKE implementation error, originally discovered by
1272 Sebastien Martini, further info and confirmation from Stefan
1273 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1276 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1278 *) Fix extension code to avoid race conditions which can result in a buffer
1279 overrun vulnerability: resumed sessions must not be modified as they can
1280 be shared by multiple threads. CVE-2010-3864
1283 *) Fix WIN32 build system to correctly link an ENGINE directory into
1287 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1289 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1291 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1293 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1295 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1296 context. The operation can be customised via the ctrl mechanism in
1297 case ENGINEs want to include additional functionality.
1300 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1303 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1304 output hashes compatible with older versions of OpenSSL.
1305 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1307 *) Fix compression algorithm handling: if resuming a session use the
1308 compression algorithm of the resumed session instead of determining
1309 it from client hello again. Don't allow server to change algorithm.
1312 *) Add load_crls() function to apps tidying load_certs() too. Add option
1313 to verify utility to allow additional CRLs to be included.
1316 *) Update OCSP request code to permit adding custom headers to the request:
1317 some responders need this.
1320 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1322 [Julia Lawall <julia@diku.dk>]
1324 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1325 needlessly dereferenced structures, used obsolete functions and
1326 didn't handle all updated verify codes correctly.
1329 *) Disable MD2 in the default configuration.
1332 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1333 indicate the initial BIO being pushed or popped. This makes it possible
1334 to determine whether the BIO is the one explicitly called or as a result
1335 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1336 it handles reference counts correctly and doesn't zero out the I/O bio
1337 when it is not being explicitly popped. WARNING: applications which
1338 included workarounds for the old buggy behaviour will need to be modified
1339 or they could free up already freed BIOs.
1342 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1343 renaming to all platforms (within the 0.9.8 branch, this was
1344 done conditionally on Netware platforms to avoid a name clash).
1345 [Guenter <lists@gknw.net>]
1347 *) Add ECDHE and PSK support to DTLS.
1348 [Michael Tuexen <tuexen@fh-muenster.de>]
1350 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1354 *) Add "missing" function EVP_MD_flags() (without this the only way to
1355 retrieve a digest flags is by accessing the structure directly. Update
1356 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1357 or cipher is registered as in the "from" argument. Print out all
1358 registered digests in the dgst usage message instead of manually
1359 attempting to work them out.
1362 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1363 this allows the use of compression and extensions. Change default cipher
1364 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1365 by default unless an application cipher string requests it.
1368 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1369 key ids to find matching certificates and keys but some PKCS#12 files
1370 don't follow the (somewhat unwritten) rules and this strategy fails.
1371 Now just gather all certificates together and the first private key
1372 then look for the first certificate that matches the key.
1375 *) Support use of registered digest and cipher names for dgst and cipher
1376 commands instead of having to add each one as a special case. So now
1383 openssl dgst -sha256 foo
1385 and this works for ENGINE based algorithms too.
1389 *) Update Gost ENGINE to support parameter files.
1390 [Victor B. Wagner <vitus@cryptocom.ru>]
1392 *) Support GeneralizedTime in ca utility.
1393 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1395 *) Enhance the hash format used for certificate directory links. The new
1396 form uses the canonical encoding (meaning equivalent names will work
1397 even if they aren't identical) and uses SHA1 instead of MD5. This form
1398 is incompatible with the older format and as a result c_rehash should
1399 be used to rebuild symbolic links.
1402 *) Make PKCS#8 the default write format for private keys, replacing the
1403 traditional format. This form is standardised, more secure and doesn't
1404 include an implicit MD5 dependency.
1407 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1408 committed to OpenSSL should pass this lot as a minimum.
1411 *) Add session ticket override functionality for use by EAP-FAST.
1412 [Jouni Malinen <j@w1.fi>]
1414 *) Modify HMAC functions to return a value. Since these can be implemented
1415 in an ENGINE errors can occur.
1418 *) Type-checked OBJ_bsearch_ex.
1421 *) Type-checked OBJ_bsearch. Also some constification necessitated
1422 by type-checking. Still to come: TXT_DB, bsearch(?),
1423 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1427 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1428 seconds to a tm structure directly, instead of going through OS
1429 specific date routines. This avoids any issues with OS routines such
1430 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1431 and X509_time_adj_ex() to cover the extended range. The existing
1432 X509_time_adj() is still usable and will no longer have any date issues.
1435 *) Delta CRL support. New use deltas option which will attempt to locate
1436 and search any appropriate delta CRLs available.
1438 This work was sponsored by Google.
1441 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1442 code and add additional score elements. Validate alternate CRL paths
1443 as part of the CRL checking and indicate a new error "CRL path validation
1444 error" in this case. Applications wanting additional details can use
1445 the verify callback and check the new "parent" field. If this is not
1446 NULL CRL path validation is taking place. Existing applications wont
1447 see this because it requires extended CRL support which is off by
1450 This work was sponsored by Google.
1453 *) Support for freshest CRL extension.
1455 This work was sponsored by Google.
1458 *) Initial indirect CRL support. Currently only supported in the CRLs
1459 passed directly and not via lookup. Process certificate issuer
1460 CRL entry extension and lookup CRL entries by bother issuer name
1461 and serial number. Check and process CRL issuer entry in IDP extension.
1463 This work was sponsored by Google.
1466 *) Add support for distinct certificate and CRL paths. The CRL issuer
1467 certificate is validated separately in this case. Only enabled if
1468 an extended CRL support flag is set: this flag will enable additional
1469 CRL functionality in future.
1471 This work was sponsored by Google.
1474 *) Add support for policy mappings extension.
1476 This work was sponsored by Google.
1479 *) Fixes to pathlength constraint, self issued certificate handling,
1480 policy processing to align with RFC3280 and PKITS tests.
1482 This work was sponsored by Google.
1485 *) Support for name constraints certificate extension. DN, email, DNS
1486 and URI types are currently supported.
1488 This work was sponsored by Google.
1491 *) To cater for systems that provide a pointer-based thread ID rather
1492 than numeric, deprecate the current numeric thread ID mechanism and
1493 replace it with a structure and associated callback type. This
1494 mechanism allows a numeric "hash" to be extracted from a thread ID in
1495 either case, and on platforms where pointers are larger than 'long',
1496 mixing is done to help ensure the numeric 'hash' is usable even if it
1497 can't be guaranteed unique. The default mechanism is to use "&errno"
1498 as a pointer-based thread ID to distinguish between threads.
1500 Applications that want to provide their own thread IDs should now use
1501 CRYPTO_THREADID_set_callback() to register a callback that will call
1502 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1504 Note that ERR_remove_state() is now deprecated, because it is tied
1505 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1506 to free the current thread's error state should be replaced by
1507 ERR_remove_thread_state(NULL).
1509 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1510 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1511 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1512 application was previously providing a numeric thread callback that
1513 was inappropriate for distinguishing threads, then uniqueness might
1514 have been obtained with &errno that happened immediately in the
1515 intermediate development versions of OpenSSL; this is no longer the
1516 case, the numeric thread callback will now override the automatic use
1518 [Geoff Thorpe, with help from Bodo Moeller]
1520 *) Initial support for different CRL issuing certificates. This covers a
1521 simple case where the self issued certificates in the chain exist and
1522 the real CRL issuer is higher in the existing chain.
1524 This work was sponsored by Google.
1527 *) Removed effectively defunct crypto/store from the build.
1530 *) Revamp of STACK to provide stronger type-checking. Still to come:
1531 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1532 ASN1_STRING, CONF_VALUE.
1535 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1536 RAM on SSL connections. This option can save about 34k per idle SSL.
1539 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1540 STACK, TXT_DB, bsearch, qsort.
1543 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1544 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1545 support for data, signedData, compressedData, digestedData and
1546 encryptedData, envelopedData types included. Scripts to check against
1547 RFC4134 examples draft and interop and consistency checks of many
1548 content types and variants.
1551 *) Add options to enc utility to support use of zlib compression BIO.
1554 *) Extend mk1mf to support importing of options and assembly language
1555 files from Configure script, currently only included in VC-WIN32.
1556 The assembly language rules can now optionally generate the source
1557 files from the associated perl scripts.
1560 *) Implement remaining functionality needed to support GOST ciphersuites.
1561 Interop testing has been performed using CryptoPro implementations.
1562 [Victor B. Wagner <vitus@cryptocom.ru>]
1564 *) s390x assembler pack.
1567 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1571 *) Implement Opaque PRF Input TLS extension as specified in
1572 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1573 official specification yet and no extension type assignment by
1574 IANA exists, this extension (for now) will have to be explicitly
1575 enabled when building OpenSSL by providing the extension number
1576 to use. For example, specify an option
1578 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1580 to the "config" or "Configure" script to enable the extension,
1581 assuming extension number 0x9527 (which is a completely arbitrary
1582 and unofficial assignment based on the MD5 hash of the Internet
1583 Draft). Note that by doing so, you potentially lose
1584 interoperability with other TLS implementations since these might
1585 be using the same extension number for other purposes.
1587 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1588 opaque PRF input value to use in the handshake. This will create
1589 an interal copy of the length-'len' string at 'src', and will
1590 return non-zero for success.
1592 To get more control and flexibility, provide a callback function
1595 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1596 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1600 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1603 Callback function 'cb' will be called in handshakes, and is
1604 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1605 Argument 'arg' is for application purposes (the value as given to
1606 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1607 be provided to the callback function). The callback function
1608 has to return non-zero to report success: usually 1 to use opaque
1609 PRF input just if possible, or 2 to enforce use of the opaque PRF
1610 input. In the latter case, the library will abort the handshake
1611 if opaque PRF input is not successfully negotiated.
1613 Arguments 'peerinput' and 'len' given to the callback function
1614 will always be NULL and 0 in the case of a client. A server will
1615 see the client's opaque PRF input through these variables if
1616 available (NULL and 0 otherwise). Note that if the server
1617 provides an opaque PRF input, the length must be the same as the
1618 length of the client's opaque PRF input.
1620 Note that the callback function will only be called when creating
1621 a new session (session resumption can resume whatever was
1622 previously negotiated), and will not be called in SSL 2.0
1623 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1624 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1625 for applications that need to enforce opaque PRF input.
1629 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1632 [Victor B. Wagner <vitus@cryptocom.ru>]
1634 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1635 RFC4507bis. The encrypted ticket format is an encrypted encoded
1636 SSL_SESSION structure, that way new session features are automatically
1639 If a client application caches session in an SSL_SESSION structure
1640 support is transparent because tickets are now stored in the encoded
1643 The SSL_CTX structure automatically generates keys for ticket
1644 protection in servers so again support should be possible
1645 with no application modification.
1647 If a client or server wishes to disable RFC4507 support then the option
1648 SSL_OP_NO_TICKET can be set.
1650 Add a TLS extension debugging callback to allow the contents of any client
1651 or server extensions to be examined.
1653 This work was sponsored by Google.
1656 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1657 OpenSSL should now compile cleanly on gcc 4.2
1658 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1660 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1661 support including streaming MAC support: this is required for GOST
1662 ciphersuite support.
1663 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1665 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1666 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1667 to output in BER and PEM format.
1670 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1671 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1672 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1673 ENGINE support for HMAC keys which are unextractable. New -mac and
1674 -macopt options to dgst utility.
1677 *) New option -sigopt to dgst utility. Update dgst to use
1678 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1679 alternative signing paramaters such as X9.31 or PSS in the dgst
1683 *) Change ssl_cipher_apply_rule(), the internal function that does
1684 the work each time a ciphersuite string requests enabling
1685 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1686 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1687 the order of disabled ciphersuites such that those ciphersuites
1688 that most recently went from enabled to disabled not only stay
1689 in order with respect to each other, but also have higher priority
1690 than other disabled ciphersuites the next time ciphersuites are
1693 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1694 the same ciphersuites as with "HIGH" alone, but in a specific
1695 order where the PSK ciphersuites come first (since they are the
1696 most recently disabled ciphersuites when "HIGH" is parsed).
1698 Also, change ssl_create_cipher_list() (using this new
1699 funcionality) such that between otherwise identical
1700 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1704 *) Change ssl_create_cipher_list() so that it automatically
1705 arranges the ciphersuites in reasonable order before starting
1706 to process the rule string. Thus, the definition for "DEFAULT"
1707 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1708 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1709 This makes it much easier to arrive at a reasonable default order
1710 in applications for which anonymous ciphers are OK (meaning
1711 that you can't actually use DEFAULT).
1712 [Bodo Moeller; suggested by Victor Duchovni]
1714 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1715 processing) into multiple integers instead of setting
1716 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1717 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1718 (These masks as well as the individual bit definitions are hidden
1719 away into the non-exported interface ssl/ssl_locl.h, so this
1720 change to the definition of the SSL_CIPHER structure shouldn't
1721 affect applications.) This give us more bits for each of these
1722 categories, so there is no longer a need to coagulate AES128 and
1723 AES256 into a single algorithm bit, and to coagulate Camellia128
1724 and Camellia256 into a single algorithm bit, which has led to all
1727 Thus, among other things, the kludge introduced in 0.9.7m and
1728 0.9.8e for masking out AES256 independently of AES128 or masking
1729 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1731 With the change, we also introduce new ciphersuite aliases that
1732 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1736 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1737 Use the leftmost N bytes of the signature input if the input is
1738 larger than the prime q (with N being the size in bytes of q).
1741 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1742 it yet and it is largely untested.
1745 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1748 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1749 some compilers (gcc 4.2 and later) reject their use. Safestack is
1750 reimplemented. Update ASN1 to avoid use of legacy functions.
1753 *) Win32/64 targets are linked with Winsock2.
1756 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1757 to external functions. This can be used to increase CRL handling
1758 efficiency especially when CRLs are very large by (for example) storing
1759 the CRL revoked certificates in a database.
1762 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1763 new CRLs added to a directory can be used. New command line option
1764 -verify_return_error to s_client and s_server. This causes real errors
1765 to be returned by the verify callback instead of carrying on no matter
1766 what. This reflects the way a "real world" verify callback would behave.
1769 *) GOST engine, supporting several GOST algorithms and public key formats.
1770 Kindly donated by Cryptocom.
1773 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1774 partitioned by DP are handled but no indirect CRL or reason partitioning
1775 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1776 selected via a scoring technique which handles IDP and AKID in CRLs.
1779 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1780 will ultimately be used for all verify operations: this will remove the
1781 X509_STORE dependency on certificate verification and allow alternative
1782 lookup methods. X509_STORE based implementations of these two callbacks.
1785 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1786 Modify get_crl() to find a valid (unexpired) CRL if possible.
1789 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1790 this would be called X509_CRL_cmp() but that name is already used by
1791 a function that just compares CRL issuer names. Cache several CRL
1792 extensions in X509_CRL structure and cache CRLDP in X509.
1795 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1796 this maps equivalent X509_NAME structures into a consistent structure.
1797 Name comparison can then be performed rapidly using memcmp().
1800 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1804 *) Allow digests to supply their own micalg string for S/MIME type using
1805 the ctrl EVP_MD_CTRL_MICALG.
1808 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1809 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1810 ctrl. It can then customise the structure before and/or after signing
1814 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1815 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1816 to free up any added signature OIDs.
1819 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1820 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1821 digest and cipher tables. New options added to openssl utility:
1822 list-message-digest-algorithms and list-cipher-algorithms.
1825 *) Change the array representation of binary polynomials: the list
1826 of degrees of non-zero coefficients is now terminated with -1.
1827 Previously it was terminated with 0, which was also part of the
1828 value; thus, the array representation was not applicable to
1829 polynomials where t^0 has coefficient zero. This change makes
1830 the array representation useful in a more general context.
1833 *) Various modifications and fixes to SSL/TLS cipher string
1834 handling. For ECC, the code now distinguishes between fixed ECDH
1835 with RSA certificates on the one hand and with ECDSA certificates
1836 on the other hand, since these are separate ciphersuites. The
1837 unused code for Fortezza ciphersuites has been removed.
1839 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1840 (not "ECDHE"). For consistency with the code for DH
1841 certificates, use of ECDH certificates is now considered ECDH
1842 authentication, not RSA or ECDSA authentication (the latter is
1843 merely the CA's signing algorithm and not actively used in the
1846 The temporary ciphersuite alias "ECCdraft" is no longer
1847 available, and ECC ciphersuites are no longer excluded from "ALL"
1848 and "DEFAULT". The following aliases now exist for RFC 4492
1849 ciphersuites, most of these by analogy with the DH case:
1851 kECDHr - ECDH cert, signed with RSA
1852 kECDHe - ECDH cert, signed with ECDSA
1853 kECDH - ECDH cert (signed with either RSA or ECDSA)
1854 kEECDH - ephemeral ECDH
1855 ECDH - ECDH cert or ephemeral ECDH
1861 AECDH - anonymous ECDH
1862 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1866 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1867 Use correct micalg parameters depending on digest(s) in signed message.
1870 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1871 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1874 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1875 an engine to register a method. Add ENGINE lookups for methods and
1876 functional reference processing.
1879 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1880 EVP_{Sign,Verify}* which allow an application to customise the signature
1884 *) New -resign option to smime utility. This adds one or more signers
1885 to an existing PKCS#7 signedData structure. Also -md option to use an
1886 alternative message digest algorithm for signing.
1889 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1890 create PKCS7 structures containing multiple signers. Update smime
1891 application to support multiple signers.
1894 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1898 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1899 Reorganize PBE internals to lookup from a static table using NIDs,
1900 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1901 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1902 PRF which will be automatically used with PBES2.
1905 *) Replace the algorithm specific calls to generate keys in "req" with the
1909 *) Update PKCS#7 enveloped data routines to use new API. This is now
1910 supported by any public key method supporting the encrypt operation. A
1911 ctrl is added to allow the public key algorithm to examine or modify
1912 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1916 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1917 a default digest type to use. In most cases this will be SHA1 but some
1918 algorithms (such as GOST) need to specify an alternative digest. The
1919 return value indicates how strong the prefernce is 1 means optional and
1920 2 is mandatory (that is it is the only supported type). Modify
1921 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1922 use the default md. Update openssl utilities to use the default digest
1923 type for signing if it is not explicitly indicated.
1926 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1927 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1928 signing method from the key type. This effectively removes the link
1929 between digests and public key types.
1932 *) Add an OID cross reference table and utility functions. Its purpose is to
1933 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1934 rsaEncryption. This will allow some of the algorithm specific hackery
1935 needed to use the correct OID to be removed.
1938 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1939 structures for PKCS7_sign(). They are now set up by the relevant public
1943 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1946 *) Add support for key derivation (agreement) in the API, DH method and
1950 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1951 public and private key formats. As a side effect these add additional
1952 command line functionality not previously available: DSA signatures can be
1953 generated and verified using pkeyutl and DH key support and generation in
1958 [Oliver Tappe <zooey@hirschkaefer.de>]
1960 *) New make target "install_html_docs" installs HTML renditions of the
1962 [Oliver Tappe <zooey@hirschkaefer.de>]
1964 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1965 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1966 support key and parameter generation and add initial key generation
1967 functionality for RSA.
1970 *) Add functions for main EVP_PKEY_method operations. The undocumented
1971 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1972 EVP_PKEY_{encrypt,decrypt}_old.
1975 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1976 key API, doesn't do much yet.
1979 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1980 public key algorithms. New option to openssl utility:
1981 "list-public-key-algorithms" to print out info.
1984 *) Implement the Supported Elliptic Curves Extension for
1985 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1988 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1989 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1992 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1993 utilities such as rsa, dsa, dsaparam etc except they process any key
1997 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1998 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1999 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2003 *) Initial support for pluggable public key ASN1.
2004 De-spaghettify the public key ASN1 handling. Move public and private
2005 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2006 algorithm specific handling to a single module within the relevant
2007 algorithm directory. Add functions to allow (near) opaque processing
2008 of public and private key structures.
2011 *) Implement the Supported Point Formats Extension for
2012 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2015 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2016 for the psk identity [hint] and the psk callback functions to the
2017 SSL_SESSION, SSL and SSL_CTX structure.
2020 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2024 SSL_CTX_use_psk_identity_hint
2025 SSL_get_psk_identity_hint
2026 SSL_get_psk_identity
2027 SSL_use_psk_identity_hint
2029 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2031 *) Add RFC 3161 compliant time stamp request creation, response generation
2032 and response verification functionality.
2033 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2035 *) Add initial support for TLS extensions, specifically for the server_name
2036 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2037 have new members for a host name. The SSL data structure has an
2038 additional member SSL_CTX *initial_ctx so that new sessions can be
2039 stored in that context to allow for session resumption, even after the
2040 SSL has been switched to a new SSL_CTX in reaction to a client's
2041 server_name extension.
2043 New functions (subject to change):
2045 SSL_get_servername()
2046 SSL_get_servername_type()
2049 New CTRL codes and macros (subject to change):
2051 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2052 - SSL_CTX_set_tlsext_servername_callback()
2053 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2054 - SSL_CTX_set_tlsext_servername_arg()
2055 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2057 openssl s_client has a new '-servername ...' option.
2059 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2060 '-key2 ...', '-servername_fatal' (subject to change). This allows
2061 testing the HostName extension for a specific single host name ('-cert'
2062 and '-key' remain fallbacks for handshakes without HostName
2063 negotiation). If the unrecogninzed_name alert has to be sent, this by
2064 default is a warning; it becomes fatal with the '-servername_fatal'
2067 [Peter Sylvester, Remy Allais, Christophe Renou]
2069 *) Whirlpool hash implementation is added.
2072 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2073 bn(64,32). Because of instruction set limitations it doesn't have
2074 any negative impact on performance. This was done mostly in order
2075 to make it possible to share assembler modules, such as bn_mul_mont
2076 implementations, between 32- and 64-bit builds without hassle.
2079 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2080 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2084 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2085 dedicated Montgomery multiplication procedure, is introduced.
2086 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2087 "64-bit" performance on certain 32-bit targets.
2090 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2091 in SSL structures. New SSL ctrl to set maximum send fragment size.
2092 Save memory by seeting the I/O buffer sizes dynamically instead of
2093 using the maximum available value.
2096 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2097 in addition to the text details.
2100 *) Very, very preliminary EXPERIMENTAL support for printing of general
2101 ASN1 structures. This currently produces rather ugly output and doesn't
2102 handle several customised structures at all.
2105 *) Integrated support for PVK file format and some related formats such
2106 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2107 these in the 'rsa' and 'dsa' utilities.
2110 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2113 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2114 place for the (very old) "NETSCAPE" format certificates which are now
2115 handled using new ASN1 code equivalents.
2118 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2119 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2120 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2123 *) Modify CRL distribution points extension code to print out previously
2124 unsupported fields. Enhance extension setting code to allow setting of
2128 *) Add print and set support for Issuing Distribution Point CRL extension.
2131 *) Change 'Configure' script to enable Camellia by default.
2134 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
2136 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2138 This addresses the flaw in CBC record processing discovered by
2139 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2140 at: http://www.isg.rhul.ac.uk/tls/
2142 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2143 Security Group at Royal Holloway, University of London
2144 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2145 Emilia Käsper for the initial patch.
2147 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2149 *) Return an error when checking OCSP signatures when key is NULL.
2150 This fixes a DoS attack. (CVE-2013-0166)
2153 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2154 the right response is stapled. Also change SSL_get_certificate()
2155 so it returns the certificate actually sent.
2156 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2157 (This is a backport)
2158 [Rob Stradling <rob.stradling@comodo.com>]
2160 *) Fix possible deadlock when decoding public keys.
2163 Changes between 0.9.8w and 0.9.8x [10 May 2012]
2165 *) Sanity check record length before skipping explicit IV in DTLS
2168 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2169 fuzzing as a service testing platform.
2173 *) Initialise tkeylen properly when encrypting CMS messages.
2174 Thanks to Solar Designer of Openwall for reporting this issue.
2177 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
2179 *) The fix for CVE-2012-2110 did not take into account that the
2180 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
2181 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
2182 rejecting negative len parameter. (CVE-2012-2131)
2183 [Tomas Hoger <thoger@redhat.com>]
2185 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
2187 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2188 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2189 in CRYPTO_realloc_clean.
2191 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2192 issue and to Adam Langley <agl@chromium.org> for fixing it.
2194 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2196 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
2198 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2199 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2200 content decryption and always return the same error. Note: this attack
2201 needs on average 2^20 messages so it only affects automated senders. The
2202 old behaviour can be reenabled in the CMS code by setting the
2203 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2204 an MMA defence is not necessary.
2205 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2206 this issue. (CVE-2012-0884)
2209 *) Fix CVE-2011-4619: make sure we really are receiving a
2210 client hello before rejecting multiple SGC restarts. Thanks to
2211 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2214 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
2216 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2217 Thanks to Antonio Martin, Enterprise Secure Access Research and
2218 Development, Cisco Systems, Inc. for discovering this bug and
2219 preparing a fix. (CVE-2012-0050)
2222 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
2224 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2225 of the Vaudenay padding oracle attack on CBC mode encryption
2226 which enables an efficient plaintext recovery attack against
2227 the OpenSSL implementation of DTLS. Their attack exploits timing
2228 differences arising during decryption processing. A research
2229 paper describing this attack can be found at:
2230 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2231 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2232 Security Group at Royal Holloway, University of London
2233 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2234 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2235 for preparing the fix. (CVE-2011-4108)
2236 [Robin Seggelmann, Michael Tuexen]
2238 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
2239 [Ben Laurie, Kasper <ekasper@google.com>]
2241 *) Clear bytes used for block padding of SSL 3.0 records.
2243 [Adam Langley (Google)]
2245 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2246 Kadianakis <desnacked@gmail.com> for discovering this issue and
2247 Adam Langley for preparing the fix. (CVE-2011-4619)
2248 [Adam Langley (Google)]
2250 *) Prevent malformed RFC3779 data triggering an assertion failure.
2251 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2252 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2253 [Rob Austein <sra@hactrn.net>]
2255 *) Fix ssl_ciph.c set-up race.
2256 [Adam Langley (Google)]
2258 *) Fix spurious failures in ecdsatest.c.
2259 [Emilia Käsper (Google)]
2261 *) Fix the BIO_f_buffer() implementation (which was mixing different
2262 interpretations of the '..._len' fields).
2263 [Adam Langley (Google)]
2265 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2266 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2267 threads won't reuse the same blinding coefficients.
2269 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2270 lock to call BN_BLINDING_invert_ex, and avoids one use of
2271 BN_BLINDING_update for each BN_BLINDING structure (previously,
2272 the last update always remained unused).
2273 [Emilia Käsper (Google)]
2275 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2276 for multi-threaded use of ECDH.
2277 [Adam Langley (Google)]
2279 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2282 *) Add protection against ECDSA timing attacks as mentioned in the paper
2283 by Billy Bob Brumley and Nicola Tuveri, see:
2285 http://eprint.iacr.org/2011/232.pdf
2287 [Billy Bob Brumley and Nicola Tuveri]
2289 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
2291 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2292 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2294 *) Fix bug in string printing code: if *any* escaping is enabled we must
2295 escape the escape character (backslash) or the resulting string is
2299 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
2301 *) Disable code workaround for ancient and obsolete Netscape browsers
2302 and servers: an attacker can use it in a ciphersuite downgrade attack.
2303 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2306 *) Fixed J-PAKE implementation error, originally discovered by
2307 Sebastien Martini, further info and confirmation from Stefan
2308 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2311 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
2313 *) Fix extension code to avoid race conditions which can result in a buffer
2314 overrun vulnerability: resumed sessions must not be modified as they can
2315 be shared by multiple threads. CVE-2010-3864
2318 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
2321 *) Don't reencode certificate when calculating signature: cache and use
2322 the original encoding instead. This makes signature verification of
2323 some broken encodings work correctly.
2326 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
2327 is also one of the inputs.
2328 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2330 *) Don't repeatedly append PBE algorithms to table if they already exist.
2331 Sort table on each new add. This effectively makes the table read only
2332 after all algorithms are added and subsequent calls to PKCS12_pbe_add
2336 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
2338 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
2341 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
2342 access or freeing data twice (CVE-2010-0742)
2343 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
2345 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
2346 common in certificates and some applications which only call
2347 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
2351 Reduce copying into .apps and .test in makevms.com
2352 Don't try to use blank CA certificate in CA.com
2353 Allow use of C files from original directories in maketests.com
2354 [Steven M. Schweda" <sms@antinode.info>]
2356 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2358 *) When rejecting SSL/TLS records due to an incorrect version number, never
2359 update s->server with a new major version number. As of
2360 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2361 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2362 the previous behavior could result in a read attempt at NULL when
2363 receiving specific incorrect SSL/TLS records once record payload
2364 protection is active. (CVE-2010-0740)
2365 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2367 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2368 could be crashed if the relevant tables were not present (e.g. chrooted).
2369 [Tomas Hoger <thoger@redhat.com>]
2371 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2373 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2374 [Martin Olsson, Neel Mehta]
2376 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2377 accommodate for stack sorting, always a write lock!).
2380 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2381 excessive delays in the RAND_poll(): over a minute. As a workaround
2382 include a time check in the inner Heap32Next loop too.
2385 *) The code that handled flushing of data in SSL/TLS originally used the
2386 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2387 the problem outlined in PR#1949. The fix suggested there however can
2388 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2389 of Apache). So instead simplify the code to flush unconditionally.
2390 This should be fine since flushing with no data to flush is a no op.
2393 *) Handle TLS versions 2.0 and later properly and correctly use the
2394 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2395 off ancient servers have a habit of sticking around for a while...
2398 *) Modify compression code so it frees up structures without using the
2399 ex_data callbacks. This works around a problem where some applications
2400 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2401 restarting) then use compression (e.g. SSL with compression) later.
2402 This results in significant per-connection memory leaks and
2403 has caused some security issues including CVE-2008-1678 and
2407 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2408 change when encrypting or decrypting.
2411 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2412 connect and renegotiate with servers which do not support RI.
2413 Until RI is more widely deployed this option is enabled by default.
2416 *) Add "missing" ssl ctrls to clear options and mode.
2419 *) If client attempts to renegotiate and doesn't support RI respond with
2420 a no_renegotiation alert as required by RFC5746. Some renegotiating
2421 TLS clients will continue a connection gracefully when they receive
2422 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2423 waiting for a server hello which it will never receive. Now we treat a
2424 received no_renegotiation alert as a fatal error. This is because
2425 applications requesting a renegotiation might well expect it to succeed
2426 and would have no code in place to handle the server denying it so the
2427 only safe thing to do is to terminate the connection.
2430 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2431 peer supports secure renegotiation and 0 otherwise. Print out peer
2432 renegotiation support in s_client/s_server.
2435 *) Replace the highly broken and deprecated SPKAC certification method with
2436 the updated NID creation version. This should correctly handle UTF8.
2439 *) Implement RFC5746. Re-enable renegotiation but require the extension
2440 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2441 turns out to be a bad idea. It has been replaced by
2442 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2443 SSL_CTX_set_options(). This is really not recommended unless you
2444 know what you are doing.
2445 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2447 *) Fixes to stateless session resumption handling. Use initial_ctx when
2448 issuing and attempting to decrypt tickets in case it has changed during
2449 servername handling. Use a non-zero length session ID when attempting
2450 stateless session resumption: this makes it possible to determine if
2451 a resumption has occurred immediately after receiving server hello
2452 (several places in OpenSSL subtly assume this) instead of later in
2456 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2457 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2458 fixes for a few places where the return code is not checked
2460 [Julia Lawall <julia@diku.dk>]
2462 *) Add --strict-warnings option to Configure script to include devteam
2463 warnings in other configurations.
2466 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2467 makes it possible to install openssl libraries in locations which
2468 have names other than "lib", for example "/usr/lib64" which some
2470 [Steve Henson, based on patch from Jeremy Utley]
2472 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2473 X690 8.9.12 and can produce some misleading textual output of OIDs.
2474 [Steve Henson, reported by Dan Kaminsky]
2476 *) Delete MD2 from algorithm tables. This follows the recommendation in
2477 several standards that it is not used in new applications due to
2478 several cryptographic weaknesses. For binary compatibility reasons
2479 the MD2 API is still compiled in by default.
2482 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2486 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2487 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2489 [Guenter <lists@gknw.net>]
2491 *) Fix the server certificate chain building code to use X509_verify_cert(),
2492 it used to have an ad-hoc builder which was unable to cope with anything
2493 other than a simple chain.
2494 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2496 *) Don't check self signed certificate signatures in X509_verify_cert()
2497 by default (a flag can override this): it just wastes time without
2498 adding any security. As a useful side effect self signed root CAs
2499 with non-FIPS digests are now usable in FIPS mode.
2502 *) In dtls1_process_out_of_seq_message() the check if the current message
2503 is already buffered was missing. For every new message was memory
2504 allocated, allowing an attacker to perform an denial of service attack
2505 with sending out of seq handshake messages until there is no memory
2506 left. Additionally every future messege was buffered, even if the
2507 sequence number made no sense and would be part of another handshake.
2508 So only messages with sequence numbers less than 10 in advance will be
2509 buffered. (CVE-2009-1378)
2510 [Robin Seggelmann, discovered by Daniel Mentz]
2512 *) Records are buffered if they arrive with a future epoch to be
2513 processed after finishing the corresponding handshake. There is
2514 currently no limitation to this buffer allowing an attacker to perform
2515 a DOS attack with sending records with future epochs until there is no
2516 memory left. This patch adds the pqueue_size() function to detemine
2517 the size of a buffer and limits the record buffer to 100 entries.
2519 [Robin Seggelmann, discovered by Daniel Mentz]
2521 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2522 parent structure is freed. (CVE-2009-1379)
2525 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2526 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2529 [Ilya O. <vrghost@gmail.com>]
2531 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2533 *) Disable renegotiation completely - this fixes a severe security
2534 problem (CVE-2009-3555) at the cost of breaking all
2535 renegotiation. Renegotiation can be re-enabled by setting
2536 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2537 run-time. This is really not recommended unless you know what
2541 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2543 *) Don't set val to NULL when freeing up structures, it is freed up by
2544 underlying code. If sizeof(void *) > sizeof(long) this can result in
2545 zeroing past the valid field. (CVE-2009-0789)
2546 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2548 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2549 checked correctly. This would allow some invalid signed attributes to
2550 appear to verify correctly. (CVE-2009-0591)
2551 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2553 *) Reject UniversalString and BMPString types with invalid lengths. This
2554 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2555 a legal length. (CVE-2009-0590)
2558 *) Set S/MIME signing as the default purpose rather than setting it
2559 unconditionally. This allows applications to override it at the store
2563 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2564 to handle some structures.
2567 *) Improve efficiency of mem_gets: don't search whole buffer each time
2569 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2571 *) New -hex option for openssl rand.
2574 *) Print out UTF8String and NumericString when parsing ASN1.
2577 *) Support NumericString type for name components.
2580 *) Allow CC in the environment to override the automatically chosen
2581 compiler. Note that nothing is done to ensure flags work with the
2585 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2587 *) Properly check EVP_VerifyFinal() and similar return values
2589 [Ben Laurie, Bodo Moeller, Google Security Team]
2591 *) Enable TLS extensions by default.
2594 *) Allow the CHIL engine to be loaded, whether the application is
2595 multithreaded or not. (This does not release the developer from the
2596 obligation to set up the dynamic locking callbacks.)
2597 [Sander Temme <sander@temme.net>]
2599 *) Use correct exit code if there is an error in dgst command.
2600 [Steve Henson; problem pointed out by Roland Dirlewanger]
2602 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2603 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2606 *) Add experimental JPAKE support, including demo authentication in
2607 s_client and s_server.
2610 *) Set the comparison function in v3_addr_canonize().
2611 [Rob Austein <sra@hactrn.net>]
2613 *) Add support for XMPP STARTTLS in s_client.
2614 [Philip Paeps <philip@freebsd.org>]
2616 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2617 to ensure that even with this option, only ciphersuites in the
2618 server's preference list will be accepted. (Note that the option
2619 applies only when resuming a session, so the earlier behavior was
2620 just about the algorithm choice for symmetric cryptography.)
2623 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2625 *) Fix NULL pointer dereference if a DTLS server received
2626 ChangeCipherSpec as first record (CVE-2009-1386).
2629 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2630 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2633 *) The fix in 0.9.8c that supposedly got rid of unsafe
2634 double-checked locking was incomplete for RSA blinding,
2635 addressing just one layer of what turns out to have been
2636 doubly unsafe triple-checked locking.
2638 So now fix this for real by retiring the MONT_HELPER macro
2639 in crypto/rsa/rsa_eay.c.
2641 [Bodo Moeller; problem pointed out by Marius Schilder]
2643 *) Various precautionary measures:
2645 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2647 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2648 (NB: This would require knowledge of the secret session ticket key
2649 to exploit, in which case you'd be SOL either way.)
2651 - Change bn_nist.c so that it will properly handle input BIGNUMs
2652 outside the expected range.
2654 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2657 [Neel Mehta, Bodo Moeller]
2659 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2660 the load fails. Useful for distros.
2661 [Ben Laurie and the FreeBSD team]
2663 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2666 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2669 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2671 This work was sponsored by Logica.
2674 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2675 keystores. Support for SSL/TLS client authentication too.
2676 Not compiled unless enable-capieng specified to Configure.
2678 This work was sponsored by Logica.
2681 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2682 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2683 attribute creation routines such as certifcate requests and PKCS#12
2687 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2689 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2690 handshake which could lead to a cilent crash as found using the
2691 Codenomicon TLS test suite (CVE-2008-1672)
2692 [Steve Henson, Mark Cox]
2694 *) Fix double free in TLS server name extensions which could lead to
2695 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2698 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2700 Clear the error queue to ensure that error entries left from
2701 older function calls do not interfere with the correct operation.
2702 [Lutz Jaenicke, Erik de Castro Lopo]
2704 *) Remove root CA certificates of commercial CAs:
2706 The OpenSSL project does not recommend any specific CA and does not
2707 have any policy with respect to including or excluding any CA.
2708 Therefore it does not make any sense to ship an arbitrary selection
2709 of root CA certificates with the OpenSSL software.
2712 *) RSA OAEP patches to fix two separate invalid memory reads.
2713 The first one involves inputs when 'lzero' is greater than
2714 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2715 before the beginning of from). The second one involves inputs where
2716 the 'db' section contains nothing but zeroes (there is a one-byte
2717 invalid read after the end of 'db').
2718 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2720 *) Partial backport from 0.9.9-dev:
2722 Introduce bn_mul_mont (dedicated Montgomery multiplication
2723 procedure) as a candidate for BIGNUM assembler implementation.
2724 While 0.9.9-dev uses assembler for various architectures, only
2725 x86_64 is available by default here in the 0.9.8 branch, and
2726 32-bit x86 is available through a compile-time setting.
2728 To try the 32-bit x86 assembler implementation, use Configure
2729 option "enable-montasm" (which exists only for this backport).
2731 As "enable-montasm" for 32-bit x86 disclaims code stability
2732 anyway, in this constellation we activate additional code
2733 backported from 0.9.9-dev for further performance improvements,
2734 namely BN_from_montgomery_word. (To enable this otherwise,
2735 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2737 [Andy Polyakov (backport partially by Bodo Moeller)]
2739 *) Add TLS session ticket callback. This allows an application to set
2740 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2741 values. This is useful for key rollover for example where several key
2742 sets may exist with different names.
2745 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2746 This was broken until now in 0.9.8 releases, such that the only way
2747 a registered ENGINE could be used (assuming it initialises
2748 successfully on the host) was to explicitly set it as the default
2749 for the relevant algorithms. This is in contradiction with 0.9.7
2750 behaviour and the documentation. With this fix, when an ENGINE is
2751 registered into a given algorithm's table of implementations, the
2752 'uptodate' flag is reset so that auto-discovery will be used next
2753 time a new context for that algorithm attempts to select an
2755 [Ian Lister (tweaked by Geoff Thorpe)]
2757 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2758 implemention in the following ways:
2760 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2763 Lack of BER streaming support means one pass streaming processing is
2764 only supported if data is detached: setting the streaming flag is
2765 ignored for embedded content.
2767 CMS support is disabled by default and must be explicitly enabled
2768 with the enable-cms configuration option.
2771 *) Update the GMP engine glue to do direct copies between BIGNUM and
2772 mpz_t when openssl and GMP use the same limb size. Otherwise the
2773 existing "conversion via a text string export" trick is still used.
2774 [Paul Sheer <paulsheer@gmail.com>]
2776 *) Zlib compression BIO. This is a filter BIO which compressed and
2777 uncompresses any data passed through it.
2780 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2781 RFC3394 compatible AES key wrapping.
2784 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2785 sets string data without copying. X509_ALGOR_set0() and
2786 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2787 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2788 from an X509_ATTRIBUTE structure optionally checking it occurs only
2789 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2793 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2794 to get the expected BN_FLG_CONSTTIME behavior.
2795 [Bodo Moeller (Google)]
2799 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2800 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2801 - added some more tests to do_tests.pl
2802 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2803 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2804 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2805 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2806 - various changes to netware.pl to enable gcc-cross builds on Win32
2808 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2809 - various changes to fix missing prototype warnings
2810 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2811 - added AES, WHIRLPOOL and CPUID assembler code to build files
2812 - added missing AES assembler make rules to mk1mf.pl
2813 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2814 [Guenter Knauf <eflash@gmx.net>]
2816 *) Implement certificate status request TLS extension defined in RFC3546.
2817 A client can set the appropriate parameters and receive the encoded
2818 OCSP response via a callback. A server can query the supplied parameters
2819 and set the encoded OCSP response in the callback. Add simplified examples
2820 to s_client and s_server.
2823 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2825 *) Fix various bugs:
2826 + Binary incompatibility of ssl_ctx_st structure
2827 + DTLS interoperation with non-compliant servers
2828 + Don't call get_session_cb() without proposed session
2829 + Fix ia64 assembler code
2830 [Andy Polyakov, Steve Henson]
2832 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2834 *) DTLS Handshake overhaul. There were longstanding issues with
2835 OpenSSL DTLS implementation, which were making it impossible for
2836 RFC 4347 compliant client to communicate with OpenSSL server.
2837 Unfortunately just fixing these incompatibilities would "cut off"
2838 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2839 server keeps tolerating non RFC compliant syntax. The opposite is
2840 not true, 0.9.8f client can not communicate with earlier server.
2841 This update even addresses CVE-2007-4995.
2844 *) Changes to avoid need for function casts in OpenSSL: some compilers
2845 (gcc 4.2 and later) reject their use.
2846 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2849 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2850 RFC4507bis. The encrypted ticket format is an encrypted encoded
2851 SSL_SESSION structure, that way new session features are automatically
2854 If a client application caches session in an SSL_SESSION structure
2855 support is transparent because tickets are now stored in the encoded
2858 The SSL_CTX structure automatically generates keys for ticket
2859 protection in servers so again support should be possible
2860 with no application modification.
2862 If a client or server wishes to disable RFC4507 support then the option
2863 SSL_OP_NO_TICKET can be set.
2865 Add a TLS extension debugging callback to allow the contents of any client
2866 or server extensions to be examined.
2868 This work was sponsored by Google.
2871 *) Add initial support for TLS extensions, specifically for the server_name
2872 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2873 have new members for a host name. The SSL data structure has an
2874 additional member SSL_CTX *initial_ctx so that new sessions can be
2875 stored in that context to allow for session resumption, even after the
2876 SSL has been switched to a new SSL_CTX in reaction to a client's
2877 server_name extension.
2879 New functions (subject to change):
2881 SSL_get_servername()
2882 SSL_get_servername_type()
2885 New CTRL codes and macros (subject to change):
2887 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2888 - SSL_CTX_set_tlsext_servername_callback()
2889 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2890 - SSL_CTX_set_tlsext_servername_arg()
2891 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2893 openssl s_client has a new '-servername ...' option.
2895 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2896 '-key2 ...', '-servername_fatal' (subject to change). This allows
2897 testing the HostName extension for a specific single host name ('-cert'
2898 and '-key' remain fallbacks for handshakes without HostName
2899 negotiation). If the unrecogninzed_name alert has to be sent, this by
2900 default is a warning; it becomes fatal with the '-servername_fatal'
2903 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2905 *) Add AES and SSE2 assembly language support to VC++ build.
2908 *) Mitigate attack on final subtraction in Montgomery reduction.
2911 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2912 (which previously caused an internal error).
2915 *) Squeeze another 10% out of IGE mode when in != out.
2918 *) AES IGE mode speedup.
2919 [Dean Gaudet (Google)]
2921 *) Add the Korean symmetric 128-bit cipher SEED (see
2922 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2923 add SEED ciphersuites from RFC 4162:
2925 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2926 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2927 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2928 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2930 To minimize changes between patchlevels in the OpenSSL 0.9.8
2931 series, SEED remains excluded from compilation unless OpenSSL
2932 is configured with 'enable-seed'.
2933 [KISA, Bodo Moeller]
2935 *) Mitigate branch prediction attacks, which can be practical if a
2936 single processor is shared, allowing a spy process to extract
2937 information. For detailed background information, see
2938 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2939 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2940 and Necessary Software Countermeasures"). The core of the change
2941 are new versions BN_div_no_branch() and
2942 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2943 respectively, which are slower, but avoid the security-relevant
2944 conditional branches. These are automatically called by BN_div()
2945 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2946 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2947 remove a conditional branch.
2949 BN_FLG_CONSTTIME is the new name for the previous
2950 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2951 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2952 in the exponent causes BN_mod_exp_mont() to use the alternative
2953 implementation in BN_mod_exp_mont_consttime().) The old name
2954 remains as a deprecated alias.
2956 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2957 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2958 constant-time implementations for more than just exponentiation.
2959 Here too the old name is kept as a deprecated alias.
2961 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2962 the BN_BLINDING structure gets an independent copy of the
2963 modulus. This means that the previous "BIGNUM *m" argument to
2964 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2965 essentially becomes "const BIGNUM *m", although we can't actually
2966 change this in the header file before 0.9.9. It allows
2967 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2968 enable BN_FLG_CONSTTIME.
2970 [Matthew D Wood (Intel Corp)]
2972 *) In the SSL/TLS server implementation, be strict about session ID
2973 context matching (which matters if an application uses a single
2974 external cache for different purposes). Previously,
2975 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2976 set. This did ensure strict client verification, but meant that,
2977 with applications using a single external cache for quite
2978 different requirements, clients could circumvent ciphersuite
2979 restrictions for a given session ID context by starting a session
2980 in a different context.
2983 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2984 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2985 authentication-only ciphersuites.
2988 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2989 not complete and could lead to a possible single byte overflow
2990 (CVE-2007-5135) [Ben Laurie]
2992 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2994 *) Since AES128 and AES256 (and similarly Camellia128 and
2995 Camellia256) share a single mask bit in the logic of
2996 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2997 kludge to work properly if AES128 is available and AES256 isn't
2998 (or if Camellia128 is available and Camellia256 isn't).
3001 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3002 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3003 When a point or a seed is encoded in a BIT STRING, we need to
3004 prevent the removal of trailing zero bits to get the proper DER
3005 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3006 of a NamedBitList, for which trailing 0 bits need to be removed.)
3009 *) Have SSL/TLS server implementation tolerate "mismatched" record
3010 protocol version while receiving ClientHello even if the
3011 ClientHello is fragmented. (The server can't insist on the
3012 particular protocol version it has chosen before the ServerHello
3013 message has informed the client about his choice.)
3016 *) Add RFC 3779 support.
3017 [Rob Austein for ARIN, Ben Laurie]
3019 *) Load error codes if they are not already present instead of using a
3020 static variable. This allows them to be cleanly unloaded and reloaded.
3021 Improve header file function name parsing.
3024 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3025 or CAPABILITY handshake as required by RFCs.
3028 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3030 *) Introduce limits to prevent malicious keys being able to
3031 cause a denial of service. (CVE-2006-2940)
3032 [Steve Henson, Bodo Moeller]
3034 *) Fix ASN.1 parsing of certain invalid structures that can result
3035 in a denial of service. (CVE-2006-2937) [Steve Henson]
3037 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3038 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3040 *) Fix SSL client code which could crash if connecting to a
3041 malicious SSLv2 server. (CVE-2006-4343)
3042 [Tavis Ormandy and Will Drewry, Google Security Team]
3044 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3045 match only those. Before that, "AES256-SHA" would be interpreted
3046 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3047 the same strength classification in 0.9.7h) as we currently only
3048 have a single AES bit in the ciphersuite description bitmap.
3049 That change, however, also applied to ciphersuite strings such as
3050 "RC4-MD5" that intentionally matched multiple ciphersuites --
3051 namely, SSL 2.0 ciphersuites in addition to the more common ones
3052 from SSL 3.0/TLS 1.0.
3054 So we change the selection algorithm again: Naming an explicit
3055 ciphersuite selects this one ciphersuite, and any other similar
3056 ciphersuite (same bitmap) from *other* protocol versions.
3057 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3058 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3060 Since SSL 2.0 does not have any ciphersuites for which the
3061 128/256 bit distinction would be relevant, this works for now.
3062 The proper fix will be to use different bits for AES128 and
3063 AES256, which would have avoided the problems from the beginning;
3064 however, bits are scarce, so we can only do this in a new release
3065 (not just a patchlevel) when we can change the SSL_CIPHER
3066 definition to split the single 'unsigned long mask' bitmap into
3067 multiple values to extend the available space.
3071 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3073 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3074 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3076 *) Add AES IGE and biIGE modes.
3079 *) Change the Unix randomness entropy gathering to use poll() when
3080 possible instead of select(), since the latter has some
3081 undesirable limitations.
3082 [Darryl Miles via Richard Levitte and Bodo Moeller]
3084 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3085 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3086 cannot be implicitly activated as part of, e.g., the "AES" alias.
3087 However, please upgrade to OpenSSL 0.9.9[-dev] for
3088 non-experimental use of the ECC ciphersuites to get TLS extension
3089 support, which is required for curve and point format negotiation
3090 to avoid potential handshake problems.
3093 *) Disable rogue ciphersuites:
3095 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3096 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3097 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3099 The latter two were purportedly from
3100 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3103 Also deactivate the remaining ciphersuites from
3104 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3105 unofficial, and the ID has long expired.
3108 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3109 dual-core machines) and other potential thread-safety issues.
3112 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3113 versions), which is now available for royalty-free use
3114 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3115 Also, add Camellia TLS ciphersuites from RFC 4132.
3117 To minimize changes between patchlevels in the OpenSSL 0.9.8
3118 series, Camellia remains excluded from compilation unless OpenSSL
3119 is configured with 'enable-camellia'.
3122 *) Disable the padding bug check when compression is in use. The padding
3123 bug check assumes the first packet is of even length, this is not
3124 necessarily true if compresssion is enabled and can result in false
3125 positives causing handshake failure. The actual bug test is ancient
3126 code so it is hoped that implementations will either have fixed it by
3127 now or any which still have the bug do not support compression.
3130 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3132 *) When applying a cipher rule check to see if string match is an explicit
3133 cipher suite and only match that one cipher suite if it is.
3136 *) Link in manifests for VC++ if needed.
3137 [Austin Ziegler <halostatue@gmail.com>]
3139 *) Update support for ECC-based TLS ciphersuites according to
3140 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3141 TLS extensions, which are supported starting with the 0.9.9
3142 branch, not in the OpenSSL 0.9.8 branch).
3145 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3146 opaque EVP_CIPHER_CTX handling.
3149 *) Fixes and enhancements to zlib compression code. We now only use
3150 "zlib1.dll" and use the default __cdecl calling convention on Win32
3151 to conform with the standards mentioned here:
3152 http://www.zlib.net/DLL_FAQ.txt
3153 Static zlib linking now works on Windows and the new --with-zlib-include
3154 --with-zlib-lib options to Configure can be used to supply the location
3155 of the headers and library. Gracefully handle case where zlib library
3159 *) Several fixes and enhancements to the OID generation code. The old code
3160 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3161 handle numbers larger than ULONG_MAX, truncated printing and had a
3162 non standard OBJ_obj2txt() behaviour.
3165 *) Add support for building of engines under engine/ as shared libraries
3166 under VC++ build system.
3169 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3170 Hopefully, we will not see any false combination of paths any more.
3173 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3175 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3176 (part of SSL_OP_ALL). This option used to disable the
3177 countermeasure against man-in-the-middle protocol-version
3178 rollback in the SSL 2.0 server implementation, which is a bad
3179 idea. (CVE-2005-2969)
3181 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3182 for Information Security, National Institute of Advanced Industrial
3183 Science and Technology [AIST], Japan)]
3185 *) Add two function to clear and return the verify parameter flags.
3188 *) Keep cipherlists sorted in the source instead of sorting them at
3189 runtime, thus removing the need for a lock.
3192 *) Avoid some small subgroup attacks in Diffie-Hellman.
3193 [Nick Mathewson and Ben Laurie]
3195 *) Add functions for well-known primes.
3198 *) Extended Windows CE support.
3199 [Satoshi Nakamura and Andy Polyakov]
3201 *) Initialize SSL_METHOD structures at compile time instead of during
3202 runtime, thus removing the need for a lock.
3205 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3206 attempting to decrypt each encrypted key in turn. Add support to
3210 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3212 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3215 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3218 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3219 key into the same file any more.
3222 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3225 *) Add -utf8 command line and config file option to 'ca'.
3226 [Stefan <stf@udoma.org]
3228 *) Removed the macro des_crypt(), as it seems to conflict with some
3229 libraries. Use DES_crypt().
3232 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3233 involves renaming the source and generated shared-libs for
3234 both. The engines will accept the corrected or legacy ids
3235 ('ncipher' and '4758_cca' respectively) when binding. NB,
3236 this only applies when building 'shared'.
3237 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3239 *) Add attribute functions to EVP_PKEY structure. Modify
3240 PKCS12_create() to recognize a CSP name attribute and
3241 use it. Make -CSP option work again in pkcs12 utility.
3244 *) Add new functionality to the bn blinding code:
3245 - automatic re-creation of the BN_BLINDING parameters after
3246 a fixed number of uses (currently 32)
3247 - add new function for parameter creation
3248 - introduce flags to control the update behaviour of the
3249 BN_BLINDING parameters
3250 - hide BN_BLINDING structure
3251 Add a second BN_BLINDING slot to the RSA structure to improve
3252 performance when a single RSA object is shared among several
3256 *) Add support for DTLS.
3257 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3259 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3260 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3263 *) Remove buggy and incompletet DH cert support from
3264 ssl/ssl_rsa.c and ssl/s3_both.c
3267 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3268 the apps/openssl applications.
3271 *) Compile clean with "-Wall -Wmissing-prototypes
3272 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3273 DEBUG_SAFESTACK must also be set.
3276 *) Change ./Configure so that certain algorithms can be disabled by default.
3277 The new counterpiece to "no-xxx" is "enable-xxx".
3279 The patented RC5 and MDC2 algorithms will now be disabled unless
3280 "enable-rc5" and "enable-mdc2", respectively, are specified.
3282 (IDEA remains enabled despite being patented. This is because IDEA
3283 is frequently required for interoperability, and there is no license
3284 fee for non-commercial use. As before, "no-idea" can be used to
3285 avoid this algorithm.)
3289 *) Add processing of proxy certificates (see RFC 3820). This work was
3290 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3291 EGEE (Enabling Grids for E-science in Europe).
3294 *) RC4 performance overhaul on modern architectures/implementations, such
3295 as Intel P4, IA-64 and AMD64.
3298 *) New utility extract-section.pl. This can be used specify an alternative
3299 section number in a pod file instead of having to treat each file as
3300 a separate case in Makefile. This can be done by adding two lines to the
3303 =for comment openssl_section:XXX
3305 The blank line is mandatory.
3309 *) New arguments -certform, -keyform and -pass for s_client and s_server
3310 to allow alternative format key and certificate files and passphrase
3314 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3315 update associated structures and add various utility functions.
3317 Add new policy related verify parameters, include policy checking in
3318 standard verify code. Enhance 'smime' application with extra parameters
3319 to support policy checking and print out.
3322 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3323 Nehemiah processors. These extensions support AES encryption in hardware
3324 as well as RNG (though RNG support is currently disabled).
3325 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3327 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3330 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3331 [Andy Polyakov and a number of other people]
3333 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3334 implementation contributed by IBM.
3335 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3337 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3338 exponent rather than 'unsigned long'. There is a corresponding change to
3339 the new 'rsa_keygen' element of the RSA_METHOD structure.
3340 [Jelte Jansen, Geoff Thorpe]
3342 *) Functionality for creating the initial serial number file is now
3343 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3345 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3346 number file to 1, which is bound to cause problems. To avoid
3347 the problems while respecting compatibility between different 0.9.7
3348 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3349 CA.pl for serial number initialization. With the new release 0.9.8,
3350 we can fix the problem directly in the 'ca' utility.)
3353 *) Reduced header interdepencies by declaring more opaque objects in
3354 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3355 give fewer recursive includes, which could break lazy source code - so
3356 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3357 developers should define this symbol when building and using openssl to
3358 ensure they track the recommended behaviour, interfaces, [etc], but
3359 backwards-compatible behaviour prevails when this isn't defined.
3362 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3365 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3366 This will generate a random key of the appropriate length based on the
3367 cipher context. The EVP_CIPHER can provide its own random key generation
3368 routine to support keys of a specific form. This is used in the des and
3369 3des routines to generate a key of the correct parity. Update S/MIME
3370 code to use new functions and hence generate correct parity DES keys.
3371 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3372 valid (weak or incorrect parity).
3375 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3376 as looking them up. This is useful when the verified structure may contain
3377 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3378 present unless the new PKCS7_NO_CRL flag is asserted.
3381 *) Extend ASN1 oid configuration module. It now additionally accepts the
3384 shortName = some long name, 1.2.3.4
3387 *) Reimplemented the BN_CTX implementation. There is now no more static
3388 limitation on the number of variables it can handle nor the depth of the
3389 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3390 information can now expand as required, and rather than having a single
3391 static array of bignums, BN_CTX now uses a linked-list of such arrays
3392 allowing it to expand on demand whilst maintaining the usefulness of
3393 BN_CTX's "bundling".
3396 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3397 to allow all RSA operations to function using a single BN_CTX.
3400 *) Preliminary support for certificate policy evaluation and checking. This
3401 is initially intended to pass the tests outlined in "Conformance Testing
3402 of Relying Party Client Certificate Path Processing Logic" v1.07.
3405 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3406 remained unused and not that useful. A variety of other little bignum
3407 tweaks and fixes have also been made continuing on from the audit (see
3411 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3412 associated ASN1, EVP and SSL functions and old ASN1 macros.
3415 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3416 and this should never fail. So the return value from the use of
3417 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3418 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3421 *) BN_CTX_get() should return zero-valued bignums, providing the same
3422 initialised value as BN_new().
3423 [Geoff Thorpe, suggested by Ulf Möller]
3425 *) Support for inhibitAnyPolicy certificate extension.
3428 *) An audit of the BIGNUM code is underway, for which debugging code is
3429 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3430 is considered valid when processing BIGNUMs, and causes execution to
3431 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3432 further steps are taken to deliberately pollute unused data in BIGNUM
3433 structures to try and expose faulty code further on. For now, openssl will
3434 (in its default mode of operation) continue to tolerate the inconsistent
3435 forms that it has tolerated in the past, but authors and packagers should
3436 consider trying openssl and their own applications when compiled with
3437 these debugging symbols defined. It will help highlight potential bugs in
3438 their own code, and will improve the test coverage for OpenSSL itself. At
3439 some point, these tighter rules will become openssl's default to improve
3440 maintainability, though the assert()s and other overheads will remain only
3441 in debugging configurations. See bn.h for more details.
3442 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3444 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3445 that can only be obtained through BN_CTX_new() (which implicitly
3446 initialises it). The presence of this function only made it possible
3447 to overwrite an existing structure (and cause memory leaks).
3450 *) Because of the callback-based approach for implementing LHASH as a
3451 template type, lh_insert() adds opaque objects to hash-tables and
3452 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3453 to clean up those corresponding objects before destroying the hash table
3454 (and losing the object pointers). So some over-zealous constifications in
3455 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3456 objects as "const" and the lh_doall[_arg] callback wrappers are not
3457 prototyped to have "const" restrictions on the object pointers they are
3458 given (and so aren't required to cast them away any more).
3461 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3462 (speed) prefers to use its own implementation. The two implementations
3463 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3464 its object type properly exposed (MS_TM) instead of casting to/from "char
3465 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3466 aren't necessarily the greatest nomenclatures - but this is what was used
3467 internally to the implementation so I've used that for now.
3470 *) Ensure that deprecated functions do not get compiled when
3471 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3472 the self-tests were still using deprecated key-generation functions so
3473 these have been updated also.
3476 *) Reorganise PKCS#7 code to separate the digest location functionality
3477 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
3478 New function PKCS7_set_digest() to set the digest type for PKCS#7
3479 digestedData type. Add additional code to correctly generate the
3480 digestedData type and add support for this type in PKCS7 initialization
3484 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3485 structure of type "other".
3488 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3489 sure the loop does correctly stop and breaking ("division by zero")
3490 modulus operations are not performed. The (pre-generated) prime
3491 table crypto/bn/bn_prime.h was already correct, but it could not be
3492 re-generated on some platforms because of the "division by zero"
3493 situation in the script.
3494 [Ralf S. Engelschall]
3496 *) Update support for ECC-based TLS ciphersuites according to
3497 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3498 SHA-1 now is only used for "small" curves (where the
3499 representation of a field element takes up to 24 bytes); for
3500 larger curves, the field element resulting from ECDH is directly
3501 used as premaster secret.
3502 [Douglas Stebila (Sun Microsystems Laboratories)]
3504 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3505 curve secp160r1 to the tests.
3506 [Douglas Stebila (Sun Microsystems Laboratories)]
3508 *) Add the possibility to load symbols globally with DSO.
3509 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3511 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3512 control of the error stack.
3515 *) Add support for STORE in ENGINE.
3518 *) Add the STORE type. The intention is to provide a common interface
3519 to certificate and key stores, be they simple file-based stores, or
3520 HSM-type store, or LDAP stores, or...
3521 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3524 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3525 pass a list of arguments to any function as well as provide a way
3526 for a function to pass data back to the caller.
3529 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3530 works like BUF_strdup() but can be used to duplicate a portion of
3531 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3535 *) Add the function sk_find_ex() which works like sk_find(), but will
3536 return an index to an element even if an exact match couldn't be
3537 found. The index is guaranteed to point at the element where the
3538 searched-for key would be inserted to preserve sorting order.
3541 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3542 takes an extra flags argument for optional functionality. Currently,
3543 the following flags are defined:
3545 OBJ_BSEARCH_VALUE_ON_NOMATCH
3546 This one gets OBJ_bsearch_ex() to return a pointer to the first
3547 element where the comparing function returns a negative or zero
3550 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3551 This one gets OBJ_bsearch_ex() to return a pointer to the first
3552 element where the comparing function returns zero. This is useful
3553 if there are more than one element where the comparing function
3557 *) Make it possible to create self-signed certificates with 'openssl ca'
3558 in such a way that the self-signed certificate becomes part of the
3559 CA database and uses the same mechanisms for serial number generation
3560 as all other certificate signing. The new flag '-selfsign' enables
3561 this functionality. Adapt CA.sh and CA.pl.in.
3564 *) Add functionality to check the public key of a certificate request
3565 against a given private. This is useful to check that a certificate
3566 request can be signed by that key (self-signing).
3569 *) Make it possible to have multiple active certificates with the same
3570 subject in the CA index file. This is done only if the keyword
3571 'unique_subject' is set to 'no' in the main CA section (default
3572 if 'CA_default') of the configuration file. The value is saved
3573 with the database itself in a separate index attribute file,
3574 named like the index file with '.attr' appended to the name.
3577 *) Generate muti valued AVAs using '+' notation in config files for
3581 *) Support for nameConstraints certificate extension.
3584 *) Support for policyConstraints certificate extension.
3587 *) Support for policyMappings certificate extension.
3590 *) Make sure the default DSA_METHOD implementation only uses its
3591 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3592 and change its own handlers to be NULL so as to remove unnecessary
3593 indirection. This lets alternative implementations fallback to the
3594 default implementation more easily.
3597 *) Support for directoryName in GeneralName related extensions
3601 *) Make it possible to link applications using Makefile.shared.
3602 Make that possible even when linking against static libraries!
3605 *) Support for single pass processing for S/MIME signing. This now
3606 means that S/MIME signing can be done from a pipe, in addition
3607 cleartext signing (multipart/signed type) is effectively streaming
3608 and the signed data does not need to be all held in memory.
3610 This is done with a new flag PKCS7_STREAM. When this flag is set
3611 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3612 is done after the data is output (and digests calculated) in
3613 SMIME_write_PKCS7().
3616 *) Add full support for -rpath/-R, both in shared libraries and
3617 applications, at least on the platforms where it's known how
3621 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3622 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3623 will now compute a table of multiples of the generator that
3624 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3625 faster (notably in the case of a single point multiplication,
3626 scalar * generator).
3627 [Nils Larsch, Bodo Moeller]
3629 *) IPv6 support for certificate extensions. The various extensions
3630 which use the IP:a.b.c.d can now take IPv6 addresses using the
3631 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3635 *) Added an ENGINE that implements RSA by performing private key
3636 exponentiations with the GMP library. The conversions to and from
3637 GMP's mpz_t format aren't optimised nor are any montgomery forms
3638 cached, and on x86 it appears OpenSSL's own performance has caught up.
3639 However there are likely to be other architectures where GMP could
3640 provide a boost. This ENGINE is not built in by default, but it can be
3641 specified at Configure time and should be accompanied by the necessary
3642 linker additions, eg;
3643 ./config -DOPENSSL_USE_GMP -lgmp
3646 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3647 testing availability of engines with "-t" - the old behaviour is
3648 produced by increasing the feature's verbosity with "-tt".
3651 *) ECDSA routines: under certain error conditions uninitialized BN objects
3652 could be freed. Solution: make sure initialization is performed early
3653 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3657 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3658 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3659 software implementations. For DSA and DH, parameter generation can
3660 also be overriden by providing the appropriate method callbacks.
3663 *) Change the "progress" mechanism used in key-generation and
3664 primality testing to functions that take a new BN_GENCB pointer in
3665 place of callback/argument pairs. The new API functions have "_ex"
3666 postfixes and the older functions are reimplemented as wrappers for
3667 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3668 declarations of the old functions to help (graceful) attempts to
3669 migrate to the new functions. Also, the new key-generation API
3670 functions operate on a caller-supplied key-structure and return
3671 success/failure rather than returning a key or NULL - this is to
3672 help make "keygen" another member function of RSA_METHOD etc.
3674 Example for using the new callback interface:
3676 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3680 BN_GENCB_set(&my_cb, my_callback, my_arg);
3682 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3683 /* For the meaning of a, b in calls to my_callback(), see the
3684 * documentation of the function that calls the callback.
3685 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3686 * my_callback should return 1 if it wants BN_is_prime_ex()
3687 * to continue, or 0 to stop.
3692 *) Change the ZLIB compression method to be stateful, and make it
3693 available to TLS with the number defined in
3694 draft-ietf-tls-compression-04.txt.
3697 *) Add the ASN.1 structures and functions for CertificatePair, which
3698 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3700 CertificatePair ::= SEQUENCE {
3701 forward [0] Certificate OPTIONAL,
3702 reverse [1] Certificate OPTIONAL,
3703 -- at least one of the pair shall be present -- }
3705 Also implement the PEM functions to read and write certificate
3706 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3708 This needed to be defined, mostly for the sake of the LDAP
3709 attribute crossCertificatePair, but may prove useful elsewhere as
3713 *) Make it possible to inhibit symlinking of shared libraries in
3714 Makefile.shared, for Cygwin's sake.
3717 *) Extend the BIGNUM API by creating a function
3718 void BN_set_negative(BIGNUM *a, int neg);
3719 and a macro that behave like
3720 int BN_is_negative(const BIGNUM *a);
3722 to avoid the need to access 'a->neg' directly in applications.
3725 *) Implement fast modular reduction for pseudo-Mersenne primes
3726 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3727 EC_GROUP_new_curve_GFp() will now automatically use this
3729 [Nils Larsch <nla@trustcenter.de>]
3731 *) Add new lock type (CRYPTO_LOCK_BN).
3734 *) Change the ENGINE framework to automatically load engines
3735 dynamically from specific directories unless they could be
3736 found to already be built in or loaded. Move all the
3737 current engines except for the cryptodev one to a new
3739 The engines in engines/ are built as shared libraries if
3740 the "shared" options was given to ./Configure or ./config.
3741 Otherwise, they are inserted in libcrypto.a.
3742 /usr/local/ssl/engines is the default directory for dynamic
3743 engines, but that can be overriden at configure time through
3744 the usual use of --prefix and/or --openssldir, and at run
3745 time with the environment variable OPENSSL_ENGINES.
3746 [Geoff Thorpe and Richard Levitte]
3748 *) Add Makefile.shared, a helper makefile to build shared
3749 libraries. Addapt Makefile.org.
3752 *) Add version info to Win32 DLLs.
3753 [Peter 'Luna' Runestig" <peter@runestig.com>]
3755 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3756 can be added using this API to created arbitrary PKCS#12
3757 files while avoiding the low level API.
3759 New options to PKCS12_create(), key or cert can be NULL and
3760 will then be omitted from the output file. The encryption
3761 algorithm NIDs can be set to -1 for no encryption, the mac
3762 iteration count can be set to 0 to omit the mac.
3764 Enhance pkcs12 utility by making the -nokeys and -nocerts
3765 options work when creating a PKCS#12 file. New option -nomac
3766 to omit the mac, NONE can be set for an encryption algorithm.
3767 New code is modified to use the enhanced PKCS12_create()
3768 instead of the low level API.
3771 *) Extend ASN1 encoder to support indefinite length constructed
3772 encoding. This can output sequences tags and octet strings in
3773 this form. Modify pk7_asn1.c to support indefinite length
3774 encoding. This is experimental and needs additional code to
3775 be useful, such as an ASN1 bio and some enhanced streaming
3778 Extend template encode functionality so that tagging is passed
3779 down to the template encoder.
3782 *) Let 'openssl req' fail if an argument to '-newkey' is not
3783 recognized instead of using RSA as a default.
3786 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3787 As these are not official, they are not included in "ALL";
3788 the "ECCdraft" ciphersuite group alias can be used to select them.
3789 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3791 *) Add ECDH engine support.
3792 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3794 *) Add ECDH in new directory crypto/ecdh/.
3795 [Douglas Stebila (Sun Microsystems Laboratories)]
3797 *) Let BN_rand_range() abort with an error after 100 iterations
3798 without success (which indicates a broken PRNG).
3801 *) Change BN_mod_sqrt() so that it verifies that the input value
3802 is really the square of the return value. (Previously,
3803 BN_mod_sqrt would show GIGO behaviour.)
3806 *) Add named elliptic curves over binary fields from X9.62, SECG,
3807 and WAP/WTLS; add OIDs that were still missing.
3809 [Sheueling Chang Shantz and Douglas Stebila
3810 (Sun Microsystems Laboratories)]
3812 *) Extend the EC library for elliptic curves over binary fields
3813 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3816 EC_GF2m_simple_method
3820 EC_GROUP_new_curve_GF2m
3821 EC_GROUP_set_curve_GF2m
3822 EC_GROUP_get_curve_GF2m
3823 EC_POINT_set_affine_coordinates_GF2m
3824 EC_POINT_get_affine_coordinates_GF2m
3825 EC_POINT_set_compressed_coordinates_GF2m
3827 Point compression for binary fields is disabled by default for
3828 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3831 As binary polynomials are represented as BIGNUMs, various members
3832 of the EC_GROUP and EC_POINT data structures can be shared
3833 between the implementations for prime fields and binary fields;
3834 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3835 are essentially identical to their ..._GFp counterparts.
3836 (For simplicity, the '..._GFp' prefix has been dropped from
3837 various internal method names.)
3839 An internal 'field_div' method (similar to 'field_mul' and
3840 'field_sqr') has been added; this is used only for binary fields.
3842 [Sheueling Chang Shantz and Douglas Stebila
3843 (Sun Microsystems Laboratories)]
3845 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3846 through methods ('mul', 'precompute_mult').
3848 The generic implementations (now internally called 'ec_wNAF_mul'
3849 and 'ec_wNAF_precomputed_mult') remain the default if these
3850 methods are undefined.
3852 [Sheueling Chang Shantz and Douglas Stebila
3853 (Sun Microsystems Laboratories)]
3855 *) New function EC_GROUP_get_degree, which is defined through
3856 EC_METHOD. For curves over prime fields, this returns the bit
3857 length of the modulus.
3859 [Sheueling Chang Shantz and Douglas Stebila
3860 (Sun Microsystems Laboratories)]
3862 *) New functions EC_GROUP_dup, EC_POINT_dup.
3863 (These simply call ..._new and ..._copy).
3865 [Sheueling Chang Shantz and Douglas Stebila
3866 (Sun Microsystems Laboratories)]
3868 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3869 Polynomials are represented as BIGNUMs (where the sign bit is not
3870 used) in the following functions [macros]:
3873 BN_GF2m_sub [= BN_GF2m_add]
3874 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3875 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3876 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3878 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3879 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3880 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3881 BN_GF2m_cmp [= BN_ucmp]
3883 (Note that only the 'mod' functions are actually for fields GF(2^m).
3884 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3886 For some functions, an the irreducible polynomial defining a
3887 field can be given as an 'unsigned int[]' with strictly
3888 decreasing elements giving the indices of those bits that are set;
3889 i.e., p[] represents the polynomial
3890 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3892 p[0] > p[1] > ... > p[k] = 0.
3893 This applies to the following functions:
3898 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3899 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3901 BN_GF2m_mod_sqrt_arr
3902 BN_GF2m_mod_solve_quad_arr
3906 Conversion can be performed by the following functions:
3911 bntest.c has additional tests for binary polynomial arithmetic.
3913 Two implementations for BN_GF2m_mod_div() are available.
3914 The default algorithm simply uses BN_GF2m_mod_inv() and
3915 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3916 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3917 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3919 [Sheueling Chang Shantz and Douglas Stebila
3920 (Sun Microsystems Laboratories)]
3922 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3923 functionality is disabled at compile-time.
3924 [Douglas Stebila <douglas.stebila@sun.com>]
3926 *) Change default behaviour of 'openssl asn1parse' so that more
3927 information is visible when viewing, e.g., a certificate:
3929 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3930 mode the content of non-printable OCTET STRINGs is output in a
3931 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3932 avoid the appearance of a printable string.
3933 [Nils Larsch <nla@trustcenter.de>]
3935 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3937 EC_GROUP_set_asn1_flag()
3938 EC_GROUP_get_asn1_flag()
3939 EC_GROUP_set_point_conversion_form()
3940 EC_GROUP_get_point_conversion_form()
3941 These control ASN1 encoding details:
3942 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3943 has been set to OPENSSL_EC_NAMED_CURVE.
3944 - Points are encoded in uncompressed form by default; options for
3945 asn1_for are as for point2oct, namely
3946 POINT_CONVERSION_COMPRESSED
3947 POINT_CONVERSION_UNCOMPRESSED
3948 POINT_CONVERSION_HYBRID
3950 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3953 EC_GROUP_get0_seed()
3954 EC_GROUP_get_seed_len()
3955 This is used only for ASN1 purposes (so far).
3956 [Nils Larsch <nla@trustcenter.de>]
3958 *) Add 'field_type' member to EC_METHOD, which holds the NID
3959 of the appropriate field type OID. The new function
3960 EC_METHOD_get_field_type() returns this value.
3961 [Nils Larsch <nla@trustcenter.de>]
3966 EC_POINT_point2hex()
3967 EC_POINT_hex2point()
3968 providing useful interfaces to EC_POINT_point2oct() and
3969 EC_POINT_oct2point().
3970 [Nils Larsch <nla@trustcenter.de>]
3972 *) Change internals of the EC library so that the functions
3973 EC_GROUP_set_generator()
3974 EC_GROUP_get_generator()
3975 EC_GROUP_get_order()
3976 EC_GROUP_get_cofactor()
3977 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3978 to methods, which would lead to unnecessary code duplication when
3979 adding different types of curves.
3980 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3982 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3983 arithmetic, and such that modified wNAFs are generated
3984 (which avoid length expansion in many cases).
3987 *) Add a function EC_GROUP_check_discriminant() (defined via
3988 EC_METHOD) that verifies that the curve discriminant is non-zero.
3990 Add a function EC_GROUP_check() that makes some sanity tests
3991 on a EC_GROUP, its generator and order. This includes
3992 EC_GROUP_check_discriminant().
3993 [Nils Larsch <nla@trustcenter.de>]
3995 *) Add ECDSA in new directory crypto/ecdsa/.
3997 Add applications 'openssl ecparam' and 'openssl ecdsa'
3998 (these are based on 'openssl dsaparam' and 'openssl dsa').
4000 ECDSA support is also included in various other files across the
4001 library. Most notably,
4002 - 'openssl req' now has a '-newkey ecdsa:file' option;
4003 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4004 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4005 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4006 them suitable for ECDSA where domain parameters must be
4007 extracted before the specific public key;
4008 - ECDSA engine support has been added.
4009 [Nils Larsch <nla@trustcenter.de>]
4011 *) Include some named elliptic curves, and add OIDs from X9.62,
4012 SECG, and WAP/WTLS. Each curve can be obtained from the new
4014 EC_GROUP_new_by_curve_name(),
4015 and the list of available named curves can be obtained with
4016 EC_get_builtin_curves().
4017 Also add a 'curve_name' member to EC_GROUP objects, which can be
4019 EC_GROUP_set_curve_name()
4020 EC_GROUP_get_curve_name()
4021 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4023 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4024 was actually never needed) and in BN_mul(). The removal in BN_mul()
4025 required a small change in bn_mul_part_recursive() and the addition
4026 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4027 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4028 bn_sub_words() and bn_add_words() except they take arrays with
4032 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4034 *) Cleanse PEM buffers before freeing them since they may contain
4036 [Benjamin Bennett <ben@psc.edu>]
4038 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4039 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4040 authentication-only ciphersuites.
4043 *) Since AES128 and AES256 share a single mask bit in the logic of
4044 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4045 kludge to work properly if AES128 is available and AES256 isn't.
4048 *) Expand security boundary to match 1.1.1 module.
4051 *) Remove redundant features: hash file source, editing of test vectors
4052 modify fipsld to use external fips_premain.c signature.
4055 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4056 run algorithm test programs.
4059 *) Make algorithm test programs more tolerant of whitespace.
4062 *) Have SSL/TLS server implementation tolerate "mismatched" record
4063 protocol version while receiving ClientHello even if the
4064 ClientHello is fragmented. (The server can't insist on the
4065 particular protocol version it has chosen before the ServerHello
4066 message has informed the client about his choice.)
4069 *) Load error codes if they are not already present instead of using a
4070 static variable. This allows them to be cleanly unloaded and reloaded.
4073 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4075 *) Introduce limits to prevent malicious keys being able to
4076 cause a denial of service. (CVE-2006-2940)
4077 [Steve Henson, Bodo Moeller]
4079 *) Fix ASN.1 parsing of certain invalid structures that can result
4080 in a denial of service. (CVE-2006-2937) [Steve Henson]
4082 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4083 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4085 *) Fix SSL client code which could crash if connecting to a
4086 malicious SSLv2 server. (CVE-2006-4343)
4087 [Tavis Ormandy and Will Drewry, Google Security Team]
4089 *) Change ciphersuite string processing so that an explicit
4090 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4091 will no longer include "AES128-SHA"), and any other similar
4092 ciphersuite (same bitmap) from *other* protocol versions (so that
4093 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4094 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4095 changes from 0.9.8b and 0.9.8d.
4098 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4100 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4101 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4103 *) Change the Unix randomness entropy gathering to use poll() when
4104 possible instead of select(), since the latter has some
4105 undesirable limitations.
4106 [Darryl Miles via Richard Levitte and Bodo Moeller]
4108 *) Disable rogue ciphersuites:
4110 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4111 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4112 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4114 The latter two were purportedly from
4115 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4118 Also deactive the remaining ciphersuites from
4119 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4120 unofficial, and the ID has long expired.
4123 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4124 dual-core machines) and other potential thread-safety issues.
4127 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4129 *) Adapt fipsld and the build system to link against the validated FIPS
4130 module in FIPS mode.
4133 *) Fixes for VC++ 2005 build under Windows.
4136 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4137 from a Windows bash shell such as MSYS. It is autodetected from the
4138 "config" script when run from a VC++ environment. Modify standard VC++
4139 build to use fipscanister.o from the GNU make build.
4142 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4144 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4145 The value now differs depending on if you build for FIPS or not.
4146 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4147 safely run with a non-FIPSed libcrypto, as it may crash because of
4148 the difference induced by this change.
4151 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4153 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4154 (part of SSL_OP_ALL). This option used to disable the
4155 countermeasure against man-in-the-middle protocol-version
4156 rollback in the SSL 2.0 server implementation, which is a bad
4157 idea. (CVE-2005-2969)
4159 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4160 for Information Security, National Institute of Advanced Industrial
4161 Science and Technology [AIST], Japan)]
4163 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4164 mainly for FIPS compliance and not fully integrated at this stage.
4167 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4168 the exponentiation using a fixed-length exponent. (Otherwise,
4169 the information leaked through timing could expose the secret key
4170 after many signatures; cf. Bleichenbacher's attack on DSA with
4174 *) Make a new fixed-window mod_exp implementation the default for
4175 RSA, DSA, and DH private-key operations so that the sequence of
4176 squares and multiplies and the memory access pattern are
4177 independent of the particular secret key. This will mitigate
4178 cache-timing and potential related attacks.
4180 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4181 and this is automatically used by BN_mod_exp_mont() if the new flag
4182 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4183 will use this BN flag for private exponents unless the flag
4184 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4185 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4187 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4189 *) Change the client implementation for SSLv23_method() and
4190 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4191 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4192 (Previously, the SSL 2.0 backwards compatible Client Hello
4193 message format would be used even with SSL_OP_NO_SSLv2.)
4196 *) Add support for smime-type MIME parameter in S/MIME messages which some
4200 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4201 a threadsafe manner. Modify rsa code to use new function and add calls
4202 to dsa and dh code (which had race conditions before).
4205 *) Include the fixed error library code in the C error file definitions
4206 instead of fixing them up at runtime. This keeps the error code
4207 structures constant.
4210 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4212 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4215 *) Fixes for newer kerberos headers. NB: the casts are needed because
4216 the 'length' field is signed on one version and unsigned on another
4217 with no (?) obvious way to tell the difference, without these VC++
4218 complains. Also the "definition" of FAR (blank) is no longer included
4219 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4220 some needed definitions.
4223 *) Undo Cygwin change.
4226 *) Added support for proxy certificates according to RFC 3820.
4227 Because they may be a security thread to unaware applications,
4228 they must be explicitely allowed in run-time. See
4229 docs/HOWTO/proxy_certificates.txt for further information.
4232 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4234 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4235 server and client random values. Previously
4236 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4237 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4239 This change has negligible security impact because:
4241 1. Server and client random values still have 24 bytes of pseudo random
4244 2. Server and client random values are sent in the clear in the initial
4247 3. The master secret is derived using the premaster secret (48 bytes in
4248 size for static RSA ciphersuites) as well as client server and random
4251 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4254 [Stephen Henson, reported by UK NISCC]
4256 *) Use Windows randomness collection on Cygwin.
4259 *) Fix hang in EGD/PRNGD query when communication socket is closed
4260 prematurely by EGD/PRNGD.
4261 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4263 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4266 *) Back-port of selected performance improvements from development
4267 branch, as well as improved support for PowerPC platforms.
4270 *) Add lots of checks for memory allocation failure, error codes to indicate
4271 failure and freeing up memory if a failure occurs.
4272 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4274 *) Add new -passin argument to dgst.
4277 *) Perform some character comparisons of different types in X509_NAME_cmp:
4278 this is needed for some certificates that reencode DNs into UTF8Strings
4279 (in violation of RFC3280) and can't or wont issue name rollover
4283 *) Make an explicit check during certificate validation to see that
4284 the CA setting in each certificate on the chain is correct. As a
4285 side effect always do the following basic checks on extensions,
4286 not just when there's an associated purpose to the check:
4288 - if there is an unhandled critical extension (unless the user
4289 has chosen to ignore this fault)
4290 - if the path length has been exceeded (if one is set at all)
4291 - that certain extensions fit the associated purpose (if one has
4295 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4297 *) Avoid a race condition when CRLs are checked in a multi threaded
4298 environment. This would happen due to the reordering of the revoked
4299 entries during signature checking and serial number lookup. Now the
4300 encoding is cached and the serial number sort performed under a lock.
4301 Add new STACK function sk_is_sorted().
4304 *) Add Delta CRL to the extension code.
4307 *) Various fixes to s3_pkt.c so alerts are sent properly.
4308 [David Holmes <d.holmes@f5.com>]
4310 *) Reduce the chances of duplicate issuer name and serial numbers (in
4311 violation of RFC3280) using the OpenSSL certificate creation utilities.
4312 This is done by creating a random 64 bit value for the initial serial
4313 number when a serial number file is created or when a self signed
4314 certificate is created using 'openssl req -x509'. The initial serial
4315 number file is created using 'openssl x509 -next_serial' in CA.pl
4316 rather than being initialized to 1.
4319 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4321 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4322 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4323 [Joe Orton, Steve Henson]
4325 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4327 [Joe Orton, Steve Henson]
4329 *) Make it possible to have multiple active certificates with the same
4330 subject in the CA index file. This is done only if the keyword
4331 'unique_subject' is set to 'no' in the main CA section (default
4332 if 'CA_default') of the configuration file. The value is saved
4333 with the database itself in a separate index attribute file,
4334 named like the index file with '.attr' appended to the name.
4337 *) X509 verify fixes. Disable broken certificate workarounds when
4338 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4339 keyUsage extension present. Don't accept CRLs with unhandled critical
4340 extensions: since verify currently doesn't process CRL extensions this
4341 rejects a CRL with *any* critical extensions. Add new verify error codes
4345 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4346 A clarification of RFC2560 will require the use of OCTET STRINGs and
4347 some implementations cannot handle the current raw format. Since OpenSSL
4348 copies and compares OCSP nonces as opaque blobs without any attempt at
4349 parsing them this should not create any compatibility issues.
4352 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4353 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4354 this HMAC (and other) operations are several times slower than OpenSSL
4358 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4359 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4361 *) Use the correct content when signing type "other".
4364 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4366 *) Fix various bugs revealed by running the NISCC test suite:
4368 Stop out of bounds reads in the ASN1 code when presented with
4369 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4371 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4373 If verify callback ignores invalid public key errors don't try to check
4374 certificate signature with the NULL public key.
4378 *) New -ignore_err option in ocsp application to stop the server
4379 exiting on the first error in a request.
4382 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4383 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4387 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4388 extra data after the compression methods not only for TLS 1.0
4389 but also for SSL 3.0 (as required by the specification).
4390 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4392 *) Change X509_certificate_type() to mark the key as exported/exportable
4393 when it's 512 *bits* long, not 512 bytes.
4396 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4397 blocks during encryption.
4400 *) Various fixes to base64 BIO and non blocking I/O. On write
4401 flushes were not handled properly if the BIO retried. On read
4402 data was not being buffered properly and had various logic bugs.
4403 This also affects blocking I/O when the data being decoded is a
4407 *) Various S/MIME bugfixes and compatibility changes:
4408 output correct application/pkcs7 MIME type if
4409 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4410 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4411 of files as .eml work). Correctly handle very long lines in MIME
4415 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4417 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4418 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4419 a protocol version number mismatch like a decryption error
4420 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4423 *) Turn on RSA blinding by default in the default implementation
4424 to avoid a timing attack. Applications that don't want it can call
4425 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4426 They would be ill-advised to do so in most cases.
4427 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4429 *) Change RSA blinding code so that it works when the PRNG is not
4430 seeded (in this case, the secret RSA exponent is abused as
4431 an unpredictable seed -- if it is not unpredictable, there
4432 is no point in blinding anyway). Make RSA blinding thread-safe
4433 by remembering the creator's thread ID in rsa->blinding and
4434 having all other threads use local one-time blinding factors
4435 (this requires more computation than sharing rsa->blinding, but
4436 avoids excessive locking; and if an RSA object is not shared
4437 between threads, blinding will still be very fast).
4440 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4441 ENGINE as defaults for all supported algorithms irrespective of
4442 the 'flags' parameter. 'flags' is now honoured, so applications
4443 should make sure they are passing it correctly.
4446 *) Target "mingw" now allows native Windows code to be generated in
4447 the Cygwin environment as well as with the MinGW compiler.
4450 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4452 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4453 via timing by performing a MAC computation even if incorrrect
4454 block cipher padding has been found. This is a countermeasure
4455 against active attacks where the attacker has to distinguish
4456 between bad padding and a MAC verification error. (CVE-2003-0078)
4458 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4459 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4460 Martin Vuagnoux (EPFL, Ilion)]
4462 *) Make the no-err option work as intended. The intention with no-err
4463 is not to have the whole error stack handling routines removed from
4464 libcrypto, it's only intended to remove all the function name and
4465 reason texts, thereby removing some of the footprint that may not
4466 be interesting if those errors aren't displayed anyway.
4468 NOTE: it's still possible for any application or module to have it's
4469 own set of error texts inserted. The routines are there, just not
4470 used by default when no-err is given.
4473 *) Add support for FreeBSD on IA64.
4474 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4476 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4477 Kerberos function mit_des_cbc_cksum(). Before this change,
4478 the value returned by DES_cbc_cksum() was like the one from
4479 mit_des_cbc_cksum(), except the bytes were swapped.
4480 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4482 *) Allow an application to disable the automatic SSL chain building.
4483 Before this a rather primitive chain build was always performed in
4484 ssl3_output_cert_chain(): an application had no way to send the
4485 correct chain if the automatic operation produced an incorrect result.
4487 Now the chain builder is disabled if either:
4489 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4491 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4493 The reasoning behind this is that an application would not want the
4494 auto chain building to take place if extra chain certificates are
4495 present and it might also want a means of sending no additional
4496 certificates (for example the chain has two certificates and the
4500 *) Add the possibility to build without the ENGINE framework.
4501 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4503 *) Under Win32 gmtime() can return NULL: check return value in
4504 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4507 *) DSA routines: under certain error conditions uninitialized BN objects
4508 could be freed. Solution: make sure initialization is performed early
4509 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4510 Nils Larsch <nla@trustcenter.de> via PR#459)
4513 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4514 checked on reconnect on the client side, therefore session resumption
4515 could still fail with a "ssl session id is different" error. This
4516 behaviour is masked when SSL_OP_ALL is used due to
4517 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4518 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4519 followup to PR #377.
4522 *) IA-32 assembler support enhancements: unified ELF targets, support
4523 for SCO/Caldera platforms, fix for Cygwin shared build.
4526 *) Add support for FreeBSD on sparc64. As a consequence, support for
4527 FreeBSD on non-x86 processors is separate from x86 processors on
4528 the config script, much like the NetBSD support.
4529 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4531 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4533 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4536 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4537 code (06) was taken as the first octet of the session ID and the last
4538 octet was ignored consequently. As a result SSLv2 client side session
4539 caching could not have worked due to the session ID mismatch between
4541 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4545 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4546 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4550 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4551 seems that in spite of existing for more than a year, many application
4552 author have done nothing to provide the necessary callbacks, which
4553 means that this particular engine will not work properly anywhere.
4554 This is a very unfortunate situation which forces us, in the name
4555 of usability, to give the hw_ncipher.c a static lock, which is part
4557 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4558 appear in 0.9.8 or later. We EXPECT application authors to have
4559 dealt properly with this when 0.9.8 is released (unless we actually
4560 make such changes in the libcrypto locking code that changes will
4561 have to be made anyway).
4564 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4565 octets have been read, EOF or an error occurs. Without this change
4566 some truncated ASN1 structures will not produce an error.
4569 *) Disable Heimdal support, since it hasn't been fully implemented.
4570 Still give the possibility to force the use of Heimdal, but with
4571 warnings and a request that patches get sent to openssl-dev.
4574 *) Add the VC-CE target, introduce the WINCE sysname, and add
4575 INSTALL.WCE and appropriate conditionals to make it build.
4576 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4578 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4579 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4580 edit numbers of the version.
4581 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4583 *) Introduce safe string copy and catenation functions
4584 (BUF_strlcpy() and BUF_strlcat()).
4585 [Ben Laurie (CHATS) and Richard Levitte]
4587 *) Avoid using fixed-size buffers for one-line DNs.
4588 [Ben Laurie (CHATS)]
4590 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4591 resizing buffers containing secrets, and use where appropriate.
4592 [Ben Laurie (CHATS)]
4594 *) Avoid using fixed size buffers for configuration file location.
4595 [Ben Laurie (CHATS)]
4597 *) Avoid filename truncation for various CA files.
4598 [Ben Laurie (CHATS)]
4600 *) Use sizeof in preference to magic numbers.
4601 [Ben Laurie (CHATS)]
4603 *) Avoid filename truncation in cert requests.
4604 [Ben Laurie (CHATS)]
4606 *) Add assertions to check for (supposedly impossible) buffer
4608 [Ben Laurie (CHATS)]
4610 *) Don't cache truncated DNS entries in the local cache (this could
4611 potentially lead to a spoofing attack).
4612 [Ben Laurie (CHATS)]
4614 *) Fix various buffers to be large enough for hex/decimal
4615 representations in a platform independent manner.
4616 [Ben Laurie (CHATS)]
4618 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4619 resizing buffers containing secrets, and use where appropriate.
4620 [Ben Laurie (CHATS)]
4622 *) Add BIO_indent() to avoid much slightly worrying code to do
4624 [Ben Laurie (CHATS)]
4626 *) Convert sprintf()/BIO_puts() to BIO_printf().
4627 [Ben Laurie (CHATS)]
4629 *) buffer_gets() could terminate with the buffer only half
4631 [Ben Laurie (CHATS)]
4633 *) Add assertions to prevent user-supplied crypto functions from
4634 overflowing internal buffers by having large block sizes, etc.
4635 [Ben Laurie (CHATS)]
4637 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4639 [Ben Laurie (CHATS)]
4641 *) Eliminate unused copy of key in RC4.
4642 [Ben Laurie (CHATS)]
4644 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4645 [Ben Laurie (CHATS)]
4647 *) Fix off-by-one error in EGD path.
4648 [Ben Laurie (CHATS)]
4650 *) If RANDFILE path is too long, ignore instead of truncating.
4651 [Ben Laurie (CHATS)]
4653 *) Eliminate unused and incorrectly sized X.509 structure
4655 [Ben Laurie (CHATS)]
4657 *) Eliminate unused and dangerous function knumber().
4658 [Ben Laurie (CHATS)]
4660 *) Eliminate unused and dangerous structure, KSSL_ERR.
4661 [Ben Laurie (CHATS)]
4663 *) Protect against overlong session ID context length in an encoded
4664 session object. Since these are local, this does not appear to be
4666 [Ben Laurie (CHATS)]
4668 *) Change from security patch (see 0.9.6e below) that did not affect
4669 the 0.9.6 release series:
4671 Remote buffer overflow in SSL3 protocol - an attacker could
4672 supply an oversized master key in Kerberos-enabled versions.
4674 [Ben Laurie (CHATS)]
4676 *) Change the SSL kerb5 codes to match RFC 2712.
4679 *) Make -nameopt work fully for req and add -reqopt switch.
4680 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4682 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4683 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4685 *) Make sure tests can be performed even if the corresponding algorithms
4686 have been removed entirely. This was also the last step to make
4687 OpenSSL compilable with DJGPP under all reasonable conditions.
4688 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4690 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4691 to allow version independent disabling of normally unselected ciphers,
4692 which may be activated as a side-effect of selecting a single cipher.
4694 (E.g., cipher list string "RSA" enables ciphersuites that are left
4695 out of "ALL" because they do not provide symmetric encryption.
4696 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4697 [Lutz Jaenicke, Bodo Moeller]
4699 *) Add appropriate support for separate platform-dependent build
4700 directories. The recommended way to make a platform-dependent
4701 build directory is the following (tested on Linux), maybe with
4704 # Place yourself outside of the OpenSSL source tree. In
4705 # this example, the environment variable OPENSSL_SOURCE
4706 # is assumed to contain the absolute OpenSSL source directory.
4707 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4708 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4709 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4710 mkdir -p `dirname $F`
4711 ln -s $OPENSSL_SOURCE/$F $F
4714 To be absolutely sure not to disturb the source tree, a "make clean"
4715 is a good thing. If it isn't successfull, don't worry about it,
4716 it probably means the source directory is very clean.
4719 *) Make sure any ENGINE control commands make local copies of string
4720 pointers passed to them whenever necessary. Otherwise it is possible
4721 the caller may have overwritten (or deallocated) the original string
4722 data when a later ENGINE operation tries to use the stored values.
4723 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4725 *) Improve diagnostics in file reading and command-line digests.
4726 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4728 *) Add AES modes CFB and OFB to the object database. Correct an
4729 error in AES-CFB decryption.
4732 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4733 allows existing EVP_CIPHER_CTX structures to be reused after
4734 calling EVP_*Final(). This behaviour is used by encryption
4735 BIOs and some applications. This has the side effect that
4736 applications must explicitly clean up cipher contexts with
4737 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4740 *) Check the values of dna and dnb in bn_mul_recursive before calling
4741 bn_mul_comba (a non zero value means the a or b arrays do not contain
4742 n2 elements) and fallback to bn_mul_normal if either is not zero.
4745 *) Fix escaping of non-ASCII characters when using the -subj option
4746 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4749 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4750 form for "surname", serialNumber has no short form.
4751 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4752 therefore remove "mail" short name for "internet 7".
4753 The OID for unique identifiers in X509 certificates is
4754 x500UniqueIdentifier, not uniqueIdentifier.
4755 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4758 *) Add an "init" command to the ENGINE config module and auto initialize
4759 ENGINEs. Without any "init" command the ENGINE will be initialized
4760 after all ctrl commands have been executed on it. If init=1 the
4761 ENGINE is initailized at that point (ctrls before that point are run
4762 on the uninitialized ENGINE and after on the initialized one). If
4763 init=0 then the ENGINE will not be iniatialized at all.
4766 *) Fix the 'app_verify_callback' interface so that the user-defined
4767 argument is actually passed to the callback: In the
4768 SSL_CTX_set_cert_verify_callback() prototype, the callback
4769 declaration has been changed from
4772 int (*cb)(X509_STORE_CTX *,void *);
4773 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4774 i=s->ctx->app_verify_callback(&ctx)
4775 has been changed into
4776 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4778 To update applications using SSL_CTX_set_cert_verify_callback(),
4779 a dummy argument can be added to their callback functions.
4780 [D. K. Smetters <smetters@parc.xerox.com>]
4782 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4783 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4785 *) Add and OPENSSL_LOAD_CONF define which will cause
4786 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4787 This allows older applications to transparently support certain
4788 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4789 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4790 load the config file and OPENSSL_add_all_algorithms_conf() which will
4791 always load it have also been added.
4794 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4795 Adjust NIDs and EVP layer.
4796 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4798 *) Config modules support in openssl utility.
4800 Most commands now load modules from the config file,
4801 though in a few (such as version) this isn't done
4802 because it couldn't be used for anything.
4804 In the case of ca and req the config file used is
4805 the same as the utility itself: that is the -config
4806 command line option can be used to specify an
4810 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4811 use "openssl_conf" if filename is NULL use default openssl config file.
4814 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4815 config section name. Add a new flag to tolerate a missing config file
4816 and move code to CONF_modules_load_file().
4819 *) Support for crypto accelerator cards from Accelerated Encryption
4820 Processing, www.aep.ie. (Use engine 'aep')
4821 The support was copied from 0.9.6c [engine] and adapted/corrected
4822 to work with the new engine framework.
4823 [AEP Inc. and Richard Levitte]
4825 *) Support for SureWare crypto accelerator cards from Baltimore
4826 Technologies. (Use engine 'sureware')
4827 The support was copied from 0.9.6c [engine] and adapted
4828 to work with the new engine framework.
4831 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4832 make the newer ENGINE framework commands for the CHIL engine work.
4833 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4835 *) Make it possible to produce shared libraries on ReliantUNIX.
4836 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4838 *) Add the configuration target debug-linux-ppro.
4839 Make 'openssl rsa' use the general key loading routines
4840 implemented in apps.c, and make those routines able to
4841 handle the key format FORMAT_NETSCAPE and the variant
4843 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4845 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4846 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4848 *) Add -keyform to rsautl, and document -engine.
4849 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4851 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4852 BIO_R_NO_SUCH_FILE error code rather than the generic
4853 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4856 *) Add new functions
4858 ERR_peek_last_error_line
4859 ERR_peek_last_error_line_data.
4860 These are similar to
4863 ERR_peek_error_line_data,
4864 but report on the latest error recorded rather than the first one
4865 still in the error queue.
4866 [Ben Laurie, Bodo Moeller]
4868 *) default_algorithms option in ENGINE config module. This allows things
4870 default_algorithms = ALL
4871 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4874 *) Prelminary ENGINE config module.
4877 *) New experimental application configuration code.
4880 *) Change the AES code to follow the same name structure as all other
4881 symmetric ciphers, and behave the same way. Move everything to
4882 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4883 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4885 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4886 [Ben Laurie and Theo de Raadt]
4888 *) Add option to output public keys in req command.
4889 [Massimiliano Pala madwolf@openca.org]
4891 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4892 (up to about 10% better than before for P-192 and P-224).
4895 *) New functions/macros
4897 SSL_CTX_set_msg_callback(ctx, cb)
4898 SSL_CTX_set_msg_callback_arg(ctx, arg)
4899 SSL_set_msg_callback(ssl, cb)
4900 SSL_set_msg_callback_arg(ssl, arg)
4902 to request calling a callback function
4904 void cb(int write_p, int version, int content_type,
4905 const void *buf, size_t len, SSL *ssl, void *arg)
4907 whenever a protocol message has been completely received
4908 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4909 protocol version according to which the SSL library interprets
4910 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4911 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4912 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4913 specification (change_cipher_spec(20), alert(21), handshake(22)).
4914 'buf' and 'len' point to the actual message, 'ssl' to the
4915 SSL object, and 'arg' is the application-defined value set by
4916 SSL[_CTX]_set_msg_callback_arg().
4918 'openssl s_client' and 'openssl s_server' have new '-msg' options
4919 to enable a callback that displays all protocol messages.
4922 *) Change the shared library support so shared libraries are built as
4923 soon as the corresponding static library is finished, and thereby get
4924 openssl and the test programs linked against the shared library.
4925 This still only happens when the keyword "shard" has been given to
4926 the configuration scripts.
4928 NOTE: shared library support is still an experimental thing, and
4929 backward binary compatibility is still not guaranteed.
4930 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4932 *) Add support for Subject Information Access extension.
4933 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4935 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4936 additional bytes when new memory had to be allocated, not just
4937 when reusing an existing buffer.
4940 *) New command line and configuration option 'utf8' for the req command.
4941 This allows field values to be specified as UTF8 strings.
4944 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4945 runs for the former and machine-readable output for the latter.
4948 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4949 of the e-mail address in the DN (i.e., it will go into a certificate
4950 extension only). The new configuration file option 'email_in_dn = no'
4951 has the same effect.
4952 [Massimiliano Pala madwolf@openca.org]
4954 *) Change all functions with names starting with des_ to be starting
4955 with DES_ instead. Add wrappers that are compatible with libdes,
4956 but are named _ossl_old_des_*. Finally, add macros that map the
4957 des_* symbols to the corresponding _ossl_old_des_* if libdes
4958 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4959 desired, the des_* symbols will be mapped to DES_*, with one
4962 Since we provide two compatibility mappings, the user needs to
4963 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4964 compatibility is desired. The default (i.e., when that macro
4965 isn't defined) is OpenSSL 0.9.6c compatibility.
4967 There are also macros that enable and disable the support of old
4968 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4969 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4970 are defined, the default will apply: to support the old des routines.
4972 In either case, one must include openssl/des.h to get the correct
4973 definitions. Do not try to just include openssl/des_old.h, that
4976 NOTE: This is a major break of an old API into a new one. Software
4977 authors are encouraged to switch to the DES_ style functions. Some
4978 time in the future, des_old.h and the libdes compatibility functions
4979 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4980 default), and then completely removed.
4983 *) Test for certificates which contain unsupported critical extensions.
4984 If such a certificate is found during a verify operation it is
4985 rejected by default: this behaviour can be overridden by either
4986 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4987 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4988 X509_supported_extension() has also been added which returns 1 if a
4989 particular extension is supported.
4992 *) Modify the behaviour of EVP cipher functions in similar way to digests
4993 to retain compatibility with existing code.
4996 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4997 compatibility with existing code. In particular the 'ctx' parameter does
4998 not have to be to be initialized before the call to EVP_DigestInit() and
4999 it is tidied up after a call to EVP_DigestFinal(). New function
5000 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5001 EVP_MD_CTX_copy() changed to not require the destination to be
5002 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5003 requires the destination to be valid.
5005 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5006 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5009 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5010 so that complete 'Handshake' protocol structures are kept in memory
5011 instead of overwriting 'msg_type' and 'length' with 'body' data.
5014 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5015 [Massimo Santin via Richard Levitte]
5017 *) Major restructuring to the underlying ENGINE code. This includes
5018 reduction of linker bloat, separation of pure "ENGINE" manipulation
5019 (initialisation, etc) from functionality dealing with implementations
5020 of specific crypto iterfaces. This change also introduces integrated
5021 support for symmetric ciphers and digest implementations - so ENGINEs
5022 can now accelerate these by providing EVP_CIPHER and EVP_MD
5023 implementations of their own. This is detailed in crypto/engine/README
5024 as it couldn't be adequately described here. However, there are a few
5025 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5026 were changed in the original introduction of ENGINE code have now
5027 reverted back - the hooking from this code to ENGINE is now a good
5028 deal more passive and at run-time, operations deal directly with
5029 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5030 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5031 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5032 they were not being used by the framework as there is no concept of a
5033 BIGNUM_METHOD and they could not be generalised to the new
5034 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5035 ENGINE_cpy() has been removed as it cannot be consistently defined in
5039 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5042 *) Change mkdef.pl to sort symbols that get the same entry number,
5043 and make sure the automatically generated functions ERR_load_*
5044 become part of libeay.num as well.
5047 *) New function SSL_renegotiate_pending(). This returns true once
5048 renegotiation has been requested (either SSL_renegotiate() call
5049 or HelloRequest/ClientHello receveived from the peer) and becomes
5050 false once a handshake has been completed.
5051 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5052 sends a HelloRequest, but does not ensure that a handshake takes
5053 place. SSL_renegotiate_pending() is useful for checking if the
5054 client has followed the request.)
5057 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5058 By default, clients may request session resumption even during
5059 renegotiation (if session ID contexts permit); with this option,
5060 session resumption is possible only in the first handshake.
5062 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5063 more bits available for options that should not be part of
5064 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5067 *) Add some demos for certificate and certificate request creation.
5070 *) Make maximum certificate chain size accepted from the peer application
5071 settable (SSL*_get/set_max_cert_list()), as proposed by
5072 "Douglas E. Engert" <deengert@anl.gov>.
5075 *) Add support for shared libraries for Unixware-7
5076 (Boyd Lynn Gerber <gerberb@zenez.com>).
5079 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5080 be done prior to destruction. Use this to unload error strings from
5081 ENGINEs that load their own error strings. NB: This adds two new API
5082 functions to "get" and "set" this destroy handler in an ENGINE.
5085 *) Alter all existing ENGINE implementations (except "openssl" and
5086 "openbsd") to dynamically instantiate their own error strings. This
5087 makes them more flexible to be built both as statically-linked ENGINEs
5088 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5089 Also, add stub code to each that makes building them as self-contained
5090 shared-libraries easier (see README.ENGINE).
5093 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5094 implementations into applications that are completely implemented in
5095 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5096 commands that can be used to configure what shared-library to load and
5097 to control aspects of the way it is handled. Also, made an update to
5098 the README.ENGINE file that brings its information up-to-date and
5099 provides some information and instructions on the "dynamic" ENGINE
5100 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5103 *) Make it possible to unload ranges of ERR strings with a new
5104 "ERR_unload_strings" function.
5107 *) Add a copy() function to EVP_MD.
5110 *) Make EVP_MD routines take a context pointer instead of just the
5111 md_data void pointer.
5114 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5115 that the digest can only process a single chunk of data
5116 (typically because it is provided by a piece of
5117 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5118 is only going to provide a single chunk of data, and hence the
5119 framework needn't accumulate the data for oneshot drivers.
5122 *) As with "ERR", make it possible to replace the underlying "ex_data"
5123 functions. This change also alters the storage and management of global
5124 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5125 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5126 index counters. The API functions that use this state have been changed
5127 to take a "class_index" rather than pointers to the class's local STACK
5128 and counter, and there is now an API function to dynamically create new
5129 classes. This centralisation allows us to (a) plug a lot of the
5130 thread-safety problems that existed, and (b) makes it possible to clean
5131 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5132 such data would previously have always leaked in application code and
5133 workarounds were in place to make the memory debugging turn a blind eye
5134 to it. Application code that doesn't use this new function will still
5135 leak as before, but their memory debugging output will announce it now
5136 rather than letting it slide.
5138 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5139 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5140 has a return value to indicate success or failure.
5143 *) Make it possible to replace the underlying "ERR" functions such that the
5144 global state (2 LHASH tables and 2 locks) is only used by the "default"
5145 implementation. This change also adds two functions to "get" and "set"
5146 the implementation prior to it being automatically set the first time
5147 any other ERR function takes place. Ie. an application can call "get",
5148 pass the return value to a module it has just loaded, and that module
5149 can call its own "set" function using that value. This means the
5150 module's "ERR" operations will use (and modify) the error state in the
5151 application and not in its own statically linked copy of OpenSSL code.
5154 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5155 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5156 the operation, and provides a more encapsulated way for external code
5157 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5158 to use these functions rather than manually incrementing the counts.
5160 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5163 *) Add EVP test program.
5166 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5169 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5170 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5171 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5172 These allow a CRL to be built without having to access X509_CRL fields
5173 directly. Modify 'ca' application to use new functions.
5176 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5177 bug workarounds. Rollback attack detection is a security feature.
5178 The problem will only arise on OpenSSL servers when TLSv1 is not
5179 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5180 Software authors not wanting to support TLSv1 will have special reasons
5181 for their choice and can explicitly enable this option.
5182 [Bodo Moeller, Lutz Jaenicke]
5184 *) Rationalise EVP so it can be extended: don't include a union of
5185 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5186 (similar to those existing for EVP_CIPHER_CTX).
5191 EVP_MD_CTX_init(&md); /* new function call */
5192 EVP_DigestInit(&md, EVP_sha1());
5193 EVP_DigestUpdate(&md, in, len);
5194 EVP_DigestFinal(&md, out, NULL);
5195 EVP_MD_CTX_cleanup(&md); /* new function call */
5199 *) Make DES key schedule conform to the usual scheme, as well as
5200 correcting its structure. This means that calls to DES functions
5201 now have to pass a pointer to a des_key_schedule instead of a
5202 plain des_key_schedule (which was actually always a pointer
5205 des_key_schedule ks;
5207 des_set_key_checked(..., &ks);
5208 des_ncbc_encrypt(..., &ks, ...);
5210 (Note that a later change renames 'des_...' into 'DES_...'.)
5213 *) Initial reduction of linker bloat: the use of some functions, such as
5214 PEM causes large amounts of unused functions to be linked in due to
5215 poor organisation. For example pem_all.c contains every PEM function
5216 which has a knock on effect of linking in large amounts of (unused)
5217 ASN1 code. Grouping together similar functions and splitting unrelated
5218 functions prevents this.
5221 *) Cleanup of EVP macros.
5224 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5225 correct _ecb suffix.
5228 *) Add initial OCSP responder support to ocsp application. The
5229 revocation information is handled using the text based index
5230 use by the ca application. The responder can either handle
5231 requests generated internally, supplied in files (for example
5232 via a CGI script) or using an internal minimal server.
5235 *) Add configuration choices to get zlib compression for TLS.
5238 *) Changes to Kerberos SSL for RFC 2712 compliance:
5239 1. Implemented real KerberosWrapper, instead of just using
5240 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5241 2. Implemented optional authenticator field of KerberosWrapper.
5243 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5244 and authenticator structs; see crypto/krb5/.
5246 Generalized Kerberos calls to support multiple Kerberos libraries.
5247 [Vern Staats <staatsvr@asc.hpc.mil>,
5248 Jeffrey Altman <jaltman@columbia.edu>
5249 via Richard Levitte]
5251 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5252 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5253 values for each of the key sizes rather than having just
5254 parameters (and 'speed' generating keys each time).
5257 *) Speed up EVP routines.
5260 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5261 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5262 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5263 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5265 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5266 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5267 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5270 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5272 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5275 *) Added the OS2-EMX target.
5276 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5278 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5279 to support NCONF routines in extension code. New function CONF_set_nconf()
5280 to allow functions which take an NCONF to also handle the old LHASH
5281 structure: this means that the old CONF compatible routines can be
5282 retained (in particular wrt extensions) without having to duplicate the
5283 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5286 *) Enhance the general user interface with mechanisms for inner control
5287 and with possibilities to have yes/no kind of prompts.
5290 *) Change all calls to low level digest routines in the library and
5291 applications to use EVP. Add missing calls to HMAC_cleanup() and
5292 don't assume HMAC_CTX can be copied using memcpy().
5293 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5295 *) Add the possibility to control engines through control names but with
5296 arbitrary arguments instead of just a string.
5297 Change the key loaders to take a UI_METHOD instead of a callback
5298 function pointer. NOTE: this breaks binary compatibility with earlier
5299 versions of OpenSSL [engine].
5300 Adapt the nCipher code for these new conditions and add a card insertion
5304 *) Enhance the general user interface with mechanisms to better support
5305 dialog box interfaces, application-defined prompts, the possibility
5306 to use defaults (for example default passwords from somewhere else)
5307 and interrupts/cancellations.
5310 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5311 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5314 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5315 tidy up some unnecessarily weird code in 'sk_new()').
5316 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5318 *) Change the key loading routines for ENGINEs to use the same kind
5319 callback (pem_password_cb) as all other routines that need this
5323 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5324 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5325 than this minimum value is recommended.
5328 *) New random seeder for OpenVMS, using the system process statistics
5329 that are easily reachable.
5332 *) Windows apparently can't transparently handle global
5333 variables defined in DLLs. Initialisations such as:
5335 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5337 wont compile. This is used by the any applications that need to
5338 declare their own ASN1 modules. This was fixed by adding the option
5339 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5340 needed for static libraries under Win32.
5343 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5344 setting of purpose and trust fields. New X509_STORE trust and
5345 purpose functions and tidy up setting in other SSL functions.
5348 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5349 structure. These are inherited by X509_STORE_CTX when it is
5350 initialised. This allows various defaults to be set in the
5351 X509_STORE structure (such as flags for CRL checking and custom
5352 purpose or trust settings) for functions which only use X509_STORE_CTX
5353 internally such as S/MIME.
5355 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5356 trust settings if they are not set in X509_STORE. This allows X509_STORE
5357 purposes and trust (in S/MIME for example) to override any set by default.
5359 Add command line options for CRL checking to smime, s_client and s_server
5363 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5364 are set then the CRL is looked up in the X509_STORE structure and
5365 its validity and signature checked, then if the certificate is found
5366 in the CRL the verify fails with a revoked error.
5368 Various new CRL related callbacks added to X509_STORE_CTX structure.
5370 Command line options added to 'verify' application to support this.
5372 This needs some additional work, such as being able to handle multiple
5373 CRLs with different times, extension based lookup (rather than just
5374 by subject name) and ultimately more complete V2 CRL extension
5378 *) Add a general user interface API (crypto/ui/). This is designed
5379 to replace things like des_read_password and friends (backward
5380 compatibility functions using this new API are provided).
5381 The purpose is to remove prompting functions from the DES code
5382 section as well as provide for prompting through dialog boxes in
5383 a window system and the like.
5386 *) Add "ex_data" support to ENGINE so implementations can add state at a
5387 per-structure level rather than having to store it globally.
5390 *) Make it possible for ENGINE structures to be copied when retrieved by
5391 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5392 This causes the "original" ENGINE structure to act like a template,
5393 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5394 operational state can be localised to each ENGINE structure, despite the
5395 fact they all share the same "methods". New ENGINE structures returned in
5396 this case have no functional references and the return value is the single
5397 structural reference. This matches the single structural reference returned
5398 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5402 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5403 needs to match any other type at all we need to manually clear the
5407 *) Changes to the "openssl engine" utility to include;
5408 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5409 about an ENGINE's available control commands.
5410 - executing control commands from command line arguments using the
5411 '-pre' and '-post' switches. '-post' is only used if '-t' is
5412 specified and the ENGINE is successfully initialised. The syntax for
5413 the individual commands are colon-separated, for example;
5414 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5417 *) New dynamic control command support for ENGINEs. ENGINEs can now
5418 declare their own commands (numbers), names (strings), descriptions,
5419 and input types for run-time discovery by calling applications. A
5420 subset of these commands are implicitly classed as "executable"
5421 depending on their input type, and only these can be invoked through
5422 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5423 can be based on user input, config files, etc). The distinction is
5424 that "executable" commands cannot return anything other than a boolean
5425 result and can only support numeric or string input, whereas some
5426 discoverable commands may only be for direct use through
5427 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5428 pointers, or other custom uses. The "executable" commands are to
5429 support parameterisations of ENGINE behaviour that can be
5430 unambiguously defined by ENGINEs and used consistently across any
5431 OpenSSL-based application. Commands have been added to all the
5432 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5433 control over shared-library paths without source code alterations.
5436 *) Changed all ENGINE implementations to dynamically allocate their
5437 ENGINEs rather than declaring them statically. Apart from this being
5438 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5439 this also allows the implementations to compile without using the
5440 internal engine_int.h header.
5443 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5444 'const' value. Any code that should be able to modify a RAND_METHOD
5445 should already have non-const pointers to it (ie. they should only
5446 modify their own ones).
5449 *) Made a variety of little tweaks to the ENGINE code.
5450 - "atalla" and "ubsec" string definitions were moved from header files
5451 to C code. "nuron" string definitions were placed in variables
5452 rather than hard-coded - allowing parameterisation of these values
5453 later on via ctrl() commands.
5454 - Removed unused "#if 0"'d code.
5455 - Fixed engine list iteration code so it uses ENGINE_free() to release
5456 structural references.
5457 - Constified the RAND_METHOD element of ENGINE structures.
5458 - Constified various get/set functions as appropriate and added
5459 missing functions (including a catch-all ENGINE_cpy that duplicates
5460 all ENGINE values onto a new ENGINE except reference counts/state).
5461 - Removed NULL parameter checks in get/set functions. Setting a method
5462 or function to NULL is a way of cancelling out a previously set
5463 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5464 and doesn't justify the extra error symbols and code.
5465 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5466 flags from engine_int.h to engine.h.
5467 - Changed prototypes for ENGINE handler functions (init(), finish(),
5468 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5471 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5472 to the algorithm using long division. The binary algorithm can be
5473 used only if the modulus is odd. On 32-bit systems, it is faster
5474 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5475 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5476 up to 450 bits. In 64-bit environments, the binary algorithm
5477 appears to be advantageous for much longer moduli; here we use it
5478 for moduli up to 2048 bits.
5481 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5482 could not support the combine flag in choice fields.
5485 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5486 extensions from a certificate request to the certificate.
5489 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5490 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5491 file: this allows the display of the certificate about to be
5492 signed to be customised, to allow certain fields to be included
5493 or excluded and extension details. The old system didn't display
5494 multicharacter strings properly, omitted fields not in the policy
5495 and couldn't display additional details such as extensions.
5498 *) Function EC_POINTs_mul for multiple scalar multiplication
5499 of an arbitrary number of elliptic curve points
5500 \sum scalars[i]*points[i],
5501 optionally including the generator defined for the EC_GROUP:
5502 scalar*generator + \sum scalars[i]*points[i].
5504 EC_POINT_mul is a simple wrapper function for the typical case
5505 that the point list has just one item (besides the optional
5509 *) First EC_METHODs for curves over GF(p):
5511 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5512 operations and provides various method functions that can also
5513 operate with faster implementations of modular arithmetic.
5515 EC_GFp_mont_method() reuses most functions that are part of
5516 EC_GFp_simple_method, but uses Montgomery arithmetic.
5518 [Bodo Moeller; point addition and point doubling
5519 implementation directly derived from source code provided by
5520 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5522 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5523 crypto/ec/ec_lib.c):
5525 Curves are EC_GROUP objects (with an optional group generator)
5526 based on EC_METHODs that are built into the library.
5528 Points are EC_POINT objects based on EC_GROUP objects.
5530 Most of the framework would be able to handle curves over arbitrary
5531 finite fields, but as there are no obvious types for fields other
5532 than GF(p), some functions are limited to that for now.
5535 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5536 that the file contains a complete HTTP response.
5539 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5540 change the def and num file printf format specifier from "%-40sXXX"
5541 to "%-39s XXX". The latter will always guarantee a space after the
5542 field while the former will cause them to run together if the field
5543 is 40 of more characters long.
5546 *) Constify the cipher and digest 'method' functions and structures
5547 and modify related functions to take constant EVP_MD and EVP_CIPHER
5551 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5552 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5555 *) Modify EVP_Digest*() routines so they now return values. Although the
5556 internal software routines can never fail additional hardware versions
5560 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5562 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5563 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5566 ERR_R_NESTED_ASN1_ERROR
5568 ERR_R_MISSING_ASN1_EOS
5569 were 4 .. 9, conflicting with
5570 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5572 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5573 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5575 Add new error code 'ERR_R_INTERNAL_ERROR'.
5578 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5582 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5583 sets the subject name for a new request or supersedes the
5584 subject name in a given request. Formats that can be parsed are
5585 'CN=Some Name, OU=myOU, C=IT'
5587 'CN=Some Name/OU=myOU/C=IT'.
5589 Add options '-batch' and '-verbose' to 'openssl req'.
5590 [Massimiliano Pala <madwolf@hackmasters.net>]
5592 *) Introduce the possibility to access global variables through
5593 functions on platform were that's the best way to handle exporting
5594 global variables in shared libraries. To enable this functionality,
5595 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5596 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5597 is normally done by Configure or something similar).
5599 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5600 in the source file (foo.c) like this:
5602 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5603 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5605 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5606 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5608 OPENSSL_DECLARE_GLOBAL(int,foo);
5609 #define foo OPENSSL_GLOBAL_REF(foo)
5610 OPENSSL_DECLARE_GLOBAL(double,bar);
5611 #define bar OPENSSL_GLOBAL_REF(bar)
5613 The #defines are very important, and therefore so is including the
5614 header file everywhere where the defined globals are used.
5616 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5617 of ASN.1 items, but that structure is a bit different.
5619 The largest change is in util/mkdef.pl which has been enhanced with
5620 better and easier to understand logic to choose which symbols should
5621 go into the Windows .def files as well as a number of fixes and code
5622 cleanup (among others, algorithm keywords are now sorted
5623 lexicographically to avoid constant rewrites).
5626 *) In BN_div() keep a copy of the sign of 'num' before writing the
5627 result to 'rm' because if rm==num the value will be overwritten
5628 and produce the wrong result if 'num' is negative: this caused
5629 problems with BN_mod() and BN_nnmod().
5632 *) Function OCSP_request_verify(). This checks the signature on an
5633 OCSP request and verifies the signer certificate. The signer
5634 certificate is just checked for a generic purpose and OCSP request
5638 *) Add OCSP_check_validity() function to check the validity of OCSP
5639 responses. OCSP responses are prepared in real time and may only
5640 be a few seconds old. Simply checking that the current time lies
5641 between thisUpdate and nextUpdate max reject otherwise valid responses
5642 caused by either OCSP responder or client clock inaccuracy. Instead
5643 we allow thisUpdate and nextUpdate to fall within a certain period of
5644 the current time. The age of the response can also optionally be
5645 checked. Two new options -validity_period and -status_age added to
5649 *) If signature or public key algorithm is unrecognized print out its
5650 OID rather that just UNKNOWN.
5653 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5654 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5655 ID to be generated from the issuer certificate alone which can then be
5656 passed to OCSP_id_issuer_cmp().
5659 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5660 ASN1 modules to export functions returning ASN1_ITEM pointers
5661 instead of the ASN1_ITEM structures themselves. This adds several
5662 new macros which allow the underlying ASN1 function/structure to
5663 be accessed transparently. As a result code should not use ASN1_ITEM
5664 references directly (such as &X509_it) but instead use the relevant
5665 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5666 use of the new ASN1 code on platforms where exporting structures
5667 is problematical (for example in shared libraries) but exporting
5668 functions returning pointers to structures is not.
5671 *) Add support for overriding the generation of SSL/TLS session IDs.
5672 These callbacks can be registered either in an SSL_CTX or per SSL.
5673 The purpose of this is to allow applications to control, if they wish,
5674 the arbitrary values chosen for use as session IDs, particularly as it
5675 can be useful for session caching in multiple-server environments. A
5676 command-line switch for testing this (and any client code that wishes
5677 to use such a feature) has been added to "s_server".
5678 [Geoff Thorpe, Lutz Jaenicke]
5680 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5681 of the form '#if defined(...) || defined(...) || ...' and
5682 '#if !defined(...) && !defined(...) && ...'. This also avoids
5683 the growing number of special cases it was previously handling.
5686 *) Make all configuration macros available for application by making
5687 sure they are available in opensslconf.h, by giving them names starting
5688 with "OPENSSL_" to avoid conflicts with other packages and by making
5689 sure e_os2.h will cover all platform-specific cases together with
5691 Additionally, it is now possible to define configuration/platform-
5692 specific names (called "system identities"). In the C code, these
5693 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5694 macro with the name beginning with "OPENSSL_SYS_", which is determined
5695 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5699 *) New option -set_serial to 'req' and 'x509' this allows the serial
5700 number to use to be specified on the command line. Previously self
5701 signed certificates were hard coded with serial number 0 and the
5702 CA options of 'x509' had to use a serial number in a file which was
5706 *) New options to 'ca' utility to support V2 CRL entry extensions.
5707 Currently CRL reason, invalidity date and hold instruction are
5708 supported. Add new CRL extensions to V3 code and some new objects.
5711 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5712 disable standard block padding (aka PKCS#5 padding) in the EVP
5713 API, which was previously mandatory. This means that the data is
5714 not padded in any way and so the total length much be a multiple
5715 of the block size, otherwise an error occurs.
5718 *) Initial (incomplete) OCSP SSL support.
5721 *) New function OCSP_parse_url(). This splits up a URL into its host,
5722 port and path components: primarily to parse OCSP URLs. New -url
5723 option to ocsp utility.
5726 *) New nonce behavior. The return value of OCSP_check_nonce() now
5727 reflects the various checks performed. Applications can decide
5728 whether to tolerate certain situations such as an absent nonce
5729 in a response when one was present in a request: the ocsp application
5730 just prints out a warning. New function OCSP_add1_basic_nonce()
5731 this is to allow responders to include a nonce in a response even if
5732 the request is nonce-less.
5735 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5736 skipped when using openssl x509 multiple times on a single input file,
5737 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5740 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5741 set string type: to handle setting ASN1_TIME structures. Fix ca
5742 utility to correctly initialize revocation date of CRLs.
5745 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5746 the clients preferred ciphersuites and rather use its own preferences.
5747 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5748 Internet Explorer by ensuring unchanged hash method during stepup.
5749 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5752 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5753 to aes and add a new 'exist' option to print out symbols that don't
5757 *) Additional options to ocsp utility to allow flags to be set and
5758 additional certificates supplied.
5761 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5762 OCSP client a number of certificate to only verify the response
5766 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5767 handle the new API. Currently only ECB, CBC modes supported. Add new
5770 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5771 Encryption Standard (AES) Ciphersuites for Transport Layer
5772 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5773 not enabled by default and were not part of the "ALL" ciphersuite
5774 alias because they were not yet official; they could be
5775 explicitly requested by specifying the "AESdraft" ciphersuite
5776 group alias. In the final release of OpenSSL 0.9.7, the group
5777 alias is called "AES" and is part of "ALL".)
5778 [Ben Laurie, Steve Henson, Bodo Moeller]
5780 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5781 request to response.
5784 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5785 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5786 extract information from a certificate request. OCSP_response_create()
5787 creates a response and optionally adds a basic response structure.
5788 OCSP_basic_add1_status() adds a complete single response to a basic
5789 response and returns the OCSP_SINGLERESP structure just added (to allow
5790 extensions to be included for example). OCSP_basic_add1_cert() adds a
5791 certificate to a basic response and OCSP_basic_sign() signs a basic
5792 response with various flags. New helper functions ASN1_TIME_check()
5793 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5794 (converts ASN1_TIME to GeneralizedTime).
5797 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5798 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5799 structure from a certificate. X509_pubkey_digest() digests the public_key
5800 contents: this is used in various key identifiers.
5803 *) Make sk_sort() tolerate a NULL argument.
5804 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5806 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5807 passed by the function are trusted implicitly. If any of them signed the
5808 response then it is assumed to be valid and is not verified.
5811 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5812 to data. This was previously part of the PKCS7 ASN1 code. This
5813 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5814 [Steve Henson, reported by Kenneth R. Robinette
5815 <support@securenetterm.com>]
5817 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5818 routines: without these tracing memory leaks is very painful.
5819 Fix leaks in PKCS12 and PKCS7 routines.
5822 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5823 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5824 effectively meant GeneralizedTime would never be used. Now it
5825 is initialised to -1 but X509_time_adj() now has to check the value
5826 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5827 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5828 [Steve Henson, reported by Kenneth R. Robinette
5829 <support@securenetterm.com>]
5831 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5832 result in a zero length in the ASN1_INTEGER structure which was
5833 not consistent with the structure when d2i_ASN1_INTEGER() was used
5834 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5835 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5836 where it did not print out a minus for negative ASN1_INTEGER.
5839 *) Add summary printout to ocsp utility. The various functions which
5840 convert status values to strings have been renamed to:
5841 OCSP_response_status_str(), OCSP_cert_status_str() and
5842 OCSP_crl_reason_str() and are no longer static. New options
5843 to verify nonce values and to disable verification. OCSP response
5844 printout format cleaned up.
5847 *) Add additional OCSP certificate checks. These are those specified
5848 in RFC2560. This consists of two separate checks: the CA of the
5849 certificate being checked must either be the OCSP signer certificate
5850 or the issuer of the OCSP signer certificate. In the latter case the
5851 OCSP signer certificate must contain the OCSP signing extended key
5852 usage. This check is performed by attempting to match the OCSP
5853 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5854 in the OCSP_CERTID structures of the response.
5857 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5858 and related routines. This uses the standard OpenSSL certificate
5859 verify routines to perform initial checks (just CA validity) and
5860 to obtain the certificate chain. Then additional checks will be
5861 performed on the chain. Currently the root CA is checked to see
5862 if it is explicitly trusted for OCSP signing. This is used to set
5863 a root CA as a global signing root: that is any certificate that
5864 chains to that CA is an acceptable OCSP signing certificate.
5867 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5868 extensions from a separate configuration file.
5869 As when reading extensions from the main configuration file,
5870 the '-extensions ...' option may be used for specifying the
5872 [Massimiliano Pala <madwolf@comune.modena.it>]
5874 *) New OCSP utility. Allows OCSP requests to be generated or
5875 read. The request can be sent to a responder and the output
5876 parsed, outputed or printed in text form. Not complete yet:
5877 still needs to check the OCSP response validity.
5880 *) New subcommands for 'openssl ca':
5881 'openssl ca -status <serial>' prints the status of the cert with
5882 the given serial number (according to the index file).
5883 'openssl ca -updatedb' updates the expiry status of certificates
5885 [Massimiliano Pala <madwolf@comune.modena.it>]
5887 *) New '-newreq-nodes' command option to CA.pl. This is like
5888 '-newreq', but calls 'openssl req' with the '-nodes' option
5889 so that the resulting key is not encrypted.
5890 [Damien Miller <djm@mindrot.org>]
5892 *) New configuration for the GNU Hurd.
5893 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5895 *) Initial code to implement OCSP basic response verify. This
5896 is currently incomplete. Currently just finds the signer's
5897 certificate and verifies the signature on the response.
5900 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5901 value of OPENSSLDIR. This is available via the new '-d' option
5902 to 'openssl version', and is also included in 'openssl version -a'.
5905 *) Allowing defining memory allocation callbacks that will be given
5906 file name and line number information in additional arguments
5907 (a const char* and an int). The basic functionality remains, as
5908 well as the original possibility to just replace malloc(),
5909 realloc() and free() by functions that do not know about these
5910 additional arguments. To register and find out the current
5911 settings for extended allocation functions, the following
5912 functions are provided:
5914 CRYPTO_set_mem_ex_functions
5915 CRYPTO_set_locked_mem_ex_functions
5916 CRYPTO_get_mem_ex_functions
5917 CRYPTO_get_locked_mem_ex_functions
5919 These work the same way as CRYPTO_set_mem_functions and friends.
5920 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5921 extended allocation function is enabled.
5922 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5923 a conventional allocation function is enabled.
5924 [Richard Levitte, Bodo Moeller]
5926 *) Finish off removing the remaining LHASH function pointer casts.
5927 There should no longer be any prototype-casting required when using
5928 the LHASH abstraction, and any casts that remain are "bugs". See
5929 the callback types and macros at the head of lhash.h for details
5930 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5933 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5934 If /dev/[u]random devices are not available or do not return enough
5935 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5937 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5938 /etc/entropy will be queried once each in this sequence, quering stops
5939 when enough entropy was collected without querying more sockets.
5942 *) Change the Unix RAND_poll() variant to be able to poll several
5943 random devices, as specified by DEVRANDOM, until a sufficient amount
5944 of data has been collected. We spend at most 10 ms on each file
5945 (select timeout) and read in non-blocking mode. DEVRANDOM now
5946 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5947 (previously it was just the string "/dev/urandom"), so on typical
5948 platforms the 10 ms delay will never occur.
5949 Also separate out the Unix variant to its own file, rand_unix.c.
5950 For VMS, there's a currently-empty rand_vms.c.
5953 *) Move OCSP client related routines to ocsp_cl.c. These
5954 provide utility functions which an application needing
5955 to issue a request to an OCSP responder and analyse the
5956 response will typically need: as opposed to those which an
5957 OCSP responder itself would need which will be added later.
5959 OCSP_request_sign() signs an OCSP request with an API similar
5960 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5961 response. OCSP_response_get1_basic() extracts basic response
5962 from response. OCSP_resp_find_status(): finds and extracts status
5963 information from an OCSP_CERTID structure (which will be created
5964 when the request structure is built). These are built from lower
5965 level functions which work on OCSP_SINGLERESP structures but
5966 wont normally be used unless the application wishes to examine
5967 extensions in the OCSP response for example.
5969 Replace nonce routines with a pair of functions.
5970 OCSP_request_add1_nonce() adds a nonce value and optionally
5971 generates a random value. OCSP_check_nonce() checks the
5972 validity of the nonce in an OCSP response.
5975 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5976 This doesn't copy the supplied OCSP_CERTID and avoids the
5977 need to free up the newly created id. Change return type
5978 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5979 This can then be used to add extensions to the request.
5980 Deleted OCSP_request_new(), since most of its functionality
5981 is now in OCSP_REQUEST_new() (and the case insensitive name
5982 clash) apart from the ability to set the request name which
5983 will be added elsewhere.
5986 *) Update OCSP API. Remove obsolete extensions argument from
5987 various functions. Extensions are now handled using the new
5988 OCSP extension code. New simple OCSP HTTP function which
5989 can be used to send requests and parse the response.
5992 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5993 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5994 uses the special reorder version of SET OF to sort the attributes
5995 and reorder them to match the encoded order. This resolves a long
5996 standing problem: a verify on a PKCS7 structure just after signing
5997 it used to fail because the attribute order did not match the
5998 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5999 it uses the received order. This is necessary to tolerate some broken
6000 software that does not order SET OF. This is handled by encoding
6001 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6002 to produce the required SET OF.
6005 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6006 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6007 files to get correct declarations of the ASN.1 item variables.
6010 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6011 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6012 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6013 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6014 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6015 ASN1_ITEM and no wrapper functions.
6018 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6019 replace the old function pointer based I/O routines. Change most of
6020 the *_d2i_bio() and *_d2i_fp() functions to use these.
6023 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6024 lines, recognice more "algorithms" that can be deselected, and make
6025 it complain about algorithm deselection that isn't recognised.
6028 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6029 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6030 to use new functions. Add NO_ASN1_OLD which can be set to remove
6031 some old style ASN1 functions: this can be used to determine if old
6032 code will still work when these eventually go away.
6035 *) New extension functions for OCSP structures, these follow the
6036 same conventions as certificates and CRLs.
6039 *) New function X509V3_add1_i2d(). This automatically encodes and
6040 adds an extension. Its behaviour can be customised with various
6041 flags to append, replace or delete. Various wrappers added for
6042 certifcates and CRLs.
6045 *) Fix to avoid calling the underlying ASN1 print routine when
6046 an extension cannot be parsed. Correct a typo in the
6047 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6050 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
6051 entries for variables.
6054 *) Add functionality to apps/openssl.c for detecting locking
6055 problems: As the program is single-threaded, all we have
6056 to do is register a locking callback using an array for
6057 storing which locks are currently held by the program.
6060 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6061 SSL_get_ex_data_X509_STORE_idx(), which is used in
6062 ssl_verify_cert_chain() and thus can be called at any time
6063 during TLS/SSL handshakes so that thread-safety is essential.
6064 Unfortunately, the ex_data design is not at all suited
6065 for multi-threaded use, so it probably should be abolished.
6068 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6069 [Broadcom, tweaked and integrated by Geoff Thorpe]
6071 *) Move common extension printing code to new function
6072 X509V3_print_extensions(). Reorganise OCSP print routines and
6073 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6076 *) New function X509_signature_print() to remove duplication in some
6080 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6081 set (this was treated exactly the same as SET OF previously). This
6082 is used to reorder the STACK representing the structure to match the
6083 encoding. This will be used to get round a problem where a PKCS7
6084 structure which was signed could not be verified because the STACK
6085 order did not reflect the encoded order.
6088 *) Reimplement the OCSP ASN1 module using the new code.
6091 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6092 for its ASN1 operations. The old style function pointers still exist
6093 for now but they will eventually go away.
6096 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6097 completely replaces the old ASN1 functionality with a table driven
6098 encoder and decoder which interprets an ASN1_ITEM structure describing
6099 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6100 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6101 has also been converted to the new form.
6104 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6105 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6106 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6107 for negative moduli.
6110 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6111 of not touching the result's sign bit.
6114 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6118 *) Changed the LHASH code to use prototypes for callbacks, and created
6119 macros to declare and implement thin (optionally static) functions
6120 that provide type-safety and avoid function pointer casting for the
6121 type-specific callbacks.
6124 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6126 [Veers Staats <staatsvr@asc.hpc.mil>,
6127 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6129 *) Reformat the FAQ so the different questions and answers can be divided
6130 in sections depending on the subject.
6133 *) Have the zlib compression code load ZLIB.DLL dynamically under
6137 *) New function BN_mod_sqrt for computing square roots modulo a prime
6138 (using the probabilistic Tonelli-Shanks algorithm unless
6139 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6140 be handled deterministically).
6141 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6143 *) Make BN_mod_inverse faster by explicitly handling small quotients
6144 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6145 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6148 *) New function BN_kronecker.
6151 *) Fix BN_gcd so that it works on negative inputs; the result is
6152 positive unless both parameters are zero.
6153 Previously something reasonably close to an infinite loop was
6154 possible because numbers could be growing instead of shrinking
6155 in the implementation of Euclid's algorithm.
6158 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6159 sign of the number in question.
6161 Fix BN_is_word(a,w) to work correctly for w == 0.
6163 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6164 because its test if the absolute value of 'a' equals 'w'.
6165 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6166 it exists mostly for use in the implementations of BN_is_zero(),
6167 BN_is_one(), and BN_is_word().
6170 *) New function BN_swap.
6173 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6174 the exponentiation functions are more likely to produce reasonable
6175 results on negative inputs.
6178 *) Change BN_mod_mul so that the result is always non-negative.
6179 Previously, it could be negative if one of the factors was negative;
6180 I don't think anyone really wanted that behaviour.
6183 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6184 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6185 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6186 and add new functions:
6195 BN_mod_lshift1_quick
6199 These functions always generate non-negative results.
6201 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6202 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6204 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6205 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6206 be reduced modulo m.
6207 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6210 The following entry accidentily appeared in the CHANGES file
6211 distributed with OpenSSL 0.9.7. The modifications described in
6212 it do *not* apply to OpenSSL 0.9.7.
6214 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6215 was actually never needed) and in BN_mul(). The removal in BN_mul()
6216 required a small change in bn_mul_part_recursive() and the addition
6217 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6218 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6219 bn_sub_words() and bn_add_words() except they take arrays with
6224 *) In 'openssl passwd', verify passwords read from the terminal
6225 unless the '-salt' option is used (which usually means that
6226 verification would just waste user's time since the resulting
6227 hash is going to be compared with some given password hash)
6228 or the new '-noverify' option is used.
6230 This is an incompatible change, but it does not affect
6231 non-interactive use of 'openssl passwd' (passwords on the command
6232 line, '-stdin' option, '-in ...' option) and thus should not
6236 *) Remove all references to RSAref, since there's no more need for it.
6239 *) Make DSO load along a path given through an environment variable
6240 (SHLIB_PATH) with shl_load().
6243 *) Constify the ENGINE code as a result of BIGNUM constification.
6244 Also constify the RSA code and most things related to it. In a
6245 few places, most notable in the depth of the ASN.1 code, ugly
6246 casts back to non-const were required (to be solved at a later
6250 *) Make it so the openssl application has all engines loaded by default.
6253 *) Constify the BIGNUM routines a little more.
6256 *) Add the following functions:
6258 ENGINE_load_cswift()
6260 ENGINE_load_atalla()
6262 ENGINE_load_builtin_engines()
6264 That way, an application can itself choose if external engines that
6265 are built-in in OpenSSL shall ever be used or not. The benefit is
6266 that applications won't have to be linked with libdl or other dso
6267 libraries unless it's really needed.
6269 Changed 'openssl engine' to load all engines on demand.
6270 Changed the engine header files to avoid the duplication of some
6271 declarations (they differed!).
6274 *) 'openssl engine' can now list capabilities.
6277 *) Better error reporting in 'openssl engine'.
6280 *) Never call load_dh_param(NULL) in s_server.
6283 *) Add engine application. It can currently list engines by name and
6284 identity, and test if they are actually available.
6287 *) Improve RPM specification file by forcing symbolic linking and making
6288 sure the installed documentation is also owned by root.root.
6289 [Damien Miller <djm@mindrot.org>]
6291 *) Give the OpenSSL applications more possibilities to make use of
6292 keys (public as well as private) handled by engines.
6295 *) Add OCSP code that comes from CertCo.
6298 *) Add VMS support for the Rijndael code.
6301 *) Added untested support for Nuron crypto accelerator.
6304 *) Add support for external cryptographic devices. This code was
6305 previously distributed separately as the "engine" branch.
6306 [Geoff Thorpe, Richard Levitte]
6308 *) Rework the filename-translation in the DSO code. It is now possible to
6309 have far greater control over how a "name" is turned into a filename
6310 depending on the operating environment and any oddities about the
6311 different shared library filenames on each system.
6314 *) Support threads on FreeBSD-elf in Configure.
6317 *) Fix for SHA1 assembly problem with MASM: it produces
6318 warnings about corrupt line number information when assembling
6319 with debugging information. This is caused by the overlapping
6321 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6324 NCONF_get_number() has no error checking at all. As a replacement,
6325 NCONF_get_number_e() is defined (_e for "error checking") and is
6326 promoted strongly. The old NCONF_get_number is kept around for
6327 binary backward compatibility.
6328 Make it possible for methods to load from something other than a BIO,
6329 by providing a function pointer that is given a name instead of a BIO.
6330 For example, this could be used to load configuration data from an
6334 *) Fix for non blocking accept BIOs. Added new I/O special reason
6335 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6336 with non blocking I/O was not possible because no retry code was
6337 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6341 *) Added the beginnings of Rijndael support.
6344 *) Fix for bug in DirectoryString mask setting. Add support for
6345 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6346 to allow certificate printing to more controllable, additional
6347 'certopt' option to 'x509' to allow new printing options to be
6351 *) Clean old EAY MD5 hack from e_os.h.
6354 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6356 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6357 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6358 [Joe Orton, Steve Henson]
6360 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6362 *) Fix additional bug revealed by the NISCC test suite:
6364 Stop bug triggering large recursion when presented with
6365 certain ASN.1 tags (CVE-2003-0851)
6368 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6370 *) Fix various bugs revealed by running the NISCC test suite:
6372 Stop out of bounds reads in the ASN1 code when presented with
6373 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6375 If verify callback ignores invalid public key errors don't try to check
6376 certificate signature with the NULL public key.
6380 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6381 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6385 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6386 extra data after the compression methods not only for TLS 1.0
6387 but also for SSL 3.0 (as required by the specification).
6388 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6390 *) Change X509_certificate_type() to mark the key as exported/exportable
6391 when it's 512 *bits* long, not 512 bytes.
6394 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6396 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6397 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6398 a protocol version number mismatch like a decryption error
6399 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6402 *) Turn on RSA blinding by default in the default implementation
6403 to avoid a timing attack. Applications that don't want it can call
6404 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6405 They would be ill-advised to do so in most cases.
6406 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6408 *) Change RSA blinding code so that it works when the PRNG is not
6409 seeded (in this case, the secret RSA exponent is abused as
6410 an unpredictable seed -- if it is not unpredictable, there
6411 is no point in blinding anyway). Make RSA blinding thread-safe
6412 by remembering the creator's thread ID in rsa->blinding and
6413 having all other threads use local one-time blinding factors
6414 (this requires more computation than sharing rsa->blinding, but
6415 avoids excessive locking; and if an RSA object is not shared
6416 between threads, blinding will still be very fast).
6419 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6421 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6422 via timing by performing a MAC computation even if incorrrect
6423 block cipher padding has been found. This is a countermeasure
6424 against active attacks where the attacker has to distinguish
6425 between bad padding and a MAC verification error. (CVE-2003-0078)
6427 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6428 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6429 Martin Vuagnoux (EPFL, Ilion)]
6431 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6433 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6434 memory from it's contents. This is done with a counter that will
6435 place alternating values in each byte. This can be used to solve
6436 two issues: 1) the removal of calls to memset() by highly optimizing
6437 compilers, and 2) cleansing with other values than 0, since those can
6438 be read through on certain media, for example a swap space on disk.
6441 *) Bugfix: client side session caching did not work with external caching,
6442 because the session->cipher setting was not restored when reloading
6443 from the external cache. This problem was masked, when
6444 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6445 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6448 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6449 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6450 [Zeev Lieber <zeev-l@yahoo.com>]
6452 *) Undo an undocumented change introduced in 0.9.6e which caused
6453 repeated calls to OpenSSL_add_all_ciphers() and
6454 OpenSSL_add_all_digests() to be ignored, even after calling
6458 *) Change the default configuration reader to deal with last line not
6459 being properly terminated.
6462 *) Change X509_NAME_cmp() so it applies the special rules on handling
6463 DN values that are of type PrintableString, as well as RDNs of type
6464 emailAddress where the value has the type ia5String.
6465 [stefank@valicert.com via Richard Levitte]
6467 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6468 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6469 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6470 the bitwise-OR of the two for use by the majority of applications
6471 wanting this behaviour, and update the docs. The documented
6472 behaviour and actual behaviour were inconsistent and had been
6473 changing anyway, so this is more a bug-fix than a behavioural
6475 [Geoff Thorpe, diagnosed by Nadav Har'El]
6477 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6478 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6481 *) Fix initialization code race conditions in
6482 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6483 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6484 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6485 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6486 ssl2_get_cipher_by_char(),
6487 ssl3_get_cipher_by_char().
6488 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6490 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6491 the cached sessions are flushed, as the remove_cb() might use ex_data
6492 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6493 (see [openssl.org #212]).
6494 [Geoff Thorpe, Lutz Jaenicke]
6496 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6497 length, instead of the encoding length to d2i_ASN1_OBJECT.
6500 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6502 *) [In 0.9.6g-engine release:]
6503 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6504 [Lynn Gazis <lgazis@rainbow.com>]
6506 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6508 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6509 and get fix the header length calculation.
6510 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6511 Alon Kantor <alonk@checkpoint.com> (and others),
6514 *) Use proper error handling instead of 'assertions' in buffer
6515 overflow checks added in 0.9.6e. This prevents DoS (the
6516 assertions could call abort()).
6517 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6519 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6521 *) Add various sanity checks to asn1_get_length() to reject
6522 the ASN1 length bytes if they exceed sizeof(long), will appear
6523 negative or the content length exceeds the length of the
6525 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6527 *) Fix cipher selection routines: ciphers without encryption had no flags
6528 for the cipher strength set and where therefore not handled correctly
6529 by the selection routines (PR #130).
6532 *) Fix EVP_dsa_sha macro.
6536 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6537 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6538 that was added in OpenSSL 0.9.6d.
6540 As the countermeasure turned out to be incompatible with some
6541 broken SSL implementations, the new option is part of SSL_OP_ALL.
6542 SSL_OP_ALL is usually employed when compatibility with weird SSL
6543 implementations is desired (e.g. '-bugs' option to 's_client' and
6544 's_server'), so the new option is automatically set in many
6548 *) Changes in security patch:
6550 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6551 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6552 Air Force Materiel Command, USAF, under agreement number
6555 *) Add various sanity checks to asn1_get_length() to reject
6556 the ASN1 length bytes if they exceed sizeof(long), will appear
6557 negative or the content length exceeds the length of the
6558 supplied buffer. (CVE-2002-0659)
6559 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6561 *) Assertions for various potential buffer overflows, not known to
6563 [Ben Laurie (CHATS)]
6565 *) Various temporary buffers to hold ASCII versions of integers were
6566 too small for 64 bit platforms. (CVE-2002-0655)
6567 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6569 *) Remote buffer overflow in SSL3 protocol - an attacker could
6570 supply an oversized session ID to a client. (CVE-2002-0656)
6571 [Ben Laurie (CHATS)]
6573 *) Remote buffer overflow in SSL2 protocol - an attacker could
6574 supply an oversized client master key. (CVE-2002-0656)
6575 [Ben Laurie (CHATS)]
6577 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6579 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6580 encoded as NULL) with id-dsa-with-sha1.
6581 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6583 *) Check various X509_...() return values in apps/req.c.
6584 [Nils Larsch <nla@trustcenter.de>]
6586 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6587 an end-of-file condition would erronously be flagged, when the CRLF
6588 was just at the end of a processed block. The bug was discovered when
6589 processing data through a buffering memory BIO handing the data to a
6590 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6591 <ptsekov@syntrex.com> and Nedelcho Stanev.
6594 *) Implement a countermeasure against a vulnerability recently found
6595 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6596 before application data chunks to avoid the use of known IVs
6597 with data potentially chosen by the attacker.
6600 *) Fix length checks in ssl3_get_client_hello().
6603 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6604 to prevent ssl3_read_internal() from incorrectly assuming that
6605 ssl3_read_bytes() found application data while handshake
6606 processing was enabled when in fact s->s3->in_read_app_data was
6607 merely automatically cleared during the initial handshake.
6608 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6610 *) Fix object definitions for Private and Enterprise: they were not
6611 recognized in their shortname (=lowercase) representation. Extend
6612 obj_dat.pl to issue an error when using undefined keywords instead
6613 of silently ignoring the problem (Svenning Sorensen
6614 <sss@sss.dnsalias.net>).
6617 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6618 generators, i.e. generators other than 2 and 5. (Previously, the
6619 code did not properly initialise the 'add' and 'rem' values to
6620 BN_generate_prime().)
6622 In the new general case, we do not insist that 'generator' is
6623 actually a primitive root: This requirement is rather pointless;
6624 a generator of the order-q subgroup is just as good, if not
6628 *) Map new X509 verification errors to alerts. Discovered and submitted by
6629 Tom Wu <tom@arcot.com>.
6632 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6633 returning non-zero before the data has been completely received
6634 when using non-blocking I/O.
6635 [Bodo Moeller; problem pointed out by John Hughes]
6637 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6638 [Ben Laurie, Lutz Jaenicke]
6640 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6641 Yoram Zahavi <YoramZ@gilian.com>).
6644 *) Add information about CygWin 1.3 and on, and preserve proper
6645 configuration for the versions before that.
6646 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6648 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6649 check whether we deal with a copy of a session and do not delete from
6650 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6651 <izhar@checkpoint.com>.
6654 *) Do not store session data into the internal session cache, if it
6655 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6656 flag is set). Proposed by Aslam <aslam@funk.com>.
6659 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6663 *) [In 0.9.6d-engine release:]
6664 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6665 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6667 *) Add the configuration target linux-s390x.
6668 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6670 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6671 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6672 variable as an indication that a ClientHello message has been
6673 received. As the flag value will be lost between multiple
6674 invocations of ssl3_accept when using non-blocking I/O, the
6675 function may not be aware that a handshake has actually taken
6676 place, thus preventing a new session from being added to the
6679 To avoid this problem, we now set s->new_session to 2 instead of
6680 using a local variable.
6681 [Lutz Jaenicke, Bodo Moeller]
6683 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6684 if the SSL_R_LENGTH_MISMATCH error is detected.
6685 [Geoff Thorpe, Bodo Moeller]
6687 *) New 'shared_ldflag' column in Configure platform table.
6690 *) Fix EVP_CIPHER_mode macro.
6691 ["Dan S. Camper" <dan@bti.net>]
6693 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6694 type, we must throw them away by setting rr->length to 0.
6695 [D P Chang <dpc@qualys.com>]
6697 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6699 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6700 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6701 worked incorrectly for those cases where range = 10..._2 and
6702 3*range is two bits longer than range.)
6705 *) Only add signing time to PKCS7 structures if it is not already
6709 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6710 OBJ_ld_ce should be OBJ_id_ce.
6711 Also some ip-pda OIDs in crypto/objects/objects.txt were
6712 incorrect (cf. RFC 3039).
6713 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6715 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6716 returns early because it has nothing to do.
6717 [Andy Schneider <andy.schneider@bjss.co.uk>]
6719 *) [In 0.9.6c-engine release:]
6720 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6721 [Andy Schneider <andy.schneider@bjss.co.uk>]
6723 *) [In 0.9.6c-engine release:]
6724 Add support for Cryptographic Appliance's keyserver technology.
6725 (Use engine 'keyclient')
6726 [Cryptographic Appliances and Geoff Thorpe]
6728 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6729 is called via tools/c89.sh because arguments have to be
6730 rearranged (all '-L' options must appear before the first object
6732 [Richard Shapiro <rshapiro@abinitio.com>]
6734 *) [In 0.9.6c-engine release:]
6735 Add support for Broadcom crypto accelerator cards, backported
6737 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6739 *) [In 0.9.6c-engine release:]
6740 Add support for SureWare crypto accelerator cards from
6741 Baltimore Technologies. (Use engine 'sureware')
6742 [Baltimore Technologies and Mark Cox]
6744 *) [In 0.9.6c-engine release:]
6745 Add support for crypto accelerator cards from Accelerated
6746 Encryption Processing, www.aep.ie. (Use engine 'aep')
6747 [AEP Inc. and Mark Cox]
6749 *) Add a configuration entry for gcc on UnixWare.
6750 [Gary Benson <gbenson@redhat.com>]
6752 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6753 messages are stored in a single piece (fixed-length part and
6754 variable-length part combined) and fix various bugs found on the way.
6757 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6758 instead. BIO_gethostbyname() does not know what timeouts are
6759 appropriate, so entries would stay in cache even when they have
6761 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6763 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6764 faced with a pathologically small ClientHello fragment that does
6765 not contain client_version: Instead of aborting with an error,
6766 simply choose the highest available protocol version (i.e.,
6767 TLS 1.0 unless it is disabled). In practice, ClientHello
6768 messages are never sent like this, but this change gives us
6769 strictly correct behaviour at least for TLS.
6772 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6773 never resets s->method to s->ctx->method when called from within
6774 one of the SSL handshake functions.
6775 [Bodo Moeller; problem pointed out by Niko Baric]
6777 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6778 (sent using the client's version number) if client_version is
6779 smaller than the protocol version in use. Also change
6780 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6781 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6782 the client will at least see that alert.
6785 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6789 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6790 client receives HelloRequest while in a handshake.
6791 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6793 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6794 should end in 'break', not 'goto end' which circuments various
6795 cleanups done in state SSL_ST_OK. But session related stuff
6796 must be disabled for SSL_ST_OK in the case that we just sent a
6799 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6800 before just sending a HelloRequest.
6801 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6803 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6804 reveal whether illegal block cipher padding was found or a MAC
6805 verification error occured. (Neither SSLerr() codes nor alerts
6806 are directly visible to potential attackers, but the information
6807 may leak via logfiles.)
6809 Similar changes are not required for the SSL 2.0 implementation
6810 because the number of padding bytes is sent in clear for SSL 2.0,
6811 and the extra bytes are just ignored. However ssl/s2_pkt.c
6812 failed to verify that the purported number of padding bytes is in
6816 *) Add OpenUNIX-8 support including shared libraries
6817 (Boyd Lynn Gerber <gerberb@zenez.com>).
6820 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6821 'wristwatch attack' using huge encoding parameters (cf.
6822 James H. Manger's CRYPTO 2001 paper). Note that the
6823 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6824 encoding parameters and hence was not vulnerable.
6827 *) BN_sqr() bug fix.
6828 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6830 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6831 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6832 followed by modular reduction.
6833 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6835 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6836 equivalent based on BN_pseudo_rand() instead of BN_rand().
6839 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6840 This function was broken, as the check for a new client hello message
6841 to handle SGC did not allow these large messages.
6842 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6845 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6848 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6849 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6852 *) Rework the configuration and shared library support for Tru64 Unix.
6853 The configuration part makes use of modern compiler features and
6854 still retains old compiler behavior for those that run older versions
6855 of the OS. The shared library support part includes a variant that
6856 uses the RPATH feature, and is available through the special
6857 configuration target "alpha-cc-rpath", which will never be selected
6859 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6861 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6862 with the same message size as in ssl3_get_certificate_request().
6863 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6864 messages might inadvertently be reject as too long.
6865 [Petr Lampa <lampa@fee.vutbr.cz>]
6867 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6870 *) Modified SSL library such that the verify_callback that has been set
6871 specificly for an SSL object with SSL_set_verify() is actually being
6872 used. Before the change, a verify_callback set with this function was
6873 ignored and the verify_callback() set in the SSL_CTX at the time of
6874 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6875 to allow the necessary settings.
6878 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6879 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6880 done automatically (in contradiction to the requirements of the C
6881 standard). This made problems when used from OpenSSH.
6884 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6885 dh->length and always used
6887 BN_rand_range(priv_key, dh->p).
6889 BN_rand_range() is not necessary for Diffie-Hellman, and this
6890 specific range makes Diffie-Hellman unnecessarily inefficient if
6891 dh->length (recommended exponent length) is much smaller than the
6892 length of dh->p. We could use BN_rand_range() if the order of
6893 the subgroup was stored in the DH structure, but we only have
6898 BN_rand(priv_key, l, ...)
6900 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6906 RSA_eay_public_encrypt
6907 RSA_eay_private_decrypt
6908 RSA_eay_private_encrypt (signing)
6909 RSA_eay_public_decrypt (signature verification)
6911 (default implementations for RSA_public_encrypt,
6912 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6913 always reject numbers >= n.
6916 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6917 to synchronize access to 'locking_thread'. This is necessary on
6918 systems where access to 'locking_thread' (an 'unsigned long'
6919 variable) is not atomic.
6922 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6923 *before* setting the 'crypto_lock_rand' flag. The previous code had
6924 a race condition if 0 is a valid thread ID.
6925 [Travis Vitek <vitek@roguewave.com>]
6927 *) Add support for shared libraries under Irix.
6928 [Albert Chin-A-Young <china@thewrittenword.com>]
6930 *) Add configuration option to build on Linux on both big-endian and
6932 [Ralf Baechle <ralf@uni-koblenz.de>]
6934 *) Add the possibility to create shared libraries on HP-UX.
6937 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6939 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6940 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6941 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6942 PRNG state recovery was possible based on the output of
6943 one PRNG request appropriately sized to gain knowledge on
6944 'md' followed by enough consecutive 1-byte PRNG requests
6945 to traverse all of 'state'.
6947 1. When updating 'md_local' (the current thread's copy of 'md')
6948 during PRNG output generation, hash all of the previous
6949 'md_local' value, not just the half used for PRNG output.
6951 2. Make the number of bytes from 'state' included into the hash
6952 independent from the number of PRNG bytes requested.
6954 The first measure alone would be sufficient to avoid
6955 Markku-Juhani's attack. (Actually it had never occurred
6956 to me that the half of 'md_local' used for chaining was the
6957 half from which PRNG output bytes were taken -- I had always
6958 assumed that the secret half would be used.) The second
6959 measure makes sure that additional data from 'state' is never
6960 mixed into 'md_local' in small portions; this heuristically
6961 further strengthens the PRNG.
6964 *) Fix crypto/bn/asm/mips3.s.
6967 *) When only the key is given to "enc", the IV is undefined. Print out
6968 an error message in this case.
6971 *) Handle special case when X509_NAME is empty in X509 printing routines.
6974 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6975 positive and less than q.
6978 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6979 used: it isn't thread safe and the add_lock_callback should handle
6981 [Paul Rose <Paul.Rose@bridge.com>]
6983 *) Verify that incoming data obeys the block size in
6984 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6988 [Ulf Möller, Bodo Möller]
6990 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6991 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6992 when fixing the server behaviour for backwards-compatible 'client
6993 hello' messages. (Note that the attack is impractical against
6994 SSL 3.0 and TLS 1.0 anyway because length and version checking
6995 means that the probability of guessing a valid ciphertext is
6996 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6999 Before 0.9.5, the countermeasure (hide the error by generating a
7000 random 'decryption result') did not work properly because
7001 ERR_clear_error() was missing, meaning that SSL_get_error() would
7002 detect the supposedly ignored error.
7004 Both problems are now fixed.
7007 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7008 (previously it was 1024).
7011 *) Fix for compatibility mode trust settings: ignore trust settings
7012 unless some valid trust or reject settings are present.
7015 *) Fix for blowfish EVP: its a variable length cipher.
7018 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7019 parameters in DSA public key structures and return an error in the
7020 DSA routines if parameters are absent.
7023 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7024 in the current directory if neither $RANDFILE nor $HOME was set.
7025 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7026 caused some confusion to Windows users who haven't defined $HOME.
7027 Thus RAND_file_name() is changed again: e_os.h can define a
7028 DEFAULT_HOME, which will be used if $HOME is not set.
7029 For Windows, we use "C:"; on other platforms, we still require
7030 environment variables.
7032 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7033 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7034 having multiple threads call RAND_poll() concurrently.
7037 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7038 combination of a flag and a thread ID variable.
7039 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7040 flag), *other* threads can enter ssleay_add_bytes without obeying
7041 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7042 that they do not hold after the first thread unsets add_do_not_lock).
7045 *) Change bctest again: '-x' expressions are not available in all
7049 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7051 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7052 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7054 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7055 the default extension for executables, if any. Also, make the perl
7056 scripts that use symlink() to test if it really exists and use "cp"
7057 if it doesn't. All this made OpenSSL compilable and installable in
7061 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7062 If SEQUENCE is length is indefinite just set c->slen to the total
7063 amount of data available.
7064 [Steve Henson, reported by shige@FreeBSD.org]
7065 [This change does not apply to 0.9.7.]
7067 *) Change bctest to avoid here-documents inside command substitution
7068 (workaround for FreeBSD /bin/sh bug).
7069 For compatibility with Ultrix, avoid shell functions (introduced
7070 in the bctest version that searches along $PATH).
7073 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7074 with des_encrypt() defined on some operating systems, like Solaris
7078 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7079 On the Importance of Eliminating Errors in Cryptographic
7080 Computations, J. Cryptology 14 (2001) 2, 101-119,
7081 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7084 *) MIPS assembler BIGNUM division bug fix.
7087 *) Disabled incorrect Alpha assembler code.
7090 *) Fix PKCS#7 decode routines so they correctly update the length
7091 after reading an EOC for the EXPLICIT tag.
7093 [This change does not apply to 0.9.7.]
7095 *) Fix bug in PKCS#12 key generation routines. This was triggered
7096 if a 3DES key was generated with a 0 initial byte. Include
7097 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7098 (but broken) behaviour.
7101 *) Enhance bctest to search for a working bc along $PATH and print
7103 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7105 *) Fix memory leaks in err.c: free err_data string if necessary;
7106 don't write to the wrong index in ERR_set_error_data.
7109 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7113 *) Replace rdtsc with _emit statements for VC++ version 5.
7114 [Jeremy Cooper <jeremy@baymoo.org>]
7116 *) Make it possible to reuse SSLv2 sessions.
7119 *) In copy_email() check for >= 0 as a return value for
7120 X509_NAME_get_index_by_NID() since 0 is a valid index.
7121 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7123 *) Avoid coredump with unsupported or invalid public keys by checking if
7124 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7125 PKCS7_verify() fails with non detached data.
7128 *) Don't use getenv in library functions when run as setuid/setgid.
7129 New function OPENSSL_issetugid().
7132 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7133 due to incorrect handling of multi-threading:
7135 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7137 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7139 3. Count how many times MemCheck_off() has been called so that
7140 nested use can be treated correctly. This also avoids
7141 inband-signalling in the previous code (which relied on the
7142 assumption that thread ID 0 is impossible).
7145 *) Add "-rand" option also to s_client and s_server.
7148 *) Fix CPU detection on Irix 6.x.
7149 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7150 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7152 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7155 [This change does not apply to 0.9.7.]
7157 *) Use the cached encoding of an X509_NAME structure rather than
7158 copying it. This is apparently the reason for the libsafe "errors"
7159 but the code is actually correct.
7162 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7163 Bleichenbacher's DSA attack.
7164 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7165 to be set and top=0 forces the highest bit to be set; top=-1 is new
7166 and leaves the highest bit random.
7167 [Ulf Moeller, Bodo Moeller]
7169 *) In the NCONF_...-based implementations for CONF_... queries
7170 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7171 a temporary CONF structure with the data component set to NULL
7172 (which gives segmentation faults in lh_retrieve).
7173 Instead, use NULL for the CONF pointer in CONF_get_string and
7174 CONF_get_number (which may use environment variables) and directly
7175 return NULL from CONF_get_section.
7178 *) Fix potential buffer overrun for EBCDIC.
7181 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7182 keyUsage if basicConstraints absent for a CA.
7185 *) Make SMIME_write_PKCS7() write mail header values with a format that
7186 is more generally accepted (no spaces before the semicolon), since
7187 some programs can't parse those values properly otherwise. Also make
7188 sure BIO's that break lines after each write do not create invalid
7192 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7193 macros previously used would not encode an empty SEQUENCE OF
7194 and break the signature.
7196 [This change does not apply to 0.9.7.]
7198 *) Zero the premaster secret after deriving the master secret in
7202 *) Add some EVP_add_digest_alias registrations (as found in
7203 OpenSSL_add_all_digests()) to SSL_library_init()
7204 aka OpenSSL_add_ssl_algorithms(). This provides improved
7205 compatibility with peers using X.509 certificates
7206 with unconventional AlgorithmIdentifier OIDs.
7209 *) Fix for Irix with NO_ASM.
7210 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7212 *) ./config script fixes.
7213 [Ulf Moeller, Richard Levitte]
7215 *) Fix 'openssl passwd -1'.
7218 *) Change PKCS12_key_gen_asc() so it can cope with non null
7219 terminated strings whose length is passed in the passlen
7220 parameter, for example from PEM callbacks. This was done
7221 by adding an extra length parameter to asc2uni().
7222 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7224 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7225 call failed, free the DSA structure.
7228 *) Fix to uni2asc() to cope with zero length Unicode strings.
7229 These are present in some PKCS#12 files.
7232 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7233 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7234 when writing a 32767 byte record.
7235 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7237 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7238 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7240 (RSA objects have a reference count access to which is protected
7241 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7242 so they are meant to be shared between threads.)
7243 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7244 "Reddie, Steven" <Steven.Reddie@ca.com>]
7246 *) Fix a deadlock in CRYPTO_mem_leaks().
7249 *) Use better test patterns in bntest.
7252 *) rand_win.c fix for Borland C.
7255 *) BN_rshift bugfix for n == 0.
7258 *) Add a 'bctest' script that checks for some known 'bc' bugs
7259 so that 'make test' does not abort just because 'bc' is broken.
7262 *) Store verify_result within SSL_SESSION also for client side to
7263 avoid potential security hole. (Re-used sessions on the client side
7264 always resulted in verify_result==X509_V_OK, not using the original
7265 result of the server certificate verification.)
7268 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7269 SSL3_RT_APPLICATION_DATA, return 0.
7270 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7274 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7275 releases, have been re-implemented by renaming the previous
7276 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7277 and ssl3_read_internal, respectively, and adding 'peek' parameters
7278 to them. The new ssl[23]_{read,peek} functions are calls to
7279 ssl[23]_read_internal with the 'peek' flag set appropriately.
7280 A 'peek' parameter has also been added to ssl3_read_bytes, which
7281 does the actual work for ssl3_read_internal.
7284 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7285 the method-specific "init()" handler. Also clean up ex_data after
7286 calling the method-specific "finish()" handler. Previously, this was
7287 happening the other way round.
7290 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7291 The previous value, 12, was not always sufficient for BN_mod_exp().
7294 *) Make sure that shared libraries get the internal name engine with
7295 the full version number and not just 0. This should mark the
7296 shared libraries as not backward compatible. Of course, this should
7297 be changed again when we can guarantee backward binary compatibility.
7300 *) Fix typo in get_cert_by_subject() in by_dir.c
7301 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7303 *) Rework the system to generate shared libraries:
7305 - Make note of the expected extension for the shared libraries and
7306 if there is a need for symbolic links from for example libcrypto.so.0
7307 to libcrypto.so.0.9.7. There is extended info in Configure for
7310 - Make as few rebuilds of the shared libraries as possible.
7312 - Still avoid linking the OpenSSL programs with the shared libraries.
7314 - When installing, install the shared libraries separately from the
7318 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7320 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7321 and not in SSL_clear because the latter is also used by the
7322 accept/connect functions; previously, the settings made by
7323 SSL_set_read_ahead would be lost during the handshake.
7324 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7326 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7327 Previously, it would create entries for disableed algorithms no
7331 *) Added several new manual pages for SSL_* function.
7334 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7336 *) In ssl23_get_client_hello, generate an error message when faced
7337 with an initial SSL 3.0/TLS record that is too small to contain the
7338 first two bytes of the ClientHello message, i.e. client_version.
7339 (Note that this is a pathologic case that probably has never happened
7340 in real life.) The previous approach was to use the version number
7341 from the record header as a substitute; but our protocol choice
7342 should not depend on that one because it is not authenticated
7343 by the Finished messages.
7346 *) More robust randomness gathering functions for Windows.
7347 [Jeffrey Altman <jaltman@columbia.edu>]
7349 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7350 not set then we don't setup the error code for issuer check errors
7351 to avoid possibly overwriting other errors which the callback does
7352 handle. If an application does set the flag then we assume it knows
7353 what it is doing and can handle the new informational codes
7357 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7358 a general "ANY" type, as such it should be able to decode anything
7359 including tagged types. However it didn't check the class so it would
7360 wrongly interpret tagged types in the same way as their universal
7361 counterpart and unknown types were just rejected. Changed so that the
7362 tagged and unknown types are handled in the same way as a SEQUENCE:
7363 that is the encoding is stored intact. There is also a new type
7364 "V_ASN1_OTHER" which is used when the class is not universal, in this
7365 case we have no idea what the actual type is so we just lump them all
7369 *) On VMS, stdout may very well lead to a file that is written to
7370 in a record-oriented fashion. That means that every write() will
7371 write a separate record, which will be read separately by the
7372 programs trying to read from it. This can be very confusing.
7374 The solution is to put a BIO filter in the way that will buffer
7375 text until a linefeed is reached, and then write everything a
7376 line at a time, so every record written will be an actual line,
7377 not chunks of lines and not (usually doesn't happen, but I've
7378 seen it once) several lines in one record. BIO_f_linebuffer() is
7381 Currently, it's a VMS-only method, because that's where it has
7382 been tested well enough.
7385 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7386 it can return incorrect results.
7387 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7388 but it was in 0.9.6-beta[12].)
7391 *) Disable the check for content being present when verifying detached
7392 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7393 include zero length content when signing messages.
7396 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7397 BIO_ctrl (for BIO pairs).
7400 *) Add DSO method for VMS.
7403 *) Bug fix: Montgomery multiplication could produce results with the
7407 *) Add RPM specification openssl.spec and modify it to build three
7408 packages. The default package contains applications, application
7409 documentation and run-time libraries. The devel package contains
7410 include files, static libraries and function documentation. The
7411 doc package contains the contents of the doc directory. The original
7412 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7415 *) Add a large number of documentation files for many SSL routines.
7416 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7418 *) Add a configuration entry for Sony News 4.
7419 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7421 *) Don't set the two most significant bits to one when generating a
7422 random number < q in the DSA library.
7425 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7426 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7427 the underlying transport is blocking) if a handshake took place.
7428 (The default behaviour is needed by applications such as s_client
7429 and s_server that use select() to determine when to use SSL_read;
7430 but for applications that know in advance when to expect data, it
7431 just makes things more complicated.)
7434 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7438 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7439 work better on such systems.
7440 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7442 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7443 Update PKCS12_parse() so it copies the friendlyName and the
7444 keyid to the certificates aux info.
7447 *) Fix bug in PKCS7_verify() which caused an infinite loop
7448 if there was more than one signature.
7449 [Sven Uszpelkat <su@celocom.de>]
7451 *) Major change in util/mkdef.pl to include extra information
7452 about each symbol, as well as presentig variables as well
7453 as functions. This change means that there's n more need
7454 to rebuild the .num files when some algorithms are excluded.
7457 *) Allow the verify time to be set by an application,
7458 rather than always using the current time.
7461 *) Phase 2 verify code reorganisation. The certificate
7462 verify code now looks up an issuer certificate by a
7463 number of criteria: subject name, authority key id
7464 and key usage. It also verifies self signed certificates
7465 by the same criteria. The main comparison function is
7466 X509_check_issued() which performs these checks.
7468 Lot of changes were necessary in order to support this
7469 without completely rewriting the lookup code.
7471 Authority and subject key identifier are now cached.
7473 The LHASH 'certs' is X509_STORE has now been replaced
7474 by a STACK_OF(X509_OBJECT). This is mainly because an
7475 LHASH can't store or retrieve multiple objects with
7476 the same hash value.
7478 As a result various functions (which were all internal
7479 use only) have changed to handle the new X509_STORE
7480 structure. This will break anything that messed round
7481 with X509_STORE internally.
7483 The functions X509_STORE_add_cert() now checks for an
7484 exact match, rather than just subject name.
7486 The X509_STORE API doesn't directly support the retrieval
7487 of multiple certificates matching a given criteria, however
7488 this can be worked round by performing a lookup first
7489 (which will fill the cache with candidate certificates)
7490 and then examining the cache for matches. This is probably
7491 the best we can do without throwing out X509_LOOKUP
7492 entirely (maybe later...).
7494 The X509_VERIFY_CTX structure has been enhanced considerably.
7496 All certificate lookup operations now go via a get_issuer()
7497 callback. Although this currently uses an X509_STORE it
7498 can be replaced by custom lookups. This is a simple way
7499 to bypass the X509_STORE hackery necessary to make this
7500 work and makes it possible to use more efficient techniques
7501 in future. A very simple version which uses a simple
7502 STACK for its trusted certificate store is also provided
7503 using X509_STORE_CTX_trusted_stack().
7505 The verify_cb() and verify() callbacks now have equivalents
7506 in the X509_STORE_CTX structure.
7508 X509_STORE_CTX also has a 'flags' field which can be used
7509 to customise the verify behaviour.
7512 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7513 excludes S/MIME capabilities.
7516 *) When a certificate request is read in keep a copy of the
7517 original encoding of the signed data and use it when outputing
7518 again. Signatures then use the original encoding rather than
7519 a decoded, encoded version which may cause problems if the
7520 request is improperly encoded.
7523 *) For consistency with other BIO_puts implementations, call
7524 buffer_write(b, ...) directly in buffer_puts instead of calling
7527 In BIO_puts, increment b->num_write as in BIO_write.
7528 [Peter.Sylvester@EdelWeb.fr]
7530 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7531 BN_zero, we may not return a BIGNUM with an array consisting of
7535 *) Avoid calling abort() from within the library when problems are
7536 detected, except if preprocessor symbols have been defined
7537 (such as REF_CHECK, BN_DEBUG etc.).
7540 *) New openssl application 'rsautl'. This utility can be
7541 used for low level RSA operations. DER public key
7542 BIO/fp routines also added.
7545 *) New Configure entry and patches for compiling on QNX 4.
7546 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7548 *) A demo state-machine implementation was sponsored by
7549 Nuron (http://www.nuron.com/) and is now available in
7550 demos/state_machine.
7553 *) New options added to the 'dgst' utility for signature
7554 generation and verification.
7557 *) Unrecognized PKCS#7 content types are now handled via a
7558 catch all ASN1_TYPE structure. This allows unsupported
7559 types to be stored as a "blob" and an application can
7560 encode and decode it manually.
7563 *) Fix various signed/unsigned issues to make a_strex.c
7565 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7567 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7568 length if passed a buffer. ASN1_INTEGER_to_BN failed
7569 if passed a NULL BN and its argument was negative.
7570 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7572 *) Modification to PKCS#7 encoding routines to output definite
7573 length encoding. Since currently the whole structures are in
7574 memory there's not real point in using indefinite length
7575 constructed encoding. However if OpenSSL is compiled with
7576 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7579 *) Added BIO_vprintf() and BIO_vsnprintf().
7582 *) Added more prefixes to parse for in the the strings written
7583 through a logging bio, to cover all the levels that are available
7584 through syslog. The prefixes are now:
7586 PANIC, EMERG, EMR => LOG_EMERG
7587 ALERT, ALR => LOG_ALERT
7588 CRIT, CRI => LOG_CRIT
7589 ERROR, ERR => LOG_ERR
7590 WARNING, WARN, WAR => LOG_WARNING
7591 NOTICE, NOTE, NOT => LOG_NOTICE
7592 INFO, INF => LOG_INFO
7593 DEBUG, DBG => LOG_DEBUG
7595 and as before, if none of those prefixes are present at the
7596 beginning of the string, LOG_ERR is chosen.
7598 On Win32, the LOG_* levels are mapped according to this:
7600 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7601 LOG_WARNING => EVENTLOG_WARNING_TYPE
7602 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7606 *) Made it possible to reconfigure with just the configuration
7607 argument "reconf" or "reconfigure". The command line arguments
7608 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7609 and are retrieved from there when reconfiguring.
7613 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7615 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7618 *) The obj_dat.pl script was messing up the sorting of object
7619 names. The reason was that it compared the quoted version
7620 of strings as a result "OCSP" > "OCSP Signing" because
7621 " > SPACE. Changed script to store unquoted versions of
7622 names and add quotes on output. It was also omitting some
7623 names from the lookup table if they were given a default
7624 value (that is if SN is missing it is given the same
7625 value as LN and vice versa), these are now added on the
7626 grounds that if an object has a name we should be able to
7627 look it up. Finally added warning output when duplicate
7628 short or long names are found.
7631 *) Changes needed for Tandem NSK.
7632 [Scott Uroff <scott@xypro.com>]
7634 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7635 RSA_padding_check_SSLv23(), special padding was never detected
7636 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7637 version rollback attacks was not effective.
7639 In s23_clnt.c, don't use special rollback-attack detection padding
7640 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7641 client; similarly, in s23_srvr.c, don't do the rollback check if
7642 SSL 2.0 is the only protocol enabled in the server.
7645 *) Make it possible to get hexdumps of unprintable data with 'openssl
7646 asn1parse'. By implication, the functions ASN1_parse_dump() and
7647 BIO_dump_indent() are added.
7650 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7651 these print out strings and name structures based on various
7652 flags including RFC2253 support and proper handling of
7653 multibyte characters. Added options to the 'x509' utility
7654 to allow the various flags to be set.
7657 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7658 Also change the functions X509_cmp_current_time() and
7659 X509_gmtime_adj() work with an ASN1_TIME structure,
7660 this will enable certificates using GeneralizedTime in validity
7661 dates to be checked.
7664 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7665 negative public key encodings) on by default,
7666 NO_NEG_PUBKEY_BUG can be set to disable it.
7669 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7670 content octets. An i2c_ASN1_OBJECT is unnecessary because
7671 the encoding can be trivially obtained from the structure.
7674 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7675 not read locks (CRYPTO_r_[un]lock).
7678 *) A first attempt at creating official support for shared
7679 libraries through configuration. I've kept it so the
7680 default is static libraries only, and the OpenSSL programs
7681 are always statically linked for now, but there are
7682 preparations for dynamic linking in place.
7683 This has been tested on Linux and Tru64.
7686 *) Randomness polling function for Win9x, as described in:
7687 Peter Gutmann, Software Generation of Practically Strong
7691 *) Fix so PRNG is seeded in req if using an already existing
7695 *) New options to smime application. -inform and -outform
7696 allow alternative formats for the S/MIME message including
7697 PEM and DER. The -content option allows the content to be
7698 specified separately. This should allow things like Netscape
7699 form signing output easier to verify.
7702 *) Fix the ASN1 encoding of tags using the 'long form'.
7705 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7706 STRING types. These convert content octets to and from the
7707 underlying type. The actual tag and length octets are
7708 already assumed to have been read in and checked. These
7709 are needed because all other string types have virtually
7710 identical handling apart from the tag. By having versions
7711 of the ASN1 functions that just operate on content octets
7712 IMPLICIT tagging can be handled properly. It also allows
7713 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7714 and ASN1_INTEGER are identical apart from the tag.
7717 *) Change the handling of OID objects as follows:
7719 - New object identifiers are inserted in objects.txt, following
7720 the syntax given in objects.README.
7721 - objects.pl is used to process obj_mac.num and create a new
7723 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7726 This is currently kind of a hack, and the perl code in objects.pl
7727 isn't very elegant, but it works as I intended. The simplest way
7728 to check that it worked correctly is to look in obj_dat.h and
7729 check the array nid_objs and make sure the objects haven't moved
7730 around (this is important!). Additions are OK, as well as
7731 consistent name changes.
7734 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7737 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7738 The given file adds to whatever has already been seeded into the
7739 random pool through the RANDFILE configuration file option or
7740 environment variable, or the default random state file.
7743 *) mkstack.pl now sorts each macro group into lexical order.
7744 Previously the output order depended on the order the files
7745 appeared in the directory, resulting in needless rewriting
7749 *) Patches to make OpenSSL compile under Win32 again. Mostly
7750 work arounds for the VC++ problem that it treats func() as
7751 func(void). Also stripped out the parts of mkdef.pl that
7752 added extra typesafe functions: these no longer exist.
7755 *) Reorganisation of the stack code. The macros are now all
7756 collected in safestack.h . Each macro is defined in terms of
7757 a "stack macro" of the form SKM_<name>(type, a, b). The
7758 DEBUG_SAFESTACK is now handled in terms of function casts,
7759 this has the advantage of retaining type safety without the
7760 use of additional functions. If DEBUG_SAFESTACK is not defined
7761 then the non typesafe macros are used instead. Also modified the
7762 mkstack.pl script to handle the new form. Needs testing to see
7763 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7764 the default if no major problems. Similar behaviour for ASN1_SET_OF
7765 and PKCS12_STACK_OF.
7768 *) When some versions of IIS use the 'NET' form of private key the
7769 key derivation algorithm is different. Normally MD5(password) is
7770 used as a 128 bit RC4 key. In the modified case
7771 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7772 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7773 as the old Netscape_RSA functions except they have an additional
7774 'sgckey' parameter which uses the modified algorithm. Also added
7775 an -sgckey command line option to the rsa utility. Thanks to
7776 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7777 algorithm to openssl-dev.
7780 *) The evp_local.h macros were using 'c.##kname' which resulted in
7781 invalid expansion on some systems (SCO 5.0.5 for example).
7782 Corrected to 'c.kname'.
7783 [Phillip Porch <root@theporch.com>]
7785 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7786 a STACK of email addresses from a certificate or request, these look
7787 in the subject name and the subject alternative name extensions and
7788 omit any duplicate addresses.
7791 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7792 This makes DSA verification about 2 % faster.
7795 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7796 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7797 plus overhead for 1024 bit moduli).
7798 This makes exponentiations about 0.5 % faster for 1024 bit
7799 exponents (as measured by "openssl speed rsa2048").
7802 *) Rename memory handling macros to avoid conflicts with other
7804 Malloc => OPENSSL_malloc
7805 Malloc_locked => OPENSSL_malloc_locked
7806 Realloc => OPENSSL_realloc
7807 Free => OPENSSL_free
7810 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7811 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7814 *) CygWin32 support.
7815 [John Jarvie <jjarvie@newsguy.com>]
7817 *) The type-safe stack code has been rejigged. It is now only compiled
7818 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7819 by default all type-specific stack functions are "#define"d back to
7820 standard stack functions. This results in more streamlined output
7821 but retains the type-safety checking possibilities of the original
7825 *) The STACK code has been cleaned up, and certain type declarations
7826 that didn't make a lot of sense have been brought in line. This has
7827 also involved a cleanup of sorts in safestack.h to more correctly
7828 map type-safe stack functions onto their plain stack counterparts.
7829 This work has also resulted in a variety of "const"ifications of
7830 lots of the code, especially "_cmp" operations which should normally
7831 be prototyped with "const" parameters anyway.
7834 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7835 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7836 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7837 where all of 'md' is used each time the PRNG is used, but 'state'
7838 is used only indexed by a cyclic counter. As entropy may not be
7839 well distributed from the beginning, 'md' is important as a
7840 chaining variable. However, the output function chains only half
7841 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7842 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7843 in all of 'state' being rewritten, with the new values depending
7844 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7847 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7848 the handshake is continued after ssl_verify_cert_chain();
7849 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7850 can lead to 'unexplainable' connection aborts later.
7851 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7853 *) Major EVP API cipher revision.
7854 Add hooks for extra EVP features. This allows various cipher
7855 parameters to be set in the EVP interface. Support added for variable
7856 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7857 setting of RC2 and RC5 parameters.
7859 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7862 Remove lots of duplicated code from the EVP library. For example *every*
7863 cipher init() function handles the 'iv' in the same way according to the
7864 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7865 for CFB and OFB modes they zero ctx->num.
7867 New functionality allows removal of S/MIME code RC2 hack.
7869 Most of the routines have the same form and so can be declared in terms
7872 By shifting this to the top level EVP_CipherInit() it can be removed from
7873 all individual ciphers. If the cipher wants to handle IVs or keys
7874 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7877 Change lots of functions like EVP_EncryptUpdate() to now return a
7878 value: although software versions of the algorithms cannot fail
7879 any installed hardware versions can.
7882 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7883 this option is set, tolerate broken clients that send the negotiated
7884 protocol version number instead of the requested protocol version
7888 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7889 i.e. non-zero for export ciphersuites, zero otherwise.
7890 Previous versions had this flag inverted, inconsistent with
7891 rsa_tmp_cb (..._TMP_RSA_CB).
7892 [Bodo Moeller; problem reported by Amit Chopra]
7894 *) Add missing DSA library text string. Work around for some IIS
7895 key files with invalid SEQUENCE encoding.
7898 *) Add a document (doc/standards.txt) that list all kinds of standards
7899 and so on that are implemented in OpenSSL.
7902 *) Enhance c_rehash script. Old version would mishandle certificates
7903 with the same subject name hash and wouldn't handle CRLs at all.
7904 Added -fingerprint option to crl utility, to support new c_rehash
7908 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7911 *) Fix for SSL server purpose checking. Server checking was
7912 rejecting certificates which had extended key usage present
7913 but no ssl client purpose.
7914 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7916 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7917 is a little unclear about how a blank password is handled.
7918 Since the password in encoded as a BMPString with terminating
7919 double NULL a zero length password would end up as just the
7920 double NULL. However no password at all is different and is
7921 handled differently in the PKCS#12 key generation code. NS
7922 treats a blank password as zero length. MSIE treats it as no
7923 password on export: but it will try both on import. We now do
7924 the same: PKCS12_parse() tries zero length and no password if
7925 the password is set to "" or NULL (NULL is now a valid password:
7926 it wasn't before) as does the pkcs12 application.
7929 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7930 perror when PEM_read_bio_X509_REQ fails, the error message must
7931 be obtained from the error queue.
7934 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7935 it in ERR_remove_state if appropriate, and change ERR_get_state
7936 accordingly to avoid race conditions (this is necessary because
7937 thread_hash is no longer constant once set).
7940 *) Bugfix for linux-elf makefile.one.
7943 *) RSA_get_default_method() will now cause a default
7944 RSA_METHOD to be chosen if one doesn't exist already.
7945 Previously this was only set during a call to RSA_new()
7946 or RSA_new_method(NULL) meaning it was possible for
7947 RSA_get_default_method() to return NULL.
7950 *) Added native name translation to the existing DSO code
7951 that will convert (if the flag to do so is set) filenames
7952 that are sufficiently small and have no path information
7953 into a canonical native form. Eg. "blah" converted to
7954 "libblah.so" or "blah.dll" etc.
7957 *) New function ERR_error_string_n(e, buf, len) which is like
7958 ERR_error_string(e, buf), but writes at most 'len' bytes
7959 including the 0 terminator. For ERR_error_string_n, 'buf'
7961 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7963 *) CONF library reworked to become more general. A new CONF
7964 configuration file reader "class" is implemented as well as a
7965 new functions (NCONF_*, for "New CONF") to handle it. The now
7966 old CONF_* functions are still there, but are reimplemented to
7967 work in terms of the new functions. Also, a set of functions
7968 to handle the internal storage of the configuration data is
7969 provided to make it easier to write new configuration file
7970 reader "classes" (I can definitely see something reading a
7971 configuration file in XML format, for example), called _CONF_*,
7972 or "the configuration storage API"...
7974 The new configuration file reading functions are:
7976 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7977 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7979 NCONF_default, NCONF_WIN32
7981 NCONF_dump_fp, NCONF_dump_bio
7983 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7984 NCONF_new creates a new CONF object. This works in the same way
7985 as other interfaces in OpenSSL, like the BIO interface.
7986 NCONF_dump_* dump the internal storage of the configuration file,
7987 which is useful for debugging. All other functions take the same
7988 arguments as the old CONF_* functions wth the exception of the
7989 first that must be a `CONF *' instead of a `LHASH *'.
7991 To make it easer to use the new classes with the old CONF_* functions,
7992 the function CONF_set_default_method is provided.
7995 *) Add '-tls1' option to 'openssl ciphers', which was already
7996 mentioned in the documentation but had not been implemented.
7997 (This option is not yet really useful because even the additional
7998 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8001 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8002 OpenSSL-based applications) load shared libraries and bind to
8003 them in a portable way.
8004 [Geoff Thorpe, with contributions from Richard Levitte]
8006 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8008 *) Make sure _lrotl and _lrotr are only used with MSVC.
8010 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8011 (the default implementation of RAND_status).
8013 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8014 to '-clrext' (= clear extensions), as intended and documented.
8015 [Bodo Moeller; inconsistency pointed out by Michael Attili
8018 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8019 was larger than the MD block size.
8020 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8022 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8023 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8024 using the passed key: if the passed key was a private key the result
8025 of X509_print(), for example, would be to print out all the private key
8029 *) des_quad_cksum() byte order bug fix.
8030 [Ulf Möller, using the problem description in krb4-0.9.7, where
8031 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8033 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8035 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8037 *) For easily testing in shell scripts whether some command
8038 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8039 returns with exit code 0 iff no command of the given name is available.
8040 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8041 the output goes to stdout and nothing is printed to stderr.
8042 Additional arguments are always ignored.
8044 Since for each cipher there is a command of the same name,
8045 the 'no-cipher' compilation switches can be tested this way.
8047 ('openssl no-XXX' is not able to detect pseudo-commands such
8048 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8051 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8054 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8055 is set; it will be thrown away anyway because each handshake creates
8057 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8058 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8059 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
8060 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8063 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8064 'Q' and 'R' lose their special meanings (quit/renegotiate).
8065 This is part of what -quiet does; unlike -quiet, -ign_eof
8066 does not suppress any output.
8069 *) Add compatibility options to the purpose and trust code. The
8070 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8071 accepts a certificate or CA, this was the previous behaviour,
8072 with all the associated security issues.
8074 X509_TRUST_COMPAT is the old trust behaviour: only and
8075 automatically trust self signed roots in certificate store. A
8076 new trust setting X509_TRUST_DEFAULT is used to specify that
8077 a purpose has no associated trust setting and it should instead
8078 use the value in the default purpose.
8081 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8082 and fix a memory leak.
8085 *) In util/mkerr.pl (which implements 'make errors'), preserve
8086 reason strings from the previous version of the .c file, as
8087 the default to have only downcase letters (and digits) in
8088 automatically generated reasons codes is not always appropriate.
8091 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8092 using strerror. Previously, ERR_reason_error_string() returned
8093 library names as reason strings for SYSerr; but SYSerr is a special
8094 case where small numbers are errno values, not library numbers.
8097 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8098 converts DSA parameters into DH parameters. (When creating parameters,
8099 DSA_generate_parameters is used.)
8102 *) Include 'length' (recommended exponent length) in C code generated
8103 by 'openssl dhparam -C'.
8106 *) The second argument to set_label in perlasm was already being used
8107 so couldn't be used as a "file scope" flag. Moved to third argument
8111 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8112 instead of RAND_bytes for encryption IVs and salts.
8115 *) Include RAND_status() into RAND_METHOD instead of implementing
8116 it only for md_rand.c Otherwise replacing the PRNG by calling
8117 RAND_set_rand_method would be impossible.
8120 *) Don't let DSA_generate_key() enter an infinite loop if the random
8121 number generation fails.
8124 *) New 'rand' application for creating pseudo-random output.
8127 *) Added configuration support for Linux/IA64
8128 [Rolf Haberrecker <rolf@suse.de>]
8130 *) Assembler module support for Mingw32.
8133 *) Shared library support for HPUX (in shlib/).
8134 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8136 *) Shared library support for Solaris gcc.
8137 [Lutz Behnke <behnke@trustcenter.de>]
8139 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8141 *) PKCS7_encrypt() was adding text MIME headers twice because they
8142 were added manually and by SMIME_crlf_copy().
8145 *) In bntest.c don't call BN_rand with zero bits argument.
8146 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8148 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8149 case was implemented. This caused BN_div_recp() to fail occasionally.
8152 *) Add an optional second argument to the set_label() in the perl
8153 assembly language builder. If this argument exists and is set
8154 to 1 it signals that the assembler should use a symbol whose
8155 scope is the entire file, not just the current function. This
8156 is needed with MASM which uses the format label:: for this scope.
8157 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8159 *) Change the ASN1 types so they are typedefs by default. Before
8160 almost all types were #define'd to ASN1_STRING which was causing
8161 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8165 *) Change names of new functions to the new get1/get0 naming
8166 convention: After 'get1', the caller owns a reference count
8167 and has to call ..._free; 'get0' returns a pointer to some
8168 data structure without incrementing reference counters.
8169 (Some of the existing 'get' functions increment a reference
8170 counter, some don't.)
8171 Similarly, 'set1' and 'add1' functions increase reference
8172 counters or duplicate objects.
8175 *) Allow for the possibility of temp RSA key generation failure:
8176 the code used to assume it always worked and crashed on failure.
8179 *) Fix potential buffer overrun problem in BIO_printf().
8180 [Ulf Möller, using public domain code by Patrick Powell; problem
8181 pointed out by David Sacerdote <das33@cornell.edu>]
8183 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8184 RAND_egd() and RAND_status(). In the command line application,
8185 the EGD socket can be specified like a seed file using RANDFILE
8189 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8190 Some CAs (e.g. Verisign) distribute certificates in this form.
8193 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8194 list to exclude them. This means that no special compilation option
8195 is needed to use anonymous DH: it just needs to be included in the
8199 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8200 EVP_MD_type. The old functionality is available in a new macro called
8201 EVP_MD_md(). Change code that uses it and update docs.
8204 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8205 where the 'void *' argument is replaced by a function pointer argument.
8206 Previously 'void *' was abused to point to functions, which works on
8207 many platforms, but is not correct. As these functions are usually
8208 called by macros defined in OpenSSL header files, most source code
8209 should work without changes.
8212 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8213 sections with information on -D... compiler switches used for
8214 compiling the library so that applications can see them. To enable
8215 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8216 must be defined. E.g.,
8217 #define OPENSSL_ALGORITHM_DEFINES
8218 #include <openssl/opensslconf.h>
8219 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8220 [Richard Levitte, Ulf and Bodo Möller]
8222 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8226 *) Change the 'other' type in certificate aux info to a STACK_OF
8227 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8228 the required ASN1 format: arbitrary types determined by an OID.
8231 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8232 argument to 'req'. This is not because the function is newer or
8233 better than others it just uses the work 'NEW' in the certificate
8234 request header lines. Some software needs this.
8237 *) Reorganise password command line arguments: now passwords can be
8238 obtained from various sources. Delete the PEM_cb function and make
8239 it the default behaviour: i.e. if the callback is NULL and the
8240 usrdata argument is not NULL interpret it as a null terminated pass
8241 phrase. If usrdata and the callback are NULL then the pass phrase
8242 is prompted for as usual.
8245 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8246 the support is automatically enabled. The resulting binaries will
8247 autodetect the card and use it if present.
8248 [Ben Laurie and Compaq Inc.]
8250 *) Work around for Netscape hang bug. This sends certificate request
8251 and server done in one record. Since this is perfectly legal in the
8252 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8253 the bugs/SSLv3 entry for more info.
8256 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8259 *) Add -rand argument to smime and pkcs12 applications and read/write
8263 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8266 *) Add command line password options to the remaining applications.
8269 *) Bug fix for BN_div_recp() for numerators with an even number of
8273 *) More tests in bntest.c, and changed test_bn output.
8276 *) ./config recognizes MacOS X now.
8279 *) Bug fix for BN_div() when the first words of num and divsor are
8280 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8283 *) Add support for various broken PKCS#8 formats, and command line
8284 options to produce them.
8287 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8288 get temporary BIGNUMs from a BN_CTX.
8291 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8295 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8296 include a #define from the old name to the new. The original intent
8297 was that statically linked binaries could for example just call
8298 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8299 link with digests. This never worked becayse SSLeay_add_all_digests()
8300 and SSLeay_add_all_ciphers() were in the same source file so calling
8301 one would link with the other. They are now in separate source files.
8304 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8307 *) Use a less unusual form of the Miller-Rabin primality test (it used
8308 a binary algorithm for exponentiation integrated into the Miller-Rabin
8309 loop, our standard modexp algorithms are faster).
8312 *) Support for the EBCDIC character set completed.
8313 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8315 *) Source code cleanups: use const where appropriate, eliminate casts,
8316 use void * instead of char * in lhash.
8319 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8320 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8321 this the server could overwrite ephemeral keys that the client
8325 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8326 using 50 iterations of the Rabin-Miller test.
8328 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8329 iterations of the Rabin-Miller test as required by the appendix
8330 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8331 As BN_is_prime_fasttest includes trial division, DSA parameter
8332 generation becomes much faster.
8334 This implies a change for the callback functions in DSA_is_prime
8335 and DSA_generate_parameters: The callback function is called once
8336 for each positive witness in the Rabin-Miller test, not just
8337 occasionally in the inner loop; and the parameters to the
8338 callback function now provide an iteration count for the outer
8339 loop rather than for the current invocation of the inner loop.
8340 DSA_generate_parameters additionally can call the callback
8341 function with an 'iteration count' of -1, meaning that a
8342 candidate has passed the trial division test (when q is generated
8343 from an application-provided seed, trial division is skipped).
8346 *) New function BN_is_prime_fasttest that optionally does trial
8347 division before starting the Rabin-Miller test and has
8348 an additional BN_CTX * argument (whereas BN_is_prime always
8349 has to allocate at least one BN_CTX).
8350 'callback(1, -1, cb_arg)' is called when a number has passed the
8351 trial division stage.
8354 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8358 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8361 *) New function BN_pseudo_rand().
8364 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8365 bignum version of BN_from_montgomery() with the working code from
8366 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8370 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8371 made it impossible to use the same SSL_SESSION data structure in
8372 SSL2 clients in multiple threads.
8375 *) The return value of RAND_load_file() no longer counts bytes obtained
8376 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8377 to seed the PRNG (previously an explicit byte count was required).
8378 [Ulf Möller, Bodo Möller]
8380 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8381 used (char *) instead of (void *) and had casts all over the place.
8384 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8387 *) Retain source code compatibility for BN_prime_checks macro:
8388 BN_is_prime(..., BN_prime_checks, ...) now uses
8389 BN_prime_checks_for_size to determine the appropriate number of
8390 Rabin-Miller iterations.
8393 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8394 DH_CHECK_P_NOT_SAFE_PRIME.
8395 (Check if this is true? OpenPGP calls them "strong".)
8398 *) Merge the functionality of "dh" and "gendh" programs into a new program
8399 "dhparam". The old programs are retained for now but will handle DH keys
8400 (instead of parameters) in future.
8403 *) Make the ciphers, s_server and s_client programs check the return values
8404 when a new cipher list is set.
8407 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8408 ciphers. Before when the 56bit ciphers were enabled the sorting was
8411 The syntax for the cipher sorting has been extended to support sorting by
8412 cipher-strength (using the strength_bits hard coded in the tables).
8413 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8415 Fix a bug in the cipher-command parser: when supplying a cipher command
8416 string with an "undefined" symbol (neither command nor alphanumeric
8417 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8418 an error is flagged.
8420 Due to the strength-sorting extension, the code of the
8421 ssl_create_cipher_list() function was completely rearranged. I hope that
8422 the readability was also increased :-)
8423 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8425 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8426 for the first serial number and places 2 in the serial number file. This
8427 avoids problems when the root CA is created with serial number zero and
8428 the first user certificate has the same issuer name and serial number
8432 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8433 the new code. Add documentation for this stuff.
8436 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8437 X509_*() to X509at_*() on the grounds that they don't handle X509
8438 structures and behave in an analagous way to the X509v3 functions:
8439 they shouldn't be called directly but wrapper functions should be used
8442 So we also now have some wrapper functions that call the X509at functions
8443 when passed certificate requests. (TO DO: similar things can be done with
8444 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8445 things. Some of these need some d2i or i2d and print functionality
8446 because they handle more complex structures.)
8449 *) Add missing #ifndefs that caused missing symbols when building libssl
8450 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8451 NO_RSA in ssl/s2*.c.
8452 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8454 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8455 has a return value which indicates the quality of the random data
8456 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8457 error queue. New function RAND_pseudo_bytes() generates output that is
8458 guaranteed to be unique but not unpredictable. RAND_add is like
8459 RAND_seed, but takes an extra argument for an entropy estimate
8460 (RAND_seed always assumes full entropy).
8463 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8464 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8465 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8466 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8467 false-positive rate of at most 2^-80 for random input.
8470 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8473 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8474 in the 0.9.5 release), this returns the chain
8475 from an X509_CTX structure with a dup of the stack and all
8476 the X509 reference counts upped: so the stack will exist
8477 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8480 Also make SSL_SESSION_print() print out the verify return
8484 *) Add manpage for the pkcs12 command. Also change the default
8485 behaviour so MAC iteration counts are used unless the new
8486 -nomaciter option is used. This improves file security and
8487 only older versions of MSIE (4.0 for example) need it.
8490 *) Honor the no-xxx Configure options when creating .DEF files.
8493 *) Add PKCS#10 attributes to field table: challengePassword,
8494 unstructuredName and unstructuredAddress. These are taken from
8495 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8496 international characters are used.
8498 More changes to X509_ATTRIBUTE code: allow the setting of types
8499 based on strings. Remove the 'loc' parameter when adding
8500 attributes because these will be a SET OF encoding which is sorted
8504 *) Initial changes to the 'req' utility to allow request generation
8505 automation. This will allow an application to just generate a template
8506 file containing all the field values and have req construct the
8509 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8510 used all over the place including certificate requests and PKCS#7
8511 structures. They are currently handled manually where necessary with
8512 some primitive wrappers for PKCS#7. The new functions behave in a
8513 manner analogous to the X509 extension functions: they allow
8514 attributes to be looked up by NID and added.
8516 Later something similar to the X509V3 code would be desirable to
8517 automatically handle the encoding, decoding and printing of the
8518 more complex types. The string types like challengePassword can
8519 be handled by the string table functions.
8521 Also modified the multi byte string table handling. Now there is
8522 a 'global mask' which masks out certain types. The table itself
8523 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8524 is useful when for example there is only one permissible type
8525 (as in countryName) and using the mask might result in no valid
8529 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8530 SSL_get_peer_finished to allow applications to obtain the latest
8531 Finished messages sent to the peer or expected from the peer,
8532 respectively. (SSL_get_peer_finished is usually the Finished message
8533 actually received from the peer, otherwise the protocol will be aborted.)
8535 As the Finished message are message digests of the complete handshake
8536 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8537 be used for external authentication procedures when the authentication
8538 provided by SSL/TLS is not desired or is not enough.
8541 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8542 the host supports BWX extension and if Compaq C is present on the
8543 $PATH. Just exploiting of the BWX extension results in 20-30%
8544 performance kick for some algorithms, e.g. DES and RC4 to mention
8545 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8549 *) Add support for MS "fast SGC". This is arguably a violation of the
8550 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8551 weak crypto and after checking the certificate is SGC a second one
8552 with strong crypto. MS SGC stops the first handshake after receiving
8553 the server certificate message and sends a second client hello. Since
8554 a server will typically do all the time consuming operations before
8555 expecting any further messages from the client (server key exchange
8556 is the most expensive) there is little difference between the two.
8558 To get OpenSSL to support MS SGC we have to permit a second client
8559 hello message after we have sent server done. In addition we have to
8560 reset the MAC if we do get this second client hello.
8563 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8564 if a DER encoded private key is RSA or DSA traditional format. Changed
8565 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8566 format DER encoded private key. Newer code should use PKCS#8 format which
8567 has the key type encoded in the ASN1 structure. Added DER private key
8568 support to pkcs8 application.
8571 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8572 ciphersuites has been selected (as required by the SSL 3/TLS 1
8573 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8574 is set, we interpret this as a request to violate the specification
8575 (the worst that can happen is a handshake failure, and 'correct'
8576 behaviour would result in a handshake failure anyway).
8579 *) In SSL_CTX_add_session, take into account that there might be multiple
8580 SSL_SESSION structures with the same session ID (e.g. when two threads
8581 concurrently obtain them from an external cache).
8582 The internal cache can handle only one SSL_SESSION with a given ID,
8583 so if there's a conflict, we now throw out the old one to achieve
8587 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8588 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8589 some routines that use cipher OIDs: some ciphers do not have OIDs
8590 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8594 *) Simplify the trust setting structure and code. Now we just have
8595 two sequences of OIDs for trusted and rejected settings. These will
8596 typically have values the same as the extended key usage extension
8597 and any application specific purposes.
8599 The trust checking code now has a default behaviour: it will just
8600 check for an object with the same NID as the passed id. Functions can
8601 be provided to override either the default behaviour or the behaviour
8602 for a given id. SSL client, server and email already have functions
8603 in place for compatibility: they check the NID and also return "trusted"
8604 if the certificate is self signed.
8607 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8608 traditional format into an EVP_PKEY structure.
8611 *) Add a password callback function PEM_cb() which either prompts for
8612 a password if usr_data is NULL or otherwise assumes it is a null
8613 terminated password. Allow passwords to be passed on command line
8614 environment or config files in a few more utilities.
8617 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8618 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8619 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8620 Update documentation.
8623 *) Support for ASN1 "NULL" type. This could be handled before by using
8624 ASN1_TYPE but there wasn't any function that would try to read a NULL
8625 and produce an error if it couldn't. For compatibility we also have
8626 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8627 don't allocate anything because they don't need to.
8630 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8632 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8634 *) Rebuild of the memory allocation routines used by OpenSSL code and
8635 possibly others as well. The purpose is to make an interface that
8636 provide hooks so anyone can build a separate set of allocation and
8637 deallocation routines to be used by OpenSSL, for example memory
8638 pool implementations, or something else, which was previously hard
8639 since Malloc(), Realloc() and Free() were defined as macros having
8640 the values malloc, realloc and free, respectively (except for Win32
8641 compilations). The same is provided for memory debugging code.
8642 OpenSSL already comes with functionality to find memory leaks, but
8643 this gives people a chance to debug other memory problems.
8645 With these changes, a new set of functions and macros have appeared:
8647 CRYPTO_set_mem_debug_functions() [F]
8648 CRYPTO_get_mem_debug_functions() [F]
8649 CRYPTO_dbg_set_options() [F]
8650 CRYPTO_dbg_get_options() [F]
8651 CRYPTO_malloc_debug_init() [M]
8653 The memory debug functions are NULL by default, unless the library
8654 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8655 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8656 gives the standard debugging functions that come with OpenSSL) or
8657 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8658 provided by the library user) must be used. When the standard
8659 debugging functions are used, CRYPTO_dbg_set_options can be used to
8660 request additional information:
8661 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8662 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8664 Also, things like CRYPTO_set_mem_functions will always give the
8665 expected result (the new set of functions is used for allocation
8666 and deallocation) at all times, regardless of platform and compiler
8669 To finish it up, some functions that were never use in any other
8670 way than through macros have a new API and new semantic:
8673 CRYPTO_dbg_realloc()
8676 All macros of value have retained their old syntax.
8677 [Richard Levitte and Bodo Moeller]
8679 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8680 ordering of SMIMECapabilities wasn't in "strength order" and there
8681 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8685 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8686 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8687 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8689 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8690 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8691 functionality to handle multipart/signed properly) and a utility
8692 called 'smime' to call all this stuff. This is based on code I
8693 originally wrote for Celo who have kindly allowed it to be
8694 included in OpenSSL.
8697 *) Add variants des_set_key_checked and des_set_key_unchecked of
8698 des_set_key (aka des_key_sched). Global variable des_check_key
8699 decides which of these is called by des_set_key; this way
8700 des_check_key behaves as it always did, but applications and
8701 the library itself, which was buggy for des_check_key == 1,
8702 have a cleaner way to pick the version they need.
8705 *) New function PKCS12_newpass() which changes the password of a
8709 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8710 dynamic mix. In both cases the ids can be used as an index into the
8711 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8712 functions so they accept a list of the field values and the
8713 application doesn't need to directly manipulate the X509_TRUST
8717 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8721 *) Modify the way the V3 extension code looks up extensions. This now
8722 works in a similar way to the object code: we have some "standard"
8723 extensions in a static table which is searched with OBJ_bsearch()
8724 and the application can add dynamic ones if needed. The file
8725 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8726 updated whenever a new extension is added to the core code and kept
8727 in ext_nid order. There is a simple program 'tabtest.c' which checks
8728 this. New extensions are not added too often so this file can readily
8729 be maintained manually.
8731 There are two big advantages in doing things this way. The extensions
8732 can be looked up immediately and no longer need to be "added" using
8733 X509V3_add_standard_extensions(): this function now does nothing.
8734 [Side note: I get *lots* of email saying the extension code doesn't
8735 work because people forget to call this function]
8736 Also no dynamic allocation is done unless new extensions are added:
8737 so if we don't add custom extensions there is no need to call
8738 X509V3_EXT_cleanup().
8741 *) Modify enc utility's salting as follows: make salting the default. Add a
8742 magic header, so unsalted files fail gracefully instead of just decrypting
8743 to garbage. This is because not salting is a big security hole, so people
8744 should be discouraged from doing it.
8747 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8748 digest to be passed on the command line but it only used this
8749 parameter when signing a certificate. Modified so all relevant
8750 operations are affected by the digest parameter including the
8751 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8752 DSA key was used because it didn't fix the digest.
8755 *) Initial certificate chain verify code. Currently tests the untrusted
8756 certificates for consistency with the verify purpose (which is set
8757 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8759 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8760 this is because it will reject chains with invalid extensions whereas
8761 every previous version of OpenSSL and SSLeay made no checks at all.
8763 Trust code: checks the root CA for the relevant trust settings. Trust
8764 settings have an initial value consistent with the verify purpose: e.g.
8765 if the verify purpose is for SSL client use it expects the CA to be
8766 trusted for SSL client use. However the default value can be changed to
8767 permit custom trust settings: one example of this would be to only trust
8768 certificates from a specific "secure" set of CAs.
8770 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8771 which should be used for version portability: especially since the
8772 verify structure is likely to change more often now.
8774 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8775 to set them. If not set then assume SSL clients will verify SSL servers
8778 Two new options to the verify program: -untrusted allows a set of
8779 untrusted certificates to be passed in and -purpose which sets the
8780 intended purpose of the certificate. If a purpose is set then the
8781 new chain verify code is used to check extension consistency.
8784 *) Support for the authority information access extension.
8787 *) Modify RSA and DSA PEM read routines to transparently handle
8788 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8789 public keys in a format compatible with certificate
8790 SubjectPublicKeyInfo structures. Unfortunately there were already
8791 functions called *_PublicKey_* which used various odd formats so
8792 these are retained for compatibility: however the DSA variants were
8793 never in a public release so they have been deleted. Changed dsa/rsa
8794 utilities to handle the new format: note no releases ever handled public
8795 keys so we should be OK.
8797 The primary motivation for this change is to avoid the same fiasco
8798 that dogs private keys: there are several incompatible private key
8799 formats some of which are standard and some OpenSSL specific and
8800 require various evil hacks to allow partial transparent handling and
8801 even then it doesn't work with DER formats. Given the option anything
8802 other than PKCS#8 should be dumped: but the other formats have to
8803 stay in the name of compatibility.
8805 With public keys and the benefit of hindsight one standard format
8806 is used which works with EVP_PKEY, RSA or DSA structures: though
8807 it clearly returns an error if you try to read the wrong kind of key.
8809 Added a -pubkey option to the 'x509' utility to output the public key.
8810 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8811 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8812 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8813 that do the same as the EVP_PKEY_assign_*() except they up the
8814 reference count of the added key (they don't "swallow" the
8818 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8819 CRLs would fail if the file contained no certificates or no CRLs:
8820 added a new function to read in both types and return the number
8821 read: this means that if none are read it will be an error. The
8822 DER versions of the certificate and CRL reader would always fail
8823 because it isn't possible to mix certificates and CRLs in DER format
8824 without choking one or the other routine. Changed this to just read
8825 a certificate: this is the best we can do. Also modified the code
8826 in apps/verify.c to take notice of return codes: it was previously
8827 attempting to read in certificates from NULL pointers and ignoring
8828 any errors: this is one reason why the cert and CRL reader seemed
8829 to work. It doesn't check return codes from the default certificate
8830 routines: these may well fail if the certificates aren't installed.
8833 *) Code to support otherName option in GeneralName.
8836 *) First update to verify code. Change the verify utility
8837 so it warns if it is passed a self signed certificate:
8838 for consistency with the normal behaviour. X509_verify
8839 has been modified to it will now verify a self signed
8840 certificate if *exactly* the same certificate appears
8841 in the store: it was previously impossible to trust a
8842 single self signed certificate. This means that:
8843 openssl verify ss.pem
8844 now gives a warning about a self signed certificate but
8845 openssl verify -CAfile ss.pem ss.pem
8849 *) For servers, store verify_result in SSL_SESSION data structure
8850 (and add it to external session representation).
8851 This is needed when client certificate verifications fails,
8852 but an application-provided verification callback (set by
8853 SSL_CTX_set_cert_verify_callback) allows accepting the session
8854 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8855 but returns 1): When the session is reused, we have to set
8856 ssl->verify_result to the appropriate error code to avoid
8858 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8860 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8861 case in PKCS7_dataInit() where the signed PKCS7 structure
8862 didn't contain any existing data because it was being created.
8863 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8865 *) Add a salt to the key derivation routines in enc.c. This
8866 forms the first 8 bytes of the encrypted file. Also add a
8867 -S option to allow a salt to be input on the command line.
8870 *) New function X509_cmp(). Oddly enough there wasn't a function
8871 to compare two certificates. We do this by working out the SHA1
8872 hash and comparing that. X509_cmp() will be needed by the trust
8876 *) SSL_get1_session() is like SSL_get_session(), but increments
8877 the reference count in the SSL_SESSION returned.
8878 [Geoff Thorpe <geoff@eu.c2.net>]
8880 *) Fix for 'req': it was adding a null to request attributes.
8881 Also change the X509_LOOKUP and X509_INFO code to handle
8882 certificate auxiliary information.
8885 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8889 *) Add the possibility to add extra information to the memory leak
8890 detecting output, to form tracebacks, showing from where each
8891 allocation was originated: CRYPTO_push_info("constant string") adds
8892 the string plus current file name and line number to a per-thread
8893 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8894 is like calling CYRPTO_pop_info() until the stack is empty.
8895 Also updated memory leak detection code to be multi-thread-safe.
8898 *) Add options -text and -noout to pkcs7 utility and delete the
8899 encryption options which never did anything. Update docs.
8902 *) Add options to some of the utilities to allow the pass phrase
8903 to be included on either the command line (not recommended on
8904 OSes like Unix) or read from the environment. Update the
8905 manpages and fix a few bugs.
8908 *) Add a few manpages for some of the openssl commands.
8911 *) Fix the -revoke option in ca. It was freeing up memory twice,
8912 leaking and not finding already revoked certificates.
8915 *) Extensive changes to support certificate auxiliary information.
8916 This involves the use of X509_CERT_AUX structure and X509_AUX
8917 functions. An X509_AUX function such as PEM_read_X509_AUX()
8918 can still read in a certificate file in the usual way but it
8919 will also read in any additional "auxiliary information". By
8920 doing things this way a fair degree of compatibility can be
8921 retained: existing certificates can have this information added
8922 using the new 'x509' options.
8924 Current auxiliary information includes an "alias" and some trust
8925 settings. The trust settings will ultimately be used in enhanced
8926 certificate chain verification routines: currently a certificate
8927 can only be trusted if it is self signed and then it is trusted
8931 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8932 The problem was that one of the replacement routines had not been working
8933 since SSLeay releases. For now the offending routine has been replaced
8934 with non-optimised assembler. Even so, this now gives around 95%
8935 performance improvement for 1024 bit RSA signs.
8938 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8939 handling. Most clients have the effective key size in bits equal to
8940 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8941 A few however don't do this and instead use the size of the decrypted key
8942 to determine the RC2 key length and the AlgorithmIdentifier to determine
8943 the effective key length. In this case the effective key length can still
8944 be 40 bits but the key length can be 168 bits for example. This is fixed
8945 by manually forcing an RC2 key into the EVP_PKEY structure because the
8946 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8947 the key length and effective key length are equal.
8950 *) Add a bunch of functions that should simplify the creation of
8951 X509_NAME structures. Now you should be able to do:
8952 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8953 and have it automatically work out the correct field type and fill in
8954 the structures. The more adventurous can try:
8955 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8956 and it will (hopefully) work out the correct multibyte encoding.
8959 *) Change the 'req' utility to use the new field handling and multibyte
8960 copy routines. Before the DN field creation was handled in an ad hoc
8961 way in req, ca, and x509 which was rather broken and didn't support
8962 BMPStrings or UTF8Strings. Since some software doesn't implement
8963 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8964 using the dirstring_type option. See the new comment in the default
8965 openssl.cnf for more info.
8968 *) Make crypto/rand/md_rand.c more robust:
8969 - Assure unique random numbers after fork().
8970 - Make sure that concurrent threads access the global counter and
8971 md serializably so that we never lose entropy in them
8972 or use exactly the same state in multiple threads.
8973 Access to the large state is not always serializable because
8974 the additional locking could be a performance killer, and
8975 md should be large enough anyway.
8978 *) New file apps/app_rand.c with commonly needed functionality
8979 for handling the random seed file.
8981 Use the random seed file in some applications that previously did not:
8983 dsaparam -genkey (which also ignored its '-rand' option),
8986 x509 (when signing).
8987 Except on systems with /dev/urandom, it is crucial to have a random
8988 seed file at least for key creation, DSA signing, and for DH exchanges;
8989 for RSA signatures we could do without one.
8991 gendh and gendsa (unlike genrsa) used to read only the first byte
8992 of each file listed in the '-rand' option. The function as previously
8993 found in genrsa is now in app_rand.c and is used by all programs
8994 that support '-rand'.
8997 *) In RAND_write_file, use mode 0600 for creating files;
8998 don't just chmod when it may be too late.
9001 *) Report an error from X509_STORE_load_locations
9002 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9005 *) New function ASN1_mbstring_copy() this copies a string in either
9006 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9007 into an ASN1_STRING type. A mask of permissible types is passed
9008 and it chooses the "minimal" type to use or an error if not type
9012 *) Add function equivalents to the various macros in asn1.h. The old
9013 macros are retained with an M_ prefix. Code inside the library can
9014 use the M_ macros. External code (including the openssl utility)
9015 should *NOT* in order to be "shared library friendly".
9018 *) Add various functions that can check a certificate's extensions
9019 to see if it usable for various purposes such as SSL client,
9020 server or S/MIME and CAs of these types. This is currently
9021 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9022 verification. Also added a -purpose flag to x509 utility to
9023 print out all the purposes.
9026 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9030 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9031 for, obtain and decode and extension and obtain its critical flag.
9032 This allows all the necessary extension code to be handled in a
9033 single function call.
9036 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9037 platforms. See crypto/rc4/rc4_enc.c for further details.
9040 *) New -noout option to asn1parse. This causes no output to be produced
9041 its main use is when combined with -strparse and -out to extract data
9042 from a file (which may not be in ASN.1 format).
9045 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9046 when producing the local key id.
9047 [Richard Levitte <levitte@stacken.kth.se>]
9049 *) New option -dhparam in s_server. This allows a DH parameter file to be
9050 stated explicitly. If it is not stated then it tries the first server
9051 certificate file. The previous behaviour hard coded the filename
9055 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9056 a public key to be input or output. For example:
9057 openssl rsa -in key.pem -pubout -out pubkey.pem
9058 Also added necessary DSA public key functions to handle this.
9061 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9062 in the message. This was handled by allowing
9063 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9064 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9066 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9067 to the end of the strings whereas this didn't. This would cause problems
9068 if strings read with d2i_ASN1_bytes() were later modified.
9069 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9071 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9072 data and it contains EOF it will end up returning an error. This is
9073 caused by input 46 bytes long. The cause is due to the way base64
9074 BIOs find the start of base64 encoded data. They do this by trying a
9075 trial decode on each line until they find one that works. When they
9076 do a flag is set and it starts again knowing it can pass all the
9077 data directly through the decoder. Unfortunately it doesn't reset
9078 the context it uses. This means that if EOF is reached an attempt
9079 is made to pass two EOFs through the context and this causes the
9080 resulting error. This can also cause other problems as well. As is
9081 usual with these problems it takes *ages* to find and the fix is
9082 trivial: move one line.
9083 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9085 *) Ugly workaround to get s_client and s_server working under Windows. The
9086 old code wouldn't work because it needed to select() on sockets and the
9087 tty (for keypresses and to see if data could be written). Win32 only
9088 supports select() on sockets so we select() with a 1s timeout on the
9089 sockets and then see if any characters are waiting to be read, if none
9090 are present then we retry, we also assume we can always write data to
9091 the tty. This isn't nice because the code then blocks until we've
9092 received a complete line of data and it is effectively polling the
9093 keyboard at 1s intervals: however it's quite a bit better than not
9094 working at all :-) A dedicated Windows application might handle this
9095 with an event loop for example.
9098 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9099 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9100 will be called when RSA_sign() and RSA_verify() are used. This is useful
9101 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9102 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9103 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9104 This necessitated the support of an extra signature type NID_md5_sha1
9105 for SSL signatures and modifications to the SSL library to use it instead
9106 of calling RSA_public_decrypt() and RSA_private_encrypt().
9109 *) Add new -verify -CAfile and -CApath options to the crl program, these
9110 will lookup a CRL issuers certificate and verify the signature in a
9111 similar way to the verify program. Tidy up the crl program so it
9112 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9113 less strict. It will now permit CRL extensions even if it is not
9114 a V2 CRL: this will allow it to tolerate some broken CRLs.
9117 *) Initialize all non-automatic variables each time one of the openssl
9118 sub-programs is started (this is necessary as they may be started
9119 multiple times from the "OpenSSL>" prompt).
9120 [Lennart Bang, Bodo Moeller]
9122 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9123 removing all other RSA functionality (this is what NO_RSA does). This
9124 is so (for example) those in the US can disable those operations covered
9125 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9129 *) Non-copying interface to BIO pairs.
9130 (still largely untested)
9133 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9134 ASCII string. This was handled independently in various places before.
9137 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9138 UTF8 strings a character at a time.
9141 *) Use client_version from client hello to select the protocol
9142 (s23_srvr.c) and for RSA client key exchange verification
9143 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9146 *) Add various utility functions to handle SPKACs, these were previously
9147 handled by poking round in the structure internals. Added new function
9148 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9149 print, verify and generate SPKACs. Based on an original idea from
9150 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9153 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9156 *) Allow the config file extension section to be overwritten on the
9157 command line. Based on an original idea from Massimiliano Pala
9158 <madwolf@comune.modena.it>. The new option is called -extensions
9159 and can be applied to ca, req and x509. Also -reqexts to override
9160 the request extensions in req and -crlexts to override the crl extensions
9164 *) Add new feature to the SPKAC handling in ca. Now you can include
9165 the same field multiple times by preceding it by "XXXX." for example:
9168 this is the same syntax as used in the req config file.
9171 *) Allow certificate extensions to be added to certificate requests. These
9172 are specified in a 'req_extensions' option of the req section of the
9173 config file. They can be printed out with the -text option to req but
9174 are otherwise ignored at present.
9177 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9178 data read consists of only the final block it would not decrypted because
9179 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9180 A misplaced 'break' also meant the decrypted final block might not be
9181 copied until the next read.
9184 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9185 a few extra parameters to the DH structure: these will be useful if
9186 for example we want the value of 'q' or implement X9.42 DH.
9189 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9190 provides hooks that allow the default DSA functions or functions on a
9191 "per key" basis to be replaced. This allows hardware acceleration and
9192 hardware key storage to be handled without major modification to the
9193 library. Also added low level modexp hooks and CRYPTO_EX structure and
9194 associated functions.
9197 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9198 as "read only": it can't be written to and the buffer it points to will
9199 not be freed. Reading from a read only BIO is much more efficient than
9200 a normal memory BIO. This was added because there are several times when
9201 an area of memory needs to be read from a BIO. The previous method was
9202 to create a memory BIO and write the data to it, this results in two
9203 copies of the data and an O(n^2) reading algorithm. There is a new
9204 function BIO_new_mem_buf() which creates a read only memory BIO from
9205 an area of memory. Also modified the PKCS#7 routines to use read only
9209 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9210 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9211 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9212 but a retry condition occured while trying to read the rest.
9215 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9216 NID_pkcs7_encrypted by default: this was wrong since this should almost
9217 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9218 the encrypted data type: this is a more sensible place to put it and it
9219 allows the PKCS#12 code to be tidied up that duplicated this
9223 *) Changed obj_dat.pl script so it takes its input and output files on
9224 the command line. This should avoid shell escape redirection problems
9228 *) Initial support for certificate extension requests, these are included
9229 in things like Xenroll certificate requests. Included functions to allow
9230 extensions to be obtained and added.
9233 *) -crlf option to s_client and s_server for sending newlines as
9234 CRLF (as required by many protocols).
9237 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9239 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9240 [Ralf S. Engelschall]
9242 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9243 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9245 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9249 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9250 DH parameters/keys (q is lost during that conversion, but the resulting
9251 DH parameters contain its length).
9253 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9254 much faster than DH_generate_parameters (which creates parameters
9255 where p = 2*q + 1), and also the smaller q makes DH computations
9256 much more efficient (160-bit exponentiation instead of 1024-bit
9257 exponentiation); so this provides a convenient way to support DHE
9258 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9259 utter importance to use
9260 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9262 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9263 when such DH parameters are used, because otherwise small subgroup
9264 attacks may become possible!
9267 *) Avoid memory leak in i2d_DHparams.
9270 *) Allow the -k option to be used more than once in the enc program:
9271 this allows the same encrypted message to be read by multiple recipients.
9274 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9275 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9276 it will always use the numerical form of the OID, even if it has a short
9280 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9281 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9282 otherwise bn_mod_exp was called. In the case of hardware keys for example
9283 no private key components need be present and it might store extra data
9284 in the RSA structure, which cannot be accessed from bn_mod_exp.
9285 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9286 private key operations.
9289 *) Added support for SPARC Linux.
9292 *) pem_password_cb function type incompatibly changed from
9293 typedef int pem_password_cb(char *buf, int size, int rwflag);
9295 ....(char *buf, int size, int rwflag, void *userdata);
9296 so that applications can pass data to their callbacks:
9297 The PEM[_ASN1]_{read,write}... functions and macros now take an
9298 additional void * argument, which is just handed through whenever
9299 the password callback is called.
9300 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9302 New function SSL_CTX_set_default_passwd_cb_userdata.
9304 Compatibility note: As many C implementations push function arguments
9305 onto the stack in reverse order, the new library version is likely to
9306 interoperate with programs that have been compiled with the old
9307 pem_password_cb definition (PEM_whatever takes some data that
9308 happens to be on the stack as its last argument, and the callback
9309 just ignores this garbage); but there is no guarantee whatsoever that
9312 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9313 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9314 problems not only on Windows, but also on some Unix platforms.
9315 To avoid problematic command lines, these definitions are now in an
9316 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9317 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9320 *) MIPS III/IV assembler module is reimplemented.
9323 *) More DES library cleanups: remove references to srand/rand and
9324 delete an unused file.
9327 *) Add support for the the free Netwide assembler (NASM) under Win32,
9328 since not many people have MASM (ml) and it can be hard to obtain.
9329 This is currently experimental but it seems to work OK and pass all
9330 the tests. Check out INSTALL.W32 for info.
9333 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9334 without temporary keys kept an extra copy of the server key,
9335 and connections with temporary keys did not free everything in case
9339 *) New function RSA_check_key and new openssl rsa option -check
9340 for verifying the consistency of RSA keys.
9341 [Ulf Moeller, Bodo Moeller]
9343 *) Various changes to make Win32 compile work:
9344 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9345 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9346 comparison" warnings.
9347 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9350 *) Add a debugging option to PKCS#5 v2 key generation function: when
9351 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9352 derived keys are printed to stderr.
9355 *) Copy the flags in ASN1_STRING_dup().
9356 [Roman E. Pavlov <pre@mo.msk.ru>]
9358 *) The x509 application mishandled signing requests containing DSA
9359 keys when the signing key was also DSA and the parameters didn't match.
9361 It was supposed to omit the parameters when they matched the signing key:
9362 the verifying software was then supposed to automatically use the CA's
9363 parameters if they were absent from the end user certificate.
9365 Omitting parameters is no longer recommended. The test was also
9366 the wrong way round! This was probably due to unusual behaviour in
9367 EVP_cmp_parameters() which returns 1 if the parameters match.
9368 This meant that parameters were omitted when they *didn't* match and
9369 the certificate was useless. Certificates signed with 'ca' didn't have
9371 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9373 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9374 The interface is as follows:
9375 Applications can use
9376 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9377 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9378 "off" is now the default.
9379 The library internally uses
9380 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9381 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9382 to disable memory-checking temporarily.
9384 Some inconsistent states that previously were possible (and were
9385 even the default) are now avoided.
9387 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9388 with each memory chunk allocated; this is occasionally more helpful
9389 than just having a counter.
9391 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9393 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9397 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9398 which largely parallels "options", but is for changing API behaviour,
9399 whereas "options" are about protocol behaviour.
9400 Initial "mode" flags are:
9402 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9403 a single record has been written.
9404 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9405 retries use the same buffer location.
9406 (But all of the contents must be
9410 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9413 *) Fix problems with no-hmac etc.
9414 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9416 *) New functions RSA_get_default_method(), RSA_set_method() and
9417 RSA_get_method(). These allows replacement of RSA_METHODs without having
9418 to mess around with the internals of an RSA structure.
9421 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9422 Also really enable memory leak checks in openssl.c and in some
9424 [Chad C. Mulligan, Bodo Moeller]
9426 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9427 up the length of negative integers. This has now been simplified to just
9428 store the length when it is first determined and use it later, rather
9429 than trying to keep track of where data is copied and updating it to
9431 [Steve Henson, reported by Brien Wheeler
9432 <bwheeler@authentica-security.com>]
9434 *) Add a new function PKCS7_signatureVerify. This allows the verification
9435 of a PKCS#7 signature but with the signing certificate passed to the
9436 function itself. This contrasts with PKCS7_dataVerify which assumes the
9437 certificate is present in the PKCS#7 structure. This isn't always the
9438 case: certificates can be omitted from a PKCS#7 structure and be
9439 distributed by "out of band" means (such as a certificate database).
9442 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9443 function prototypes in pem.h, also change util/mkdef.pl to add the
9444 necessary function names.
9447 *) mk1mf.pl (used by Windows builds) did not properly read the
9448 options set by Configure in the top level Makefile, and Configure
9449 was not even able to write more than one option correctly.
9450 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9453 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9454 file to be loaded from a BIO or FILE pointer. The BIO version will
9455 for example allow memory BIOs to contain config info.
9458 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9459 Whoever hopes to achieve shared-library compatibility across versions
9460 must use this, not the compile-time macro.
9461 (Exercise 0.9.4: Which is the minimum library version required by
9463 Note: All this applies only to multi-threaded programs, others don't
9467 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9468 through a BIO pair triggered the default case, i.e.
9469 SSLerr(...,SSL_R_UNKNOWN_STATE).
9472 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9473 can use the SSL library even if none of the specific BIOs is
9477 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9478 for the encoded length.
9479 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9481 *) Add initial documentation of the X509V3 functions.
9484 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9485 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9486 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9487 secure PKCS#8 private key format with a high iteration count.
9490 *) Fix determination of Perl interpreter: A perl or perl5
9491 _directory_ in $PATH was also accepted as the interpreter.
9492 [Ralf S. Engelschall]
9494 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9495 wrong with it but it was very old and did things like calling
9496 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9500 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9501 to use the new extension code.
9504 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9505 with macros. This should make it easier to change their form, add extra
9506 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9510 *) Add to configuration table a new entry that can specify an alternative
9511 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9512 according to Mark Crispin <MRC@Panda.COM>.
9516 *) DES CBC did not update the IV. Weird.
9519 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9520 Changing the behaviour of the former might break existing programs --
9521 where IV updating is needed, des_ncbc_encrypt can be used.
9524 *) When bntest is run from "make test" it drives bc to check its
9525 calculations, as well as internally checking them. If an internal check
9526 fails, it needs to cause bc to give a non-zero result or make test carries
9527 on without noticing the failure. Fixed.
9530 *) DES library cleanups.
9533 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9534 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9535 ciphers. NOTE: although the key derivation function has been verified
9536 against some published test vectors it has not been extensively tested
9537 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9541 *) Instead of "mkdir -p", which is not fully portable, use new
9542 Perl script "util/mkdir-p.pl".
9545 *) Rewrite the way password based encryption (PBE) is handled. It used to
9546 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9547 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9548 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9549 the 'parameter' field of the AlgorithmIdentifier is passed to the
9550 underlying key generation function so it must do its own ASN1 parsing.
9551 This has also changed the EVP_PBE_CipherInit() function which now has a
9552 'parameter' argument instead of literal salt and iteration count values
9553 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9556 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9557 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9558 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9559 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9560 value was just used as a "magic string" and not used directly its
9561 value doesn't matter.
9564 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9568 *) "linux-sparc64" configuration (ultrapenguin).
9569 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9570 "linux-sparc" configuration.
9571 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9573 *) config now generates no-xxx options for missing ciphers.
9576 *) Support the EBCDIC character set (work in progress).
9577 File ebcdic.c not yet included because it has a different license.
9578 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9580 *) Support BS2000/OSD-POSIX.
9581 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9583 *) Make callbacks for key generation use void * instead of char *.
9586 *) Make S/MIME samples compile (not yet tested).
9589 *) Additional typesafe stacks.
9592 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9596 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9598 *) New configuration variant "sco5-gcc".
9600 *) Updated some demos.
9601 [Sean O Riordain, Wade Scholine]
9603 *) Add missing BIO_free at exit of pkcs12 application.
9606 *) Fix memory leak in conf.c.
9609 *) Updates for Win32 to assembler version of MD5.
9612 *) Set #! path to perl in apps/der_chop to where we found it
9613 instead of using a fixed path.
9616 *) SHA library changes for irix64-mips4-cc.
9619 *) Improvements for VMS support.
9623 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9625 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9626 This also avoids the problems with SC4.2 and unpatched SC5.
9627 [Andy Polyakov <appro@fy.chalmers.se>]
9629 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9630 These are required because of the typesafe stack would otherwise break
9631 existing code. If old code used a structure member which used to be STACK
9632 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9633 sk_num or sk_value it would produce an error because the num, data members
9634 are not present in STACK_OF. Now it just produces a warning. sk_set
9635 replaces the old method of assigning a value to sk_value
9636 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9637 that does this will no longer work (and should use sk_set instead) but
9638 this could be regarded as a "questionable" behaviour anyway.
9641 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9642 correctly handle encrypted S/MIME data.
9645 *) Change type of various DES function arguments from des_cblock
9646 (which means, in function argument declarations, pointer to char)
9647 to des_cblock * (meaning pointer to array with 8 char elements),
9648 which allows the compiler to do more typechecking; it was like
9649 that back in SSLeay, but with lots of ugly casts.
9651 Introduce new type const_des_cblock.
9654 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9655 problems: find RecipientInfo structure that matches recipient certificate
9656 and initialise the ASN1 structures properly based on passed cipher.
9659 *) Belatedly make the BN tests actually check the results.
9662 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9663 to and from BNs: it was completely broken. New compilation option
9664 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9665 key elements as negative integers.
9668 *) Reorganize and speed up MD5.
9669 [Andy Polyakov <appro@fy.chalmers.se>]
9672 [Richard Levitte <richard@levitte.org>]
9674 *) New option -out to asn1parse to allow the parsed structure to be
9675 output to a file. This is most useful when combined with the -strparse
9676 option to examine the output of things like OCTET STRINGS.
9679 *) Make SSL library a little more fool-proof by not requiring any longer
9680 that SSL_set_{accept,connect}_state be called before
9681 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9682 in many applications because usually everything *appeared* to work as
9683 intended anyway -- now it really works as intended).
9686 *) Move openssl.cnf out of lib/.
9689 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9690 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9691 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9692 [Ralf S. Engelschall]
9694 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9695 handle PKCS#7 enveloped data properly.
9696 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9698 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9699 copying pointers. The cert_st handling is changed by this in
9700 various ways (and thus what used to be known as ctx->default_cert
9701 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9702 any longer when s->cert does not give us what we need).
9703 ssl_cert_instantiate becomes obsolete by this change.
9704 As soon as we've got the new code right (possibly it already is?),
9705 we have solved a couple of bugs of the earlier code where s->cert
9706 was used as if it could not have been shared with other SSL structures.
9708 Note that using the SSL API in certain dirty ways now will result
9709 in different behaviour than observed with earlier library versions:
9710 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9711 does not influence s as it used to.
9713 In order to clean up things more thoroughly, inside SSL_SESSION
9714 we don't use CERT any longer, but a new structure SESS_CERT
9715 that holds per-session data (if available); currently, this is
9716 the peer's certificate chain and, for clients, the server's certificate
9717 and temporary key. CERT holds only those values that can have
9718 meaningful defaults in an SSL_CTX.
9721 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9722 from the internal representation. Various PKCS#7 fixes: remove some
9723 evil casts and set the enc_dig_alg field properly based on the signing
9727 *) Allow PKCS#12 password to be set from the command line or the
9728 environment. Let 'ca' get its config file name from the environment
9729 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9733 *) Allow certificate policies extension to use an IA5STRING for the
9734 organization field. This is contrary to the PKIX definition but
9735 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9739 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9740 without disallowing inline assembler and the like for non-pedantic builds.
9743 *) Support Borland C++ builder.
9744 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9749 *) SHA-1 cleanups and performance enhancements.
9750 [Andy Polyakov <appro@fy.chalmers.se>]
9752 *) Sparc v8plus assembler for the bignum library.
9753 [Andy Polyakov <appro@fy.chalmers.se>]
9755 *) Accept any -xxx and +xxx compiler options in Configure.
9758 *) Update HPUX configuration.
9761 *) Add missing sk_<type>_unshift() function to safestack.h
9762 [Ralf S. Engelschall]
9764 *) New function SSL_CTX_use_certificate_chain_file that sets the
9765 "extra_cert"s in addition to the certificate. (This makes sense
9766 only for "PEM" format files, as chains as a whole are not
9770 *) Support verify_depth from the SSL API.
9771 x509_vfy.c had what can be considered an off-by-one-error:
9772 Its depth (which was not part of the external interface)
9773 was actually counting the number of certificates in a chain;
9774 now it really counts the depth.
9777 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9778 instead of X509err, which often resulted in confusing error
9779 messages since the error codes are not globally unique
9780 (e.g. an alleged error in ssl3_accept when a certificate
9781 didn't match the private key).
9783 *) New function SSL_CTX_set_session_id_context that allows to set a default
9784 value (so that you don't need SSL_set_session_id_context for each
9785 connection using the SSL_CTX).
9788 *) OAEP decoding bug fix.
9791 *) Support INSTALL_PREFIX for package builders, as proposed by
9795 *) New Configure options "threads" and "no-threads". For systems
9796 where the proper compiler options are known (currently Solaris
9797 and Linux), "threads" is the default.
9800 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9803 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9804 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9805 such as /usr/local/bin.
9808 *) "make linux-shared" to build shared libraries.
9809 [Niels Poppe <niels@netbox.org>]
9811 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9814 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9815 extension adding in x509 utility.
9818 *) Remove NOPROTO sections and error code comments.
9821 *) Partial rewrite of the DEF file generator to now parse the ANSI
9825 *) New Configure options --prefix=DIR and --openssldir=DIR.
9828 *) Complete rewrite of the error code script(s). It is all now handled
9829 by one script at the top level which handles error code gathering,
9830 header rewriting and C source file generation. It should be much better
9831 than the old method: it now uses a modified version of Ulf's parser to
9832 read the ANSI prototypes in all header files (thus the old K&R definitions
9833 aren't needed for error creation any more) and do a better job of
9834 translating function codes into names. The old 'ASN1 error code imbedded
9835 in a comment' is no longer necessary and it doesn't use .err files which
9836 have now been deleted. Also the error code call doesn't have to appear all
9837 on one line (which resulted in some large lines...).
9840 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9843 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9844 0 (which usually indicates a closed connection), but continue reading.
9847 *) Fix some race conditions.
9850 *) Add support for CRL distribution points extension. Add Certificate
9851 Policies and CRL distribution points documentation.
9854 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9857 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9858 8 of keying material. Merlin has also confirmed interop with this fix
9859 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9860 [Merlin Hughes <merlin@baltimore.ie>]
9862 *) Fix lots of warnings.
9863 [Richard Levitte <levitte@stacken.kth.se>]
9865 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9866 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9867 [Richard Levitte <levitte@stacken.kth.se>]
9869 *) Fix problems with sizeof(long) == 8.
9870 [Andy Polyakov <appro@fy.chalmers.se>]
9872 *) Change functions to ANSI C.
9875 *) Fix typos in error codes.
9876 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9878 *) Remove defunct assembler files from Configure.
9881 *) SPARC v8 assembler BIGNUM implementation.
9882 [Andy Polyakov <appro@fy.chalmers.se>]
9884 *) Support for Certificate Policies extension: both print and set.
9885 Various additions to support the r2i method this uses.
9888 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9889 return a const string when you are expecting an allocated buffer.
9892 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9893 types DirectoryString and DisplayText.
9896 *) Add code to allow r2i extensions to access the configuration database,
9897 add an LHASH database driver and add several ctx helper functions.
9900 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9901 fail when they extended the size of a BIGNUM.
9904 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9905 support typesafe stack.
9908 *) Fix typo in SSL_[gs]et_options().
9909 [Nils Frostberg <nils@medcom.se>]
9911 *) Delete various functions and files that belonged to the (now obsolete)
9912 old X509V3 handling code.
9915 *) New Configure option "rsaref".
9918 *) Don't auto-generate pem.h.
9921 *) Introduce type-safe ASN.1 SETs.
9924 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9925 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9927 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9928 that links with OpenSSL (well at least cause lots of warnings), but fear
9929 not: the conversion is trivial, and it eliminates loads of evil casts. A
9930 few STACKed things have been converted already. Feel free to convert more.
9931 In the fullness of time, I'll do away with the STACK type altogether.
9934 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9935 specified in <certfile> by updating the entry in the index.txt file.
9936 This way one no longer has to edit the index.txt file manually for
9937 revoking a certificate. The -revoke option does the gory details now.
9938 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9940 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9941 `-text' option at all and this way the `-noout -text' combination was
9942 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9943 [Ralf S. Engelschall]
9945 *) Make sure a corresponding plain text error message exists for the
9946 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9947 verify callback function determined that a certificate was revoked.
9948 [Ralf S. Engelschall]
9950 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9951 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9952 all available cipers including rc5, which was forgotten until now.
9953 In order to let the testing shell script know which algorithms
9954 are available, a new (up to now undocumented) command
9955 "openssl list-cipher-commands" is used.
9958 *) Bugfix: s_client occasionally would sleep in select() when
9959 it should have checked SSL_pending() first.
9962 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9963 the raw DSA values prior to ASN.1 encoding.
9966 *) Tweaks to Configure
9967 [Niels Poppe <niels@netbox.org>]
9969 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9973 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9976 *) New config option to avoid instructions that are illegal on the 80386.
9977 The default code is faster, but requires at least a 486.
9980 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9981 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9982 same as SSL2_VERSION anyway.
9985 *) New "-showcerts" option for s_client.
9988 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9989 application. Various cleanups and fixes.
9992 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9993 modify error routines to work internally. Add error codes and PBE init
9994 to library startup routines.
9997 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9998 packing functions to asn1 and evp. Changed function names and error
9999 codes along the way.
10002 *) PKCS12 integration: and so it begins... First of several patches to
10003 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10004 objects to objects.h
10007 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10008 and display support for Thawte strong extranet extension.
10011 *) Add LinuxPPC support.
10012 [Jeff Dubrule <igor@pobox.org>]
10014 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10015 bn_div_words in alpha.s.
10016 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10018 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10019 OAEP isn't supported when OpenSSL is built with RSAref.
10020 [Ulf Moeller <ulf@fitug.de>]
10022 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10023 so they no longer are missing under -DNOPROTO.
10024 [Soren S. Jorvang <soren@t.dk>]
10027 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10029 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10030 doesn't work when the session is reused. Coming soon!
10033 *) Fix a security hole, that allows sessions to be reused in the wrong
10034 context thus bypassing client cert protection! All software that uses
10035 client certs and session caches in multiple contexts NEEDS PATCHING to
10036 allow session reuse! A fuller solution is in the works.
10037 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10039 *) Some more source tree cleanups (removed obsolete files
10040 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10041 permission on "config" script to be executable) and a fix for the INSTALL
10043 [Ulf Moeller <ulf@fitug.de>]
10045 *) Remove some legacy and erroneous uses of malloc, free instead of
10047 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10049 *) Make rsa_oaep_test return non-zero on error.
10050 [Ulf Moeller <ulf@fitug.de>]
10052 *) Add support for native Solaris shared libraries. Configure
10053 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10054 if someone would make that last step automatic.
10055 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10057 *) ctx_size was not built with the right compiler during "make links". Fixed.
10060 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10061 except NULL ciphers". This means the default cipher list will no longer
10062 enable NULL ciphers. They need to be specifically enabled e.g. with
10063 the string "DEFAULT:eNULL".
10066 *) Fix to RSA private encryption routines: if p < q then it would
10067 occasionally produce an invalid result. This will only happen with
10068 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10071 *) Be less restrictive and allow also `perl util/perlpath.pl
10072 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10073 because this way one can also use an interpreter named `perl5' (which is
10074 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10075 installed as `perl').
10076 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10078 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10079 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10081 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10082 advapi32.lib to Win32 build and change the pem test comparision
10083 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10084 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10085 and crypto/des/ede_cbcm_enc.c.
10088 *) DES quad checksum was broken on big-endian architectures. Fixed.
10091 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10092 Win32 test batch file so it (might) work again. The Win32 test batch file
10093 is horrible: I feel ill....
10096 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10097 in e_os.h. Audit of header files to check ANSI and non ANSI
10098 sections: 10 functions were absent from non ANSI section and not exported
10099 from Windows DLLs. Fixed up libeay.num for new functions.
10102 *) Make `openssl version' output lines consistent.
10103 [Ralf S. Engelschall]
10105 *) Fix Win32 symbol export lists for BIO functions: Added
10106 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10107 to ms/libeay{16,32}.def.
10108 [Ralf S. Engelschall]
10110 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10111 fine under Unix and passes some trivial tests I've now added. But the
10112 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10113 added to make sure no one expects that this stuff really works in the
10114 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10115 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10117 [Ralf S. Engelschall]
10119 *) Fix the generation of two part addresses in perl.
10120 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10122 *) Add config entry for Linux on MIPS.
10123 [John Tobey <jtobey@channel1.com>]
10125 *) Make links whenever Configure is run, unless we are on Windoze.
10128 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10129 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10133 *) Add a useful kludge to allow package maintainers to specify compiler and
10134 other platforms details on the command line without having to patch the
10135 Configure script everytime: One now can use ``perl Configure
10136 <id>:<details>'', i.e. platform ids are allowed to have details appended
10137 to them (seperated by colons). This is treated as there would be a static
10138 pre-configured entry in Configure's %table under key <id> with value
10139 <details> and ``perl Configure <id>'' is called. So, when you want to
10140 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10141 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10142 now, which overrides the FreeBSD-elf entry on-the-fly.
10143 [Ralf S. Engelschall]
10145 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10148 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10149 on the `perl Configure ...' command line. This way one can compile
10150 OpenSSL libraries with Position Independent Code (PIC) which is needed
10151 for linking it into DSOs.
10152 [Ralf S. Engelschall]
10154 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10158 *) Cleaned up the LICENSE document: The official contact for any license
10159 questions now is the OpenSSL core team under openssl-core@openssl.org.
10160 And add a paragraph about the dual-license situation to make sure people
10161 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10162 to the OpenSSL toolkit.
10163 [Ralf S. Engelschall]
10165 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10166 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10167 Additonally cleaned up the `make links' target: Remove unnecessary
10168 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10169 to speed processing and no longer clutter the display with confusing
10170 stuff. Instead only the actually done links are displayed.
10171 [Ralf S. Engelschall]
10173 *) Permit null encryption ciphersuites, used for authentication only. It used
10174 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10175 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10179 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10180 signed attributes when verifying signatures (this would break them),
10181 the detached data encoding was wrong and public keys obtained using
10182 X509_get_pubkey() weren't freed.
10185 *) Add text documentation for the BUFFER functions. Also added a work around
10186 to a Win95 console bug. This was triggered by the password read stuff: the
10187 last character typed gets carried over to the next fread(). If you were
10188 generating a new cert request using 'req' for example then the last
10189 character of the passphrase would be CR which would then enter the first
10193 *) Added the new `Includes OpenSSL Cryptography Software' button as
10194 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10195 button and can be used by applications based on OpenSSL to show the
10196 relationship to the OpenSSL project.
10197 [Ralf S. Engelschall]
10199 *) Remove confusing variables in function signatures in files
10200 ssl/ssl_lib.c and ssl/ssl.h.
10201 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10203 *) Don't install bss_file.c under PREFIX/include/
10204 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10206 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10207 functions that return function pointers and has support for NT specific
10208 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10209 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10210 unsigned to signed types: this was killing the Win32 compile.
10213 *) Add new certificate file to stack functions,
10214 SSL_add_dir_cert_subjects_to_stack() and
10215 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10216 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10217 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10218 This means that Apache-SSL and similar packages don't have to mess around
10219 to add as many CAs as they want to the preferred list.
10222 *) Experiment with doxygen documentation. Currently only partially applied to
10224 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10225 openssl.doxy as the configuration file.
10228 *) Get rid of remaining C++-style comments which strict C compilers hate.
10229 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10231 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10232 compiled in by default: it has problems with large keys.
10235 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10236 DH private keys and/or callback functions which directly correspond to
10237 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10238 is needed for applications which have to configure certificates on a
10239 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10241 For the RSA certificate situation is makes no difference, but
10242 for the DSA certificate situation this fixes the "no shared cipher"
10243 problem where the OpenSSL cipher selection procedure failed because the
10244 temporary keys were not overtaken from the context and the API provided
10245 no way to reconfigure them.
10246 The new functions now let applications reconfigure the stuff and they
10247 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10248 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10249 non-public-API function ssl_cert_instantiate() is used as a helper
10250 function and also to reduce code redundancy inside ssl_rsa.c.
10251 [Ralf S. Engelschall]
10253 *) Move s_server -dcert and -dkey options out of the undocumented feature
10254 area because they are useful for the DSA situation and should be
10255 recognized by the users.
10256 [Ralf S. Engelschall]
10258 *) Fix the cipher decision scheme for export ciphers: the export bits are
10259 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10260 SSL_EXP_MASK. So, the original variable has to be used instead of the
10261 already masked variable.
10262 [Richard Levitte <levitte@stacken.kth.se>]
10264 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10265 [Richard Levitte <levitte@stacken.kth.se>]
10267 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10268 from `int' to `unsigned int' because it's a length and initialized by
10269 EVP_DigestFinal() which expects an `unsigned int *'.
10270 [Richard Levitte <levitte@stacken.kth.se>]
10272 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10273 script. Instead use the usual Shell->Perl transition trick.
10274 [Ralf S. Engelschall]
10276 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10277 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10278 -noout -modulus' as it's already the case for `openssl rsa -noout
10279 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10280 currently the public key is printed (a decision which was already done by
10281 `openssl dsa -modulus' in the past) which serves a similar purpose.
10282 Additionally the NO_RSA no longer completely removes the whole -modulus
10283 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10285 [Ralf S. Engelschall]
10287 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10288 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10289 [Arne Ansper <arne@ats.cyber.ee>]
10291 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10292 to be added. Now both 'req' and 'ca' can use new objects defined in the
10296 *) Add cool BIO that does syslog (or event log on NT).
10297 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10299 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10300 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10301 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10302 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10305 *) Add preliminary config info for new extension code.
10308 *) Make RSA_NO_PADDING really use no padding.
10309 [Ulf Moeller <ulf@fitug.de>]
10311 *) Generate errors when private/public key check is done.
10314 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10315 for some CRL extensions and new objects added.
10318 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10319 key usage extension and fuller support for authority key id.
10322 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10323 padding method for RSA, which is recommended for new applications in PKCS
10324 #1 v2.0 (RFC 2437, October 1998).
10325 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10326 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10327 against Bleichbacher's attack on RSA.
10328 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10331 *) Updates to the new SSL compression code
10332 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10334 *) Fix so that the version number in the master secret, when passed
10335 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10336 (because the server will not accept higher), that the version number
10337 is 0x03,0x01, not 0x03,0x00
10338 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10340 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10341 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10342 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10345 *) Support for RAW extensions where an arbitrary extension can be
10346 created by including its DER encoding. See apps/openssl.cnf for
10350 *) Make sure latest Perl versions don't interpret some generated C array
10351 code as Perl array code in the crypto/err/err_genc.pl script.
10352 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10354 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10355 not many people have the assembler. Various Win32 compilation fixes and
10356 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10357 build instructions.
10360 *) Modify configure script 'Configure' to automatically create crypto/date.h
10361 file under Win32 and also build pem.h from pem.org. New script
10362 util/mkfiles.pl to create the MINFO file on environments that can't do a
10363 'make files': perl util/mkfiles.pl >MINFO should work.
10366 *) Major rework of DES function declarations, in the pursuit of correctness
10367 and purity. As a result, many evil casts evaporated, and some weirdness,
10368 too. You may find this causes warnings in your code. Zapping your evil
10369 casts will probably fix them. Mostly.
10372 *) Fix for a typo in asn1.h. Bug fix to object creation script
10373 obj_dat.pl. It considered a zero in an object definition to mean
10374 "end of object": none of the objects in objects.h have any zeros
10375 so it wasn't spotted.
10376 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10378 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10379 Masking (CBCM). In the absence of test vectors, the best I have been able
10380 to do is check that the decrypt undoes the encrypt, so far. Send me test
10381 vectors if you have them.
10384 *) Correct calculation of key length for export ciphers (too much space was
10385 allocated for null ciphers). This has not been tested!
10388 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10389 message is now correct (it understands "crypto" and "ssl" on its
10390 command line). There is also now an "update" option. This will update
10391 the util/ssleay.num and util/libeay.num files with any new functions.
10393 perl util/mkdef.pl crypto ssl update
10394 it will update them.
10397 *) Overhauled the Perl interface (perl/*):
10398 - ported BN stuff to OpenSSL's different BN library
10399 - made the perl/ source tree CVS-aware
10400 - renamed the package from SSLeay to OpenSSL (the files still contain
10401 their history because I've copied them in the repository)
10402 - removed obsolete files (the test scripts will be replaced
10403 by better Test::Harness variants in the future)
10404 [Ralf S. Engelschall]
10406 *) First cut for a very conservative source tree cleanup:
10407 1. merge various obsolete readme texts into doc/ssleay.txt
10408 where we collect the old documents and readme texts.
10409 2. remove the first part of files where I'm already sure that we no
10410 longer need them because of three reasons: either they are just temporary
10411 files which were left by Eric or they are preserved original files where
10412 I've verified that the diff is also available in the CVS via "cvs diff
10413 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10414 the crypto/md/ stuff).
10415 [Ralf S. Engelschall]
10417 *) More extension code. Incomplete support for subject and issuer alt
10418 name, issuer and authority key id. Change the i2v function parameters
10419 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10420 what that's for :-) Fix to ASN1 macro which messed up
10421 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10424 *) Preliminary support for ENUMERATED type. This is largely copied from the
10428 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10429 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10431 *) Make sure `make rehash' target really finds the `openssl' program.
10432 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10434 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10435 like to hear about it if this slows down other processors.
10438 *) Add CygWin32 platform information to Configure script.
10439 [Alan Batie <batie@aahz.jf.intel.com>]
10441 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10442 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10444 *) New program nseq to manipulate netscape certificate sequences
10447 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10451 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10452 but the BN code had some problems that would cause failures when
10453 doing certificate verification and some other functions.
10454 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10456 *) Add ASN1 and PEM code to support netscape certificate sequences.
10459 *) Add ASN1 and PEM code to support netscape certificate sequences.
10462 *) Add several PKIX and private extended key usage OIDs.
10465 *) Modify the 'ca' program to handle the new extension code. Modify
10466 openssl.cnf for new extension format, add comments.
10469 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10470 and add a sample to openssl.cnf so req -x509 now adds appropriate
10474 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10475 error code, add initial support to X509_print() and x509 application.
10478 *) Takes a deep breath and start addding X509 V3 extension support code. Add
10479 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10480 stuff is currently isolated and isn't even compiled yet.
10483 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10484 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10485 Removed the versions check from X509 routines when loading extensions:
10486 this allows certain broken certificates that don't set the version
10487 properly to be processed.
10490 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10491 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10492 can still be regenerated with "make depend".
10495 *) Spelling mistake in C version of CAST-128.
10496 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10498 *) Changes to the error generation code. The perl script err-code.pl
10499 now reads in the old error codes and retains the old numbers, only
10500 adding new ones if necessary. It also only changes the .err files if new
10501 codes are added. The makefiles have been modified to only insert errors
10502 when needed (to avoid needlessly modifying header files). This is done
10503 by only inserting errors if the .err file is newer than the auto generated
10504 C file. To rebuild all the error codes from scratch (the old behaviour)
10505 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10506 or delete all the .err files.
10509 *) CAST-128 was incorrectly implemented for short keys. The C version has
10510 been fixed, but is untested. The assembler versions are also fixed, but
10511 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10512 to regenerate it if needed.
10513 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10514 Hagino <itojun@kame.net>]
10516 *) File was opened incorrectly in randfile.c.
10517 [Ulf Möller <ulf@fitug.de>]
10519 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10520 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10521 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10522 al: it's just almost always a UTCTime. Note this patch adds new error
10523 codes so do a "make errors" if there are problems.
10526 *) Correct Linux 1 recognition in config.
10527 [Ulf Möller <ulf@fitug.de>]
10529 *) Remove pointless MD5 hash when using DSA keys in ca.
10530 [Anonymous <nobody@replay.com>]
10532 *) Generate an error if given an empty string as a cert directory. Also
10533 generate an error if handed NULL (previously returned 0 to indicate an
10534 error, but didn't set one).
10535 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10537 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10540 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10541 parameters. This was causing a warning which killed off the Win32 compile.
10544 *) Remove C++ style comments from crypto/bn/bn_local.h.
10545 [Neil Costigan <neil.costigan@celocom.com>]
10547 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10548 based on a text string, looking up short and long names and finally
10549 "dot" format. The "dot" format stuff didn't work. Added new function
10550 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10551 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10552 OID is not part of the table.
10555 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10556 X509_LOOKUP_by_alias().
10559 *) Sort openssl functions by name.
10562 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10563 encryption from sample DSA keys (in case anyone is interested the password
10567 *) Make _all_ *_free functions accept a NULL pointer.
10568 [Frans Heymans <fheymans@isaserver.be>]
10570 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10572 [Anonymous <nobody@replay.com>]
10574 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10575 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10577 *) Don't blow it for numeric -newkey arguments to apps/req.
10578 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10580 *) Temp key "for export" tests were wrong in s3_srvr.c.
10581 [Anonymous <nobody@replay.com>]
10583 *) Add prototype for temp key callback functions
10584 SSL_CTX_set_tmp_{rsa,dh}_callback().
10587 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10588 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10591 *) X509_name_add_entry() freed the wrong thing after an error.
10592 [Arne Ansper <arne@ats.cyber.ee>]
10594 *) rsa_eay.c would attempt to free a NULL context.
10595 [Arne Ansper <arne@ats.cyber.ee>]
10597 *) BIO_s_socket() had a broken should_retry() on Windoze.
10598 [Arne Ansper <arne@ats.cyber.ee>]
10600 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10601 [Arne Ansper <arne@ats.cyber.ee>]
10603 *) Make sure the already existing X509_STORE->depth variable is initialized
10604 in X509_STORE_new(), but document the fact that this variable is still
10605 unused in the certificate verification process.
10606 [Ralf S. Engelschall]
10608 *) Fix the various library and apps files to free up pkeys obtained from
10609 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10612 *) Fix reference counting in X509_PUBKEY_get(). This makes
10613 demos/maurice/example2.c work, amongst others, probably.
10614 [Steve Henson and Ben Laurie]
10616 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10617 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10618 are no longer created. This way we have a single and consistent command
10619 line interface `openssl <command>', similar to `cvs <command>'.
10620 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10622 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10623 BIT STRING wrapper always have zero unused bits.
10626 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10629 *) Make the top-level INSTALL documentation easier to understand.
10632 *) Makefiles updated to exit if an error occurs in a sub-directory
10633 make (including if user presses ^C) [Paul Sutton]
10635 *) Make Montgomery context stuff explicit in RSA data structure.
10638 *) Fix build order of pem and err to allow for generated pem.h.
10641 *) Fix renumbering bug in X509_NAME_delete_entry().
10644 *) Enhanced the err-ins.pl script so it makes the error library number
10645 global and can add a library name. This is needed for external ASN1 and
10646 other error libraries.
10649 *) Fixed sk_insert which never worked properly.
10652 *) Fix ASN1 macros so they can handle indefinite length construted
10653 EXPLICIT tags. Some non standard certificates use these: they can now
10657 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10658 into a single doc/ssleay.txt bundle. This way the information is still
10659 preserved but no longer messes up this directory. Now it's new room for
10660 the new set of documenation files.
10661 [Ralf S. Engelschall]
10663 *) SETs were incorrectly DER encoded. This was a major pain, because they
10664 shared code with SEQUENCEs, which aren't coded the same. This means that
10665 almost everything to do with SETs or SEQUENCEs has either changed name or
10666 number of arguments.
10667 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10669 *) Fix test data to work with the above.
10672 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10673 was already fixed by Eric for 0.9.1 it seems.
10674 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10676 *) Autodetect FreeBSD3.
10679 *) Fix various bugs in Configure. This affects the following platforms:
10683 unixware-2.0-pentium
10687 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10688 before they are needed.
10691 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10695 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10697 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10698 changed SSLeay to OpenSSL in version strings.
10699 [Ralf S. Engelschall]
10701 *) Some fixups to the top-level documents.
10704 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10705 because the symlink to include/ was missing.
10706 [Ralf S. Engelschall]
10708 *) Incorporated the popular no-RSA/DSA-only patches
10709 which allow to compile a RSA-free SSLeay.
10710 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10712 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10713 when "ssleay" is still not found.
10714 [Ralf S. Engelschall]
10716 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10717 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10719 *) Updated the README file.
10720 [Ralf S. Engelschall]
10722 *) Added various .cvsignore files in the CVS repository subdirs
10723 to make a "cvs update" really silent.
10724 [Ralf S. Engelschall]
10726 *) Recompiled the error-definition header files and added
10727 missing symbols to the Win32 linker tables.
10728 [Ralf S. Engelschall]
10730 *) Cleaned up the top-level documents;
10731 o new files: CHANGES and LICENSE
10732 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10733 o merged COPYRIGHT into LICENSE
10734 o removed obsolete TODO file
10735 o renamed MICROSOFT to INSTALL.W32
10736 [Ralf S. Engelschall]
10738 *) Removed dummy files from the 0.9.1b source tree:
10739 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10740 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10741 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10742 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10743 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10744 [Ralf S. Engelschall]
10746 *) Added various platform portability fixes.
10749 *) The Genesis of the OpenSSL rpject:
10750 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10751 Young and Tim J. Hudson created while they were working for C2Net until
10753 [The OpenSSL Project]
10756 Changes between 0.9.0b and 0.9.1b [not released]
10758 *) Updated a few CA certificates under certs/
10761 *) Changed some BIGNUM api stuff.
10764 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10765 DGUX x86, Linux Alpha, etc.
10768 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10769 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10773 *) Add -strparse option to asn1pars program which parses nested
10775 [Dr Stephen Henson <shenson@bigfoot.com>]
10777 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10780 *) DSA fix for "ca" program.
10783 *) Added "-genkey" option to "dsaparam" program.
10786 *) Added RIPE MD160 (rmd160) message digest.
10789 *) Added -a (all) option to "ssleay version" command.
10792 *) Added PLATFORM define which is the id given to Configure.
10795 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10798 *) Extended the ASN.1 parser routines.
10801 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10804 *) Added a BN_CTX to the BN library.
10807 *) Fixed the weak key values in DES library
10810 *) Changed API in EVP library for cipher aliases.
10813 *) Added support for RC2/64bit cipher.
10816 *) Converted the lhash library to the crypto/mem.c functions.
10819 *) Added more recognized ASN.1 object ids.
10822 *) Added more RSA padding checks for SSL/TLS.
10825 *) Added BIO proxy/filter functionality.
10828 *) Added extra_certs to SSL_CTX which can be used
10829 send extra CA certificates to the client in the CA cert chain sending
10830 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10833 *) Now Fortezza is denied in the authentication phase because
10834 this is key exchange mechanism is not supported by SSLeay at all.
10837 *) Additional PKCS1 checks.
10840 *) Support the string "TLSv1" for all TLS v1 ciphers.
10843 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10844 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10847 *) Fixed a few memory leaks.
10850 *) Fixed various code and comment typos.
10853 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10854 bytes sent in the client random.
10855 [Edward Bishop <ebishop@spyglass.com>]