5 Changes between 1.0.0q and 1.0.0r [xx XXX xxxx]
7 *) Removed the export ciphers from the DEFAULT ciphers
10 Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
12 *) Build fixes for the Windows and OpenVMS platforms
13 [Matt Caswell and Richard Levitte]
15 Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
17 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
18 message can cause a segmentation fault in OpenSSL due to a NULL pointer
19 dereference. This could lead to a Denial Of Service attack. Thanks to
20 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
24 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
25 dtls1_buffer_record function under certain conditions. In particular this
26 could occur if an attacker sent repeated DTLS records with the same
27 sequence number but for the next epoch. The memory leak could be exploited
28 by an attacker in a Denial of Service attack through memory exhaustion.
29 Thanks to Chris Mueller for reporting this issue.
33 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
34 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
35 method would be set to NULL which could later result in a NULL pointer
36 dereference. Thanks to Frank Schmirler for reporting this issue.
40 *) Abort handshake if server key exchange message is omitted for ephemeral
43 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
48 *) Remove non-export ephemeral RSA code on client and server. This code
49 violated the TLS standard by allowing the use of temporary RSA keys in
50 non-export ciphersuites and could be used by a server to effectively
51 downgrade the RSA key length used to a value smaller than the server
52 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
53 INRIA or reporting this issue.
57 *) Fixed issue where DH client certificates are accepted without verification.
58 An OpenSSL server will accept a DH certificate for client authentication
59 without the certificate verify message. This effectively allows a client to
60 authenticate without the use of a private key. This only affects servers
61 which trust a client certificate authority which issues certificates
62 containing DH keys: these are extremely rare and hardly ever encountered.
63 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
68 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
69 results on some platforms, including x86_64. This bug occurs at random
70 with a very low probability, and is not known to be exploitable in any
71 way, though its exact impact is difficult to determine. Thanks to Pieter
72 Wuille (Blockstream) who reported this issue and also suggested an initial
73 fix. Further analysis was conducted by the OpenSSL development team and
74 Adam Langley of Google. The final fix was developed by Andy Polyakov of
75 the OpenSSL core team.
79 *) Fix various certificate fingerprint issues.
81 By using non-DER or invalid encodings outside the signed portion of a
82 certificate the fingerprint can be changed without breaking the signature.
83 Although no details of the signed portion of the certificate can be changed
84 this can cause problems with some applications: e.g. those using the
85 certificate fingerprint for blacklists.
87 1. Reject signatures with non zero unused bits.
89 If the BIT STRING containing the signature has non zero unused bits reject
90 the signature. All current signature algorithms require zero unused bits.
92 2. Check certificate algorithm consistency.
94 Check the AlgorithmIdentifier inside TBS matches the one in the
95 certificate signature. NB: this will result in signature failure
96 errors for some broken certificates.
98 Thanks to Konrad Kraszewski from Google for reporting this issue.
100 3. Check DSA/ECDSA signatures use DER.
102 Reencode DSA/ECDSA signatures and compare with the original received
103 signature. Return an error if there is a mismatch.
105 This will reject various cases including garbage after signature
106 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
107 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
108 (negative or with leading zeroes).
110 Further analysis was conducted and fixes were developed by Stephen Henson
111 of the OpenSSL core team.
116 Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
118 *) Session Ticket Memory Leak.
120 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
121 integrity of that ticket is first verified. In the event of a session
122 ticket integrity check failing, OpenSSL will fail to free memory
123 causing a memory leak. By sending a large number of invalid session
124 tickets an attacker could exploit this issue in a Denial Of Service
129 *) Build option no-ssl3 is incomplete.
131 When OpenSSL is configured with "no-ssl3" as a build option, servers
132 could accept and complete a SSL 3.0 handshake, and clients could be
133 configured to send them.
135 [Akamai and the OpenSSL team]
137 *) Add support for TLS_FALLBACK_SCSV.
138 Client applications doing fallback retries should call
139 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
141 [Adam Langley, Bodo Moeller]
143 *) Add additional DigestInfo checks.
145 Reencode DigestInto in DER and check against the original when
146 verifying RSA signature: this will reject any improperly encoded
147 DigestInfo structures.
149 Note: this is a precautionary measure and no attacks are currently known.
153 Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
155 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
156 to a denial of service attack. A malicious server can crash the client
157 with a null pointer dereference (read) by specifying an anonymous (EC)DH
158 ciphersuite and sending carefully crafted handshake messages.
160 Thanks to Felix Gröbert (Google) for discovering and researching this
165 *) By sending carefully crafted DTLS packets an attacker could cause openssl
166 to leak memory. This can be exploited through a Denial of Service attack.
167 Thanks to Adam Langley for discovering and researching this issue.
171 *) An attacker can force openssl to consume large amounts of memory whilst
172 processing DTLS handshake messages. This can be exploited through a
173 Denial of Service attack.
174 Thanks to Adam Langley for discovering and researching this issue.
178 *) An attacker can force an error condition which causes openssl to crash
179 whilst processing DTLS packets due to memory being freed twice. This
180 can be exploited through a Denial of Service attack.
181 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
186 *) If a multithreaded client connects to a malicious server using a resumed
187 session and the server sends an ec point format extension it could write
188 up to 255 bytes to freed memory.
190 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
195 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
196 X509_name_oneline, X509_name_print_ex et al. to leak some information
197 from the stack. Applications may be affected if they echo pretty printing
198 output to the attacker.
200 Thanks to Ivan Fratric (Google) for discovering this issue.
202 [Emilia Käsper, and Steve Henson]
204 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
205 for corner cases. (Certain input points at infinity could lead to
206 bogus results, with non-infinity inputs mapped to infinity too.)
209 Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
211 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
212 handshake can force the use of weak keying material in OpenSSL
213 SSL/TLS clients and servers.
215 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
216 researching this issue. (CVE-2014-0224)
217 [KIKUCHI Masashi, Steve Henson]
219 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
220 OpenSSL DTLS client the code can be made to recurse eventually crashing
223 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
225 [Imre Rad, Steve Henson]
227 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
228 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
229 client or server. This is potentially exploitable to run arbitrary
230 code on a vulnerable client or server.
232 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
233 [Jüri Aedla, Steve Henson]
235 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
236 are subject to a denial of service attack.
238 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
239 this issue. (CVE-2014-3470)
240 [Felix Gröbert, Ivan Fratric, Steve Henson]
242 *) Harmonize version and its documentation. -f flag is used to display
244 [mancha <mancha1@zoho.com>]
246 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
248 [mancha <mancha1@zoho.com>]
250 *) Fix some double frees. These are not thought to be exploitable.
251 [mancha <mancha1@zoho.com>]
253 *) Fix for the attack described in the paper "Recovering OpenSSL
254 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
255 by Yuval Yarom and Naomi Benger. Details can be obtained from:
256 http://eprint.iacr.org/2014/140
258 Thanks to Yuval Yarom and Naomi Benger for discovering this
259 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
260 [Yuval Yarom and Naomi Benger]
262 Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
264 *) Keep original DTLS digest and encryption contexts in retransmission
265 structures so we can use the previous session parameters if they need
266 to be resent. (CVE-2013-6450)
269 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
270 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
271 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
272 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
273 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
274 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
275 [Rob Stradling, Adam Langley]
277 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
279 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
281 This addresses the flaw in CBC record processing discovered by
282 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
283 at: http://www.isg.rhul.ac.uk/tls/
285 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
286 Security Group at Royal Holloway, University of London
287 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
288 Emilia Käsper for the initial patch.
290 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
292 *) Return an error when checking OCSP signatures when key is NULL.
293 This fixes a DoS attack. (CVE-2013-0166)
296 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
297 the right response is stapled. Also change SSL_get_certificate()
298 so it returns the certificate actually sent.
299 See http://rt.openssl.org/Ticket/Display.html?id=2836.
301 [Rob Stradling <rob.stradling@comodo.com>]
303 *) Fix possible deadlock when decoding public keys.
306 Changes between 1.0.0i and 1.0.0j [10 May 2012]
308 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
311 *) Sanity check record length before skipping explicit IV in DTLS
314 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
315 fuzzing as a service testing platform.
319 *) Initialise tkeylen properly when encrypting CMS messages.
320 Thanks to Solar Designer of Openwall for reporting this issue.
323 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
325 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
326 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
327 in CRYPTO_realloc_clean.
329 Thanks to Tavis Ormandy, Google Security Team, for discovering this
330 issue and to Adam Langley <agl@chromium.org> for fixing it.
332 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
334 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
336 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
337 in CMS and PKCS7 code. When RSA decryption fails use a random key for
338 content decryption and always return the same error. Note: this attack
339 needs on average 2^20 messages so it only affects automated senders. The
340 old behaviour can be reenabled in the CMS code by setting the
341 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
342 an MMA defence is not necessary.
343 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
344 this issue. (CVE-2012-0884)
347 *) Fix CVE-2011-4619: make sure we really are receiving a
348 client hello before rejecting multiple SGC restarts. Thanks to
349 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
352 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
354 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
355 Thanks to Antonio Martin, Enterprise Secure Access Research and
356 Development, Cisco Systems, Inc. for discovering this bug and
357 preparing a fix. (CVE-2012-0050)
360 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
362 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
363 of the Vaudenay padding oracle attack on CBC mode encryption
364 which enables an efficient plaintext recovery attack against
365 the OpenSSL implementation of DTLS. Their attack exploits timing
366 differences arising during decryption processing. A research
367 paper describing this attack can be found at:
368 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
369 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
370 Security Group at Royal Holloway, University of London
371 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
372 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
373 for preparing the fix. (CVE-2011-4108)
374 [Robin Seggelmann, Michael Tuexen]
376 *) Clear bytes used for block padding of SSL 3.0 records.
378 [Adam Langley (Google)]
380 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
381 Kadianakis <desnacked@gmail.com> for discovering this issue and
382 Adam Langley for preparing the fix. (CVE-2011-4619)
383 [Adam Langley (Google)]
385 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
386 [Andrey Kulikov <amdeich@gmail.com>]
388 *) Prevent malformed RFC3779 data triggering an assertion failure.
389 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
390 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
391 [Rob Austein <sra@hactrn.net>]
393 *) Improved PRNG seeding for VOS.
394 [Paul Green <Paul.Green@stratus.com>]
396 *) Fix ssl_ciph.c set-up race.
397 [Adam Langley (Google)]
399 *) Fix spurious failures in ecdsatest.c.
400 [Emilia Käsper (Google)]
402 *) Fix the BIO_f_buffer() implementation (which was mixing different
403 interpretations of the '..._len' fields).
404 [Adam Langley (Google)]
406 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
407 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
408 threads won't reuse the same blinding coefficients.
410 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
411 lock to call BN_BLINDING_invert_ex, and avoids one use of
412 BN_BLINDING_update for each BN_BLINDING structure (previously,
413 the last update always remained unused).
414 [Emilia Käsper (Google)]
416 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
417 [Bob Buckholz (Google)]
419 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
421 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
422 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
423 [Kaspar Brand <ossl@velox.ch>]
425 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
426 for multi-threaded use of ECDH. (CVE-2011-3210)
427 [Adam Langley (Google)]
429 *) Fix x509_name_ex_d2i memory leak on bad inputs.
432 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
433 signature public key algorithm by using OID xref utilities instead.
434 Before this you could only use some ECC ciphersuites with SHA1 only.
437 *) Add protection against ECDSA timing attacks as mentioned in the paper
438 by Billy Bob Brumley and Nicola Tuveri, see:
440 http://eprint.iacr.org/2011/232.pdf
442 [Billy Bob Brumley and Nicola Tuveri]
444 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
446 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
447 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
449 *) Fix bug in string printing code: if *any* escaping is enabled we must
450 escape the escape character (backslash) or the resulting string is
454 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
456 *) Disable code workaround for ancient and obsolete Netscape browsers
457 and servers: an attacker can use it in a ciphersuite downgrade attack.
458 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
461 *) Fixed J-PAKE implementation error, originally discovered by
462 Sebastien Martini, further info and confirmation from Stefan
463 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
466 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
468 *) Fix extension code to avoid race conditions which can result in a buffer
469 overrun vulnerability: resumed sessions must not be modified as they can
470 be shared by multiple threads. CVE-2010-3864
473 *) Fix WIN32 build system to correctly link an ENGINE directory into
477 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
479 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
481 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
483 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
485 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
486 context. The operation can be customised via the ctrl mechanism in
487 case ENGINEs want to include additional functionality.
490 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
493 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
494 output hashes compatible with older versions of OpenSSL.
495 [Willy Weisz <weisz@vcpc.univie.ac.at>]
497 *) Fix compression algorithm handling: if resuming a session use the
498 compression algorithm of the resumed session instead of determining
499 it from client hello again. Don't allow server to change algorithm.
502 *) Add load_crls() function to apps tidying load_certs() too. Add option
503 to verify utility to allow additional CRLs to be included.
506 *) Update OCSP request code to permit adding custom headers to the request:
507 some responders need this.
510 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
512 [Julia Lawall <julia@diku.dk>]
514 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
515 needlessly dereferenced structures, used obsolete functions and
516 didn't handle all updated verify codes correctly.
519 *) Disable MD2 in the default configuration.
522 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
523 indicate the initial BIO being pushed or popped. This makes it possible
524 to determine whether the BIO is the one explicitly called or as a result
525 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
526 it handles reference counts correctly and doesn't zero out the I/O bio
527 when it is not being explicitly popped. WARNING: applications which
528 included workarounds for the old buggy behaviour will need to be modified
529 or they could free up already freed BIOs.
532 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
533 renaming to all platforms (within the 0.9.8 branch, this was
534 done conditionally on Netware platforms to avoid a name clash).
535 [Guenter <lists@gknw.net>]
537 *) Add ECDHE and PSK support to DTLS.
538 [Michael Tuexen <tuexen@fh-muenster.de>]
540 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
544 *) Add "missing" function EVP_MD_flags() (without this the only way to
545 retrieve a digest flags is by accessing the structure directly. Update
546 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
547 or cipher is registered as in the "from" argument. Print out all
548 registered digests in the dgst usage message instead of manually
549 attempting to work them out.
552 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
553 this allows the use of compression and extensions. Change default cipher
554 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
555 by default unless an application cipher string requests it.
558 *) Alter match criteria in PKCS12_parse(). It used to try to use local
559 key ids to find matching certificates and keys but some PKCS#12 files
560 don't follow the (somewhat unwritten) rules and this strategy fails.
561 Now just gather all certificates together and the first private key
562 then look for the first certificate that matches the key.
565 *) Support use of registered digest and cipher names for dgst and cipher
566 commands instead of having to add each one as a special case. So now
573 openssl dgst -sha256 foo
575 and this works for ENGINE based algorithms too.
579 *) Update Gost ENGINE to support parameter files.
580 [Victor B. Wagner <vitus@cryptocom.ru>]
582 *) Support GeneralizedTime in ca utility.
583 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
585 *) Enhance the hash format used for certificate directory links. The new
586 form uses the canonical encoding (meaning equivalent names will work
587 even if they aren't identical) and uses SHA1 instead of MD5. This form
588 is incompatible with the older format and as a result c_rehash should
589 be used to rebuild symbolic links.
592 *) Make PKCS#8 the default write format for private keys, replacing the
593 traditional format. This form is standardised, more secure and doesn't
594 include an implicit MD5 dependency.
597 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
598 committed to OpenSSL should pass this lot as a minimum.
601 *) Add session ticket override functionality for use by EAP-FAST.
602 [Jouni Malinen <j@w1.fi>]
604 *) Modify HMAC functions to return a value. Since these can be implemented
605 in an ENGINE errors can occur.
608 *) Type-checked OBJ_bsearch_ex.
611 *) Type-checked OBJ_bsearch. Also some constification necessitated
612 by type-checking. Still to come: TXT_DB, bsearch(?),
613 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
617 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
618 seconds to a tm structure directly, instead of going through OS
619 specific date routines. This avoids any issues with OS routines such
620 as the year 2038 bug. New *_adj() functions for ASN1 time structures
621 and X509_time_adj_ex() to cover the extended range. The existing
622 X509_time_adj() is still usable and will no longer have any date issues.
625 *) Delta CRL support. New use deltas option which will attempt to locate
626 and search any appropriate delta CRLs available.
628 This work was sponsored by Google.
631 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
632 code and add additional score elements. Validate alternate CRL paths
633 as part of the CRL checking and indicate a new error "CRL path validation
634 error" in this case. Applications wanting additional details can use
635 the verify callback and check the new "parent" field. If this is not
636 NULL CRL path validation is taking place. Existing applications wont
637 see this because it requires extended CRL support which is off by
640 This work was sponsored by Google.
643 *) Support for freshest CRL extension.
645 This work was sponsored by Google.
648 *) Initial indirect CRL support. Currently only supported in the CRLs
649 passed directly and not via lookup. Process certificate issuer
650 CRL entry extension and lookup CRL entries by bother issuer name
651 and serial number. Check and process CRL issuer entry in IDP extension.
653 This work was sponsored by Google.
656 *) Add support for distinct certificate and CRL paths. The CRL issuer
657 certificate is validated separately in this case. Only enabled if
658 an extended CRL support flag is set: this flag will enable additional
659 CRL functionality in future.
661 This work was sponsored by Google.
664 *) Add support for policy mappings extension.
666 This work was sponsored by Google.
669 *) Fixes to pathlength constraint, self issued certificate handling,
670 policy processing to align with RFC3280 and PKITS tests.
672 This work was sponsored by Google.
675 *) Support for name constraints certificate extension. DN, email, DNS
676 and URI types are currently supported.
678 This work was sponsored by Google.
681 *) To cater for systems that provide a pointer-based thread ID rather
682 than numeric, deprecate the current numeric thread ID mechanism and
683 replace it with a structure and associated callback type. This
684 mechanism allows a numeric "hash" to be extracted from a thread ID in
685 either case, and on platforms where pointers are larger than 'long',
686 mixing is done to help ensure the numeric 'hash' is usable even if it
687 can't be guaranteed unique. The default mechanism is to use "&errno"
688 as a pointer-based thread ID to distinguish between threads.
690 Applications that want to provide their own thread IDs should now use
691 CRYPTO_THREADID_set_callback() to register a callback that will call
692 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
694 Note that ERR_remove_state() is now deprecated, because it is tied
695 to the assumption that thread IDs are numeric. ERR_remove_state(0)
696 to free the current thread's error state should be replaced by
697 ERR_remove_thread_state(NULL).
699 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
700 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
701 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
702 application was previously providing a numeric thread callback that
703 was inappropriate for distinguishing threads, then uniqueness might
704 have been obtained with &errno that happened immediately in the
705 intermediate development versions of OpenSSL; this is no longer the
706 case, the numeric thread callback will now override the automatic use
708 [Geoff Thorpe, with help from Bodo Moeller]
710 *) Initial support for different CRL issuing certificates. This covers a
711 simple case where the self issued certificates in the chain exist and
712 the real CRL issuer is higher in the existing chain.
714 This work was sponsored by Google.
717 *) Removed effectively defunct crypto/store from the build.
720 *) Revamp of STACK to provide stronger type-checking. Still to come:
721 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
722 ASN1_STRING, CONF_VALUE.
725 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
726 RAM on SSL connections. This option can save about 34k per idle SSL.
729 *) Revamp of LHASH to provide stronger type-checking. Still to come:
730 STACK, TXT_DB, bsearch, qsort.
733 *) Initial support for Cryptographic Message Syntax (aka CMS) based
734 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
735 support for data, signedData, compressedData, digestedData and
736 encryptedData, envelopedData types included. Scripts to check against
737 RFC4134 examples draft and interop and consistency checks of many
738 content types and variants.
741 *) Add options to enc utility to support use of zlib compression BIO.
744 *) Extend mk1mf to support importing of options and assembly language
745 files from Configure script, currently only included in VC-WIN32.
746 The assembly language rules can now optionally generate the source
747 files from the associated perl scripts.
750 *) Implement remaining functionality needed to support GOST ciphersuites.
751 Interop testing has been performed using CryptoPro implementations.
752 [Victor B. Wagner <vitus@cryptocom.ru>]
754 *) s390x assembler pack.
757 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
761 *) Implement Opaque PRF Input TLS extension as specified in
762 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
763 official specification yet and no extension type assignment by
764 IANA exists, this extension (for now) will have to be explicitly
765 enabled when building OpenSSL by providing the extension number
766 to use. For example, specify an option
768 -DTLSEXT_TYPE_opaque_prf_input=0x9527
770 to the "config" or "Configure" script to enable the extension,
771 assuming extension number 0x9527 (which is a completely arbitrary
772 and unofficial assignment based on the MD5 hash of the Internet
773 Draft). Note that by doing so, you potentially lose
774 interoperability with other TLS implementations since these might
775 be using the same extension number for other purposes.
777 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
778 opaque PRF input value to use in the handshake. This will create
779 an interal copy of the length-'len' string at 'src', and will
780 return non-zero for success.
782 To get more control and flexibility, provide a callback function
785 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
786 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
790 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
793 Callback function 'cb' will be called in handshakes, and is
794 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
795 Argument 'arg' is for application purposes (the value as given to
796 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
797 be provided to the callback function). The callback function
798 has to return non-zero to report success: usually 1 to use opaque
799 PRF input just if possible, or 2 to enforce use of the opaque PRF
800 input. In the latter case, the library will abort the handshake
801 if opaque PRF input is not successfully negotiated.
803 Arguments 'peerinput' and 'len' given to the callback function
804 will always be NULL and 0 in the case of a client. A server will
805 see the client's opaque PRF input through these variables if
806 available (NULL and 0 otherwise). Note that if the server
807 provides an opaque PRF input, the length must be the same as the
808 length of the client's opaque PRF input.
810 Note that the callback function will only be called when creating
811 a new session (session resumption can resume whatever was
812 previously negotiated), and will not be called in SSL 2.0
813 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
814 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
815 for applications that need to enforce opaque PRF input.
819 *) Update ssl code to support digests other than SHA1+MD5 for handshake
822 [Victor B. Wagner <vitus@cryptocom.ru>]
824 *) Add RFC4507 support to OpenSSL. This includes the corrections in
825 RFC4507bis. The encrypted ticket format is an encrypted encoded
826 SSL_SESSION structure, that way new session features are automatically
829 If a client application caches session in an SSL_SESSION structure
830 support is transparent because tickets are now stored in the encoded
833 The SSL_CTX structure automatically generates keys for ticket
834 protection in servers so again support should be possible
835 with no application modification.
837 If a client or server wishes to disable RFC4507 support then the option
838 SSL_OP_NO_TICKET can be set.
840 Add a TLS extension debugging callback to allow the contents of any client
841 or server extensions to be examined.
843 This work was sponsored by Google.
846 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
847 OpenSSL should now compile cleanly on gcc 4.2
848 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
850 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
851 support including streaming MAC support: this is required for GOST
853 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
855 *) Add option -stream to use PKCS#7 streaming in smime utility. New
856 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
857 to output in BER and PEM format.
860 *) Experimental support for use of HMAC via EVP_PKEY interface. This
861 allows HMAC to be handled via the EVP_DigestSign*() interface. The
862 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
863 ENGINE support for HMAC keys which are unextractable. New -mac and
864 -macopt options to dgst utility.
867 *) New option -sigopt to dgst utility. Update dgst to use
868 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
869 alternative signing paramaters such as X9.31 or PSS in the dgst
873 *) Change ssl_cipher_apply_rule(), the internal function that does
874 the work each time a ciphersuite string requests enabling
875 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
876 removing ("!foo+bar") a class of ciphersuites: Now it maintains
877 the order of disabled ciphersuites such that those ciphersuites
878 that most recently went from enabled to disabled not only stay
879 in order with respect to each other, but also have higher priority
880 than other disabled ciphersuites the next time ciphersuites are
883 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
884 the same ciphersuites as with "HIGH" alone, but in a specific
885 order where the PSK ciphersuites come first (since they are the
886 most recently disabled ciphersuites when "HIGH" is parsed).
888 Also, change ssl_create_cipher_list() (using this new
889 funcionality) such that between otherwise identical
890 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
894 *) Change ssl_create_cipher_list() so that it automatically
895 arranges the ciphersuites in reasonable order before starting
896 to process the rule string. Thus, the definition for "DEFAULT"
897 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
898 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
899 This makes it much easier to arrive at a reasonable default order
900 in applications for which anonymous ciphers are OK (meaning
901 that you can't actually use DEFAULT).
902 [Bodo Moeller; suggested by Victor Duchovni]
904 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
905 processing) into multiple integers instead of setting
906 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
907 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
908 (These masks as well as the individual bit definitions are hidden
909 away into the non-exported interface ssl/ssl_locl.h, so this
910 change to the definition of the SSL_CIPHER structure shouldn't
911 affect applications.) This give us more bits for each of these
912 categories, so there is no longer a need to coagulate AES128 and
913 AES256 into a single algorithm bit, and to coagulate Camellia128
914 and Camellia256 into a single algorithm bit, which has led to all
917 Thus, among other things, the kludge introduced in 0.9.7m and
918 0.9.8e for masking out AES256 independently of AES128 or masking
919 out Camellia256 independently of AES256 is not needed here in 0.9.9.
921 With the change, we also introduce new ciphersuite aliases that
922 so far were missing: "AES128", "AES256", "CAMELLIA128", and
926 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
927 Use the leftmost N bytes of the signature input if the input is
928 larger than the prime q (with N being the size in bytes of q).
931 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
932 it yet and it is largely untested.
935 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
938 *) Initial incomplete changes to avoid need for function casts in OpenSSL
939 some compilers (gcc 4.2 and later) reject their use. Safestack is
940 reimplemented. Update ASN1 to avoid use of legacy functions.
943 *) Win32/64 targets are linked with Winsock2.
946 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
947 to external functions. This can be used to increase CRL handling
948 efficiency especially when CRLs are very large by (for example) storing
949 the CRL revoked certificates in a database.
952 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
953 new CRLs added to a directory can be used. New command line option
954 -verify_return_error to s_client and s_server. This causes real errors
955 to be returned by the verify callback instead of carrying on no matter
956 what. This reflects the way a "real world" verify callback would behave.
959 *) GOST engine, supporting several GOST algorithms and public key formats.
960 Kindly donated by Cryptocom.
963 *) Partial support for Issuing Distribution Point CRL extension. CRLs
964 partitioned by DP are handled but no indirect CRL or reason partitioning
965 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
966 selected via a scoring technique which handles IDP and AKID in CRLs.
969 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
970 will ultimately be used for all verify operations: this will remove the
971 X509_STORE dependency on certificate verification and allow alternative
972 lookup methods. X509_STORE based implementations of these two callbacks.
975 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
976 Modify get_crl() to find a valid (unexpired) CRL if possible.
979 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
980 this would be called X509_CRL_cmp() but that name is already used by
981 a function that just compares CRL issuer names. Cache several CRL
982 extensions in X509_CRL structure and cache CRLDP in X509.
985 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
986 this maps equivalent X509_NAME structures into a consistent structure.
987 Name comparison can then be performed rapidly using memcmp().
990 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
994 *) Allow digests to supply their own micalg string for S/MIME type using
995 the ctrl EVP_MD_CTRL_MICALG.
998 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
999 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1000 ctrl. It can then customise the structure before and/or after signing
1004 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1005 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1006 to free up any added signature OIDs.
1009 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1010 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1011 digest and cipher tables. New options added to openssl utility:
1012 list-message-digest-algorithms and list-cipher-algorithms.
1015 *) Change the array representation of binary polynomials: the list
1016 of degrees of non-zero coefficients is now terminated with -1.
1017 Previously it was terminated with 0, which was also part of the
1018 value; thus, the array representation was not applicable to
1019 polynomials where t^0 has coefficient zero. This change makes
1020 the array representation useful in a more general context.
1023 *) Various modifications and fixes to SSL/TLS cipher string
1024 handling. For ECC, the code now distinguishes between fixed ECDH
1025 with RSA certificates on the one hand and with ECDSA certificates
1026 on the other hand, since these are separate ciphersuites. The
1027 unused code for Fortezza ciphersuites has been removed.
1029 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1030 (not "ECDHE"). For consistency with the code for DH
1031 certificates, use of ECDH certificates is now considered ECDH
1032 authentication, not RSA or ECDSA authentication (the latter is
1033 merely the CA's signing algorithm and not actively used in the
1036 The temporary ciphersuite alias "ECCdraft" is no longer
1037 available, and ECC ciphersuites are no longer excluded from "ALL"
1038 and "DEFAULT". The following aliases now exist for RFC 4492
1039 ciphersuites, most of these by analogy with the DH case:
1041 kECDHr - ECDH cert, signed with RSA
1042 kECDHe - ECDH cert, signed with ECDSA
1043 kECDH - ECDH cert (signed with either RSA or ECDSA)
1044 kEECDH - ephemeral ECDH
1045 ECDH - ECDH cert or ephemeral ECDH
1051 AECDH - anonymous ECDH
1052 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1056 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1057 Use correct micalg parameters depending on digest(s) in signed message.
1060 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1061 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1064 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1065 an engine to register a method. Add ENGINE lookups for methods and
1066 functional reference processing.
1069 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1070 EVP_{Sign,Verify}* which allow an application to customise the signature
1074 *) New -resign option to smime utility. This adds one or more signers
1075 to an existing PKCS#7 signedData structure. Also -md option to use an
1076 alternative message digest algorithm for signing.
1079 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1080 create PKCS7 structures containing multiple signers. Update smime
1081 application to support multiple signers.
1084 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1088 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1089 Reorganize PBE internals to lookup from a static table using NIDs,
1090 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1091 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1092 PRF which will be automatically used with PBES2.
1095 *) Replace the algorithm specific calls to generate keys in "req" with the
1099 *) Update PKCS#7 enveloped data routines to use new API. This is now
1100 supported by any public key method supporting the encrypt operation. A
1101 ctrl is added to allow the public key algorithm to examine or modify
1102 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1106 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1107 a default digest type to use. In most cases this will be SHA1 but some
1108 algorithms (such as GOST) need to specify an alternative digest. The
1109 return value indicates how strong the prefernce is 1 means optional and
1110 2 is mandatory (that is it is the only supported type). Modify
1111 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1112 use the default md. Update openssl utilities to use the default digest
1113 type for signing if it is not explicitly indicated.
1116 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1117 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1118 signing method from the key type. This effectively removes the link
1119 between digests and public key types.
1122 *) Add an OID cross reference table and utility functions. Its purpose is to
1123 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1124 rsaEncryption. This will allow some of the algorithm specific hackery
1125 needed to use the correct OID to be removed.
1128 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1129 structures for PKCS7_sign(). They are now set up by the relevant public
1133 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1136 *) Add support for key derivation (agreement) in the API, DH method and
1140 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1141 public and private key formats. As a side effect these add additional
1142 command line functionality not previously available: DSA signatures can be
1143 generated and verified using pkeyutl and DH key support and generation in
1148 [Oliver Tappe <zooey@hirschkaefer.de>]
1150 *) New make target "install_html_docs" installs HTML renditions of the
1152 [Oliver Tappe <zooey@hirschkaefer.de>]
1154 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1155 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1156 support key and parameter generation and add initial key generation
1157 functionality for RSA.
1160 *) Add functions for main EVP_PKEY_method operations. The undocumented
1161 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1162 EVP_PKEY_{encrypt,decrypt}_old.
1165 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1166 key API, doesn't do much yet.
1169 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1170 public key algorithms. New option to openssl utility:
1171 "list-public-key-algorithms" to print out info.
1174 *) Implement the Supported Elliptic Curves Extension for
1175 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1178 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1179 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1182 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1183 utilities such as rsa, dsa, dsaparam etc except they process any key
1187 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1188 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1189 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1193 *) Initial support for pluggable public key ASN1.
1194 De-spaghettify the public key ASN1 handling. Move public and private
1195 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1196 algorithm specific handling to a single module within the relevant
1197 algorithm directory. Add functions to allow (near) opaque processing
1198 of public and private key structures.
1201 *) Implement the Supported Point Formats Extension for
1202 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1205 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1206 for the psk identity [hint] and the psk callback functions to the
1207 SSL_SESSION, SSL and SSL_CTX structure.
1210 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1214 SSL_CTX_use_psk_identity_hint
1215 SSL_get_psk_identity_hint
1216 SSL_get_psk_identity
1217 SSL_use_psk_identity_hint
1219 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1221 *) Add RFC 3161 compliant time stamp request creation, response generation
1222 and response verification functionality.
1223 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1225 *) Add initial support for TLS extensions, specifically for the server_name
1226 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1227 have new members for a host name. The SSL data structure has an
1228 additional member SSL_CTX *initial_ctx so that new sessions can be
1229 stored in that context to allow for session resumption, even after the
1230 SSL has been switched to a new SSL_CTX in reaction to a client's
1231 server_name extension.
1233 New functions (subject to change):
1235 SSL_get_servername()
1236 SSL_get_servername_type()
1239 New CTRL codes and macros (subject to change):
1241 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1242 - SSL_CTX_set_tlsext_servername_callback()
1243 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1244 - SSL_CTX_set_tlsext_servername_arg()
1245 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1247 openssl s_client has a new '-servername ...' option.
1249 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1250 '-key2 ...', '-servername_fatal' (subject to change). This allows
1251 testing the HostName extension for a specific single host name ('-cert'
1252 and '-key' remain fallbacks for handshakes without HostName
1253 negotiation). If the unrecogninzed_name alert has to be sent, this by
1254 default is a warning; it becomes fatal with the '-servername_fatal'
1257 [Peter Sylvester, Remy Allais, Christophe Renou]
1259 *) Whirlpool hash implementation is added.
1262 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1263 bn(64,32). Because of instruction set limitations it doesn't have
1264 any negative impact on performance. This was done mostly in order
1265 to make it possible to share assembler modules, such as bn_mul_mont
1266 implementations, between 32- and 64-bit builds without hassle.
1269 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1270 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1274 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1275 dedicated Montgomery multiplication procedure, is introduced.
1276 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1277 "64-bit" performance on certain 32-bit targets.
1280 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1281 in SSL structures. New SSL ctrl to set maximum send fragment size.
1282 Save memory by seeting the I/O buffer sizes dynamically instead of
1283 using the maximum available value.
1286 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1287 in addition to the text details.
1290 *) Very, very preliminary EXPERIMENTAL support for printing of general
1291 ASN1 structures. This currently produces rather ugly output and doesn't
1292 handle several customised structures at all.
1295 *) Integrated support for PVK file format and some related formats such
1296 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1297 these in the 'rsa' and 'dsa' utilities.
1300 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1303 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1304 place for the (very old) "NETSCAPE" format certificates which are now
1305 handled using new ASN1 code equivalents.
1308 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1309 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1310 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1313 *) Modify CRL distribution points extension code to print out previously
1314 unsupported fields. Enhance extension setting code to allow setting of
1318 *) Add print and set support for Issuing Distribution Point CRL extension.
1321 *) Change 'Configure' script to enable Camellia by default.
1324 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
1326 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1328 This addresses the flaw in CBC record processing discovered by
1329 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1330 at: http://www.isg.rhul.ac.uk/tls/
1332 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1333 Security Group at Royal Holloway, University of London
1334 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1335 Emilia Käsper for the initial patch.
1337 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1339 *) Return an error when checking OCSP signatures when key is NULL.
1340 This fixes a DoS attack. (CVE-2013-0166)
1343 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1344 the right response is stapled. Also change SSL_get_certificate()
1345 so it returns the certificate actually sent.
1346 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1347 (This is a backport)
1348 [Rob Stradling <rob.stradling@comodo.com>]
1350 *) Fix possible deadlock when decoding public keys.
1353 Changes between 0.9.8w and 0.9.8x [10 May 2012]
1355 *) Sanity check record length before skipping explicit IV in DTLS
1358 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1359 fuzzing as a service testing platform.
1363 *) Initialise tkeylen properly when encrypting CMS messages.
1364 Thanks to Solar Designer of Openwall for reporting this issue.
1367 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
1369 *) The fix for CVE-2012-2110 did not take into account that the
1370 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
1371 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
1372 rejecting negative len parameter. (CVE-2012-2131)
1373 [Tomas Hoger <thoger@redhat.com>]
1375 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
1377 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1378 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1379 in CRYPTO_realloc_clean.
1381 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1382 issue and to Adam Langley <agl@chromium.org> for fixing it.
1384 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1386 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
1388 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1389 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1390 content decryption and always return the same error. Note: this attack
1391 needs on average 2^20 messages so it only affects automated senders. The
1392 old behaviour can be reenabled in the CMS code by setting the
1393 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1394 an MMA defence is not necessary.
1395 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1396 this issue. (CVE-2012-0884)
1399 *) Fix CVE-2011-4619: make sure we really are receiving a
1400 client hello before rejecting multiple SGC restarts. Thanks to
1401 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1404 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1406 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1407 Thanks to Antonio Martin, Enterprise Secure Access Research and
1408 Development, Cisco Systems, Inc. for discovering this bug and
1409 preparing a fix. (CVE-2012-0050)
1412 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1414 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1415 of the Vaudenay padding oracle attack on CBC mode encryption
1416 which enables an efficient plaintext recovery attack against
1417 the OpenSSL implementation of DTLS. Their attack exploits timing
1418 differences arising during decryption processing. A research
1419 paper describing this attack can be found at:
1420 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1421 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1422 Security Group at Royal Holloway, University of London
1423 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1424 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1425 for preparing the fix. (CVE-2011-4108)
1426 [Robin Seggelmann, Michael Tuexen]
1428 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1429 [Ben Laurie, Kasper <ekasper@google.com>]
1431 *) Clear bytes used for block padding of SSL 3.0 records.
1433 [Adam Langley (Google)]
1435 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1436 Kadianakis <desnacked@gmail.com> for discovering this issue and
1437 Adam Langley for preparing the fix. (CVE-2011-4619)
1438 [Adam Langley (Google)]
1440 *) Prevent malformed RFC3779 data triggering an assertion failure.
1441 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1442 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1443 [Rob Austein <sra@hactrn.net>]
1445 *) Fix ssl_ciph.c set-up race.
1446 [Adam Langley (Google)]
1448 *) Fix spurious failures in ecdsatest.c.
1449 [Emilia Käsper (Google)]
1451 *) Fix the BIO_f_buffer() implementation (which was mixing different
1452 interpretations of the '..._len' fields).
1453 [Adam Langley (Google)]
1455 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1456 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1457 threads won't reuse the same blinding coefficients.
1459 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1460 lock to call BN_BLINDING_invert_ex, and avoids one use of
1461 BN_BLINDING_update for each BN_BLINDING structure (previously,
1462 the last update always remained unused).
1463 [Emilia Käsper (Google)]
1465 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1466 for multi-threaded use of ECDH.
1467 [Adam Langley (Google)]
1469 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1472 *) Add protection against ECDSA timing attacks as mentioned in the paper
1473 by Billy Bob Brumley and Nicola Tuveri, see:
1475 http://eprint.iacr.org/2011/232.pdf
1477 [Billy Bob Brumley and Nicola Tuveri]
1479 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1481 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1482 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1484 *) Fix bug in string printing code: if *any* escaping is enabled we must
1485 escape the escape character (backslash) or the resulting string is
1489 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1491 *) Disable code workaround for ancient and obsolete Netscape browsers
1492 and servers: an attacker can use it in a ciphersuite downgrade attack.
1493 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1496 *) Fixed J-PAKE implementation error, originally discovered by
1497 Sebastien Martini, further info and confirmation from Stefan
1498 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1501 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1503 *) Fix extension code to avoid race conditions which can result in a buffer
1504 overrun vulnerability: resumed sessions must not be modified as they can
1505 be shared by multiple threads. CVE-2010-3864
1508 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1511 *) Don't reencode certificate when calculating signature: cache and use
1512 the original encoding instead. This makes signature verification of
1513 some broken encodings work correctly.
1516 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1517 is also one of the inputs.
1518 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1520 *) Don't repeatedly append PBE algorithms to table if they already exist.
1521 Sort table on each new add. This effectively makes the table read only
1522 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1526 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1528 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1531 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1532 access or freeing data twice (CVE-2010-0742)
1533 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1535 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1536 common in certificates and some applications which only call
1537 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1541 Reduce copying into .apps and .test in makevms.com
1542 Don't try to use blank CA certificate in CA.com
1543 Allow use of C files from original directories in maketests.com
1544 [Steven M. Schweda" <sms@antinode.info>]
1546 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1548 *) When rejecting SSL/TLS records due to an incorrect version number, never
1549 update s->server with a new major version number. As of
1550 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1551 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1552 the previous behavior could result in a read attempt at NULL when
1553 receiving specific incorrect SSL/TLS records once record payload
1554 protection is active. (CVE-2010-0740)
1555 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1557 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1558 could be crashed if the relevant tables were not present (e.g. chrooted).
1559 [Tomas Hoger <thoger@redhat.com>]
1561 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1563 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1564 [Martin Olsson, Neel Mehta]
1566 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1567 accommodate for stack sorting, always a write lock!).
1570 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1571 excessive delays in the RAND_poll(): over a minute. As a workaround
1572 include a time check in the inner Heap32Next loop too.
1575 *) The code that handled flushing of data in SSL/TLS originally used the
1576 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1577 the problem outlined in PR#1949. The fix suggested there however can
1578 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1579 of Apache). So instead simplify the code to flush unconditionally.
1580 This should be fine since flushing with no data to flush is a no op.
1583 *) Handle TLS versions 2.0 and later properly and correctly use the
1584 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1585 off ancient servers have a habit of sticking around for a while...
1588 *) Modify compression code so it frees up structures without using the
1589 ex_data callbacks. This works around a problem where some applications
1590 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1591 restarting) then use compression (e.g. SSL with compression) later.
1592 This results in significant per-connection memory leaks and
1593 has caused some security issues including CVE-2008-1678 and
1597 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1598 change when encrypting or decrypting.
1601 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1602 connect and renegotiate with servers which do not support RI.
1603 Until RI is more widely deployed this option is enabled by default.
1606 *) Add "missing" ssl ctrls to clear options and mode.
1609 *) If client attempts to renegotiate and doesn't support RI respond with
1610 a no_renegotiation alert as required by RFC5746. Some renegotiating
1611 TLS clients will continue a connection gracefully when they receive
1612 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1613 waiting for a server hello which it will never receive. Now we treat a
1614 received no_renegotiation alert as a fatal error. This is because
1615 applications requesting a renegotiation might well expect it to succeed
1616 and would have no code in place to handle the server denying it so the
1617 only safe thing to do is to terminate the connection.
1620 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1621 peer supports secure renegotiation and 0 otherwise. Print out peer
1622 renegotiation support in s_client/s_server.
1625 *) Replace the highly broken and deprecated SPKAC certification method with
1626 the updated NID creation version. This should correctly handle UTF8.
1629 *) Implement RFC5746. Re-enable renegotiation but require the extension
1630 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1631 turns out to be a bad idea. It has been replaced by
1632 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1633 SSL_CTX_set_options(). This is really not recommended unless you
1634 know what you are doing.
1635 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1637 *) Fixes to stateless session resumption handling. Use initial_ctx when
1638 issuing and attempting to decrypt tickets in case it has changed during
1639 servername handling. Use a non-zero length session ID when attempting
1640 stateless session resumption: this makes it possible to determine if
1641 a resumption has occurred immediately after receiving server hello
1642 (several places in OpenSSL subtly assume this) instead of later in
1646 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1647 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1648 fixes for a few places where the return code is not checked
1650 [Julia Lawall <julia@diku.dk>]
1652 *) Add --strict-warnings option to Configure script to include devteam
1653 warnings in other configurations.
1656 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1657 makes it possible to install openssl libraries in locations which
1658 have names other than "lib", for example "/usr/lib64" which some
1660 [Steve Henson, based on patch from Jeremy Utley]
1662 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1663 X690 8.9.12 and can produce some misleading textual output of OIDs.
1664 [Steve Henson, reported by Dan Kaminsky]
1666 *) Delete MD2 from algorithm tables. This follows the recommendation in
1667 several standards that it is not used in new applications due to
1668 several cryptographic weaknesses. For binary compatibility reasons
1669 the MD2 API is still compiled in by default.
1672 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1676 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1677 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1679 [Guenter <lists@gknw.net>]
1681 *) Fix the server certificate chain building code to use X509_verify_cert(),
1682 it used to have an ad-hoc builder which was unable to cope with anything
1683 other than a simple chain.
1684 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1686 *) Don't check self signed certificate signatures in X509_verify_cert()
1687 by default (a flag can override this): it just wastes time without
1688 adding any security. As a useful side effect self signed root CAs
1689 with non-FIPS digests are now usable in FIPS mode.
1692 *) In dtls1_process_out_of_seq_message() the check if the current message
1693 is already buffered was missing. For every new message was memory
1694 allocated, allowing an attacker to perform an denial of service attack
1695 with sending out of seq handshake messages until there is no memory
1696 left. Additionally every future messege was buffered, even if the
1697 sequence number made no sense and would be part of another handshake.
1698 So only messages with sequence numbers less than 10 in advance will be
1699 buffered. (CVE-2009-1378)
1700 [Robin Seggelmann, discovered by Daniel Mentz]
1702 *) Records are buffered if they arrive with a future epoch to be
1703 processed after finishing the corresponding handshake. There is
1704 currently no limitation to this buffer allowing an attacker to perform
1705 a DOS attack with sending records with future epochs until there is no
1706 memory left. This patch adds the pqueue_size() function to detemine
1707 the size of a buffer and limits the record buffer to 100 entries.
1709 [Robin Seggelmann, discovered by Daniel Mentz]
1711 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1712 parent structure is freed. (CVE-2009-1379)
1715 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1716 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1719 [Ilya O. <vrghost@gmail.com>]
1721 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1723 *) Disable renegotiation completely - this fixes a severe security
1724 problem (CVE-2009-3555) at the cost of breaking all
1725 renegotiation. Renegotiation can be re-enabled by setting
1726 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1727 run-time. This is really not recommended unless you know what
1731 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1733 *) Don't set val to NULL when freeing up structures, it is freed up by
1734 underlying code. If sizeof(void *) > sizeof(long) this can result in
1735 zeroing past the valid field. (CVE-2009-0789)
1736 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1738 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1739 checked correctly. This would allow some invalid signed attributes to
1740 appear to verify correctly. (CVE-2009-0591)
1741 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1743 *) Reject UniversalString and BMPString types with invalid lengths. This
1744 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1745 a legal length. (CVE-2009-0590)
1748 *) Set S/MIME signing as the default purpose rather than setting it
1749 unconditionally. This allows applications to override it at the store
1753 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1754 to handle some structures.
1757 *) Improve efficiency of mem_gets: don't search whole buffer each time
1759 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1761 *) New -hex option for openssl rand.
1764 *) Print out UTF8String and NumericString when parsing ASN1.
1767 *) Support NumericString type for name components.
1770 *) Allow CC in the environment to override the automatically chosen
1771 compiler. Note that nothing is done to ensure flags work with the
1775 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1777 *) Properly check EVP_VerifyFinal() and similar return values
1779 [Ben Laurie, Bodo Moeller, Google Security Team]
1781 *) Enable TLS extensions by default.
1784 *) Allow the CHIL engine to be loaded, whether the application is
1785 multithreaded or not. (This does not release the developer from the
1786 obligation to set up the dynamic locking callbacks.)
1787 [Sander Temme <sander@temme.net>]
1789 *) Use correct exit code if there is an error in dgst command.
1790 [Steve Henson; problem pointed out by Roland Dirlewanger]
1792 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1793 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1796 *) Add experimental JPAKE support, including demo authentication in
1797 s_client and s_server.
1800 *) Set the comparison function in v3_addr_canonize().
1801 [Rob Austein <sra@hactrn.net>]
1803 *) Add support for XMPP STARTTLS in s_client.
1804 [Philip Paeps <philip@freebsd.org>]
1806 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1807 to ensure that even with this option, only ciphersuites in the
1808 server's preference list will be accepted. (Note that the option
1809 applies only when resuming a session, so the earlier behavior was
1810 just about the algorithm choice for symmetric cryptography.)
1813 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1815 *) Fix NULL pointer dereference if a DTLS server received
1816 ChangeCipherSpec as first record (CVE-2009-1386).
1819 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1820 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1823 *) The fix in 0.9.8c that supposedly got rid of unsafe
1824 double-checked locking was incomplete for RSA blinding,
1825 addressing just one layer of what turns out to have been
1826 doubly unsafe triple-checked locking.
1828 So now fix this for real by retiring the MONT_HELPER macro
1829 in crypto/rsa/rsa_eay.c.
1831 [Bodo Moeller; problem pointed out by Marius Schilder]
1833 *) Various precautionary measures:
1835 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1837 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1838 (NB: This would require knowledge of the secret session ticket key
1839 to exploit, in which case you'd be SOL either way.)
1841 - Change bn_nist.c so that it will properly handle input BIGNUMs
1842 outside the expected range.
1844 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1847 [Neel Mehta, Bodo Moeller]
1849 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1850 the load fails. Useful for distros.
1851 [Ben Laurie and the FreeBSD team]
1853 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1856 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1859 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1861 This work was sponsored by Logica.
1864 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1865 keystores. Support for SSL/TLS client authentication too.
1866 Not compiled unless enable-capieng specified to Configure.
1868 This work was sponsored by Logica.
1871 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1872 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1873 attribute creation routines such as certifcate requests and PKCS#12
1877 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1879 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1880 handshake which could lead to a cilent crash as found using the
1881 Codenomicon TLS test suite (CVE-2008-1672)
1882 [Steve Henson, Mark Cox]
1884 *) Fix double free in TLS server name extensions which could lead to
1885 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1888 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1890 Clear the error queue to ensure that error entries left from
1891 older function calls do not interfere with the correct operation.
1892 [Lutz Jaenicke, Erik de Castro Lopo]
1894 *) Remove root CA certificates of commercial CAs:
1896 The OpenSSL project does not recommend any specific CA and does not
1897 have any policy with respect to including or excluding any CA.
1898 Therefore it does not make any sense to ship an arbitrary selection
1899 of root CA certificates with the OpenSSL software.
1902 *) RSA OAEP patches to fix two separate invalid memory reads.
1903 The first one involves inputs when 'lzero' is greater than
1904 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1905 before the beginning of from). The second one involves inputs where
1906 the 'db' section contains nothing but zeroes (there is a one-byte
1907 invalid read after the end of 'db').
1908 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1910 *) Partial backport from 0.9.9-dev:
1912 Introduce bn_mul_mont (dedicated Montgomery multiplication
1913 procedure) as a candidate for BIGNUM assembler implementation.
1914 While 0.9.9-dev uses assembler for various architectures, only
1915 x86_64 is available by default here in the 0.9.8 branch, and
1916 32-bit x86 is available through a compile-time setting.
1918 To try the 32-bit x86 assembler implementation, use Configure
1919 option "enable-montasm" (which exists only for this backport).
1921 As "enable-montasm" for 32-bit x86 disclaims code stability
1922 anyway, in this constellation we activate additional code
1923 backported from 0.9.9-dev for further performance improvements,
1924 namely BN_from_montgomery_word. (To enable this otherwise,
1925 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1927 [Andy Polyakov (backport partially by Bodo Moeller)]
1929 *) Add TLS session ticket callback. This allows an application to set
1930 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1931 values. This is useful for key rollover for example where several key
1932 sets may exist with different names.
1935 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1936 This was broken until now in 0.9.8 releases, such that the only way
1937 a registered ENGINE could be used (assuming it initialises
1938 successfully on the host) was to explicitly set it as the default
1939 for the relevant algorithms. This is in contradiction with 0.9.7
1940 behaviour and the documentation. With this fix, when an ENGINE is
1941 registered into a given algorithm's table of implementations, the
1942 'uptodate' flag is reset so that auto-discovery will be used next
1943 time a new context for that algorithm attempts to select an
1945 [Ian Lister (tweaked by Geoff Thorpe)]
1947 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1948 implemention in the following ways:
1950 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1953 Lack of BER streaming support means one pass streaming processing is
1954 only supported if data is detached: setting the streaming flag is
1955 ignored for embedded content.
1957 CMS support is disabled by default and must be explicitly enabled
1958 with the enable-cms configuration option.
1961 *) Update the GMP engine glue to do direct copies between BIGNUM and
1962 mpz_t when openssl and GMP use the same limb size. Otherwise the
1963 existing "conversion via a text string export" trick is still used.
1964 [Paul Sheer <paulsheer@gmail.com>]
1966 *) Zlib compression BIO. This is a filter BIO which compressed and
1967 uncompresses any data passed through it.
1970 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1971 RFC3394 compatible AES key wrapping.
1974 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1975 sets string data without copying. X509_ALGOR_set0() and
1976 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1977 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1978 from an X509_ATTRIBUTE structure optionally checking it occurs only
1979 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1983 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1984 to get the expected BN_FLG_CONSTTIME behavior.
1985 [Bodo Moeller (Google)]
1989 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1990 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1991 - added some more tests to do_tests.pl
1992 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1993 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1994 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1995 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1996 - various changes to netware.pl to enable gcc-cross builds on Win32
1998 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1999 - various changes to fix missing prototype warnings
2000 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2001 - added AES, WHIRLPOOL and CPUID assembler code to build files
2002 - added missing AES assembler make rules to mk1mf.pl
2003 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2004 [Guenter Knauf <eflash@gmx.net>]
2006 *) Implement certificate status request TLS extension defined in RFC3546.
2007 A client can set the appropriate parameters and receive the encoded
2008 OCSP response via a callback. A server can query the supplied parameters
2009 and set the encoded OCSP response in the callback. Add simplified examples
2010 to s_client and s_server.
2013 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2015 *) Fix various bugs:
2016 + Binary incompatibility of ssl_ctx_st structure
2017 + DTLS interoperation with non-compliant servers
2018 + Don't call get_session_cb() without proposed session
2019 + Fix ia64 assembler code
2020 [Andy Polyakov, Steve Henson]
2022 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2024 *) DTLS Handshake overhaul. There were longstanding issues with
2025 OpenSSL DTLS implementation, which were making it impossible for
2026 RFC 4347 compliant client to communicate with OpenSSL server.
2027 Unfortunately just fixing these incompatibilities would "cut off"
2028 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2029 server keeps tolerating non RFC compliant syntax. The opposite is
2030 not true, 0.9.8f client can not communicate with earlier server.
2031 This update even addresses CVE-2007-4995.
2034 *) Changes to avoid need for function casts in OpenSSL: some compilers
2035 (gcc 4.2 and later) reject their use.
2036 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2039 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2040 RFC4507bis. The encrypted ticket format is an encrypted encoded
2041 SSL_SESSION structure, that way new session features are automatically
2044 If a client application caches session in an SSL_SESSION structure
2045 support is transparent because tickets are now stored in the encoded
2048 The SSL_CTX structure automatically generates keys for ticket
2049 protection in servers so again support should be possible
2050 with no application modification.
2052 If a client or server wishes to disable RFC4507 support then the option
2053 SSL_OP_NO_TICKET can be set.
2055 Add a TLS extension debugging callback to allow the contents of any client
2056 or server extensions to be examined.
2058 This work was sponsored by Google.
2061 *) Add initial support for TLS extensions, specifically for the server_name
2062 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2063 have new members for a host name. The SSL data structure has an
2064 additional member SSL_CTX *initial_ctx so that new sessions can be
2065 stored in that context to allow for session resumption, even after the
2066 SSL has been switched to a new SSL_CTX in reaction to a client's
2067 server_name extension.
2069 New functions (subject to change):
2071 SSL_get_servername()
2072 SSL_get_servername_type()
2075 New CTRL codes and macros (subject to change):
2077 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2078 - SSL_CTX_set_tlsext_servername_callback()
2079 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2080 - SSL_CTX_set_tlsext_servername_arg()
2081 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2083 openssl s_client has a new '-servername ...' option.
2085 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2086 '-key2 ...', '-servername_fatal' (subject to change). This allows
2087 testing the HostName extension for a specific single host name ('-cert'
2088 and '-key' remain fallbacks for handshakes without HostName
2089 negotiation). If the unrecogninzed_name alert has to be sent, this by
2090 default is a warning; it becomes fatal with the '-servername_fatal'
2093 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2095 *) Add AES and SSE2 assembly language support to VC++ build.
2098 *) Mitigate attack on final subtraction in Montgomery reduction.
2101 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2102 (which previously caused an internal error).
2105 *) Squeeze another 10% out of IGE mode when in != out.
2108 *) AES IGE mode speedup.
2109 [Dean Gaudet (Google)]
2111 *) Add the Korean symmetric 128-bit cipher SEED (see
2112 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2113 add SEED ciphersuites from RFC 4162:
2115 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2116 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2117 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2118 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2120 To minimize changes between patchlevels in the OpenSSL 0.9.8
2121 series, SEED remains excluded from compilation unless OpenSSL
2122 is configured with 'enable-seed'.
2123 [KISA, Bodo Moeller]
2125 *) Mitigate branch prediction attacks, which can be practical if a
2126 single processor is shared, allowing a spy process to extract
2127 information. For detailed background information, see
2128 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2129 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2130 and Necessary Software Countermeasures"). The core of the change
2131 are new versions BN_div_no_branch() and
2132 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2133 respectively, which are slower, but avoid the security-relevant
2134 conditional branches. These are automatically called by BN_div()
2135 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2136 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2137 remove a conditional branch.
2139 BN_FLG_CONSTTIME is the new name for the previous
2140 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2141 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2142 in the exponent causes BN_mod_exp_mont() to use the alternative
2143 implementation in BN_mod_exp_mont_consttime().) The old name
2144 remains as a deprecated alias.
2146 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2147 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2148 constant-time implementations for more than just exponentiation.
2149 Here too the old name is kept as a deprecated alias.
2151 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2152 the BN_BLINDING structure gets an independent copy of the
2153 modulus. This means that the previous "BIGNUM *m" argument to
2154 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2155 essentially becomes "const BIGNUM *m", although we can't actually
2156 change this in the header file before 0.9.9. It allows
2157 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2158 enable BN_FLG_CONSTTIME.
2160 [Matthew D Wood (Intel Corp)]
2162 *) In the SSL/TLS server implementation, be strict about session ID
2163 context matching (which matters if an application uses a single
2164 external cache for different purposes). Previously,
2165 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2166 set. This did ensure strict client verification, but meant that,
2167 with applications using a single external cache for quite
2168 different requirements, clients could circumvent ciphersuite
2169 restrictions for a given session ID context by starting a session
2170 in a different context.
2173 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2174 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2175 authentication-only ciphersuites.
2178 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2179 not complete and could lead to a possible single byte overflow
2180 (CVE-2007-5135) [Ben Laurie]
2182 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2184 *) Since AES128 and AES256 (and similarly Camellia128 and
2185 Camellia256) share a single mask bit in the logic of
2186 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2187 kludge to work properly if AES128 is available and AES256 isn't
2188 (or if Camellia128 is available and Camellia256 isn't).
2191 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2192 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2193 When a point or a seed is encoded in a BIT STRING, we need to
2194 prevent the removal of trailing zero bits to get the proper DER
2195 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2196 of a NamedBitList, for which trailing 0 bits need to be removed.)
2199 *) Have SSL/TLS server implementation tolerate "mismatched" record
2200 protocol version while receiving ClientHello even if the
2201 ClientHello is fragmented. (The server can't insist on the
2202 particular protocol version it has chosen before the ServerHello
2203 message has informed the client about his choice.)
2206 *) Add RFC 3779 support.
2207 [Rob Austein for ARIN, Ben Laurie]
2209 *) Load error codes if they are not already present instead of using a
2210 static variable. This allows them to be cleanly unloaded and reloaded.
2211 Improve header file function name parsing.
2214 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2215 or CAPABILITY handshake as required by RFCs.
2218 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2220 *) Introduce limits to prevent malicious keys being able to
2221 cause a denial of service. (CVE-2006-2940)
2222 [Steve Henson, Bodo Moeller]
2224 *) Fix ASN.1 parsing of certain invalid structures that can result
2225 in a denial of service. (CVE-2006-2937) [Steve Henson]
2227 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2228 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2230 *) Fix SSL client code which could crash if connecting to a
2231 malicious SSLv2 server. (CVE-2006-4343)
2232 [Tavis Ormandy and Will Drewry, Google Security Team]
2234 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2235 match only those. Before that, "AES256-SHA" would be interpreted
2236 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2237 the same strength classification in 0.9.7h) as we currently only
2238 have a single AES bit in the ciphersuite description bitmap.
2239 That change, however, also applied to ciphersuite strings such as
2240 "RC4-MD5" that intentionally matched multiple ciphersuites --
2241 namely, SSL 2.0 ciphersuites in addition to the more common ones
2242 from SSL 3.0/TLS 1.0.
2244 So we change the selection algorithm again: Naming an explicit
2245 ciphersuite selects this one ciphersuite, and any other similar
2246 ciphersuite (same bitmap) from *other* protocol versions.
2247 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2248 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2250 Since SSL 2.0 does not have any ciphersuites for which the
2251 128/256 bit distinction would be relevant, this works for now.
2252 The proper fix will be to use different bits for AES128 and
2253 AES256, which would have avoided the problems from the beginning;
2254 however, bits are scarce, so we can only do this in a new release
2255 (not just a patchlevel) when we can change the SSL_CIPHER
2256 definition to split the single 'unsigned long mask' bitmap into
2257 multiple values to extend the available space.
2261 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2263 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2264 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2266 *) Add AES IGE and biIGE modes.
2269 *) Change the Unix randomness entropy gathering to use poll() when
2270 possible instead of select(), since the latter has some
2271 undesirable limitations.
2272 [Darryl Miles via Richard Levitte and Bodo Moeller]
2274 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2275 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2276 cannot be implicitly activated as part of, e.g., the "AES" alias.
2277 However, please upgrade to OpenSSL 0.9.9[-dev] for
2278 non-experimental use of the ECC ciphersuites to get TLS extension
2279 support, which is required for curve and point format negotiation
2280 to avoid potential handshake problems.
2283 *) Disable rogue ciphersuites:
2285 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2286 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2287 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2289 The latter two were purportedly from
2290 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2293 Also deactivate the remaining ciphersuites from
2294 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2295 unofficial, and the ID has long expired.
2298 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2299 dual-core machines) and other potential thread-safety issues.
2302 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2303 versions), which is now available for royalty-free use
2304 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2305 Also, add Camellia TLS ciphersuites from RFC 4132.
2307 To minimize changes between patchlevels in the OpenSSL 0.9.8
2308 series, Camellia remains excluded from compilation unless OpenSSL
2309 is configured with 'enable-camellia'.
2312 *) Disable the padding bug check when compression is in use. The padding
2313 bug check assumes the first packet is of even length, this is not
2314 necessarily true if compresssion is enabled and can result in false
2315 positives causing handshake failure. The actual bug test is ancient
2316 code so it is hoped that implementations will either have fixed it by
2317 now or any which still have the bug do not support compression.
2320 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2322 *) When applying a cipher rule check to see if string match is an explicit
2323 cipher suite and only match that one cipher suite if it is.
2326 *) Link in manifests for VC++ if needed.
2327 [Austin Ziegler <halostatue@gmail.com>]
2329 *) Update support for ECC-based TLS ciphersuites according to
2330 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2331 TLS extensions, which are supported starting with the 0.9.9
2332 branch, not in the OpenSSL 0.9.8 branch).
2335 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2336 opaque EVP_CIPHER_CTX handling.
2339 *) Fixes and enhancements to zlib compression code. We now only use
2340 "zlib1.dll" and use the default __cdecl calling convention on Win32
2341 to conform with the standards mentioned here:
2342 http://www.zlib.net/DLL_FAQ.txt
2343 Static zlib linking now works on Windows and the new --with-zlib-include
2344 --with-zlib-lib options to Configure can be used to supply the location
2345 of the headers and library. Gracefully handle case where zlib library
2349 *) Several fixes and enhancements to the OID generation code. The old code
2350 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2351 handle numbers larger than ULONG_MAX, truncated printing and had a
2352 non standard OBJ_obj2txt() behaviour.
2355 *) Add support for building of engines under engine/ as shared libraries
2356 under VC++ build system.
2359 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2360 Hopefully, we will not see any false combination of paths any more.
2363 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2365 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2366 (part of SSL_OP_ALL). This option used to disable the
2367 countermeasure against man-in-the-middle protocol-version
2368 rollback in the SSL 2.0 server implementation, which is a bad
2369 idea. (CVE-2005-2969)
2371 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2372 for Information Security, National Institute of Advanced Industrial
2373 Science and Technology [AIST], Japan)]
2375 *) Add two function to clear and return the verify parameter flags.
2378 *) Keep cipherlists sorted in the source instead of sorting them at
2379 runtime, thus removing the need for a lock.
2382 *) Avoid some small subgroup attacks in Diffie-Hellman.
2383 [Nick Mathewson and Ben Laurie]
2385 *) Add functions for well-known primes.
2388 *) Extended Windows CE support.
2389 [Satoshi Nakamura and Andy Polyakov]
2391 *) Initialize SSL_METHOD structures at compile time instead of during
2392 runtime, thus removing the need for a lock.
2395 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2396 attempting to decrypt each encrypted key in turn. Add support to
2400 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2402 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2405 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2408 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2409 key into the same file any more.
2412 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2415 *) Add -utf8 command line and config file option to 'ca'.
2416 [Stefan <stf@udoma.org]
2418 *) Removed the macro des_crypt(), as it seems to conflict with some
2419 libraries. Use DES_crypt().
2422 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2423 involves renaming the source and generated shared-libs for
2424 both. The engines will accept the corrected or legacy ids
2425 ('ncipher' and '4758_cca' respectively) when binding. NB,
2426 this only applies when building 'shared'.
2427 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2429 *) Add attribute functions to EVP_PKEY structure. Modify
2430 PKCS12_create() to recognize a CSP name attribute and
2431 use it. Make -CSP option work again in pkcs12 utility.
2434 *) Add new functionality to the bn blinding code:
2435 - automatic re-creation of the BN_BLINDING parameters after
2436 a fixed number of uses (currently 32)
2437 - add new function for parameter creation
2438 - introduce flags to control the update behaviour of the
2439 BN_BLINDING parameters
2440 - hide BN_BLINDING structure
2441 Add a second BN_BLINDING slot to the RSA structure to improve
2442 performance when a single RSA object is shared among several
2446 *) Add support for DTLS.
2447 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2449 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2450 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2453 *) Remove buggy and incompletet DH cert support from
2454 ssl/ssl_rsa.c and ssl/s3_both.c
2457 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2458 the apps/openssl applications.
2461 *) Compile clean with "-Wall -Wmissing-prototypes
2462 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2463 DEBUG_SAFESTACK must also be set.
2466 *) Change ./Configure so that certain algorithms can be disabled by default.
2467 The new counterpiece to "no-xxx" is "enable-xxx".
2469 The patented RC5 and MDC2 algorithms will now be disabled unless
2470 "enable-rc5" and "enable-mdc2", respectively, are specified.
2472 (IDEA remains enabled despite being patented. This is because IDEA
2473 is frequently required for interoperability, and there is no license
2474 fee for non-commercial use. As before, "no-idea" can be used to
2475 avoid this algorithm.)
2479 *) Add processing of proxy certificates (see RFC 3820). This work was
2480 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2481 EGEE (Enabling Grids for E-science in Europe).
2484 *) RC4 performance overhaul on modern architectures/implementations, such
2485 as Intel P4, IA-64 and AMD64.
2488 *) New utility extract-section.pl. This can be used specify an alternative
2489 section number in a pod file instead of having to treat each file as
2490 a separate case in Makefile. This can be done by adding two lines to the
2493 =for comment openssl_section:XXX
2495 The blank line is mandatory.
2499 *) New arguments -certform, -keyform and -pass for s_client and s_server
2500 to allow alternative format key and certificate files and passphrase
2504 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2505 update associated structures and add various utility functions.
2507 Add new policy related verify parameters, include policy checking in
2508 standard verify code. Enhance 'smime' application with extra parameters
2509 to support policy checking and print out.
2512 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2513 Nehemiah processors. These extensions support AES encryption in hardware
2514 as well as RNG (though RNG support is currently disabled).
2515 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2517 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2520 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2521 [Andy Polyakov and a number of other people]
2523 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2524 implementation contributed by IBM.
2525 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2527 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2528 exponent rather than 'unsigned long'. There is a corresponding change to
2529 the new 'rsa_keygen' element of the RSA_METHOD structure.
2530 [Jelte Jansen, Geoff Thorpe]
2532 *) Functionality for creating the initial serial number file is now
2533 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2535 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2536 number file to 1, which is bound to cause problems. To avoid
2537 the problems while respecting compatibility between different 0.9.7
2538 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2539 CA.pl for serial number initialization. With the new release 0.9.8,
2540 we can fix the problem directly in the 'ca' utility.)
2543 *) Reduced header interdepencies by declaring more opaque objects in
2544 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2545 give fewer recursive includes, which could break lazy source code - so
2546 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2547 developers should define this symbol when building and using openssl to
2548 ensure they track the recommended behaviour, interfaces, [etc], but
2549 backwards-compatible behaviour prevails when this isn't defined.
2552 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2555 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2556 This will generate a random key of the appropriate length based on the
2557 cipher context. The EVP_CIPHER can provide its own random key generation
2558 routine to support keys of a specific form. This is used in the des and
2559 3des routines to generate a key of the correct parity. Update S/MIME
2560 code to use new functions and hence generate correct parity DES keys.
2561 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2562 valid (weak or incorrect parity).
2565 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2566 as looking them up. This is useful when the verified structure may contain
2567 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2568 present unless the new PKCS7_NO_CRL flag is asserted.
2571 *) Extend ASN1 oid configuration module. It now additionally accepts the
2574 shortName = some long name, 1.2.3.4
2577 *) Reimplemented the BN_CTX implementation. There is now no more static
2578 limitation on the number of variables it can handle nor the depth of the
2579 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2580 information can now expand as required, and rather than having a single
2581 static array of bignums, BN_CTX now uses a linked-list of such arrays
2582 allowing it to expand on demand whilst maintaining the usefulness of
2583 BN_CTX's "bundling".
2586 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2587 to allow all RSA operations to function using a single BN_CTX.
2590 *) Preliminary support for certificate policy evaluation and checking. This
2591 is initially intended to pass the tests outlined in "Conformance Testing
2592 of Relying Party Client Certificate Path Processing Logic" v1.07.
2595 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2596 remained unused and not that useful. A variety of other little bignum
2597 tweaks and fixes have also been made continuing on from the audit (see
2601 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2602 associated ASN1, EVP and SSL functions and old ASN1 macros.
2605 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2606 and this should never fail. So the return value from the use of
2607 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2608 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2611 *) BN_CTX_get() should return zero-valued bignums, providing the same
2612 initialised value as BN_new().
2613 [Geoff Thorpe, suggested by Ulf Möller]
2615 *) Support for inhibitAnyPolicy certificate extension.
2618 *) An audit of the BIGNUM code is underway, for which debugging code is
2619 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2620 is considered valid when processing BIGNUMs, and causes execution to
2621 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2622 further steps are taken to deliberately pollute unused data in BIGNUM
2623 structures to try and expose faulty code further on. For now, openssl will
2624 (in its default mode of operation) continue to tolerate the inconsistent
2625 forms that it has tolerated in the past, but authors and packagers should
2626 consider trying openssl and their own applications when compiled with
2627 these debugging symbols defined. It will help highlight potential bugs in
2628 their own code, and will improve the test coverage for OpenSSL itself. At
2629 some point, these tighter rules will become openssl's default to improve
2630 maintainability, though the assert()s and other overheads will remain only
2631 in debugging configurations. See bn.h for more details.
2632 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2634 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2635 that can only be obtained through BN_CTX_new() (which implicitly
2636 initialises it). The presence of this function only made it possible
2637 to overwrite an existing structure (and cause memory leaks).
2640 *) Because of the callback-based approach for implementing LHASH as a
2641 template type, lh_insert() adds opaque objects to hash-tables and
2642 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2643 to clean up those corresponding objects before destroying the hash table
2644 (and losing the object pointers). So some over-zealous constifications in
2645 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2646 objects as "const" and the lh_doall[_arg] callback wrappers are not
2647 prototyped to have "const" restrictions on the object pointers they are
2648 given (and so aren't required to cast them away any more).
2651 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2652 (speed) prefers to use its own implementation. The two implementations
2653 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2654 its object type properly exposed (MS_TM) instead of casting to/from "char
2655 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2656 aren't necessarily the greatest nomenclatures - but this is what was used
2657 internally to the implementation so I've used that for now.
2660 *) Ensure that deprecated functions do not get compiled when
2661 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2662 the self-tests were still using deprecated key-generation functions so
2663 these have been updated also.
2666 *) Reorganise PKCS#7 code to separate the digest location functionality
2667 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2668 New function PKCS7_set_digest() to set the digest type for PKCS#7
2669 digestedData type. Add additional code to correctly generate the
2670 digestedData type and add support for this type in PKCS7 initialization
2674 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2675 structure of type "other".
2678 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2679 sure the loop does correctly stop and breaking ("division by zero")
2680 modulus operations are not performed. The (pre-generated) prime
2681 table crypto/bn/bn_prime.h was already correct, but it could not be
2682 re-generated on some platforms because of the "division by zero"
2683 situation in the script.
2684 [Ralf S. Engelschall]
2686 *) Update support for ECC-based TLS ciphersuites according to
2687 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2688 SHA-1 now is only used for "small" curves (where the
2689 representation of a field element takes up to 24 bytes); for
2690 larger curves, the field element resulting from ECDH is directly
2691 used as premaster secret.
2692 [Douglas Stebila (Sun Microsystems Laboratories)]
2694 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2695 curve secp160r1 to the tests.
2696 [Douglas Stebila (Sun Microsystems Laboratories)]
2698 *) Add the possibility to load symbols globally with DSO.
2699 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2701 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2702 control of the error stack.
2705 *) Add support for STORE in ENGINE.
2708 *) Add the STORE type. The intention is to provide a common interface
2709 to certificate and key stores, be they simple file-based stores, or
2710 HSM-type store, or LDAP stores, or...
2711 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2714 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2715 pass a list of arguments to any function as well as provide a way
2716 for a function to pass data back to the caller.
2719 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2720 works like BUF_strdup() but can be used to duplicate a portion of
2721 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2725 *) Add the function sk_find_ex() which works like sk_find(), but will
2726 return an index to an element even if an exact match couldn't be
2727 found. The index is guaranteed to point at the element where the
2728 searched-for key would be inserted to preserve sorting order.
2731 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2732 takes an extra flags argument for optional functionality. Currently,
2733 the following flags are defined:
2735 OBJ_BSEARCH_VALUE_ON_NOMATCH
2736 This one gets OBJ_bsearch_ex() to return a pointer to the first
2737 element where the comparing function returns a negative or zero
2740 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2741 This one gets OBJ_bsearch_ex() to return a pointer to the first
2742 element where the comparing function returns zero. This is useful
2743 if there are more than one element where the comparing function
2747 *) Make it possible to create self-signed certificates with 'openssl ca'
2748 in such a way that the self-signed certificate becomes part of the
2749 CA database and uses the same mechanisms for serial number generation
2750 as all other certificate signing. The new flag '-selfsign' enables
2751 this functionality. Adapt CA.sh and CA.pl.in.
2754 *) Add functionality to check the public key of a certificate request
2755 against a given private. This is useful to check that a certificate
2756 request can be signed by that key (self-signing).
2759 *) Make it possible to have multiple active certificates with the same
2760 subject in the CA index file. This is done only if the keyword
2761 'unique_subject' is set to 'no' in the main CA section (default
2762 if 'CA_default') of the configuration file. The value is saved
2763 with the database itself in a separate index attribute file,
2764 named like the index file with '.attr' appended to the name.
2767 *) Generate muti valued AVAs using '+' notation in config files for
2771 *) Support for nameConstraints certificate extension.
2774 *) Support for policyConstraints certificate extension.
2777 *) Support for policyMappings certificate extension.
2780 *) Make sure the default DSA_METHOD implementation only uses its
2781 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2782 and change its own handlers to be NULL so as to remove unnecessary
2783 indirection. This lets alternative implementations fallback to the
2784 default implementation more easily.
2787 *) Support for directoryName in GeneralName related extensions
2791 *) Make it possible to link applications using Makefile.shared.
2792 Make that possible even when linking against static libraries!
2795 *) Support for single pass processing for S/MIME signing. This now
2796 means that S/MIME signing can be done from a pipe, in addition
2797 cleartext signing (multipart/signed type) is effectively streaming
2798 and the signed data does not need to be all held in memory.
2800 This is done with a new flag PKCS7_STREAM. When this flag is set
2801 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2802 is done after the data is output (and digests calculated) in
2803 SMIME_write_PKCS7().
2806 *) Add full support for -rpath/-R, both in shared libraries and
2807 applications, at least on the platforms where it's known how
2811 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2812 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2813 will now compute a table of multiples of the generator that
2814 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2815 faster (notably in the case of a single point multiplication,
2816 scalar * generator).
2817 [Nils Larsch, Bodo Moeller]
2819 *) IPv6 support for certificate extensions. The various extensions
2820 which use the IP:a.b.c.d can now take IPv6 addresses using the
2821 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2825 *) Added an ENGINE that implements RSA by performing private key
2826 exponentiations with the GMP library. The conversions to and from
2827 GMP's mpz_t format aren't optimised nor are any montgomery forms
2828 cached, and on x86 it appears OpenSSL's own performance has caught up.
2829 However there are likely to be other architectures where GMP could
2830 provide a boost. This ENGINE is not built in by default, but it can be
2831 specified at Configure time and should be accompanied by the necessary
2832 linker additions, eg;
2833 ./config -DOPENSSL_USE_GMP -lgmp
2836 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2837 testing availability of engines with "-t" - the old behaviour is
2838 produced by increasing the feature's verbosity with "-tt".
2841 *) ECDSA routines: under certain error conditions uninitialized BN objects
2842 could be freed. Solution: make sure initialization is performed early
2843 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2847 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2848 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2849 software implementations. For DSA and DH, parameter generation can
2850 also be overriden by providing the appropriate method callbacks.
2853 *) Change the "progress" mechanism used in key-generation and
2854 primality testing to functions that take a new BN_GENCB pointer in
2855 place of callback/argument pairs. The new API functions have "_ex"
2856 postfixes and the older functions are reimplemented as wrappers for
2857 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2858 declarations of the old functions to help (graceful) attempts to
2859 migrate to the new functions. Also, the new key-generation API
2860 functions operate on a caller-supplied key-structure and return
2861 success/failure rather than returning a key or NULL - this is to
2862 help make "keygen" another member function of RSA_METHOD etc.
2864 Example for using the new callback interface:
2866 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2870 BN_GENCB_set(&my_cb, my_callback, my_arg);
2872 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2873 /* For the meaning of a, b in calls to my_callback(), see the
2874 * documentation of the function that calls the callback.
2875 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2876 * my_callback should return 1 if it wants BN_is_prime_ex()
2877 * to continue, or 0 to stop.
2882 *) Change the ZLIB compression method to be stateful, and make it
2883 available to TLS with the number defined in
2884 draft-ietf-tls-compression-04.txt.
2887 *) Add the ASN.1 structures and functions for CertificatePair, which
2888 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2890 CertificatePair ::= SEQUENCE {
2891 forward [0] Certificate OPTIONAL,
2892 reverse [1] Certificate OPTIONAL,
2893 -- at least one of the pair shall be present -- }
2895 Also implement the PEM functions to read and write certificate
2896 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2898 This needed to be defined, mostly for the sake of the LDAP
2899 attribute crossCertificatePair, but may prove useful elsewhere as
2903 *) Make it possible to inhibit symlinking of shared libraries in
2904 Makefile.shared, for Cygwin's sake.
2907 *) Extend the BIGNUM API by creating a function
2908 void BN_set_negative(BIGNUM *a, int neg);
2909 and a macro that behave like
2910 int BN_is_negative(const BIGNUM *a);
2912 to avoid the need to access 'a->neg' directly in applications.
2915 *) Implement fast modular reduction for pseudo-Mersenne primes
2916 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2917 EC_GROUP_new_curve_GFp() will now automatically use this
2919 [Nils Larsch <nla@trustcenter.de>]
2921 *) Add new lock type (CRYPTO_LOCK_BN).
2924 *) Change the ENGINE framework to automatically load engines
2925 dynamically from specific directories unless they could be
2926 found to already be built in or loaded. Move all the
2927 current engines except for the cryptodev one to a new
2929 The engines in engines/ are built as shared libraries if
2930 the "shared" options was given to ./Configure or ./config.
2931 Otherwise, they are inserted in libcrypto.a.
2932 /usr/local/ssl/engines is the default directory for dynamic
2933 engines, but that can be overriden at configure time through
2934 the usual use of --prefix and/or --openssldir, and at run
2935 time with the environment variable OPENSSL_ENGINES.
2936 [Geoff Thorpe and Richard Levitte]
2938 *) Add Makefile.shared, a helper makefile to build shared
2939 libraries. Addapt Makefile.org.
2942 *) Add version info to Win32 DLLs.
2943 [Peter 'Luna' Runestig" <peter@runestig.com>]
2945 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2946 can be added using this API to created arbitrary PKCS#12
2947 files while avoiding the low level API.
2949 New options to PKCS12_create(), key or cert can be NULL and
2950 will then be omitted from the output file. The encryption
2951 algorithm NIDs can be set to -1 for no encryption, the mac
2952 iteration count can be set to 0 to omit the mac.
2954 Enhance pkcs12 utility by making the -nokeys and -nocerts
2955 options work when creating a PKCS#12 file. New option -nomac
2956 to omit the mac, NONE can be set for an encryption algorithm.
2957 New code is modified to use the enhanced PKCS12_create()
2958 instead of the low level API.
2961 *) Extend ASN1 encoder to support indefinite length constructed
2962 encoding. This can output sequences tags and octet strings in
2963 this form. Modify pk7_asn1.c to support indefinite length
2964 encoding. This is experimental and needs additional code to
2965 be useful, such as an ASN1 bio and some enhanced streaming
2968 Extend template encode functionality so that tagging is passed
2969 down to the template encoder.
2972 *) Let 'openssl req' fail if an argument to '-newkey' is not
2973 recognized instead of using RSA as a default.
2976 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2977 As these are not official, they are not included in "ALL";
2978 the "ECCdraft" ciphersuite group alias can be used to select them.
2979 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2981 *) Add ECDH engine support.
2982 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2984 *) Add ECDH in new directory crypto/ecdh/.
2985 [Douglas Stebila (Sun Microsystems Laboratories)]
2987 *) Let BN_rand_range() abort with an error after 100 iterations
2988 without success (which indicates a broken PRNG).
2991 *) Change BN_mod_sqrt() so that it verifies that the input value
2992 is really the square of the return value. (Previously,
2993 BN_mod_sqrt would show GIGO behaviour.)
2996 *) Add named elliptic curves over binary fields from X9.62, SECG,
2997 and WAP/WTLS; add OIDs that were still missing.
2999 [Sheueling Chang Shantz and Douglas Stebila
3000 (Sun Microsystems Laboratories)]
3002 *) Extend the EC library for elliptic curves over binary fields
3003 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3006 EC_GF2m_simple_method
3010 EC_GROUP_new_curve_GF2m
3011 EC_GROUP_set_curve_GF2m
3012 EC_GROUP_get_curve_GF2m
3013 EC_POINT_set_affine_coordinates_GF2m
3014 EC_POINT_get_affine_coordinates_GF2m
3015 EC_POINT_set_compressed_coordinates_GF2m
3017 Point compression for binary fields is disabled by default for
3018 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3021 As binary polynomials are represented as BIGNUMs, various members
3022 of the EC_GROUP and EC_POINT data structures can be shared
3023 between the implementations for prime fields and binary fields;
3024 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3025 are essentially identical to their ..._GFp counterparts.
3026 (For simplicity, the '..._GFp' prefix has been dropped from
3027 various internal method names.)
3029 An internal 'field_div' method (similar to 'field_mul' and
3030 'field_sqr') has been added; this is used only for binary fields.
3032 [Sheueling Chang Shantz and Douglas Stebila
3033 (Sun Microsystems Laboratories)]
3035 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3036 through methods ('mul', 'precompute_mult').
3038 The generic implementations (now internally called 'ec_wNAF_mul'
3039 and 'ec_wNAF_precomputed_mult') remain the default if these
3040 methods are undefined.
3042 [Sheueling Chang Shantz and Douglas Stebila
3043 (Sun Microsystems Laboratories)]
3045 *) New function EC_GROUP_get_degree, which is defined through
3046 EC_METHOD. For curves over prime fields, this returns the bit
3047 length of the modulus.
3049 [Sheueling Chang Shantz and Douglas Stebila
3050 (Sun Microsystems Laboratories)]
3052 *) New functions EC_GROUP_dup, EC_POINT_dup.
3053 (These simply call ..._new and ..._copy).
3055 [Sheueling Chang Shantz and Douglas Stebila
3056 (Sun Microsystems Laboratories)]
3058 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3059 Polynomials are represented as BIGNUMs (where the sign bit is not
3060 used) in the following functions [macros]:
3063 BN_GF2m_sub [= BN_GF2m_add]
3064 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3065 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3066 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3068 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3069 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3070 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3071 BN_GF2m_cmp [= BN_ucmp]
3073 (Note that only the 'mod' functions are actually for fields GF(2^m).
3074 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3076 For some functions, an the irreducible polynomial defining a
3077 field can be given as an 'unsigned int[]' with strictly
3078 decreasing elements giving the indices of those bits that are set;
3079 i.e., p[] represents the polynomial
3080 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3082 p[0] > p[1] > ... > p[k] = 0.
3083 This applies to the following functions:
3088 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3089 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3091 BN_GF2m_mod_sqrt_arr
3092 BN_GF2m_mod_solve_quad_arr
3096 Conversion can be performed by the following functions:
3101 bntest.c has additional tests for binary polynomial arithmetic.
3103 Two implementations for BN_GF2m_mod_div() are available.
3104 The default algorithm simply uses BN_GF2m_mod_inv() and
3105 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3106 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3107 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3109 [Sheueling Chang Shantz and Douglas Stebila
3110 (Sun Microsystems Laboratories)]
3112 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3113 functionality is disabled at compile-time.
3114 [Douglas Stebila <douglas.stebila@sun.com>]
3116 *) Change default behaviour of 'openssl asn1parse' so that more
3117 information is visible when viewing, e.g., a certificate:
3119 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3120 mode the content of non-printable OCTET STRINGs is output in a
3121 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3122 avoid the appearance of a printable string.
3123 [Nils Larsch <nla@trustcenter.de>]
3125 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3127 EC_GROUP_set_asn1_flag()
3128 EC_GROUP_get_asn1_flag()
3129 EC_GROUP_set_point_conversion_form()
3130 EC_GROUP_get_point_conversion_form()
3131 These control ASN1 encoding details:
3132 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3133 has been set to OPENSSL_EC_NAMED_CURVE.
3134 - Points are encoded in uncompressed form by default; options for
3135 asn1_for are as for point2oct, namely
3136 POINT_CONVERSION_COMPRESSED
3137 POINT_CONVERSION_UNCOMPRESSED
3138 POINT_CONVERSION_HYBRID
3140 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3143 EC_GROUP_get0_seed()
3144 EC_GROUP_get_seed_len()
3145 This is used only for ASN1 purposes (so far).
3146 [Nils Larsch <nla@trustcenter.de>]
3148 *) Add 'field_type' member to EC_METHOD, which holds the NID
3149 of the appropriate field type OID. The new function
3150 EC_METHOD_get_field_type() returns this value.
3151 [Nils Larsch <nla@trustcenter.de>]
3156 EC_POINT_point2hex()
3157 EC_POINT_hex2point()
3158 providing useful interfaces to EC_POINT_point2oct() and
3159 EC_POINT_oct2point().
3160 [Nils Larsch <nla@trustcenter.de>]
3162 *) Change internals of the EC library so that the functions
3163 EC_GROUP_set_generator()
3164 EC_GROUP_get_generator()
3165 EC_GROUP_get_order()
3166 EC_GROUP_get_cofactor()
3167 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3168 to methods, which would lead to unnecessary code duplication when
3169 adding different types of curves.
3170 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3172 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3173 arithmetic, and such that modified wNAFs are generated
3174 (which avoid length expansion in many cases).
3177 *) Add a function EC_GROUP_check_discriminant() (defined via
3178 EC_METHOD) that verifies that the curve discriminant is non-zero.
3180 Add a function EC_GROUP_check() that makes some sanity tests
3181 on a EC_GROUP, its generator and order. This includes
3182 EC_GROUP_check_discriminant().
3183 [Nils Larsch <nla@trustcenter.de>]
3185 *) Add ECDSA in new directory crypto/ecdsa/.
3187 Add applications 'openssl ecparam' and 'openssl ecdsa'
3188 (these are based on 'openssl dsaparam' and 'openssl dsa').
3190 ECDSA support is also included in various other files across the
3191 library. Most notably,
3192 - 'openssl req' now has a '-newkey ecdsa:file' option;
3193 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3194 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3195 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3196 them suitable for ECDSA where domain parameters must be
3197 extracted before the specific public key;
3198 - ECDSA engine support has been added.
3199 [Nils Larsch <nla@trustcenter.de>]
3201 *) Include some named elliptic curves, and add OIDs from X9.62,
3202 SECG, and WAP/WTLS. Each curve can be obtained from the new
3204 EC_GROUP_new_by_curve_name(),
3205 and the list of available named curves can be obtained with
3206 EC_get_builtin_curves().
3207 Also add a 'curve_name' member to EC_GROUP objects, which can be
3209 EC_GROUP_set_curve_name()
3210 EC_GROUP_get_curve_name()
3211 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3213 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3214 was actually never needed) and in BN_mul(). The removal in BN_mul()
3215 required a small change in bn_mul_part_recursive() and the addition
3216 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3217 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3218 bn_sub_words() and bn_add_words() except they take arrays with
3222 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3224 *) Cleanse PEM buffers before freeing them since they may contain
3226 [Benjamin Bennett <ben@psc.edu>]
3228 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3229 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3230 authentication-only ciphersuites.
3233 *) Since AES128 and AES256 share a single mask bit in the logic of
3234 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3235 kludge to work properly if AES128 is available and AES256 isn't.
3238 *) Expand security boundary to match 1.1.1 module.
3241 *) Remove redundant features: hash file source, editing of test vectors
3242 modify fipsld to use external fips_premain.c signature.
3245 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3246 run algorithm test programs.
3249 *) Make algorithm test programs more tolerant of whitespace.
3252 *) Have SSL/TLS server implementation tolerate "mismatched" record
3253 protocol version while receiving ClientHello even if the
3254 ClientHello is fragmented. (The server can't insist on the
3255 particular protocol version it has chosen before the ServerHello
3256 message has informed the client about his choice.)
3259 *) Load error codes if they are not already present instead of using a
3260 static variable. This allows them to be cleanly unloaded and reloaded.
3263 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3265 *) Introduce limits to prevent malicious keys being able to
3266 cause a denial of service. (CVE-2006-2940)
3267 [Steve Henson, Bodo Moeller]
3269 *) Fix ASN.1 parsing of certain invalid structures that can result
3270 in a denial of service. (CVE-2006-2937) [Steve Henson]
3272 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3273 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3275 *) Fix SSL client code which could crash if connecting to a
3276 malicious SSLv2 server. (CVE-2006-4343)
3277 [Tavis Ormandy and Will Drewry, Google Security Team]
3279 *) Change ciphersuite string processing so that an explicit
3280 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3281 will no longer include "AES128-SHA"), and any other similar
3282 ciphersuite (same bitmap) from *other* protocol versions (so that
3283 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3284 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3285 changes from 0.9.8b and 0.9.8d.
3288 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3290 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3291 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3293 *) Change the Unix randomness entropy gathering to use poll() when
3294 possible instead of select(), since the latter has some
3295 undesirable limitations.
3296 [Darryl Miles via Richard Levitte and Bodo Moeller]
3298 *) Disable rogue ciphersuites:
3300 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3301 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3302 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3304 The latter two were purportedly from
3305 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3308 Also deactive the remaining ciphersuites from
3309 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3310 unofficial, and the ID has long expired.
3313 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3314 dual-core machines) and other potential thread-safety issues.
3317 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3319 *) Adapt fipsld and the build system to link against the validated FIPS
3320 module in FIPS mode.
3323 *) Fixes for VC++ 2005 build under Windows.
3326 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3327 from a Windows bash shell such as MSYS. It is autodetected from the
3328 "config" script when run from a VC++ environment. Modify standard VC++
3329 build to use fipscanister.o from the GNU make build.
3332 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3334 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3335 The value now differs depending on if you build for FIPS or not.
3336 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3337 safely run with a non-FIPSed libcrypto, as it may crash because of
3338 the difference induced by this change.
3341 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3343 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3344 (part of SSL_OP_ALL). This option used to disable the
3345 countermeasure against man-in-the-middle protocol-version
3346 rollback in the SSL 2.0 server implementation, which is a bad
3347 idea. (CVE-2005-2969)
3349 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3350 for Information Security, National Institute of Advanced Industrial
3351 Science and Technology [AIST], Japan)]
3353 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3354 mainly for FIPS compliance and not fully integrated at this stage.
3357 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3358 the exponentiation using a fixed-length exponent. (Otherwise,
3359 the information leaked through timing could expose the secret key
3360 after many signatures; cf. Bleichenbacher's attack on DSA with
3364 *) Make a new fixed-window mod_exp implementation the default for
3365 RSA, DSA, and DH private-key operations so that the sequence of
3366 squares and multiplies and the memory access pattern are
3367 independent of the particular secret key. This will mitigate
3368 cache-timing and potential related attacks.
3370 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3371 and this is automatically used by BN_mod_exp_mont() if the new flag
3372 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3373 will use this BN flag for private exponents unless the flag
3374 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3375 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3377 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3379 *) Change the client implementation for SSLv23_method() and
3380 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3381 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3382 (Previously, the SSL 2.0 backwards compatible Client Hello
3383 message format would be used even with SSL_OP_NO_SSLv2.)
3386 *) Add support for smime-type MIME parameter in S/MIME messages which some
3390 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3391 a threadsafe manner. Modify rsa code to use new function and add calls
3392 to dsa and dh code (which had race conditions before).
3395 *) Include the fixed error library code in the C error file definitions
3396 instead of fixing them up at runtime. This keeps the error code
3397 structures constant.
3400 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3402 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3405 *) Fixes for newer kerberos headers. NB: the casts are needed because
3406 the 'length' field is signed on one version and unsigned on another
3407 with no (?) obvious way to tell the difference, without these VC++
3408 complains. Also the "definition" of FAR (blank) is no longer included
3409 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3410 some needed definitions.
3413 *) Undo Cygwin change.
3416 *) Added support for proxy certificates according to RFC 3820.
3417 Because they may be a security thread to unaware applications,
3418 they must be explicitely allowed in run-time. See
3419 docs/HOWTO/proxy_certificates.txt for further information.
3422 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3424 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3425 server and client random values. Previously
3426 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3427 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3429 This change has negligible security impact because:
3431 1. Server and client random values still have 24 bytes of pseudo random
3434 2. Server and client random values are sent in the clear in the initial
3437 3. The master secret is derived using the premaster secret (48 bytes in
3438 size for static RSA ciphersuites) as well as client server and random
3441 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3444 [Stephen Henson, reported by UK NISCC]
3446 *) Use Windows randomness collection on Cygwin.
3449 *) Fix hang in EGD/PRNGD query when communication socket is closed
3450 prematurely by EGD/PRNGD.
3451 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3453 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3456 *) Back-port of selected performance improvements from development
3457 branch, as well as improved support for PowerPC platforms.
3460 *) Add lots of checks for memory allocation failure, error codes to indicate
3461 failure and freeing up memory if a failure occurs.
3462 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3464 *) Add new -passin argument to dgst.
3467 *) Perform some character comparisons of different types in X509_NAME_cmp:
3468 this is needed for some certificates that reencode DNs into UTF8Strings
3469 (in violation of RFC3280) and can't or wont issue name rollover
3473 *) Make an explicit check during certificate validation to see that
3474 the CA setting in each certificate on the chain is correct. As a
3475 side effect always do the following basic checks on extensions,
3476 not just when there's an associated purpose to the check:
3478 - if there is an unhandled critical extension (unless the user
3479 has chosen to ignore this fault)
3480 - if the path length has been exceeded (if one is set at all)
3481 - that certain extensions fit the associated purpose (if one has
3485 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3487 *) Avoid a race condition when CRLs are checked in a multi threaded
3488 environment. This would happen due to the reordering of the revoked
3489 entries during signature checking and serial number lookup. Now the
3490 encoding is cached and the serial number sort performed under a lock.
3491 Add new STACK function sk_is_sorted().
3494 *) Add Delta CRL to the extension code.
3497 *) Various fixes to s3_pkt.c so alerts are sent properly.
3498 [David Holmes <d.holmes@f5.com>]
3500 *) Reduce the chances of duplicate issuer name and serial numbers (in
3501 violation of RFC3280) using the OpenSSL certificate creation utilities.
3502 This is done by creating a random 64 bit value for the initial serial
3503 number when a serial number file is created or when a self signed
3504 certificate is created using 'openssl req -x509'. The initial serial
3505 number file is created using 'openssl x509 -next_serial' in CA.pl
3506 rather than being initialized to 1.
3509 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3511 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3512 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3513 [Joe Orton, Steve Henson]
3515 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3517 [Joe Orton, Steve Henson]
3519 *) Make it possible to have multiple active certificates with the same
3520 subject in the CA index file. This is done only if the keyword
3521 'unique_subject' is set to 'no' in the main CA section (default
3522 if 'CA_default') of the configuration file. The value is saved
3523 with the database itself in a separate index attribute file,
3524 named like the index file with '.attr' appended to the name.
3527 *) X509 verify fixes. Disable broken certificate workarounds when
3528 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3529 keyUsage extension present. Don't accept CRLs with unhandled critical
3530 extensions: since verify currently doesn't process CRL extensions this
3531 rejects a CRL with *any* critical extensions. Add new verify error codes
3535 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3536 A clarification of RFC2560 will require the use of OCTET STRINGs and
3537 some implementations cannot handle the current raw format. Since OpenSSL
3538 copies and compares OCSP nonces as opaque blobs without any attempt at
3539 parsing them this should not create any compatibility issues.
3542 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3543 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3544 this HMAC (and other) operations are several times slower than OpenSSL
3548 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3549 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3551 *) Use the correct content when signing type "other".
3554 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3556 *) Fix various bugs revealed by running the NISCC test suite:
3558 Stop out of bounds reads in the ASN1 code when presented with
3559 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3561 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3563 If verify callback ignores invalid public key errors don't try to check
3564 certificate signature with the NULL public key.
3568 *) New -ignore_err option in ocsp application to stop the server
3569 exiting on the first error in a request.
3572 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3573 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3577 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3578 extra data after the compression methods not only for TLS 1.0
3579 but also for SSL 3.0 (as required by the specification).
3580 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3582 *) Change X509_certificate_type() to mark the key as exported/exportable
3583 when it's 512 *bits* long, not 512 bytes.
3586 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3587 blocks during encryption.
3590 *) Various fixes to base64 BIO and non blocking I/O. On write
3591 flushes were not handled properly if the BIO retried. On read
3592 data was not being buffered properly and had various logic bugs.
3593 This also affects blocking I/O when the data being decoded is a
3597 *) Various S/MIME bugfixes and compatibility changes:
3598 output correct application/pkcs7 MIME type if
3599 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3600 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3601 of files as .eml work). Correctly handle very long lines in MIME
3605 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3607 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3608 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3609 a protocol version number mismatch like a decryption error
3610 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3613 *) Turn on RSA blinding by default in the default implementation
3614 to avoid a timing attack. Applications that don't want it can call
3615 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3616 They would be ill-advised to do so in most cases.
3617 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3619 *) Change RSA blinding code so that it works when the PRNG is not
3620 seeded (in this case, the secret RSA exponent is abused as
3621 an unpredictable seed -- if it is not unpredictable, there
3622 is no point in blinding anyway). Make RSA blinding thread-safe
3623 by remembering the creator's thread ID in rsa->blinding and
3624 having all other threads use local one-time blinding factors
3625 (this requires more computation than sharing rsa->blinding, but
3626 avoids excessive locking; and if an RSA object is not shared
3627 between threads, blinding will still be very fast).
3630 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3631 ENGINE as defaults for all supported algorithms irrespective of
3632 the 'flags' parameter. 'flags' is now honoured, so applications
3633 should make sure they are passing it correctly.
3636 *) Target "mingw" now allows native Windows code to be generated in
3637 the Cygwin environment as well as with the MinGW compiler.
3640 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3642 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3643 via timing by performing a MAC computation even if incorrrect
3644 block cipher padding has been found. This is a countermeasure
3645 against active attacks where the attacker has to distinguish
3646 between bad padding and a MAC verification error. (CVE-2003-0078)
3648 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3649 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3650 Martin Vuagnoux (EPFL, Ilion)]
3652 *) Make the no-err option work as intended. The intention with no-err
3653 is not to have the whole error stack handling routines removed from
3654 libcrypto, it's only intended to remove all the function name and
3655 reason texts, thereby removing some of the footprint that may not
3656 be interesting if those errors aren't displayed anyway.
3658 NOTE: it's still possible for any application or module to have it's
3659 own set of error texts inserted. The routines are there, just not
3660 used by default when no-err is given.
3663 *) Add support for FreeBSD on IA64.
3664 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3666 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3667 Kerberos function mit_des_cbc_cksum(). Before this change,
3668 the value returned by DES_cbc_cksum() was like the one from
3669 mit_des_cbc_cksum(), except the bytes were swapped.
3670 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3672 *) Allow an application to disable the automatic SSL chain building.
3673 Before this a rather primitive chain build was always performed in
3674 ssl3_output_cert_chain(): an application had no way to send the
3675 correct chain if the automatic operation produced an incorrect result.
3677 Now the chain builder is disabled if either:
3679 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3681 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3683 The reasoning behind this is that an application would not want the
3684 auto chain building to take place if extra chain certificates are
3685 present and it might also want a means of sending no additional
3686 certificates (for example the chain has two certificates and the
3690 *) Add the possibility to build without the ENGINE framework.
3691 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3693 *) Under Win32 gmtime() can return NULL: check return value in
3694 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3697 *) DSA routines: under certain error conditions uninitialized BN objects
3698 could be freed. Solution: make sure initialization is performed early
3699 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3700 Nils Larsch <nla@trustcenter.de> via PR#459)
3703 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3704 checked on reconnect on the client side, therefore session resumption
3705 could still fail with a "ssl session id is different" error. This
3706 behaviour is masked when SSL_OP_ALL is used due to
3707 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3708 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3709 followup to PR #377.
3712 *) IA-32 assembler support enhancements: unified ELF targets, support
3713 for SCO/Caldera platforms, fix for Cygwin shared build.
3716 *) Add support for FreeBSD on sparc64. As a consequence, support for
3717 FreeBSD on non-x86 processors is separate from x86 processors on
3718 the config script, much like the NetBSD support.
3719 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3721 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3723 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3726 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3727 code (06) was taken as the first octet of the session ID and the last
3728 octet was ignored consequently. As a result SSLv2 client side session
3729 caching could not have worked due to the session ID mismatch between
3731 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3735 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3736 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3740 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3741 seems that in spite of existing for more than a year, many application
3742 author have done nothing to provide the necessary callbacks, which
3743 means that this particular engine will not work properly anywhere.
3744 This is a very unfortunate situation which forces us, in the name
3745 of usability, to give the hw_ncipher.c a static lock, which is part
3747 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3748 appear in 0.9.8 or later. We EXPECT application authors to have
3749 dealt properly with this when 0.9.8 is released (unless we actually
3750 make such changes in the libcrypto locking code that changes will
3751 have to be made anyway).
3754 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3755 octets have been read, EOF or an error occurs. Without this change
3756 some truncated ASN1 structures will not produce an error.
3759 *) Disable Heimdal support, since it hasn't been fully implemented.
3760 Still give the possibility to force the use of Heimdal, but with
3761 warnings and a request that patches get sent to openssl-dev.
3764 *) Add the VC-CE target, introduce the WINCE sysname, and add
3765 INSTALL.WCE and appropriate conditionals to make it build.
3766 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3768 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3769 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3770 edit numbers of the version.
3771 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3773 *) Introduce safe string copy and catenation functions
3774 (BUF_strlcpy() and BUF_strlcat()).
3775 [Ben Laurie (CHATS) and Richard Levitte]
3777 *) Avoid using fixed-size buffers for one-line DNs.
3778 [Ben Laurie (CHATS)]
3780 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3781 resizing buffers containing secrets, and use where appropriate.
3782 [Ben Laurie (CHATS)]
3784 *) Avoid using fixed size buffers for configuration file location.
3785 [Ben Laurie (CHATS)]
3787 *) Avoid filename truncation for various CA files.
3788 [Ben Laurie (CHATS)]
3790 *) Use sizeof in preference to magic numbers.
3791 [Ben Laurie (CHATS)]
3793 *) Avoid filename truncation in cert requests.
3794 [Ben Laurie (CHATS)]
3796 *) Add assertions to check for (supposedly impossible) buffer
3798 [Ben Laurie (CHATS)]
3800 *) Don't cache truncated DNS entries in the local cache (this could
3801 potentially lead to a spoofing attack).
3802 [Ben Laurie (CHATS)]
3804 *) Fix various buffers to be large enough for hex/decimal
3805 representations in a platform independent manner.
3806 [Ben Laurie (CHATS)]
3808 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3809 resizing buffers containing secrets, and use where appropriate.
3810 [Ben Laurie (CHATS)]
3812 *) Add BIO_indent() to avoid much slightly worrying code to do
3814 [Ben Laurie (CHATS)]
3816 *) Convert sprintf()/BIO_puts() to BIO_printf().
3817 [Ben Laurie (CHATS)]
3819 *) buffer_gets() could terminate with the buffer only half
3821 [Ben Laurie (CHATS)]
3823 *) Add assertions to prevent user-supplied crypto functions from
3824 overflowing internal buffers by having large block sizes, etc.
3825 [Ben Laurie (CHATS)]
3827 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3829 [Ben Laurie (CHATS)]
3831 *) Eliminate unused copy of key in RC4.
3832 [Ben Laurie (CHATS)]
3834 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3835 [Ben Laurie (CHATS)]
3837 *) Fix off-by-one error in EGD path.
3838 [Ben Laurie (CHATS)]
3840 *) If RANDFILE path is too long, ignore instead of truncating.
3841 [Ben Laurie (CHATS)]
3843 *) Eliminate unused and incorrectly sized X.509 structure
3845 [Ben Laurie (CHATS)]
3847 *) Eliminate unused and dangerous function knumber().
3848 [Ben Laurie (CHATS)]
3850 *) Eliminate unused and dangerous structure, KSSL_ERR.
3851 [Ben Laurie (CHATS)]
3853 *) Protect against overlong session ID context length in an encoded
3854 session object. Since these are local, this does not appear to be
3856 [Ben Laurie (CHATS)]
3858 *) Change from security patch (see 0.9.6e below) that did not affect
3859 the 0.9.6 release series:
3861 Remote buffer overflow in SSL3 protocol - an attacker could
3862 supply an oversized master key in Kerberos-enabled versions.
3864 [Ben Laurie (CHATS)]
3866 *) Change the SSL kerb5 codes to match RFC 2712.
3869 *) Make -nameopt work fully for req and add -reqopt switch.
3870 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3872 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3873 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3875 *) Make sure tests can be performed even if the corresponding algorithms
3876 have been removed entirely. This was also the last step to make
3877 OpenSSL compilable with DJGPP under all reasonable conditions.
3878 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3880 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3881 to allow version independent disabling of normally unselected ciphers,
3882 which may be activated as a side-effect of selecting a single cipher.
3884 (E.g., cipher list string "RSA" enables ciphersuites that are left
3885 out of "ALL" because they do not provide symmetric encryption.
3886 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3887 [Lutz Jaenicke, Bodo Moeller]
3889 *) Add appropriate support for separate platform-dependent build
3890 directories. The recommended way to make a platform-dependent
3891 build directory is the following (tested on Linux), maybe with
3894 # Place yourself outside of the OpenSSL source tree. In
3895 # this example, the environment variable OPENSSL_SOURCE
3896 # is assumed to contain the absolute OpenSSL source directory.
3897 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3898 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3899 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3900 mkdir -p `dirname $F`
3901 ln -s $OPENSSL_SOURCE/$F $F
3904 To be absolutely sure not to disturb the source tree, a "make clean"
3905 is a good thing. If it isn't successfull, don't worry about it,
3906 it probably means the source directory is very clean.
3909 *) Make sure any ENGINE control commands make local copies of string
3910 pointers passed to them whenever necessary. Otherwise it is possible
3911 the caller may have overwritten (or deallocated) the original string
3912 data when a later ENGINE operation tries to use the stored values.
3913 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3915 *) Improve diagnostics in file reading and command-line digests.
3916 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3918 *) Add AES modes CFB and OFB to the object database. Correct an
3919 error in AES-CFB decryption.
3922 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3923 allows existing EVP_CIPHER_CTX structures to be reused after
3924 calling EVP_*Final(). This behaviour is used by encryption
3925 BIOs and some applications. This has the side effect that
3926 applications must explicitly clean up cipher contexts with
3927 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3930 *) Check the values of dna and dnb in bn_mul_recursive before calling
3931 bn_mul_comba (a non zero value means the a or b arrays do not contain
3932 n2 elements) and fallback to bn_mul_normal if either is not zero.
3935 *) Fix escaping of non-ASCII characters when using the -subj option
3936 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3939 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3940 form for "surname", serialNumber has no short form.
3941 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3942 therefore remove "mail" short name for "internet 7".
3943 The OID for unique identifiers in X509 certificates is
3944 x500UniqueIdentifier, not uniqueIdentifier.
3945 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3948 *) Add an "init" command to the ENGINE config module and auto initialize
3949 ENGINEs. Without any "init" command the ENGINE will be initialized
3950 after all ctrl commands have been executed on it. If init=1 the
3951 ENGINE is initailized at that point (ctrls before that point are run
3952 on the uninitialized ENGINE and after on the initialized one). If
3953 init=0 then the ENGINE will not be iniatialized at all.
3956 *) Fix the 'app_verify_callback' interface so that the user-defined
3957 argument is actually passed to the callback: In the
3958 SSL_CTX_set_cert_verify_callback() prototype, the callback
3959 declaration has been changed from
3962 int (*cb)(X509_STORE_CTX *,void *);
3963 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3964 i=s->ctx->app_verify_callback(&ctx)
3965 has been changed into
3966 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3968 To update applications using SSL_CTX_set_cert_verify_callback(),
3969 a dummy argument can be added to their callback functions.
3970 [D. K. Smetters <smetters@parc.xerox.com>]
3972 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3973 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3975 *) Add and OPENSSL_LOAD_CONF define which will cause
3976 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3977 This allows older applications to transparently support certain
3978 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3979 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3980 load the config file and OPENSSL_add_all_algorithms_conf() which will
3981 always load it have also been added.
3984 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3985 Adjust NIDs and EVP layer.
3986 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3988 *) Config modules support in openssl utility.
3990 Most commands now load modules from the config file,
3991 though in a few (such as version) this isn't done
3992 because it couldn't be used for anything.
3994 In the case of ca and req the config file used is
3995 the same as the utility itself: that is the -config
3996 command line option can be used to specify an
4000 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4001 use "openssl_conf" if filename is NULL use default openssl config file.
4004 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4005 config section name. Add a new flag to tolerate a missing config file
4006 and move code to CONF_modules_load_file().
4009 *) Support for crypto accelerator cards from Accelerated Encryption
4010 Processing, www.aep.ie. (Use engine 'aep')
4011 The support was copied from 0.9.6c [engine] and adapted/corrected
4012 to work with the new engine framework.
4013 [AEP Inc. and Richard Levitte]
4015 *) Support for SureWare crypto accelerator cards from Baltimore
4016 Technologies. (Use engine 'sureware')
4017 The support was copied from 0.9.6c [engine] and adapted
4018 to work with the new engine framework.
4021 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4022 make the newer ENGINE framework commands for the CHIL engine work.
4023 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4025 *) Make it possible to produce shared libraries on ReliantUNIX.
4026 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4028 *) Add the configuration target debug-linux-ppro.
4029 Make 'openssl rsa' use the general key loading routines
4030 implemented in apps.c, and make those routines able to
4031 handle the key format FORMAT_NETSCAPE and the variant
4033 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4035 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4036 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4038 *) Add -keyform to rsautl, and document -engine.
4039 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4041 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4042 BIO_R_NO_SUCH_FILE error code rather than the generic
4043 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4046 *) Add new functions
4048 ERR_peek_last_error_line
4049 ERR_peek_last_error_line_data.
4050 These are similar to
4053 ERR_peek_error_line_data,
4054 but report on the latest error recorded rather than the first one
4055 still in the error queue.
4056 [Ben Laurie, Bodo Moeller]
4058 *) default_algorithms option in ENGINE config module. This allows things
4060 default_algorithms = ALL
4061 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4064 *) Prelminary ENGINE config module.
4067 *) New experimental application configuration code.
4070 *) Change the AES code to follow the same name structure as all other
4071 symmetric ciphers, and behave the same way. Move everything to
4072 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4073 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4075 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4076 [Ben Laurie and Theo de Raadt]
4078 *) Add option to output public keys in req command.
4079 [Massimiliano Pala madwolf@openca.org]
4081 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4082 (up to about 10% better than before for P-192 and P-224).
4085 *) New functions/macros
4087 SSL_CTX_set_msg_callback(ctx, cb)
4088 SSL_CTX_set_msg_callback_arg(ctx, arg)
4089 SSL_set_msg_callback(ssl, cb)
4090 SSL_set_msg_callback_arg(ssl, arg)
4092 to request calling a callback function
4094 void cb(int write_p, int version, int content_type,
4095 const void *buf, size_t len, SSL *ssl, void *arg)
4097 whenever a protocol message has been completely received
4098 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4099 protocol version according to which the SSL library interprets
4100 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4101 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4102 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4103 specification (change_cipher_spec(20), alert(21), handshake(22)).
4104 'buf' and 'len' point to the actual message, 'ssl' to the
4105 SSL object, and 'arg' is the application-defined value set by
4106 SSL[_CTX]_set_msg_callback_arg().
4108 'openssl s_client' and 'openssl s_server' have new '-msg' options
4109 to enable a callback that displays all protocol messages.
4112 *) Change the shared library support so shared libraries are built as
4113 soon as the corresponding static library is finished, and thereby get
4114 openssl and the test programs linked against the shared library.
4115 This still only happens when the keyword "shard" has been given to
4116 the configuration scripts.
4118 NOTE: shared library support is still an experimental thing, and
4119 backward binary compatibility is still not guaranteed.
4120 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4122 *) Add support for Subject Information Access extension.
4123 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4125 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4126 additional bytes when new memory had to be allocated, not just
4127 when reusing an existing buffer.
4130 *) New command line and configuration option 'utf8' for the req command.
4131 This allows field values to be specified as UTF8 strings.
4134 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4135 runs for the former and machine-readable output for the latter.
4138 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4139 of the e-mail address in the DN (i.e., it will go into a certificate
4140 extension only). The new configuration file option 'email_in_dn = no'
4141 has the same effect.
4142 [Massimiliano Pala madwolf@openca.org]
4144 *) Change all functions with names starting with des_ to be starting
4145 with DES_ instead. Add wrappers that are compatible with libdes,
4146 but are named _ossl_old_des_*. Finally, add macros that map the
4147 des_* symbols to the corresponding _ossl_old_des_* if libdes
4148 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4149 desired, the des_* symbols will be mapped to DES_*, with one
4152 Since we provide two compatibility mappings, the user needs to
4153 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4154 compatibility is desired. The default (i.e., when that macro
4155 isn't defined) is OpenSSL 0.9.6c compatibility.
4157 There are also macros that enable and disable the support of old
4158 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4159 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4160 are defined, the default will apply: to support the old des routines.
4162 In either case, one must include openssl/des.h to get the correct
4163 definitions. Do not try to just include openssl/des_old.h, that
4166 NOTE: This is a major break of an old API into a new one. Software
4167 authors are encouraged to switch to the DES_ style functions. Some
4168 time in the future, des_old.h and the libdes compatibility functions
4169 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4170 default), and then completely removed.
4173 *) Test for certificates which contain unsupported critical extensions.
4174 If such a certificate is found during a verify operation it is
4175 rejected by default: this behaviour can be overridden by either
4176 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4177 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4178 X509_supported_extension() has also been added which returns 1 if a
4179 particular extension is supported.
4182 *) Modify the behaviour of EVP cipher functions in similar way to digests
4183 to retain compatibility with existing code.
4186 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4187 compatibility with existing code. In particular the 'ctx' parameter does
4188 not have to be to be initialized before the call to EVP_DigestInit() and
4189 it is tidied up after a call to EVP_DigestFinal(). New function
4190 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4191 EVP_MD_CTX_copy() changed to not require the destination to be
4192 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4193 requires the destination to be valid.
4195 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4196 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4199 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4200 so that complete 'Handshake' protocol structures are kept in memory
4201 instead of overwriting 'msg_type' and 'length' with 'body' data.
4204 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4205 [Massimo Santin via Richard Levitte]
4207 *) Major restructuring to the underlying ENGINE code. This includes
4208 reduction of linker bloat, separation of pure "ENGINE" manipulation
4209 (initialisation, etc) from functionality dealing with implementations
4210 of specific crypto iterfaces. This change also introduces integrated
4211 support for symmetric ciphers and digest implementations - so ENGINEs
4212 can now accelerate these by providing EVP_CIPHER and EVP_MD
4213 implementations of their own. This is detailed in crypto/engine/README
4214 as it couldn't be adequately described here. However, there are a few
4215 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4216 were changed in the original introduction of ENGINE code have now
4217 reverted back - the hooking from this code to ENGINE is now a good
4218 deal more passive and at run-time, operations deal directly with
4219 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4220 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4221 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4222 they were not being used by the framework as there is no concept of a
4223 BIGNUM_METHOD and they could not be generalised to the new
4224 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4225 ENGINE_cpy() has been removed as it cannot be consistently defined in
4229 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4232 *) Change mkdef.pl to sort symbols that get the same entry number,
4233 and make sure the automatically generated functions ERR_load_*
4234 become part of libeay.num as well.
4237 *) New function SSL_renegotiate_pending(). This returns true once
4238 renegotiation has been requested (either SSL_renegotiate() call
4239 or HelloRequest/ClientHello receveived from the peer) and becomes
4240 false once a handshake has been completed.
4241 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4242 sends a HelloRequest, but does not ensure that a handshake takes
4243 place. SSL_renegotiate_pending() is useful for checking if the
4244 client has followed the request.)
4247 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4248 By default, clients may request session resumption even during
4249 renegotiation (if session ID contexts permit); with this option,
4250 session resumption is possible only in the first handshake.
4252 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4253 more bits available for options that should not be part of
4254 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4257 *) Add some demos for certificate and certificate request creation.
4260 *) Make maximum certificate chain size accepted from the peer application
4261 settable (SSL*_get/set_max_cert_list()), as proposed by
4262 "Douglas E. Engert" <deengert@anl.gov>.
4265 *) Add support for shared libraries for Unixware-7
4266 (Boyd Lynn Gerber <gerberb@zenez.com>).
4269 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4270 be done prior to destruction. Use this to unload error strings from
4271 ENGINEs that load their own error strings. NB: This adds two new API
4272 functions to "get" and "set" this destroy handler in an ENGINE.
4275 *) Alter all existing ENGINE implementations (except "openssl" and
4276 "openbsd") to dynamically instantiate their own error strings. This
4277 makes them more flexible to be built both as statically-linked ENGINEs
4278 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4279 Also, add stub code to each that makes building them as self-contained
4280 shared-libraries easier (see README.ENGINE).
4283 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4284 implementations into applications that are completely implemented in
4285 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4286 commands that can be used to configure what shared-library to load and
4287 to control aspects of the way it is handled. Also, made an update to
4288 the README.ENGINE file that brings its information up-to-date and
4289 provides some information and instructions on the "dynamic" ENGINE
4290 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4293 *) Make it possible to unload ranges of ERR strings with a new
4294 "ERR_unload_strings" function.
4297 *) Add a copy() function to EVP_MD.
4300 *) Make EVP_MD routines take a context pointer instead of just the
4301 md_data void pointer.
4304 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4305 that the digest can only process a single chunk of data
4306 (typically because it is provided by a piece of
4307 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4308 is only going to provide a single chunk of data, and hence the
4309 framework needn't accumulate the data for oneshot drivers.
4312 *) As with "ERR", make it possible to replace the underlying "ex_data"
4313 functions. This change also alters the storage and management of global
4314 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4315 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4316 index counters. The API functions that use this state have been changed
4317 to take a "class_index" rather than pointers to the class's local STACK
4318 and counter, and there is now an API function to dynamically create new
4319 classes. This centralisation allows us to (a) plug a lot of the
4320 thread-safety problems that existed, and (b) makes it possible to clean
4321 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4322 such data would previously have always leaked in application code and
4323 workarounds were in place to make the memory debugging turn a blind eye
4324 to it. Application code that doesn't use this new function will still
4325 leak as before, but their memory debugging output will announce it now
4326 rather than letting it slide.
4328 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4329 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4330 has a return value to indicate success or failure.
4333 *) Make it possible to replace the underlying "ERR" functions such that the
4334 global state (2 LHASH tables and 2 locks) is only used by the "default"
4335 implementation. This change also adds two functions to "get" and "set"
4336 the implementation prior to it being automatically set the first time
4337 any other ERR function takes place. Ie. an application can call "get",
4338 pass the return value to a module it has just loaded, and that module
4339 can call its own "set" function using that value. This means the
4340 module's "ERR" operations will use (and modify) the error state in the
4341 application and not in its own statically linked copy of OpenSSL code.
4344 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4345 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4346 the operation, and provides a more encapsulated way for external code
4347 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4348 to use these functions rather than manually incrementing the counts.
4350 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4353 *) Add EVP test program.
4356 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4359 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4360 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4361 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4362 These allow a CRL to be built without having to access X509_CRL fields
4363 directly. Modify 'ca' application to use new functions.
4366 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4367 bug workarounds. Rollback attack detection is a security feature.
4368 The problem will only arise on OpenSSL servers when TLSv1 is not
4369 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4370 Software authors not wanting to support TLSv1 will have special reasons
4371 for their choice and can explicitly enable this option.
4372 [Bodo Moeller, Lutz Jaenicke]
4374 *) Rationalise EVP so it can be extended: don't include a union of
4375 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4376 (similar to those existing for EVP_CIPHER_CTX).
4381 EVP_MD_CTX_init(&md); /* new function call */
4382 EVP_DigestInit(&md, EVP_sha1());
4383 EVP_DigestUpdate(&md, in, len);
4384 EVP_DigestFinal(&md, out, NULL);
4385 EVP_MD_CTX_cleanup(&md); /* new function call */
4389 *) Make DES key schedule conform to the usual scheme, as well as
4390 correcting its structure. This means that calls to DES functions
4391 now have to pass a pointer to a des_key_schedule instead of a
4392 plain des_key_schedule (which was actually always a pointer
4395 des_key_schedule ks;
4397 des_set_key_checked(..., &ks);
4398 des_ncbc_encrypt(..., &ks, ...);
4400 (Note that a later change renames 'des_...' into 'DES_...'.)
4403 *) Initial reduction of linker bloat: the use of some functions, such as
4404 PEM causes large amounts of unused functions to be linked in due to
4405 poor organisation. For example pem_all.c contains every PEM function
4406 which has a knock on effect of linking in large amounts of (unused)
4407 ASN1 code. Grouping together similar functions and splitting unrelated
4408 functions prevents this.
4411 *) Cleanup of EVP macros.
4414 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4415 correct _ecb suffix.
4418 *) Add initial OCSP responder support to ocsp application. The
4419 revocation information is handled using the text based index
4420 use by the ca application. The responder can either handle
4421 requests generated internally, supplied in files (for example
4422 via a CGI script) or using an internal minimal server.
4425 *) Add configuration choices to get zlib compression for TLS.
4428 *) Changes to Kerberos SSL for RFC 2712 compliance:
4429 1. Implemented real KerberosWrapper, instead of just using
4430 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4431 2. Implemented optional authenticator field of KerberosWrapper.
4433 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4434 and authenticator structs; see crypto/krb5/.
4436 Generalized Kerberos calls to support multiple Kerberos libraries.
4437 [Vern Staats <staatsvr@asc.hpc.mil>,
4438 Jeffrey Altman <jaltman@columbia.edu>
4439 via Richard Levitte]
4441 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4442 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4443 values for each of the key sizes rather than having just
4444 parameters (and 'speed' generating keys each time).
4447 *) Speed up EVP routines.
4450 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4451 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4452 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4453 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4455 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4456 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4457 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4460 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4462 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4465 *) Added the OS2-EMX target.
4466 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4468 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4469 to support NCONF routines in extension code. New function CONF_set_nconf()
4470 to allow functions which take an NCONF to also handle the old LHASH
4471 structure: this means that the old CONF compatible routines can be
4472 retained (in particular wrt extensions) without having to duplicate the
4473 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4476 *) Enhance the general user interface with mechanisms for inner control
4477 and with possibilities to have yes/no kind of prompts.
4480 *) Change all calls to low level digest routines in the library and
4481 applications to use EVP. Add missing calls to HMAC_cleanup() and
4482 don't assume HMAC_CTX can be copied using memcpy().
4483 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4485 *) Add the possibility to control engines through control names but with
4486 arbitrary arguments instead of just a string.
4487 Change the key loaders to take a UI_METHOD instead of a callback
4488 function pointer. NOTE: this breaks binary compatibility with earlier
4489 versions of OpenSSL [engine].
4490 Adapt the nCipher code for these new conditions and add a card insertion
4494 *) Enhance the general user interface with mechanisms to better support
4495 dialog box interfaces, application-defined prompts, the possibility
4496 to use defaults (for example default passwords from somewhere else)
4497 and interrupts/cancellations.
4500 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4501 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4504 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4505 tidy up some unnecessarily weird code in 'sk_new()').
4506 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4508 *) Change the key loading routines for ENGINEs to use the same kind
4509 callback (pem_password_cb) as all other routines that need this
4513 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4514 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4515 than this minimum value is recommended.
4518 *) New random seeder for OpenVMS, using the system process statistics
4519 that are easily reachable.
4522 *) Windows apparently can't transparently handle global
4523 variables defined in DLLs. Initialisations such as:
4525 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4527 wont compile. This is used by the any applications that need to
4528 declare their own ASN1 modules. This was fixed by adding the option
4529 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4530 needed for static libraries under Win32.
4533 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4534 setting of purpose and trust fields. New X509_STORE trust and
4535 purpose functions and tidy up setting in other SSL functions.
4538 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4539 structure. These are inherited by X509_STORE_CTX when it is
4540 initialised. This allows various defaults to be set in the
4541 X509_STORE structure (such as flags for CRL checking and custom
4542 purpose or trust settings) for functions which only use X509_STORE_CTX
4543 internally such as S/MIME.
4545 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4546 trust settings if they are not set in X509_STORE. This allows X509_STORE
4547 purposes and trust (in S/MIME for example) to override any set by default.
4549 Add command line options for CRL checking to smime, s_client and s_server
4553 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4554 are set then the CRL is looked up in the X509_STORE structure and
4555 its validity and signature checked, then if the certificate is found
4556 in the CRL the verify fails with a revoked error.
4558 Various new CRL related callbacks added to X509_STORE_CTX structure.
4560 Command line options added to 'verify' application to support this.
4562 This needs some additional work, such as being able to handle multiple
4563 CRLs with different times, extension based lookup (rather than just
4564 by subject name) and ultimately more complete V2 CRL extension
4568 *) Add a general user interface API (crypto/ui/). This is designed
4569 to replace things like des_read_password and friends (backward
4570 compatibility functions using this new API are provided).
4571 The purpose is to remove prompting functions from the DES code
4572 section as well as provide for prompting through dialog boxes in
4573 a window system and the like.
4576 *) Add "ex_data" support to ENGINE so implementations can add state at a
4577 per-structure level rather than having to store it globally.
4580 *) Make it possible for ENGINE structures to be copied when retrieved by
4581 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4582 This causes the "original" ENGINE structure to act like a template,
4583 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4584 operational state can be localised to each ENGINE structure, despite the
4585 fact they all share the same "methods". New ENGINE structures returned in
4586 this case have no functional references and the return value is the single
4587 structural reference. This matches the single structural reference returned
4588 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4592 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4593 needs to match any other type at all we need to manually clear the
4597 *) Changes to the "openssl engine" utility to include;
4598 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4599 about an ENGINE's available control commands.
4600 - executing control commands from command line arguments using the
4601 '-pre' and '-post' switches. '-post' is only used if '-t' is
4602 specified and the ENGINE is successfully initialised. The syntax for
4603 the individual commands are colon-separated, for example;
4604 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4607 *) New dynamic control command support for ENGINEs. ENGINEs can now
4608 declare their own commands (numbers), names (strings), descriptions,
4609 and input types for run-time discovery by calling applications. A
4610 subset of these commands are implicitly classed as "executable"
4611 depending on their input type, and only these can be invoked through
4612 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4613 can be based on user input, config files, etc). The distinction is
4614 that "executable" commands cannot return anything other than a boolean
4615 result and can only support numeric or string input, whereas some
4616 discoverable commands may only be for direct use through
4617 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4618 pointers, or other custom uses. The "executable" commands are to
4619 support parameterisations of ENGINE behaviour that can be
4620 unambiguously defined by ENGINEs and used consistently across any
4621 OpenSSL-based application. Commands have been added to all the
4622 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4623 control over shared-library paths without source code alterations.
4626 *) Changed all ENGINE implementations to dynamically allocate their
4627 ENGINEs rather than declaring them statically. Apart from this being
4628 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4629 this also allows the implementations to compile without using the
4630 internal engine_int.h header.
4633 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4634 'const' value. Any code that should be able to modify a RAND_METHOD
4635 should already have non-const pointers to it (ie. they should only
4636 modify their own ones).
4639 *) Made a variety of little tweaks to the ENGINE code.
4640 - "atalla" and "ubsec" string definitions were moved from header files
4641 to C code. "nuron" string definitions were placed in variables
4642 rather than hard-coded - allowing parameterisation of these values
4643 later on via ctrl() commands.
4644 - Removed unused "#if 0"'d code.
4645 - Fixed engine list iteration code so it uses ENGINE_free() to release
4646 structural references.
4647 - Constified the RAND_METHOD element of ENGINE structures.
4648 - Constified various get/set functions as appropriate and added
4649 missing functions (including a catch-all ENGINE_cpy that duplicates
4650 all ENGINE values onto a new ENGINE except reference counts/state).
4651 - Removed NULL parameter checks in get/set functions. Setting a method
4652 or function to NULL is a way of cancelling out a previously set
4653 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4654 and doesn't justify the extra error symbols and code.
4655 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4656 flags from engine_int.h to engine.h.
4657 - Changed prototypes for ENGINE handler functions (init(), finish(),
4658 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4661 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4662 to the algorithm using long division. The binary algorithm can be
4663 used only if the modulus is odd. On 32-bit systems, it is faster
4664 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4665 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4666 up to 450 bits. In 64-bit environments, the binary algorithm
4667 appears to be advantageous for much longer moduli; here we use it
4668 for moduli up to 2048 bits.
4671 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4672 could not support the combine flag in choice fields.
4675 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4676 extensions from a certificate request to the certificate.
4679 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4680 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4681 file: this allows the display of the certificate about to be
4682 signed to be customised, to allow certain fields to be included
4683 or excluded and extension details. The old system didn't display
4684 multicharacter strings properly, omitted fields not in the policy
4685 and couldn't display additional details such as extensions.
4688 *) Function EC_POINTs_mul for multiple scalar multiplication
4689 of an arbitrary number of elliptic curve points
4690 \sum scalars[i]*points[i],
4691 optionally including the generator defined for the EC_GROUP:
4692 scalar*generator + \sum scalars[i]*points[i].
4694 EC_POINT_mul is a simple wrapper function for the typical case
4695 that the point list has just one item (besides the optional
4699 *) First EC_METHODs for curves over GF(p):
4701 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4702 operations and provides various method functions that can also
4703 operate with faster implementations of modular arithmetic.
4705 EC_GFp_mont_method() reuses most functions that are part of
4706 EC_GFp_simple_method, but uses Montgomery arithmetic.
4708 [Bodo Moeller; point addition and point doubling
4709 implementation directly derived from source code provided by
4710 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4712 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4713 crypto/ec/ec_lib.c):
4715 Curves are EC_GROUP objects (with an optional group generator)
4716 based on EC_METHODs that are built into the library.
4718 Points are EC_POINT objects based on EC_GROUP objects.
4720 Most of the framework would be able to handle curves over arbitrary
4721 finite fields, but as there are no obvious types for fields other
4722 than GF(p), some functions are limited to that for now.
4725 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4726 that the file contains a complete HTTP response.
4729 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4730 change the def and num file printf format specifier from "%-40sXXX"
4731 to "%-39s XXX". The latter will always guarantee a space after the
4732 field while the former will cause them to run together if the field
4733 is 40 of more characters long.
4736 *) Constify the cipher and digest 'method' functions and structures
4737 and modify related functions to take constant EVP_MD and EVP_CIPHER
4741 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4742 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4745 *) Modify EVP_Digest*() routines so they now return values. Although the
4746 internal software routines can never fail additional hardware versions
4750 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4752 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4753 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4756 ERR_R_NESTED_ASN1_ERROR
4758 ERR_R_MISSING_ASN1_EOS
4759 were 4 .. 9, conflicting with
4760 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4762 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4763 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4765 Add new error code 'ERR_R_INTERNAL_ERROR'.
4768 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4772 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4773 sets the subject name for a new request or supersedes the
4774 subject name in a given request. Formats that can be parsed are
4775 'CN=Some Name, OU=myOU, C=IT'
4777 'CN=Some Name/OU=myOU/C=IT'.
4779 Add options '-batch' and '-verbose' to 'openssl req'.
4780 [Massimiliano Pala <madwolf@hackmasters.net>]
4782 *) Introduce the possibility to access global variables through
4783 functions on platform were that's the best way to handle exporting
4784 global variables in shared libraries. To enable this functionality,
4785 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4786 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4787 is normally done by Configure or something similar).
4789 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4790 in the source file (foo.c) like this:
4792 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4793 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4795 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4796 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4798 OPENSSL_DECLARE_GLOBAL(int,foo);
4799 #define foo OPENSSL_GLOBAL_REF(foo)
4800 OPENSSL_DECLARE_GLOBAL(double,bar);
4801 #define bar OPENSSL_GLOBAL_REF(bar)
4803 The #defines are very important, and therefore so is including the
4804 header file everywhere where the defined globals are used.
4806 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4807 of ASN.1 items, but that structure is a bit different.
4809 The largest change is in util/mkdef.pl which has been enhanced with
4810 better and easier to understand logic to choose which symbols should
4811 go into the Windows .def files as well as a number of fixes and code
4812 cleanup (among others, algorithm keywords are now sorted
4813 lexicographically to avoid constant rewrites).
4816 *) In BN_div() keep a copy of the sign of 'num' before writing the
4817 result to 'rm' because if rm==num the value will be overwritten
4818 and produce the wrong result if 'num' is negative: this caused
4819 problems with BN_mod() and BN_nnmod().
4822 *) Function OCSP_request_verify(). This checks the signature on an
4823 OCSP request and verifies the signer certificate. The signer
4824 certificate is just checked for a generic purpose and OCSP request
4828 *) Add OCSP_check_validity() function to check the validity of OCSP
4829 responses. OCSP responses are prepared in real time and may only
4830 be a few seconds old. Simply checking that the current time lies
4831 between thisUpdate and nextUpdate max reject otherwise valid responses
4832 caused by either OCSP responder or client clock inaccuracy. Instead
4833 we allow thisUpdate and nextUpdate to fall within a certain period of
4834 the current time. The age of the response can also optionally be
4835 checked. Two new options -validity_period and -status_age added to
4839 *) If signature or public key algorithm is unrecognized print out its
4840 OID rather that just UNKNOWN.
4843 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4844 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4845 ID to be generated from the issuer certificate alone which can then be
4846 passed to OCSP_id_issuer_cmp().
4849 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4850 ASN1 modules to export functions returning ASN1_ITEM pointers
4851 instead of the ASN1_ITEM structures themselves. This adds several
4852 new macros which allow the underlying ASN1 function/structure to
4853 be accessed transparently. As a result code should not use ASN1_ITEM
4854 references directly (such as &X509_it) but instead use the relevant
4855 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4856 use of the new ASN1 code on platforms where exporting structures
4857 is problematical (for example in shared libraries) but exporting
4858 functions returning pointers to structures is not.
4861 *) Add support for overriding the generation of SSL/TLS session IDs.
4862 These callbacks can be registered either in an SSL_CTX or per SSL.
4863 The purpose of this is to allow applications to control, if they wish,
4864 the arbitrary values chosen for use as session IDs, particularly as it
4865 can be useful for session caching in multiple-server environments. A
4866 command-line switch for testing this (and any client code that wishes
4867 to use such a feature) has been added to "s_server".
4868 [Geoff Thorpe, Lutz Jaenicke]
4870 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4871 of the form '#if defined(...) || defined(...) || ...' and
4872 '#if !defined(...) && !defined(...) && ...'. This also avoids
4873 the growing number of special cases it was previously handling.
4876 *) Make all configuration macros available for application by making
4877 sure they are available in opensslconf.h, by giving them names starting
4878 with "OPENSSL_" to avoid conflicts with other packages and by making
4879 sure e_os2.h will cover all platform-specific cases together with
4881 Additionally, it is now possible to define configuration/platform-
4882 specific names (called "system identities"). In the C code, these
4883 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4884 macro with the name beginning with "OPENSSL_SYS_", which is determined
4885 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4889 *) New option -set_serial to 'req' and 'x509' this allows the serial
4890 number to use to be specified on the command line. Previously self
4891 signed certificates were hard coded with serial number 0 and the
4892 CA options of 'x509' had to use a serial number in a file which was
4896 *) New options to 'ca' utility to support V2 CRL entry extensions.
4897 Currently CRL reason, invalidity date and hold instruction are
4898 supported. Add new CRL extensions to V3 code and some new objects.
4901 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4902 disable standard block padding (aka PKCS#5 padding) in the EVP
4903 API, which was previously mandatory. This means that the data is
4904 not padded in any way and so the total length much be a multiple
4905 of the block size, otherwise an error occurs.
4908 *) Initial (incomplete) OCSP SSL support.
4911 *) New function OCSP_parse_url(). This splits up a URL into its host,
4912 port and path components: primarily to parse OCSP URLs. New -url
4913 option to ocsp utility.
4916 *) New nonce behavior. The return value of OCSP_check_nonce() now
4917 reflects the various checks performed. Applications can decide
4918 whether to tolerate certain situations such as an absent nonce
4919 in a response when one was present in a request: the ocsp application
4920 just prints out a warning. New function OCSP_add1_basic_nonce()
4921 this is to allow responders to include a nonce in a response even if
4922 the request is nonce-less.
4925 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4926 skipped when using openssl x509 multiple times on a single input file,
4927 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4930 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4931 set string type: to handle setting ASN1_TIME structures. Fix ca
4932 utility to correctly initialize revocation date of CRLs.
4935 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4936 the clients preferred ciphersuites and rather use its own preferences.
4937 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4938 Internet Explorer by ensuring unchanged hash method during stepup.
4939 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4942 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4943 to aes and add a new 'exist' option to print out symbols that don't
4947 *) Additional options to ocsp utility to allow flags to be set and
4948 additional certificates supplied.
4951 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4952 OCSP client a number of certificate to only verify the response
4956 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4957 handle the new API. Currently only ECB, CBC modes supported. Add new
4960 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4961 Encryption Standard (AES) Ciphersuites for Transport Layer
4962 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4963 not enabled by default and were not part of the "ALL" ciphersuite
4964 alias because they were not yet official; they could be
4965 explicitly requested by specifying the "AESdraft" ciphersuite
4966 group alias. In the final release of OpenSSL 0.9.7, the group
4967 alias is called "AES" and is part of "ALL".)
4968 [Ben Laurie, Steve Henson, Bodo Moeller]
4970 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4971 request to response.
4974 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4975 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4976 extract information from a certificate request. OCSP_response_create()
4977 creates a response and optionally adds a basic response structure.
4978 OCSP_basic_add1_status() adds a complete single response to a basic
4979 response and returns the OCSP_SINGLERESP structure just added (to allow
4980 extensions to be included for example). OCSP_basic_add1_cert() adds a
4981 certificate to a basic response and OCSP_basic_sign() signs a basic
4982 response with various flags. New helper functions ASN1_TIME_check()
4983 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4984 (converts ASN1_TIME to GeneralizedTime).
4987 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4988 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4989 structure from a certificate. X509_pubkey_digest() digests the public_key
4990 contents: this is used in various key identifiers.
4993 *) Make sk_sort() tolerate a NULL argument.
4994 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4996 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4997 passed by the function are trusted implicitly. If any of them signed the
4998 response then it is assumed to be valid and is not verified.
5001 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5002 to data. This was previously part of the PKCS7 ASN1 code. This
5003 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5004 [Steve Henson, reported by Kenneth R. Robinette
5005 <support@securenetterm.com>]
5007 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5008 routines: without these tracing memory leaks is very painful.
5009 Fix leaks in PKCS12 and PKCS7 routines.
5012 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5013 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5014 effectively meant GeneralizedTime would never be used. Now it
5015 is initialised to -1 but X509_time_adj() now has to check the value
5016 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5017 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5018 [Steve Henson, reported by Kenneth R. Robinette
5019 <support@securenetterm.com>]
5021 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5022 result in a zero length in the ASN1_INTEGER structure which was
5023 not consistent with the structure when d2i_ASN1_INTEGER() was used
5024 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5025 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5026 where it did not print out a minus for negative ASN1_INTEGER.
5029 *) Add summary printout to ocsp utility. The various functions which
5030 convert status values to strings have been renamed to:
5031 OCSP_response_status_str(), OCSP_cert_status_str() and
5032 OCSP_crl_reason_str() and are no longer static. New options
5033 to verify nonce values and to disable verification. OCSP response
5034 printout format cleaned up.
5037 *) Add additional OCSP certificate checks. These are those specified
5038 in RFC2560. This consists of two separate checks: the CA of the
5039 certificate being checked must either be the OCSP signer certificate
5040 or the issuer of the OCSP signer certificate. In the latter case the
5041 OCSP signer certificate must contain the OCSP signing extended key
5042 usage. This check is performed by attempting to match the OCSP
5043 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5044 in the OCSP_CERTID structures of the response.
5047 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5048 and related routines. This uses the standard OpenSSL certificate
5049 verify routines to perform initial checks (just CA validity) and
5050 to obtain the certificate chain. Then additional checks will be
5051 performed on the chain. Currently the root CA is checked to see
5052 if it is explicitly trusted for OCSP signing. This is used to set
5053 a root CA as a global signing root: that is any certificate that
5054 chains to that CA is an acceptable OCSP signing certificate.
5057 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5058 extensions from a separate configuration file.
5059 As when reading extensions from the main configuration file,
5060 the '-extensions ...' option may be used for specifying the
5062 [Massimiliano Pala <madwolf@comune.modena.it>]
5064 *) New OCSP utility. Allows OCSP requests to be generated or
5065 read. The request can be sent to a responder and the output
5066 parsed, outputed or printed in text form. Not complete yet:
5067 still needs to check the OCSP response validity.
5070 *) New subcommands for 'openssl ca':
5071 'openssl ca -status <serial>' prints the status of the cert with
5072 the given serial number (according to the index file).
5073 'openssl ca -updatedb' updates the expiry status of certificates
5075 [Massimiliano Pala <madwolf@comune.modena.it>]
5077 *) New '-newreq-nodes' command option to CA.pl. This is like
5078 '-newreq', but calls 'openssl req' with the '-nodes' option
5079 so that the resulting key is not encrypted.
5080 [Damien Miller <djm@mindrot.org>]
5082 *) New configuration for the GNU Hurd.
5083 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5085 *) Initial code to implement OCSP basic response verify. This
5086 is currently incomplete. Currently just finds the signer's
5087 certificate and verifies the signature on the response.
5090 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5091 value of OPENSSLDIR. This is available via the new '-d' option
5092 to 'openssl version', and is also included in 'openssl version -a'.
5095 *) Allowing defining memory allocation callbacks that will be given
5096 file name and line number information in additional arguments
5097 (a const char* and an int). The basic functionality remains, as
5098 well as the original possibility to just replace malloc(),
5099 realloc() and free() by functions that do not know about these
5100 additional arguments. To register and find out the current
5101 settings for extended allocation functions, the following
5102 functions are provided:
5104 CRYPTO_set_mem_ex_functions
5105 CRYPTO_set_locked_mem_ex_functions
5106 CRYPTO_get_mem_ex_functions
5107 CRYPTO_get_locked_mem_ex_functions
5109 These work the same way as CRYPTO_set_mem_functions and friends.
5110 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5111 extended allocation function is enabled.
5112 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5113 a conventional allocation function is enabled.
5114 [Richard Levitte, Bodo Moeller]
5116 *) Finish off removing the remaining LHASH function pointer casts.
5117 There should no longer be any prototype-casting required when using
5118 the LHASH abstraction, and any casts that remain are "bugs". See
5119 the callback types and macros at the head of lhash.h for details
5120 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5123 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5124 If /dev/[u]random devices are not available or do not return enough
5125 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5127 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5128 /etc/entropy will be queried once each in this sequence, quering stops
5129 when enough entropy was collected without querying more sockets.
5132 *) Change the Unix RAND_poll() variant to be able to poll several
5133 random devices, as specified by DEVRANDOM, until a sufficient amount
5134 of data has been collected. We spend at most 10 ms on each file
5135 (select timeout) and read in non-blocking mode. DEVRANDOM now
5136 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5137 (previously it was just the string "/dev/urandom"), so on typical
5138 platforms the 10 ms delay will never occur.
5139 Also separate out the Unix variant to its own file, rand_unix.c.
5140 For VMS, there's a currently-empty rand_vms.c.
5143 *) Move OCSP client related routines to ocsp_cl.c. These
5144 provide utility functions which an application needing
5145 to issue a request to an OCSP responder and analyse the
5146 response will typically need: as opposed to those which an
5147 OCSP responder itself would need which will be added later.
5149 OCSP_request_sign() signs an OCSP request with an API similar
5150 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5151 response. OCSP_response_get1_basic() extracts basic response
5152 from response. OCSP_resp_find_status(): finds and extracts status
5153 information from an OCSP_CERTID structure (which will be created
5154 when the request structure is built). These are built from lower
5155 level functions which work on OCSP_SINGLERESP structures but
5156 wont normally be used unless the application wishes to examine
5157 extensions in the OCSP response for example.
5159 Replace nonce routines with a pair of functions.
5160 OCSP_request_add1_nonce() adds a nonce value and optionally
5161 generates a random value. OCSP_check_nonce() checks the
5162 validity of the nonce in an OCSP response.
5165 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5166 This doesn't copy the supplied OCSP_CERTID and avoids the
5167 need to free up the newly created id. Change return type
5168 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5169 This can then be used to add extensions to the request.
5170 Deleted OCSP_request_new(), since most of its functionality
5171 is now in OCSP_REQUEST_new() (and the case insensitive name
5172 clash) apart from the ability to set the request name which
5173 will be added elsewhere.
5176 *) Update OCSP API. Remove obsolete extensions argument from
5177 various functions. Extensions are now handled using the new
5178 OCSP extension code. New simple OCSP HTTP function which
5179 can be used to send requests and parse the response.
5182 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5183 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5184 uses the special reorder version of SET OF to sort the attributes
5185 and reorder them to match the encoded order. This resolves a long
5186 standing problem: a verify on a PKCS7 structure just after signing
5187 it used to fail because the attribute order did not match the
5188 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5189 it uses the received order. This is necessary to tolerate some broken
5190 software that does not order SET OF. This is handled by encoding
5191 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5192 to produce the required SET OF.
5195 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5196 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5197 files to get correct declarations of the ASN.1 item variables.
5200 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5201 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5202 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5203 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5204 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5205 ASN1_ITEM and no wrapper functions.
5208 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5209 replace the old function pointer based I/O routines. Change most of
5210 the *_d2i_bio() and *_d2i_fp() functions to use these.
5213 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5214 lines, recognice more "algorithms" that can be deselected, and make
5215 it complain about algorithm deselection that isn't recognised.
5218 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5219 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5220 to use new functions. Add NO_ASN1_OLD which can be set to remove
5221 some old style ASN1 functions: this can be used to determine if old
5222 code will still work when these eventually go away.
5225 *) New extension functions for OCSP structures, these follow the
5226 same conventions as certificates and CRLs.
5229 *) New function X509V3_add1_i2d(). This automatically encodes and
5230 adds an extension. Its behaviour can be customised with various
5231 flags to append, replace or delete. Various wrappers added for
5232 certifcates and CRLs.
5235 *) Fix to avoid calling the underlying ASN1 print routine when
5236 an extension cannot be parsed. Correct a typo in the
5237 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5240 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5241 entries for variables.
5244 *) Add functionality to apps/openssl.c for detecting locking
5245 problems: As the program is single-threaded, all we have
5246 to do is register a locking callback using an array for
5247 storing which locks are currently held by the program.
5250 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5251 SSL_get_ex_data_X509_STORE_idx(), which is used in
5252 ssl_verify_cert_chain() and thus can be called at any time
5253 during TLS/SSL handshakes so that thread-safety is essential.
5254 Unfortunately, the ex_data design is not at all suited
5255 for multi-threaded use, so it probably should be abolished.
5258 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5259 [Broadcom, tweaked and integrated by Geoff Thorpe]
5261 *) Move common extension printing code to new function
5262 X509V3_print_extensions(). Reorganise OCSP print routines and
5263 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5266 *) New function X509_signature_print() to remove duplication in some
5270 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5271 set (this was treated exactly the same as SET OF previously). This
5272 is used to reorder the STACK representing the structure to match the
5273 encoding. This will be used to get round a problem where a PKCS7
5274 structure which was signed could not be verified because the STACK
5275 order did not reflect the encoded order.
5278 *) Reimplement the OCSP ASN1 module using the new code.
5281 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5282 for its ASN1 operations. The old style function pointers still exist
5283 for now but they will eventually go away.
5286 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5287 completely replaces the old ASN1 functionality with a table driven
5288 encoder and decoder which interprets an ASN1_ITEM structure describing
5289 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5290 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5291 has also been converted to the new form.
5294 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5295 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5296 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5297 for negative moduli.
5300 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5301 of not touching the result's sign bit.
5304 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5308 *) Changed the LHASH code to use prototypes for callbacks, and created
5309 macros to declare and implement thin (optionally static) functions
5310 that provide type-safety and avoid function pointer casting for the
5311 type-specific callbacks.
5314 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5316 [Veers Staats <staatsvr@asc.hpc.mil>,
5317 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5319 *) Reformat the FAQ so the different questions and answers can be divided
5320 in sections depending on the subject.
5323 *) Have the zlib compression code load ZLIB.DLL dynamically under
5327 *) New function BN_mod_sqrt for computing square roots modulo a prime
5328 (using the probabilistic Tonelli-Shanks algorithm unless
5329 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5330 be handled deterministically).
5331 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5333 *) Make BN_mod_inverse faster by explicitly handling small quotients
5334 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5335 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5338 *) New function BN_kronecker.
5341 *) Fix BN_gcd so that it works on negative inputs; the result is
5342 positive unless both parameters are zero.
5343 Previously something reasonably close to an infinite loop was
5344 possible because numbers could be growing instead of shrinking
5345 in the implementation of Euclid's algorithm.
5348 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5349 sign of the number in question.
5351 Fix BN_is_word(a,w) to work correctly for w == 0.
5353 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5354 because its test if the absolute value of 'a' equals 'w'.
5355 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5356 it exists mostly for use in the implementations of BN_is_zero(),
5357 BN_is_one(), and BN_is_word().
5360 *) New function BN_swap.
5363 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5364 the exponentiation functions are more likely to produce reasonable
5365 results on negative inputs.
5368 *) Change BN_mod_mul so that the result is always non-negative.
5369 Previously, it could be negative if one of the factors was negative;
5370 I don't think anyone really wanted that behaviour.
5373 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5374 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5375 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5376 and add new functions:
5385 BN_mod_lshift1_quick
5389 These functions always generate non-negative results.
5391 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5392 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5394 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5395 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5396 be reduced modulo m.
5397 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5400 The following entry accidentily appeared in the CHANGES file
5401 distributed with OpenSSL 0.9.7. The modifications described in
5402 it do *not* apply to OpenSSL 0.9.7.
5404 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5405 was actually never needed) and in BN_mul(). The removal in BN_mul()
5406 required a small change in bn_mul_part_recursive() and the addition
5407 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5408 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5409 bn_sub_words() and bn_add_words() except they take arrays with
5414 *) In 'openssl passwd', verify passwords read from the terminal
5415 unless the '-salt' option is used (which usually means that
5416 verification would just waste user's time since the resulting
5417 hash is going to be compared with some given password hash)
5418 or the new '-noverify' option is used.
5420 This is an incompatible change, but it does not affect
5421 non-interactive use of 'openssl passwd' (passwords on the command
5422 line, '-stdin' option, '-in ...' option) and thus should not
5426 *) Remove all references to RSAref, since there's no more need for it.
5429 *) Make DSO load along a path given through an environment variable
5430 (SHLIB_PATH) with shl_load().
5433 *) Constify the ENGINE code as a result of BIGNUM constification.
5434 Also constify the RSA code and most things related to it. In a
5435 few places, most notable in the depth of the ASN.1 code, ugly
5436 casts back to non-const were required (to be solved at a later
5440 *) Make it so the openssl application has all engines loaded by default.
5443 *) Constify the BIGNUM routines a little more.
5446 *) Add the following functions:
5448 ENGINE_load_cswift()
5450 ENGINE_load_atalla()
5452 ENGINE_load_builtin_engines()
5454 That way, an application can itself choose if external engines that
5455 are built-in in OpenSSL shall ever be used or not. The benefit is
5456 that applications won't have to be linked with libdl or other dso
5457 libraries unless it's really needed.
5459 Changed 'openssl engine' to load all engines on demand.
5460 Changed the engine header files to avoid the duplication of some
5461 declarations (they differed!).
5464 *) 'openssl engine' can now list capabilities.
5467 *) Better error reporting in 'openssl engine'.
5470 *) Never call load_dh_param(NULL) in s_server.
5473 *) Add engine application. It can currently list engines by name and
5474 identity, and test if they are actually available.
5477 *) Improve RPM specification file by forcing symbolic linking and making
5478 sure the installed documentation is also owned by root.root.
5479 [Damien Miller <djm@mindrot.org>]
5481 *) Give the OpenSSL applications more possibilities to make use of
5482 keys (public as well as private) handled by engines.
5485 *) Add OCSP code that comes from CertCo.
5488 *) Add VMS support for the Rijndael code.
5491 *) Added untested support for Nuron crypto accelerator.
5494 *) Add support for external cryptographic devices. This code was
5495 previously distributed separately as the "engine" branch.
5496 [Geoff Thorpe, Richard Levitte]
5498 *) Rework the filename-translation in the DSO code. It is now possible to
5499 have far greater control over how a "name" is turned into a filename
5500 depending on the operating environment and any oddities about the
5501 different shared library filenames on each system.
5504 *) Support threads on FreeBSD-elf in Configure.
5507 *) Fix for SHA1 assembly problem with MASM: it produces
5508 warnings about corrupt line number information when assembling
5509 with debugging information. This is caused by the overlapping
5511 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5514 NCONF_get_number() has no error checking at all. As a replacement,
5515 NCONF_get_number_e() is defined (_e for "error checking") and is
5516 promoted strongly. The old NCONF_get_number is kept around for
5517 binary backward compatibility.
5518 Make it possible for methods to load from something other than a BIO,
5519 by providing a function pointer that is given a name instead of a BIO.
5520 For example, this could be used to load configuration data from an
5524 *) Fix for non blocking accept BIOs. Added new I/O special reason
5525 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5526 with non blocking I/O was not possible because no retry code was
5527 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5531 *) Added the beginnings of Rijndael support.
5534 *) Fix for bug in DirectoryString mask setting. Add support for
5535 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5536 to allow certificate printing to more controllable, additional
5537 'certopt' option to 'x509' to allow new printing options to be
5541 *) Clean old EAY MD5 hack from e_os.h.
5544 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5546 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5547 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5548 [Joe Orton, Steve Henson]
5550 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5552 *) Fix additional bug revealed by the NISCC test suite:
5554 Stop bug triggering large recursion when presented with
5555 certain ASN.1 tags (CVE-2003-0851)
5558 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5560 *) Fix various bugs revealed by running the NISCC test suite:
5562 Stop out of bounds reads in the ASN1 code when presented with
5563 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5565 If verify callback ignores invalid public key errors don't try to check
5566 certificate signature with the NULL public key.
5570 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5571 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5575 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5576 extra data after the compression methods not only for TLS 1.0
5577 but also for SSL 3.0 (as required by the specification).
5578 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5580 *) Change X509_certificate_type() to mark the key as exported/exportable
5581 when it's 512 *bits* long, not 512 bytes.
5584 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5586 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5587 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5588 a protocol version number mismatch like a decryption error
5589 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5592 *) Turn on RSA blinding by default in the default implementation
5593 to avoid a timing attack. Applications that don't want it can call
5594 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5595 They would be ill-advised to do so in most cases.
5596 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5598 *) Change RSA blinding code so that it works when the PRNG is not
5599 seeded (in this case, the secret RSA exponent is abused as
5600 an unpredictable seed -- if it is not unpredictable, there
5601 is no point in blinding anyway). Make RSA blinding thread-safe
5602 by remembering the creator's thread ID in rsa->blinding and
5603 having all other threads use local one-time blinding factors
5604 (this requires more computation than sharing rsa->blinding, but
5605 avoids excessive locking; and if an RSA object is not shared
5606 between threads, blinding will still be very fast).
5609 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5611 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5612 via timing by performing a MAC computation even if incorrrect
5613 block cipher padding has been found. This is a countermeasure
5614 against active attacks where the attacker has to distinguish
5615 between bad padding and a MAC verification error. (CVE-2003-0078)
5617 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5618 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5619 Martin Vuagnoux (EPFL, Ilion)]
5621 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5623 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5624 memory from it's contents. This is done with a counter that will
5625 place alternating values in each byte. This can be used to solve
5626 two issues: 1) the removal of calls to memset() by highly optimizing
5627 compilers, and 2) cleansing with other values than 0, since those can
5628 be read through on certain media, for example a swap space on disk.
5631 *) Bugfix: client side session caching did not work with external caching,
5632 because the session->cipher setting was not restored when reloading
5633 from the external cache. This problem was masked, when
5634 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5635 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5638 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5639 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5640 [Zeev Lieber <zeev-l@yahoo.com>]
5642 *) Undo an undocumented change introduced in 0.9.6e which caused
5643 repeated calls to OpenSSL_add_all_ciphers() and
5644 OpenSSL_add_all_digests() to be ignored, even after calling
5648 *) Change the default configuration reader to deal with last line not
5649 being properly terminated.
5652 *) Change X509_NAME_cmp() so it applies the special rules on handling
5653 DN values that are of type PrintableString, as well as RDNs of type
5654 emailAddress where the value has the type ia5String.
5655 [stefank@valicert.com via Richard Levitte]
5657 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5658 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5659 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5660 the bitwise-OR of the two for use by the majority of applications
5661 wanting this behaviour, and update the docs. The documented
5662 behaviour and actual behaviour were inconsistent and had been
5663 changing anyway, so this is more a bug-fix than a behavioural
5665 [Geoff Thorpe, diagnosed by Nadav Har'El]
5667 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5668 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5671 *) Fix initialization code race conditions in
5672 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5673 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5674 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5675 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5676 ssl2_get_cipher_by_char(),
5677 ssl3_get_cipher_by_char().
5678 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5680 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5681 the cached sessions are flushed, as the remove_cb() might use ex_data
5682 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5683 (see [openssl.org #212]).
5684 [Geoff Thorpe, Lutz Jaenicke]
5686 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5687 length, instead of the encoding length to d2i_ASN1_OBJECT.
5690 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5692 *) [In 0.9.6g-engine release:]
5693 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5694 [Lynn Gazis <lgazis@rainbow.com>]
5696 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5698 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5699 and get fix the header length calculation.
5700 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5701 Alon Kantor <alonk@checkpoint.com> (and others),
5704 *) Use proper error handling instead of 'assertions' in buffer
5705 overflow checks added in 0.9.6e. This prevents DoS (the
5706 assertions could call abort()).
5707 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5709 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5711 *) Add various sanity checks to asn1_get_length() to reject
5712 the ASN1 length bytes if they exceed sizeof(long), will appear
5713 negative or the content length exceeds the length of the
5715 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5717 *) Fix cipher selection routines: ciphers without encryption had no flags
5718 for the cipher strength set and where therefore not handled correctly
5719 by the selection routines (PR #130).
5722 *) Fix EVP_dsa_sha macro.
5726 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5727 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5728 that was added in OpenSSL 0.9.6d.
5730 As the countermeasure turned out to be incompatible with some
5731 broken SSL implementations, the new option is part of SSL_OP_ALL.
5732 SSL_OP_ALL is usually employed when compatibility with weird SSL
5733 implementations is desired (e.g. '-bugs' option to 's_client' and
5734 's_server'), so the new option is automatically set in many
5738 *) Changes in security patch:
5740 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5741 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5742 Air Force Materiel Command, USAF, under agreement number
5745 *) Add various sanity checks to asn1_get_length() to reject
5746 the ASN1 length bytes if they exceed sizeof(long), will appear
5747 negative or the content length exceeds the length of the
5748 supplied buffer. (CVE-2002-0659)
5749 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5751 *) Assertions for various potential buffer overflows, not known to
5753 [Ben Laurie (CHATS)]
5755 *) Various temporary buffers to hold ASCII versions of integers were
5756 too small for 64 bit platforms. (CVE-2002-0655)
5757 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5759 *) Remote buffer overflow in SSL3 protocol - an attacker could
5760 supply an oversized session ID to a client. (CVE-2002-0656)
5761 [Ben Laurie (CHATS)]
5763 *) Remote buffer overflow in SSL2 protocol - an attacker could
5764 supply an oversized client master key. (CVE-2002-0656)
5765 [Ben Laurie (CHATS)]
5767 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5769 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5770 encoded as NULL) with id-dsa-with-sha1.
5771 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5773 *) Check various X509_...() return values in apps/req.c.
5774 [Nils Larsch <nla@trustcenter.de>]
5776 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5777 an end-of-file condition would erronously be flagged, when the CRLF
5778 was just at the end of a processed block. The bug was discovered when
5779 processing data through a buffering memory BIO handing the data to a
5780 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5781 <ptsekov@syntrex.com> and Nedelcho Stanev.
5784 *) Implement a countermeasure against a vulnerability recently found
5785 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5786 before application data chunks to avoid the use of known IVs
5787 with data potentially chosen by the attacker.
5790 *) Fix length checks in ssl3_get_client_hello().
5793 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5794 to prevent ssl3_read_internal() from incorrectly assuming that
5795 ssl3_read_bytes() found application data while handshake
5796 processing was enabled when in fact s->s3->in_read_app_data was
5797 merely automatically cleared during the initial handshake.
5798 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5800 *) Fix object definitions for Private and Enterprise: they were not
5801 recognized in their shortname (=lowercase) representation. Extend
5802 obj_dat.pl to issue an error when using undefined keywords instead
5803 of silently ignoring the problem (Svenning Sorensen
5804 <sss@sss.dnsalias.net>).
5807 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5808 generators, i.e. generators other than 2 and 5. (Previously, the
5809 code did not properly initialise the 'add' and 'rem' values to
5810 BN_generate_prime().)
5812 In the new general case, we do not insist that 'generator' is
5813 actually a primitive root: This requirement is rather pointless;
5814 a generator of the order-q subgroup is just as good, if not
5818 *) Map new X509 verification errors to alerts. Discovered and submitted by
5819 Tom Wu <tom@arcot.com>.
5822 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5823 returning non-zero before the data has been completely received
5824 when using non-blocking I/O.
5825 [Bodo Moeller; problem pointed out by John Hughes]
5827 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5828 [Ben Laurie, Lutz Jaenicke]
5830 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5831 Yoram Zahavi <YoramZ@gilian.com>).
5834 *) Add information about CygWin 1.3 and on, and preserve proper
5835 configuration for the versions before that.
5836 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5838 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5839 check whether we deal with a copy of a session and do not delete from
5840 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5841 <izhar@checkpoint.com>.
5844 *) Do not store session data into the internal session cache, if it
5845 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5846 flag is set). Proposed by Aslam <aslam@funk.com>.
5849 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5853 *) [In 0.9.6d-engine release:]
5854 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5855 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5857 *) Add the configuration target linux-s390x.
5858 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5860 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5861 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5862 variable as an indication that a ClientHello message has been
5863 received. As the flag value will be lost between multiple
5864 invocations of ssl3_accept when using non-blocking I/O, the
5865 function may not be aware that a handshake has actually taken
5866 place, thus preventing a new session from being added to the
5869 To avoid this problem, we now set s->new_session to 2 instead of
5870 using a local variable.
5871 [Lutz Jaenicke, Bodo Moeller]
5873 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5874 if the SSL_R_LENGTH_MISMATCH error is detected.
5875 [Geoff Thorpe, Bodo Moeller]
5877 *) New 'shared_ldflag' column in Configure platform table.
5880 *) Fix EVP_CIPHER_mode macro.
5881 ["Dan S. Camper" <dan@bti.net>]
5883 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5884 type, we must throw them away by setting rr->length to 0.
5885 [D P Chang <dpc@qualys.com>]
5887 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5889 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5890 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5891 worked incorrectly for those cases where range = 10..._2 and
5892 3*range is two bits longer than range.)
5895 *) Only add signing time to PKCS7 structures if it is not already
5899 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5900 OBJ_ld_ce should be OBJ_id_ce.
5901 Also some ip-pda OIDs in crypto/objects/objects.txt were
5902 incorrect (cf. RFC 3039).
5903 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5905 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5906 returns early because it has nothing to do.
5907 [Andy Schneider <andy.schneider@bjss.co.uk>]
5909 *) [In 0.9.6c-engine release:]
5910 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5911 [Andy Schneider <andy.schneider@bjss.co.uk>]
5913 *) [In 0.9.6c-engine release:]
5914 Add support for Cryptographic Appliance's keyserver technology.
5915 (Use engine 'keyclient')
5916 [Cryptographic Appliances and Geoff Thorpe]
5918 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5919 is called via tools/c89.sh because arguments have to be
5920 rearranged (all '-L' options must appear before the first object
5922 [Richard Shapiro <rshapiro@abinitio.com>]
5924 *) [In 0.9.6c-engine release:]
5925 Add support for Broadcom crypto accelerator cards, backported
5927 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5929 *) [In 0.9.6c-engine release:]
5930 Add support for SureWare crypto accelerator cards from
5931 Baltimore Technologies. (Use engine 'sureware')
5932 [Baltimore Technologies and Mark Cox]
5934 *) [In 0.9.6c-engine release:]
5935 Add support for crypto accelerator cards from Accelerated
5936 Encryption Processing, www.aep.ie. (Use engine 'aep')
5937 [AEP Inc. and Mark Cox]
5939 *) Add a configuration entry for gcc on UnixWare.
5940 [Gary Benson <gbenson@redhat.com>]
5942 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5943 messages are stored in a single piece (fixed-length part and
5944 variable-length part combined) and fix various bugs found on the way.
5947 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5948 instead. BIO_gethostbyname() does not know what timeouts are
5949 appropriate, so entries would stay in cache even when they have
5951 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5953 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5954 faced with a pathologically small ClientHello fragment that does
5955 not contain client_version: Instead of aborting with an error,
5956 simply choose the highest available protocol version (i.e.,
5957 TLS 1.0 unless it is disabled). In practice, ClientHello
5958 messages are never sent like this, but this change gives us
5959 strictly correct behaviour at least for TLS.
5962 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5963 never resets s->method to s->ctx->method when called from within
5964 one of the SSL handshake functions.
5965 [Bodo Moeller; problem pointed out by Niko Baric]
5967 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5968 (sent using the client's version number) if client_version is
5969 smaller than the protocol version in use. Also change
5970 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5971 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5972 the client will at least see that alert.
5975 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5979 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5980 client receives HelloRequest while in a handshake.
5981 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5983 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5984 should end in 'break', not 'goto end' which circuments various
5985 cleanups done in state SSL_ST_OK. But session related stuff
5986 must be disabled for SSL_ST_OK in the case that we just sent a
5989 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5990 before just sending a HelloRequest.
5991 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5993 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5994 reveal whether illegal block cipher padding was found or a MAC
5995 verification error occured. (Neither SSLerr() codes nor alerts
5996 are directly visible to potential attackers, but the information
5997 may leak via logfiles.)
5999 Similar changes are not required for the SSL 2.0 implementation
6000 because the number of padding bytes is sent in clear for SSL 2.0,
6001 and the extra bytes are just ignored. However ssl/s2_pkt.c
6002 failed to verify that the purported number of padding bytes is in
6006 *) Add OpenUNIX-8 support including shared libraries
6007 (Boyd Lynn Gerber <gerberb@zenez.com>).
6010 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6011 'wristwatch attack' using huge encoding parameters (cf.
6012 James H. Manger's CRYPTO 2001 paper). Note that the
6013 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6014 encoding parameters and hence was not vulnerable.
6017 *) BN_sqr() bug fix.
6018 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6020 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6021 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6022 followed by modular reduction.
6023 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6025 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6026 equivalent based on BN_pseudo_rand() instead of BN_rand().
6029 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6030 This function was broken, as the check for a new client hello message
6031 to handle SGC did not allow these large messages.
6032 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6035 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6038 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6039 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6042 *) Rework the configuration and shared library support for Tru64 Unix.
6043 The configuration part makes use of modern compiler features and
6044 still retains old compiler behavior for those that run older versions
6045 of the OS. The shared library support part includes a variant that
6046 uses the RPATH feature, and is available through the special
6047 configuration target "alpha-cc-rpath", which will never be selected
6049 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6051 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6052 with the same message size as in ssl3_get_certificate_request().
6053 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6054 messages might inadvertently be reject as too long.
6055 [Petr Lampa <lampa@fee.vutbr.cz>]
6057 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6060 *) Modified SSL library such that the verify_callback that has been set
6061 specificly for an SSL object with SSL_set_verify() is actually being
6062 used. Before the change, a verify_callback set with this function was
6063 ignored and the verify_callback() set in the SSL_CTX at the time of
6064 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6065 to allow the necessary settings.
6068 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6069 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6070 done automatically (in contradiction to the requirements of the C
6071 standard). This made problems when used from OpenSSH.
6074 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6075 dh->length and always used
6077 BN_rand_range(priv_key, dh->p).
6079 BN_rand_range() is not necessary for Diffie-Hellman, and this
6080 specific range makes Diffie-Hellman unnecessarily inefficient if
6081 dh->length (recommended exponent length) is much smaller than the
6082 length of dh->p. We could use BN_rand_range() if the order of
6083 the subgroup was stored in the DH structure, but we only have
6088 BN_rand(priv_key, l, ...)
6090 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6096 RSA_eay_public_encrypt
6097 RSA_eay_private_decrypt
6098 RSA_eay_private_encrypt (signing)
6099 RSA_eay_public_decrypt (signature verification)
6101 (default implementations for RSA_public_encrypt,
6102 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6103 always reject numbers >= n.
6106 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6107 to synchronize access to 'locking_thread'. This is necessary on
6108 systems where access to 'locking_thread' (an 'unsigned long'
6109 variable) is not atomic.
6112 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6113 *before* setting the 'crypto_lock_rand' flag. The previous code had
6114 a race condition if 0 is a valid thread ID.
6115 [Travis Vitek <vitek@roguewave.com>]
6117 *) Add support for shared libraries under Irix.
6118 [Albert Chin-A-Young <china@thewrittenword.com>]
6120 *) Add configuration option to build on Linux on both big-endian and
6122 [Ralf Baechle <ralf@uni-koblenz.de>]
6124 *) Add the possibility to create shared libraries on HP-UX.
6127 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6129 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6130 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6131 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6132 PRNG state recovery was possible based on the output of
6133 one PRNG request appropriately sized to gain knowledge on
6134 'md' followed by enough consecutive 1-byte PRNG requests
6135 to traverse all of 'state'.
6137 1. When updating 'md_local' (the current thread's copy of 'md')
6138 during PRNG output generation, hash all of the previous
6139 'md_local' value, not just the half used for PRNG output.
6141 2. Make the number of bytes from 'state' included into the hash
6142 independent from the number of PRNG bytes requested.
6144 The first measure alone would be sufficient to avoid
6145 Markku-Juhani's attack. (Actually it had never occurred
6146 to me that the half of 'md_local' used for chaining was the
6147 half from which PRNG output bytes were taken -- I had always
6148 assumed that the secret half would be used.) The second
6149 measure makes sure that additional data from 'state' is never
6150 mixed into 'md_local' in small portions; this heuristically
6151 further strengthens the PRNG.
6154 *) Fix crypto/bn/asm/mips3.s.
6157 *) When only the key is given to "enc", the IV is undefined. Print out
6158 an error message in this case.
6161 *) Handle special case when X509_NAME is empty in X509 printing routines.
6164 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6165 positive and less than q.
6168 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6169 used: it isn't thread safe and the add_lock_callback should handle
6171 [Paul Rose <Paul.Rose@bridge.com>]
6173 *) Verify that incoming data obeys the block size in
6174 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6178 [Ulf Möller, Bodo Möller]
6180 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6181 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6182 when fixing the server behaviour for backwards-compatible 'client
6183 hello' messages. (Note that the attack is impractical against
6184 SSL 3.0 and TLS 1.0 anyway because length and version checking
6185 means that the probability of guessing a valid ciphertext is
6186 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6189 Before 0.9.5, the countermeasure (hide the error by generating a
6190 random 'decryption result') did not work properly because
6191 ERR_clear_error() was missing, meaning that SSL_get_error() would
6192 detect the supposedly ignored error.
6194 Both problems are now fixed.
6197 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6198 (previously it was 1024).
6201 *) Fix for compatibility mode trust settings: ignore trust settings
6202 unless some valid trust or reject settings are present.
6205 *) Fix for blowfish EVP: its a variable length cipher.
6208 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6209 parameters in DSA public key structures and return an error in the
6210 DSA routines if parameters are absent.
6213 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6214 in the current directory if neither $RANDFILE nor $HOME was set.
6215 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6216 caused some confusion to Windows users who haven't defined $HOME.
6217 Thus RAND_file_name() is changed again: e_os.h can define a
6218 DEFAULT_HOME, which will be used if $HOME is not set.
6219 For Windows, we use "C:"; on other platforms, we still require
6220 environment variables.
6222 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6223 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6224 having multiple threads call RAND_poll() concurrently.
6227 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6228 combination of a flag and a thread ID variable.
6229 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6230 flag), *other* threads can enter ssleay_add_bytes without obeying
6231 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6232 that they do not hold after the first thread unsets add_do_not_lock).
6235 *) Change bctest again: '-x' expressions are not available in all
6239 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6241 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6242 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6244 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6245 the default extension for executables, if any. Also, make the perl
6246 scripts that use symlink() to test if it really exists and use "cp"
6247 if it doesn't. All this made OpenSSL compilable and installable in
6251 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6252 If SEQUENCE is length is indefinite just set c->slen to the total
6253 amount of data available.
6254 [Steve Henson, reported by shige@FreeBSD.org]
6255 [This change does not apply to 0.9.7.]
6257 *) Change bctest to avoid here-documents inside command substitution
6258 (workaround for FreeBSD /bin/sh bug).
6259 For compatibility with Ultrix, avoid shell functions (introduced
6260 in the bctest version that searches along $PATH).
6263 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6264 with des_encrypt() defined on some operating systems, like Solaris
6268 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6269 On the Importance of Eliminating Errors in Cryptographic
6270 Computations, J. Cryptology 14 (2001) 2, 101-119,
6271 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6274 *) MIPS assembler BIGNUM division bug fix.
6277 *) Disabled incorrect Alpha assembler code.
6280 *) Fix PKCS#7 decode routines so they correctly update the length
6281 after reading an EOC for the EXPLICIT tag.
6283 [This change does not apply to 0.9.7.]
6285 *) Fix bug in PKCS#12 key generation routines. This was triggered
6286 if a 3DES key was generated with a 0 initial byte. Include
6287 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6288 (but broken) behaviour.
6291 *) Enhance bctest to search for a working bc along $PATH and print
6293 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6295 *) Fix memory leaks in err.c: free err_data string if necessary;
6296 don't write to the wrong index in ERR_set_error_data.
6299 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6303 *) Replace rdtsc with _emit statements for VC++ version 5.
6304 [Jeremy Cooper <jeremy@baymoo.org>]
6306 *) Make it possible to reuse SSLv2 sessions.
6309 *) In copy_email() check for >= 0 as a return value for
6310 X509_NAME_get_index_by_NID() since 0 is a valid index.
6311 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6313 *) Avoid coredump with unsupported or invalid public keys by checking if
6314 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6315 PKCS7_verify() fails with non detached data.
6318 *) Don't use getenv in library functions when run as setuid/setgid.
6319 New function OPENSSL_issetugid().
6322 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6323 due to incorrect handling of multi-threading:
6325 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6327 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6329 3. Count how many times MemCheck_off() has been called so that
6330 nested use can be treated correctly. This also avoids
6331 inband-signalling in the previous code (which relied on the
6332 assumption that thread ID 0 is impossible).
6335 *) Add "-rand" option also to s_client and s_server.
6338 *) Fix CPU detection on Irix 6.x.
6339 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6340 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6342 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6345 [This change does not apply to 0.9.7.]
6347 *) Use the cached encoding of an X509_NAME structure rather than
6348 copying it. This is apparently the reason for the libsafe "errors"
6349 but the code is actually correct.
6352 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6353 Bleichenbacher's DSA attack.
6354 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6355 to be set and top=0 forces the highest bit to be set; top=-1 is new
6356 and leaves the highest bit random.
6357 [Ulf Moeller, Bodo Moeller]
6359 *) In the NCONF_...-based implementations for CONF_... queries
6360 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6361 a temporary CONF structure with the data component set to NULL
6362 (which gives segmentation faults in lh_retrieve).
6363 Instead, use NULL for the CONF pointer in CONF_get_string and
6364 CONF_get_number (which may use environment variables) and directly
6365 return NULL from CONF_get_section.
6368 *) Fix potential buffer overrun for EBCDIC.
6371 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6372 keyUsage if basicConstraints absent for a CA.
6375 *) Make SMIME_write_PKCS7() write mail header values with a format that
6376 is more generally accepted (no spaces before the semicolon), since
6377 some programs can't parse those values properly otherwise. Also make
6378 sure BIO's that break lines after each write do not create invalid
6382 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6383 macros previously used would not encode an empty SEQUENCE OF
6384 and break the signature.
6386 [This change does not apply to 0.9.7.]
6388 *) Zero the premaster secret after deriving the master secret in
6392 *) Add some EVP_add_digest_alias registrations (as found in
6393 OpenSSL_add_all_digests()) to SSL_library_init()
6394 aka OpenSSL_add_ssl_algorithms(). This provides improved
6395 compatibility with peers using X.509 certificates
6396 with unconventional AlgorithmIdentifier OIDs.
6399 *) Fix for Irix with NO_ASM.
6400 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6402 *) ./config script fixes.
6403 [Ulf Moeller, Richard Levitte]
6405 *) Fix 'openssl passwd -1'.
6408 *) Change PKCS12_key_gen_asc() so it can cope with non null
6409 terminated strings whose length is passed in the passlen
6410 parameter, for example from PEM callbacks. This was done
6411 by adding an extra length parameter to asc2uni().
6412 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6414 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6415 call failed, free the DSA structure.
6418 *) Fix to uni2asc() to cope with zero length Unicode strings.
6419 These are present in some PKCS#12 files.
6422 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6423 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6424 when writing a 32767 byte record.
6425 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6427 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6428 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6430 (RSA objects have a reference count access to which is protected
6431 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6432 so they are meant to be shared between threads.)
6433 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6434 "Reddie, Steven" <Steven.Reddie@ca.com>]
6436 *) Fix a deadlock in CRYPTO_mem_leaks().
6439 *) Use better test patterns in bntest.
6442 *) rand_win.c fix for Borland C.
6445 *) BN_rshift bugfix for n == 0.
6448 *) Add a 'bctest' script that checks for some known 'bc' bugs
6449 so that 'make test' does not abort just because 'bc' is broken.
6452 *) Store verify_result within SSL_SESSION also for client side to
6453 avoid potential security hole. (Re-used sessions on the client side
6454 always resulted in verify_result==X509_V_OK, not using the original
6455 result of the server certificate verification.)
6458 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6459 SSL3_RT_APPLICATION_DATA, return 0.
6460 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6464 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6465 releases, have been re-implemented by renaming the previous
6466 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6467 and ssl3_read_internal, respectively, and adding 'peek' parameters
6468 to them. The new ssl[23]_{read,peek} functions are calls to
6469 ssl[23]_read_internal with the 'peek' flag set appropriately.
6470 A 'peek' parameter has also been added to ssl3_read_bytes, which
6471 does the actual work for ssl3_read_internal.
6474 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6475 the method-specific "init()" handler. Also clean up ex_data after
6476 calling the method-specific "finish()" handler. Previously, this was
6477 happening the other way round.
6480 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6481 The previous value, 12, was not always sufficient for BN_mod_exp().
6484 *) Make sure that shared libraries get the internal name engine with
6485 the full version number and not just 0. This should mark the
6486 shared libraries as not backward compatible. Of course, this should
6487 be changed again when we can guarantee backward binary compatibility.
6490 *) Fix typo in get_cert_by_subject() in by_dir.c
6491 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6493 *) Rework the system to generate shared libraries:
6495 - Make note of the expected extension for the shared libraries and
6496 if there is a need for symbolic links from for example libcrypto.so.0
6497 to libcrypto.so.0.9.7. There is extended info in Configure for
6500 - Make as few rebuilds of the shared libraries as possible.
6502 - Still avoid linking the OpenSSL programs with the shared libraries.
6504 - When installing, install the shared libraries separately from the
6508 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6510 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6511 and not in SSL_clear because the latter is also used by the
6512 accept/connect functions; previously, the settings made by
6513 SSL_set_read_ahead would be lost during the handshake.
6514 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6516 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6517 Previously, it would create entries for disableed algorithms no
6521 *) Added several new manual pages for SSL_* function.
6524 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6526 *) In ssl23_get_client_hello, generate an error message when faced
6527 with an initial SSL 3.0/TLS record that is too small to contain the
6528 first two bytes of the ClientHello message, i.e. client_version.
6529 (Note that this is a pathologic case that probably has never happened
6530 in real life.) The previous approach was to use the version number
6531 from the record header as a substitute; but our protocol choice
6532 should not depend on that one because it is not authenticated
6533 by the Finished messages.
6536 *) More robust randomness gathering functions for Windows.
6537 [Jeffrey Altman <jaltman@columbia.edu>]
6539 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6540 not set then we don't setup the error code for issuer check errors
6541 to avoid possibly overwriting other errors which the callback does
6542 handle. If an application does set the flag then we assume it knows
6543 what it is doing and can handle the new informational codes
6547 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6548 a general "ANY" type, as such it should be able to decode anything
6549 including tagged types. However it didn't check the class so it would
6550 wrongly interpret tagged types in the same way as their universal
6551 counterpart and unknown types were just rejected. Changed so that the
6552 tagged and unknown types are handled in the same way as a SEQUENCE:
6553 that is the encoding is stored intact. There is also a new type
6554 "V_ASN1_OTHER" which is used when the class is not universal, in this
6555 case we have no idea what the actual type is so we just lump them all
6559 *) On VMS, stdout may very well lead to a file that is written to
6560 in a record-oriented fashion. That means that every write() will
6561 write a separate record, which will be read separately by the
6562 programs trying to read from it. This can be very confusing.
6564 The solution is to put a BIO filter in the way that will buffer
6565 text until a linefeed is reached, and then write everything a
6566 line at a time, so every record written will be an actual line,
6567 not chunks of lines and not (usually doesn't happen, but I've
6568 seen it once) several lines in one record. BIO_f_linebuffer() is
6571 Currently, it's a VMS-only method, because that's where it has
6572 been tested well enough.
6575 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6576 it can return incorrect results.
6577 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6578 but it was in 0.9.6-beta[12].)
6581 *) Disable the check for content being present when verifying detached
6582 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6583 include zero length content when signing messages.
6586 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6587 BIO_ctrl (for BIO pairs).
6590 *) Add DSO method for VMS.
6593 *) Bug fix: Montgomery multiplication could produce results with the
6597 *) Add RPM specification openssl.spec and modify it to build three
6598 packages. The default package contains applications, application
6599 documentation and run-time libraries. The devel package contains
6600 include files, static libraries and function documentation. The
6601 doc package contains the contents of the doc directory. The original
6602 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6605 *) Add a large number of documentation files for many SSL routines.
6606 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6608 *) Add a configuration entry for Sony News 4.
6609 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6611 *) Don't set the two most significant bits to one when generating a
6612 random number < q in the DSA library.
6615 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6616 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6617 the underlying transport is blocking) if a handshake took place.
6618 (The default behaviour is needed by applications such as s_client
6619 and s_server that use select() to determine when to use SSL_read;
6620 but for applications that know in advance when to expect data, it
6621 just makes things more complicated.)
6624 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6628 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6629 work better on such systems.
6630 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6632 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6633 Update PKCS12_parse() so it copies the friendlyName and the
6634 keyid to the certificates aux info.
6637 *) Fix bug in PKCS7_verify() which caused an infinite loop
6638 if there was more than one signature.
6639 [Sven Uszpelkat <su@celocom.de>]
6641 *) Major change in util/mkdef.pl to include extra information
6642 about each symbol, as well as presentig variables as well
6643 as functions. This change means that there's n more need
6644 to rebuild the .num files when some algorithms are excluded.
6647 *) Allow the verify time to be set by an application,
6648 rather than always using the current time.
6651 *) Phase 2 verify code reorganisation. The certificate
6652 verify code now looks up an issuer certificate by a
6653 number of criteria: subject name, authority key id
6654 and key usage. It also verifies self signed certificates
6655 by the same criteria. The main comparison function is
6656 X509_check_issued() which performs these checks.
6658 Lot of changes were necessary in order to support this
6659 without completely rewriting the lookup code.
6661 Authority and subject key identifier are now cached.
6663 The LHASH 'certs' is X509_STORE has now been replaced
6664 by a STACK_OF(X509_OBJECT). This is mainly because an
6665 LHASH can't store or retrieve multiple objects with
6666 the same hash value.
6668 As a result various functions (which were all internal
6669 use only) have changed to handle the new X509_STORE
6670 structure. This will break anything that messed round
6671 with X509_STORE internally.
6673 The functions X509_STORE_add_cert() now checks for an
6674 exact match, rather than just subject name.
6676 The X509_STORE API doesn't directly support the retrieval
6677 of multiple certificates matching a given criteria, however
6678 this can be worked round by performing a lookup first
6679 (which will fill the cache with candidate certificates)
6680 and then examining the cache for matches. This is probably
6681 the best we can do without throwing out X509_LOOKUP
6682 entirely (maybe later...).
6684 The X509_VERIFY_CTX structure has been enhanced considerably.
6686 All certificate lookup operations now go via a get_issuer()
6687 callback. Although this currently uses an X509_STORE it
6688 can be replaced by custom lookups. This is a simple way
6689 to bypass the X509_STORE hackery necessary to make this
6690 work and makes it possible to use more efficient techniques
6691 in future. A very simple version which uses a simple
6692 STACK for its trusted certificate store is also provided
6693 using X509_STORE_CTX_trusted_stack().
6695 The verify_cb() and verify() callbacks now have equivalents
6696 in the X509_STORE_CTX structure.
6698 X509_STORE_CTX also has a 'flags' field which can be used
6699 to customise the verify behaviour.
6702 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6703 excludes S/MIME capabilities.
6706 *) When a certificate request is read in keep a copy of the
6707 original encoding of the signed data and use it when outputing
6708 again. Signatures then use the original encoding rather than
6709 a decoded, encoded version which may cause problems if the
6710 request is improperly encoded.
6713 *) For consistency with other BIO_puts implementations, call
6714 buffer_write(b, ...) directly in buffer_puts instead of calling
6717 In BIO_puts, increment b->num_write as in BIO_write.
6718 [Peter.Sylvester@EdelWeb.fr]
6720 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6721 BN_zero, we may not return a BIGNUM with an array consisting of
6725 *) Avoid calling abort() from within the library when problems are
6726 detected, except if preprocessor symbols have been defined
6727 (such as REF_CHECK, BN_DEBUG etc.).
6730 *) New openssl application 'rsautl'. This utility can be
6731 used for low level RSA operations. DER public key
6732 BIO/fp routines also added.
6735 *) New Configure entry and patches for compiling on QNX 4.
6736 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6738 *) A demo state-machine implementation was sponsored by
6739 Nuron (http://www.nuron.com/) and is now available in
6740 demos/state_machine.
6743 *) New options added to the 'dgst' utility for signature
6744 generation and verification.
6747 *) Unrecognized PKCS#7 content types are now handled via a
6748 catch all ASN1_TYPE structure. This allows unsupported
6749 types to be stored as a "blob" and an application can
6750 encode and decode it manually.
6753 *) Fix various signed/unsigned issues to make a_strex.c
6755 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6757 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6758 length if passed a buffer. ASN1_INTEGER_to_BN failed
6759 if passed a NULL BN and its argument was negative.
6760 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6762 *) Modification to PKCS#7 encoding routines to output definite
6763 length encoding. Since currently the whole structures are in
6764 memory there's not real point in using indefinite length
6765 constructed encoding. However if OpenSSL is compiled with
6766 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6769 *) Added BIO_vprintf() and BIO_vsnprintf().
6772 *) Added more prefixes to parse for in the the strings written
6773 through a logging bio, to cover all the levels that are available
6774 through syslog. The prefixes are now:
6776 PANIC, EMERG, EMR => LOG_EMERG
6777 ALERT, ALR => LOG_ALERT
6778 CRIT, CRI => LOG_CRIT
6779 ERROR, ERR => LOG_ERR
6780 WARNING, WARN, WAR => LOG_WARNING
6781 NOTICE, NOTE, NOT => LOG_NOTICE
6782 INFO, INF => LOG_INFO
6783 DEBUG, DBG => LOG_DEBUG
6785 and as before, if none of those prefixes are present at the
6786 beginning of the string, LOG_ERR is chosen.
6788 On Win32, the LOG_* levels are mapped according to this:
6790 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6791 LOG_WARNING => EVENTLOG_WARNING_TYPE
6792 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6796 *) Made it possible to reconfigure with just the configuration
6797 argument "reconf" or "reconfigure". The command line arguments
6798 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6799 and are retrieved from there when reconfiguring.
6803 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6805 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6808 *) The obj_dat.pl script was messing up the sorting of object
6809 names. The reason was that it compared the quoted version
6810 of strings as a result "OCSP" > "OCSP Signing" because
6811 " > SPACE. Changed script to store unquoted versions of
6812 names and add quotes on output. It was also omitting some
6813 names from the lookup table if they were given a default
6814 value (that is if SN is missing it is given the same
6815 value as LN and vice versa), these are now added on the
6816 grounds that if an object has a name we should be able to
6817 look it up. Finally added warning output when duplicate
6818 short or long names are found.
6821 *) Changes needed for Tandem NSK.
6822 [Scott Uroff <scott@xypro.com>]
6824 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6825 RSA_padding_check_SSLv23(), special padding was never detected
6826 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6827 version rollback attacks was not effective.
6829 In s23_clnt.c, don't use special rollback-attack detection padding
6830 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6831 client; similarly, in s23_srvr.c, don't do the rollback check if
6832 SSL 2.0 is the only protocol enabled in the server.
6835 *) Make it possible to get hexdumps of unprintable data with 'openssl
6836 asn1parse'. By implication, the functions ASN1_parse_dump() and
6837 BIO_dump_indent() are added.
6840 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6841 these print out strings and name structures based on various
6842 flags including RFC2253 support and proper handling of
6843 multibyte characters. Added options to the 'x509' utility
6844 to allow the various flags to be set.
6847 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6848 Also change the functions X509_cmp_current_time() and
6849 X509_gmtime_adj() work with an ASN1_TIME structure,
6850 this will enable certificates using GeneralizedTime in validity
6851 dates to be checked.
6854 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6855 negative public key encodings) on by default,
6856 NO_NEG_PUBKEY_BUG can be set to disable it.
6859 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6860 content octets. An i2c_ASN1_OBJECT is unnecessary because
6861 the encoding can be trivially obtained from the structure.
6864 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6865 not read locks (CRYPTO_r_[un]lock).
6868 *) A first attempt at creating official support for shared
6869 libraries through configuration. I've kept it so the
6870 default is static libraries only, and the OpenSSL programs
6871 are always statically linked for now, but there are
6872 preparations for dynamic linking in place.
6873 This has been tested on Linux and Tru64.
6876 *) Randomness polling function for Win9x, as described in:
6877 Peter Gutmann, Software Generation of Practically Strong
6881 *) Fix so PRNG is seeded in req if using an already existing
6885 *) New options to smime application. -inform and -outform
6886 allow alternative formats for the S/MIME message including
6887 PEM and DER. The -content option allows the content to be
6888 specified separately. This should allow things like Netscape
6889 form signing output easier to verify.
6892 *) Fix the ASN1 encoding of tags using the 'long form'.
6895 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6896 STRING types. These convert content octets to and from the
6897 underlying type. The actual tag and length octets are
6898 already assumed to have been read in and checked. These
6899 are needed because all other string types have virtually
6900 identical handling apart from the tag. By having versions
6901 of the ASN1 functions that just operate on content octets
6902 IMPLICIT tagging can be handled properly. It also allows
6903 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6904 and ASN1_INTEGER are identical apart from the tag.
6907 *) Change the handling of OID objects as follows:
6909 - New object identifiers are inserted in objects.txt, following
6910 the syntax given in objects.README.
6911 - objects.pl is used to process obj_mac.num and create a new
6913 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6916 This is currently kind of a hack, and the perl code in objects.pl
6917 isn't very elegant, but it works as I intended. The simplest way
6918 to check that it worked correctly is to look in obj_dat.h and
6919 check the array nid_objs and make sure the objects haven't moved
6920 around (this is important!). Additions are OK, as well as
6921 consistent name changes.
6924 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6927 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6928 The given file adds to whatever has already been seeded into the
6929 random pool through the RANDFILE configuration file option or
6930 environment variable, or the default random state file.
6933 *) mkstack.pl now sorts each macro group into lexical order.
6934 Previously the output order depended on the order the files
6935 appeared in the directory, resulting in needless rewriting
6939 *) Patches to make OpenSSL compile under Win32 again. Mostly
6940 work arounds for the VC++ problem that it treats func() as
6941 func(void). Also stripped out the parts of mkdef.pl that
6942 added extra typesafe functions: these no longer exist.
6945 *) Reorganisation of the stack code. The macros are now all
6946 collected in safestack.h . Each macro is defined in terms of
6947 a "stack macro" of the form SKM_<name>(type, a, b). The
6948 DEBUG_SAFESTACK is now handled in terms of function casts,
6949 this has the advantage of retaining type safety without the
6950 use of additional functions. If DEBUG_SAFESTACK is not defined
6951 then the non typesafe macros are used instead. Also modified the
6952 mkstack.pl script to handle the new form. Needs testing to see
6953 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6954 the default if no major problems. Similar behaviour for ASN1_SET_OF
6955 and PKCS12_STACK_OF.
6958 *) When some versions of IIS use the 'NET' form of private key the
6959 key derivation algorithm is different. Normally MD5(password) is
6960 used as a 128 bit RC4 key. In the modified case
6961 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6962 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6963 as the old Netscape_RSA functions except they have an additional
6964 'sgckey' parameter which uses the modified algorithm. Also added
6965 an -sgckey command line option to the rsa utility. Thanks to
6966 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6967 algorithm to openssl-dev.
6970 *) The evp_local.h macros were using 'c.##kname' which resulted in
6971 invalid expansion on some systems (SCO 5.0.5 for example).
6972 Corrected to 'c.kname'.
6973 [Phillip Porch <root@theporch.com>]
6975 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6976 a STACK of email addresses from a certificate or request, these look
6977 in the subject name and the subject alternative name extensions and
6978 omit any duplicate addresses.
6981 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6982 This makes DSA verification about 2 % faster.
6985 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6986 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6987 plus overhead for 1024 bit moduli).
6988 This makes exponentiations about 0.5 % faster for 1024 bit
6989 exponents (as measured by "openssl speed rsa2048").
6992 *) Rename memory handling macros to avoid conflicts with other
6994 Malloc => OPENSSL_malloc
6995 Malloc_locked => OPENSSL_malloc_locked
6996 Realloc => OPENSSL_realloc
6997 Free => OPENSSL_free
7000 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7001 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7004 *) CygWin32 support.
7005 [John Jarvie <jjarvie@newsguy.com>]
7007 *) The type-safe stack code has been rejigged. It is now only compiled
7008 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7009 by default all type-specific stack functions are "#define"d back to
7010 standard stack functions. This results in more streamlined output
7011 but retains the type-safety checking possibilities of the original
7015 *) The STACK code has been cleaned up, and certain type declarations
7016 that didn't make a lot of sense have been brought in line. This has
7017 also involved a cleanup of sorts in safestack.h to more correctly
7018 map type-safe stack functions onto their plain stack counterparts.
7019 This work has also resulted in a variety of "const"ifications of
7020 lots of the code, especially "_cmp" operations which should normally
7021 be prototyped with "const" parameters anyway.
7024 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7025 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7026 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7027 where all of 'md' is used each time the PRNG is used, but 'state'
7028 is used only indexed by a cyclic counter. As entropy may not be
7029 well distributed from the beginning, 'md' is important as a
7030 chaining variable. However, the output function chains only half
7031 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7032 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7033 in all of 'state' being rewritten, with the new values depending
7034 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7037 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7038 the handshake is continued after ssl_verify_cert_chain();
7039 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7040 can lead to 'unexplainable' connection aborts later.
7041 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7043 *) Major EVP API cipher revision.
7044 Add hooks for extra EVP features. This allows various cipher
7045 parameters to be set in the EVP interface. Support added for variable
7046 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7047 setting of RC2 and RC5 parameters.
7049 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7052 Remove lots of duplicated code from the EVP library. For example *every*
7053 cipher init() function handles the 'iv' in the same way according to the
7054 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7055 for CFB and OFB modes they zero ctx->num.
7057 New functionality allows removal of S/MIME code RC2 hack.
7059 Most of the routines have the same form and so can be declared in terms
7062 By shifting this to the top level EVP_CipherInit() it can be removed from
7063 all individual ciphers. If the cipher wants to handle IVs or keys
7064 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7067 Change lots of functions like EVP_EncryptUpdate() to now return a
7068 value: although software versions of the algorithms cannot fail
7069 any installed hardware versions can.
7072 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7073 this option is set, tolerate broken clients that send the negotiated
7074 protocol version number instead of the requested protocol version
7078 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7079 i.e. non-zero for export ciphersuites, zero otherwise.
7080 Previous versions had this flag inverted, inconsistent with
7081 rsa_tmp_cb (..._TMP_RSA_CB).
7082 [Bodo Moeller; problem reported by Amit Chopra]
7084 *) Add missing DSA library text string. Work around for some IIS
7085 key files with invalid SEQUENCE encoding.
7088 *) Add a document (doc/standards.txt) that list all kinds of standards
7089 and so on that are implemented in OpenSSL.
7092 *) Enhance c_rehash script. Old version would mishandle certificates
7093 with the same subject name hash and wouldn't handle CRLs at all.
7094 Added -fingerprint option to crl utility, to support new c_rehash
7098 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7101 *) Fix for SSL server purpose checking. Server checking was
7102 rejecting certificates which had extended key usage present
7103 but no ssl client purpose.
7104 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7106 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7107 is a little unclear about how a blank password is handled.
7108 Since the password in encoded as a BMPString with terminating
7109 double NULL a zero length password would end up as just the
7110 double NULL. However no password at all is different and is
7111 handled differently in the PKCS#12 key generation code. NS
7112 treats a blank password as zero length. MSIE treats it as no
7113 password on export: but it will try both on import. We now do
7114 the same: PKCS12_parse() tries zero length and no password if
7115 the password is set to "" or NULL (NULL is now a valid password:
7116 it wasn't before) as does the pkcs12 application.
7119 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7120 perror when PEM_read_bio_X509_REQ fails, the error message must
7121 be obtained from the error queue.
7124 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7125 it in ERR_remove_state if appropriate, and change ERR_get_state
7126 accordingly to avoid race conditions (this is necessary because
7127 thread_hash is no longer constant once set).
7130 *) Bugfix for linux-elf makefile.one.
7133 *) RSA_get_default_method() will now cause a default
7134 RSA_METHOD to be chosen if one doesn't exist already.
7135 Previously this was only set during a call to RSA_new()
7136 or RSA_new_method(NULL) meaning it was possible for
7137 RSA_get_default_method() to return NULL.
7140 *) Added native name translation to the existing DSO code
7141 that will convert (if the flag to do so is set) filenames
7142 that are sufficiently small and have no path information
7143 into a canonical native form. Eg. "blah" converted to
7144 "libblah.so" or "blah.dll" etc.
7147 *) New function ERR_error_string_n(e, buf, len) which is like
7148 ERR_error_string(e, buf), but writes at most 'len' bytes
7149 including the 0 terminator. For ERR_error_string_n, 'buf'
7151 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7153 *) CONF library reworked to become more general. A new CONF
7154 configuration file reader "class" is implemented as well as a
7155 new functions (NCONF_*, for "New CONF") to handle it. The now
7156 old CONF_* functions are still there, but are reimplemented to
7157 work in terms of the new functions. Also, a set of functions
7158 to handle the internal storage of the configuration data is
7159 provided to make it easier to write new configuration file
7160 reader "classes" (I can definitely see something reading a
7161 configuration file in XML format, for example), called _CONF_*,
7162 or "the configuration storage API"...
7164 The new configuration file reading functions are:
7166 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7167 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7169 NCONF_default, NCONF_WIN32
7171 NCONF_dump_fp, NCONF_dump_bio
7173 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7174 NCONF_new creates a new CONF object. This works in the same way
7175 as other interfaces in OpenSSL, like the BIO interface.
7176 NCONF_dump_* dump the internal storage of the configuration file,
7177 which is useful for debugging. All other functions take the same
7178 arguments as the old CONF_* functions wth the exception of the
7179 first that must be a `CONF *' instead of a `LHASH *'.
7181 To make it easer to use the new classes with the old CONF_* functions,
7182 the function CONF_set_default_method is provided.
7185 *) Add '-tls1' option to 'openssl ciphers', which was already
7186 mentioned in the documentation but had not been implemented.
7187 (This option is not yet really useful because even the additional
7188 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7191 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7192 OpenSSL-based applications) load shared libraries and bind to
7193 them in a portable way.
7194 [Geoff Thorpe, with contributions from Richard Levitte]
7196 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7198 *) Make sure _lrotl and _lrotr are only used with MSVC.
7200 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7201 (the default implementation of RAND_status).
7203 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7204 to '-clrext' (= clear extensions), as intended and documented.
7205 [Bodo Moeller; inconsistency pointed out by Michael Attili
7208 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7209 was larger than the MD block size.
7210 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7212 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7213 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7214 using the passed key: if the passed key was a private key the result
7215 of X509_print(), for example, would be to print out all the private key
7219 *) des_quad_cksum() byte order bug fix.
7220 [Ulf Möller, using the problem description in krb4-0.9.7, where
7221 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7223 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7225 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7227 *) For easily testing in shell scripts whether some command
7228 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7229 returns with exit code 0 iff no command of the given name is available.
7230 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7231 the output goes to stdout and nothing is printed to stderr.
7232 Additional arguments are always ignored.
7234 Since for each cipher there is a command of the same name,
7235 the 'no-cipher' compilation switches can be tested this way.
7237 ('openssl no-XXX' is not able to detect pseudo-commands such
7238 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7241 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7244 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7245 is set; it will be thrown away anyway because each handshake creates
7247 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7248 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7249 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7250 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7253 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7254 'Q' and 'R' lose their special meanings (quit/renegotiate).
7255 This is part of what -quiet does; unlike -quiet, -ign_eof
7256 does not suppress any output.
7259 *) Add compatibility options to the purpose and trust code. The
7260 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7261 accepts a certificate or CA, this was the previous behaviour,
7262 with all the associated security issues.
7264 X509_TRUST_COMPAT is the old trust behaviour: only and
7265 automatically trust self signed roots in certificate store. A
7266 new trust setting X509_TRUST_DEFAULT is used to specify that
7267 a purpose has no associated trust setting and it should instead
7268 use the value in the default purpose.
7271 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7272 and fix a memory leak.
7275 *) In util/mkerr.pl (which implements 'make errors'), preserve
7276 reason strings from the previous version of the .c file, as
7277 the default to have only downcase letters (and digits) in
7278 automatically generated reasons codes is not always appropriate.
7281 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7282 using strerror. Previously, ERR_reason_error_string() returned
7283 library names as reason strings for SYSerr; but SYSerr is a special
7284 case where small numbers are errno values, not library numbers.
7287 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7288 converts DSA parameters into DH parameters. (When creating parameters,
7289 DSA_generate_parameters is used.)
7292 *) Include 'length' (recommended exponent length) in C code generated
7293 by 'openssl dhparam -C'.
7296 *) The second argument to set_label in perlasm was already being used
7297 so couldn't be used as a "file scope" flag. Moved to third argument
7301 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7302 instead of RAND_bytes for encryption IVs and salts.
7305 *) Include RAND_status() into RAND_METHOD instead of implementing
7306 it only for md_rand.c Otherwise replacing the PRNG by calling
7307 RAND_set_rand_method would be impossible.
7310 *) Don't let DSA_generate_key() enter an infinite loop if the random
7311 number generation fails.
7314 *) New 'rand' application for creating pseudo-random output.
7317 *) Added configuration support for Linux/IA64
7318 [Rolf Haberrecker <rolf@suse.de>]
7320 *) Assembler module support for Mingw32.
7323 *) Shared library support for HPUX (in shlib/).
7324 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7326 *) Shared library support for Solaris gcc.
7327 [Lutz Behnke <behnke@trustcenter.de>]
7329 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7331 *) PKCS7_encrypt() was adding text MIME headers twice because they
7332 were added manually and by SMIME_crlf_copy().
7335 *) In bntest.c don't call BN_rand with zero bits argument.
7336 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7338 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7339 case was implemented. This caused BN_div_recp() to fail occasionally.
7342 *) Add an optional second argument to the set_label() in the perl
7343 assembly language builder. If this argument exists and is set
7344 to 1 it signals that the assembler should use a symbol whose
7345 scope is the entire file, not just the current function. This
7346 is needed with MASM which uses the format label:: for this scope.
7347 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7349 *) Change the ASN1 types so they are typedefs by default. Before
7350 almost all types were #define'd to ASN1_STRING which was causing
7351 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7355 *) Change names of new functions to the new get1/get0 naming
7356 convention: After 'get1', the caller owns a reference count
7357 and has to call ..._free; 'get0' returns a pointer to some
7358 data structure without incrementing reference counters.
7359 (Some of the existing 'get' functions increment a reference
7360 counter, some don't.)
7361 Similarly, 'set1' and 'add1' functions increase reference
7362 counters or duplicate objects.
7365 *) Allow for the possibility of temp RSA key generation failure:
7366 the code used to assume it always worked and crashed on failure.
7369 *) Fix potential buffer overrun problem in BIO_printf().
7370 [Ulf Möller, using public domain code by Patrick Powell; problem
7371 pointed out by David Sacerdote <das33@cornell.edu>]
7373 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7374 RAND_egd() and RAND_status(). In the command line application,
7375 the EGD socket can be specified like a seed file using RANDFILE
7379 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7380 Some CAs (e.g. Verisign) distribute certificates in this form.
7383 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7384 list to exclude them. This means that no special compilation option
7385 is needed to use anonymous DH: it just needs to be included in the
7389 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7390 EVP_MD_type. The old functionality is available in a new macro called
7391 EVP_MD_md(). Change code that uses it and update docs.
7394 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7395 where the 'void *' argument is replaced by a function pointer argument.
7396 Previously 'void *' was abused to point to functions, which works on
7397 many platforms, but is not correct. As these functions are usually
7398 called by macros defined in OpenSSL header files, most source code
7399 should work without changes.
7402 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7403 sections with information on -D... compiler switches used for
7404 compiling the library so that applications can see them. To enable
7405 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7406 must be defined. E.g.,
7407 #define OPENSSL_ALGORITHM_DEFINES
7408 #include <openssl/opensslconf.h>
7409 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7410 [Richard Levitte, Ulf and Bodo Möller]
7412 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7416 *) Change the 'other' type in certificate aux info to a STACK_OF
7417 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7418 the required ASN1 format: arbitrary types determined by an OID.
7421 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7422 argument to 'req'. This is not because the function is newer or
7423 better than others it just uses the work 'NEW' in the certificate
7424 request header lines. Some software needs this.
7427 *) Reorganise password command line arguments: now passwords can be
7428 obtained from various sources. Delete the PEM_cb function and make
7429 it the default behaviour: i.e. if the callback is NULL and the
7430 usrdata argument is not NULL interpret it as a null terminated pass
7431 phrase. If usrdata and the callback are NULL then the pass phrase
7432 is prompted for as usual.
7435 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7436 the support is automatically enabled. The resulting binaries will
7437 autodetect the card and use it if present.
7438 [Ben Laurie and Compaq Inc.]
7440 *) Work around for Netscape hang bug. This sends certificate request
7441 and server done in one record. Since this is perfectly legal in the
7442 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7443 the bugs/SSLv3 entry for more info.
7446 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7449 *) Add -rand argument to smime and pkcs12 applications and read/write
7453 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7456 *) Add command line password options to the remaining applications.
7459 *) Bug fix for BN_div_recp() for numerators with an even number of
7463 *) More tests in bntest.c, and changed test_bn output.
7466 *) ./config recognizes MacOS X now.
7469 *) Bug fix for BN_div() when the first words of num and divsor are
7470 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7473 *) Add support for various broken PKCS#8 formats, and command line
7474 options to produce them.
7477 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7478 get temporary BIGNUMs from a BN_CTX.
7481 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7485 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7486 include a #define from the old name to the new. The original intent
7487 was that statically linked binaries could for example just call
7488 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7489 link with digests. This never worked becayse SSLeay_add_all_digests()
7490 and SSLeay_add_all_ciphers() were in the same source file so calling
7491 one would link with the other. They are now in separate source files.
7494 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7497 *) Use a less unusual form of the Miller-Rabin primality test (it used
7498 a binary algorithm for exponentiation integrated into the Miller-Rabin
7499 loop, our standard modexp algorithms are faster).
7502 *) Support for the EBCDIC character set completed.
7503 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7505 *) Source code cleanups: use const where appropriate, eliminate casts,
7506 use void * instead of char * in lhash.
7509 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7510 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7511 this the server could overwrite ephemeral keys that the client
7515 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7516 using 50 iterations of the Rabin-Miller test.
7518 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7519 iterations of the Rabin-Miller test as required by the appendix
7520 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7521 As BN_is_prime_fasttest includes trial division, DSA parameter
7522 generation becomes much faster.
7524 This implies a change for the callback functions in DSA_is_prime
7525 and DSA_generate_parameters: The callback function is called once
7526 for each positive witness in the Rabin-Miller test, not just
7527 occasionally in the inner loop; and the parameters to the
7528 callback function now provide an iteration count for the outer
7529 loop rather than for the current invocation of the inner loop.
7530 DSA_generate_parameters additionally can call the callback
7531 function with an 'iteration count' of -1, meaning that a
7532 candidate has passed the trial division test (when q is generated
7533 from an application-provided seed, trial division is skipped).
7536 *) New function BN_is_prime_fasttest that optionally does trial
7537 division before starting the Rabin-Miller test and has
7538 an additional BN_CTX * argument (whereas BN_is_prime always
7539 has to allocate at least one BN_CTX).
7540 'callback(1, -1, cb_arg)' is called when a number has passed the
7541 trial division stage.
7544 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7548 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7551 *) New function BN_pseudo_rand().
7554 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7555 bignum version of BN_from_montgomery() with the working code from
7556 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7560 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7561 made it impossible to use the same SSL_SESSION data structure in
7562 SSL2 clients in multiple threads.
7565 *) The return value of RAND_load_file() no longer counts bytes obtained
7566 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7567 to seed the PRNG (previously an explicit byte count was required).
7568 [Ulf Möller, Bodo Möller]
7570 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7571 used (char *) instead of (void *) and had casts all over the place.
7574 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7577 *) Retain source code compatibility for BN_prime_checks macro:
7578 BN_is_prime(..., BN_prime_checks, ...) now uses
7579 BN_prime_checks_for_size to determine the appropriate number of
7580 Rabin-Miller iterations.
7583 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7584 DH_CHECK_P_NOT_SAFE_PRIME.
7585 (Check if this is true? OpenPGP calls them "strong".)
7588 *) Merge the functionality of "dh" and "gendh" programs into a new program
7589 "dhparam". The old programs are retained for now but will handle DH keys
7590 (instead of parameters) in future.
7593 *) Make the ciphers, s_server and s_client programs check the return values
7594 when a new cipher list is set.
7597 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7598 ciphers. Before when the 56bit ciphers were enabled the sorting was
7601 The syntax for the cipher sorting has been extended to support sorting by
7602 cipher-strength (using the strength_bits hard coded in the tables).
7603 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7605 Fix a bug in the cipher-command parser: when supplying a cipher command
7606 string with an "undefined" symbol (neither command nor alphanumeric
7607 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7608 an error is flagged.
7610 Due to the strength-sorting extension, the code of the
7611 ssl_create_cipher_list() function was completely rearranged. I hope that
7612 the readability was also increased :-)
7613 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7615 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7616 for the first serial number and places 2 in the serial number file. This
7617 avoids problems when the root CA is created with serial number zero and
7618 the first user certificate has the same issuer name and serial number
7622 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7623 the new code. Add documentation for this stuff.
7626 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7627 X509_*() to X509at_*() on the grounds that they don't handle X509
7628 structures and behave in an analagous way to the X509v3 functions:
7629 they shouldn't be called directly but wrapper functions should be used
7632 So we also now have some wrapper functions that call the X509at functions
7633 when passed certificate requests. (TO DO: similar things can be done with
7634 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7635 things. Some of these need some d2i or i2d and print functionality
7636 because they handle more complex structures.)
7639 *) Add missing #ifndefs that caused missing symbols when building libssl
7640 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7641 NO_RSA in ssl/s2*.c.
7642 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7644 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7645 has a return value which indicates the quality of the random data
7646 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7647 error queue. New function RAND_pseudo_bytes() generates output that is
7648 guaranteed to be unique but not unpredictable. RAND_add is like
7649 RAND_seed, but takes an extra argument for an entropy estimate
7650 (RAND_seed always assumes full entropy).
7653 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7654 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7655 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7656 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7657 false-positive rate of at most 2^-80 for random input.
7660 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7663 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7664 in the 0.9.5 release), this returns the chain
7665 from an X509_CTX structure with a dup of the stack and all
7666 the X509 reference counts upped: so the stack will exist
7667 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7670 Also make SSL_SESSION_print() print out the verify return
7674 *) Add manpage for the pkcs12 command. Also change the default
7675 behaviour so MAC iteration counts are used unless the new
7676 -nomaciter option is used. This improves file security and
7677 only older versions of MSIE (4.0 for example) need it.
7680 *) Honor the no-xxx Configure options when creating .DEF files.
7683 *) Add PKCS#10 attributes to field table: challengePassword,
7684 unstructuredName and unstructuredAddress. These are taken from
7685 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7686 international characters are used.
7688 More changes to X509_ATTRIBUTE code: allow the setting of types
7689 based on strings. Remove the 'loc' parameter when adding
7690 attributes because these will be a SET OF encoding which is sorted
7694 *) Initial changes to the 'req' utility to allow request generation
7695 automation. This will allow an application to just generate a template
7696 file containing all the field values and have req construct the
7699 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7700 used all over the place including certificate requests and PKCS#7
7701 structures. They are currently handled manually where necessary with
7702 some primitive wrappers for PKCS#7. The new functions behave in a
7703 manner analogous to the X509 extension functions: they allow
7704 attributes to be looked up by NID and added.
7706 Later something similar to the X509V3 code would be desirable to
7707 automatically handle the encoding, decoding and printing of the
7708 more complex types. The string types like challengePassword can
7709 be handled by the string table functions.
7711 Also modified the multi byte string table handling. Now there is
7712 a 'global mask' which masks out certain types. The table itself
7713 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7714 is useful when for example there is only one permissible type
7715 (as in countryName) and using the mask might result in no valid
7719 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7720 SSL_get_peer_finished to allow applications to obtain the latest
7721 Finished messages sent to the peer or expected from the peer,
7722 respectively. (SSL_get_peer_finished is usually the Finished message
7723 actually received from the peer, otherwise the protocol will be aborted.)
7725 As the Finished message are message digests of the complete handshake
7726 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7727 be used for external authentication procedures when the authentication
7728 provided by SSL/TLS is not desired or is not enough.
7731 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7732 the host supports BWX extension and if Compaq C is present on the
7733 $PATH. Just exploiting of the BWX extension results in 20-30%
7734 performance kick for some algorithms, e.g. DES and RC4 to mention
7735 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7739 *) Add support for MS "fast SGC". This is arguably a violation of the
7740 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7741 weak crypto and after checking the certificate is SGC a second one
7742 with strong crypto. MS SGC stops the first handshake after receiving
7743 the server certificate message and sends a second client hello. Since
7744 a server will typically do all the time consuming operations before
7745 expecting any further messages from the client (server key exchange
7746 is the most expensive) there is little difference between the two.
7748 To get OpenSSL to support MS SGC we have to permit a second client
7749 hello message after we have sent server done. In addition we have to
7750 reset the MAC if we do get this second client hello.
7753 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7754 if a DER encoded private key is RSA or DSA traditional format. Changed
7755 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7756 format DER encoded private key. Newer code should use PKCS#8 format which
7757 has the key type encoded in the ASN1 structure. Added DER private key
7758 support to pkcs8 application.
7761 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7762 ciphersuites has been selected (as required by the SSL 3/TLS 1
7763 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7764 is set, we interpret this as a request to violate the specification
7765 (the worst that can happen is a handshake failure, and 'correct'
7766 behaviour would result in a handshake failure anyway).
7769 *) In SSL_CTX_add_session, take into account that there might be multiple
7770 SSL_SESSION structures with the same session ID (e.g. when two threads
7771 concurrently obtain them from an external cache).
7772 The internal cache can handle only one SSL_SESSION with a given ID,
7773 so if there's a conflict, we now throw out the old one to achieve
7777 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7778 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7779 some routines that use cipher OIDs: some ciphers do not have OIDs
7780 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7784 *) Simplify the trust setting structure and code. Now we just have
7785 two sequences of OIDs for trusted and rejected settings. These will
7786 typically have values the same as the extended key usage extension
7787 and any application specific purposes.
7789 The trust checking code now has a default behaviour: it will just
7790 check for an object with the same NID as the passed id. Functions can
7791 be provided to override either the default behaviour or the behaviour
7792 for a given id. SSL client, server and email already have functions
7793 in place for compatibility: they check the NID and also return "trusted"
7794 if the certificate is self signed.
7797 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7798 traditional format into an EVP_PKEY structure.
7801 *) Add a password callback function PEM_cb() which either prompts for
7802 a password if usr_data is NULL or otherwise assumes it is a null
7803 terminated password. Allow passwords to be passed on command line
7804 environment or config files in a few more utilities.
7807 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7808 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7809 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7810 Update documentation.
7813 *) Support for ASN1 "NULL" type. This could be handled before by using
7814 ASN1_TYPE but there wasn't any function that would try to read a NULL
7815 and produce an error if it couldn't. For compatibility we also have
7816 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7817 don't allocate anything because they don't need to.
7820 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7822 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7824 *) Rebuild of the memory allocation routines used by OpenSSL code and
7825 possibly others as well. The purpose is to make an interface that
7826 provide hooks so anyone can build a separate set of allocation and
7827 deallocation routines to be used by OpenSSL, for example memory
7828 pool implementations, or something else, which was previously hard
7829 since Malloc(), Realloc() and Free() were defined as macros having
7830 the values malloc, realloc and free, respectively (except for Win32
7831 compilations). The same is provided for memory debugging code.
7832 OpenSSL already comes with functionality to find memory leaks, but
7833 this gives people a chance to debug other memory problems.
7835 With these changes, a new set of functions and macros have appeared:
7837 CRYPTO_set_mem_debug_functions() [F]
7838 CRYPTO_get_mem_debug_functions() [F]
7839 CRYPTO_dbg_set_options() [F]
7840 CRYPTO_dbg_get_options() [F]
7841 CRYPTO_malloc_debug_init() [M]
7843 The memory debug functions are NULL by default, unless the library
7844 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7845 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7846 gives the standard debugging functions that come with OpenSSL) or
7847 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7848 provided by the library user) must be used. When the standard
7849 debugging functions are used, CRYPTO_dbg_set_options can be used to
7850 request additional information:
7851 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7852 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7854 Also, things like CRYPTO_set_mem_functions will always give the
7855 expected result (the new set of functions is used for allocation
7856 and deallocation) at all times, regardless of platform and compiler
7859 To finish it up, some functions that were never use in any other
7860 way than through macros have a new API and new semantic:
7863 CRYPTO_dbg_realloc()
7866 All macros of value have retained their old syntax.
7867 [Richard Levitte and Bodo Moeller]
7869 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7870 ordering of SMIMECapabilities wasn't in "strength order" and there
7871 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7875 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7876 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7877 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7879 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7880 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7881 functionality to handle multipart/signed properly) and a utility
7882 called 'smime' to call all this stuff. This is based on code I
7883 originally wrote for Celo who have kindly allowed it to be
7884 included in OpenSSL.
7887 *) Add variants des_set_key_checked and des_set_key_unchecked of
7888 des_set_key (aka des_key_sched). Global variable des_check_key
7889 decides which of these is called by des_set_key; this way
7890 des_check_key behaves as it always did, but applications and
7891 the library itself, which was buggy for des_check_key == 1,
7892 have a cleaner way to pick the version they need.
7895 *) New function PKCS12_newpass() which changes the password of a
7899 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7900 dynamic mix. In both cases the ids can be used as an index into the
7901 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7902 functions so they accept a list of the field values and the
7903 application doesn't need to directly manipulate the X509_TRUST
7907 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7911 *) Modify the way the V3 extension code looks up extensions. This now
7912 works in a similar way to the object code: we have some "standard"
7913 extensions in a static table which is searched with OBJ_bsearch()
7914 and the application can add dynamic ones if needed. The file
7915 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7916 updated whenever a new extension is added to the core code and kept
7917 in ext_nid order. There is a simple program 'tabtest.c' which checks
7918 this. New extensions are not added too often so this file can readily
7919 be maintained manually.
7921 There are two big advantages in doing things this way. The extensions
7922 can be looked up immediately and no longer need to be "added" using
7923 X509V3_add_standard_extensions(): this function now does nothing.
7924 [Side note: I get *lots* of email saying the extension code doesn't
7925 work because people forget to call this function]
7926 Also no dynamic allocation is done unless new extensions are added:
7927 so if we don't add custom extensions there is no need to call
7928 X509V3_EXT_cleanup().
7931 *) Modify enc utility's salting as follows: make salting the default. Add a
7932 magic header, so unsalted files fail gracefully instead of just decrypting
7933 to garbage. This is because not salting is a big security hole, so people
7934 should be discouraged from doing it.
7937 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7938 digest to be passed on the command line but it only used this
7939 parameter when signing a certificate. Modified so all relevant
7940 operations are affected by the digest parameter including the
7941 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7942 DSA key was used because it didn't fix the digest.
7945 *) Initial certificate chain verify code. Currently tests the untrusted
7946 certificates for consistency with the verify purpose (which is set
7947 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7949 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7950 this is because it will reject chains with invalid extensions whereas
7951 every previous version of OpenSSL and SSLeay made no checks at all.
7953 Trust code: checks the root CA for the relevant trust settings. Trust
7954 settings have an initial value consistent with the verify purpose: e.g.
7955 if the verify purpose is for SSL client use it expects the CA to be
7956 trusted for SSL client use. However the default value can be changed to
7957 permit custom trust settings: one example of this would be to only trust
7958 certificates from a specific "secure" set of CAs.
7960 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7961 which should be used for version portability: especially since the
7962 verify structure is likely to change more often now.
7964 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7965 to set them. If not set then assume SSL clients will verify SSL servers
7968 Two new options to the verify program: -untrusted allows a set of
7969 untrusted certificates to be passed in and -purpose which sets the
7970 intended purpose of the certificate. If a purpose is set then the
7971 new chain verify code is used to check extension consistency.
7974 *) Support for the authority information access extension.
7977 *) Modify RSA and DSA PEM read routines to transparently handle
7978 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7979 public keys in a format compatible with certificate
7980 SubjectPublicKeyInfo structures. Unfortunately there were already
7981 functions called *_PublicKey_* which used various odd formats so
7982 these are retained for compatibility: however the DSA variants were
7983 never in a public release so they have been deleted. Changed dsa/rsa
7984 utilities to handle the new format: note no releases ever handled public
7985 keys so we should be OK.
7987 The primary motivation for this change is to avoid the same fiasco
7988 that dogs private keys: there are several incompatible private key
7989 formats some of which are standard and some OpenSSL specific and
7990 require various evil hacks to allow partial transparent handling and
7991 even then it doesn't work with DER formats. Given the option anything
7992 other than PKCS#8 should be dumped: but the other formats have to
7993 stay in the name of compatibility.
7995 With public keys and the benefit of hindsight one standard format
7996 is used which works with EVP_PKEY, RSA or DSA structures: though
7997 it clearly returns an error if you try to read the wrong kind of key.
7999 Added a -pubkey option to the 'x509' utility to output the public key.
8000 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8001 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8002 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8003 that do the same as the EVP_PKEY_assign_*() except they up the
8004 reference count of the added key (they don't "swallow" the
8008 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8009 CRLs would fail if the file contained no certificates or no CRLs:
8010 added a new function to read in both types and return the number
8011 read: this means that if none are read it will be an error. The
8012 DER versions of the certificate and CRL reader would always fail
8013 because it isn't possible to mix certificates and CRLs in DER format
8014 without choking one or the other routine. Changed this to just read
8015 a certificate: this is the best we can do. Also modified the code
8016 in apps/verify.c to take notice of return codes: it was previously
8017 attempting to read in certificates from NULL pointers and ignoring
8018 any errors: this is one reason why the cert and CRL reader seemed
8019 to work. It doesn't check return codes from the default certificate
8020 routines: these may well fail if the certificates aren't installed.
8023 *) Code to support otherName option in GeneralName.
8026 *) First update to verify code. Change the verify utility
8027 so it warns if it is passed a self signed certificate:
8028 for consistency with the normal behaviour. X509_verify
8029 has been modified to it will now verify a self signed
8030 certificate if *exactly* the same certificate appears
8031 in the store: it was previously impossible to trust a
8032 single self signed certificate. This means that:
8033 openssl verify ss.pem
8034 now gives a warning about a self signed certificate but
8035 openssl verify -CAfile ss.pem ss.pem
8039 *) For servers, store verify_result in SSL_SESSION data structure
8040 (and add it to external session representation).
8041 This is needed when client certificate verifications fails,
8042 but an application-provided verification callback (set by
8043 SSL_CTX_set_cert_verify_callback) allows accepting the session
8044 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8045 but returns 1): When the session is reused, we have to set
8046 ssl->verify_result to the appropriate error code to avoid
8048 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8050 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8051 case in PKCS7_dataInit() where the signed PKCS7 structure
8052 didn't contain any existing data because it was being created.
8053 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8055 *) Add a salt to the key derivation routines in enc.c. This
8056 forms the first 8 bytes of the encrypted file. Also add a
8057 -S option to allow a salt to be input on the command line.
8060 *) New function X509_cmp(). Oddly enough there wasn't a function
8061 to compare two certificates. We do this by working out the SHA1
8062 hash and comparing that. X509_cmp() will be needed by the trust
8066 *) SSL_get1_session() is like SSL_get_session(), but increments
8067 the reference count in the SSL_SESSION returned.
8068 [Geoff Thorpe <geoff@eu.c2.net>]
8070 *) Fix for 'req': it was adding a null to request attributes.
8071 Also change the X509_LOOKUP and X509_INFO code to handle
8072 certificate auxiliary information.
8075 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8079 *) Add the possibility to add extra information to the memory leak
8080 detecting output, to form tracebacks, showing from where each
8081 allocation was originated: CRYPTO_push_info("constant string") adds
8082 the string plus current file name and line number to a per-thread
8083 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8084 is like calling CYRPTO_pop_info() until the stack is empty.
8085 Also updated memory leak detection code to be multi-thread-safe.
8088 *) Add options -text and -noout to pkcs7 utility and delete the
8089 encryption options which never did anything. Update docs.
8092 *) Add options to some of the utilities to allow the pass phrase
8093 to be included on either the command line (not recommended on
8094 OSes like Unix) or read from the environment. Update the
8095 manpages and fix a few bugs.
8098 *) Add a few manpages for some of the openssl commands.
8101 *) Fix the -revoke option in ca. It was freeing up memory twice,
8102 leaking and not finding already revoked certificates.
8105 *) Extensive changes to support certificate auxiliary information.
8106 This involves the use of X509_CERT_AUX structure and X509_AUX
8107 functions. An X509_AUX function such as PEM_read_X509_AUX()
8108 can still read in a certificate file in the usual way but it
8109 will also read in any additional "auxiliary information". By
8110 doing things this way a fair degree of compatibility can be
8111 retained: existing certificates can have this information added
8112 using the new 'x509' options.
8114 Current auxiliary information includes an "alias" and some trust
8115 settings. The trust settings will ultimately be used in enhanced
8116 certificate chain verification routines: currently a certificate
8117 can only be trusted if it is self signed and then it is trusted
8121 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8122 The problem was that one of the replacement routines had not been working
8123 since SSLeay releases. For now the offending routine has been replaced
8124 with non-optimised assembler. Even so, this now gives around 95%
8125 performance improvement for 1024 bit RSA signs.
8128 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8129 handling. Most clients have the effective key size in bits equal to
8130 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8131 A few however don't do this and instead use the size of the decrypted key
8132 to determine the RC2 key length and the AlgorithmIdentifier to determine
8133 the effective key length. In this case the effective key length can still
8134 be 40 bits but the key length can be 168 bits for example. This is fixed
8135 by manually forcing an RC2 key into the EVP_PKEY structure because the
8136 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8137 the key length and effective key length are equal.
8140 *) Add a bunch of functions that should simplify the creation of
8141 X509_NAME structures. Now you should be able to do:
8142 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8143 and have it automatically work out the correct field type and fill in
8144 the structures. The more adventurous can try:
8145 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8146 and it will (hopefully) work out the correct multibyte encoding.
8149 *) Change the 'req' utility to use the new field handling and multibyte
8150 copy routines. Before the DN field creation was handled in an ad hoc
8151 way in req, ca, and x509 which was rather broken and didn't support
8152 BMPStrings or UTF8Strings. Since some software doesn't implement
8153 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8154 using the dirstring_type option. See the new comment in the default
8155 openssl.cnf for more info.
8158 *) Make crypto/rand/md_rand.c more robust:
8159 - Assure unique random numbers after fork().
8160 - Make sure that concurrent threads access the global counter and
8161 md serializably so that we never lose entropy in them
8162 or use exactly the same state in multiple threads.
8163 Access to the large state is not always serializable because
8164 the additional locking could be a performance killer, and
8165 md should be large enough anyway.
8168 *) New file apps/app_rand.c with commonly needed functionality
8169 for handling the random seed file.
8171 Use the random seed file in some applications that previously did not:
8173 dsaparam -genkey (which also ignored its '-rand' option),
8176 x509 (when signing).
8177 Except on systems with /dev/urandom, it is crucial to have a random
8178 seed file at least for key creation, DSA signing, and for DH exchanges;
8179 for RSA signatures we could do without one.
8181 gendh and gendsa (unlike genrsa) used to read only the first byte
8182 of each file listed in the '-rand' option. The function as previously
8183 found in genrsa is now in app_rand.c and is used by all programs
8184 that support '-rand'.
8187 *) In RAND_write_file, use mode 0600 for creating files;
8188 don't just chmod when it may be too late.
8191 *) Report an error from X509_STORE_load_locations
8192 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8195 *) New function ASN1_mbstring_copy() this copies a string in either
8196 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8197 into an ASN1_STRING type. A mask of permissible types is passed
8198 and it chooses the "minimal" type to use or an error if not type
8202 *) Add function equivalents to the various macros in asn1.h. The old
8203 macros are retained with an M_ prefix. Code inside the library can
8204 use the M_ macros. External code (including the openssl utility)
8205 should *NOT* in order to be "shared library friendly".
8208 *) Add various functions that can check a certificate's extensions
8209 to see if it usable for various purposes such as SSL client,
8210 server or S/MIME and CAs of these types. This is currently
8211 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8212 verification. Also added a -purpose flag to x509 utility to
8213 print out all the purposes.
8216 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8220 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8221 for, obtain and decode and extension and obtain its critical flag.
8222 This allows all the necessary extension code to be handled in a
8223 single function call.
8226 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8227 platforms. See crypto/rc4/rc4_enc.c for further details.
8230 *) New -noout option to asn1parse. This causes no output to be produced
8231 its main use is when combined with -strparse and -out to extract data
8232 from a file (which may not be in ASN.1 format).
8235 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8236 when producing the local key id.
8237 [Richard Levitte <levitte@stacken.kth.se>]
8239 *) New option -dhparam in s_server. This allows a DH parameter file to be
8240 stated explicitly. If it is not stated then it tries the first server
8241 certificate file. The previous behaviour hard coded the filename
8245 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8246 a public key to be input or output. For example:
8247 openssl rsa -in key.pem -pubout -out pubkey.pem
8248 Also added necessary DSA public key functions to handle this.
8251 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8252 in the message. This was handled by allowing
8253 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8254 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8256 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8257 to the end of the strings whereas this didn't. This would cause problems
8258 if strings read with d2i_ASN1_bytes() were later modified.
8259 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8261 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8262 data and it contains EOF it will end up returning an error. This is
8263 caused by input 46 bytes long. The cause is due to the way base64
8264 BIOs find the start of base64 encoded data. They do this by trying a
8265 trial decode on each line until they find one that works. When they
8266 do a flag is set and it starts again knowing it can pass all the
8267 data directly through the decoder. Unfortunately it doesn't reset
8268 the context it uses. This means that if EOF is reached an attempt
8269 is made to pass two EOFs through the context and this causes the
8270 resulting error. This can also cause other problems as well. As is
8271 usual with these problems it takes *ages* to find and the fix is
8272 trivial: move one line.
8273 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8275 *) Ugly workaround to get s_client and s_server working under Windows. The
8276 old code wouldn't work because it needed to select() on sockets and the
8277 tty (for keypresses and to see if data could be written). Win32 only
8278 supports select() on sockets so we select() with a 1s timeout on the
8279 sockets and then see if any characters are waiting to be read, if none
8280 are present then we retry, we also assume we can always write data to
8281 the tty. This isn't nice because the code then blocks until we've
8282 received a complete line of data and it is effectively polling the
8283 keyboard at 1s intervals: however it's quite a bit better than not
8284 working at all :-) A dedicated Windows application might handle this
8285 with an event loop for example.
8288 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8289 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8290 will be called when RSA_sign() and RSA_verify() are used. This is useful
8291 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8292 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8293 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8294 This necessitated the support of an extra signature type NID_md5_sha1
8295 for SSL signatures and modifications to the SSL library to use it instead
8296 of calling RSA_public_decrypt() and RSA_private_encrypt().
8299 *) Add new -verify -CAfile and -CApath options to the crl program, these
8300 will lookup a CRL issuers certificate and verify the signature in a
8301 similar way to the verify program. Tidy up the crl program so it
8302 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8303 less strict. It will now permit CRL extensions even if it is not
8304 a V2 CRL: this will allow it to tolerate some broken CRLs.
8307 *) Initialize all non-automatic variables each time one of the openssl
8308 sub-programs is started (this is necessary as they may be started
8309 multiple times from the "OpenSSL>" prompt).
8310 [Lennart Bang, Bodo Moeller]
8312 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8313 removing all other RSA functionality (this is what NO_RSA does). This
8314 is so (for example) those in the US can disable those operations covered
8315 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8319 *) Non-copying interface to BIO pairs.
8320 (still largely untested)
8323 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8324 ASCII string. This was handled independently in various places before.
8327 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8328 UTF8 strings a character at a time.
8331 *) Use client_version from client hello to select the protocol
8332 (s23_srvr.c) and for RSA client key exchange verification
8333 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8336 *) Add various utility functions to handle SPKACs, these were previously
8337 handled by poking round in the structure internals. Added new function
8338 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8339 print, verify and generate SPKACs. Based on an original idea from
8340 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8343 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8346 *) Allow the config file extension section to be overwritten on the
8347 command line. Based on an original idea from Massimiliano Pala
8348 <madwolf@comune.modena.it>. The new option is called -extensions
8349 and can be applied to ca, req and x509. Also -reqexts to override
8350 the request extensions in req and -crlexts to override the crl extensions
8354 *) Add new feature to the SPKAC handling in ca. Now you can include
8355 the same field multiple times by preceding it by "XXXX." for example:
8358 this is the same syntax as used in the req config file.
8361 *) Allow certificate extensions to be added to certificate requests. These
8362 are specified in a 'req_extensions' option of the req section of the
8363 config file. They can be printed out with the -text option to req but
8364 are otherwise ignored at present.
8367 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8368 data read consists of only the final block it would not decrypted because
8369 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8370 A misplaced 'break' also meant the decrypted final block might not be
8371 copied until the next read.
8374 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8375 a few extra parameters to the DH structure: these will be useful if
8376 for example we want the value of 'q' or implement X9.42 DH.
8379 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8380 provides hooks that allow the default DSA functions or functions on a
8381 "per key" basis to be replaced. This allows hardware acceleration and
8382 hardware key storage to be handled without major modification to the
8383 library. Also added low level modexp hooks and CRYPTO_EX structure and
8384 associated functions.
8387 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8388 as "read only": it can't be written to and the buffer it points to will
8389 not be freed. Reading from a read only BIO is much more efficient than
8390 a normal memory BIO. This was added because there are several times when
8391 an area of memory needs to be read from a BIO. The previous method was
8392 to create a memory BIO and write the data to it, this results in two
8393 copies of the data and an O(n^2) reading algorithm. There is a new
8394 function BIO_new_mem_buf() which creates a read only memory BIO from
8395 an area of memory. Also modified the PKCS#7 routines to use read only
8399 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8400 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8401 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8402 but a retry condition occured while trying to read the rest.
8405 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8406 NID_pkcs7_encrypted by default: this was wrong since this should almost
8407 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8408 the encrypted data type: this is a more sensible place to put it and it
8409 allows the PKCS#12 code to be tidied up that duplicated this
8413 *) Changed obj_dat.pl script so it takes its input and output files on
8414 the command line. This should avoid shell escape redirection problems
8418 *) Initial support for certificate extension requests, these are included
8419 in things like Xenroll certificate requests. Included functions to allow
8420 extensions to be obtained and added.
8423 *) -crlf option to s_client and s_server for sending newlines as
8424 CRLF (as required by many protocols).
8427 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8429 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8430 [Ralf S. Engelschall]
8432 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8433 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8435 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8439 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8440 DH parameters/keys (q is lost during that conversion, but the resulting
8441 DH parameters contain its length).
8443 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8444 much faster than DH_generate_parameters (which creates parameters
8445 where p = 2*q + 1), and also the smaller q makes DH computations
8446 much more efficient (160-bit exponentiation instead of 1024-bit
8447 exponentiation); so this provides a convenient way to support DHE
8448 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8449 utter importance to use
8450 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8452 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8453 when such DH parameters are used, because otherwise small subgroup
8454 attacks may become possible!
8457 *) Avoid memory leak in i2d_DHparams.
8460 *) Allow the -k option to be used more than once in the enc program:
8461 this allows the same encrypted message to be read by multiple recipients.
8464 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8465 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8466 it will always use the numerical form of the OID, even if it has a short
8470 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8471 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8472 otherwise bn_mod_exp was called. In the case of hardware keys for example
8473 no private key components need be present and it might store extra data
8474 in the RSA structure, which cannot be accessed from bn_mod_exp.
8475 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8476 private key operations.
8479 *) Added support for SPARC Linux.
8482 *) pem_password_cb function type incompatibly changed from
8483 typedef int pem_password_cb(char *buf, int size, int rwflag);
8485 ....(char *buf, int size, int rwflag, void *userdata);
8486 so that applications can pass data to their callbacks:
8487 The PEM[_ASN1]_{read,write}... functions and macros now take an
8488 additional void * argument, which is just handed through whenever
8489 the password callback is called.
8490 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8492 New function SSL_CTX_set_default_passwd_cb_userdata.
8494 Compatibility note: As many C implementations push function arguments
8495 onto the stack in reverse order, the new library version is likely to
8496 interoperate with programs that have been compiled with the old
8497 pem_password_cb definition (PEM_whatever takes some data that
8498 happens to be on the stack as its last argument, and the callback
8499 just ignores this garbage); but there is no guarantee whatsoever that
8502 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8503 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8504 problems not only on Windows, but also on some Unix platforms.
8505 To avoid problematic command lines, these definitions are now in an
8506 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8507 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8510 *) MIPS III/IV assembler module is reimplemented.
8513 *) More DES library cleanups: remove references to srand/rand and
8514 delete an unused file.
8517 *) Add support for the the free Netwide assembler (NASM) under Win32,
8518 since not many people have MASM (ml) and it can be hard to obtain.
8519 This is currently experimental but it seems to work OK and pass all
8520 the tests. Check out INSTALL.W32 for info.
8523 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8524 without temporary keys kept an extra copy of the server key,
8525 and connections with temporary keys did not free everything in case
8529 *) New function RSA_check_key and new openssl rsa option -check
8530 for verifying the consistency of RSA keys.
8531 [Ulf Moeller, Bodo Moeller]
8533 *) Various changes to make Win32 compile work:
8534 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8535 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8536 comparison" warnings.
8537 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8540 *) Add a debugging option to PKCS#5 v2 key generation function: when
8541 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8542 derived keys are printed to stderr.
8545 *) Copy the flags in ASN1_STRING_dup().
8546 [Roman E. Pavlov <pre@mo.msk.ru>]
8548 *) The x509 application mishandled signing requests containing DSA
8549 keys when the signing key was also DSA and the parameters didn't match.
8551 It was supposed to omit the parameters when they matched the signing key:
8552 the verifying software was then supposed to automatically use the CA's
8553 parameters if they were absent from the end user certificate.
8555 Omitting parameters is no longer recommended. The test was also
8556 the wrong way round! This was probably due to unusual behaviour in
8557 EVP_cmp_parameters() which returns 1 if the parameters match.
8558 This meant that parameters were omitted when they *didn't* match and
8559 the certificate was useless. Certificates signed with 'ca' didn't have
8561 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8563 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8564 The interface is as follows:
8565 Applications can use
8566 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8567 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8568 "off" is now the default.
8569 The library internally uses
8570 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8571 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8572 to disable memory-checking temporarily.
8574 Some inconsistent states that previously were possible (and were
8575 even the default) are now avoided.
8577 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8578 with each memory chunk allocated; this is occasionally more helpful
8579 than just having a counter.
8581 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8583 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8587 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8588 which largely parallels "options", but is for changing API behaviour,
8589 whereas "options" are about protocol behaviour.
8590 Initial "mode" flags are:
8592 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8593 a single record has been written.
8594 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8595 retries use the same buffer location.
8596 (But all of the contents must be
8600 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8603 *) Fix problems with no-hmac etc.
8604 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8606 *) New functions RSA_get_default_method(), RSA_set_method() and
8607 RSA_get_method(). These allows replacement of RSA_METHODs without having
8608 to mess around with the internals of an RSA structure.
8611 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8612 Also really enable memory leak checks in openssl.c and in some
8614 [Chad C. Mulligan, Bodo Moeller]
8616 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8617 up the length of negative integers. This has now been simplified to just
8618 store the length when it is first determined and use it later, rather
8619 than trying to keep track of where data is copied and updating it to
8621 [Steve Henson, reported by Brien Wheeler
8622 <bwheeler@authentica-security.com>]
8624 *) Add a new function PKCS7_signatureVerify. This allows the verification
8625 of a PKCS#7 signature but with the signing certificate passed to the
8626 function itself. This contrasts with PKCS7_dataVerify which assumes the
8627 certificate is present in the PKCS#7 structure. This isn't always the
8628 case: certificates can be omitted from a PKCS#7 structure and be
8629 distributed by "out of band" means (such as a certificate database).
8632 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8633 function prototypes in pem.h, also change util/mkdef.pl to add the
8634 necessary function names.
8637 *) mk1mf.pl (used by Windows builds) did not properly read the
8638 options set by Configure in the top level Makefile, and Configure
8639 was not even able to write more than one option correctly.
8640 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8643 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8644 file to be loaded from a BIO or FILE pointer. The BIO version will
8645 for example allow memory BIOs to contain config info.
8648 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8649 Whoever hopes to achieve shared-library compatibility across versions
8650 must use this, not the compile-time macro.
8651 (Exercise 0.9.4: Which is the minimum library version required by
8653 Note: All this applies only to multi-threaded programs, others don't
8657 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8658 through a BIO pair triggered the default case, i.e.
8659 SSLerr(...,SSL_R_UNKNOWN_STATE).
8662 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8663 can use the SSL library even if none of the specific BIOs is
8667 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8668 for the encoded length.
8669 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8671 *) Add initial documentation of the X509V3 functions.
8674 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8675 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8676 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8677 secure PKCS#8 private key format with a high iteration count.
8680 *) Fix determination of Perl interpreter: A perl or perl5
8681 _directory_ in $PATH was also accepted as the interpreter.
8682 [Ralf S. Engelschall]
8684 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8685 wrong with it but it was very old and did things like calling
8686 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8690 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8691 to use the new extension code.
8694 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8695 with macros. This should make it easier to change their form, add extra
8696 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8700 *) Add to configuration table a new entry that can specify an alternative
8701 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8702 according to Mark Crispin <MRC@Panda.COM>.
8706 *) DES CBC did not update the IV. Weird.
8709 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8710 Changing the behaviour of the former might break existing programs --
8711 where IV updating is needed, des_ncbc_encrypt can be used.
8714 *) When bntest is run from "make test" it drives bc to check its
8715 calculations, as well as internally checking them. If an internal check
8716 fails, it needs to cause bc to give a non-zero result or make test carries
8717 on without noticing the failure. Fixed.
8720 *) DES library cleanups.
8723 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8724 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8725 ciphers. NOTE: although the key derivation function has been verified
8726 against some published test vectors it has not been extensively tested
8727 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8731 *) Instead of "mkdir -p", which is not fully portable, use new
8732 Perl script "util/mkdir-p.pl".
8735 *) Rewrite the way password based encryption (PBE) is handled. It used to
8736 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8737 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8738 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8739 the 'parameter' field of the AlgorithmIdentifier is passed to the
8740 underlying key generation function so it must do its own ASN1 parsing.
8741 This has also changed the EVP_PBE_CipherInit() function which now has a
8742 'parameter' argument instead of literal salt and iteration count values
8743 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8746 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8747 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8748 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8749 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8750 value was just used as a "magic string" and not used directly its
8751 value doesn't matter.
8754 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8758 *) "linux-sparc64" configuration (ultrapenguin).
8759 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8760 "linux-sparc" configuration.
8761 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8763 *) config now generates no-xxx options for missing ciphers.
8766 *) Support the EBCDIC character set (work in progress).
8767 File ebcdic.c not yet included because it has a different license.
8768 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8770 *) Support BS2000/OSD-POSIX.
8771 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8773 *) Make callbacks for key generation use void * instead of char *.
8776 *) Make S/MIME samples compile (not yet tested).
8779 *) Additional typesafe stacks.
8782 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8786 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8788 *) New configuration variant "sco5-gcc".
8790 *) Updated some demos.
8791 [Sean O Riordain, Wade Scholine]
8793 *) Add missing BIO_free at exit of pkcs12 application.
8796 *) Fix memory leak in conf.c.
8799 *) Updates for Win32 to assembler version of MD5.
8802 *) Set #! path to perl in apps/der_chop to where we found it
8803 instead of using a fixed path.
8806 *) SHA library changes for irix64-mips4-cc.
8809 *) Improvements for VMS support.
8813 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8815 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8816 This also avoids the problems with SC4.2 and unpatched SC5.
8817 [Andy Polyakov <appro@fy.chalmers.se>]
8819 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8820 These are required because of the typesafe stack would otherwise break
8821 existing code. If old code used a structure member which used to be STACK
8822 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8823 sk_num or sk_value it would produce an error because the num, data members
8824 are not present in STACK_OF. Now it just produces a warning. sk_set
8825 replaces the old method of assigning a value to sk_value
8826 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8827 that does this will no longer work (and should use sk_set instead) but
8828 this could be regarded as a "questionable" behaviour anyway.
8831 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8832 correctly handle encrypted S/MIME data.
8835 *) Change type of various DES function arguments from des_cblock
8836 (which means, in function argument declarations, pointer to char)
8837 to des_cblock * (meaning pointer to array with 8 char elements),
8838 which allows the compiler to do more typechecking; it was like
8839 that back in SSLeay, but with lots of ugly casts.
8841 Introduce new type const_des_cblock.
8844 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8845 problems: find RecipientInfo structure that matches recipient certificate
8846 and initialise the ASN1 structures properly based on passed cipher.
8849 *) Belatedly make the BN tests actually check the results.
8852 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8853 to and from BNs: it was completely broken. New compilation option
8854 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8855 key elements as negative integers.
8858 *) Reorganize and speed up MD5.
8859 [Andy Polyakov <appro@fy.chalmers.se>]
8862 [Richard Levitte <richard@levitte.org>]
8864 *) New option -out to asn1parse to allow the parsed structure to be
8865 output to a file. This is most useful when combined with the -strparse
8866 option to examine the output of things like OCTET STRINGS.
8869 *) Make SSL library a little more fool-proof by not requiring any longer
8870 that SSL_set_{accept,connect}_state be called before
8871 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8872 in many applications because usually everything *appeared* to work as
8873 intended anyway -- now it really works as intended).
8876 *) Move openssl.cnf out of lib/.
8879 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8880 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8881 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8882 [Ralf S. Engelschall]
8884 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8885 handle PKCS#7 enveloped data properly.
8886 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8888 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8889 copying pointers. The cert_st handling is changed by this in
8890 various ways (and thus what used to be known as ctx->default_cert
8891 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8892 any longer when s->cert does not give us what we need).
8893 ssl_cert_instantiate becomes obsolete by this change.
8894 As soon as we've got the new code right (possibly it already is?),
8895 we have solved a couple of bugs of the earlier code where s->cert
8896 was used as if it could not have been shared with other SSL structures.
8898 Note that using the SSL API in certain dirty ways now will result
8899 in different behaviour than observed with earlier library versions:
8900 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8901 does not influence s as it used to.
8903 In order to clean up things more thoroughly, inside SSL_SESSION
8904 we don't use CERT any longer, but a new structure SESS_CERT
8905 that holds per-session data (if available); currently, this is
8906 the peer's certificate chain and, for clients, the server's certificate
8907 and temporary key. CERT holds only those values that can have
8908 meaningful defaults in an SSL_CTX.
8911 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8912 from the internal representation. Various PKCS#7 fixes: remove some
8913 evil casts and set the enc_dig_alg field properly based on the signing
8917 *) Allow PKCS#12 password to be set from the command line or the
8918 environment. Let 'ca' get its config file name from the environment
8919 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8923 *) Allow certificate policies extension to use an IA5STRING for the
8924 organization field. This is contrary to the PKIX definition but
8925 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8929 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8930 without disallowing inline assembler and the like for non-pedantic builds.
8933 *) Support Borland C++ builder.
8934 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8939 *) SHA-1 cleanups and performance enhancements.
8940 [Andy Polyakov <appro@fy.chalmers.se>]
8942 *) Sparc v8plus assembler for the bignum library.
8943 [Andy Polyakov <appro@fy.chalmers.se>]
8945 *) Accept any -xxx and +xxx compiler options in Configure.
8948 *) Update HPUX configuration.
8951 *) Add missing sk_<type>_unshift() function to safestack.h
8952 [Ralf S. Engelschall]
8954 *) New function SSL_CTX_use_certificate_chain_file that sets the
8955 "extra_cert"s in addition to the certificate. (This makes sense
8956 only for "PEM" format files, as chains as a whole are not
8960 *) Support verify_depth from the SSL API.
8961 x509_vfy.c had what can be considered an off-by-one-error:
8962 Its depth (which was not part of the external interface)
8963 was actually counting the number of certificates in a chain;
8964 now it really counts the depth.
8967 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8968 instead of X509err, which often resulted in confusing error
8969 messages since the error codes are not globally unique
8970 (e.g. an alleged error in ssl3_accept when a certificate
8971 didn't match the private key).
8973 *) New function SSL_CTX_set_session_id_context that allows to set a default
8974 value (so that you don't need SSL_set_session_id_context for each
8975 connection using the SSL_CTX).
8978 *) OAEP decoding bug fix.
8981 *) Support INSTALL_PREFIX for package builders, as proposed by
8985 *) New Configure options "threads" and "no-threads". For systems
8986 where the proper compiler options are known (currently Solaris
8987 and Linux), "threads" is the default.
8990 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8993 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8994 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8995 such as /usr/local/bin.
8998 *) "make linux-shared" to build shared libraries.
8999 [Niels Poppe <niels@netbox.org>]
9001 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9004 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9005 extension adding in x509 utility.
9008 *) Remove NOPROTO sections and error code comments.
9011 *) Partial rewrite of the DEF file generator to now parse the ANSI
9015 *) New Configure options --prefix=DIR and --openssldir=DIR.
9018 *) Complete rewrite of the error code script(s). It is all now handled
9019 by one script at the top level which handles error code gathering,
9020 header rewriting and C source file generation. It should be much better
9021 than the old method: it now uses a modified version of Ulf's parser to
9022 read the ANSI prototypes in all header files (thus the old K&R definitions
9023 aren't needed for error creation any more) and do a better job of
9024 translating function codes into names. The old 'ASN1 error code imbedded
9025 in a comment' is no longer necessary and it doesn't use .err files which
9026 have now been deleted. Also the error code call doesn't have to appear all
9027 on one line (which resulted in some large lines...).
9030 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9033 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9034 0 (which usually indicates a closed connection), but continue reading.
9037 *) Fix some race conditions.
9040 *) Add support for CRL distribution points extension. Add Certificate
9041 Policies and CRL distribution points documentation.
9044 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9047 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9048 8 of keying material. Merlin has also confirmed interop with this fix
9049 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9050 [Merlin Hughes <merlin@baltimore.ie>]
9052 *) Fix lots of warnings.
9053 [Richard Levitte <levitte@stacken.kth.se>]
9055 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9056 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9057 [Richard Levitte <levitte@stacken.kth.se>]
9059 *) Fix problems with sizeof(long) == 8.
9060 [Andy Polyakov <appro@fy.chalmers.se>]
9062 *) Change functions to ANSI C.
9065 *) Fix typos in error codes.
9066 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9068 *) Remove defunct assembler files from Configure.
9071 *) SPARC v8 assembler BIGNUM implementation.
9072 [Andy Polyakov <appro@fy.chalmers.se>]
9074 *) Support for Certificate Policies extension: both print and set.
9075 Various additions to support the r2i method this uses.
9078 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9079 return a const string when you are expecting an allocated buffer.
9082 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9083 types DirectoryString and DisplayText.
9086 *) Add code to allow r2i extensions to access the configuration database,
9087 add an LHASH database driver and add several ctx helper functions.
9090 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9091 fail when they extended the size of a BIGNUM.
9094 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9095 support typesafe stack.
9098 *) Fix typo in SSL_[gs]et_options().
9099 [Nils Frostberg <nils@medcom.se>]
9101 *) Delete various functions and files that belonged to the (now obsolete)
9102 old X509V3 handling code.
9105 *) New Configure option "rsaref".
9108 *) Don't auto-generate pem.h.
9111 *) Introduce type-safe ASN.1 SETs.
9114 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9115 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9117 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9118 that links with OpenSSL (well at least cause lots of warnings), but fear
9119 not: the conversion is trivial, and it eliminates loads of evil casts. A
9120 few STACKed things have been converted already. Feel free to convert more.
9121 In the fullness of time, I'll do away with the STACK type altogether.
9124 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9125 specified in <certfile> by updating the entry in the index.txt file.
9126 This way one no longer has to edit the index.txt file manually for
9127 revoking a certificate. The -revoke option does the gory details now.
9128 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9130 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9131 `-text' option at all and this way the `-noout -text' combination was
9132 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9133 [Ralf S. Engelschall]
9135 *) Make sure a corresponding plain text error message exists for the
9136 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9137 verify callback function determined that a certificate was revoked.
9138 [Ralf S. Engelschall]
9140 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9141 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9142 all available cipers including rc5, which was forgotten until now.
9143 In order to let the testing shell script know which algorithms
9144 are available, a new (up to now undocumented) command
9145 "openssl list-cipher-commands" is used.
9148 *) Bugfix: s_client occasionally would sleep in select() when
9149 it should have checked SSL_pending() first.
9152 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9153 the raw DSA values prior to ASN.1 encoding.
9156 *) Tweaks to Configure
9157 [Niels Poppe <niels@netbox.org>]
9159 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9163 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9166 *) New config option to avoid instructions that are illegal on the 80386.
9167 The default code is faster, but requires at least a 486.
9170 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9171 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9172 same as SSL2_VERSION anyway.
9175 *) New "-showcerts" option for s_client.
9178 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9179 application. Various cleanups and fixes.
9182 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9183 modify error routines to work internally. Add error codes and PBE init
9184 to library startup routines.
9187 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9188 packing functions to asn1 and evp. Changed function names and error
9189 codes along the way.
9192 *) PKCS12 integration: and so it begins... First of several patches to
9193 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9194 objects to objects.h
9197 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9198 and display support for Thawte strong extranet extension.
9201 *) Add LinuxPPC support.
9202 [Jeff Dubrule <igor@pobox.org>]
9204 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9205 bn_div_words in alpha.s.
9206 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9208 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9209 OAEP isn't supported when OpenSSL is built with RSAref.
9210 [Ulf Moeller <ulf@fitug.de>]
9212 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9213 so they no longer are missing under -DNOPROTO.
9214 [Soren S. Jorvang <soren@t.dk>]
9217 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9219 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9220 doesn't work when the session is reused. Coming soon!
9223 *) Fix a security hole, that allows sessions to be reused in the wrong
9224 context thus bypassing client cert protection! All software that uses
9225 client certs and session caches in multiple contexts NEEDS PATCHING to
9226 allow session reuse! A fuller solution is in the works.
9227 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9229 *) Some more source tree cleanups (removed obsolete files
9230 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9231 permission on "config" script to be executable) and a fix for the INSTALL
9233 [Ulf Moeller <ulf@fitug.de>]
9235 *) Remove some legacy and erroneous uses of malloc, free instead of
9237 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9239 *) Make rsa_oaep_test return non-zero on error.
9240 [Ulf Moeller <ulf@fitug.de>]
9242 *) Add support for native Solaris shared libraries. Configure
9243 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9244 if someone would make that last step automatic.
9245 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9247 *) ctx_size was not built with the right compiler during "make links". Fixed.
9250 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9251 except NULL ciphers". This means the default cipher list will no longer
9252 enable NULL ciphers. They need to be specifically enabled e.g. with
9253 the string "DEFAULT:eNULL".
9256 *) Fix to RSA private encryption routines: if p < q then it would
9257 occasionally produce an invalid result. This will only happen with
9258 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9261 *) Be less restrictive and allow also `perl util/perlpath.pl
9262 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9263 because this way one can also use an interpreter named `perl5' (which is
9264 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9265 installed as `perl').
9266 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9268 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9269 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9271 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9272 advapi32.lib to Win32 build and change the pem test comparision
9273 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9274 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9275 and crypto/des/ede_cbcm_enc.c.
9278 *) DES quad checksum was broken on big-endian architectures. Fixed.
9281 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9282 Win32 test batch file so it (might) work again. The Win32 test batch file
9283 is horrible: I feel ill....
9286 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9287 in e_os.h. Audit of header files to check ANSI and non ANSI
9288 sections: 10 functions were absent from non ANSI section and not exported
9289 from Windows DLLs. Fixed up libeay.num for new functions.
9292 *) Make `openssl version' output lines consistent.
9293 [Ralf S. Engelschall]
9295 *) Fix Win32 symbol export lists for BIO functions: Added
9296 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9297 to ms/libeay{16,32}.def.
9298 [Ralf S. Engelschall]
9300 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9301 fine under Unix and passes some trivial tests I've now added. But the
9302 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9303 added to make sure no one expects that this stuff really works in the
9304 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9305 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9307 [Ralf S. Engelschall]
9309 *) Fix the generation of two part addresses in perl.
9310 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9312 *) Add config entry for Linux on MIPS.
9313 [John Tobey <jtobey@channel1.com>]
9315 *) Make links whenever Configure is run, unless we are on Windoze.
9318 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9319 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9323 *) Add a useful kludge to allow package maintainers to specify compiler and
9324 other platforms details on the command line without having to patch the
9325 Configure script everytime: One now can use ``perl Configure
9326 <id>:<details>'', i.e. platform ids are allowed to have details appended
9327 to them (seperated by colons). This is treated as there would be a static
9328 pre-configured entry in Configure's %table under key <id> with value
9329 <details> and ``perl Configure <id>'' is called. So, when you want to
9330 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9331 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9332 now, which overrides the FreeBSD-elf entry on-the-fly.
9333 [Ralf S. Engelschall]
9335 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9338 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9339 on the `perl Configure ...' command line. This way one can compile
9340 OpenSSL libraries with Position Independent Code (PIC) which is needed
9341 for linking it into DSOs.
9342 [Ralf S. Engelschall]
9344 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9348 *) Cleaned up the LICENSE document: The official contact for any license
9349 questions now is the OpenSSL core team under openssl-core@openssl.org.
9350 And add a paragraph about the dual-license situation to make sure people
9351 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9352 to the OpenSSL toolkit.
9353 [Ralf S. Engelschall]
9355 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9356 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9357 Additonally cleaned up the `make links' target: Remove unnecessary
9358 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9359 to speed processing and no longer clutter the display with confusing
9360 stuff. Instead only the actually done links are displayed.
9361 [Ralf S. Engelschall]
9363 *) Permit null encryption ciphersuites, used for authentication only. It used
9364 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9365 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9369 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9370 signed attributes when verifying signatures (this would break them),
9371 the detached data encoding was wrong and public keys obtained using
9372 X509_get_pubkey() weren't freed.
9375 *) Add text documentation for the BUFFER functions. Also added a work around
9376 to a Win95 console bug. This was triggered by the password read stuff: the
9377 last character typed gets carried over to the next fread(). If you were
9378 generating a new cert request using 'req' for example then the last
9379 character of the passphrase would be CR which would then enter the first
9383 *) Added the new `Includes OpenSSL Cryptography Software' button as
9384 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9385 button and can be used by applications based on OpenSSL to show the
9386 relationship to the OpenSSL project.
9387 [Ralf S. Engelschall]
9389 *) Remove confusing variables in function signatures in files
9390 ssl/ssl_lib.c and ssl/ssl.h.
9391 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9393 *) Don't install bss_file.c under PREFIX/include/
9394 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9396 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9397 functions that return function pointers and has support for NT specific
9398 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9399 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9400 unsigned to signed types: this was killing the Win32 compile.
9403 *) Add new certificate file to stack functions,
9404 SSL_add_dir_cert_subjects_to_stack() and
9405 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9406 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9407 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9408 This means that Apache-SSL and similar packages don't have to mess around
9409 to add as many CAs as they want to the preferred list.
9412 *) Experiment with doxygen documentation. Currently only partially applied to
9414 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9415 openssl.doxy as the configuration file.
9418 *) Get rid of remaining C++-style comments which strict C compilers hate.
9419 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9421 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9422 compiled in by default: it has problems with large keys.
9425 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9426 DH private keys and/or callback functions which directly correspond to
9427 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9428 is needed for applications which have to configure certificates on a
9429 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9431 For the RSA certificate situation is makes no difference, but
9432 for the DSA certificate situation this fixes the "no shared cipher"
9433 problem where the OpenSSL cipher selection procedure failed because the
9434 temporary keys were not overtaken from the context and the API provided
9435 no way to reconfigure them.
9436 The new functions now let applications reconfigure the stuff and they
9437 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9438 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9439 non-public-API function ssl_cert_instantiate() is used as a helper
9440 function and also to reduce code redundancy inside ssl_rsa.c.
9441 [Ralf S. Engelschall]
9443 *) Move s_server -dcert and -dkey options out of the undocumented feature
9444 area because they are useful for the DSA situation and should be
9445 recognized by the users.
9446 [Ralf S. Engelschall]
9448 *) Fix the cipher decision scheme for export ciphers: the export bits are
9449 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9450 SSL_EXP_MASK. So, the original variable has to be used instead of the
9451 already masked variable.
9452 [Richard Levitte <levitte@stacken.kth.se>]
9454 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9455 [Richard Levitte <levitte@stacken.kth.se>]
9457 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9458 from `int' to `unsigned int' because it's a length and initialized by
9459 EVP_DigestFinal() which expects an `unsigned int *'.
9460 [Richard Levitte <levitte@stacken.kth.se>]
9462 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9463 script. Instead use the usual Shell->Perl transition trick.
9464 [Ralf S. Engelschall]
9466 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9467 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9468 -noout -modulus' as it's already the case for `openssl rsa -noout
9469 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9470 currently the public key is printed (a decision which was already done by
9471 `openssl dsa -modulus' in the past) which serves a similar purpose.
9472 Additionally the NO_RSA no longer completely removes the whole -modulus
9473 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9475 [Ralf S. Engelschall]
9477 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9478 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9479 [Arne Ansper <arne@ats.cyber.ee>]
9481 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9482 to be added. Now both 'req' and 'ca' can use new objects defined in the
9486 *) Add cool BIO that does syslog (or event log on NT).
9487 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9489 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9490 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9491 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9492 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9495 *) Add preliminary config info for new extension code.
9498 *) Make RSA_NO_PADDING really use no padding.
9499 [Ulf Moeller <ulf@fitug.de>]
9501 *) Generate errors when private/public key check is done.
9504 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9505 for some CRL extensions and new objects added.
9508 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9509 key usage extension and fuller support for authority key id.
9512 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9513 padding method for RSA, which is recommended for new applications in PKCS
9514 #1 v2.0 (RFC 2437, October 1998).
9515 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9516 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9517 against Bleichbacher's attack on RSA.
9518 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9521 *) Updates to the new SSL compression code
9522 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9524 *) Fix so that the version number in the master secret, when passed
9525 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9526 (because the server will not accept higher), that the version number
9527 is 0x03,0x01, not 0x03,0x00
9528 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9530 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9531 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9532 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9535 *) Support for RAW extensions where an arbitrary extension can be
9536 created by including its DER encoding. See apps/openssl.cnf for
9540 *) Make sure latest Perl versions don't interpret some generated C array
9541 code as Perl array code in the crypto/err/err_genc.pl script.
9542 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9544 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9545 not many people have the assembler. Various Win32 compilation fixes and
9546 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9550 *) Modify configure script 'Configure' to automatically create crypto/date.h
9551 file under Win32 and also build pem.h from pem.org. New script
9552 util/mkfiles.pl to create the MINFO file on environments that can't do a
9553 'make files': perl util/mkfiles.pl >MINFO should work.
9556 *) Major rework of DES function declarations, in the pursuit of correctness
9557 and purity. As a result, many evil casts evaporated, and some weirdness,
9558 too. You may find this causes warnings in your code. Zapping your evil
9559 casts will probably fix them. Mostly.
9562 *) Fix for a typo in asn1.h. Bug fix to object creation script
9563 obj_dat.pl. It considered a zero in an object definition to mean
9564 "end of object": none of the objects in objects.h have any zeros
9565 so it wasn't spotted.
9566 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9568 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9569 Masking (CBCM). In the absence of test vectors, the best I have been able
9570 to do is check that the decrypt undoes the encrypt, so far. Send me test
9571 vectors if you have them.
9574 *) Correct calculation of key length for export ciphers (too much space was
9575 allocated for null ciphers). This has not been tested!
9578 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9579 message is now correct (it understands "crypto" and "ssl" on its
9580 command line). There is also now an "update" option. This will update
9581 the util/ssleay.num and util/libeay.num files with any new functions.
9583 perl util/mkdef.pl crypto ssl update
9584 it will update them.
9587 *) Overhauled the Perl interface (perl/*):
9588 - ported BN stuff to OpenSSL's different BN library
9589 - made the perl/ source tree CVS-aware
9590 - renamed the package from SSLeay to OpenSSL (the files still contain
9591 their history because I've copied them in the repository)
9592 - removed obsolete files (the test scripts will be replaced
9593 by better Test::Harness variants in the future)
9594 [Ralf S. Engelschall]
9596 *) First cut for a very conservative source tree cleanup:
9597 1. merge various obsolete readme texts into doc/ssleay.txt
9598 where we collect the old documents and readme texts.
9599 2. remove the first part of files where I'm already sure that we no
9600 longer need them because of three reasons: either they are just temporary
9601 files which were left by Eric or they are preserved original files where
9602 I've verified that the diff is also available in the CVS via "cvs diff
9603 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9604 the crypto/md/ stuff).
9605 [Ralf S. Engelschall]
9607 *) More extension code. Incomplete support for subject and issuer alt
9608 name, issuer and authority key id. Change the i2v function parameters
9609 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9610 what that's for :-) Fix to ASN1 macro which messed up
9611 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9614 *) Preliminary support for ENUMERATED type. This is largely copied from the
9618 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9619 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9621 *) Make sure `make rehash' target really finds the `openssl' program.
9622 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9624 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9625 like to hear about it if this slows down other processors.
9628 *) Add CygWin32 platform information to Configure script.
9629 [Alan Batie <batie@aahz.jf.intel.com>]
9631 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9632 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9634 *) New program nseq to manipulate netscape certificate sequences
9637 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9641 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9642 but the BN code had some problems that would cause failures when
9643 doing certificate verification and some other functions.
9644 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9646 *) Add ASN1 and PEM code to support netscape certificate sequences.
9649 *) Add ASN1 and PEM code to support netscape certificate sequences.
9652 *) Add several PKIX and private extended key usage OIDs.
9655 *) Modify the 'ca' program to handle the new extension code. Modify
9656 openssl.cnf for new extension format, add comments.
9659 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9660 and add a sample to openssl.cnf so req -x509 now adds appropriate
9664 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9665 error code, add initial support to X509_print() and x509 application.
9668 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9669 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9670 stuff is currently isolated and isn't even compiled yet.
9673 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9674 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9675 Removed the versions check from X509 routines when loading extensions:
9676 this allows certain broken certificates that don't set the version
9677 properly to be processed.
9680 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9681 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9682 can still be regenerated with "make depend".
9685 *) Spelling mistake in C version of CAST-128.
9686 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9688 *) Changes to the error generation code. The perl script err-code.pl
9689 now reads in the old error codes and retains the old numbers, only
9690 adding new ones if necessary. It also only changes the .err files if new
9691 codes are added. The makefiles have been modified to only insert errors
9692 when needed (to avoid needlessly modifying header files). This is done
9693 by only inserting errors if the .err file is newer than the auto generated
9694 C file. To rebuild all the error codes from scratch (the old behaviour)
9695 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9696 or delete all the .err files.
9699 *) CAST-128 was incorrectly implemented for short keys. The C version has
9700 been fixed, but is untested. The assembler versions are also fixed, but
9701 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9702 to regenerate it if needed.
9703 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9704 Hagino <itojun@kame.net>]
9706 *) File was opened incorrectly in randfile.c.
9707 [Ulf Möller <ulf@fitug.de>]
9709 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9710 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9711 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9712 al: it's just almost always a UTCTime. Note this patch adds new error
9713 codes so do a "make errors" if there are problems.
9716 *) Correct Linux 1 recognition in config.
9717 [Ulf Möller <ulf@fitug.de>]
9719 *) Remove pointless MD5 hash when using DSA keys in ca.
9720 [Anonymous <nobody@replay.com>]
9722 *) Generate an error if given an empty string as a cert directory. Also
9723 generate an error if handed NULL (previously returned 0 to indicate an
9724 error, but didn't set one).
9725 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9727 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9730 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9731 parameters. This was causing a warning which killed off the Win32 compile.
9734 *) Remove C++ style comments from crypto/bn/bn_local.h.
9735 [Neil Costigan <neil.costigan@celocom.com>]
9737 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9738 based on a text string, looking up short and long names and finally
9739 "dot" format. The "dot" format stuff didn't work. Added new function
9740 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9741 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9742 OID is not part of the table.
9745 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9746 X509_LOOKUP_by_alias().
9749 *) Sort openssl functions by name.
9752 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9753 encryption from sample DSA keys (in case anyone is interested the password
9757 *) Make _all_ *_free functions accept a NULL pointer.
9758 [Frans Heymans <fheymans@isaserver.be>]
9760 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9762 [Anonymous <nobody@replay.com>]
9764 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9765 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9767 *) Don't blow it for numeric -newkey arguments to apps/req.
9768 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9770 *) Temp key "for export" tests were wrong in s3_srvr.c.
9771 [Anonymous <nobody@replay.com>]
9773 *) Add prototype for temp key callback functions
9774 SSL_CTX_set_tmp_{rsa,dh}_callback().
9777 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9778 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9781 *) X509_name_add_entry() freed the wrong thing after an error.
9782 [Arne Ansper <arne@ats.cyber.ee>]
9784 *) rsa_eay.c would attempt to free a NULL context.
9785 [Arne Ansper <arne@ats.cyber.ee>]
9787 *) BIO_s_socket() had a broken should_retry() on Windoze.
9788 [Arne Ansper <arne@ats.cyber.ee>]
9790 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9791 [Arne Ansper <arne@ats.cyber.ee>]
9793 *) Make sure the already existing X509_STORE->depth variable is initialized
9794 in X509_STORE_new(), but document the fact that this variable is still
9795 unused in the certificate verification process.
9796 [Ralf S. Engelschall]
9798 *) Fix the various library and apps files to free up pkeys obtained from
9799 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9802 *) Fix reference counting in X509_PUBKEY_get(). This makes
9803 demos/maurice/example2.c work, amongst others, probably.
9804 [Steve Henson and Ben Laurie]
9806 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9807 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9808 are no longer created. This way we have a single and consistent command
9809 line interface `openssl <command>', similar to `cvs <command>'.
9810 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9812 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9813 BIT STRING wrapper always have zero unused bits.
9816 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9819 *) Make the top-level INSTALL documentation easier to understand.
9822 *) Makefiles updated to exit if an error occurs in a sub-directory
9823 make (including if user presses ^C) [Paul Sutton]
9825 *) Make Montgomery context stuff explicit in RSA data structure.
9828 *) Fix build order of pem and err to allow for generated pem.h.
9831 *) Fix renumbering bug in X509_NAME_delete_entry().
9834 *) Enhanced the err-ins.pl script so it makes the error library number
9835 global and can add a library name. This is needed for external ASN1 and
9836 other error libraries.
9839 *) Fixed sk_insert which never worked properly.
9842 *) Fix ASN1 macros so they can handle indefinite length construted
9843 EXPLICIT tags. Some non standard certificates use these: they can now
9847 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9848 into a single doc/ssleay.txt bundle. This way the information is still
9849 preserved but no longer messes up this directory. Now it's new room for
9850 the new set of documenation files.
9851 [Ralf S. Engelschall]
9853 *) SETs were incorrectly DER encoded. This was a major pain, because they
9854 shared code with SEQUENCEs, which aren't coded the same. This means that
9855 almost everything to do with SETs or SEQUENCEs has either changed name or
9856 number of arguments.
9857 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9859 *) Fix test data to work with the above.
9862 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9863 was already fixed by Eric for 0.9.1 it seems.
9864 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9866 *) Autodetect FreeBSD3.
9869 *) Fix various bugs in Configure. This affects the following platforms:
9873 unixware-2.0-pentium
9877 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9878 before they are needed.
9881 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9885 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9887 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9888 changed SSLeay to OpenSSL in version strings.
9889 [Ralf S. Engelschall]
9891 *) Some fixups to the top-level documents.
9894 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9895 because the symlink to include/ was missing.
9896 [Ralf S. Engelschall]
9898 *) Incorporated the popular no-RSA/DSA-only patches
9899 which allow to compile a RSA-free SSLeay.
9900 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9902 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9903 when "ssleay" is still not found.
9904 [Ralf S. Engelschall]
9906 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9907 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9909 *) Updated the README file.
9910 [Ralf S. Engelschall]
9912 *) Added various .cvsignore files in the CVS repository subdirs
9913 to make a "cvs update" really silent.
9914 [Ralf S. Engelschall]
9916 *) Recompiled the error-definition header files and added
9917 missing symbols to the Win32 linker tables.
9918 [Ralf S. Engelschall]
9920 *) Cleaned up the top-level documents;
9921 o new files: CHANGES and LICENSE
9922 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9923 o merged COPYRIGHT into LICENSE
9924 o removed obsolete TODO file
9925 o renamed MICROSOFT to INSTALL.W32
9926 [Ralf S. Engelschall]
9928 *) Removed dummy files from the 0.9.1b source tree:
9929 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9930 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9931 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9932 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9933 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9934 [Ralf S. Engelschall]
9936 *) Added various platform portability fixes.
9939 *) The Genesis of the OpenSSL rpject:
9940 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9941 Young and Tim J. Hudson created while they were working for C2Net until
9943 [The OpenSSL Project]
9946 Changes between 0.9.0b and 0.9.1b [not released]
9948 *) Updated a few CA certificates under certs/
9951 *) Changed some BIGNUM api stuff.
9954 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9955 DGUX x86, Linux Alpha, etc.
9958 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9959 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9963 *) Add -strparse option to asn1pars program which parses nested
9965 [Dr Stephen Henson <shenson@bigfoot.com>]
9967 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9970 *) DSA fix for "ca" program.
9973 *) Added "-genkey" option to "dsaparam" program.
9976 *) Added RIPE MD160 (rmd160) message digest.
9979 *) Added -a (all) option to "ssleay version" command.
9982 *) Added PLATFORM define which is the id given to Configure.
9985 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9988 *) Extended the ASN.1 parser routines.
9991 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9994 *) Added a BN_CTX to the BN library.
9997 *) Fixed the weak key values in DES library
10000 *) Changed API in EVP library for cipher aliases.
10003 *) Added support for RC2/64bit cipher.
10006 *) Converted the lhash library to the crypto/mem.c functions.
10009 *) Added more recognized ASN.1 object ids.
10012 *) Added more RSA padding checks for SSL/TLS.
10015 *) Added BIO proxy/filter functionality.
10018 *) Added extra_certs to SSL_CTX which can be used
10019 send extra CA certificates to the client in the CA cert chain sending
10020 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10023 *) Now Fortezza is denied in the authentication phase because
10024 this is key exchange mechanism is not supported by SSLeay at all.
10027 *) Additional PKCS1 checks.
10030 *) Support the string "TLSv1" for all TLS v1 ciphers.
10033 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10034 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10037 *) Fixed a few memory leaks.
10040 *) Fixed various code and comment typos.
10043 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10044 bytes sent in the client random.
10045 [Edward Bishop <ebishop@spyglass.com>]