5 Changes between 1.0.0o and 1.0.0p [xx XXX xxxx]
7 *) Abort handshake if server key exchange message is omitted for ephemeral
10 Thanks to Karthikeyan Bhargavan for reporting this issue.
14 *) Remove non-export ephemeral RSA code on client and server. This code
15 violated the TLS standard by allowing the use of temporary RSA keys in
16 non-export ciphersuites and could be used by a server to effectively
17 downgrade the RSA key length used to a value smaller than the server
18 certificate. Thanks for Karthikeyan Bhargavan for reporting this issue.
22 *) Fix various certificate fingerprint issues.
24 By using non-DER or invalid encodings outside the signed portion of a
25 certificate the fingerprint can be changed without breaking the signature.
26 Although no details of the signed portion of the certificate can be changed
27 this can cause problems with some applications: e.g. those using the
28 certificate fingerprint for blacklists.
30 1. Reject signatures with non zero unused bits.
32 If the BIT STRING containing the signature has non zero unused bits reject
33 the signature. All current signature algorithms require zero unused bits.
35 2. Check certificate algorithm consistency.
37 Check the AlgorithmIdentifier inside TBS matches the one in the
38 certificate signature. NB: this will result in signature failure
39 errors for some broken certificates.
41 Thanks to Konrad Kraszewski from Google for reporting this issue.
43 3. Check DSA/ECDSA signatures use DER.
45 Reencode DSA/ECDSA signatures and compare with the original received
46 signature. Return an error if there is a mismatch.
48 This will reject various cases including garbage after signature
49 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
50 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
51 (negative or with leading zeroes).
53 Further analysis was conducted and fixes were developed by Stephen Henson
54 of the OpenSSL core team.
59 Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
61 *) Session Ticket Memory Leak.
63 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
64 integrity of that ticket is first verified. In the event of a session
65 ticket integrity check failing, OpenSSL will fail to free memory
66 causing a memory leak. By sending a large number of invalid session
67 tickets an attacker could exploit this issue in a Denial Of Service
72 *) Build option no-ssl3 is incomplete.
74 When OpenSSL is configured with "no-ssl3" as a build option, servers
75 could accept and complete a SSL 3.0 handshake, and clients could be
76 configured to send them.
78 [Akamai and the OpenSSL team]
80 *) Add support for TLS_FALLBACK_SCSV.
81 Client applications doing fallback retries should call
82 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
84 [Adam Langley, Bodo Moeller]
86 *) Add additional DigestInfo checks.
88 Reencode DigestInto in DER and check against the original when
89 verifying RSA signature: this will reject any improperly encoded
90 DigestInfo structures.
92 Note: this is a precautionary measure and no attacks are currently known.
96 Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
98 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
99 to a denial of service attack. A malicious server can crash the client
100 with a null pointer dereference (read) by specifying an anonymous (EC)DH
101 ciphersuite and sending carefully crafted handshake messages.
103 Thanks to Felix Gröbert (Google) for discovering and researching this
108 *) By sending carefully crafted DTLS packets an attacker could cause openssl
109 to leak memory. This can be exploited through a Denial of Service attack.
110 Thanks to Adam Langley for discovering and researching this issue.
114 *) An attacker can force openssl to consume large amounts of memory whilst
115 processing DTLS handshake messages. This can be exploited through a
116 Denial of Service attack.
117 Thanks to Adam Langley for discovering and researching this issue.
121 *) An attacker can force an error condition which causes openssl to crash
122 whilst processing DTLS packets due to memory being freed twice. This
123 can be exploited through a Denial of Service attack.
124 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
129 *) If a multithreaded client connects to a malicious server using a resumed
130 session and the server sends an ec point format extension it could write
131 up to 255 bytes to freed memory.
133 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
138 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
139 X509_name_oneline, X509_name_print_ex et al. to leak some information
140 from the stack. Applications may be affected if they echo pretty printing
141 output to the attacker.
143 Thanks to Ivan Fratric (Google) for discovering this issue.
145 [Emilia Käsper, and Steve Henson]
147 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
148 for corner cases. (Certain input points at infinity could lead to
149 bogus results, with non-infinity inputs mapped to infinity too.)
152 Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
154 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
155 handshake can force the use of weak keying material in OpenSSL
156 SSL/TLS clients and servers.
158 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
159 researching this issue. (CVE-2014-0224)
160 [KIKUCHI Masashi, Steve Henson]
162 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
163 OpenSSL DTLS client the code can be made to recurse eventually crashing
166 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
168 [Imre Rad, Steve Henson]
170 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
171 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
172 client or server. This is potentially exploitable to run arbitrary
173 code on a vulnerable client or server.
175 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
176 [Jüri Aedla, Steve Henson]
178 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
179 are subject to a denial of service attack.
181 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
182 this issue. (CVE-2014-3470)
183 [Felix Gröbert, Ivan Fratric, Steve Henson]
185 *) Harmonize version and its documentation. -f flag is used to display
187 [mancha <mancha1@zoho.com>]
189 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
191 [mancha <mancha1@zoho.com>]
193 *) Fix some double frees. These are not thought to be exploitable.
194 [mancha <mancha1@zoho.com>]
196 *) Fix for the attack described in the paper "Recovering OpenSSL
197 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
198 by Yuval Yarom and Naomi Benger. Details can be obtained from:
199 http://eprint.iacr.org/2014/140
201 Thanks to Yuval Yarom and Naomi Benger for discovering this
202 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
203 [Yuval Yarom and Naomi Benger]
205 Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
207 *) Keep original DTLS digest and encryption contexts in retransmission
208 structures so we can use the previous session parameters if they need
209 to be resent. (CVE-2013-6450)
212 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
213 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
214 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
215 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
216 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
217 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
218 [Rob Stradling, Adam Langley]
220 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
222 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
224 This addresses the flaw in CBC record processing discovered by
225 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
226 at: http://www.isg.rhul.ac.uk/tls/
228 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
229 Security Group at Royal Holloway, University of London
230 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
231 Emilia Käsper for the initial patch.
233 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
235 *) Return an error when checking OCSP signatures when key is NULL.
236 This fixes a DoS attack. (CVE-2013-0166)
239 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
240 the right response is stapled. Also change SSL_get_certificate()
241 so it returns the certificate actually sent.
242 See http://rt.openssl.org/Ticket/Display.html?id=2836.
244 [Rob Stradling <rob.stradling@comodo.com>]
246 *) Fix possible deadlock when decoding public keys.
249 Changes between 1.0.0i and 1.0.0j [10 May 2012]
251 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
254 *) Sanity check record length before skipping explicit IV in DTLS
257 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
258 fuzzing as a service testing platform.
262 *) Initialise tkeylen properly when encrypting CMS messages.
263 Thanks to Solar Designer of Openwall for reporting this issue.
266 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
268 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
269 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
270 in CRYPTO_realloc_clean.
272 Thanks to Tavis Ormandy, Google Security Team, for discovering this
273 issue and to Adam Langley <agl@chromium.org> for fixing it.
275 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
277 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
279 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
280 in CMS and PKCS7 code. When RSA decryption fails use a random key for
281 content decryption and always return the same error. Note: this attack
282 needs on average 2^20 messages so it only affects automated senders. The
283 old behaviour can be reenabled in the CMS code by setting the
284 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
285 an MMA defence is not necessary.
286 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
287 this issue. (CVE-2012-0884)
290 *) Fix CVE-2011-4619: make sure we really are receiving a
291 client hello before rejecting multiple SGC restarts. Thanks to
292 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
295 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
297 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
298 Thanks to Antonio Martin, Enterprise Secure Access Research and
299 Development, Cisco Systems, Inc. for discovering this bug and
300 preparing a fix. (CVE-2012-0050)
303 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
305 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
306 of the Vaudenay padding oracle attack on CBC mode encryption
307 which enables an efficient plaintext recovery attack against
308 the OpenSSL implementation of DTLS. Their attack exploits timing
309 differences arising during decryption processing. A research
310 paper describing this attack can be found at:
311 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
312 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
313 Security Group at Royal Holloway, University of London
314 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
315 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
316 for preparing the fix. (CVE-2011-4108)
317 [Robin Seggelmann, Michael Tuexen]
319 *) Clear bytes used for block padding of SSL 3.0 records.
321 [Adam Langley (Google)]
323 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
324 Kadianakis <desnacked@gmail.com> for discovering this issue and
325 Adam Langley for preparing the fix. (CVE-2011-4619)
326 [Adam Langley (Google)]
328 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
329 [Andrey Kulikov <amdeich@gmail.com>]
331 *) Prevent malformed RFC3779 data triggering an assertion failure.
332 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
333 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
334 [Rob Austein <sra@hactrn.net>]
336 *) Improved PRNG seeding for VOS.
337 [Paul Green <Paul.Green@stratus.com>]
339 *) Fix ssl_ciph.c set-up race.
340 [Adam Langley (Google)]
342 *) Fix spurious failures in ecdsatest.c.
343 [Emilia Käsper (Google)]
345 *) Fix the BIO_f_buffer() implementation (which was mixing different
346 interpretations of the '..._len' fields).
347 [Adam Langley (Google)]
349 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
350 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
351 threads won't reuse the same blinding coefficients.
353 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
354 lock to call BN_BLINDING_invert_ex, and avoids one use of
355 BN_BLINDING_update for each BN_BLINDING structure (previously,
356 the last update always remained unused).
357 [Emilia Käsper (Google)]
359 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
360 [Bob Buckholz (Google)]
362 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
364 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
365 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
366 [Kaspar Brand <ossl@velox.ch>]
368 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
369 for multi-threaded use of ECDH. (CVE-2011-3210)
370 [Adam Langley (Google)]
372 *) Fix x509_name_ex_d2i memory leak on bad inputs.
375 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
376 signature public key algorithm by using OID xref utilities instead.
377 Before this you could only use some ECC ciphersuites with SHA1 only.
380 *) Add protection against ECDSA timing attacks as mentioned in the paper
381 by Billy Bob Brumley and Nicola Tuveri, see:
383 http://eprint.iacr.org/2011/232.pdf
385 [Billy Bob Brumley and Nicola Tuveri]
387 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
389 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
390 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
392 *) Fix bug in string printing code: if *any* escaping is enabled we must
393 escape the escape character (backslash) or the resulting string is
397 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
399 *) Disable code workaround for ancient and obsolete Netscape browsers
400 and servers: an attacker can use it in a ciphersuite downgrade attack.
401 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
404 *) Fixed J-PAKE implementation error, originally discovered by
405 Sebastien Martini, further info and confirmation from Stefan
406 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
409 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
411 *) Fix extension code to avoid race conditions which can result in a buffer
412 overrun vulnerability: resumed sessions must not be modified as they can
413 be shared by multiple threads. CVE-2010-3864
416 *) Fix WIN32 build system to correctly link an ENGINE directory into
420 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
422 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
424 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
426 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
428 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
429 context. The operation can be customised via the ctrl mechanism in
430 case ENGINEs want to include additional functionality.
433 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
436 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
437 output hashes compatible with older versions of OpenSSL.
438 [Willy Weisz <weisz@vcpc.univie.ac.at>]
440 *) Fix compression algorithm handling: if resuming a session use the
441 compression algorithm of the resumed session instead of determining
442 it from client hello again. Don't allow server to change algorithm.
445 *) Add load_crls() function to apps tidying load_certs() too. Add option
446 to verify utility to allow additional CRLs to be included.
449 *) Update OCSP request code to permit adding custom headers to the request:
450 some responders need this.
453 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
455 [Julia Lawall <julia@diku.dk>]
457 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
458 needlessly dereferenced structures, used obsolete functions and
459 didn't handle all updated verify codes correctly.
462 *) Disable MD2 in the default configuration.
465 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
466 indicate the initial BIO being pushed or popped. This makes it possible
467 to determine whether the BIO is the one explicitly called or as a result
468 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
469 it handles reference counts correctly and doesn't zero out the I/O bio
470 when it is not being explicitly popped. WARNING: applications which
471 included workarounds for the old buggy behaviour will need to be modified
472 or they could free up already freed BIOs.
475 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
476 renaming to all platforms (within the 0.9.8 branch, this was
477 done conditionally on Netware platforms to avoid a name clash).
478 [Guenter <lists@gknw.net>]
480 *) Add ECDHE and PSK support to DTLS.
481 [Michael Tuexen <tuexen@fh-muenster.de>]
483 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
487 *) Add "missing" function EVP_MD_flags() (without this the only way to
488 retrieve a digest flags is by accessing the structure directly. Update
489 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
490 or cipher is registered as in the "from" argument. Print out all
491 registered digests in the dgst usage message instead of manually
492 attempting to work them out.
495 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
496 this allows the use of compression and extensions. Change default cipher
497 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
498 by default unless an application cipher string requests it.
501 *) Alter match criteria in PKCS12_parse(). It used to try to use local
502 key ids to find matching certificates and keys but some PKCS#12 files
503 don't follow the (somewhat unwritten) rules and this strategy fails.
504 Now just gather all certificates together and the first private key
505 then look for the first certificate that matches the key.
508 *) Support use of registered digest and cipher names for dgst and cipher
509 commands instead of having to add each one as a special case. So now
516 openssl dgst -sha256 foo
518 and this works for ENGINE based algorithms too.
522 *) Update Gost ENGINE to support parameter files.
523 [Victor B. Wagner <vitus@cryptocom.ru>]
525 *) Support GeneralizedTime in ca utility.
526 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
528 *) Enhance the hash format used for certificate directory links. The new
529 form uses the canonical encoding (meaning equivalent names will work
530 even if they aren't identical) and uses SHA1 instead of MD5. This form
531 is incompatible with the older format and as a result c_rehash should
532 be used to rebuild symbolic links.
535 *) Make PKCS#8 the default write format for private keys, replacing the
536 traditional format. This form is standardised, more secure and doesn't
537 include an implicit MD5 dependency.
540 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
541 committed to OpenSSL should pass this lot as a minimum.
544 *) Add session ticket override functionality for use by EAP-FAST.
545 [Jouni Malinen <j@w1.fi>]
547 *) Modify HMAC functions to return a value. Since these can be implemented
548 in an ENGINE errors can occur.
551 *) Type-checked OBJ_bsearch_ex.
554 *) Type-checked OBJ_bsearch. Also some constification necessitated
555 by type-checking. Still to come: TXT_DB, bsearch(?),
556 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
560 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
561 seconds to a tm structure directly, instead of going through OS
562 specific date routines. This avoids any issues with OS routines such
563 as the year 2038 bug. New *_adj() functions for ASN1 time structures
564 and X509_time_adj_ex() to cover the extended range. The existing
565 X509_time_adj() is still usable and will no longer have any date issues.
568 *) Delta CRL support. New use deltas option which will attempt to locate
569 and search any appropriate delta CRLs available.
571 This work was sponsored by Google.
574 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
575 code and add additional score elements. Validate alternate CRL paths
576 as part of the CRL checking and indicate a new error "CRL path validation
577 error" in this case. Applications wanting additional details can use
578 the verify callback and check the new "parent" field. If this is not
579 NULL CRL path validation is taking place. Existing applications wont
580 see this because it requires extended CRL support which is off by
583 This work was sponsored by Google.
586 *) Support for freshest CRL extension.
588 This work was sponsored by Google.
591 *) Initial indirect CRL support. Currently only supported in the CRLs
592 passed directly and not via lookup. Process certificate issuer
593 CRL entry extension and lookup CRL entries by bother issuer name
594 and serial number. Check and process CRL issuer entry in IDP extension.
596 This work was sponsored by Google.
599 *) Add support for distinct certificate and CRL paths. The CRL issuer
600 certificate is validated separately in this case. Only enabled if
601 an extended CRL support flag is set: this flag will enable additional
602 CRL functionality in future.
604 This work was sponsored by Google.
607 *) Add support for policy mappings extension.
609 This work was sponsored by Google.
612 *) Fixes to pathlength constraint, self issued certificate handling,
613 policy processing to align with RFC3280 and PKITS tests.
615 This work was sponsored by Google.
618 *) Support for name constraints certificate extension. DN, email, DNS
619 and URI types are currently supported.
621 This work was sponsored by Google.
624 *) To cater for systems that provide a pointer-based thread ID rather
625 than numeric, deprecate the current numeric thread ID mechanism and
626 replace it with a structure and associated callback type. This
627 mechanism allows a numeric "hash" to be extracted from a thread ID in
628 either case, and on platforms where pointers are larger than 'long',
629 mixing is done to help ensure the numeric 'hash' is usable even if it
630 can't be guaranteed unique. The default mechanism is to use "&errno"
631 as a pointer-based thread ID to distinguish between threads.
633 Applications that want to provide their own thread IDs should now use
634 CRYPTO_THREADID_set_callback() to register a callback that will call
635 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
637 Note that ERR_remove_state() is now deprecated, because it is tied
638 to the assumption that thread IDs are numeric. ERR_remove_state(0)
639 to free the current thread's error state should be replaced by
640 ERR_remove_thread_state(NULL).
642 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
643 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
644 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
645 application was previously providing a numeric thread callback that
646 was inappropriate for distinguishing threads, then uniqueness might
647 have been obtained with &errno that happened immediately in the
648 intermediate development versions of OpenSSL; this is no longer the
649 case, the numeric thread callback will now override the automatic use
651 [Geoff Thorpe, with help from Bodo Moeller]
653 *) Initial support for different CRL issuing certificates. This covers a
654 simple case where the self issued certificates in the chain exist and
655 the real CRL issuer is higher in the existing chain.
657 This work was sponsored by Google.
660 *) Removed effectively defunct crypto/store from the build.
663 *) Revamp of STACK to provide stronger type-checking. Still to come:
664 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
665 ASN1_STRING, CONF_VALUE.
668 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
669 RAM on SSL connections. This option can save about 34k per idle SSL.
672 *) Revamp of LHASH to provide stronger type-checking. Still to come:
673 STACK, TXT_DB, bsearch, qsort.
676 *) Initial support for Cryptographic Message Syntax (aka CMS) based
677 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
678 support for data, signedData, compressedData, digestedData and
679 encryptedData, envelopedData types included. Scripts to check against
680 RFC4134 examples draft and interop and consistency checks of many
681 content types and variants.
684 *) Add options to enc utility to support use of zlib compression BIO.
687 *) Extend mk1mf to support importing of options and assembly language
688 files from Configure script, currently only included in VC-WIN32.
689 The assembly language rules can now optionally generate the source
690 files from the associated perl scripts.
693 *) Implement remaining functionality needed to support GOST ciphersuites.
694 Interop testing has been performed using CryptoPro implementations.
695 [Victor B. Wagner <vitus@cryptocom.ru>]
697 *) s390x assembler pack.
700 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
704 *) Implement Opaque PRF Input TLS extension as specified in
705 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
706 official specification yet and no extension type assignment by
707 IANA exists, this extension (for now) will have to be explicitly
708 enabled when building OpenSSL by providing the extension number
709 to use. For example, specify an option
711 -DTLSEXT_TYPE_opaque_prf_input=0x9527
713 to the "config" or "Configure" script to enable the extension,
714 assuming extension number 0x9527 (which is a completely arbitrary
715 and unofficial assignment based on the MD5 hash of the Internet
716 Draft). Note that by doing so, you potentially lose
717 interoperability with other TLS implementations since these might
718 be using the same extension number for other purposes.
720 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
721 opaque PRF input value to use in the handshake. This will create
722 an interal copy of the length-'len' string at 'src', and will
723 return non-zero for success.
725 To get more control and flexibility, provide a callback function
728 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
729 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
733 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
736 Callback function 'cb' will be called in handshakes, and is
737 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
738 Argument 'arg' is for application purposes (the value as given to
739 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
740 be provided to the callback function). The callback function
741 has to return non-zero to report success: usually 1 to use opaque
742 PRF input just if possible, or 2 to enforce use of the opaque PRF
743 input. In the latter case, the library will abort the handshake
744 if opaque PRF input is not successfully negotiated.
746 Arguments 'peerinput' and 'len' given to the callback function
747 will always be NULL and 0 in the case of a client. A server will
748 see the client's opaque PRF input through these variables if
749 available (NULL and 0 otherwise). Note that if the server
750 provides an opaque PRF input, the length must be the same as the
751 length of the client's opaque PRF input.
753 Note that the callback function will only be called when creating
754 a new session (session resumption can resume whatever was
755 previously negotiated), and will not be called in SSL 2.0
756 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
757 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
758 for applications that need to enforce opaque PRF input.
762 *) Update ssl code to support digests other than SHA1+MD5 for handshake
765 [Victor B. Wagner <vitus@cryptocom.ru>]
767 *) Add RFC4507 support to OpenSSL. This includes the corrections in
768 RFC4507bis. The encrypted ticket format is an encrypted encoded
769 SSL_SESSION structure, that way new session features are automatically
772 If a client application caches session in an SSL_SESSION structure
773 support is transparent because tickets are now stored in the encoded
776 The SSL_CTX structure automatically generates keys for ticket
777 protection in servers so again support should be possible
778 with no application modification.
780 If a client or server wishes to disable RFC4507 support then the option
781 SSL_OP_NO_TICKET can be set.
783 Add a TLS extension debugging callback to allow the contents of any client
784 or server extensions to be examined.
786 This work was sponsored by Google.
789 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
790 OpenSSL should now compile cleanly on gcc 4.2
791 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
793 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
794 support including streaming MAC support: this is required for GOST
796 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
798 *) Add option -stream to use PKCS#7 streaming in smime utility. New
799 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
800 to output in BER and PEM format.
803 *) Experimental support for use of HMAC via EVP_PKEY interface. This
804 allows HMAC to be handled via the EVP_DigestSign*() interface. The
805 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
806 ENGINE support for HMAC keys which are unextractable. New -mac and
807 -macopt options to dgst utility.
810 *) New option -sigopt to dgst utility. Update dgst to use
811 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
812 alternative signing paramaters such as X9.31 or PSS in the dgst
816 *) Change ssl_cipher_apply_rule(), the internal function that does
817 the work each time a ciphersuite string requests enabling
818 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
819 removing ("!foo+bar") a class of ciphersuites: Now it maintains
820 the order of disabled ciphersuites such that those ciphersuites
821 that most recently went from enabled to disabled not only stay
822 in order with respect to each other, but also have higher priority
823 than other disabled ciphersuites the next time ciphersuites are
826 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
827 the same ciphersuites as with "HIGH" alone, but in a specific
828 order where the PSK ciphersuites come first (since they are the
829 most recently disabled ciphersuites when "HIGH" is parsed).
831 Also, change ssl_create_cipher_list() (using this new
832 funcionality) such that between otherwise identical
833 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
837 *) Change ssl_create_cipher_list() so that it automatically
838 arranges the ciphersuites in reasonable order before starting
839 to process the rule string. Thus, the definition for "DEFAULT"
840 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
841 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
842 This makes it much easier to arrive at a reasonable default order
843 in applications for which anonymous ciphers are OK (meaning
844 that you can't actually use DEFAULT).
845 [Bodo Moeller; suggested by Victor Duchovni]
847 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
848 processing) into multiple integers instead of setting
849 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
850 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
851 (These masks as well as the individual bit definitions are hidden
852 away into the non-exported interface ssl/ssl_locl.h, so this
853 change to the definition of the SSL_CIPHER structure shouldn't
854 affect applications.) This give us more bits for each of these
855 categories, so there is no longer a need to coagulate AES128 and
856 AES256 into a single algorithm bit, and to coagulate Camellia128
857 and Camellia256 into a single algorithm bit, which has led to all
860 Thus, among other things, the kludge introduced in 0.9.7m and
861 0.9.8e for masking out AES256 independently of AES128 or masking
862 out Camellia256 independently of AES256 is not needed here in 0.9.9.
864 With the change, we also introduce new ciphersuite aliases that
865 so far were missing: "AES128", "AES256", "CAMELLIA128", and
869 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
870 Use the leftmost N bytes of the signature input if the input is
871 larger than the prime q (with N being the size in bytes of q).
874 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
875 it yet and it is largely untested.
878 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
881 *) Initial incomplete changes to avoid need for function casts in OpenSSL
882 some compilers (gcc 4.2 and later) reject their use. Safestack is
883 reimplemented. Update ASN1 to avoid use of legacy functions.
886 *) Win32/64 targets are linked with Winsock2.
889 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
890 to external functions. This can be used to increase CRL handling
891 efficiency especially when CRLs are very large by (for example) storing
892 the CRL revoked certificates in a database.
895 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
896 new CRLs added to a directory can be used. New command line option
897 -verify_return_error to s_client and s_server. This causes real errors
898 to be returned by the verify callback instead of carrying on no matter
899 what. This reflects the way a "real world" verify callback would behave.
902 *) GOST engine, supporting several GOST algorithms and public key formats.
903 Kindly donated by Cryptocom.
906 *) Partial support for Issuing Distribution Point CRL extension. CRLs
907 partitioned by DP are handled but no indirect CRL or reason partitioning
908 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
909 selected via a scoring technique which handles IDP and AKID in CRLs.
912 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
913 will ultimately be used for all verify operations: this will remove the
914 X509_STORE dependency on certificate verification and allow alternative
915 lookup methods. X509_STORE based implementations of these two callbacks.
918 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
919 Modify get_crl() to find a valid (unexpired) CRL if possible.
922 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
923 this would be called X509_CRL_cmp() but that name is already used by
924 a function that just compares CRL issuer names. Cache several CRL
925 extensions in X509_CRL structure and cache CRLDP in X509.
928 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
929 this maps equivalent X509_NAME structures into a consistent structure.
930 Name comparison can then be performed rapidly using memcmp().
933 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
937 *) Allow digests to supply their own micalg string for S/MIME type using
938 the ctrl EVP_MD_CTRL_MICALG.
941 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
942 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
943 ctrl. It can then customise the structure before and/or after signing
947 *) New function OBJ_add_sigid() to allow application defined signature OIDs
948 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
949 to free up any added signature OIDs.
952 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
953 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
954 digest and cipher tables. New options added to openssl utility:
955 list-message-digest-algorithms and list-cipher-algorithms.
958 *) Change the array representation of binary polynomials: the list
959 of degrees of non-zero coefficients is now terminated with -1.
960 Previously it was terminated with 0, which was also part of the
961 value; thus, the array representation was not applicable to
962 polynomials where t^0 has coefficient zero. This change makes
963 the array representation useful in a more general context.
966 *) Various modifications and fixes to SSL/TLS cipher string
967 handling. For ECC, the code now distinguishes between fixed ECDH
968 with RSA certificates on the one hand and with ECDSA certificates
969 on the other hand, since these are separate ciphersuites. The
970 unused code for Fortezza ciphersuites has been removed.
972 For consistency with EDH, ephemeral ECDH is now called "EECDH"
973 (not "ECDHE"). For consistency with the code for DH
974 certificates, use of ECDH certificates is now considered ECDH
975 authentication, not RSA or ECDSA authentication (the latter is
976 merely the CA's signing algorithm and not actively used in the
979 The temporary ciphersuite alias "ECCdraft" is no longer
980 available, and ECC ciphersuites are no longer excluded from "ALL"
981 and "DEFAULT". The following aliases now exist for RFC 4492
982 ciphersuites, most of these by analogy with the DH case:
984 kECDHr - ECDH cert, signed with RSA
985 kECDHe - ECDH cert, signed with ECDSA
986 kECDH - ECDH cert (signed with either RSA or ECDSA)
987 kEECDH - ephemeral ECDH
988 ECDH - ECDH cert or ephemeral ECDH
994 AECDH - anonymous ECDH
995 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
999 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1000 Use correct micalg parameters depending on digest(s) in signed message.
1003 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1004 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1007 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1008 an engine to register a method. Add ENGINE lookups for methods and
1009 functional reference processing.
1012 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1013 EVP_{Sign,Verify}* which allow an application to customise the signature
1017 *) New -resign option to smime utility. This adds one or more signers
1018 to an existing PKCS#7 signedData structure. Also -md option to use an
1019 alternative message digest algorithm for signing.
1022 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1023 create PKCS7 structures containing multiple signers. Update smime
1024 application to support multiple signers.
1027 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1031 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1032 Reorganize PBE internals to lookup from a static table using NIDs,
1033 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1034 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1035 PRF which will be automatically used with PBES2.
1038 *) Replace the algorithm specific calls to generate keys in "req" with the
1042 *) Update PKCS#7 enveloped data routines to use new API. This is now
1043 supported by any public key method supporting the encrypt operation. A
1044 ctrl is added to allow the public key algorithm to examine or modify
1045 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1049 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1050 a default digest type to use. In most cases this will be SHA1 but some
1051 algorithms (such as GOST) need to specify an alternative digest. The
1052 return value indicates how strong the prefernce is 1 means optional and
1053 2 is mandatory (that is it is the only supported type). Modify
1054 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1055 use the default md. Update openssl utilities to use the default digest
1056 type for signing if it is not explicitly indicated.
1059 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1060 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1061 signing method from the key type. This effectively removes the link
1062 between digests and public key types.
1065 *) Add an OID cross reference table and utility functions. Its purpose is to
1066 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1067 rsaEncryption. This will allow some of the algorithm specific hackery
1068 needed to use the correct OID to be removed.
1071 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1072 structures for PKCS7_sign(). They are now set up by the relevant public
1076 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1079 *) Add support for key derivation (agreement) in the API, DH method and
1083 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1084 public and private key formats. As a side effect these add additional
1085 command line functionality not previously available: DSA signatures can be
1086 generated and verified using pkeyutl and DH key support and generation in
1091 [Oliver Tappe <zooey@hirschkaefer.de>]
1093 *) New make target "install_html_docs" installs HTML renditions of the
1095 [Oliver Tappe <zooey@hirschkaefer.de>]
1097 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1098 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1099 support key and parameter generation and add initial key generation
1100 functionality for RSA.
1103 *) Add functions for main EVP_PKEY_method operations. The undocumented
1104 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1105 EVP_PKEY_{encrypt,decrypt}_old.
1108 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1109 key API, doesn't do much yet.
1112 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1113 public key algorithms. New option to openssl utility:
1114 "list-public-key-algorithms" to print out info.
1117 *) Implement the Supported Elliptic Curves Extension for
1118 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1121 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1122 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1125 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1126 utilities such as rsa, dsa, dsaparam etc except they process any key
1130 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1131 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1132 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1136 *) Initial support for pluggable public key ASN1.
1137 De-spaghettify the public key ASN1 handling. Move public and private
1138 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1139 algorithm specific handling to a single module within the relevant
1140 algorithm directory. Add functions to allow (near) opaque processing
1141 of public and private key structures.
1144 *) Implement the Supported Point Formats Extension for
1145 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1148 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1149 for the psk identity [hint] and the psk callback functions to the
1150 SSL_SESSION, SSL and SSL_CTX structure.
1153 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1157 SSL_CTX_use_psk_identity_hint
1158 SSL_get_psk_identity_hint
1159 SSL_get_psk_identity
1160 SSL_use_psk_identity_hint
1162 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1164 *) Add RFC 3161 compliant time stamp request creation, response generation
1165 and response verification functionality.
1166 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1168 *) Add initial support for TLS extensions, specifically for the server_name
1169 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1170 have new members for a host name. The SSL data structure has an
1171 additional member SSL_CTX *initial_ctx so that new sessions can be
1172 stored in that context to allow for session resumption, even after the
1173 SSL has been switched to a new SSL_CTX in reaction to a client's
1174 server_name extension.
1176 New functions (subject to change):
1178 SSL_get_servername()
1179 SSL_get_servername_type()
1182 New CTRL codes and macros (subject to change):
1184 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1185 - SSL_CTX_set_tlsext_servername_callback()
1186 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1187 - SSL_CTX_set_tlsext_servername_arg()
1188 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1190 openssl s_client has a new '-servername ...' option.
1192 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1193 '-key2 ...', '-servername_fatal' (subject to change). This allows
1194 testing the HostName extension for a specific single host name ('-cert'
1195 and '-key' remain fallbacks for handshakes without HostName
1196 negotiation). If the unrecogninzed_name alert has to be sent, this by
1197 default is a warning; it becomes fatal with the '-servername_fatal'
1200 [Peter Sylvester, Remy Allais, Christophe Renou]
1202 *) Whirlpool hash implementation is added.
1205 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1206 bn(64,32). Because of instruction set limitations it doesn't have
1207 any negative impact on performance. This was done mostly in order
1208 to make it possible to share assembler modules, such as bn_mul_mont
1209 implementations, between 32- and 64-bit builds without hassle.
1212 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1213 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1217 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1218 dedicated Montgomery multiplication procedure, is introduced.
1219 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1220 "64-bit" performance on certain 32-bit targets.
1223 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1224 in SSL structures. New SSL ctrl to set maximum send fragment size.
1225 Save memory by seeting the I/O buffer sizes dynamically instead of
1226 using the maximum available value.
1229 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1230 in addition to the text details.
1233 *) Very, very preliminary EXPERIMENTAL support for printing of general
1234 ASN1 structures. This currently produces rather ugly output and doesn't
1235 handle several customised structures at all.
1238 *) Integrated support for PVK file format and some related formats such
1239 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1240 these in the 'rsa' and 'dsa' utilities.
1243 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1246 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1247 place for the (very old) "NETSCAPE" format certificates which are now
1248 handled using new ASN1 code equivalents.
1251 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1252 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1253 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1256 *) Modify CRL distribution points extension code to print out previously
1257 unsupported fields. Enhance extension setting code to allow setting of
1261 *) Add print and set support for Issuing Distribution Point CRL extension.
1264 *) Change 'Configure' script to enable Camellia by default.
1267 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
1269 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1271 This addresses the flaw in CBC record processing discovered by
1272 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1273 at: http://www.isg.rhul.ac.uk/tls/
1275 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1276 Security Group at Royal Holloway, University of London
1277 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1278 Emilia Käsper for the initial patch.
1280 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1282 *) Return an error when checking OCSP signatures when key is NULL.
1283 This fixes a DoS attack. (CVE-2013-0166)
1286 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1287 the right response is stapled. Also change SSL_get_certificate()
1288 so it returns the certificate actually sent.
1289 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1290 (This is a backport)
1291 [Rob Stradling <rob.stradling@comodo.com>]
1293 *) Fix possible deadlock when decoding public keys.
1296 Changes between 0.9.8w and 0.9.8x [10 May 2012]
1298 *) Sanity check record length before skipping explicit IV in DTLS
1301 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1302 fuzzing as a service testing platform.
1306 *) Initialise tkeylen properly when encrypting CMS messages.
1307 Thanks to Solar Designer of Openwall for reporting this issue.
1310 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
1312 *) The fix for CVE-2012-2110 did not take into account that the
1313 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
1314 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
1315 rejecting negative len parameter. (CVE-2012-2131)
1316 [Tomas Hoger <thoger@redhat.com>]
1318 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
1320 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1321 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1322 in CRYPTO_realloc_clean.
1324 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1325 issue and to Adam Langley <agl@chromium.org> for fixing it.
1327 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1329 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
1331 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1332 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1333 content decryption and always return the same error. Note: this attack
1334 needs on average 2^20 messages so it only affects automated senders. The
1335 old behaviour can be reenabled in the CMS code by setting the
1336 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1337 an MMA defence is not necessary.
1338 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1339 this issue. (CVE-2012-0884)
1342 *) Fix CVE-2011-4619: make sure we really are receiving a
1343 client hello before rejecting multiple SGC restarts. Thanks to
1344 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1347 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1349 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1350 Thanks to Antonio Martin, Enterprise Secure Access Research and
1351 Development, Cisco Systems, Inc. for discovering this bug and
1352 preparing a fix. (CVE-2012-0050)
1355 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1357 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1358 of the Vaudenay padding oracle attack on CBC mode encryption
1359 which enables an efficient plaintext recovery attack against
1360 the OpenSSL implementation of DTLS. Their attack exploits timing
1361 differences arising during decryption processing. A research
1362 paper describing this attack can be found at:
1363 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1364 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1365 Security Group at Royal Holloway, University of London
1366 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1367 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1368 for preparing the fix. (CVE-2011-4108)
1369 [Robin Seggelmann, Michael Tuexen]
1371 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1372 [Ben Laurie, Kasper <ekasper@google.com>]
1374 *) Clear bytes used for block padding of SSL 3.0 records.
1376 [Adam Langley (Google)]
1378 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1379 Kadianakis <desnacked@gmail.com> for discovering this issue and
1380 Adam Langley for preparing the fix. (CVE-2011-4619)
1381 [Adam Langley (Google)]
1383 *) Prevent malformed RFC3779 data triggering an assertion failure.
1384 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1385 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1386 [Rob Austein <sra@hactrn.net>]
1388 *) Fix ssl_ciph.c set-up race.
1389 [Adam Langley (Google)]
1391 *) Fix spurious failures in ecdsatest.c.
1392 [Emilia Käsper (Google)]
1394 *) Fix the BIO_f_buffer() implementation (which was mixing different
1395 interpretations of the '..._len' fields).
1396 [Adam Langley (Google)]
1398 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1399 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1400 threads won't reuse the same blinding coefficients.
1402 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1403 lock to call BN_BLINDING_invert_ex, and avoids one use of
1404 BN_BLINDING_update for each BN_BLINDING structure (previously,
1405 the last update always remained unused).
1406 [Emilia Käsper (Google)]
1408 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1409 for multi-threaded use of ECDH.
1410 [Adam Langley (Google)]
1412 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1415 *) Add protection against ECDSA timing attacks as mentioned in the paper
1416 by Billy Bob Brumley and Nicola Tuveri, see:
1418 http://eprint.iacr.org/2011/232.pdf
1420 [Billy Bob Brumley and Nicola Tuveri]
1422 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1424 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1425 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1427 *) Fix bug in string printing code: if *any* escaping is enabled we must
1428 escape the escape character (backslash) or the resulting string is
1432 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1434 *) Disable code workaround for ancient and obsolete Netscape browsers
1435 and servers: an attacker can use it in a ciphersuite downgrade attack.
1436 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1439 *) Fixed J-PAKE implementation error, originally discovered by
1440 Sebastien Martini, further info and confirmation from Stefan
1441 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1444 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1446 *) Fix extension code to avoid race conditions which can result in a buffer
1447 overrun vulnerability: resumed sessions must not be modified as they can
1448 be shared by multiple threads. CVE-2010-3864
1451 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1454 *) Don't reencode certificate when calculating signature: cache and use
1455 the original encoding instead. This makes signature verification of
1456 some broken encodings work correctly.
1459 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1460 is also one of the inputs.
1461 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1463 *) Don't repeatedly append PBE algorithms to table if they already exist.
1464 Sort table on each new add. This effectively makes the table read only
1465 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1469 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1471 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1474 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1475 access or freeing data twice (CVE-2010-0742)
1476 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1478 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1479 common in certificates and some applications which only call
1480 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1484 Reduce copying into .apps and .test in makevms.com
1485 Don't try to use blank CA certificate in CA.com
1486 Allow use of C files from original directories in maketests.com
1487 [Steven M. Schweda" <sms@antinode.info>]
1489 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1491 *) When rejecting SSL/TLS records due to an incorrect version number, never
1492 update s->server with a new major version number. As of
1493 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1494 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1495 the previous behavior could result in a read attempt at NULL when
1496 receiving specific incorrect SSL/TLS records once record payload
1497 protection is active. (CVE-2010-0740)
1498 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1500 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1501 could be crashed if the relevant tables were not present (e.g. chrooted).
1502 [Tomas Hoger <thoger@redhat.com>]
1504 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1506 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1507 [Martin Olsson, Neel Mehta]
1509 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1510 accommodate for stack sorting, always a write lock!).
1513 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1514 excessive delays in the RAND_poll(): over a minute. As a workaround
1515 include a time check in the inner Heap32Next loop too.
1518 *) The code that handled flushing of data in SSL/TLS originally used the
1519 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1520 the problem outlined in PR#1949. The fix suggested there however can
1521 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1522 of Apache). So instead simplify the code to flush unconditionally.
1523 This should be fine since flushing with no data to flush is a no op.
1526 *) Handle TLS versions 2.0 and later properly and correctly use the
1527 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1528 off ancient servers have a habit of sticking around for a while...
1531 *) Modify compression code so it frees up structures without using the
1532 ex_data callbacks. This works around a problem where some applications
1533 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1534 restarting) then use compression (e.g. SSL with compression) later.
1535 This results in significant per-connection memory leaks and
1536 has caused some security issues including CVE-2008-1678 and
1540 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1541 change when encrypting or decrypting.
1544 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1545 connect and renegotiate with servers which do not support RI.
1546 Until RI is more widely deployed this option is enabled by default.
1549 *) Add "missing" ssl ctrls to clear options and mode.
1552 *) If client attempts to renegotiate and doesn't support RI respond with
1553 a no_renegotiation alert as required by RFC5746. Some renegotiating
1554 TLS clients will continue a connection gracefully when they receive
1555 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1556 waiting for a server hello which it will never receive. Now we treat a
1557 received no_renegotiation alert as a fatal error. This is because
1558 applications requesting a renegotiation might well expect it to succeed
1559 and would have no code in place to handle the server denying it so the
1560 only safe thing to do is to terminate the connection.
1563 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1564 peer supports secure renegotiation and 0 otherwise. Print out peer
1565 renegotiation support in s_client/s_server.
1568 *) Replace the highly broken and deprecated SPKAC certification method with
1569 the updated NID creation version. This should correctly handle UTF8.
1572 *) Implement RFC5746. Re-enable renegotiation but require the extension
1573 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1574 turns out to be a bad idea. It has been replaced by
1575 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1576 SSL_CTX_set_options(). This is really not recommended unless you
1577 know what you are doing.
1578 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1580 *) Fixes to stateless session resumption handling. Use initial_ctx when
1581 issuing and attempting to decrypt tickets in case it has changed during
1582 servername handling. Use a non-zero length session ID when attempting
1583 stateless session resumption: this makes it possible to determine if
1584 a resumption has occurred immediately after receiving server hello
1585 (several places in OpenSSL subtly assume this) instead of later in
1589 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1590 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1591 fixes for a few places where the return code is not checked
1593 [Julia Lawall <julia@diku.dk>]
1595 *) Add --strict-warnings option to Configure script to include devteam
1596 warnings in other configurations.
1599 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1600 makes it possible to install openssl libraries in locations which
1601 have names other than "lib", for example "/usr/lib64" which some
1603 [Steve Henson, based on patch from Jeremy Utley]
1605 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1606 X690 8.9.12 and can produce some misleading textual output of OIDs.
1607 [Steve Henson, reported by Dan Kaminsky]
1609 *) Delete MD2 from algorithm tables. This follows the recommendation in
1610 several standards that it is not used in new applications due to
1611 several cryptographic weaknesses. For binary compatibility reasons
1612 the MD2 API is still compiled in by default.
1615 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1619 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1620 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1622 [Guenter <lists@gknw.net>]
1624 *) Fix the server certificate chain building code to use X509_verify_cert(),
1625 it used to have an ad-hoc builder which was unable to cope with anything
1626 other than a simple chain.
1627 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1629 *) Don't check self signed certificate signatures in X509_verify_cert()
1630 by default (a flag can override this): it just wastes time without
1631 adding any security. As a useful side effect self signed root CAs
1632 with non-FIPS digests are now usable in FIPS mode.
1635 *) In dtls1_process_out_of_seq_message() the check if the current message
1636 is already buffered was missing. For every new message was memory
1637 allocated, allowing an attacker to perform an denial of service attack
1638 with sending out of seq handshake messages until there is no memory
1639 left. Additionally every future messege was buffered, even if the
1640 sequence number made no sense and would be part of another handshake.
1641 So only messages with sequence numbers less than 10 in advance will be
1642 buffered. (CVE-2009-1378)
1643 [Robin Seggelmann, discovered by Daniel Mentz]
1645 *) Records are buffered if they arrive with a future epoch to be
1646 processed after finishing the corresponding handshake. There is
1647 currently no limitation to this buffer allowing an attacker to perform
1648 a DOS attack with sending records with future epochs until there is no
1649 memory left. This patch adds the pqueue_size() function to detemine
1650 the size of a buffer and limits the record buffer to 100 entries.
1652 [Robin Seggelmann, discovered by Daniel Mentz]
1654 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1655 parent structure is freed. (CVE-2009-1379)
1658 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1659 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1662 [Ilya O. <vrghost@gmail.com>]
1664 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1666 *) Disable renegotiation completely - this fixes a severe security
1667 problem (CVE-2009-3555) at the cost of breaking all
1668 renegotiation. Renegotiation can be re-enabled by setting
1669 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1670 run-time. This is really not recommended unless you know what
1674 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1676 *) Don't set val to NULL when freeing up structures, it is freed up by
1677 underlying code. If sizeof(void *) > sizeof(long) this can result in
1678 zeroing past the valid field. (CVE-2009-0789)
1679 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1681 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1682 checked correctly. This would allow some invalid signed attributes to
1683 appear to verify correctly. (CVE-2009-0591)
1684 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1686 *) Reject UniversalString and BMPString types with invalid lengths. This
1687 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1688 a legal length. (CVE-2009-0590)
1691 *) Set S/MIME signing as the default purpose rather than setting it
1692 unconditionally. This allows applications to override it at the store
1696 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1697 to handle some structures.
1700 *) Improve efficiency of mem_gets: don't search whole buffer each time
1702 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1704 *) New -hex option for openssl rand.
1707 *) Print out UTF8String and NumericString when parsing ASN1.
1710 *) Support NumericString type for name components.
1713 *) Allow CC in the environment to override the automatically chosen
1714 compiler. Note that nothing is done to ensure flags work with the
1718 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1720 *) Properly check EVP_VerifyFinal() and similar return values
1722 [Ben Laurie, Bodo Moeller, Google Security Team]
1724 *) Enable TLS extensions by default.
1727 *) Allow the CHIL engine to be loaded, whether the application is
1728 multithreaded or not. (This does not release the developer from the
1729 obligation to set up the dynamic locking callbacks.)
1730 [Sander Temme <sander@temme.net>]
1732 *) Use correct exit code if there is an error in dgst command.
1733 [Steve Henson; problem pointed out by Roland Dirlewanger]
1735 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1736 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1739 *) Add experimental JPAKE support, including demo authentication in
1740 s_client and s_server.
1743 *) Set the comparison function in v3_addr_canonize().
1744 [Rob Austein <sra@hactrn.net>]
1746 *) Add support for XMPP STARTTLS in s_client.
1747 [Philip Paeps <philip@freebsd.org>]
1749 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1750 to ensure that even with this option, only ciphersuites in the
1751 server's preference list will be accepted. (Note that the option
1752 applies only when resuming a session, so the earlier behavior was
1753 just about the algorithm choice for symmetric cryptography.)
1756 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1758 *) Fix NULL pointer dereference if a DTLS server received
1759 ChangeCipherSpec as first record (CVE-2009-1386).
1762 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1763 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1766 *) The fix in 0.9.8c that supposedly got rid of unsafe
1767 double-checked locking was incomplete for RSA blinding,
1768 addressing just one layer of what turns out to have been
1769 doubly unsafe triple-checked locking.
1771 So now fix this for real by retiring the MONT_HELPER macro
1772 in crypto/rsa/rsa_eay.c.
1774 [Bodo Moeller; problem pointed out by Marius Schilder]
1776 *) Various precautionary measures:
1778 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1780 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1781 (NB: This would require knowledge of the secret session ticket key
1782 to exploit, in which case you'd be SOL either way.)
1784 - Change bn_nist.c so that it will properly handle input BIGNUMs
1785 outside the expected range.
1787 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1790 [Neel Mehta, Bodo Moeller]
1792 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1793 the load fails. Useful for distros.
1794 [Ben Laurie and the FreeBSD team]
1796 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1799 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1802 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1804 This work was sponsored by Logica.
1807 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1808 keystores. Support for SSL/TLS client authentication too.
1809 Not compiled unless enable-capieng specified to Configure.
1811 This work was sponsored by Logica.
1814 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1815 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1816 attribute creation routines such as certifcate requests and PKCS#12
1820 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1822 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1823 handshake which could lead to a cilent crash as found using the
1824 Codenomicon TLS test suite (CVE-2008-1672)
1825 [Steve Henson, Mark Cox]
1827 *) Fix double free in TLS server name extensions which could lead to
1828 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1831 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1833 Clear the error queue to ensure that error entries left from
1834 older function calls do not interfere with the correct operation.
1835 [Lutz Jaenicke, Erik de Castro Lopo]
1837 *) Remove root CA certificates of commercial CAs:
1839 The OpenSSL project does not recommend any specific CA and does not
1840 have any policy with respect to including or excluding any CA.
1841 Therefore it does not make any sense to ship an arbitrary selection
1842 of root CA certificates with the OpenSSL software.
1845 *) RSA OAEP patches to fix two separate invalid memory reads.
1846 The first one involves inputs when 'lzero' is greater than
1847 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1848 before the beginning of from). The second one involves inputs where
1849 the 'db' section contains nothing but zeroes (there is a one-byte
1850 invalid read after the end of 'db').
1851 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1853 *) Partial backport from 0.9.9-dev:
1855 Introduce bn_mul_mont (dedicated Montgomery multiplication
1856 procedure) as a candidate for BIGNUM assembler implementation.
1857 While 0.9.9-dev uses assembler for various architectures, only
1858 x86_64 is available by default here in the 0.9.8 branch, and
1859 32-bit x86 is available through a compile-time setting.
1861 To try the 32-bit x86 assembler implementation, use Configure
1862 option "enable-montasm" (which exists only for this backport).
1864 As "enable-montasm" for 32-bit x86 disclaims code stability
1865 anyway, in this constellation we activate additional code
1866 backported from 0.9.9-dev for further performance improvements,
1867 namely BN_from_montgomery_word. (To enable this otherwise,
1868 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1870 [Andy Polyakov (backport partially by Bodo Moeller)]
1872 *) Add TLS session ticket callback. This allows an application to set
1873 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1874 values. This is useful for key rollover for example where several key
1875 sets may exist with different names.
1878 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1879 This was broken until now in 0.9.8 releases, such that the only way
1880 a registered ENGINE could be used (assuming it initialises
1881 successfully on the host) was to explicitly set it as the default
1882 for the relevant algorithms. This is in contradiction with 0.9.7
1883 behaviour and the documentation. With this fix, when an ENGINE is
1884 registered into a given algorithm's table of implementations, the
1885 'uptodate' flag is reset so that auto-discovery will be used next
1886 time a new context for that algorithm attempts to select an
1888 [Ian Lister (tweaked by Geoff Thorpe)]
1890 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1891 implemention in the following ways:
1893 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1896 Lack of BER streaming support means one pass streaming processing is
1897 only supported if data is detached: setting the streaming flag is
1898 ignored for embedded content.
1900 CMS support is disabled by default and must be explicitly enabled
1901 with the enable-cms configuration option.
1904 *) Update the GMP engine glue to do direct copies between BIGNUM and
1905 mpz_t when openssl and GMP use the same limb size. Otherwise the
1906 existing "conversion via a text string export" trick is still used.
1907 [Paul Sheer <paulsheer@gmail.com>]
1909 *) Zlib compression BIO. This is a filter BIO which compressed and
1910 uncompresses any data passed through it.
1913 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1914 RFC3394 compatible AES key wrapping.
1917 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1918 sets string data without copying. X509_ALGOR_set0() and
1919 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1920 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1921 from an X509_ATTRIBUTE structure optionally checking it occurs only
1922 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1926 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1927 to get the expected BN_FLG_CONSTTIME behavior.
1928 [Bodo Moeller (Google)]
1932 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1933 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1934 - added some more tests to do_tests.pl
1935 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1936 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1937 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1938 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1939 - various changes to netware.pl to enable gcc-cross builds on Win32
1941 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1942 - various changes to fix missing prototype warnings
1943 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1944 - added AES, WHIRLPOOL and CPUID assembler code to build files
1945 - added missing AES assembler make rules to mk1mf.pl
1946 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1947 [Guenter Knauf <eflash@gmx.net>]
1949 *) Implement certificate status request TLS extension defined in RFC3546.
1950 A client can set the appropriate parameters and receive the encoded
1951 OCSP response via a callback. A server can query the supplied parameters
1952 and set the encoded OCSP response in the callback. Add simplified examples
1953 to s_client and s_server.
1956 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1958 *) Fix various bugs:
1959 + Binary incompatibility of ssl_ctx_st structure
1960 + DTLS interoperation with non-compliant servers
1961 + Don't call get_session_cb() without proposed session
1962 + Fix ia64 assembler code
1963 [Andy Polyakov, Steve Henson]
1965 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1967 *) DTLS Handshake overhaul. There were longstanding issues with
1968 OpenSSL DTLS implementation, which were making it impossible for
1969 RFC 4347 compliant client to communicate with OpenSSL server.
1970 Unfortunately just fixing these incompatibilities would "cut off"
1971 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1972 server keeps tolerating non RFC compliant syntax. The opposite is
1973 not true, 0.9.8f client can not communicate with earlier server.
1974 This update even addresses CVE-2007-4995.
1977 *) Changes to avoid need for function casts in OpenSSL: some compilers
1978 (gcc 4.2 and later) reject their use.
1979 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1982 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1983 RFC4507bis. The encrypted ticket format is an encrypted encoded
1984 SSL_SESSION structure, that way new session features are automatically
1987 If a client application caches session in an SSL_SESSION structure
1988 support is transparent because tickets are now stored in the encoded
1991 The SSL_CTX structure automatically generates keys for ticket
1992 protection in servers so again support should be possible
1993 with no application modification.
1995 If a client or server wishes to disable RFC4507 support then the option
1996 SSL_OP_NO_TICKET can be set.
1998 Add a TLS extension debugging callback to allow the contents of any client
1999 or server extensions to be examined.
2001 This work was sponsored by Google.
2004 *) Add initial support for TLS extensions, specifically for the server_name
2005 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2006 have new members for a host name. The SSL data structure has an
2007 additional member SSL_CTX *initial_ctx so that new sessions can be
2008 stored in that context to allow for session resumption, even after the
2009 SSL has been switched to a new SSL_CTX in reaction to a client's
2010 server_name extension.
2012 New functions (subject to change):
2014 SSL_get_servername()
2015 SSL_get_servername_type()
2018 New CTRL codes and macros (subject to change):
2020 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2021 - SSL_CTX_set_tlsext_servername_callback()
2022 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2023 - SSL_CTX_set_tlsext_servername_arg()
2024 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2026 openssl s_client has a new '-servername ...' option.
2028 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2029 '-key2 ...', '-servername_fatal' (subject to change). This allows
2030 testing the HostName extension for a specific single host name ('-cert'
2031 and '-key' remain fallbacks for handshakes without HostName
2032 negotiation). If the unrecogninzed_name alert has to be sent, this by
2033 default is a warning; it becomes fatal with the '-servername_fatal'
2036 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2038 *) Add AES and SSE2 assembly language support to VC++ build.
2041 *) Mitigate attack on final subtraction in Montgomery reduction.
2044 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2045 (which previously caused an internal error).
2048 *) Squeeze another 10% out of IGE mode when in != out.
2051 *) AES IGE mode speedup.
2052 [Dean Gaudet (Google)]
2054 *) Add the Korean symmetric 128-bit cipher SEED (see
2055 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2056 add SEED ciphersuites from RFC 4162:
2058 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2059 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2060 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2061 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2063 To minimize changes between patchlevels in the OpenSSL 0.9.8
2064 series, SEED remains excluded from compilation unless OpenSSL
2065 is configured with 'enable-seed'.
2066 [KISA, Bodo Moeller]
2068 *) Mitigate branch prediction attacks, which can be practical if a
2069 single processor is shared, allowing a spy process to extract
2070 information. For detailed background information, see
2071 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2072 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2073 and Necessary Software Countermeasures"). The core of the change
2074 are new versions BN_div_no_branch() and
2075 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2076 respectively, which are slower, but avoid the security-relevant
2077 conditional branches. These are automatically called by BN_div()
2078 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2079 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2080 remove a conditional branch.
2082 BN_FLG_CONSTTIME is the new name for the previous
2083 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2084 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2085 in the exponent causes BN_mod_exp_mont() to use the alternative
2086 implementation in BN_mod_exp_mont_consttime().) The old name
2087 remains as a deprecated alias.
2089 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2090 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2091 constant-time implementations for more than just exponentiation.
2092 Here too the old name is kept as a deprecated alias.
2094 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2095 the BN_BLINDING structure gets an independent copy of the
2096 modulus. This means that the previous "BIGNUM *m" argument to
2097 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2098 essentially becomes "const BIGNUM *m", although we can't actually
2099 change this in the header file before 0.9.9. It allows
2100 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2101 enable BN_FLG_CONSTTIME.
2103 [Matthew D Wood (Intel Corp)]
2105 *) In the SSL/TLS server implementation, be strict about session ID
2106 context matching (which matters if an application uses a single
2107 external cache for different purposes). Previously,
2108 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2109 set. This did ensure strict client verification, but meant that,
2110 with applications using a single external cache for quite
2111 different requirements, clients could circumvent ciphersuite
2112 restrictions for a given session ID context by starting a session
2113 in a different context.
2116 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2117 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2118 authentication-only ciphersuites.
2121 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2122 not complete and could lead to a possible single byte overflow
2123 (CVE-2007-5135) [Ben Laurie]
2125 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2127 *) Since AES128 and AES256 (and similarly Camellia128 and
2128 Camellia256) share a single mask bit in the logic of
2129 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2130 kludge to work properly if AES128 is available and AES256 isn't
2131 (or if Camellia128 is available and Camellia256 isn't).
2134 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2135 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2136 When a point or a seed is encoded in a BIT STRING, we need to
2137 prevent the removal of trailing zero bits to get the proper DER
2138 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2139 of a NamedBitList, for which trailing 0 bits need to be removed.)
2142 *) Have SSL/TLS server implementation tolerate "mismatched" record
2143 protocol version while receiving ClientHello even if the
2144 ClientHello is fragmented. (The server can't insist on the
2145 particular protocol version it has chosen before the ServerHello
2146 message has informed the client about his choice.)
2149 *) Add RFC 3779 support.
2150 [Rob Austein for ARIN, Ben Laurie]
2152 *) Load error codes if they are not already present instead of using a
2153 static variable. This allows them to be cleanly unloaded and reloaded.
2154 Improve header file function name parsing.
2157 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2158 or CAPABILITY handshake as required by RFCs.
2161 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2163 *) Introduce limits to prevent malicious keys being able to
2164 cause a denial of service. (CVE-2006-2940)
2165 [Steve Henson, Bodo Moeller]
2167 *) Fix ASN.1 parsing of certain invalid structures that can result
2168 in a denial of service. (CVE-2006-2937) [Steve Henson]
2170 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2171 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2173 *) Fix SSL client code which could crash if connecting to a
2174 malicious SSLv2 server. (CVE-2006-4343)
2175 [Tavis Ormandy and Will Drewry, Google Security Team]
2177 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2178 match only those. Before that, "AES256-SHA" would be interpreted
2179 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2180 the same strength classification in 0.9.7h) as we currently only
2181 have a single AES bit in the ciphersuite description bitmap.
2182 That change, however, also applied to ciphersuite strings such as
2183 "RC4-MD5" that intentionally matched multiple ciphersuites --
2184 namely, SSL 2.0 ciphersuites in addition to the more common ones
2185 from SSL 3.0/TLS 1.0.
2187 So we change the selection algorithm again: Naming an explicit
2188 ciphersuite selects this one ciphersuite, and any other similar
2189 ciphersuite (same bitmap) from *other* protocol versions.
2190 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2191 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2193 Since SSL 2.0 does not have any ciphersuites for which the
2194 128/256 bit distinction would be relevant, this works for now.
2195 The proper fix will be to use different bits for AES128 and
2196 AES256, which would have avoided the problems from the beginning;
2197 however, bits are scarce, so we can only do this in a new release
2198 (not just a patchlevel) when we can change the SSL_CIPHER
2199 definition to split the single 'unsigned long mask' bitmap into
2200 multiple values to extend the available space.
2204 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2206 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2207 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2209 *) Add AES IGE and biIGE modes.
2212 *) Change the Unix randomness entropy gathering to use poll() when
2213 possible instead of select(), since the latter has some
2214 undesirable limitations.
2215 [Darryl Miles via Richard Levitte and Bodo Moeller]
2217 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2218 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2219 cannot be implicitly activated as part of, e.g., the "AES" alias.
2220 However, please upgrade to OpenSSL 0.9.9[-dev] for
2221 non-experimental use of the ECC ciphersuites to get TLS extension
2222 support, which is required for curve and point format negotiation
2223 to avoid potential handshake problems.
2226 *) Disable rogue ciphersuites:
2228 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2229 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2230 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2232 The latter two were purportedly from
2233 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2236 Also deactivate the remaining ciphersuites from
2237 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2238 unofficial, and the ID has long expired.
2241 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2242 dual-core machines) and other potential thread-safety issues.
2245 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2246 versions), which is now available for royalty-free use
2247 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2248 Also, add Camellia TLS ciphersuites from RFC 4132.
2250 To minimize changes between patchlevels in the OpenSSL 0.9.8
2251 series, Camellia remains excluded from compilation unless OpenSSL
2252 is configured with 'enable-camellia'.
2255 *) Disable the padding bug check when compression is in use. The padding
2256 bug check assumes the first packet is of even length, this is not
2257 necessarily true if compresssion is enabled and can result in false
2258 positives causing handshake failure. The actual bug test is ancient
2259 code so it is hoped that implementations will either have fixed it by
2260 now or any which still have the bug do not support compression.
2263 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2265 *) When applying a cipher rule check to see if string match is an explicit
2266 cipher suite and only match that one cipher suite if it is.
2269 *) Link in manifests for VC++ if needed.
2270 [Austin Ziegler <halostatue@gmail.com>]
2272 *) Update support for ECC-based TLS ciphersuites according to
2273 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2274 TLS extensions, which are supported starting with the 0.9.9
2275 branch, not in the OpenSSL 0.9.8 branch).
2278 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2279 opaque EVP_CIPHER_CTX handling.
2282 *) Fixes and enhancements to zlib compression code. We now only use
2283 "zlib1.dll" and use the default __cdecl calling convention on Win32
2284 to conform with the standards mentioned here:
2285 http://www.zlib.net/DLL_FAQ.txt
2286 Static zlib linking now works on Windows and the new --with-zlib-include
2287 --with-zlib-lib options to Configure can be used to supply the location
2288 of the headers and library. Gracefully handle case where zlib library
2292 *) Several fixes and enhancements to the OID generation code. The old code
2293 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2294 handle numbers larger than ULONG_MAX, truncated printing and had a
2295 non standard OBJ_obj2txt() behaviour.
2298 *) Add support for building of engines under engine/ as shared libraries
2299 under VC++ build system.
2302 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2303 Hopefully, we will not see any false combination of paths any more.
2306 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2308 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2309 (part of SSL_OP_ALL). This option used to disable the
2310 countermeasure against man-in-the-middle protocol-version
2311 rollback in the SSL 2.0 server implementation, which is a bad
2312 idea. (CVE-2005-2969)
2314 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2315 for Information Security, National Institute of Advanced Industrial
2316 Science and Technology [AIST], Japan)]
2318 *) Add two function to clear and return the verify parameter flags.
2321 *) Keep cipherlists sorted in the source instead of sorting them at
2322 runtime, thus removing the need for a lock.
2325 *) Avoid some small subgroup attacks in Diffie-Hellman.
2326 [Nick Mathewson and Ben Laurie]
2328 *) Add functions for well-known primes.
2331 *) Extended Windows CE support.
2332 [Satoshi Nakamura and Andy Polyakov]
2334 *) Initialize SSL_METHOD structures at compile time instead of during
2335 runtime, thus removing the need for a lock.
2338 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2339 attempting to decrypt each encrypted key in turn. Add support to
2343 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2345 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2348 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2351 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2352 key into the same file any more.
2355 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2358 *) Add -utf8 command line and config file option to 'ca'.
2359 [Stefan <stf@udoma.org]
2361 *) Removed the macro des_crypt(), as it seems to conflict with some
2362 libraries. Use DES_crypt().
2365 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2366 involves renaming the source and generated shared-libs for
2367 both. The engines will accept the corrected or legacy ids
2368 ('ncipher' and '4758_cca' respectively) when binding. NB,
2369 this only applies when building 'shared'.
2370 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2372 *) Add attribute functions to EVP_PKEY structure. Modify
2373 PKCS12_create() to recognize a CSP name attribute and
2374 use it. Make -CSP option work again in pkcs12 utility.
2377 *) Add new functionality to the bn blinding code:
2378 - automatic re-creation of the BN_BLINDING parameters after
2379 a fixed number of uses (currently 32)
2380 - add new function for parameter creation
2381 - introduce flags to control the update behaviour of the
2382 BN_BLINDING parameters
2383 - hide BN_BLINDING structure
2384 Add a second BN_BLINDING slot to the RSA structure to improve
2385 performance when a single RSA object is shared among several
2389 *) Add support for DTLS.
2390 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2392 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2393 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2396 *) Remove buggy and incompletet DH cert support from
2397 ssl/ssl_rsa.c and ssl/s3_both.c
2400 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2401 the apps/openssl applications.
2404 *) Compile clean with "-Wall -Wmissing-prototypes
2405 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2406 DEBUG_SAFESTACK must also be set.
2409 *) Change ./Configure so that certain algorithms can be disabled by default.
2410 The new counterpiece to "no-xxx" is "enable-xxx".
2412 The patented RC5 and MDC2 algorithms will now be disabled unless
2413 "enable-rc5" and "enable-mdc2", respectively, are specified.
2415 (IDEA remains enabled despite being patented. This is because IDEA
2416 is frequently required for interoperability, and there is no license
2417 fee for non-commercial use. As before, "no-idea" can be used to
2418 avoid this algorithm.)
2422 *) Add processing of proxy certificates (see RFC 3820). This work was
2423 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2424 EGEE (Enabling Grids for E-science in Europe).
2427 *) RC4 performance overhaul on modern architectures/implementations, such
2428 as Intel P4, IA-64 and AMD64.
2431 *) New utility extract-section.pl. This can be used specify an alternative
2432 section number in a pod file instead of having to treat each file as
2433 a separate case in Makefile. This can be done by adding two lines to the
2436 =for comment openssl_section:XXX
2438 The blank line is mandatory.
2442 *) New arguments -certform, -keyform and -pass for s_client and s_server
2443 to allow alternative format key and certificate files and passphrase
2447 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2448 update associated structures and add various utility functions.
2450 Add new policy related verify parameters, include policy checking in
2451 standard verify code. Enhance 'smime' application with extra parameters
2452 to support policy checking and print out.
2455 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2456 Nehemiah processors. These extensions support AES encryption in hardware
2457 as well as RNG (though RNG support is currently disabled).
2458 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2460 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2463 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2464 [Andy Polyakov and a number of other people]
2466 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2467 implementation contributed by IBM.
2468 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2470 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2471 exponent rather than 'unsigned long'. There is a corresponding change to
2472 the new 'rsa_keygen' element of the RSA_METHOD structure.
2473 [Jelte Jansen, Geoff Thorpe]
2475 *) Functionality for creating the initial serial number file is now
2476 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2478 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2479 number file to 1, which is bound to cause problems. To avoid
2480 the problems while respecting compatibility between different 0.9.7
2481 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2482 CA.pl for serial number initialization. With the new release 0.9.8,
2483 we can fix the problem directly in the 'ca' utility.)
2486 *) Reduced header interdepencies by declaring more opaque objects in
2487 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2488 give fewer recursive includes, which could break lazy source code - so
2489 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2490 developers should define this symbol when building and using openssl to
2491 ensure they track the recommended behaviour, interfaces, [etc], but
2492 backwards-compatible behaviour prevails when this isn't defined.
2495 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2498 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2499 This will generate a random key of the appropriate length based on the
2500 cipher context. The EVP_CIPHER can provide its own random key generation
2501 routine to support keys of a specific form. This is used in the des and
2502 3des routines to generate a key of the correct parity. Update S/MIME
2503 code to use new functions and hence generate correct parity DES keys.
2504 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2505 valid (weak or incorrect parity).
2508 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2509 as looking them up. This is useful when the verified structure may contain
2510 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2511 present unless the new PKCS7_NO_CRL flag is asserted.
2514 *) Extend ASN1 oid configuration module. It now additionally accepts the
2517 shortName = some long name, 1.2.3.4
2520 *) Reimplemented the BN_CTX implementation. There is now no more static
2521 limitation on the number of variables it can handle nor the depth of the
2522 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2523 information can now expand as required, and rather than having a single
2524 static array of bignums, BN_CTX now uses a linked-list of such arrays
2525 allowing it to expand on demand whilst maintaining the usefulness of
2526 BN_CTX's "bundling".
2529 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2530 to allow all RSA operations to function using a single BN_CTX.
2533 *) Preliminary support for certificate policy evaluation and checking. This
2534 is initially intended to pass the tests outlined in "Conformance Testing
2535 of Relying Party Client Certificate Path Processing Logic" v1.07.
2538 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2539 remained unused and not that useful. A variety of other little bignum
2540 tweaks and fixes have also been made continuing on from the audit (see
2544 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2545 associated ASN1, EVP and SSL functions and old ASN1 macros.
2548 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2549 and this should never fail. So the return value from the use of
2550 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2551 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2554 *) BN_CTX_get() should return zero-valued bignums, providing the same
2555 initialised value as BN_new().
2556 [Geoff Thorpe, suggested by Ulf Möller]
2558 *) Support for inhibitAnyPolicy certificate extension.
2561 *) An audit of the BIGNUM code is underway, for which debugging code is
2562 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2563 is considered valid when processing BIGNUMs, and causes execution to
2564 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2565 further steps are taken to deliberately pollute unused data in BIGNUM
2566 structures to try and expose faulty code further on. For now, openssl will
2567 (in its default mode of operation) continue to tolerate the inconsistent
2568 forms that it has tolerated in the past, but authors and packagers should
2569 consider trying openssl and their own applications when compiled with
2570 these debugging symbols defined. It will help highlight potential bugs in
2571 their own code, and will improve the test coverage for OpenSSL itself. At
2572 some point, these tighter rules will become openssl's default to improve
2573 maintainability, though the assert()s and other overheads will remain only
2574 in debugging configurations. See bn.h for more details.
2575 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2577 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2578 that can only be obtained through BN_CTX_new() (which implicitly
2579 initialises it). The presence of this function only made it possible
2580 to overwrite an existing structure (and cause memory leaks).
2583 *) Because of the callback-based approach for implementing LHASH as a
2584 template type, lh_insert() adds opaque objects to hash-tables and
2585 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2586 to clean up those corresponding objects before destroying the hash table
2587 (and losing the object pointers). So some over-zealous constifications in
2588 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2589 objects as "const" and the lh_doall[_arg] callback wrappers are not
2590 prototyped to have "const" restrictions on the object pointers they are
2591 given (and so aren't required to cast them away any more).
2594 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2595 (speed) prefers to use its own implementation. The two implementations
2596 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2597 its object type properly exposed (MS_TM) instead of casting to/from "char
2598 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2599 aren't necessarily the greatest nomenclatures - but this is what was used
2600 internally to the implementation so I've used that for now.
2603 *) Ensure that deprecated functions do not get compiled when
2604 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2605 the self-tests were still using deprecated key-generation functions so
2606 these have been updated also.
2609 *) Reorganise PKCS#7 code to separate the digest location functionality
2610 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2611 New function PKCS7_set_digest() to set the digest type for PKCS#7
2612 digestedData type. Add additional code to correctly generate the
2613 digestedData type and add support for this type in PKCS7 initialization
2617 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2618 structure of type "other".
2621 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2622 sure the loop does correctly stop and breaking ("division by zero")
2623 modulus operations are not performed. The (pre-generated) prime
2624 table crypto/bn/bn_prime.h was already correct, but it could not be
2625 re-generated on some platforms because of the "division by zero"
2626 situation in the script.
2627 [Ralf S. Engelschall]
2629 *) Update support for ECC-based TLS ciphersuites according to
2630 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2631 SHA-1 now is only used for "small" curves (where the
2632 representation of a field element takes up to 24 bytes); for
2633 larger curves, the field element resulting from ECDH is directly
2634 used as premaster secret.
2635 [Douglas Stebila (Sun Microsystems Laboratories)]
2637 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2638 curve secp160r1 to the tests.
2639 [Douglas Stebila (Sun Microsystems Laboratories)]
2641 *) Add the possibility to load symbols globally with DSO.
2642 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2644 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2645 control of the error stack.
2648 *) Add support for STORE in ENGINE.
2651 *) Add the STORE type. The intention is to provide a common interface
2652 to certificate and key stores, be they simple file-based stores, or
2653 HSM-type store, or LDAP stores, or...
2654 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2657 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2658 pass a list of arguments to any function as well as provide a way
2659 for a function to pass data back to the caller.
2662 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2663 works like BUF_strdup() but can be used to duplicate a portion of
2664 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2668 *) Add the function sk_find_ex() which works like sk_find(), but will
2669 return an index to an element even if an exact match couldn't be
2670 found. The index is guaranteed to point at the element where the
2671 searched-for key would be inserted to preserve sorting order.
2674 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2675 takes an extra flags argument for optional functionality. Currently,
2676 the following flags are defined:
2678 OBJ_BSEARCH_VALUE_ON_NOMATCH
2679 This one gets OBJ_bsearch_ex() to return a pointer to the first
2680 element where the comparing function returns a negative or zero
2683 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2684 This one gets OBJ_bsearch_ex() to return a pointer to the first
2685 element where the comparing function returns zero. This is useful
2686 if there are more than one element where the comparing function
2690 *) Make it possible to create self-signed certificates with 'openssl ca'
2691 in such a way that the self-signed certificate becomes part of the
2692 CA database and uses the same mechanisms for serial number generation
2693 as all other certificate signing. The new flag '-selfsign' enables
2694 this functionality. Adapt CA.sh and CA.pl.in.
2697 *) Add functionality to check the public key of a certificate request
2698 against a given private. This is useful to check that a certificate
2699 request can be signed by that key (self-signing).
2702 *) Make it possible to have multiple active certificates with the same
2703 subject in the CA index file. This is done only if the keyword
2704 'unique_subject' is set to 'no' in the main CA section (default
2705 if 'CA_default') of the configuration file. The value is saved
2706 with the database itself in a separate index attribute file,
2707 named like the index file with '.attr' appended to the name.
2710 *) Generate muti valued AVAs using '+' notation in config files for
2714 *) Support for nameConstraints certificate extension.
2717 *) Support for policyConstraints certificate extension.
2720 *) Support for policyMappings certificate extension.
2723 *) Make sure the default DSA_METHOD implementation only uses its
2724 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2725 and change its own handlers to be NULL so as to remove unnecessary
2726 indirection. This lets alternative implementations fallback to the
2727 default implementation more easily.
2730 *) Support for directoryName in GeneralName related extensions
2734 *) Make it possible to link applications using Makefile.shared.
2735 Make that possible even when linking against static libraries!
2738 *) Support for single pass processing for S/MIME signing. This now
2739 means that S/MIME signing can be done from a pipe, in addition
2740 cleartext signing (multipart/signed type) is effectively streaming
2741 and the signed data does not need to be all held in memory.
2743 This is done with a new flag PKCS7_STREAM. When this flag is set
2744 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2745 is done after the data is output (and digests calculated) in
2746 SMIME_write_PKCS7().
2749 *) Add full support for -rpath/-R, both in shared libraries and
2750 applications, at least on the platforms where it's known how
2754 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2755 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2756 will now compute a table of multiples of the generator that
2757 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2758 faster (notably in the case of a single point multiplication,
2759 scalar * generator).
2760 [Nils Larsch, Bodo Moeller]
2762 *) IPv6 support for certificate extensions. The various extensions
2763 which use the IP:a.b.c.d can now take IPv6 addresses using the
2764 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2768 *) Added an ENGINE that implements RSA by performing private key
2769 exponentiations with the GMP library. The conversions to and from
2770 GMP's mpz_t format aren't optimised nor are any montgomery forms
2771 cached, and on x86 it appears OpenSSL's own performance has caught up.
2772 However there are likely to be other architectures where GMP could
2773 provide a boost. This ENGINE is not built in by default, but it can be
2774 specified at Configure time and should be accompanied by the necessary
2775 linker additions, eg;
2776 ./config -DOPENSSL_USE_GMP -lgmp
2779 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2780 testing availability of engines with "-t" - the old behaviour is
2781 produced by increasing the feature's verbosity with "-tt".
2784 *) ECDSA routines: under certain error conditions uninitialized BN objects
2785 could be freed. Solution: make sure initialization is performed early
2786 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2790 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2791 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2792 software implementations. For DSA and DH, parameter generation can
2793 also be overriden by providing the appropriate method callbacks.
2796 *) Change the "progress" mechanism used in key-generation and
2797 primality testing to functions that take a new BN_GENCB pointer in
2798 place of callback/argument pairs. The new API functions have "_ex"
2799 postfixes and the older functions are reimplemented as wrappers for
2800 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2801 declarations of the old functions to help (graceful) attempts to
2802 migrate to the new functions. Also, the new key-generation API
2803 functions operate on a caller-supplied key-structure and return
2804 success/failure rather than returning a key or NULL - this is to
2805 help make "keygen" another member function of RSA_METHOD etc.
2807 Example for using the new callback interface:
2809 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2813 BN_GENCB_set(&my_cb, my_callback, my_arg);
2815 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2816 /* For the meaning of a, b in calls to my_callback(), see the
2817 * documentation of the function that calls the callback.
2818 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2819 * my_callback should return 1 if it wants BN_is_prime_ex()
2820 * to continue, or 0 to stop.
2825 *) Change the ZLIB compression method to be stateful, and make it
2826 available to TLS with the number defined in
2827 draft-ietf-tls-compression-04.txt.
2830 *) Add the ASN.1 structures and functions for CertificatePair, which
2831 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2833 CertificatePair ::= SEQUENCE {
2834 forward [0] Certificate OPTIONAL,
2835 reverse [1] Certificate OPTIONAL,
2836 -- at least one of the pair shall be present -- }
2838 Also implement the PEM functions to read and write certificate
2839 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2841 This needed to be defined, mostly for the sake of the LDAP
2842 attribute crossCertificatePair, but may prove useful elsewhere as
2846 *) Make it possible to inhibit symlinking of shared libraries in
2847 Makefile.shared, for Cygwin's sake.
2850 *) Extend the BIGNUM API by creating a function
2851 void BN_set_negative(BIGNUM *a, int neg);
2852 and a macro that behave like
2853 int BN_is_negative(const BIGNUM *a);
2855 to avoid the need to access 'a->neg' directly in applications.
2858 *) Implement fast modular reduction for pseudo-Mersenne primes
2859 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2860 EC_GROUP_new_curve_GFp() will now automatically use this
2862 [Nils Larsch <nla@trustcenter.de>]
2864 *) Add new lock type (CRYPTO_LOCK_BN).
2867 *) Change the ENGINE framework to automatically load engines
2868 dynamically from specific directories unless they could be
2869 found to already be built in or loaded. Move all the
2870 current engines except for the cryptodev one to a new
2872 The engines in engines/ are built as shared libraries if
2873 the "shared" options was given to ./Configure or ./config.
2874 Otherwise, they are inserted in libcrypto.a.
2875 /usr/local/ssl/engines is the default directory for dynamic
2876 engines, but that can be overriden at configure time through
2877 the usual use of --prefix and/or --openssldir, and at run
2878 time with the environment variable OPENSSL_ENGINES.
2879 [Geoff Thorpe and Richard Levitte]
2881 *) Add Makefile.shared, a helper makefile to build shared
2882 libraries. Addapt Makefile.org.
2885 *) Add version info to Win32 DLLs.
2886 [Peter 'Luna' Runestig" <peter@runestig.com>]
2888 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2889 can be added using this API to created arbitrary PKCS#12
2890 files while avoiding the low level API.
2892 New options to PKCS12_create(), key or cert can be NULL and
2893 will then be omitted from the output file. The encryption
2894 algorithm NIDs can be set to -1 for no encryption, the mac
2895 iteration count can be set to 0 to omit the mac.
2897 Enhance pkcs12 utility by making the -nokeys and -nocerts
2898 options work when creating a PKCS#12 file. New option -nomac
2899 to omit the mac, NONE can be set for an encryption algorithm.
2900 New code is modified to use the enhanced PKCS12_create()
2901 instead of the low level API.
2904 *) Extend ASN1 encoder to support indefinite length constructed
2905 encoding. This can output sequences tags and octet strings in
2906 this form. Modify pk7_asn1.c to support indefinite length
2907 encoding. This is experimental and needs additional code to
2908 be useful, such as an ASN1 bio and some enhanced streaming
2911 Extend template encode functionality so that tagging is passed
2912 down to the template encoder.
2915 *) Let 'openssl req' fail if an argument to '-newkey' is not
2916 recognized instead of using RSA as a default.
2919 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2920 As these are not official, they are not included in "ALL";
2921 the "ECCdraft" ciphersuite group alias can be used to select them.
2922 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2924 *) Add ECDH engine support.
2925 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2927 *) Add ECDH in new directory crypto/ecdh/.
2928 [Douglas Stebila (Sun Microsystems Laboratories)]
2930 *) Let BN_rand_range() abort with an error after 100 iterations
2931 without success (which indicates a broken PRNG).
2934 *) Change BN_mod_sqrt() so that it verifies that the input value
2935 is really the square of the return value. (Previously,
2936 BN_mod_sqrt would show GIGO behaviour.)
2939 *) Add named elliptic curves over binary fields from X9.62, SECG,
2940 and WAP/WTLS; add OIDs that were still missing.
2942 [Sheueling Chang Shantz and Douglas Stebila
2943 (Sun Microsystems Laboratories)]
2945 *) Extend the EC library for elliptic curves over binary fields
2946 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2949 EC_GF2m_simple_method
2953 EC_GROUP_new_curve_GF2m
2954 EC_GROUP_set_curve_GF2m
2955 EC_GROUP_get_curve_GF2m
2956 EC_POINT_set_affine_coordinates_GF2m
2957 EC_POINT_get_affine_coordinates_GF2m
2958 EC_POINT_set_compressed_coordinates_GF2m
2960 Point compression for binary fields is disabled by default for
2961 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2964 As binary polynomials are represented as BIGNUMs, various members
2965 of the EC_GROUP and EC_POINT data structures can be shared
2966 between the implementations for prime fields and binary fields;
2967 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2968 are essentially identical to their ..._GFp counterparts.
2969 (For simplicity, the '..._GFp' prefix has been dropped from
2970 various internal method names.)
2972 An internal 'field_div' method (similar to 'field_mul' and
2973 'field_sqr') has been added; this is used only for binary fields.
2975 [Sheueling Chang Shantz and Douglas Stebila
2976 (Sun Microsystems Laboratories)]
2978 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2979 through methods ('mul', 'precompute_mult').
2981 The generic implementations (now internally called 'ec_wNAF_mul'
2982 and 'ec_wNAF_precomputed_mult') remain the default if these
2983 methods are undefined.
2985 [Sheueling Chang Shantz and Douglas Stebila
2986 (Sun Microsystems Laboratories)]
2988 *) New function EC_GROUP_get_degree, which is defined through
2989 EC_METHOD. For curves over prime fields, this returns the bit
2990 length of the modulus.
2992 [Sheueling Chang Shantz and Douglas Stebila
2993 (Sun Microsystems Laboratories)]
2995 *) New functions EC_GROUP_dup, EC_POINT_dup.
2996 (These simply call ..._new and ..._copy).
2998 [Sheueling Chang Shantz and Douglas Stebila
2999 (Sun Microsystems Laboratories)]
3001 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3002 Polynomials are represented as BIGNUMs (where the sign bit is not
3003 used) in the following functions [macros]:
3006 BN_GF2m_sub [= BN_GF2m_add]
3007 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3008 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3009 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3011 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3012 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3013 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3014 BN_GF2m_cmp [= BN_ucmp]
3016 (Note that only the 'mod' functions are actually for fields GF(2^m).
3017 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3019 For some functions, an the irreducible polynomial defining a
3020 field can be given as an 'unsigned int[]' with strictly
3021 decreasing elements giving the indices of those bits that are set;
3022 i.e., p[] represents the polynomial
3023 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3025 p[0] > p[1] > ... > p[k] = 0.
3026 This applies to the following functions:
3031 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3032 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3034 BN_GF2m_mod_sqrt_arr
3035 BN_GF2m_mod_solve_quad_arr
3039 Conversion can be performed by the following functions:
3044 bntest.c has additional tests for binary polynomial arithmetic.
3046 Two implementations for BN_GF2m_mod_div() are available.
3047 The default algorithm simply uses BN_GF2m_mod_inv() and
3048 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3049 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3050 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3052 [Sheueling Chang Shantz and Douglas Stebila
3053 (Sun Microsystems Laboratories)]
3055 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3056 functionality is disabled at compile-time.
3057 [Douglas Stebila <douglas.stebila@sun.com>]
3059 *) Change default behaviour of 'openssl asn1parse' so that more
3060 information is visible when viewing, e.g., a certificate:
3062 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3063 mode the content of non-printable OCTET STRINGs is output in a
3064 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3065 avoid the appearance of a printable string.
3066 [Nils Larsch <nla@trustcenter.de>]
3068 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3070 EC_GROUP_set_asn1_flag()
3071 EC_GROUP_get_asn1_flag()
3072 EC_GROUP_set_point_conversion_form()
3073 EC_GROUP_get_point_conversion_form()
3074 These control ASN1 encoding details:
3075 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3076 has been set to OPENSSL_EC_NAMED_CURVE.
3077 - Points are encoded in uncompressed form by default; options for
3078 asn1_for are as for point2oct, namely
3079 POINT_CONVERSION_COMPRESSED
3080 POINT_CONVERSION_UNCOMPRESSED
3081 POINT_CONVERSION_HYBRID
3083 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3086 EC_GROUP_get0_seed()
3087 EC_GROUP_get_seed_len()
3088 This is used only for ASN1 purposes (so far).
3089 [Nils Larsch <nla@trustcenter.de>]
3091 *) Add 'field_type' member to EC_METHOD, which holds the NID
3092 of the appropriate field type OID. The new function
3093 EC_METHOD_get_field_type() returns this value.
3094 [Nils Larsch <nla@trustcenter.de>]
3099 EC_POINT_point2hex()
3100 EC_POINT_hex2point()
3101 providing useful interfaces to EC_POINT_point2oct() and
3102 EC_POINT_oct2point().
3103 [Nils Larsch <nla@trustcenter.de>]
3105 *) Change internals of the EC library so that the functions
3106 EC_GROUP_set_generator()
3107 EC_GROUP_get_generator()
3108 EC_GROUP_get_order()
3109 EC_GROUP_get_cofactor()
3110 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3111 to methods, which would lead to unnecessary code duplication when
3112 adding different types of curves.
3113 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3115 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3116 arithmetic, and such that modified wNAFs are generated
3117 (which avoid length expansion in many cases).
3120 *) Add a function EC_GROUP_check_discriminant() (defined via
3121 EC_METHOD) that verifies that the curve discriminant is non-zero.
3123 Add a function EC_GROUP_check() that makes some sanity tests
3124 on a EC_GROUP, its generator and order. This includes
3125 EC_GROUP_check_discriminant().
3126 [Nils Larsch <nla@trustcenter.de>]
3128 *) Add ECDSA in new directory crypto/ecdsa/.
3130 Add applications 'openssl ecparam' and 'openssl ecdsa'
3131 (these are based on 'openssl dsaparam' and 'openssl dsa').
3133 ECDSA support is also included in various other files across the
3134 library. Most notably,
3135 - 'openssl req' now has a '-newkey ecdsa:file' option;
3136 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3137 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3138 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3139 them suitable for ECDSA where domain parameters must be
3140 extracted before the specific public key;
3141 - ECDSA engine support has been added.
3142 [Nils Larsch <nla@trustcenter.de>]
3144 *) Include some named elliptic curves, and add OIDs from X9.62,
3145 SECG, and WAP/WTLS. Each curve can be obtained from the new
3147 EC_GROUP_new_by_curve_name(),
3148 and the list of available named curves can be obtained with
3149 EC_get_builtin_curves().
3150 Also add a 'curve_name' member to EC_GROUP objects, which can be
3152 EC_GROUP_set_curve_name()
3153 EC_GROUP_get_curve_name()
3154 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3156 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3157 was actually never needed) and in BN_mul(). The removal in BN_mul()
3158 required a small change in bn_mul_part_recursive() and the addition
3159 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3160 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3161 bn_sub_words() and bn_add_words() except they take arrays with
3165 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3167 *) Cleanse PEM buffers before freeing them since they may contain
3169 [Benjamin Bennett <ben@psc.edu>]
3171 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3172 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3173 authentication-only ciphersuites.
3176 *) Since AES128 and AES256 share a single mask bit in the logic of
3177 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3178 kludge to work properly if AES128 is available and AES256 isn't.
3181 *) Expand security boundary to match 1.1.1 module.
3184 *) Remove redundant features: hash file source, editing of test vectors
3185 modify fipsld to use external fips_premain.c signature.
3188 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3189 run algorithm test programs.
3192 *) Make algorithm test programs more tolerant of whitespace.
3195 *) Have SSL/TLS server implementation tolerate "mismatched" record
3196 protocol version while receiving ClientHello even if the
3197 ClientHello is fragmented. (The server can't insist on the
3198 particular protocol version it has chosen before the ServerHello
3199 message has informed the client about his choice.)
3202 *) Load error codes if they are not already present instead of using a
3203 static variable. This allows them to be cleanly unloaded and reloaded.
3206 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3208 *) Introduce limits to prevent malicious keys being able to
3209 cause a denial of service. (CVE-2006-2940)
3210 [Steve Henson, Bodo Moeller]
3212 *) Fix ASN.1 parsing of certain invalid structures that can result
3213 in a denial of service. (CVE-2006-2937) [Steve Henson]
3215 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3216 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3218 *) Fix SSL client code which could crash if connecting to a
3219 malicious SSLv2 server. (CVE-2006-4343)
3220 [Tavis Ormandy and Will Drewry, Google Security Team]
3222 *) Change ciphersuite string processing so that an explicit
3223 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3224 will no longer include "AES128-SHA"), and any other similar
3225 ciphersuite (same bitmap) from *other* protocol versions (so that
3226 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3227 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3228 changes from 0.9.8b and 0.9.8d.
3231 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3233 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3234 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3236 *) Change the Unix randomness entropy gathering to use poll() when
3237 possible instead of select(), since the latter has some
3238 undesirable limitations.
3239 [Darryl Miles via Richard Levitte and Bodo Moeller]
3241 *) Disable rogue ciphersuites:
3243 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3244 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3245 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3247 The latter two were purportedly from
3248 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3251 Also deactive the remaining ciphersuites from
3252 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3253 unofficial, and the ID has long expired.
3256 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3257 dual-core machines) and other potential thread-safety issues.
3260 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3262 *) Adapt fipsld and the build system to link against the validated FIPS
3263 module in FIPS mode.
3266 *) Fixes for VC++ 2005 build under Windows.
3269 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3270 from a Windows bash shell such as MSYS. It is autodetected from the
3271 "config" script when run from a VC++ environment. Modify standard VC++
3272 build to use fipscanister.o from the GNU make build.
3275 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3277 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3278 The value now differs depending on if you build for FIPS or not.
3279 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3280 safely run with a non-FIPSed libcrypto, as it may crash because of
3281 the difference induced by this change.
3284 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3286 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3287 (part of SSL_OP_ALL). This option used to disable the
3288 countermeasure against man-in-the-middle protocol-version
3289 rollback in the SSL 2.0 server implementation, which is a bad
3290 idea. (CVE-2005-2969)
3292 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3293 for Information Security, National Institute of Advanced Industrial
3294 Science and Technology [AIST], Japan)]
3296 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3297 mainly for FIPS compliance and not fully integrated at this stage.
3300 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3301 the exponentiation using a fixed-length exponent. (Otherwise,
3302 the information leaked through timing could expose the secret key
3303 after many signatures; cf. Bleichenbacher's attack on DSA with
3307 *) Make a new fixed-window mod_exp implementation the default for
3308 RSA, DSA, and DH private-key operations so that the sequence of
3309 squares and multiplies and the memory access pattern are
3310 independent of the particular secret key. This will mitigate
3311 cache-timing and potential related attacks.
3313 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3314 and this is automatically used by BN_mod_exp_mont() if the new flag
3315 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3316 will use this BN flag for private exponents unless the flag
3317 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3318 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3320 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3322 *) Change the client implementation for SSLv23_method() and
3323 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3324 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3325 (Previously, the SSL 2.0 backwards compatible Client Hello
3326 message format would be used even with SSL_OP_NO_SSLv2.)
3329 *) Add support for smime-type MIME parameter in S/MIME messages which some
3333 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3334 a threadsafe manner. Modify rsa code to use new function and add calls
3335 to dsa and dh code (which had race conditions before).
3338 *) Include the fixed error library code in the C error file definitions
3339 instead of fixing them up at runtime. This keeps the error code
3340 structures constant.
3343 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3345 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3348 *) Fixes for newer kerberos headers. NB: the casts are needed because
3349 the 'length' field is signed on one version and unsigned on another
3350 with no (?) obvious way to tell the difference, without these VC++
3351 complains. Also the "definition" of FAR (blank) is no longer included
3352 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3353 some needed definitions.
3356 *) Undo Cygwin change.
3359 *) Added support for proxy certificates according to RFC 3820.
3360 Because they may be a security thread to unaware applications,
3361 they must be explicitely allowed in run-time. See
3362 docs/HOWTO/proxy_certificates.txt for further information.
3365 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3367 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3368 server and client random values. Previously
3369 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3370 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3372 This change has negligible security impact because:
3374 1. Server and client random values still have 24 bytes of pseudo random
3377 2. Server and client random values are sent in the clear in the initial
3380 3. The master secret is derived using the premaster secret (48 bytes in
3381 size for static RSA ciphersuites) as well as client server and random
3384 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3387 [Stephen Henson, reported by UK NISCC]
3389 *) Use Windows randomness collection on Cygwin.
3392 *) Fix hang in EGD/PRNGD query when communication socket is closed
3393 prematurely by EGD/PRNGD.
3394 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3396 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3399 *) Back-port of selected performance improvements from development
3400 branch, as well as improved support for PowerPC platforms.
3403 *) Add lots of checks for memory allocation failure, error codes to indicate
3404 failure and freeing up memory if a failure occurs.
3405 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3407 *) Add new -passin argument to dgst.
3410 *) Perform some character comparisons of different types in X509_NAME_cmp:
3411 this is needed for some certificates that reencode DNs into UTF8Strings
3412 (in violation of RFC3280) and can't or wont issue name rollover
3416 *) Make an explicit check during certificate validation to see that
3417 the CA setting in each certificate on the chain is correct. As a
3418 side effect always do the following basic checks on extensions,
3419 not just when there's an associated purpose to the check:
3421 - if there is an unhandled critical extension (unless the user
3422 has chosen to ignore this fault)
3423 - if the path length has been exceeded (if one is set at all)
3424 - that certain extensions fit the associated purpose (if one has
3428 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3430 *) Avoid a race condition when CRLs are checked in a multi threaded
3431 environment. This would happen due to the reordering of the revoked
3432 entries during signature checking and serial number lookup. Now the
3433 encoding is cached and the serial number sort performed under a lock.
3434 Add new STACK function sk_is_sorted().
3437 *) Add Delta CRL to the extension code.
3440 *) Various fixes to s3_pkt.c so alerts are sent properly.
3441 [David Holmes <d.holmes@f5.com>]
3443 *) Reduce the chances of duplicate issuer name and serial numbers (in
3444 violation of RFC3280) using the OpenSSL certificate creation utilities.
3445 This is done by creating a random 64 bit value for the initial serial
3446 number when a serial number file is created or when a self signed
3447 certificate is created using 'openssl req -x509'. The initial serial
3448 number file is created using 'openssl x509 -next_serial' in CA.pl
3449 rather than being initialized to 1.
3452 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3454 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3455 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3456 [Joe Orton, Steve Henson]
3458 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3460 [Joe Orton, Steve Henson]
3462 *) Make it possible to have multiple active certificates with the same
3463 subject in the CA index file. This is done only if the keyword
3464 'unique_subject' is set to 'no' in the main CA section (default
3465 if 'CA_default') of the configuration file. The value is saved
3466 with the database itself in a separate index attribute file,
3467 named like the index file with '.attr' appended to the name.
3470 *) X509 verify fixes. Disable broken certificate workarounds when
3471 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3472 keyUsage extension present. Don't accept CRLs with unhandled critical
3473 extensions: since verify currently doesn't process CRL extensions this
3474 rejects a CRL with *any* critical extensions. Add new verify error codes
3478 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3479 A clarification of RFC2560 will require the use of OCTET STRINGs and
3480 some implementations cannot handle the current raw format. Since OpenSSL
3481 copies and compares OCSP nonces as opaque blobs without any attempt at
3482 parsing them this should not create any compatibility issues.
3485 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3486 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3487 this HMAC (and other) operations are several times slower than OpenSSL
3491 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3492 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3494 *) Use the correct content when signing type "other".
3497 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3499 *) Fix various bugs revealed by running the NISCC test suite:
3501 Stop out of bounds reads in the ASN1 code when presented with
3502 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3504 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3506 If verify callback ignores invalid public key errors don't try to check
3507 certificate signature with the NULL public key.
3511 *) New -ignore_err option in ocsp application to stop the server
3512 exiting on the first error in a request.
3515 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3516 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3520 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3521 extra data after the compression methods not only for TLS 1.0
3522 but also for SSL 3.0 (as required by the specification).
3523 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3525 *) Change X509_certificate_type() to mark the key as exported/exportable
3526 when it's 512 *bits* long, not 512 bytes.
3529 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3530 blocks during encryption.
3533 *) Various fixes to base64 BIO and non blocking I/O. On write
3534 flushes were not handled properly if the BIO retried. On read
3535 data was not being buffered properly and had various logic bugs.
3536 This also affects blocking I/O when the data being decoded is a
3540 *) Various S/MIME bugfixes and compatibility changes:
3541 output correct application/pkcs7 MIME type if
3542 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3543 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3544 of files as .eml work). Correctly handle very long lines in MIME
3548 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3550 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3551 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3552 a protocol version number mismatch like a decryption error
3553 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3556 *) Turn on RSA blinding by default in the default implementation
3557 to avoid a timing attack. Applications that don't want it can call
3558 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3559 They would be ill-advised to do so in most cases.
3560 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3562 *) Change RSA blinding code so that it works when the PRNG is not
3563 seeded (in this case, the secret RSA exponent is abused as
3564 an unpredictable seed -- if it is not unpredictable, there
3565 is no point in blinding anyway). Make RSA blinding thread-safe
3566 by remembering the creator's thread ID in rsa->blinding and
3567 having all other threads use local one-time blinding factors
3568 (this requires more computation than sharing rsa->blinding, but
3569 avoids excessive locking; and if an RSA object is not shared
3570 between threads, blinding will still be very fast).
3573 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3574 ENGINE as defaults for all supported algorithms irrespective of
3575 the 'flags' parameter. 'flags' is now honoured, so applications
3576 should make sure they are passing it correctly.
3579 *) Target "mingw" now allows native Windows code to be generated in
3580 the Cygwin environment as well as with the MinGW compiler.
3583 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3585 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3586 via timing by performing a MAC computation even if incorrrect
3587 block cipher padding has been found. This is a countermeasure
3588 against active attacks where the attacker has to distinguish
3589 between bad padding and a MAC verification error. (CVE-2003-0078)
3591 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3592 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3593 Martin Vuagnoux (EPFL, Ilion)]
3595 *) Make the no-err option work as intended. The intention with no-err
3596 is not to have the whole error stack handling routines removed from
3597 libcrypto, it's only intended to remove all the function name and
3598 reason texts, thereby removing some of the footprint that may not
3599 be interesting if those errors aren't displayed anyway.
3601 NOTE: it's still possible for any application or module to have it's
3602 own set of error texts inserted. The routines are there, just not
3603 used by default when no-err is given.
3606 *) Add support for FreeBSD on IA64.
3607 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3609 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3610 Kerberos function mit_des_cbc_cksum(). Before this change,
3611 the value returned by DES_cbc_cksum() was like the one from
3612 mit_des_cbc_cksum(), except the bytes were swapped.
3613 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3615 *) Allow an application to disable the automatic SSL chain building.
3616 Before this a rather primitive chain build was always performed in
3617 ssl3_output_cert_chain(): an application had no way to send the
3618 correct chain if the automatic operation produced an incorrect result.
3620 Now the chain builder is disabled if either:
3622 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3624 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3626 The reasoning behind this is that an application would not want the
3627 auto chain building to take place if extra chain certificates are
3628 present and it might also want a means of sending no additional
3629 certificates (for example the chain has two certificates and the
3633 *) Add the possibility to build without the ENGINE framework.
3634 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3636 *) Under Win32 gmtime() can return NULL: check return value in
3637 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3640 *) DSA routines: under certain error conditions uninitialized BN objects
3641 could be freed. Solution: make sure initialization is performed early
3642 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3643 Nils Larsch <nla@trustcenter.de> via PR#459)
3646 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3647 checked on reconnect on the client side, therefore session resumption
3648 could still fail with a "ssl session id is different" error. This
3649 behaviour is masked when SSL_OP_ALL is used due to
3650 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3651 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3652 followup to PR #377.
3655 *) IA-32 assembler support enhancements: unified ELF targets, support
3656 for SCO/Caldera platforms, fix for Cygwin shared build.
3659 *) Add support for FreeBSD on sparc64. As a consequence, support for
3660 FreeBSD on non-x86 processors is separate from x86 processors on
3661 the config script, much like the NetBSD support.
3662 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3664 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3666 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3669 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3670 code (06) was taken as the first octet of the session ID and the last
3671 octet was ignored consequently. As a result SSLv2 client side session
3672 caching could not have worked due to the session ID mismatch between
3674 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3678 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3679 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3683 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3684 seems that in spite of existing for more than a year, many application
3685 author have done nothing to provide the necessary callbacks, which
3686 means that this particular engine will not work properly anywhere.
3687 This is a very unfortunate situation which forces us, in the name
3688 of usability, to give the hw_ncipher.c a static lock, which is part
3690 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3691 appear in 0.9.8 or later. We EXPECT application authors to have
3692 dealt properly with this when 0.9.8 is released (unless we actually
3693 make such changes in the libcrypto locking code that changes will
3694 have to be made anyway).
3697 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3698 octets have been read, EOF or an error occurs. Without this change
3699 some truncated ASN1 structures will not produce an error.
3702 *) Disable Heimdal support, since it hasn't been fully implemented.
3703 Still give the possibility to force the use of Heimdal, but with
3704 warnings and a request that patches get sent to openssl-dev.
3707 *) Add the VC-CE target, introduce the WINCE sysname, and add
3708 INSTALL.WCE and appropriate conditionals to make it build.
3709 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3711 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3712 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3713 edit numbers of the version.
3714 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3716 *) Introduce safe string copy and catenation functions
3717 (BUF_strlcpy() and BUF_strlcat()).
3718 [Ben Laurie (CHATS) and Richard Levitte]
3720 *) Avoid using fixed-size buffers for one-line DNs.
3721 [Ben Laurie (CHATS)]
3723 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3724 resizing buffers containing secrets, and use where appropriate.
3725 [Ben Laurie (CHATS)]
3727 *) Avoid using fixed size buffers for configuration file location.
3728 [Ben Laurie (CHATS)]
3730 *) Avoid filename truncation for various CA files.
3731 [Ben Laurie (CHATS)]
3733 *) Use sizeof in preference to magic numbers.
3734 [Ben Laurie (CHATS)]
3736 *) Avoid filename truncation in cert requests.
3737 [Ben Laurie (CHATS)]
3739 *) Add assertions to check for (supposedly impossible) buffer
3741 [Ben Laurie (CHATS)]
3743 *) Don't cache truncated DNS entries in the local cache (this could
3744 potentially lead to a spoofing attack).
3745 [Ben Laurie (CHATS)]
3747 *) Fix various buffers to be large enough for hex/decimal
3748 representations in a platform independent manner.
3749 [Ben Laurie (CHATS)]
3751 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3752 resizing buffers containing secrets, and use where appropriate.
3753 [Ben Laurie (CHATS)]
3755 *) Add BIO_indent() to avoid much slightly worrying code to do
3757 [Ben Laurie (CHATS)]
3759 *) Convert sprintf()/BIO_puts() to BIO_printf().
3760 [Ben Laurie (CHATS)]
3762 *) buffer_gets() could terminate with the buffer only half
3764 [Ben Laurie (CHATS)]
3766 *) Add assertions to prevent user-supplied crypto functions from
3767 overflowing internal buffers by having large block sizes, etc.
3768 [Ben Laurie (CHATS)]
3770 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3772 [Ben Laurie (CHATS)]
3774 *) Eliminate unused copy of key in RC4.
3775 [Ben Laurie (CHATS)]
3777 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3778 [Ben Laurie (CHATS)]
3780 *) Fix off-by-one error in EGD path.
3781 [Ben Laurie (CHATS)]
3783 *) If RANDFILE path is too long, ignore instead of truncating.
3784 [Ben Laurie (CHATS)]
3786 *) Eliminate unused and incorrectly sized X.509 structure
3788 [Ben Laurie (CHATS)]
3790 *) Eliminate unused and dangerous function knumber().
3791 [Ben Laurie (CHATS)]
3793 *) Eliminate unused and dangerous structure, KSSL_ERR.
3794 [Ben Laurie (CHATS)]
3796 *) Protect against overlong session ID context length in an encoded
3797 session object. Since these are local, this does not appear to be
3799 [Ben Laurie (CHATS)]
3801 *) Change from security patch (see 0.9.6e below) that did not affect
3802 the 0.9.6 release series:
3804 Remote buffer overflow in SSL3 protocol - an attacker could
3805 supply an oversized master key in Kerberos-enabled versions.
3807 [Ben Laurie (CHATS)]
3809 *) Change the SSL kerb5 codes to match RFC 2712.
3812 *) Make -nameopt work fully for req and add -reqopt switch.
3813 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3815 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3816 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3818 *) Make sure tests can be performed even if the corresponding algorithms
3819 have been removed entirely. This was also the last step to make
3820 OpenSSL compilable with DJGPP under all reasonable conditions.
3821 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3823 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3824 to allow version independent disabling of normally unselected ciphers,
3825 which may be activated as a side-effect of selecting a single cipher.
3827 (E.g., cipher list string "RSA" enables ciphersuites that are left
3828 out of "ALL" because they do not provide symmetric encryption.
3829 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3830 [Lutz Jaenicke, Bodo Moeller]
3832 *) Add appropriate support for separate platform-dependent build
3833 directories. The recommended way to make a platform-dependent
3834 build directory is the following (tested on Linux), maybe with
3837 # Place yourself outside of the OpenSSL source tree. In
3838 # this example, the environment variable OPENSSL_SOURCE
3839 # is assumed to contain the absolute OpenSSL source directory.
3840 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3841 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3842 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3843 mkdir -p `dirname $F`
3844 ln -s $OPENSSL_SOURCE/$F $F
3847 To be absolutely sure not to disturb the source tree, a "make clean"
3848 is a good thing. If it isn't successfull, don't worry about it,
3849 it probably means the source directory is very clean.
3852 *) Make sure any ENGINE control commands make local copies of string
3853 pointers passed to them whenever necessary. Otherwise it is possible
3854 the caller may have overwritten (or deallocated) the original string
3855 data when a later ENGINE operation tries to use the stored values.
3856 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3858 *) Improve diagnostics in file reading and command-line digests.
3859 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3861 *) Add AES modes CFB and OFB to the object database. Correct an
3862 error in AES-CFB decryption.
3865 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3866 allows existing EVP_CIPHER_CTX structures to be reused after
3867 calling EVP_*Final(). This behaviour is used by encryption
3868 BIOs and some applications. This has the side effect that
3869 applications must explicitly clean up cipher contexts with
3870 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3873 *) Check the values of dna and dnb in bn_mul_recursive before calling
3874 bn_mul_comba (a non zero value means the a or b arrays do not contain
3875 n2 elements) and fallback to bn_mul_normal if either is not zero.
3878 *) Fix escaping of non-ASCII characters when using the -subj option
3879 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3882 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3883 form for "surname", serialNumber has no short form.
3884 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3885 therefore remove "mail" short name for "internet 7".
3886 The OID for unique identifiers in X509 certificates is
3887 x500UniqueIdentifier, not uniqueIdentifier.
3888 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3891 *) Add an "init" command to the ENGINE config module and auto initialize
3892 ENGINEs. Without any "init" command the ENGINE will be initialized
3893 after all ctrl commands have been executed on it. If init=1 the
3894 ENGINE is initailized at that point (ctrls before that point are run
3895 on the uninitialized ENGINE and after on the initialized one). If
3896 init=0 then the ENGINE will not be iniatialized at all.
3899 *) Fix the 'app_verify_callback' interface so that the user-defined
3900 argument is actually passed to the callback: In the
3901 SSL_CTX_set_cert_verify_callback() prototype, the callback
3902 declaration has been changed from
3905 int (*cb)(X509_STORE_CTX *,void *);
3906 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3907 i=s->ctx->app_verify_callback(&ctx)
3908 has been changed into
3909 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3911 To update applications using SSL_CTX_set_cert_verify_callback(),
3912 a dummy argument can be added to their callback functions.
3913 [D. K. Smetters <smetters@parc.xerox.com>]
3915 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3916 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3918 *) Add and OPENSSL_LOAD_CONF define which will cause
3919 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3920 This allows older applications to transparently support certain
3921 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3922 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3923 load the config file and OPENSSL_add_all_algorithms_conf() which will
3924 always load it have also been added.
3927 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3928 Adjust NIDs and EVP layer.
3929 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3931 *) Config modules support in openssl utility.
3933 Most commands now load modules from the config file,
3934 though in a few (such as version) this isn't done
3935 because it couldn't be used for anything.
3937 In the case of ca and req the config file used is
3938 the same as the utility itself: that is the -config
3939 command line option can be used to specify an
3943 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3944 use "openssl_conf" if filename is NULL use default openssl config file.
3947 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3948 config section name. Add a new flag to tolerate a missing config file
3949 and move code to CONF_modules_load_file().
3952 *) Support for crypto accelerator cards from Accelerated Encryption
3953 Processing, www.aep.ie. (Use engine 'aep')
3954 The support was copied from 0.9.6c [engine] and adapted/corrected
3955 to work with the new engine framework.
3956 [AEP Inc. and Richard Levitte]
3958 *) Support for SureWare crypto accelerator cards from Baltimore
3959 Technologies. (Use engine 'sureware')
3960 The support was copied from 0.9.6c [engine] and adapted
3961 to work with the new engine framework.
3964 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3965 make the newer ENGINE framework commands for the CHIL engine work.
3966 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3968 *) Make it possible to produce shared libraries on ReliantUNIX.
3969 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3971 *) Add the configuration target debug-linux-ppro.
3972 Make 'openssl rsa' use the general key loading routines
3973 implemented in apps.c, and make those routines able to
3974 handle the key format FORMAT_NETSCAPE and the variant
3976 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3978 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3979 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3981 *) Add -keyform to rsautl, and document -engine.
3982 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3984 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3985 BIO_R_NO_SUCH_FILE error code rather than the generic
3986 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3989 *) Add new functions
3991 ERR_peek_last_error_line
3992 ERR_peek_last_error_line_data.
3993 These are similar to
3996 ERR_peek_error_line_data,
3997 but report on the latest error recorded rather than the first one
3998 still in the error queue.
3999 [Ben Laurie, Bodo Moeller]
4001 *) default_algorithms option in ENGINE config module. This allows things
4003 default_algorithms = ALL
4004 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4007 *) Prelminary ENGINE config module.
4010 *) New experimental application configuration code.
4013 *) Change the AES code to follow the same name structure as all other
4014 symmetric ciphers, and behave the same way. Move everything to
4015 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4016 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4018 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4019 [Ben Laurie and Theo de Raadt]
4021 *) Add option to output public keys in req command.
4022 [Massimiliano Pala madwolf@openca.org]
4024 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4025 (up to about 10% better than before for P-192 and P-224).
4028 *) New functions/macros
4030 SSL_CTX_set_msg_callback(ctx, cb)
4031 SSL_CTX_set_msg_callback_arg(ctx, arg)
4032 SSL_set_msg_callback(ssl, cb)
4033 SSL_set_msg_callback_arg(ssl, arg)
4035 to request calling a callback function
4037 void cb(int write_p, int version, int content_type,
4038 const void *buf, size_t len, SSL *ssl, void *arg)
4040 whenever a protocol message has been completely received
4041 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4042 protocol version according to which the SSL library interprets
4043 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4044 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4045 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4046 specification (change_cipher_spec(20), alert(21), handshake(22)).
4047 'buf' and 'len' point to the actual message, 'ssl' to the
4048 SSL object, and 'arg' is the application-defined value set by
4049 SSL[_CTX]_set_msg_callback_arg().
4051 'openssl s_client' and 'openssl s_server' have new '-msg' options
4052 to enable a callback that displays all protocol messages.
4055 *) Change the shared library support so shared libraries are built as
4056 soon as the corresponding static library is finished, and thereby get
4057 openssl and the test programs linked against the shared library.
4058 This still only happens when the keyword "shard" has been given to
4059 the configuration scripts.
4061 NOTE: shared library support is still an experimental thing, and
4062 backward binary compatibility is still not guaranteed.
4063 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4065 *) Add support for Subject Information Access extension.
4066 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4068 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4069 additional bytes when new memory had to be allocated, not just
4070 when reusing an existing buffer.
4073 *) New command line and configuration option 'utf8' for the req command.
4074 This allows field values to be specified as UTF8 strings.
4077 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4078 runs for the former and machine-readable output for the latter.
4081 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4082 of the e-mail address in the DN (i.e., it will go into a certificate
4083 extension only). The new configuration file option 'email_in_dn = no'
4084 has the same effect.
4085 [Massimiliano Pala madwolf@openca.org]
4087 *) Change all functions with names starting with des_ to be starting
4088 with DES_ instead. Add wrappers that are compatible with libdes,
4089 but are named _ossl_old_des_*. Finally, add macros that map the
4090 des_* symbols to the corresponding _ossl_old_des_* if libdes
4091 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4092 desired, the des_* symbols will be mapped to DES_*, with one
4095 Since we provide two compatibility mappings, the user needs to
4096 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4097 compatibility is desired. The default (i.e., when that macro
4098 isn't defined) is OpenSSL 0.9.6c compatibility.
4100 There are also macros that enable and disable the support of old
4101 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4102 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4103 are defined, the default will apply: to support the old des routines.
4105 In either case, one must include openssl/des.h to get the correct
4106 definitions. Do not try to just include openssl/des_old.h, that
4109 NOTE: This is a major break of an old API into a new one. Software
4110 authors are encouraged to switch to the DES_ style functions. Some
4111 time in the future, des_old.h and the libdes compatibility functions
4112 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4113 default), and then completely removed.
4116 *) Test for certificates which contain unsupported critical extensions.
4117 If such a certificate is found during a verify operation it is
4118 rejected by default: this behaviour can be overridden by either
4119 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4120 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4121 X509_supported_extension() has also been added which returns 1 if a
4122 particular extension is supported.
4125 *) Modify the behaviour of EVP cipher functions in similar way to digests
4126 to retain compatibility with existing code.
4129 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4130 compatibility with existing code. In particular the 'ctx' parameter does
4131 not have to be to be initialized before the call to EVP_DigestInit() and
4132 it is tidied up after a call to EVP_DigestFinal(). New function
4133 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4134 EVP_MD_CTX_copy() changed to not require the destination to be
4135 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4136 requires the destination to be valid.
4138 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4139 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4142 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4143 so that complete 'Handshake' protocol structures are kept in memory
4144 instead of overwriting 'msg_type' and 'length' with 'body' data.
4147 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4148 [Massimo Santin via Richard Levitte]
4150 *) Major restructuring to the underlying ENGINE code. This includes
4151 reduction of linker bloat, separation of pure "ENGINE" manipulation
4152 (initialisation, etc) from functionality dealing with implementations
4153 of specific crypto iterfaces. This change also introduces integrated
4154 support for symmetric ciphers and digest implementations - so ENGINEs
4155 can now accelerate these by providing EVP_CIPHER and EVP_MD
4156 implementations of their own. This is detailed in crypto/engine/README
4157 as it couldn't be adequately described here. However, there are a few
4158 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4159 were changed in the original introduction of ENGINE code have now
4160 reverted back - the hooking from this code to ENGINE is now a good
4161 deal more passive and at run-time, operations deal directly with
4162 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4163 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4164 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4165 they were not being used by the framework as there is no concept of a
4166 BIGNUM_METHOD and they could not be generalised to the new
4167 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4168 ENGINE_cpy() has been removed as it cannot be consistently defined in
4172 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4175 *) Change mkdef.pl to sort symbols that get the same entry number,
4176 and make sure the automatically generated functions ERR_load_*
4177 become part of libeay.num as well.
4180 *) New function SSL_renegotiate_pending(). This returns true once
4181 renegotiation has been requested (either SSL_renegotiate() call
4182 or HelloRequest/ClientHello receveived from the peer) and becomes
4183 false once a handshake has been completed.
4184 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4185 sends a HelloRequest, but does not ensure that a handshake takes
4186 place. SSL_renegotiate_pending() is useful for checking if the
4187 client has followed the request.)
4190 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4191 By default, clients may request session resumption even during
4192 renegotiation (if session ID contexts permit); with this option,
4193 session resumption is possible only in the first handshake.
4195 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4196 more bits available for options that should not be part of
4197 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4200 *) Add some demos for certificate and certificate request creation.
4203 *) Make maximum certificate chain size accepted from the peer application
4204 settable (SSL*_get/set_max_cert_list()), as proposed by
4205 "Douglas E. Engert" <deengert@anl.gov>.
4208 *) Add support for shared libraries for Unixware-7
4209 (Boyd Lynn Gerber <gerberb@zenez.com>).
4212 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4213 be done prior to destruction. Use this to unload error strings from
4214 ENGINEs that load their own error strings. NB: This adds two new API
4215 functions to "get" and "set" this destroy handler in an ENGINE.
4218 *) Alter all existing ENGINE implementations (except "openssl" and
4219 "openbsd") to dynamically instantiate their own error strings. This
4220 makes them more flexible to be built both as statically-linked ENGINEs
4221 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4222 Also, add stub code to each that makes building them as self-contained
4223 shared-libraries easier (see README.ENGINE).
4226 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4227 implementations into applications that are completely implemented in
4228 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4229 commands that can be used to configure what shared-library to load and
4230 to control aspects of the way it is handled. Also, made an update to
4231 the README.ENGINE file that brings its information up-to-date and
4232 provides some information and instructions on the "dynamic" ENGINE
4233 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4236 *) Make it possible to unload ranges of ERR strings with a new
4237 "ERR_unload_strings" function.
4240 *) Add a copy() function to EVP_MD.
4243 *) Make EVP_MD routines take a context pointer instead of just the
4244 md_data void pointer.
4247 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4248 that the digest can only process a single chunk of data
4249 (typically because it is provided by a piece of
4250 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4251 is only going to provide a single chunk of data, and hence the
4252 framework needn't accumulate the data for oneshot drivers.
4255 *) As with "ERR", make it possible to replace the underlying "ex_data"
4256 functions. This change also alters the storage and management of global
4257 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4258 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4259 index counters. The API functions that use this state have been changed
4260 to take a "class_index" rather than pointers to the class's local STACK
4261 and counter, and there is now an API function to dynamically create new
4262 classes. This centralisation allows us to (a) plug a lot of the
4263 thread-safety problems that existed, and (b) makes it possible to clean
4264 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4265 such data would previously have always leaked in application code and
4266 workarounds were in place to make the memory debugging turn a blind eye
4267 to it. Application code that doesn't use this new function will still
4268 leak as before, but their memory debugging output will announce it now
4269 rather than letting it slide.
4271 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4272 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4273 has a return value to indicate success or failure.
4276 *) Make it possible to replace the underlying "ERR" functions such that the
4277 global state (2 LHASH tables and 2 locks) is only used by the "default"
4278 implementation. This change also adds two functions to "get" and "set"
4279 the implementation prior to it being automatically set the first time
4280 any other ERR function takes place. Ie. an application can call "get",
4281 pass the return value to a module it has just loaded, and that module
4282 can call its own "set" function using that value. This means the
4283 module's "ERR" operations will use (and modify) the error state in the
4284 application and not in its own statically linked copy of OpenSSL code.
4287 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4288 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4289 the operation, and provides a more encapsulated way for external code
4290 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4291 to use these functions rather than manually incrementing the counts.
4293 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4296 *) Add EVP test program.
4299 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4302 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4303 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4304 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4305 These allow a CRL to be built without having to access X509_CRL fields
4306 directly. Modify 'ca' application to use new functions.
4309 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4310 bug workarounds. Rollback attack detection is a security feature.
4311 The problem will only arise on OpenSSL servers when TLSv1 is not
4312 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4313 Software authors not wanting to support TLSv1 will have special reasons
4314 for their choice and can explicitly enable this option.
4315 [Bodo Moeller, Lutz Jaenicke]
4317 *) Rationalise EVP so it can be extended: don't include a union of
4318 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4319 (similar to those existing for EVP_CIPHER_CTX).
4324 EVP_MD_CTX_init(&md); /* new function call */
4325 EVP_DigestInit(&md, EVP_sha1());
4326 EVP_DigestUpdate(&md, in, len);
4327 EVP_DigestFinal(&md, out, NULL);
4328 EVP_MD_CTX_cleanup(&md); /* new function call */
4332 *) Make DES key schedule conform to the usual scheme, as well as
4333 correcting its structure. This means that calls to DES functions
4334 now have to pass a pointer to a des_key_schedule instead of a
4335 plain des_key_schedule (which was actually always a pointer
4338 des_key_schedule ks;
4340 des_set_key_checked(..., &ks);
4341 des_ncbc_encrypt(..., &ks, ...);
4343 (Note that a later change renames 'des_...' into 'DES_...'.)
4346 *) Initial reduction of linker bloat: the use of some functions, such as
4347 PEM causes large amounts of unused functions to be linked in due to
4348 poor organisation. For example pem_all.c contains every PEM function
4349 which has a knock on effect of linking in large amounts of (unused)
4350 ASN1 code. Grouping together similar functions and splitting unrelated
4351 functions prevents this.
4354 *) Cleanup of EVP macros.
4357 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4358 correct _ecb suffix.
4361 *) Add initial OCSP responder support to ocsp application. The
4362 revocation information is handled using the text based index
4363 use by the ca application. The responder can either handle
4364 requests generated internally, supplied in files (for example
4365 via a CGI script) or using an internal minimal server.
4368 *) Add configuration choices to get zlib compression for TLS.
4371 *) Changes to Kerberos SSL for RFC 2712 compliance:
4372 1. Implemented real KerberosWrapper, instead of just using
4373 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4374 2. Implemented optional authenticator field of KerberosWrapper.
4376 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4377 and authenticator structs; see crypto/krb5/.
4379 Generalized Kerberos calls to support multiple Kerberos libraries.
4380 [Vern Staats <staatsvr@asc.hpc.mil>,
4381 Jeffrey Altman <jaltman@columbia.edu>
4382 via Richard Levitte]
4384 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4385 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4386 values for each of the key sizes rather than having just
4387 parameters (and 'speed' generating keys each time).
4390 *) Speed up EVP routines.
4393 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4394 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4395 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4396 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4398 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4399 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4400 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4403 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4405 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4408 *) Added the OS2-EMX target.
4409 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4411 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4412 to support NCONF routines in extension code. New function CONF_set_nconf()
4413 to allow functions which take an NCONF to also handle the old LHASH
4414 structure: this means that the old CONF compatible routines can be
4415 retained (in particular wrt extensions) without having to duplicate the
4416 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4419 *) Enhance the general user interface with mechanisms for inner control
4420 and with possibilities to have yes/no kind of prompts.
4423 *) Change all calls to low level digest routines in the library and
4424 applications to use EVP. Add missing calls to HMAC_cleanup() and
4425 don't assume HMAC_CTX can be copied using memcpy().
4426 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4428 *) Add the possibility to control engines through control names but with
4429 arbitrary arguments instead of just a string.
4430 Change the key loaders to take a UI_METHOD instead of a callback
4431 function pointer. NOTE: this breaks binary compatibility with earlier
4432 versions of OpenSSL [engine].
4433 Adapt the nCipher code for these new conditions and add a card insertion
4437 *) Enhance the general user interface with mechanisms to better support
4438 dialog box interfaces, application-defined prompts, the possibility
4439 to use defaults (for example default passwords from somewhere else)
4440 and interrupts/cancellations.
4443 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4444 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4447 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4448 tidy up some unnecessarily weird code in 'sk_new()').
4449 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4451 *) Change the key loading routines for ENGINEs to use the same kind
4452 callback (pem_password_cb) as all other routines that need this
4456 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4457 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4458 than this minimum value is recommended.
4461 *) New random seeder for OpenVMS, using the system process statistics
4462 that are easily reachable.
4465 *) Windows apparently can't transparently handle global
4466 variables defined in DLLs. Initialisations such as:
4468 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4470 wont compile. This is used by the any applications that need to
4471 declare their own ASN1 modules. This was fixed by adding the option
4472 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4473 needed for static libraries under Win32.
4476 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4477 setting of purpose and trust fields. New X509_STORE trust and
4478 purpose functions and tidy up setting in other SSL functions.
4481 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4482 structure. These are inherited by X509_STORE_CTX when it is
4483 initialised. This allows various defaults to be set in the
4484 X509_STORE structure (such as flags for CRL checking and custom
4485 purpose or trust settings) for functions which only use X509_STORE_CTX
4486 internally such as S/MIME.
4488 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4489 trust settings if they are not set in X509_STORE. This allows X509_STORE
4490 purposes and trust (in S/MIME for example) to override any set by default.
4492 Add command line options for CRL checking to smime, s_client and s_server
4496 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4497 are set then the CRL is looked up in the X509_STORE structure and
4498 its validity and signature checked, then if the certificate is found
4499 in the CRL the verify fails with a revoked error.
4501 Various new CRL related callbacks added to X509_STORE_CTX structure.
4503 Command line options added to 'verify' application to support this.
4505 This needs some additional work, such as being able to handle multiple
4506 CRLs with different times, extension based lookup (rather than just
4507 by subject name) and ultimately more complete V2 CRL extension
4511 *) Add a general user interface API (crypto/ui/). This is designed
4512 to replace things like des_read_password and friends (backward
4513 compatibility functions using this new API are provided).
4514 The purpose is to remove prompting functions from the DES code
4515 section as well as provide for prompting through dialog boxes in
4516 a window system and the like.
4519 *) Add "ex_data" support to ENGINE so implementations can add state at a
4520 per-structure level rather than having to store it globally.
4523 *) Make it possible for ENGINE structures to be copied when retrieved by
4524 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4525 This causes the "original" ENGINE structure to act like a template,
4526 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4527 operational state can be localised to each ENGINE structure, despite the
4528 fact they all share the same "methods". New ENGINE structures returned in
4529 this case have no functional references and the return value is the single
4530 structural reference. This matches the single structural reference returned
4531 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4535 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4536 needs to match any other type at all we need to manually clear the
4540 *) Changes to the "openssl engine" utility to include;
4541 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4542 about an ENGINE's available control commands.
4543 - executing control commands from command line arguments using the
4544 '-pre' and '-post' switches. '-post' is only used if '-t' is
4545 specified and the ENGINE is successfully initialised. The syntax for
4546 the individual commands are colon-separated, for example;
4547 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4550 *) New dynamic control command support for ENGINEs. ENGINEs can now
4551 declare their own commands (numbers), names (strings), descriptions,
4552 and input types for run-time discovery by calling applications. A
4553 subset of these commands are implicitly classed as "executable"
4554 depending on their input type, and only these can be invoked through
4555 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4556 can be based on user input, config files, etc). The distinction is
4557 that "executable" commands cannot return anything other than a boolean
4558 result and can only support numeric or string input, whereas some
4559 discoverable commands may only be for direct use through
4560 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4561 pointers, or other custom uses. The "executable" commands are to
4562 support parameterisations of ENGINE behaviour that can be
4563 unambiguously defined by ENGINEs and used consistently across any
4564 OpenSSL-based application. Commands have been added to all the
4565 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4566 control over shared-library paths without source code alterations.
4569 *) Changed all ENGINE implementations to dynamically allocate their
4570 ENGINEs rather than declaring them statically. Apart from this being
4571 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4572 this also allows the implementations to compile without using the
4573 internal engine_int.h header.
4576 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4577 'const' value. Any code that should be able to modify a RAND_METHOD
4578 should already have non-const pointers to it (ie. they should only
4579 modify their own ones).
4582 *) Made a variety of little tweaks to the ENGINE code.
4583 - "atalla" and "ubsec" string definitions were moved from header files
4584 to C code. "nuron" string definitions were placed in variables
4585 rather than hard-coded - allowing parameterisation of these values
4586 later on via ctrl() commands.
4587 - Removed unused "#if 0"'d code.
4588 - Fixed engine list iteration code so it uses ENGINE_free() to release
4589 structural references.
4590 - Constified the RAND_METHOD element of ENGINE structures.
4591 - Constified various get/set functions as appropriate and added
4592 missing functions (including a catch-all ENGINE_cpy that duplicates
4593 all ENGINE values onto a new ENGINE except reference counts/state).
4594 - Removed NULL parameter checks in get/set functions. Setting a method
4595 or function to NULL is a way of cancelling out a previously set
4596 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4597 and doesn't justify the extra error symbols and code.
4598 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4599 flags from engine_int.h to engine.h.
4600 - Changed prototypes for ENGINE handler functions (init(), finish(),
4601 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4604 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4605 to the algorithm using long division. The binary algorithm can be
4606 used only if the modulus is odd. On 32-bit systems, it is faster
4607 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4608 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4609 up to 450 bits. In 64-bit environments, the binary algorithm
4610 appears to be advantageous for much longer moduli; here we use it
4611 for moduli up to 2048 bits.
4614 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4615 could not support the combine flag in choice fields.
4618 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4619 extensions from a certificate request to the certificate.
4622 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4623 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4624 file: this allows the display of the certificate about to be
4625 signed to be customised, to allow certain fields to be included
4626 or excluded and extension details. The old system didn't display
4627 multicharacter strings properly, omitted fields not in the policy
4628 and couldn't display additional details such as extensions.
4631 *) Function EC_POINTs_mul for multiple scalar multiplication
4632 of an arbitrary number of elliptic curve points
4633 \sum scalars[i]*points[i],
4634 optionally including the generator defined for the EC_GROUP:
4635 scalar*generator + \sum scalars[i]*points[i].
4637 EC_POINT_mul is a simple wrapper function for the typical case
4638 that the point list has just one item (besides the optional
4642 *) First EC_METHODs for curves over GF(p):
4644 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4645 operations and provides various method functions that can also
4646 operate with faster implementations of modular arithmetic.
4648 EC_GFp_mont_method() reuses most functions that are part of
4649 EC_GFp_simple_method, but uses Montgomery arithmetic.
4651 [Bodo Moeller; point addition and point doubling
4652 implementation directly derived from source code provided by
4653 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4655 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4656 crypto/ec/ec_lib.c):
4658 Curves are EC_GROUP objects (with an optional group generator)
4659 based on EC_METHODs that are built into the library.
4661 Points are EC_POINT objects based on EC_GROUP objects.
4663 Most of the framework would be able to handle curves over arbitrary
4664 finite fields, but as there are no obvious types for fields other
4665 than GF(p), some functions are limited to that for now.
4668 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4669 that the file contains a complete HTTP response.
4672 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4673 change the def and num file printf format specifier from "%-40sXXX"
4674 to "%-39s XXX". The latter will always guarantee a space after the
4675 field while the former will cause them to run together if the field
4676 is 40 of more characters long.
4679 *) Constify the cipher and digest 'method' functions and structures
4680 and modify related functions to take constant EVP_MD and EVP_CIPHER
4684 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4685 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4688 *) Modify EVP_Digest*() routines so they now return values. Although the
4689 internal software routines can never fail additional hardware versions
4693 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4695 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4696 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4699 ERR_R_NESTED_ASN1_ERROR
4701 ERR_R_MISSING_ASN1_EOS
4702 were 4 .. 9, conflicting with
4703 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4705 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4706 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4708 Add new error code 'ERR_R_INTERNAL_ERROR'.
4711 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4715 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4716 sets the subject name for a new request or supersedes the
4717 subject name in a given request. Formats that can be parsed are
4718 'CN=Some Name, OU=myOU, C=IT'
4720 'CN=Some Name/OU=myOU/C=IT'.
4722 Add options '-batch' and '-verbose' to 'openssl req'.
4723 [Massimiliano Pala <madwolf@hackmasters.net>]
4725 *) Introduce the possibility to access global variables through
4726 functions on platform were that's the best way to handle exporting
4727 global variables in shared libraries. To enable this functionality,
4728 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4729 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4730 is normally done by Configure or something similar).
4732 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4733 in the source file (foo.c) like this:
4735 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4736 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4738 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4739 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4741 OPENSSL_DECLARE_GLOBAL(int,foo);
4742 #define foo OPENSSL_GLOBAL_REF(foo)
4743 OPENSSL_DECLARE_GLOBAL(double,bar);
4744 #define bar OPENSSL_GLOBAL_REF(bar)
4746 The #defines are very important, and therefore so is including the
4747 header file everywhere where the defined globals are used.
4749 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4750 of ASN.1 items, but that structure is a bit different.
4752 The largest change is in util/mkdef.pl which has been enhanced with
4753 better and easier to understand logic to choose which symbols should
4754 go into the Windows .def files as well as a number of fixes and code
4755 cleanup (among others, algorithm keywords are now sorted
4756 lexicographically to avoid constant rewrites).
4759 *) In BN_div() keep a copy of the sign of 'num' before writing the
4760 result to 'rm' because if rm==num the value will be overwritten
4761 and produce the wrong result if 'num' is negative: this caused
4762 problems with BN_mod() and BN_nnmod().
4765 *) Function OCSP_request_verify(). This checks the signature on an
4766 OCSP request and verifies the signer certificate. The signer
4767 certificate is just checked for a generic purpose and OCSP request
4771 *) Add OCSP_check_validity() function to check the validity of OCSP
4772 responses. OCSP responses are prepared in real time and may only
4773 be a few seconds old. Simply checking that the current time lies
4774 between thisUpdate and nextUpdate max reject otherwise valid responses
4775 caused by either OCSP responder or client clock inaccuracy. Instead
4776 we allow thisUpdate and nextUpdate to fall within a certain period of
4777 the current time. The age of the response can also optionally be
4778 checked. Two new options -validity_period and -status_age added to
4782 *) If signature or public key algorithm is unrecognized print out its
4783 OID rather that just UNKNOWN.
4786 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4787 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4788 ID to be generated from the issuer certificate alone which can then be
4789 passed to OCSP_id_issuer_cmp().
4792 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4793 ASN1 modules to export functions returning ASN1_ITEM pointers
4794 instead of the ASN1_ITEM structures themselves. This adds several
4795 new macros which allow the underlying ASN1 function/structure to
4796 be accessed transparently. As a result code should not use ASN1_ITEM
4797 references directly (such as &X509_it) but instead use the relevant
4798 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4799 use of the new ASN1 code on platforms where exporting structures
4800 is problematical (for example in shared libraries) but exporting
4801 functions returning pointers to structures is not.
4804 *) Add support for overriding the generation of SSL/TLS session IDs.
4805 These callbacks can be registered either in an SSL_CTX or per SSL.
4806 The purpose of this is to allow applications to control, if they wish,
4807 the arbitrary values chosen for use as session IDs, particularly as it
4808 can be useful for session caching in multiple-server environments. A
4809 command-line switch for testing this (and any client code that wishes
4810 to use such a feature) has been added to "s_server".
4811 [Geoff Thorpe, Lutz Jaenicke]
4813 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4814 of the form '#if defined(...) || defined(...) || ...' and
4815 '#if !defined(...) && !defined(...) && ...'. This also avoids
4816 the growing number of special cases it was previously handling.
4819 *) Make all configuration macros available for application by making
4820 sure they are available in opensslconf.h, by giving them names starting
4821 with "OPENSSL_" to avoid conflicts with other packages and by making
4822 sure e_os2.h will cover all platform-specific cases together with
4824 Additionally, it is now possible to define configuration/platform-
4825 specific names (called "system identities"). In the C code, these
4826 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4827 macro with the name beginning with "OPENSSL_SYS_", which is determined
4828 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4832 *) New option -set_serial to 'req' and 'x509' this allows the serial
4833 number to use to be specified on the command line. Previously self
4834 signed certificates were hard coded with serial number 0 and the
4835 CA options of 'x509' had to use a serial number in a file which was
4839 *) New options to 'ca' utility to support V2 CRL entry extensions.
4840 Currently CRL reason, invalidity date and hold instruction are
4841 supported. Add new CRL extensions to V3 code and some new objects.
4844 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4845 disable standard block padding (aka PKCS#5 padding) in the EVP
4846 API, which was previously mandatory. This means that the data is
4847 not padded in any way and so the total length much be a multiple
4848 of the block size, otherwise an error occurs.
4851 *) Initial (incomplete) OCSP SSL support.
4854 *) New function OCSP_parse_url(). This splits up a URL into its host,
4855 port and path components: primarily to parse OCSP URLs. New -url
4856 option to ocsp utility.
4859 *) New nonce behavior. The return value of OCSP_check_nonce() now
4860 reflects the various checks performed. Applications can decide
4861 whether to tolerate certain situations such as an absent nonce
4862 in a response when one was present in a request: the ocsp application
4863 just prints out a warning. New function OCSP_add1_basic_nonce()
4864 this is to allow responders to include a nonce in a response even if
4865 the request is nonce-less.
4868 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4869 skipped when using openssl x509 multiple times on a single input file,
4870 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4873 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4874 set string type: to handle setting ASN1_TIME structures. Fix ca
4875 utility to correctly initialize revocation date of CRLs.
4878 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4879 the clients preferred ciphersuites and rather use its own preferences.
4880 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4881 Internet Explorer by ensuring unchanged hash method during stepup.
4882 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4885 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4886 to aes and add a new 'exist' option to print out symbols that don't
4890 *) Additional options to ocsp utility to allow flags to be set and
4891 additional certificates supplied.
4894 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4895 OCSP client a number of certificate to only verify the response
4899 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4900 handle the new API. Currently only ECB, CBC modes supported. Add new
4903 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4904 Encryption Standard (AES) Ciphersuites for Transport Layer
4905 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4906 not enabled by default and were not part of the "ALL" ciphersuite
4907 alias because they were not yet official; they could be
4908 explicitly requested by specifying the "AESdraft" ciphersuite
4909 group alias. In the final release of OpenSSL 0.9.7, the group
4910 alias is called "AES" and is part of "ALL".)
4911 [Ben Laurie, Steve Henson, Bodo Moeller]
4913 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4914 request to response.
4917 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4918 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4919 extract information from a certificate request. OCSP_response_create()
4920 creates a response and optionally adds a basic response structure.
4921 OCSP_basic_add1_status() adds a complete single response to a basic
4922 response and returns the OCSP_SINGLERESP structure just added (to allow
4923 extensions to be included for example). OCSP_basic_add1_cert() adds a
4924 certificate to a basic response and OCSP_basic_sign() signs a basic
4925 response with various flags. New helper functions ASN1_TIME_check()
4926 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4927 (converts ASN1_TIME to GeneralizedTime).
4930 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4931 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4932 structure from a certificate. X509_pubkey_digest() digests the public_key
4933 contents: this is used in various key identifiers.
4936 *) Make sk_sort() tolerate a NULL argument.
4937 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4939 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4940 passed by the function are trusted implicitly. If any of them signed the
4941 response then it is assumed to be valid and is not verified.
4944 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4945 to data. This was previously part of the PKCS7 ASN1 code. This
4946 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4947 [Steve Henson, reported by Kenneth R. Robinette
4948 <support@securenetterm.com>]
4950 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4951 routines: without these tracing memory leaks is very painful.
4952 Fix leaks in PKCS12 and PKCS7 routines.
4955 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4956 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4957 effectively meant GeneralizedTime would never be used. Now it
4958 is initialised to -1 but X509_time_adj() now has to check the value
4959 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4960 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4961 [Steve Henson, reported by Kenneth R. Robinette
4962 <support@securenetterm.com>]
4964 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4965 result in a zero length in the ASN1_INTEGER structure which was
4966 not consistent with the structure when d2i_ASN1_INTEGER() was used
4967 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4968 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4969 where it did not print out a minus for negative ASN1_INTEGER.
4972 *) Add summary printout to ocsp utility. The various functions which
4973 convert status values to strings have been renamed to:
4974 OCSP_response_status_str(), OCSP_cert_status_str() and
4975 OCSP_crl_reason_str() and are no longer static. New options
4976 to verify nonce values and to disable verification. OCSP response
4977 printout format cleaned up.
4980 *) Add additional OCSP certificate checks. These are those specified
4981 in RFC2560. This consists of two separate checks: the CA of the
4982 certificate being checked must either be the OCSP signer certificate
4983 or the issuer of the OCSP signer certificate. In the latter case the
4984 OCSP signer certificate must contain the OCSP signing extended key
4985 usage. This check is performed by attempting to match the OCSP
4986 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4987 in the OCSP_CERTID structures of the response.
4990 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4991 and related routines. This uses the standard OpenSSL certificate
4992 verify routines to perform initial checks (just CA validity) and
4993 to obtain the certificate chain. Then additional checks will be
4994 performed on the chain. Currently the root CA is checked to see
4995 if it is explicitly trusted for OCSP signing. This is used to set
4996 a root CA as a global signing root: that is any certificate that
4997 chains to that CA is an acceptable OCSP signing certificate.
5000 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5001 extensions from a separate configuration file.
5002 As when reading extensions from the main configuration file,
5003 the '-extensions ...' option may be used for specifying the
5005 [Massimiliano Pala <madwolf@comune.modena.it>]
5007 *) New OCSP utility. Allows OCSP requests to be generated or
5008 read. The request can be sent to a responder and the output
5009 parsed, outputed or printed in text form. Not complete yet:
5010 still needs to check the OCSP response validity.
5013 *) New subcommands for 'openssl ca':
5014 'openssl ca -status <serial>' prints the status of the cert with
5015 the given serial number (according to the index file).
5016 'openssl ca -updatedb' updates the expiry status of certificates
5018 [Massimiliano Pala <madwolf@comune.modena.it>]
5020 *) New '-newreq-nodes' command option to CA.pl. This is like
5021 '-newreq', but calls 'openssl req' with the '-nodes' option
5022 so that the resulting key is not encrypted.
5023 [Damien Miller <djm@mindrot.org>]
5025 *) New configuration for the GNU Hurd.
5026 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5028 *) Initial code to implement OCSP basic response verify. This
5029 is currently incomplete. Currently just finds the signer's
5030 certificate and verifies the signature on the response.
5033 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5034 value of OPENSSLDIR. This is available via the new '-d' option
5035 to 'openssl version', and is also included in 'openssl version -a'.
5038 *) Allowing defining memory allocation callbacks that will be given
5039 file name and line number information in additional arguments
5040 (a const char* and an int). The basic functionality remains, as
5041 well as the original possibility to just replace malloc(),
5042 realloc() and free() by functions that do not know about these
5043 additional arguments. To register and find out the current
5044 settings for extended allocation functions, the following
5045 functions are provided:
5047 CRYPTO_set_mem_ex_functions
5048 CRYPTO_set_locked_mem_ex_functions
5049 CRYPTO_get_mem_ex_functions
5050 CRYPTO_get_locked_mem_ex_functions
5052 These work the same way as CRYPTO_set_mem_functions and friends.
5053 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5054 extended allocation function is enabled.
5055 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5056 a conventional allocation function is enabled.
5057 [Richard Levitte, Bodo Moeller]
5059 *) Finish off removing the remaining LHASH function pointer casts.
5060 There should no longer be any prototype-casting required when using
5061 the LHASH abstraction, and any casts that remain are "bugs". See
5062 the callback types and macros at the head of lhash.h for details
5063 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5066 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5067 If /dev/[u]random devices are not available or do not return enough
5068 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5070 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5071 /etc/entropy will be queried once each in this sequence, quering stops
5072 when enough entropy was collected without querying more sockets.
5075 *) Change the Unix RAND_poll() variant to be able to poll several
5076 random devices, as specified by DEVRANDOM, until a sufficient amount
5077 of data has been collected. We spend at most 10 ms on each file
5078 (select timeout) and read in non-blocking mode. DEVRANDOM now
5079 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5080 (previously it was just the string "/dev/urandom"), so on typical
5081 platforms the 10 ms delay will never occur.
5082 Also separate out the Unix variant to its own file, rand_unix.c.
5083 For VMS, there's a currently-empty rand_vms.c.
5086 *) Move OCSP client related routines to ocsp_cl.c. These
5087 provide utility functions which an application needing
5088 to issue a request to an OCSP responder and analyse the
5089 response will typically need: as opposed to those which an
5090 OCSP responder itself would need which will be added later.
5092 OCSP_request_sign() signs an OCSP request with an API similar
5093 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5094 response. OCSP_response_get1_basic() extracts basic response
5095 from response. OCSP_resp_find_status(): finds and extracts status
5096 information from an OCSP_CERTID structure (which will be created
5097 when the request structure is built). These are built from lower
5098 level functions which work on OCSP_SINGLERESP structures but
5099 wont normally be used unless the application wishes to examine
5100 extensions in the OCSP response for example.
5102 Replace nonce routines with a pair of functions.
5103 OCSP_request_add1_nonce() adds a nonce value and optionally
5104 generates a random value. OCSP_check_nonce() checks the
5105 validity of the nonce in an OCSP response.
5108 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5109 This doesn't copy the supplied OCSP_CERTID and avoids the
5110 need to free up the newly created id. Change return type
5111 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5112 This can then be used to add extensions to the request.
5113 Deleted OCSP_request_new(), since most of its functionality
5114 is now in OCSP_REQUEST_new() (and the case insensitive name
5115 clash) apart from the ability to set the request name which
5116 will be added elsewhere.
5119 *) Update OCSP API. Remove obsolete extensions argument from
5120 various functions. Extensions are now handled using the new
5121 OCSP extension code. New simple OCSP HTTP function which
5122 can be used to send requests and parse the response.
5125 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5126 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5127 uses the special reorder version of SET OF to sort the attributes
5128 and reorder them to match the encoded order. This resolves a long
5129 standing problem: a verify on a PKCS7 structure just after signing
5130 it used to fail because the attribute order did not match the
5131 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5132 it uses the received order. This is necessary to tolerate some broken
5133 software that does not order SET OF. This is handled by encoding
5134 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5135 to produce the required SET OF.
5138 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5139 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5140 files to get correct declarations of the ASN.1 item variables.
5143 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5144 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5145 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5146 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5147 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5148 ASN1_ITEM and no wrapper functions.
5151 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5152 replace the old function pointer based I/O routines. Change most of
5153 the *_d2i_bio() and *_d2i_fp() functions to use these.
5156 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5157 lines, recognice more "algorithms" that can be deselected, and make
5158 it complain about algorithm deselection that isn't recognised.
5161 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5162 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5163 to use new functions. Add NO_ASN1_OLD which can be set to remove
5164 some old style ASN1 functions: this can be used to determine if old
5165 code will still work when these eventually go away.
5168 *) New extension functions for OCSP structures, these follow the
5169 same conventions as certificates and CRLs.
5172 *) New function X509V3_add1_i2d(). This automatically encodes and
5173 adds an extension. Its behaviour can be customised with various
5174 flags to append, replace or delete. Various wrappers added for
5175 certifcates and CRLs.
5178 *) Fix to avoid calling the underlying ASN1 print routine when
5179 an extension cannot be parsed. Correct a typo in the
5180 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5183 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5184 entries for variables.
5187 *) Add functionality to apps/openssl.c for detecting locking
5188 problems: As the program is single-threaded, all we have
5189 to do is register a locking callback using an array for
5190 storing which locks are currently held by the program.
5193 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5194 SSL_get_ex_data_X509_STORE_idx(), which is used in
5195 ssl_verify_cert_chain() and thus can be called at any time
5196 during TLS/SSL handshakes so that thread-safety is essential.
5197 Unfortunately, the ex_data design is not at all suited
5198 for multi-threaded use, so it probably should be abolished.
5201 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5202 [Broadcom, tweaked and integrated by Geoff Thorpe]
5204 *) Move common extension printing code to new function
5205 X509V3_print_extensions(). Reorganise OCSP print routines and
5206 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5209 *) New function X509_signature_print() to remove duplication in some
5213 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5214 set (this was treated exactly the same as SET OF previously). This
5215 is used to reorder the STACK representing the structure to match the
5216 encoding. This will be used to get round a problem where a PKCS7
5217 structure which was signed could not be verified because the STACK
5218 order did not reflect the encoded order.
5221 *) Reimplement the OCSP ASN1 module using the new code.
5224 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5225 for its ASN1 operations. The old style function pointers still exist
5226 for now but they will eventually go away.
5229 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5230 completely replaces the old ASN1 functionality with a table driven
5231 encoder and decoder which interprets an ASN1_ITEM structure describing
5232 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5233 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5234 has also been converted to the new form.
5237 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5238 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5239 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5240 for negative moduli.
5243 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5244 of not touching the result's sign bit.
5247 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5251 *) Changed the LHASH code to use prototypes for callbacks, and created
5252 macros to declare and implement thin (optionally static) functions
5253 that provide type-safety and avoid function pointer casting for the
5254 type-specific callbacks.
5257 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5259 [Veers Staats <staatsvr@asc.hpc.mil>,
5260 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5262 *) Reformat the FAQ so the different questions and answers can be divided
5263 in sections depending on the subject.
5266 *) Have the zlib compression code load ZLIB.DLL dynamically under
5270 *) New function BN_mod_sqrt for computing square roots modulo a prime
5271 (using the probabilistic Tonelli-Shanks algorithm unless
5272 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5273 be handled deterministically).
5274 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5276 *) Make BN_mod_inverse faster by explicitly handling small quotients
5277 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5278 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5281 *) New function BN_kronecker.
5284 *) Fix BN_gcd so that it works on negative inputs; the result is
5285 positive unless both parameters are zero.
5286 Previously something reasonably close to an infinite loop was
5287 possible because numbers could be growing instead of shrinking
5288 in the implementation of Euclid's algorithm.
5291 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5292 sign of the number in question.
5294 Fix BN_is_word(a,w) to work correctly for w == 0.
5296 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5297 because its test if the absolute value of 'a' equals 'w'.
5298 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5299 it exists mostly for use in the implementations of BN_is_zero(),
5300 BN_is_one(), and BN_is_word().
5303 *) New function BN_swap.
5306 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5307 the exponentiation functions are more likely to produce reasonable
5308 results on negative inputs.
5311 *) Change BN_mod_mul so that the result is always non-negative.
5312 Previously, it could be negative if one of the factors was negative;
5313 I don't think anyone really wanted that behaviour.
5316 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5317 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5318 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5319 and add new functions:
5328 BN_mod_lshift1_quick
5332 These functions always generate non-negative results.
5334 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5335 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5337 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5338 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5339 be reduced modulo m.
5340 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5343 The following entry accidentily appeared in the CHANGES file
5344 distributed with OpenSSL 0.9.7. The modifications described in
5345 it do *not* apply to OpenSSL 0.9.7.
5347 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5348 was actually never needed) and in BN_mul(). The removal in BN_mul()
5349 required a small change in bn_mul_part_recursive() and the addition
5350 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5351 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5352 bn_sub_words() and bn_add_words() except they take arrays with
5357 *) In 'openssl passwd', verify passwords read from the terminal
5358 unless the '-salt' option is used (which usually means that
5359 verification would just waste user's time since the resulting
5360 hash is going to be compared with some given password hash)
5361 or the new '-noverify' option is used.
5363 This is an incompatible change, but it does not affect
5364 non-interactive use of 'openssl passwd' (passwords on the command
5365 line, '-stdin' option, '-in ...' option) and thus should not
5369 *) Remove all references to RSAref, since there's no more need for it.
5372 *) Make DSO load along a path given through an environment variable
5373 (SHLIB_PATH) with shl_load().
5376 *) Constify the ENGINE code as a result of BIGNUM constification.
5377 Also constify the RSA code and most things related to it. In a
5378 few places, most notable in the depth of the ASN.1 code, ugly
5379 casts back to non-const were required (to be solved at a later
5383 *) Make it so the openssl application has all engines loaded by default.
5386 *) Constify the BIGNUM routines a little more.
5389 *) Add the following functions:
5391 ENGINE_load_cswift()
5393 ENGINE_load_atalla()
5395 ENGINE_load_builtin_engines()
5397 That way, an application can itself choose if external engines that
5398 are built-in in OpenSSL shall ever be used or not. The benefit is
5399 that applications won't have to be linked with libdl or other dso
5400 libraries unless it's really needed.
5402 Changed 'openssl engine' to load all engines on demand.
5403 Changed the engine header files to avoid the duplication of some
5404 declarations (they differed!).
5407 *) 'openssl engine' can now list capabilities.
5410 *) Better error reporting in 'openssl engine'.
5413 *) Never call load_dh_param(NULL) in s_server.
5416 *) Add engine application. It can currently list engines by name and
5417 identity, and test if they are actually available.
5420 *) Improve RPM specification file by forcing symbolic linking and making
5421 sure the installed documentation is also owned by root.root.
5422 [Damien Miller <djm@mindrot.org>]
5424 *) Give the OpenSSL applications more possibilities to make use of
5425 keys (public as well as private) handled by engines.
5428 *) Add OCSP code that comes from CertCo.
5431 *) Add VMS support for the Rijndael code.
5434 *) Added untested support for Nuron crypto accelerator.
5437 *) Add support for external cryptographic devices. This code was
5438 previously distributed separately as the "engine" branch.
5439 [Geoff Thorpe, Richard Levitte]
5441 *) Rework the filename-translation in the DSO code. It is now possible to
5442 have far greater control over how a "name" is turned into a filename
5443 depending on the operating environment and any oddities about the
5444 different shared library filenames on each system.
5447 *) Support threads on FreeBSD-elf in Configure.
5450 *) Fix for SHA1 assembly problem with MASM: it produces
5451 warnings about corrupt line number information when assembling
5452 with debugging information. This is caused by the overlapping
5454 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5457 NCONF_get_number() has no error checking at all. As a replacement,
5458 NCONF_get_number_e() is defined (_e for "error checking") and is
5459 promoted strongly. The old NCONF_get_number is kept around for
5460 binary backward compatibility.
5461 Make it possible for methods to load from something other than a BIO,
5462 by providing a function pointer that is given a name instead of a BIO.
5463 For example, this could be used to load configuration data from an
5467 *) Fix for non blocking accept BIOs. Added new I/O special reason
5468 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5469 with non blocking I/O was not possible because no retry code was
5470 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5474 *) Added the beginnings of Rijndael support.
5477 *) Fix for bug in DirectoryString mask setting. Add support for
5478 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5479 to allow certificate printing to more controllable, additional
5480 'certopt' option to 'x509' to allow new printing options to be
5484 *) Clean old EAY MD5 hack from e_os.h.
5487 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5489 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5490 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5491 [Joe Orton, Steve Henson]
5493 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5495 *) Fix additional bug revealed by the NISCC test suite:
5497 Stop bug triggering large recursion when presented with
5498 certain ASN.1 tags (CVE-2003-0851)
5501 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5503 *) Fix various bugs revealed by running the NISCC test suite:
5505 Stop out of bounds reads in the ASN1 code when presented with
5506 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5508 If verify callback ignores invalid public key errors don't try to check
5509 certificate signature with the NULL public key.
5513 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5514 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5518 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5519 extra data after the compression methods not only for TLS 1.0
5520 but also for SSL 3.0 (as required by the specification).
5521 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5523 *) Change X509_certificate_type() to mark the key as exported/exportable
5524 when it's 512 *bits* long, not 512 bytes.
5527 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5529 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5530 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5531 a protocol version number mismatch like a decryption error
5532 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5535 *) Turn on RSA blinding by default in the default implementation
5536 to avoid a timing attack. Applications that don't want it can call
5537 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5538 They would be ill-advised to do so in most cases.
5539 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5541 *) Change RSA blinding code so that it works when the PRNG is not
5542 seeded (in this case, the secret RSA exponent is abused as
5543 an unpredictable seed -- if it is not unpredictable, there
5544 is no point in blinding anyway). Make RSA blinding thread-safe
5545 by remembering the creator's thread ID in rsa->blinding and
5546 having all other threads use local one-time blinding factors
5547 (this requires more computation than sharing rsa->blinding, but
5548 avoids excessive locking; and if an RSA object is not shared
5549 between threads, blinding will still be very fast).
5552 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5554 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5555 via timing by performing a MAC computation even if incorrrect
5556 block cipher padding has been found. This is a countermeasure
5557 against active attacks where the attacker has to distinguish
5558 between bad padding and a MAC verification error. (CVE-2003-0078)
5560 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5561 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5562 Martin Vuagnoux (EPFL, Ilion)]
5564 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5566 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5567 memory from it's contents. This is done with a counter that will
5568 place alternating values in each byte. This can be used to solve
5569 two issues: 1) the removal of calls to memset() by highly optimizing
5570 compilers, and 2) cleansing with other values than 0, since those can
5571 be read through on certain media, for example a swap space on disk.
5574 *) Bugfix: client side session caching did not work with external caching,
5575 because the session->cipher setting was not restored when reloading
5576 from the external cache. This problem was masked, when
5577 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5578 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5581 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5582 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5583 [Zeev Lieber <zeev-l@yahoo.com>]
5585 *) Undo an undocumented change introduced in 0.9.6e which caused
5586 repeated calls to OpenSSL_add_all_ciphers() and
5587 OpenSSL_add_all_digests() to be ignored, even after calling
5591 *) Change the default configuration reader to deal with last line not
5592 being properly terminated.
5595 *) Change X509_NAME_cmp() so it applies the special rules on handling
5596 DN values that are of type PrintableString, as well as RDNs of type
5597 emailAddress where the value has the type ia5String.
5598 [stefank@valicert.com via Richard Levitte]
5600 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5601 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5602 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5603 the bitwise-OR of the two for use by the majority of applications
5604 wanting this behaviour, and update the docs. The documented
5605 behaviour and actual behaviour were inconsistent and had been
5606 changing anyway, so this is more a bug-fix than a behavioural
5608 [Geoff Thorpe, diagnosed by Nadav Har'El]
5610 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5611 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5614 *) Fix initialization code race conditions in
5615 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5616 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5617 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5618 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5619 ssl2_get_cipher_by_char(),
5620 ssl3_get_cipher_by_char().
5621 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5623 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5624 the cached sessions are flushed, as the remove_cb() might use ex_data
5625 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5626 (see [openssl.org #212]).
5627 [Geoff Thorpe, Lutz Jaenicke]
5629 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5630 length, instead of the encoding length to d2i_ASN1_OBJECT.
5633 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5635 *) [In 0.9.6g-engine release:]
5636 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5637 [Lynn Gazis <lgazis@rainbow.com>]
5639 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5641 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5642 and get fix the header length calculation.
5643 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5644 Alon Kantor <alonk@checkpoint.com> (and others),
5647 *) Use proper error handling instead of 'assertions' in buffer
5648 overflow checks added in 0.9.6e. This prevents DoS (the
5649 assertions could call abort()).
5650 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5652 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5654 *) Add various sanity checks to asn1_get_length() to reject
5655 the ASN1 length bytes if they exceed sizeof(long), will appear
5656 negative or the content length exceeds the length of the
5658 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5660 *) Fix cipher selection routines: ciphers without encryption had no flags
5661 for the cipher strength set and where therefore not handled correctly
5662 by the selection routines (PR #130).
5665 *) Fix EVP_dsa_sha macro.
5669 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5670 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5671 that was added in OpenSSL 0.9.6d.
5673 As the countermeasure turned out to be incompatible with some
5674 broken SSL implementations, the new option is part of SSL_OP_ALL.
5675 SSL_OP_ALL is usually employed when compatibility with weird SSL
5676 implementations is desired (e.g. '-bugs' option to 's_client' and
5677 's_server'), so the new option is automatically set in many
5681 *) Changes in security patch:
5683 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5684 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5685 Air Force Materiel Command, USAF, under agreement number
5688 *) Add various sanity checks to asn1_get_length() to reject
5689 the ASN1 length bytes if they exceed sizeof(long), will appear
5690 negative or the content length exceeds the length of the
5691 supplied buffer. (CVE-2002-0659)
5692 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5694 *) Assertions for various potential buffer overflows, not known to
5696 [Ben Laurie (CHATS)]
5698 *) Various temporary buffers to hold ASCII versions of integers were
5699 too small for 64 bit platforms. (CVE-2002-0655)
5700 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5702 *) Remote buffer overflow in SSL3 protocol - an attacker could
5703 supply an oversized session ID to a client. (CVE-2002-0656)
5704 [Ben Laurie (CHATS)]
5706 *) Remote buffer overflow in SSL2 protocol - an attacker could
5707 supply an oversized client master key. (CVE-2002-0656)
5708 [Ben Laurie (CHATS)]
5710 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5712 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5713 encoded as NULL) with id-dsa-with-sha1.
5714 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5716 *) Check various X509_...() return values in apps/req.c.
5717 [Nils Larsch <nla@trustcenter.de>]
5719 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5720 an end-of-file condition would erronously be flagged, when the CRLF
5721 was just at the end of a processed block. The bug was discovered when
5722 processing data through a buffering memory BIO handing the data to a
5723 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5724 <ptsekov@syntrex.com> and Nedelcho Stanev.
5727 *) Implement a countermeasure against a vulnerability recently found
5728 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5729 before application data chunks to avoid the use of known IVs
5730 with data potentially chosen by the attacker.
5733 *) Fix length checks in ssl3_get_client_hello().
5736 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5737 to prevent ssl3_read_internal() from incorrectly assuming that
5738 ssl3_read_bytes() found application data while handshake
5739 processing was enabled when in fact s->s3->in_read_app_data was
5740 merely automatically cleared during the initial handshake.
5741 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5743 *) Fix object definitions for Private and Enterprise: they were not
5744 recognized in their shortname (=lowercase) representation. Extend
5745 obj_dat.pl to issue an error when using undefined keywords instead
5746 of silently ignoring the problem (Svenning Sorensen
5747 <sss@sss.dnsalias.net>).
5750 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5751 generators, i.e. generators other than 2 and 5. (Previously, the
5752 code did not properly initialise the 'add' and 'rem' values to
5753 BN_generate_prime().)
5755 In the new general case, we do not insist that 'generator' is
5756 actually a primitive root: This requirement is rather pointless;
5757 a generator of the order-q subgroup is just as good, if not
5761 *) Map new X509 verification errors to alerts. Discovered and submitted by
5762 Tom Wu <tom@arcot.com>.
5765 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5766 returning non-zero before the data has been completely received
5767 when using non-blocking I/O.
5768 [Bodo Moeller; problem pointed out by John Hughes]
5770 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5771 [Ben Laurie, Lutz Jaenicke]
5773 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5774 Yoram Zahavi <YoramZ@gilian.com>).
5777 *) Add information about CygWin 1.3 and on, and preserve proper
5778 configuration for the versions before that.
5779 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5781 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5782 check whether we deal with a copy of a session and do not delete from
5783 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5784 <izhar@checkpoint.com>.
5787 *) Do not store session data into the internal session cache, if it
5788 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5789 flag is set). Proposed by Aslam <aslam@funk.com>.
5792 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5796 *) [In 0.9.6d-engine release:]
5797 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5798 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5800 *) Add the configuration target linux-s390x.
5801 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5803 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5804 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5805 variable as an indication that a ClientHello message has been
5806 received. As the flag value will be lost between multiple
5807 invocations of ssl3_accept when using non-blocking I/O, the
5808 function may not be aware that a handshake has actually taken
5809 place, thus preventing a new session from being added to the
5812 To avoid this problem, we now set s->new_session to 2 instead of
5813 using a local variable.
5814 [Lutz Jaenicke, Bodo Moeller]
5816 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5817 if the SSL_R_LENGTH_MISMATCH error is detected.
5818 [Geoff Thorpe, Bodo Moeller]
5820 *) New 'shared_ldflag' column in Configure platform table.
5823 *) Fix EVP_CIPHER_mode macro.
5824 ["Dan S. Camper" <dan@bti.net>]
5826 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5827 type, we must throw them away by setting rr->length to 0.
5828 [D P Chang <dpc@qualys.com>]
5830 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5832 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5833 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5834 worked incorrectly for those cases where range = 10..._2 and
5835 3*range is two bits longer than range.)
5838 *) Only add signing time to PKCS7 structures if it is not already
5842 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5843 OBJ_ld_ce should be OBJ_id_ce.
5844 Also some ip-pda OIDs in crypto/objects/objects.txt were
5845 incorrect (cf. RFC 3039).
5846 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5848 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5849 returns early because it has nothing to do.
5850 [Andy Schneider <andy.schneider@bjss.co.uk>]
5852 *) [In 0.9.6c-engine release:]
5853 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5854 [Andy Schneider <andy.schneider@bjss.co.uk>]
5856 *) [In 0.9.6c-engine release:]
5857 Add support for Cryptographic Appliance's keyserver technology.
5858 (Use engine 'keyclient')
5859 [Cryptographic Appliances and Geoff Thorpe]
5861 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5862 is called via tools/c89.sh because arguments have to be
5863 rearranged (all '-L' options must appear before the first object
5865 [Richard Shapiro <rshapiro@abinitio.com>]
5867 *) [In 0.9.6c-engine release:]
5868 Add support for Broadcom crypto accelerator cards, backported
5870 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5872 *) [In 0.9.6c-engine release:]
5873 Add support for SureWare crypto accelerator cards from
5874 Baltimore Technologies. (Use engine 'sureware')
5875 [Baltimore Technologies and Mark Cox]
5877 *) [In 0.9.6c-engine release:]
5878 Add support for crypto accelerator cards from Accelerated
5879 Encryption Processing, www.aep.ie. (Use engine 'aep')
5880 [AEP Inc. and Mark Cox]
5882 *) Add a configuration entry for gcc on UnixWare.
5883 [Gary Benson <gbenson@redhat.com>]
5885 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5886 messages are stored in a single piece (fixed-length part and
5887 variable-length part combined) and fix various bugs found on the way.
5890 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5891 instead. BIO_gethostbyname() does not know what timeouts are
5892 appropriate, so entries would stay in cache even when they have
5894 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5896 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5897 faced with a pathologically small ClientHello fragment that does
5898 not contain client_version: Instead of aborting with an error,
5899 simply choose the highest available protocol version (i.e.,
5900 TLS 1.0 unless it is disabled). In practice, ClientHello
5901 messages are never sent like this, but this change gives us
5902 strictly correct behaviour at least for TLS.
5905 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5906 never resets s->method to s->ctx->method when called from within
5907 one of the SSL handshake functions.
5908 [Bodo Moeller; problem pointed out by Niko Baric]
5910 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5911 (sent using the client's version number) if client_version is
5912 smaller than the protocol version in use. Also change
5913 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5914 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5915 the client will at least see that alert.
5918 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5922 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5923 client receives HelloRequest while in a handshake.
5924 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5926 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5927 should end in 'break', not 'goto end' which circuments various
5928 cleanups done in state SSL_ST_OK. But session related stuff
5929 must be disabled for SSL_ST_OK in the case that we just sent a
5932 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5933 before just sending a HelloRequest.
5934 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5936 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5937 reveal whether illegal block cipher padding was found or a MAC
5938 verification error occured. (Neither SSLerr() codes nor alerts
5939 are directly visible to potential attackers, but the information
5940 may leak via logfiles.)
5942 Similar changes are not required for the SSL 2.0 implementation
5943 because the number of padding bytes is sent in clear for SSL 2.0,
5944 and the extra bytes are just ignored. However ssl/s2_pkt.c
5945 failed to verify that the purported number of padding bytes is in
5949 *) Add OpenUNIX-8 support including shared libraries
5950 (Boyd Lynn Gerber <gerberb@zenez.com>).
5953 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5954 'wristwatch attack' using huge encoding parameters (cf.
5955 James H. Manger's CRYPTO 2001 paper). Note that the
5956 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5957 encoding parameters and hence was not vulnerable.
5960 *) BN_sqr() bug fix.
5961 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5963 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5964 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5965 followed by modular reduction.
5966 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5968 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5969 equivalent based on BN_pseudo_rand() instead of BN_rand().
5972 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5973 This function was broken, as the check for a new client hello message
5974 to handle SGC did not allow these large messages.
5975 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5978 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5981 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5982 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5985 *) Rework the configuration and shared library support for Tru64 Unix.
5986 The configuration part makes use of modern compiler features and
5987 still retains old compiler behavior for those that run older versions
5988 of the OS. The shared library support part includes a variant that
5989 uses the RPATH feature, and is available through the special
5990 configuration target "alpha-cc-rpath", which will never be selected
5992 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5994 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5995 with the same message size as in ssl3_get_certificate_request().
5996 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5997 messages might inadvertently be reject as too long.
5998 [Petr Lampa <lampa@fee.vutbr.cz>]
6000 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6003 *) Modified SSL library such that the verify_callback that has been set
6004 specificly for an SSL object with SSL_set_verify() is actually being
6005 used. Before the change, a verify_callback set with this function was
6006 ignored and the verify_callback() set in the SSL_CTX at the time of
6007 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6008 to allow the necessary settings.
6011 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6012 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6013 done automatically (in contradiction to the requirements of the C
6014 standard). This made problems when used from OpenSSH.
6017 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6018 dh->length and always used
6020 BN_rand_range(priv_key, dh->p).
6022 BN_rand_range() is not necessary for Diffie-Hellman, and this
6023 specific range makes Diffie-Hellman unnecessarily inefficient if
6024 dh->length (recommended exponent length) is much smaller than the
6025 length of dh->p. We could use BN_rand_range() if the order of
6026 the subgroup was stored in the DH structure, but we only have
6031 BN_rand(priv_key, l, ...)
6033 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6039 RSA_eay_public_encrypt
6040 RSA_eay_private_decrypt
6041 RSA_eay_private_encrypt (signing)
6042 RSA_eay_public_decrypt (signature verification)
6044 (default implementations for RSA_public_encrypt,
6045 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6046 always reject numbers >= n.
6049 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6050 to synchronize access to 'locking_thread'. This is necessary on
6051 systems where access to 'locking_thread' (an 'unsigned long'
6052 variable) is not atomic.
6055 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6056 *before* setting the 'crypto_lock_rand' flag. The previous code had
6057 a race condition if 0 is a valid thread ID.
6058 [Travis Vitek <vitek@roguewave.com>]
6060 *) Add support for shared libraries under Irix.
6061 [Albert Chin-A-Young <china@thewrittenword.com>]
6063 *) Add configuration option to build on Linux on both big-endian and
6065 [Ralf Baechle <ralf@uni-koblenz.de>]
6067 *) Add the possibility to create shared libraries on HP-UX.
6070 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6072 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6073 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6074 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6075 PRNG state recovery was possible based on the output of
6076 one PRNG request appropriately sized to gain knowledge on
6077 'md' followed by enough consecutive 1-byte PRNG requests
6078 to traverse all of 'state'.
6080 1. When updating 'md_local' (the current thread's copy of 'md')
6081 during PRNG output generation, hash all of the previous
6082 'md_local' value, not just the half used for PRNG output.
6084 2. Make the number of bytes from 'state' included into the hash
6085 independent from the number of PRNG bytes requested.
6087 The first measure alone would be sufficient to avoid
6088 Markku-Juhani's attack. (Actually it had never occurred
6089 to me that the half of 'md_local' used for chaining was the
6090 half from which PRNG output bytes were taken -- I had always
6091 assumed that the secret half would be used.) The second
6092 measure makes sure that additional data from 'state' is never
6093 mixed into 'md_local' in small portions; this heuristically
6094 further strengthens the PRNG.
6097 *) Fix crypto/bn/asm/mips3.s.
6100 *) When only the key is given to "enc", the IV is undefined. Print out
6101 an error message in this case.
6104 *) Handle special case when X509_NAME is empty in X509 printing routines.
6107 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6108 positive and less than q.
6111 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6112 used: it isn't thread safe and the add_lock_callback should handle
6114 [Paul Rose <Paul.Rose@bridge.com>]
6116 *) Verify that incoming data obeys the block size in
6117 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6121 [Ulf Möller, Bodo Möller]
6123 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6124 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6125 when fixing the server behaviour for backwards-compatible 'client
6126 hello' messages. (Note that the attack is impractical against
6127 SSL 3.0 and TLS 1.0 anyway because length and version checking
6128 means that the probability of guessing a valid ciphertext is
6129 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6132 Before 0.9.5, the countermeasure (hide the error by generating a
6133 random 'decryption result') did not work properly because
6134 ERR_clear_error() was missing, meaning that SSL_get_error() would
6135 detect the supposedly ignored error.
6137 Both problems are now fixed.
6140 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6141 (previously it was 1024).
6144 *) Fix for compatibility mode trust settings: ignore trust settings
6145 unless some valid trust or reject settings are present.
6148 *) Fix for blowfish EVP: its a variable length cipher.
6151 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6152 parameters in DSA public key structures and return an error in the
6153 DSA routines if parameters are absent.
6156 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6157 in the current directory if neither $RANDFILE nor $HOME was set.
6158 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6159 caused some confusion to Windows users who haven't defined $HOME.
6160 Thus RAND_file_name() is changed again: e_os.h can define a
6161 DEFAULT_HOME, which will be used if $HOME is not set.
6162 For Windows, we use "C:"; on other platforms, we still require
6163 environment variables.
6165 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6166 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6167 having multiple threads call RAND_poll() concurrently.
6170 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6171 combination of a flag and a thread ID variable.
6172 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6173 flag), *other* threads can enter ssleay_add_bytes without obeying
6174 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6175 that they do not hold after the first thread unsets add_do_not_lock).
6178 *) Change bctest again: '-x' expressions are not available in all
6182 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6184 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6185 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6187 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6188 the default extension for executables, if any. Also, make the perl
6189 scripts that use symlink() to test if it really exists and use "cp"
6190 if it doesn't. All this made OpenSSL compilable and installable in
6194 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6195 If SEQUENCE is length is indefinite just set c->slen to the total
6196 amount of data available.
6197 [Steve Henson, reported by shige@FreeBSD.org]
6198 [This change does not apply to 0.9.7.]
6200 *) Change bctest to avoid here-documents inside command substitution
6201 (workaround for FreeBSD /bin/sh bug).
6202 For compatibility with Ultrix, avoid shell functions (introduced
6203 in the bctest version that searches along $PATH).
6206 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6207 with des_encrypt() defined on some operating systems, like Solaris
6211 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6212 On the Importance of Eliminating Errors in Cryptographic
6213 Computations, J. Cryptology 14 (2001) 2, 101-119,
6214 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6217 *) MIPS assembler BIGNUM division bug fix.
6220 *) Disabled incorrect Alpha assembler code.
6223 *) Fix PKCS#7 decode routines so they correctly update the length
6224 after reading an EOC for the EXPLICIT tag.
6226 [This change does not apply to 0.9.7.]
6228 *) Fix bug in PKCS#12 key generation routines. This was triggered
6229 if a 3DES key was generated with a 0 initial byte. Include
6230 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6231 (but broken) behaviour.
6234 *) Enhance bctest to search for a working bc along $PATH and print
6236 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6238 *) Fix memory leaks in err.c: free err_data string if necessary;
6239 don't write to the wrong index in ERR_set_error_data.
6242 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6246 *) Replace rdtsc with _emit statements for VC++ version 5.
6247 [Jeremy Cooper <jeremy@baymoo.org>]
6249 *) Make it possible to reuse SSLv2 sessions.
6252 *) In copy_email() check for >= 0 as a return value for
6253 X509_NAME_get_index_by_NID() since 0 is a valid index.
6254 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6256 *) Avoid coredump with unsupported or invalid public keys by checking if
6257 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6258 PKCS7_verify() fails with non detached data.
6261 *) Don't use getenv in library functions when run as setuid/setgid.
6262 New function OPENSSL_issetugid().
6265 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6266 due to incorrect handling of multi-threading:
6268 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6270 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6272 3. Count how many times MemCheck_off() has been called so that
6273 nested use can be treated correctly. This also avoids
6274 inband-signalling in the previous code (which relied on the
6275 assumption that thread ID 0 is impossible).
6278 *) Add "-rand" option also to s_client and s_server.
6281 *) Fix CPU detection on Irix 6.x.
6282 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6283 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6285 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6288 [This change does not apply to 0.9.7.]
6290 *) Use the cached encoding of an X509_NAME structure rather than
6291 copying it. This is apparently the reason for the libsafe "errors"
6292 but the code is actually correct.
6295 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6296 Bleichenbacher's DSA attack.
6297 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6298 to be set and top=0 forces the highest bit to be set; top=-1 is new
6299 and leaves the highest bit random.
6300 [Ulf Moeller, Bodo Moeller]
6302 *) In the NCONF_...-based implementations for CONF_... queries
6303 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6304 a temporary CONF structure with the data component set to NULL
6305 (which gives segmentation faults in lh_retrieve).
6306 Instead, use NULL for the CONF pointer in CONF_get_string and
6307 CONF_get_number (which may use environment variables) and directly
6308 return NULL from CONF_get_section.
6311 *) Fix potential buffer overrun for EBCDIC.
6314 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6315 keyUsage if basicConstraints absent for a CA.
6318 *) Make SMIME_write_PKCS7() write mail header values with a format that
6319 is more generally accepted (no spaces before the semicolon), since
6320 some programs can't parse those values properly otherwise. Also make
6321 sure BIO's that break lines after each write do not create invalid
6325 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6326 macros previously used would not encode an empty SEQUENCE OF
6327 and break the signature.
6329 [This change does not apply to 0.9.7.]
6331 *) Zero the premaster secret after deriving the master secret in
6335 *) Add some EVP_add_digest_alias registrations (as found in
6336 OpenSSL_add_all_digests()) to SSL_library_init()
6337 aka OpenSSL_add_ssl_algorithms(). This provides improved
6338 compatibility with peers using X.509 certificates
6339 with unconventional AlgorithmIdentifier OIDs.
6342 *) Fix for Irix with NO_ASM.
6343 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6345 *) ./config script fixes.
6346 [Ulf Moeller, Richard Levitte]
6348 *) Fix 'openssl passwd -1'.
6351 *) Change PKCS12_key_gen_asc() so it can cope with non null
6352 terminated strings whose length is passed in the passlen
6353 parameter, for example from PEM callbacks. This was done
6354 by adding an extra length parameter to asc2uni().
6355 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6357 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6358 call failed, free the DSA structure.
6361 *) Fix to uni2asc() to cope with zero length Unicode strings.
6362 These are present in some PKCS#12 files.
6365 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6366 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6367 when writing a 32767 byte record.
6368 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6370 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6371 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6373 (RSA objects have a reference count access to which is protected
6374 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6375 so they are meant to be shared between threads.)
6376 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6377 "Reddie, Steven" <Steven.Reddie@ca.com>]
6379 *) Fix a deadlock in CRYPTO_mem_leaks().
6382 *) Use better test patterns in bntest.
6385 *) rand_win.c fix for Borland C.
6388 *) BN_rshift bugfix for n == 0.
6391 *) Add a 'bctest' script that checks for some known 'bc' bugs
6392 so that 'make test' does not abort just because 'bc' is broken.
6395 *) Store verify_result within SSL_SESSION also for client side to
6396 avoid potential security hole. (Re-used sessions on the client side
6397 always resulted in verify_result==X509_V_OK, not using the original
6398 result of the server certificate verification.)
6401 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6402 SSL3_RT_APPLICATION_DATA, return 0.
6403 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6407 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6408 releases, have been re-implemented by renaming the previous
6409 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6410 and ssl3_read_internal, respectively, and adding 'peek' parameters
6411 to them. The new ssl[23]_{read,peek} functions are calls to
6412 ssl[23]_read_internal with the 'peek' flag set appropriately.
6413 A 'peek' parameter has also been added to ssl3_read_bytes, which
6414 does the actual work for ssl3_read_internal.
6417 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6418 the method-specific "init()" handler. Also clean up ex_data after
6419 calling the method-specific "finish()" handler. Previously, this was
6420 happening the other way round.
6423 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6424 The previous value, 12, was not always sufficient for BN_mod_exp().
6427 *) Make sure that shared libraries get the internal name engine with
6428 the full version number and not just 0. This should mark the
6429 shared libraries as not backward compatible. Of course, this should
6430 be changed again when we can guarantee backward binary compatibility.
6433 *) Fix typo in get_cert_by_subject() in by_dir.c
6434 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6436 *) Rework the system to generate shared libraries:
6438 - Make note of the expected extension for the shared libraries and
6439 if there is a need for symbolic links from for example libcrypto.so.0
6440 to libcrypto.so.0.9.7. There is extended info in Configure for
6443 - Make as few rebuilds of the shared libraries as possible.
6445 - Still avoid linking the OpenSSL programs with the shared libraries.
6447 - When installing, install the shared libraries separately from the
6451 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6453 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6454 and not in SSL_clear because the latter is also used by the
6455 accept/connect functions; previously, the settings made by
6456 SSL_set_read_ahead would be lost during the handshake.
6457 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6459 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6460 Previously, it would create entries for disableed algorithms no
6464 *) Added several new manual pages for SSL_* function.
6467 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6469 *) In ssl23_get_client_hello, generate an error message when faced
6470 with an initial SSL 3.0/TLS record that is too small to contain the
6471 first two bytes of the ClientHello message, i.e. client_version.
6472 (Note that this is a pathologic case that probably has never happened
6473 in real life.) The previous approach was to use the version number
6474 from the record header as a substitute; but our protocol choice
6475 should not depend on that one because it is not authenticated
6476 by the Finished messages.
6479 *) More robust randomness gathering functions for Windows.
6480 [Jeffrey Altman <jaltman@columbia.edu>]
6482 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6483 not set then we don't setup the error code for issuer check errors
6484 to avoid possibly overwriting other errors which the callback does
6485 handle. If an application does set the flag then we assume it knows
6486 what it is doing and can handle the new informational codes
6490 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6491 a general "ANY" type, as such it should be able to decode anything
6492 including tagged types. However it didn't check the class so it would
6493 wrongly interpret tagged types in the same way as their universal
6494 counterpart and unknown types were just rejected. Changed so that the
6495 tagged and unknown types are handled in the same way as a SEQUENCE:
6496 that is the encoding is stored intact. There is also a new type
6497 "V_ASN1_OTHER" which is used when the class is not universal, in this
6498 case we have no idea what the actual type is so we just lump them all
6502 *) On VMS, stdout may very well lead to a file that is written to
6503 in a record-oriented fashion. That means that every write() will
6504 write a separate record, which will be read separately by the
6505 programs trying to read from it. This can be very confusing.
6507 The solution is to put a BIO filter in the way that will buffer
6508 text until a linefeed is reached, and then write everything a
6509 line at a time, so every record written will be an actual line,
6510 not chunks of lines and not (usually doesn't happen, but I've
6511 seen it once) several lines in one record. BIO_f_linebuffer() is
6514 Currently, it's a VMS-only method, because that's where it has
6515 been tested well enough.
6518 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6519 it can return incorrect results.
6520 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6521 but it was in 0.9.6-beta[12].)
6524 *) Disable the check for content being present when verifying detached
6525 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6526 include zero length content when signing messages.
6529 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6530 BIO_ctrl (for BIO pairs).
6533 *) Add DSO method for VMS.
6536 *) Bug fix: Montgomery multiplication could produce results with the
6540 *) Add RPM specification openssl.spec and modify it to build three
6541 packages. The default package contains applications, application
6542 documentation and run-time libraries. The devel package contains
6543 include files, static libraries and function documentation. The
6544 doc package contains the contents of the doc directory. The original
6545 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6548 *) Add a large number of documentation files for many SSL routines.
6549 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6551 *) Add a configuration entry for Sony News 4.
6552 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6554 *) Don't set the two most significant bits to one when generating a
6555 random number < q in the DSA library.
6558 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6559 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6560 the underlying transport is blocking) if a handshake took place.
6561 (The default behaviour is needed by applications such as s_client
6562 and s_server that use select() to determine when to use SSL_read;
6563 but for applications that know in advance when to expect data, it
6564 just makes things more complicated.)
6567 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6571 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6572 work better on such systems.
6573 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6575 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6576 Update PKCS12_parse() so it copies the friendlyName and the
6577 keyid to the certificates aux info.
6580 *) Fix bug in PKCS7_verify() which caused an infinite loop
6581 if there was more than one signature.
6582 [Sven Uszpelkat <su@celocom.de>]
6584 *) Major change in util/mkdef.pl to include extra information
6585 about each symbol, as well as presentig variables as well
6586 as functions. This change means that there's n more need
6587 to rebuild the .num files when some algorithms are excluded.
6590 *) Allow the verify time to be set by an application,
6591 rather than always using the current time.
6594 *) Phase 2 verify code reorganisation. The certificate
6595 verify code now looks up an issuer certificate by a
6596 number of criteria: subject name, authority key id
6597 and key usage. It also verifies self signed certificates
6598 by the same criteria. The main comparison function is
6599 X509_check_issued() which performs these checks.
6601 Lot of changes were necessary in order to support this
6602 without completely rewriting the lookup code.
6604 Authority and subject key identifier are now cached.
6606 The LHASH 'certs' is X509_STORE has now been replaced
6607 by a STACK_OF(X509_OBJECT). This is mainly because an
6608 LHASH can't store or retrieve multiple objects with
6609 the same hash value.
6611 As a result various functions (which were all internal
6612 use only) have changed to handle the new X509_STORE
6613 structure. This will break anything that messed round
6614 with X509_STORE internally.
6616 The functions X509_STORE_add_cert() now checks for an
6617 exact match, rather than just subject name.
6619 The X509_STORE API doesn't directly support the retrieval
6620 of multiple certificates matching a given criteria, however
6621 this can be worked round by performing a lookup first
6622 (which will fill the cache with candidate certificates)
6623 and then examining the cache for matches. This is probably
6624 the best we can do without throwing out X509_LOOKUP
6625 entirely (maybe later...).
6627 The X509_VERIFY_CTX structure has been enhanced considerably.
6629 All certificate lookup operations now go via a get_issuer()
6630 callback. Although this currently uses an X509_STORE it
6631 can be replaced by custom lookups. This is a simple way
6632 to bypass the X509_STORE hackery necessary to make this
6633 work and makes it possible to use more efficient techniques
6634 in future. A very simple version which uses a simple
6635 STACK for its trusted certificate store is also provided
6636 using X509_STORE_CTX_trusted_stack().
6638 The verify_cb() and verify() callbacks now have equivalents
6639 in the X509_STORE_CTX structure.
6641 X509_STORE_CTX also has a 'flags' field which can be used
6642 to customise the verify behaviour.
6645 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6646 excludes S/MIME capabilities.
6649 *) When a certificate request is read in keep a copy of the
6650 original encoding of the signed data and use it when outputing
6651 again. Signatures then use the original encoding rather than
6652 a decoded, encoded version which may cause problems if the
6653 request is improperly encoded.
6656 *) For consistency with other BIO_puts implementations, call
6657 buffer_write(b, ...) directly in buffer_puts instead of calling
6660 In BIO_puts, increment b->num_write as in BIO_write.
6661 [Peter.Sylvester@EdelWeb.fr]
6663 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6664 BN_zero, we may not return a BIGNUM with an array consisting of
6668 *) Avoid calling abort() from within the library when problems are
6669 detected, except if preprocessor symbols have been defined
6670 (such as REF_CHECK, BN_DEBUG etc.).
6673 *) New openssl application 'rsautl'. This utility can be
6674 used for low level RSA operations. DER public key
6675 BIO/fp routines also added.
6678 *) New Configure entry and patches for compiling on QNX 4.
6679 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6681 *) A demo state-machine implementation was sponsored by
6682 Nuron (http://www.nuron.com/) and is now available in
6683 demos/state_machine.
6686 *) New options added to the 'dgst' utility for signature
6687 generation and verification.
6690 *) Unrecognized PKCS#7 content types are now handled via a
6691 catch all ASN1_TYPE structure. This allows unsupported
6692 types to be stored as a "blob" and an application can
6693 encode and decode it manually.
6696 *) Fix various signed/unsigned issues to make a_strex.c
6698 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6700 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6701 length if passed a buffer. ASN1_INTEGER_to_BN failed
6702 if passed a NULL BN and its argument was negative.
6703 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6705 *) Modification to PKCS#7 encoding routines to output definite
6706 length encoding. Since currently the whole structures are in
6707 memory there's not real point in using indefinite length
6708 constructed encoding. However if OpenSSL is compiled with
6709 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6712 *) Added BIO_vprintf() and BIO_vsnprintf().
6715 *) Added more prefixes to parse for in the the strings written
6716 through a logging bio, to cover all the levels that are available
6717 through syslog. The prefixes are now:
6719 PANIC, EMERG, EMR => LOG_EMERG
6720 ALERT, ALR => LOG_ALERT
6721 CRIT, CRI => LOG_CRIT
6722 ERROR, ERR => LOG_ERR
6723 WARNING, WARN, WAR => LOG_WARNING
6724 NOTICE, NOTE, NOT => LOG_NOTICE
6725 INFO, INF => LOG_INFO
6726 DEBUG, DBG => LOG_DEBUG
6728 and as before, if none of those prefixes are present at the
6729 beginning of the string, LOG_ERR is chosen.
6731 On Win32, the LOG_* levels are mapped according to this:
6733 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6734 LOG_WARNING => EVENTLOG_WARNING_TYPE
6735 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6739 *) Made it possible to reconfigure with just the configuration
6740 argument "reconf" or "reconfigure". The command line arguments
6741 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6742 and are retrieved from there when reconfiguring.
6746 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6748 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6751 *) The obj_dat.pl script was messing up the sorting of object
6752 names. The reason was that it compared the quoted version
6753 of strings as a result "OCSP" > "OCSP Signing" because
6754 " > SPACE. Changed script to store unquoted versions of
6755 names and add quotes on output. It was also omitting some
6756 names from the lookup table if they were given a default
6757 value (that is if SN is missing it is given the same
6758 value as LN and vice versa), these are now added on the
6759 grounds that if an object has a name we should be able to
6760 look it up. Finally added warning output when duplicate
6761 short or long names are found.
6764 *) Changes needed for Tandem NSK.
6765 [Scott Uroff <scott@xypro.com>]
6767 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6768 RSA_padding_check_SSLv23(), special padding was never detected
6769 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6770 version rollback attacks was not effective.
6772 In s23_clnt.c, don't use special rollback-attack detection padding
6773 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6774 client; similarly, in s23_srvr.c, don't do the rollback check if
6775 SSL 2.0 is the only protocol enabled in the server.
6778 *) Make it possible to get hexdumps of unprintable data with 'openssl
6779 asn1parse'. By implication, the functions ASN1_parse_dump() and
6780 BIO_dump_indent() are added.
6783 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6784 these print out strings and name structures based on various
6785 flags including RFC2253 support and proper handling of
6786 multibyte characters. Added options to the 'x509' utility
6787 to allow the various flags to be set.
6790 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6791 Also change the functions X509_cmp_current_time() and
6792 X509_gmtime_adj() work with an ASN1_TIME structure,
6793 this will enable certificates using GeneralizedTime in validity
6794 dates to be checked.
6797 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6798 negative public key encodings) on by default,
6799 NO_NEG_PUBKEY_BUG can be set to disable it.
6802 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6803 content octets. An i2c_ASN1_OBJECT is unnecessary because
6804 the encoding can be trivially obtained from the structure.
6807 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6808 not read locks (CRYPTO_r_[un]lock).
6811 *) A first attempt at creating official support for shared
6812 libraries through configuration. I've kept it so the
6813 default is static libraries only, and the OpenSSL programs
6814 are always statically linked for now, but there are
6815 preparations for dynamic linking in place.
6816 This has been tested on Linux and Tru64.
6819 *) Randomness polling function for Win9x, as described in:
6820 Peter Gutmann, Software Generation of Practically Strong
6824 *) Fix so PRNG is seeded in req if using an already existing
6828 *) New options to smime application. -inform and -outform
6829 allow alternative formats for the S/MIME message including
6830 PEM and DER. The -content option allows the content to be
6831 specified separately. This should allow things like Netscape
6832 form signing output easier to verify.
6835 *) Fix the ASN1 encoding of tags using the 'long form'.
6838 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6839 STRING types. These convert content octets to and from the
6840 underlying type. The actual tag and length octets are
6841 already assumed to have been read in and checked. These
6842 are needed because all other string types have virtually
6843 identical handling apart from the tag. By having versions
6844 of the ASN1 functions that just operate on content octets
6845 IMPLICIT tagging can be handled properly. It also allows
6846 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6847 and ASN1_INTEGER are identical apart from the tag.
6850 *) Change the handling of OID objects as follows:
6852 - New object identifiers are inserted in objects.txt, following
6853 the syntax given in objects.README.
6854 - objects.pl is used to process obj_mac.num and create a new
6856 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6859 This is currently kind of a hack, and the perl code in objects.pl
6860 isn't very elegant, but it works as I intended. The simplest way
6861 to check that it worked correctly is to look in obj_dat.h and
6862 check the array nid_objs and make sure the objects haven't moved
6863 around (this is important!). Additions are OK, as well as
6864 consistent name changes.
6867 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6870 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6871 The given file adds to whatever has already been seeded into the
6872 random pool through the RANDFILE configuration file option or
6873 environment variable, or the default random state file.
6876 *) mkstack.pl now sorts each macro group into lexical order.
6877 Previously the output order depended on the order the files
6878 appeared in the directory, resulting in needless rewriting
6882 *) Patches to make OpenSSL compile under Win32 again. Mostly
6883 work arounds for the VC++ problem that it treats func() as
6884 func(void). Also stripped out the parts of mkdef.pl that
6885 added extra typesafe functions: these no longer exist.
6888 *) Reorganisation of the stack code. The macros are now all
6889 collected in safestack.h . Each macro is defined in terms of
6890 a "stack macro" of the form SKM_<name>(type, a, b). The
6891 DEBUG_SAFESTACK is now handled in terms of function casts,
6892 this has the advantage of retaining type safety without the
6893 use of additional functions. If DEBUG_SAFESTACK is not defined
6894 then the non typesafe macros are used instead. Also modified the
6895 mkstack.pl script to handle the new form. Needs testing to see
6896 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6897 the default if no major problems. Similar behaviour for ASN1_SET_OF
6898 and PKCS12_STACK_OF.
6901 *) When some versions of IIS use the 'NET' form of private key the
6902 key derivation algorithm is different. Normally MD5(password) is
6903 used as a 128 bit RC4 key. In the modified case
6904 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6905 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6906 as the old Netscape_RSA functions except they have an additional
6907 'sgckey' parameter which uses the modified algorithm. Also added
6908 an -sgckey command line option to the rsa utility. Thanks to
6909 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6910 algorithm to openssl-dev.
6913 *) The evp_local.h macros were using 'c.##kname' which resulted in
6914 invalid expansion on some systems (SCO 5.0.5 for example).
6915 Corrected to 'c.kname'.
6916 [Phillip Porch <root@theporch.com>]
6918 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6919 a STACK of email addresses from a certificate or request, these look
6920 in the subject name and the subject alternative name extensions and
6921 omit any duplicate addresses.
6924 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6925 This makes DSA verification about 2 % faster.
6928 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6929 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6930 plus overhead for 1024 bit moduli).
6931 This makes exponentiations about 0.5 % faster for 1024 bit
6932 exponents (as measured by "openssl speed rsa2048").
6935 *) Rename memory handling macros to avoid conflicts with other
6937 Malloc => OPENSSL_malloc
6938 Malloc_locked => OPENSSL_malloc_locked
6939 Realloc => OPENSSL_realloc
6940 Free => OPENSSL_free
6943 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6944 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6947 *) CygWin32 support.
6948 [John Jarvie <jjarvie@newsguy.com>]
6950 *) The type-safe stack code has been rejigged. It is now only compiled
6951 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6952 by default all type-specific stack functions are "#define"d back to
6953 standard stack functions. This results in more streamlined output
6954 but retains the type-safety checking possibilities of the original
6958 *) The STACK code has been cleaned up, and certain type declarations
6959 that didn't make a lot of sense have been brought in line. This has
6960 also involved a cleanup of sorts in safestack.h to more correctly
6961 map type-safe stack functions onto their plain stack counterparts.
6962 This work has also resulted in a variety of "const"ifications of
6963 lots of the code, especially "_cmp" operations which should normally
6964 be prototyped with "const" parameters anyway.
6967 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6968 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6969 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6970 where all of 'md' is used each time the PRNG is used, but 'state'
6971 is used only indexed by a cyclic counter. As entropy may not be
6972 well distributed from the beginning, 'md' is important as a
6973 chaining variable. However, the output function chains only half
6974 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6975 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6976 in all of 'state' being rewritten, with the new values depending
6977 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6980 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6981 the handshake is continued after ssl_verify_cert_chain();
6982 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6983 can lead to 'unexplainable' connection aborts later.
6984 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6986 *) Major EVP API cipher revision.
6987 Add hooks for extra EVP features. This allows various cipher
6988 parameters to be set in the EVP interface. Support added for variable
6989 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6990 setting of RC2 and RC5 parameters.
6992 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6995 Remove lots of duplicated code from the EVP library. For example *every*
6996 cipher init() function handles the 'iv' in the same way according to the
6997 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6998 for CFB and OFB modes they zero ctx->num.
7000 New functionality allows removal of S/MIME code RC2 hack.
7002 Most of the routines have the same form and so can be declared in terms
7005 By shifting this to the top level EVP_CipherInit() it can be removed from
7006 all individual ciphers. If the cipher wants to handle IVs or keys
7007 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7010 Change lots of functions like EVP_EncryptUpdate() to now return a
7011 value: although software versions of the algorithms cannot fail
7012 any installed hardware versions can.
7015 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7016 this option is set, tolerate broken clients that send the negotiated
7017 protocol version number instead of the requested protocol version
7021 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7022 i.e. non-zero for export ciphersuites, zero otherwise.
7023 Previous versions had this flag inverted, inconsistent with
7024 rsa_tmp_cb (..._TMP_RSA_CB).
7025 [Bodo Moeller; problem reported by Amit Chopra]
7027 *) Add missing DSA library text string. Work around for some IIS
7028 key files with invalid SEQUENCE encoding.
7031 *) Add a document (doc/standards.txt) that list all kinds of standards
7032 and so on that are implemented in OpenSSL.
7035 *) Enhance c_rehash script. Old version would mishandle certificates
7036 with the same subject name hash and wouldn't handle CRLs at all.
7037 Added -fingerprint option to crl utility, to support new c_rehash
7041 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7044 *) Fix for SSL server purpose checking. Server checking was
7045 rejecting certificates which had extended key usage present
7046 but no ssl client purpose.
7047 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7049 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7050 is a little unclear about how a blank password is handled.
7051 Since the password in encoded as a BMPString with terminating
7052 double NULL a zero length password would end up as just the
7053 double NULL. However no password at all is different and is
7054 handled differently in the PKCS#12 key generation code. NS
7055 treats a blank password as zero length. MSIE treats it as no
7056 password on export: but it will try both on import. We now do
7057 the same: PKCS12_parse() tries zero length and no password if
7058 the password is set to "" or NULL (NULL is now a valid password:
7059 it wasn't before) as does the pkcs12 application.
7062 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7063 perror when PEM_read_bio_X509_REQ fails, the error message must
7064 be obtained from the error queue.
7067 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7068 it in ERR_remove_state if appropriate, and change ERR_get_state
7069 accordingly to avoid race conditions (this is necessary because
7070 thread_hash is no longer constant once set).
7073 *) Bugfix for linux-elf makefile.one.
7076 *) RSA_get_default_method() will now cause a default
7077 RSA_METHOD to be chosen if one doesn't exist already.
7078 Previously this was only set during a call to RSA_new()
7079 or RSA_new_method(NULL) meaning it was possible for
7080 RSA_get_default_method() to return NULL.
7083 *) Added native name translation to the existing DSO code
7084 that will convert (if the flag to do so is set) filenames
7085 that are sufficiently small and have no path information
7086 into a canonical native form. Eg. "blah" converted to
7087 "libblah.so" or "blah.dll" etc.
7090 *) New function ERR_error_string_n(e, buf, len) which is like
7091 ERR_error_string(e, buf), but writes at most 'len' bytes
7092 including the 0 terminator. For ERR_error_string_n, 'buf'
7094 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7096 *) CONF library reworked to become more general. A new CONF
7097 configuration file reader "class" is implemented as well as a
7098 new functions (NCONF_*, for "New CONF") to handle it. The now
7099 old CONF_* functions are still there, but are reimplemented to
7100 work in terms of the new functions. Also, a set of functions
7101 to handle the internal storage of the configuration data is
7102 provided to make it easier to write new configuration file
7103 reader "classes" (I can definitely see something reading a
7104 configuration file in XML format, for example), called _CONF_*,
7105 or "the configuration storage API"...
7107 The new configuration file reading functions are:
7109 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7110 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7112 NCONF_default, NCONF_WIN32
7114 NCONF_dump_fp, NCONF_dump_bio
7116 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7117 NCONF_new creates a new CONF object. This works in the same way
7118 as other interfaces in OpenSSL, like the BIO interface.
7119 NCONF_dump_* dump the internal storage of the configuration file,
7120 which is useful for debugging. All other functions take the same
7121 arguments as the old CONF_* functions wth the exception of the
7122 first that must be a `CONF *' instead of a `LHASH *'.
7124 To make it easer to use the new classes with the old CONF_* functions,
7125 the function CONF_set_default_method is provided.
7128 *) Add '-tls1' option to 'openssl ciphers', which was already
7129 mentioned in the documentation but had not been implemented.
7130 (This option is not yet really useful because even the additional
7131 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7134 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7135 OpenSSL-based applications) load shared libraries and bind to
7136 them in a portable way.
7137 [Geoff Thorpe, with contributions from Richard Levitte]
7139 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7141 *) Make sure _lrotl and _lrotr are only used with MSVC.
7143 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7144 (the default implementation of RAND_status).
7146 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7147 to '-clrext' (= clear extensions), as intended and documented.
7148 [Bodo Moeller; inconsistency pointed out by Michael Attili
7151 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7152 was larger than the MD block size.
7153 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7155 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7156 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7157 using the passed key: if the passed key was a private key the result
7158 of X509_print(), for example, would be to print out all the private key
7162 *) des_quad_cksum() byte order bug fix.
7163 [Ulf Möller, using the problem description in krb4-0.9.7, where
7164 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7166 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7168 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7170 *) For easily testing in shell scripts whether some command
7171 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7172 returns with exit code 0 iff no command of the given name is available.
7173 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7174 the output goes to stdout and nothing is printed to stderr.
7175 Additional arguments are always ignored.
7177 Since for each cipher there is a command of the same name,
7178 the 'no-cipher' compilation switches can be tested this way.
7180 ('openssl no-XXX' is not able to detect pseudo-commands such
7181 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7184 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7187 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7188 is set; it will be thrown away anyway because each handshake creates
7190 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7191 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7192 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7193 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7196 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7197 'Q' and 'R' lose their special meanings (quit/renegotiate).
7198 This is part of what -quiet does; unlike -quiet, -ign_eof
7199 does not suppress any output.
7202 *) Add compatibility options to the purpose and trust code. The
7203 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7204 accepts a certificate or CA, this was the previous behaviour,
7205 with all the associated security issues.
7207 X509_TRUST_COMPAT is the old trust behaviour: only and
7208 automatically trust self signed roots in certificate store. A
7209 new trust setting X509_TRUST_DEFAULT is used to specify that
7210 a purpose has no associated trust setting and it should instead
7211 use the value in the default purpose.
7214 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7215 and fix a memory leak.
7218 *) In util/mkerr.pl (which implements 'make errors'), preserve
7219 reason strings from the previous version of the .c file, as
7220 the default to have only downcase letters (and digits) in
7221 automatically generated reasons codes is not always appropriate.
7224 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7225 using strerror. Previously, ERR_reason_error_string() returned
7226 library names as reason strings for SYSerr; but SYSerr is a special
7227 case where small numbers are errno values, not library numbers.
7230 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7231 converts DSA parameters into DH parameters. (When creating parameters,
7232 DSA_generate_parameters is used.)
7235 *) Include 'length' (recommended exponent length) in C code generated
7236 by 'openssl dhparam -C'.
7239 *) The second argument to set_label in perlasm was already being used
7240 so couldn't be used as a "file scope" flag. Moved to third argument
7244 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7245 instead of RAND_bytes for encryption IVs and salts.
7248 *) Include RAND_status() into RAND_METHOD instead of implementing
7249 it only for md_rand.c Otherwise replacing the PRNG by calling
7250 RAND_set_rand_method would be impossible.
7253 *) Don't let DSA_generate_key() enter an infinite loop if the random
7254 number generation fails.
7257 *) New 'rand' application for creating pseudo-random output.
7260 *) Added configuration support for Linux/IA64
7261 [Rolf Haberrecker <rolf@suse.de>]
7263 *) Assembler module support for Mingw32.
7266 *) Shared library support for HPUX (in shlib/).
7267 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7269 *) Shared library support for Solaris gcc.
7270 [Lutz Behnke <behnke@trustcenter.de>]
7272 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7274 *) PKCS7_encrypt() was adding text MIME headers twice because they
7275 were added manually and by SMIME_crlf_copy().
7278 *) In bntest.c don't call BN_rand with zero bits argument.
7279 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7281 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7282 case was implemented. This caused BN_div_recp() to fail occasionally.
7285 *) Add an optional second argument to the set_label() in the perl
7286 assembly language builder. If this argument exists and is set
7287 to 1 it signals that the assembler should use a symbol whose
7288 scope is the entire file, not just the current function. This
7289 is needed with MASM which uses the format label:: for this scope.
7290 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7292 *) Change the ASN1 types so they are typedefs by default. Before
7293 almost all types were #define'd to ASN1_STRING which was causing
7294 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7298 *) Change names of new functions to the new get1/get0 naming
7299 convention: After 'get1', the caller owns a reference count
7300 and has to call ..._free; 'get0' returns a pointer to some
7301 data structure without incrementing reference counters.
7302 (Some of the existing 'get' functions increment a reference
7303 counter, some don't.)
7304 Similarly, 'set1' and 'add1' functions increase reference
7305 counters or duplicate objects.
7308 *) Allow for the possibility of temp RSA key generation failure:
7309 the code used to assume it always worked and crashed on failure.
7312 *) Fix potential buffer overrun problem in BIO_printf().
7313 [Ulf Möller, using public domain code by Patrick Powell; problem
7314 pointed out by David Sacerdote <das33@cornell.edu>]
7316 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7317 RAND_egd() and RAND_status(). In the command line application,
7318 the EGD socket can be specified like a seed file using RANDFILE
7322 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7323 Some CAs (e.g. Verisign) distribute certificates in this form.
7326 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7327 list to exclude them. This means that no special compilation option
7328 is needed to use anonymous DH: it just needs to be included in the
7332 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7333 EVP_MD_type. The old functionality is available in a new macro called
7334 EVP_MD_md(). Change code that uses it and update docs.
7337 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7338 where the 'void *' argument is replaced by a function pointer argument.
7339 Previously 'void *' was abused to point to functions, which works on
7340 many platforms, but is not correct. As these functions are usually
7341 called by macros defined in OpenSSL header files, most source code
7342 should work without changes.
7345 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7346 sections with information on -D... compiler switches used for
7347 compiling the library so that applications can see them. To enable
7348 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7349 must be defined. E.g.,
7350 #define OPENSSL_ALGORITHM_DEFINES
7351 #include <openssl/opensslconf.h>
7352 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7353 [Richard Levitte, Ulf and Bodo Möller]
7355 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7359 *) Change the 'other' type in certificate aux info to a STACK_OF
7360 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7361 the required ASN1 format: arbitrary types determined by an OID.
7364 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7365 argument to 'req'. This is not because the function is newer or
7366 better than others it just uses the work 'NEW' in the certificate
7367 request header lines. Some software needs this.
7370 *) Reorganise password command line arguments: now passwords can be
7371 obtained from various sources. Delete the PEM_cb function and make
7372 it the default behaviour: i.e. if the callback is NULL and the
7373 usrdata argument is not NULL interpret it as a null terminated pass
7374 phrase. If usrdata and the callback are NULL then the pass phrase
7375 is prompted for as usual.
7378 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7379 the support is automatically enabled. The resulting binaries will
7380 autodetect the card and use it if present.
7381 [Ben Laurie and Compaq Inc.]
7383 *) Work around for Netscape hang bug. This sends certificate request
7384 and server done in one record. Since this is perfectly legal in the
7385 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7386 the bugs/SSLv3 entry for more info.
7389 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7392 *) Add -rand argument to smime and pkcs12 applications and read/write
7396 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7399 *) Add command line password options to the remaining applications.
7402 *) Bug fix for BN_div_recp() for numerators with an even number of
7406 *) More tests in bntest.c, and changed test_bn output.
7409 *) ./config recognizes MacOS X now.
7412 *) Bug fix for BN_div() when the first words of num and divsor are
7413 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7416 *) Add support for various broken PKCS#8 formats, and command line
7417 options to produce them.
7420 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7421 get temporary BIGNUMs from a BN_CTX.
7424 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7428 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7429 include a #define from the old name to the new. The original intent
7430 was that statically linked binaries could for example just call
7431 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7432 link with digests. This never worked becayse SSLeay_add_all_digests()
7433 and SSLeay_add_all_ciphers() were in the same source file so calling
7434 one would link with the other. They are now in separate source files.
7437 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7440 *) Use a less unusual form of the Miller-Rabin primality test (it used
7441 a binary algorithm for exponentiation integrated into the Miller-Rabin
7442 loop, our standard modexp algorithms are faster).
7445 *) Support for the EBCDIC character set completed.
7446 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7448 *) Source code cleanups: use const where appropriate, eliminate casts,
7449 use void * instead of char * in lhash.
7452 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7453 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7454 this the server could overwrite ephemeral keys that the client
7458 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7459 using 50 iterations of the Rabin-Miller test.
7461 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7462 iterations of the Rabin-Miller test as required by the appendix
7463 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7464 As BN_is_prime_fasttest includes trial division, DSA parameter
7465 generation becomes much faster.
7467 This implies a change for the callback functions in DSA_is_prime
7468 and DSA_generate_parameters: The callback function is called once
7469 for each positive witness in the Rabin-Miller test, not just
7470 occasionally in the inner loop; and the parameters to the
7471 callback function now provide an iteration count for the outer
7472 loop rather than for the current invocation of the inner loop.
7473 DSA_generate_parameters additionally can call the callback
7474 function with an 'iteration count' of -1, meaning that a
7475 candidate has passed the trial division test (when q is generated
7476 from an application-provided seed, trial division is skipped).
7479 *) New function BN_is_prime_fasttest that optionally does trial
7480 division before starting the Rabin-Miller test and has
7481 an additional BN_CTX * argument (whereas BN_is_prime always
7482 has to allocate at least one BN_CTX).
7483 'callback(1, -1, cb_arg)' is called when a number has passed the
7484 trial division stage.
7487 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7491 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7494 *) New function BN_pseudo_rand().
7497 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7498 bignum version of BN_from_montgomery() with the working code from
7499 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7503 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7504 made it impossible to use the same SSL_SESSION data structure in
7505 SSL2 clients in multiple threads.
7508 *) The return value of RAND_load_file() no longer counts bytes obtained
7509 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7510 to seed the PRNG (previously an explicit byte count was required).
7511 [Ulf Möller, Bodo Möller]
7513 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7514 used (char *) instead of (void *) and had casts all over the place.
7517 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7520 *) Retain source code compatibility for BN_prime_checks macro:
7521 BN_is_prime(..., BN_prime_checks, ...) now uses
7522 BN_prime_checks_for_size to determine the appropriate number of
7523 Rabin-Miller iterations.
7526 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7527 DH_CHECK_P_NOT_SAFE_PRIME.
7528 (Check if this is true? OpenPGP calls them "strong".)
7531 *) Merge the functionality of "dh" and "gendh" programs into a new program
7532 "dhparam". The old programs are retained for now but will handle DH keys
7533 (instead of parameters) in future.
7536 *) Make the ciphers, s_server and s_client programs check the return values
7537 when a new cipher list is set.
7540 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7541 ciphers. Before when the 56bit ciphers were enabled the sorting was
7544 The syntax for the cipher sorting has been extended to support sorting by
7545 cipher-strength (using the strength_bits hard coded in the tables).
7546 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7548 Fix a bug in the cipher-command parser: when supplying a cipher command
7549 string with an "undefined" symbol (neither command nor alphanumeric
7550 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7551 an error is flagged.
7553 Due to the strength-sorting extension, the code of the
7554 ssl_create_cipher_list() function was completely rearranged. I hope that
7555 the readability was also increased :-)
7556 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7558 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7559 for the first serial number and places 2 in the serial number file. This
7560 avoids problems when the root CA is created with serial number zero and
7561 the first user certificate has the same issuer name and serial number
7565 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7566 the new code. Add documentation for this stuff.
7569 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7570 X509_*() to X509at_*() on the grounds that they don't handle X509
7571 structures and behave in an analagous way to the X509v3 functions:
7572 they shouldn't be called directly but wrapper functions should be used
7575 So we also now have some wrapper functions that call the X509at functions
7576 when passed certificate requests. (TO DO: similar things can be done with
7577 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7578 things. Some of these need some d2i or i2d and print functionality
7579 because they handle more complex structures.)
7582 *) Add missing #ifndefs that caused missing symbols when building libssl
7583 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7584 NO_RSA in ssl/s2*.c.
7585 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7587 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7588 has a return value which indicates the quality of the random data
7589 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7590 error queue. New function RAND_pseudo_bytes() generates output that is
7591 guaranteed to be unique but not unpredictable. RAND_add is like
7592 RAND_seed, but takes an extra argument for an entropy estimate
7593 (RAND_seed always assumes full entropy).
7596 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7597 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7598 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7599 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7600 false-positive rate of at most 2^-80 for random input.
7603 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7606 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7607 in the 0.9.5 release), this returns the chain
7608 from an X509_CTX structure with a dup of the stack and all
7609 the X509 reference counts upped: so the stack will exist
7610 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7613 Also make SSL_SESSION_print() print out the verify return
7617 *) Add manpage for the pkcs12 command. Also change the default
7618 behaviour so MAC iteration counts are used unless the new
7619 -nomaciter option is used. This improves file security and
7620 only older versions of MSIE (4.0 for example) need it.
7623 *) Honor the no-xxx Configure options when creating .DEF files.
7626 *) Add PKCS#10 attributes to field table: challengePassword,
7627 unstructuredName and unstructuredAddress. These are taken from
7628 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7629 international characters are used.
7631 More changes to X509_ATTRIBUTE code: allow the setting of types
7632 based on strings. Remove the 'loc' parameter when adding
7633 attributes because these will be a SET OF encoding which is sorted
7637 *) Initial changes to the 'req' utility to allow request generation
7638 automation. This will allow an application to just generate a template
7639 file containing all the field values and have req construct the
7642 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7643 used all over the place including certificate requests and PKCS#7
7644 structures. They are currently handled manually where necessary with
7645 some primitive wrappers for PKCS#7. The new functions behave in a
7646 manner analogous to the X509 extension functions: they allow
7647 attributes to be looked up by NID and added.
7649 Later something similar to the X509V3 code would be desirable to
7650 automatically handle the encoding, decoding and printing of the
7651 more complex types. The string types like challengePassword can
7652 be handled by the string table functions.
7654 Also modified the multi byte string table handling. Now there is
7655 a 'global mask' which masks out certain types. The table itself
7656 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7657 is useful when for example there is only one permissible type
7658 (as in countryName) and using the mask might result in no valid
7662 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7663 SSL_get_peer_finished to allow applications to obtain the latest
7664 Finished messages sent to the peer or expected from the peer,
7665 respectively. (SSL_get_peer_finished is usually the Finished message
7666 actually received from the peer, otherwise the protocol will be aborted.)
7668 As the Finished message are message digests of the complete handshake
7669 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7670 be used for external authentication procedures when the authentication
7671 provided by SSL/TLS is not desired or is not enough.
7674 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7675 the host supports BWX extension and if Compaq C is present on the
7676 $PATH. Just exploiting of the BWX extension results in 20-30%
7677 performance kick for some algorithms, e.g. DES and RC4 to mention
7678 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7682 *) Add support for MS "fast SGC". This is arguably a violation of the
7683 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7684 weak crypto and after checking the certificate is SGC a second one
7685 with strong crypto. MS SGC stops the first handshake after receiving
7686 the server certificate message and sends a second client hello. Since
7687 a server will typically do all the time consuming operations before
7688 expecting any further messages from the client (server key exchange
7689 is the most expensive) there is little difference between the two.
7691 To get OpenSSL to support MS SGC we have to permit a second client
7692 hello message after we have sent server done. In addition we have to
7693 reset the MAC if we do get this second client hello.
7696 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7697 if a DER encoded private key is RSA or DSA traditional format. Changed
7698 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7699 format DER encoded private key. Newer code should use PKCS#8 format which
7700 has the key type encoded in the ASN1 structure. Added DER private key
7701 support to pkcs8 application.
7704 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7705 ciphersuites has been selected (as required by the SSL 3/TLS 1
7706 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7707 is set, we interpret this as a request to violate the specification
7708 (the worst that can happen is a handshake failure, and 'correct'
7709 behaviour would result in a handshake failure anyway).
7712 *) In SSL_CTX_add_session, take into account that there might be multiple
7713 SSL_SESSION structures with the same session ID (e.g. when two threads
7714 concurrently obtain them from an external cache).
7715 The internal cache can handle only one SSL_SESSION with a given ID,
7716 so if there's a conflict, we now throw out the old one to achieve
7720 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7721 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7722 some routines that use cipher OIDs: some ciphers do not have OIDs
7723 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7727 *) Simplify the trust setting structure and code. Now we just have
7728 two sequences of OIDs for trusted and rejected settings. These will
7729 typically have values the same as the extended key usage extension
7730 and any application specific purposes.
7732 The trust checking code now has a default behaviour: it will just
7733 check for an object with the same NID as the passed id. Functions can
7734 be provided to override either the default behaviour or the behaviour
7735 for a given id. SSL client, server and email already have functions
7736 in place for compatibility: they check the NID and also return "trusted"
7737 if the certificate is self signed.
7740 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7741 traditional format into an EVP_PKEY structure.
7744 *) Add a password callback function PEM_cb() which either prompts for
7745 a password if usr_data is NULL or otherwise assumes it is a null
7746 terminated password. Allow passwords to be passed on command line
7747 environment or config files in a few more utilities.
7750 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7751 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7752 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7753 Update documentation.
7756 *) Support for ASN1 "NULL" type. This could be handled before by using
7757 ASN1_TYPE but there wasn't any function that would try to read a NULL
7758 and produce an error if it couldn't. For compatibility we also have
7759 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7760 don't allocate anything because they don't need to.
7763 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7765 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7767 *) Rebuild of the memory allocation routines used by OpenSSL code and
7768 possibly others as well. The purpose is to make an interface that
7769 provide hooks so anyone can build a separate set of allocation and
7770 deallocation routines to be used by OpenSSL, for example memory
7771 pool implementations, or something else, which was previously hard
7772 since Malloc(), Realloc() and Free() were defined as macros having
7773 the values malloc, realloc and free, respectively (except for Win32
7774 compilations). The same is provided for memory debugging code.
7775 OpenSSL already comes with functionality to find memory leaks, but
7776 this gives people a chance to debug other memory problems.
7778 With these changes, a new set of functions and macros have appeared:
7780 CRYPTO_set_mem_debug_functions() [F]
7781 CRYPTO_get_mem_debug_functions() [F]
7782 CRYPTO_dbg_set_options() [F]
7783 CRYPTO_dbg_get_options() [F]
7784 CRYPTO_malloc_debug_init() [M]
7786 The memory debug functions are NULL by default, unless the library
7787 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7788 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7789 gives the standard debugging functions that come with OpenSSL) or
7790 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7791 provided by the library user) must be used. When the standard
7792 debugging functions are used, CRYPTO_dbg_set_options can be used to
7793 request additional information:
7794 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7795 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7797 Also, things like CRYPTO_set_mem_functions will always give the
7798 expected result (the new set of functions is used for allocation
7799 and deallocation) at all times, regardless of platform and compiler
7802 To finish it up, some functions that were never use in any other
7803 way than through macros have a new API and new semantic:
7806 CRYPTO_dbg_realloc()
7809 All macros of value have retained their old syntax.
7810 [Richard Levitte and Bodo Moeller]
7812 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7813 ordering of SMIMECapabilities wasn't in "strength order" and there
7814 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7818 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7819 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7820 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7822 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7823 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7824 functionality to handle multipart/signed properly) and a utility
7825 called 'smime' to call all this stuff. This is based on code I
7826 originally wrote for Celo who have kindly allowed it to be
7827 included in OpenSSL.
7830 *) Add variants des_set_key_checked and des_set_key_unchecked of
7831 des_set_key (aka des_key_sched). Global variable des_check_key
7832 decides which of these is called by des_set_key; this way
7833 des_check_key behaves as it always did, but applications and
7834 the library itself, which was buggy for des_check_key == 1,
7835 have a cleaner way to pick the version they need.
7838 *) New function PKCS12_newpass() which changes the password of a
7842 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7843 dynamic mix. In both cases the ids can be used as an index into the
7844 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7845 functions so they accept a list of the field values and the
7846 application doesn't need to directly manipulate the X509_TRUST
7850 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7854 *) Modify the way the V3 extension code looks up extensions. This now
7855 works in a similar way to the object code: we have some "standard"
7856 extensions in a static table which is searched with OBJ_bsearch()
7857 and the application can add dynamic ones if needed. The file
7858 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7859 updated whenever a new extension is added to the core code and kept
7860 in ext_nid order. There is a simple program 'tabtest.c' which checks
7861 this. New extensions are not added too often so this file can readily
7862 be maintained manually.
7864 There are two big advantages in doing things this way. The extensions
7865 can be looked up immediately and no longer need to be "added" using
7866 X509V3_add_standard_extensions(): this function now does nothing.
7867 [Side note: I get *lots* of email saying the extension code doesn't
7868 work because people forget to call this function]
7869 Also no dynamic allocation is done unless new extensions are added:
7870 so if we don't add custom extensions there is no need to call
7871 X509V3_EXT_cleanup().
7874 *) Modify enc utility's salting as follows: make salting the default. Add a
7875 magic header, so unsalted files fail gracefully instead of just decrypting
7876 to garbage. This is because not salting is a big security hole, so people
7877 should be discouraged from doing it.
7880 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7881 digest to be passed on the command line but it only used this
7882 parameter when signing a certificate. Modified so all relevant
7883 operations are affected by the digest parameter including the
7884 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7885 DSA key was used because it didn't fix the digest.
7888 *) Initial certificate chain verify code. Currently tests the untrusted
7889 certificates for consistency with the verify purpose (which is set
7890 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7892 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7893 this is because it will reject chains with invalid extensions whereas
7894 every previous version of OpenSSL and SSLeay made no checks at all.
7896 Trust code: checks the root CA for the relevant trust settings. Trust
7897 settings have an initial value consistent with the verify purpose: e.g.
7898 if the verify purpose is for SSL client use it expects the CA to be
7899 trusted for SSL client use. However the default value can be changed to
7900 permit custom trust settings: one example of this would be to only trust
7901 certificates from a specific "secure" set of CAs.
7903 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7904 which should be used for version portability: especially since the
7905 verify structure is likely to change more often now.
7907 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7908 to set them. If not set then assume SSL clients will verify SSL servers
7911 Two new options to the verify program: -untrusted allows a set of
7912 untrusted certificates to be passed in and -purpose which sets the
7913 intended purpose of the certificate. If a purpose is set then the
7914 new chain verify code is used to check extension consistency.
7917 *) Support for the authority information access extension.
7920 *) Modify RSA and DSA PEM read routines to transparently handle
7921 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7922 public keys in a format compatible with certificate
7923 SubjectPublicKeyInfo structures. Unfortunately there were already
7924 functions called *_PublicKey_* which used various odd formats so
7925 these are retained for compatibility: however the DSA variants were
7926 never in a public release so they have been deleted. Changed dsa/rsa
7927 utilities to handle the new format: note no releases ever handled public
7928 keys so we should be OK.
7930 The primary motivation for this change is to avoid the same fiasco
7931 that dogs private keys: there are several incompatible private key
7932 formats some of which are standard and some OpenSSL specific and
7933 require various evil hacks to allow partial transparent handling and
7934 even then it doesn't work with DER formats. Given the option anything
7935 other than PKCS#8 should be dumped: but the other formats have to
7936 stay in the name of compatibility.
7938 With public keys and the benefit of hindsight one standard format
7939 is used which works with EVP_PKEY, RSA or DSA structures: though
7940 it clearly returns an error if you try to read the wrong kind of key.
7942 Added a -pubkey option to the 'x509' utility to output the public key.
7943 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7944 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7945 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7946 that do the same as the EVP_PKEY_assign_*() except they up the
7947 reference count of the added key (they don't "swallow" the
7951 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7952 CRLs would fail if the file contained no certificates or no CRLs:
7953 added a new function to read in both types and return the number
7954 read: this means that if none are read it will be an error. The
7955 DER versions of the certificate and CRL reader would always fail
7956 because it isn't possible to mix certificates and CRLs in DER format
7957 without choking one or the other routine. Changed this to just read
7958 a certificate: this is the best we can do. Also modified the code
7959 in apps/verify.c to take notice of return codes: it was previously
7960 attempting to read in certificates from NULL pointers and ignoring
7961 any errors: this is one reason why the cert and CRL reader seemed
7962 to work. It doesn't check return codes from the default certificate
7963 routines: these may well fail if the certificates aren't installed.
7966 *) Code to support otherName option in GeneralName.
7969 *) First update to verify code. Change the verify utility
7970 so it warns if it is passed a self signed certificate:
7971 for consistency with the normal behaviour. X509_verify
7972 has been modified to it will now verify a self signed
7973 certificate if *exactly* the same certificate appears
7974 in the store: it was previously impossible to trust a
7975 single self signed certificate. This means that:
7976 openssl verify ss.pem
7977 now gives a warning about a self signed certificate but
7978 openssl verify -CAfile ss.pem ss.pem
7982 *) For servers, store verify_result in SSL_SESSION data structure
7983 (and add it to external session representation).
7984 This is needed when client certificate verifications fails,
7985 but an application-provided verification callback (set by
7986 SSL_CTX_set_cert_verify_callback) allows accepting the session
7987 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7988 but returns 1): When the session is reused, we have to set
7989 ssl->verify_result to the appropriate error code to avoid
7991 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7993 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7994 case in PKCS7_dataInit() where the signed PKCS7 structure
7995 didn't contain any existing data because it was being created.
7996 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7998 *) Add a salt to the key derivation routines in enc.c. This
7999 forms the first 8 bytes of the encrypted file. Also add a
8000 -S option to allow a salt to be input on the command line.
8003 *) New function X509_cmp(). Oddly enough there wasn't a function
8004 to compare two certificates. We do this by working out the SHA1
8005 hash and comparing that. X509_cmp() will be needed by the trust
8009 *) SSL_get1_session() is like SSL_get_session(), but increments
8010 the reference count in the SSL_SESSION returned.
8011 [Geoff Thorpe <geoff@eu.c2.net>]
8013 *) Fix for 'req': it was adding a null to request attributes.
8014 Also change the X509_LOOKUP and X509_INFO code to handle
8015 certificate auxiliary information.
8018 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8022 *) Add the possibility to add extra information to the memory leak
8023 detecting output, to form tracebacks, showing from where each
8024 allocation was originated: CRYPTO_push_info("constant string") adds
8025 the string plus current file name and line number to a per-thread
8026 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8027 is like calling CYRPTO_pop_info() until the stack is empty.
8028 Also updated memory leak detection code to be multi-thread-safe.
8031 *) Add options -text and -noout to pkcs7 utility and delete the
8032 encryption options which never did anything. Update docs.
8035 *) Add options to some of the utilities to allow the pass phrase
8036 to be included on either the command line (not recommended on
8037 OSes like Unix) or read from the environment. Update the
8038 manpages and fix a few bugs.
8041 *) Add a few manpages for some of the openssl commands.
8044 *) Fix the -revoke option in ca. It was freeing up memory twice,
8045 leaking and not finding already revoked certificates.
8048 *) Extensive changes to support certificate auxiliary information.
8049 This involves the use of X509_CERT_AUX structure and X509_AUX
8050 functions. An X509_AUX function such as PEM_read_X509_AUX()
8051 can still read in a certificate file in the usual way but it
8052 will also read in any additional "auxiliary information". By
8053 doing things this way a fair degree of compatibility can be
8054 retained: existing certificates can have this information added
8055 using the new 'x509' options.
8057 Current auxiliary information includes an "alias" and some trust
8058 settings. The trust settings will ultimately be used in enhanced
8059 certificate chain verification routines: currently a certificate
8060 can only be trusted if it is self signed and then it is trusted
8064 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8065 The problem was that one of the replacement routines had not been working
8066 since SSLeay releases. For now the offending routine has been replaced
8067 with non-optimised assembler. Even so, this now gives around 95%
8068 performance improvement for 1024 bit RSA signs.
8071 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8072 handling. Most clients have the effective key size in bits equal to
8073 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8074 A few however don't do this and instead use the size of the decrypted key
8075 to determine the RC2 key length and the AlgorithmIdentifier to determine
8076 the effective key length. In this case the effective key length can still
8077 be 40 bits but the key length can be 168 bits for example. This is fixed
8078 by manually forcing an RC2 key into the EVP_PKEY structure because the
8079 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8080 the key length and effective key length are equal.
8083 *) Add a bunch of functions that should simplify the creation of
8084 X509_NAME structures. Now you should be able to do:
8085 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8086 and have it automatically work out the correct field type and fill in
8087 the structures. The more adventurous can try:
8088 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8089 and it will (hopefully) work out the correct multibyte encoding.
8092 *) Change the 'req' utility to use the new field handling and multibyte
8093 copy routines. Before the DN field creation was handled in an ad hoc
8094 way in req, ca, and x509 which was rather broken and didn't support
8095 BMPStrings or UTF8Strings. Since some software doesn't implement
8096 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8097 using the dirstring_type option. See the new comment in the default
8098 openssl.cnf for more info.
8101 *) Make crypto/rand/md_rand.c more robust:
8102 - Assure unique random numbers after fork().
8103 - Make sure that concurrent threads access the global counter and
8104 md serializably so that we never lose entropy in them
8105 or use exactly the same state in multiple threads.
8106 Access to the large state is not always serializable because
8107 the additional locking could be a performance killer, and
8108 md should be large enough anyway.
8111 *) New file apps/app_rand.c with commonly needed functionality
8112 for handling the random seed file.
8114 Use the random seed file in some applications that previously did not:
8116 dsaparam -genkey (which also ignored its '-rand' option),
8119 x509 (when signing).
8120 Except on systems with /dev/urandom, it is crucial to have a random
8121 seed file at least for key creation, DSA signing, and for DH exchanges;
8122 for RSA signatures we could do without one.
8124 gendh and gendsa (unlike genrsa) used to read only the first byte
8125 of each file listed in the '-rand' option. The function as previously
8126 found in genrsa is now in app_rand.c and is used by all programs
8127 that support '-rand'.
8130 *) In RAND_write_file, use mode 0600 for creating files;
8131 don't just chmod when it may be too late.
8134 *) Report an error from X509_STORE_load_locations
8135 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8138 *) New function ASN1_mbstring_copy() this copies a string in either
8139 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8140 into an ASN1_STRING type. A mask of permissible types is passed
8141 and it chooses the "minimal" type to use or an error if not type
8145 *) Add function equivalents to the various macros in asn1.h. The old
8146 macros are retained with an M_ prefix. Code inside the library can
8147 use the M_ macros. External code (including the openssl utility)
8148 should *NOT* in order to be "shared library friendly".
8151 *) Add various functions that can check a certificate's extensions
8152 to see if it usable for various purposes such as SSL client,
8153 server or S/MIME and CAs of these types. This is currently
8154 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8155 verification. Also added a -purpose flag to x509 utility to
8156 print out all the purposes.
8159 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8163 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8164 for, obtain and decode and extension and obtain its critical flag.
8165 This allows all the necessary extension code to be handled in a
8166 single function call.
8169 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8170 platforms. See crypto/rc4/rc4_enc.c for further details.
8173 *) New -noout option to asn1parse. This causes no output to be produced
8174 its main use is when combined with -strparse and -out to extract data
8175 from a file (which may not be in ASN.1 format).
8178 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8179 when producing the local key id.
8180 [Richard Levitte <levitte@stacken.kth.se>]
8182 *) New option -dhparam in s_server. This allows a DH parameter file to be
8183 stated explicitly. If it is not stated then it tries the first server
8184 certificate file. The previous behaviour hard coded the filename
8188 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8189 a public key to be input or output. For example:
8190 openssl rsa -in key.pem -pubout -out pubkey.pem
8191 Also added necessary DSA public key functions to handle this.
8194 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8195 in the message. This was handled by allowing
8196 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8197 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8199 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8200 to the end of the strings whereas this didn't. This would cause problems
8201 if strings read with d2i_ASN1_bytes() were later modified.
8202 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8204 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8205 data and it contains EOF it will end up returning an error. This is
8206 caused by input 46 bytes long. The cause is due to the way base64
8207 BIOs find the start of base64 encoded data. They do this by trying a
8208 trial decode on each line until they find one that works. When they
8209 do a flag is set and it starts again knowing it can pass all the
8210 data directly through the decoder. Unfortunately it doesn't reset
8211 the context it uses. This means that if EOF is reached an attempt
8212 is made to pass two EOFs through the context and this causes the
8213 resulting error. This can also cause other problems as well. As is
8214 usual with these problems it takes *ages* to find and the fix is
8215 trivial: move one line.
8216 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8218 *) Ugly workaround to get s_client and s_server working under Windows. The
8219 old code wouldn't work because it needed to select() on sockets and the
8220 tty (for keypresses and to see if data could be written). Win32 only
8221 supports select() on sockets so we select() with a 1s timeout on the
8222 sockets and then see if any characters are waiting to be read, if none
8223 are present then we retry, we also assume we can always write data to
8224 the tty. This isn't nice because the code then blocks until we've
8225 received a complete line of data and it is effectively polling the
8226 keyboard at 1s intervals: however it's quite a bit better than not
8227 working at all :-) A dedicated Windows application might handle this
8228 with an event loop for example.
8231 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8232 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8233 will be called when RSA_sign() and RSA_verify() are used. This is useful
8234 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8235 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8236 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8237 This necessitated the support of an extra signature type NID_md5_sha1
8238 for SSL signatures and modifications to the SSL library to use it instead
8239 of calling RSA_public_decrypt() and RSA_private_encrypt().
8242 *) Add new -verify -CAfile and -CApath options to the crl program, these
8243 will lookup a CRL issuers certificate and verify the signature in a
8244 similar way to the verify program. Tidy up the crl program so it
8245 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8246 less strict. It will now permit CRL extensions even if it is not
8247 a V2 CRL: this will allow it to tolerate some broken CRLs.
8250 *) Initialize all non-automatic variables each time one of the openssl
8251 sub-programs is started (this is necessary as they may be started
8252 multiple times from the "OpenSSL>" prompt).
8253 [Lennart Bang, Bodo Moeller]
8255 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8256 removing all other RSA functionality (this is what NO_RSA does). This
8257 is so (for example) those in the US can disable those operations covered
8258 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8262 *) Non-copying interface to BIO pairs.
8263 (still largely untested)
8266 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8267 ASCII string. This was handled independently in various places before.
8270 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8271 UTF8 strings a character at a time.
8274 *) Use client_version from client hello to select the protocol
8275 (s23_srvr.c) and for RSA client key exchange verification
8276 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8279 *) Add various utility functions to handle SPKACs, these were previously
8280 handled by poking round in the structure internals. Added new function
8281 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8282 print, verify and generate SPKACs. Based on an original idea from
8283 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8286 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8289 *) Allow the config file extension section to be overwritten on the
8290 command line. Based on an original idea from Massimiliano Pala
8291 <madwolf@comune.modena.it>. The new option is called -extensions
8292 and can be applied to ca, req and x509. Also -reqexts to override
8293 the request extensions in req and -crlexts to override the crl extensions
8297 *) Add new feature to the SPKAC handling in ca. Now you can include
8298 the same field multiple times by preceding it by "XXXX." for example:
8301 this is the same syntax as used in the req config file.
8304 *) Allow certificate extensions to be added to certificate requests. These
8305 are specified in a 'req_extensions' option of the req section of the
8306 config file. They can be printed out with the -text option to req but
8307 are otherwise ignored at present.
8310 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8311 data read consists of only the final block it would not decrypted because
8312 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8313 A misplaced 'break' also meant the decrypted final block might not be
8314 copied until the next read.
8317 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8318 a few extra parameters to the DH structure: these will be useful if
8319 for example we want the value of 'q' or implement X9.42 DH.
8322 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8323 provides hooks that allow the default DSA functions or functions on a
8324 "per key" basis to be replaced. This allows hardware acceleration and
8325 hardware key storage to be handled without major modification to the
8326 library. Also added low level modexp hooks and CRYPTO_EX structure and
8327 associated functions.
8330 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8331 as "read only": it can't be written to and the buffer it points to will
8332 not be freed. Reading from a read only BIO is much more efficient than
8333 a normal memory BIO. This was added because there are several times when
8334 an area of memory needs to be read from a BIO. The previous method was
8335 to create a memory BIO and write the data to it, this results in two
8336 copies of the data and an O(n^2) reading algorithm. There is a new
8337 function BIO_new_mem_buf() which creates a read only memory BIO from
8338 an area of memory. Also modified the PKCS#7 routines to use read only
8342 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8343 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8344 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8345 but a retry condition occured while trying to read the rest.
8348 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8349 NID_pkcs7_encrypted by default: this was wrong since this should almost
8350 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8351 the encrypted data type: this is a more sensible place to put it and it
8352 allows the PKCS#12 code to be tidied up that duplicated this
8356 *) Changed obj_dat.pl script so it takes its input and output files on
8357 the command line. This should avoid shell escape redirection problems
8361 *) Initial support for certificate extension requests, these are included
8362 in things like Xenroll certificate requests. Included functions to allow
8363 extensions to be obtained and added.
8366 *) -crlf option to s_client and s_server for sending newlines as
8367 CRLF (as required by many protocols).
8370 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8372 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8373 [Ralf S. Engelschall]
8375 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8376 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8378 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8382 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8383 DH parameters/keys (q is lost during that conversion, but the resulting
8384 DH parameters contain its length).
8386 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8387 much faster than DH_generate_parameters (which creates parameters
8388 where p = 2*q + 1), and also the smaller q makes DH computations
8389 much more efficient (160-bit exponentiation instead of 1024-bit
8390 exponentiation); so this provides a convenient way to support DHE
8391 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8392 utter importance to use
8393 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8395 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8396 when such DH parameters are used, because otherwise small subgroup
8397 attacks may become possible!
8400 *) Avoid memory leak in i2d_DHparams.
8403 *) Allow the -k option to be used more than once in the enc program:
8404 this allows the same encrypted message to be read by multiple recipients.
8407 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8408 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8409 it will always use the numerical form of the OID, even if it has a short
8413 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8414 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8415 otherwise bn_mod_exp was called. In the case of hardware keys for example
8416 no private key components need be present and it might store extra data
8417 in the RSA structure, which cannot be accessed from bn_mod_exp.
8418 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8419 private key operations.
8422 *) Added support for SPARC Linux.
8425 *) pem_password_cb function type incompatibly changed from
8426 typedef int pem_password_cb(char *buf, int size, int rwflag);
8428 ....(char *buf, int size, int rwflag, void *userdata);
8429 so that applications can pass data to their callbacks:
8430 The PEM[_ASN1]_{read,write}... functions and macros now take an
8431 additional void * argument, which is just handed through whenever
8432 the password callback is called.
8433 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8435 New function SSL_CTX_set_default_passwd_cb_userdata.
8437 Compatibility note: As many C implementations push function arguments
8438 onto the stack in reverse order, the new library version is likely to
8439 interoperate with programs that have been compiled with the old
8440 pem_password_cb definition (PEM_whatever takes some data that
8441 happens to be on the stack as its last argument, and the callback
8442 just ignores this garbage); but there is no guarantee whatsoever that
8445 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8446 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8447 problems not only on Windows, but also on some Unix platforms.
8448 To avoid problematic command lines, these definitions are now in an
8449 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8450 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8453 *) MIPS III/IV assembler module is reimplemented.
8456 *) More DES library cleanups: remove references to srand/rand and
8457 delete an unused file.
8460 *) Add support for the the free Netwide assembler (NASM) under Win32,
8461 since not many people have MASM (ml) and it can be hard to obtain.
8462 This is currently experimental but it seems to work OK and pass all
8463 the tests. Check out INSTALL.W32 for info.
8466 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8467 without temporary keys kept an extra copy of the server key,
8468 and connections with temporary keys did not free everything in case
8472 *) New function RSA_check_key and new openssl rsa option -check
8473 for verifying the consistency of RSA keys.
8474 [Ulf Moeller, Bodo Moeller]
8476 *) Various changes to make Win32 compile work:
8477 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8478 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8479 comparison" warnings.
8480 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8483 *) Add a debugging option to PKCS#5 v2 key generation function: when
8484 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8485 derived keys are printed to stderr.
8488 *) Copy the flags in ASN1_STRING_dup().
8489 [Roman E. Pavlov <pre@mo.msk.ru>]
8491 *) The x509 application mishandled signing requests containing DSA
8492 keys when the signing key was also DSA and the parameters didn't match.
8494 It was supposed to omit the parameters when they matched the signing key:
8495 the verifying software was then supposed to automatically use the CA's
8496 parameters if they were absent from the end user certificate.
8498 Omitting parameters is no longer recommended. The test was also
8499 the wrong way round! This was probably due to unusual behaviour in
8500 EVP_cmp_parameters() which returns 1 if the parameters match.
8501 This meant that parameters were omitted when they *didn't* match and
8502 the certificate was useless. Certificates signed with 'ca' didn't have
8504 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8506 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8507 The interface is as follows:
8508 Applications can use
8509 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8510 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8511 "off" is now the default.
8512 The library internally uses
8513 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8514 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8515 to disable memory-checking temporarily.
8517 Some inconsistent states that previously were possible (and were
8518 even the default) are now avoided.
8520 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8521 with each memory chunk allocated; this is occasionally more helpful
8522 than just having a counter.
8524 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8526 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8530 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8531 which largely parallels "options", but is for changing API behaviour,
8532 whereas "options" are about protocol behaviour.
8533 Initial "mode" flags are:
8535 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8536 a single record has been written.
8537 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8538 retries use the same buffer location.
8539 (But all of the contents must be
8543 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8546 *) Fix problems with no-hmac etc.
8547 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8549 *) New functions RSA_get_default_method(), RSA_set_method() and
8550 RSA_get_method(). These allows replacement of RSA_METHODs without having
8551 to mess around with the internals of an RSA structure.
8554 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8555 Also really enable memory leak checks in openssl.c and in some
8557 [Chad C. Mulligan, Bodo Moeller]
8559 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8560 up the length of negative integers. This has now been simplified to just
8561 store the length when it is first determined and use it later, rather
8562 than trying to keep track of where data is copied and updating it to
8564 [Steve Henson, reported by Brien Wheeler
8565 <bwheeler@authentica-security.com>]
8567 *) Add a new function PKCS7_signatureVerify. This allows the verification
8568 of a PKCS#7 signature but with the signing certificate passed to the
8569 function itself. This contrasts with PKCS7_dataVerify which assumes the
8570 certificate is present in the PKCS#7 structure. This isn't always the
8571 case: certificates can be omitted from a PKCS#7 structure and be
8572 distributed by "out of band" means (such as a certificate database).
8575 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8576 function prototypes in pem.h, also change util/mkdef.pl to add the
8577 necessary function names.
8580 *) mk1mf.pl (used by Windows builds) did not properly read the
8581 options set by Configure in the top level Makefile, and Configure
8582 was not even able to write more than one option correctly.
8583 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8586 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8587 file to be loaded from a BIO or FILE pointer. The BIO version will
8588 for example allow memory BIOs to contain config info.
8591 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8592 Whoever hopes to achieve shared-library compatibility across versions
8593 must use this, not the compile-time macro.
8594 (Exercise 0.9.4: Which is the minimum library version required by
8596 Note: All this applies only to multi-threaded programs, others don't
8600 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8601 through a BIO pair triggered the default case, i.e.
8602 SSLerr(...,SSL_R_UNKNOWN_STATE).
8605 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8606 can use the SSL library even if none of the specific BIOs is
8610 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8611 for the encoded length.
8612 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8614 *) Add initial documentation of the X509V3 functions.
8617 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8618 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8619 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8620 secure PKCS#8 private key format with a high iteration count.
8623 *) Fix determination of Perl interpreter: A perl or perl5
8624 _directory_ in $PATH was also accepted as the interpreter.
8625 [Ralf S. Engelschall]
8627 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8628 wrong with it but it was very old and did things like calling
8629 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8633 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8634 to use the new extension code.
8637 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8638 with macros. This should make it easier to change their form, add extra
8639 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8643 *) Add to configuration table a new entry that can specify an alternative
8644 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8645 according to Mark Crispin <MRC@Panda.COM>.
8649 *) DES CBC did not update the IV. Weird.
8652 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8653 Changing the behaviour of the former might break existing programs --
8654 where IV updating is needed, des_ncbc_encrypt can be used.
8657 *) When bntest is run from "make test" it drives bc to check its
8658 calculations, as well as internally checking them. If an internal check
8659 fails, it needs to cause bc to give a non-zero result or make test carries
8660 on without noticing the failure. Fixed.
8663 *) DES library cleanups.
8666 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8667 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8668 ciphers. NOTE: although the key derivation function has been verified
8669 against some published test vectors it has not been extensively tested
8670 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8674 *) Instead of "mkdir -p", which is not fully portable, use new
8675 Perl script "util/mkdir-p.pl".
8678 *) Rewrite the way password based encryption (PBE) is handled. It used to
8679 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8680 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8681 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8682 the 'parameter' field of the AlgorithmIdentifier is passed to the
8683 underlying key generation function so it must do its own ASN1 parsing.
8684 This has also changed the EVP_PBE_CipherInit() function which now has a
8685 'parameter' argument instead of literal salt and iteration count values
8686 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8689 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8690 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8691 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8692 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8693 value was just used as a "magic string" and not used directly its
8694 value doesn't matter.
8697 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8701 *) "linux-sparc64" configuration (ultrapenguin).
8702 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8703 "linux-sparc" configuration.
8704 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8706 *) config now generates no-xxx options for missing ciphers.
8709 *) Support the EBCDIC character set (work in progress).
8710 File ebcdic.c not yet included because it has a different license.
8711 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8713 *) Support BS2000/OSD-POSIX.
8714 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8716 *) Make callbacks for key generation use void * instead of char *.
8719 *) Make S/MIME samples compile (not yet tested).
8722 *) Additional typesafe stacks.
8725 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8729 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8731 *) New configuration variant "sco5-gcc".
8733 *) Updated some demos.
8734 [Sean O Riordain, Wade Scholine]
8736 *) Add missing BIO_free at exit of pkcs12 application.
8739 *) Fix memory leak in conf.c.
8742 *) Updates for Win32 to assembler version of MD5.
8745 *) Set #! path to perl in apps/der_chop to where we found it
8746 instead of using a fixed path.
8749 *) SHA library changes for irix64-mips4-cc.
8752 *) Improvements for VMS support.
8756 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8758 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8759 This also avoids the problems with SC4.2 and unpatched SC5.
8760 [Andy Polyakov <appro@fy.chalmers.se>]
8762 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8763 These are required because of the typesafe stack would otherwise break
8764 existing code. If old code used a structure member which used to be STACK
8765 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8766 sk_num or sk_value it would produce an error because the num, data members
8767 are not present in STACK_OF. Now it just produces a warning. sk_set
8768 replaces the old method of assigning a value to sk_value
8769 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8770 that does this will no longer work (and should use sk_set instead) but
8771 this could be regarded as a "questionable" behaviour anyway.
8774 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8775 correctly handle encrypted S/MIME data.
8778 *) Change type of various DES function arguments from des_cblock
8779 (which means, in function argument declarations, pointer to char)
8780 to des_cblock * (meaning pointer to array with 8 char elements),
8781 which allows the compiler to do more typechecking; it was like
8782 that back in SSLeay, but with lots of ugly casts.
8784 Introduce new type const_des_cblock.
8787 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8788 problems: find RecipientInfo structure that matches recipient certificate
8789 and initialise the ASN1 structures properly based on passed cipher.
8792 *) Belatedly make the BN tests actually check the results.
8795 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8796 to and from BNs: it was completely broken. New compilation option
8797 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8798 key elements as negative integers.
8801 *) Reorganize and speed up MD5.
8802 [Andy Polyakov <appro@fy.chalmers.se>]
8805 [Richard Levitte <richard@levitte.org>]
8807 *) New option -out to asn1parse to allow the parsed structure to be
8808 output to a file. This is most useful when combined with the -strparse
8809 option to examine the output of things like OCTET STRINGS.
8812 *) Make SSL library a little more fool-proof by not requiring any longer
8813 that SSL_set_{accept,connect}_state be called before
8814 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8815 in many applications because usually everything *appeared* to work as
8816 intended anyway -- now it really works as intended).
8819 *) Move openssl.cnf out of lib/.
8822 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8823 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8824 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8825 [Ralf S. Engelschall]
8827 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8828 handle PKCS#7 enveloped data properly.
8829 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8831 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8832 copying pointers. The cert_st handling is changed by this in
8833 various ways (and thus what used to be known as ctx->default_cert
8834 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8835 any longer when s->cert does not give us what we need).
8836 ssl_cert_instantiate becomes obsolete by this change.
8837 As soon as we've got the new code right (possibly it already is?),
8838 we have solved a couple of bugs of the earlier code where s->cert
8839 was used as if it could not have been shared with other SSL structures.
8841 Note that using the SSL API in certain dirty ways now will result
8842 in different behaviour than observed with earlier library versions:
8843 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8844 does not influence s as it used to.
8846 In order to clean up things more thoroughly, inside SSL_SESSION
8847 we don't use CERT any longer, but a new structure SESS_CERT
8848 that holds per-session data (if available); currently, this is
8849 the peer's certificate chain and, for clients, the server's certificate
8850 and temporary key. CERT holds only those values that can have
8851 meaningful defaults in an SSL_CTX.
8854 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8855 from the internal representation. Various PKCS#7 fixes: remove some
8856 evil casts and set the enc_dig_alg field properly based on the signing
8860 *) Allow PKCS#12 password to be set from the command line or the
8861 environment. Let 'ca' get its config file name from the environment
8862 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8866 *) Allow certificate policies extension to use an IA5STRING for the
8867 organization field. This is contrary to the PKIX definition but
8868 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8872 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8873 without disallowing inline assembler and the like for non-pedantic builds.
8876 *) Support Borland C++ builder.
8877 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8882 *) SHA-1 cleanups and performance enhancements.
8883 [Andy Polyakov <appro@fy.chalmers.se>]
8885 *) Sparc v8plus assembler for the bignum library.
8886 [Andy Polyakov <appro@fy.chalmers.se>]
8888 *) Accept any -xxx and +xxx compiler options in Configure.
8891 *) Update HPUX configuration.
8894 *) Add missing sk_<type>_unshift() function to safestack.h
8895 [Ralf S. Engelschall]
8897 *) New function SSL_CTX_use_certificate_chain_file that sets the
8898 "extra_cert"s in addition to the certificate. (This makes sense
8899 only for "PEM" format files, as chains as a whole are not
8903 *) Support verify_depth from the SSL API.
8904 x509_vfy.c had what can be considered an off-by-one-error:
8905 Its depth (which was not part of the external interface)
8906 was actually counting the number of certificates in a chain;
8907 now it really counts the depth.
8910 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8911 instead of X509err, which often resulted in confusing error
8912 messages since the error codes are not globally unique
8913 (e.g. an alleged error in ssl3_accept when a certificate
8914 didn't match the private key).
8916 *) New function SSL_CTX_set_session_id_context that allows to set a default
8917 value (so that you don't need SSL_set_session_id_context for each
8918 connection using the SSL_CTX).
8921 *) OAEP decoding bug fix.
8924 *) Support INSTALL_PREFIX for package builders, as proposed by
8928 *) New Configure options "threads" and "no-threads". For systems
8929 where the proper compiler options are known (currently Solaris
8930 and Linux), "threads" is the default.
8933 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8936 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8937 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8938 such as /usr/local/bin.
8941 *) "make linux-shared" to build shared libraries.
8942 [Niels Poppe <niels@netbox.org>]
8944 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8947 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8948 extension adding in x509 utility.
8951 *) Remove NOPROTO sections and error code comments.
8954 *) Partial rewrite of the DEF file generator to now parse the ANSI
8958 *) New Configure options --prefix=DIR and --openssldir=DIR.
8961 *) Complete rewrite of the error code script(s). It is all now handled
8962 by one script at the top level which handles error code gathering,
8963 header rewriting and C source file generation. It should be much better
8964 than the old method: it now uses a modified version of Ulf's parser to
8965 read the ANSI prototypes in all header files (thus the old K&R definitions
8966 aren't needed for error creation any more) and do a better job of
8967 translating function codes into names. The old 'ASN1 error code imbedded
8968 in a comment' is no longer necessary and it doesn't use .err files which
8969 have now been deleted. Also the error code call doesn't have to appear all
8970 on one line (which resulted in some large lines...).
8973 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8976 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8977 0 (which usually indicates a closed connection), but continue reading.
8980 *) Fix some race conditions.
8983 *) Add support for CRL distribution points extension. Add Certificate
8984 Policies and CRL distribution points documentation.
8987 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8990 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8991 8 of keying material. Merlin has also confirmed interop with this fix
8992 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8993 [Merlin Hughes <merlin@baltimore.ie>]
8995 *) Fix lots of warnings.
8996 [Richard Levitte <levitte@stacken.kth.se>]
8998 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8999 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9000 [Richard Levitte <levitte@stacken.kth.se>]
9002 *) Fix problems with sizeof(long) == 8.
9003 [Andy Polyakov <appro@fy.chalmers.se>]
9005 *) Change functions to ANSI C.
9008 *) Fix typos in error codes.
9009 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9011 *) Remove defunct assembler files from Configure.
9014 *) SPARC v8 assembler BIGNUM implementation.
9015 [Andy Polyakov <appro@fy.chalmers.se>]
9017 *) Support for Certificate Policies extension: both print and set.
9018 Various additions to support the r2i method this uses.
9021 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9022 return a const string when you are expecting an allocated buffer.
9025 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9026 types DirectoryString and DisplayText.
9029 *) Add code to allow r2i extensions to access the configuration database,
9030 add an LHASH database driver and add several ctx helper functions.
9033 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9034 fail when they extended the size of a BIGNUM.
9037 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9038 support typesafe stack.
9041 *) Fix typo in SSL_[gs]et_options().
9042 [Nils Frostberg <nils@medcom.se>]
9044 *) Delete various functions and files that belonged to the (now obsolete)
9045 old X509V3 handling code.
9048 *) New Configure option "rsaref".
9051 *) Don't auto-generate pem.h.
9054 *) Introduce type-safe ASN.1 SETs.
9057 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9058 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9060 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9061 that links with OpenSSL (well at least cause lots of warnings), but fear
9062 not: the conversion is trivial, and it eliminates loads of evil casts. A
9063 few STACKed things have been converted already. Feel free to convert more.
9064 In the fullness of time, I'll do away with the STACK type altogether.
9067 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9068 specified in <certfile> by updating the entry in the index.txt file.
9069 This way one no longer has to edit the index.txt file manually for
9070 revoking a certificate. The -revoke option does the gory details now.
9071 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9073 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9074 `-text' option at all and this way the `-noout -text' combination was
9075 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9076 [Ralf S. Engelschall]
9078 *) Make sure a corresponding plain text error message exists for the
9079 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9080 verify callback function determined that a certificate was revoked.
9081 [Ralf S. Engelschall]
9083 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9084 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9085 all available cipers including rc5, which was forgotten until now.
9086 In order to let the testing shell script know which algorithms
9087 are available, a new (up to now undocumented) command
9088 "openssl list-cipher-commands" is used.
9091 *) Bugfix: s_client occasionally would sleep in select() when
9092 it should have checked SSL_pending() first.
9095 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9096 the raw DSA values prior to ASN.1 encoding.
9099 *) Tweaks to Configure
9100 [Niels Poppe <niels@netbox.org>]
9102 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9106 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9109 *) New config option to avoid instructions that are illegal on the 80386.
9110 The default code is faster, but requires at least a 486.
9113 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9114 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9115 same as SSL2_VERSION anyway.
9118 *) New "-showcerts" option for s_client.
9121 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9122 application. Various cleanups and fixes.
9125 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9126 modify error routines to work internally. Add error codes and PBE init
9127 to library startup routines.
9130 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9131 packing functions to asn1 and evp. Changed function names and error
9132 codes along the way.
9135 *) PKCS12 integration: and so it begins... First of several patches to
9136 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9137 objects to objects.h
9140 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9141 and display support for Thawte strong extranet extension.
9144 *) Add LinuxPPC support.
9145 [Jeff Dubrule <igor@pobox.org>]
9147 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9148 bn_div_words in alpha.s.
9149 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9151 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9152 OAEP isn't supported when OpenSSL is built with RSAref.
9153 [Ulf Moeller <ulf@fitug.de>]
9155 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9156 so they no longer are missing under -DNOPROTO.
9157 [Soren S. Jorvang <soren@t.dk>]
9160 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9162 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9163 doesn't work when the session is reused. Coming soon!
9166 *) Fix a security hole, that allows sessions to be reused in the wrong
9167 context thus bypassing client cert protection! All software that uses
9168 client certs and session caches in multiple contexts NEEDS PATCHING to
9169 allow session reuse! A fuller solution is in the works.
9170 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9172 *) Some more source tree cleanups (removed obsolete files
9173 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9174 permission on "config" script to be executable) and a fix for the INSTALL
9176 [Ulf Moeller <ulf@fitug.de>]
9178 *) Remove some legacy and erroneous uses of malloc, free instead of
9180 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9182 *) Make rsa_oaep_test return non-zero on error.
9183 [Ulf Moeller <ulf@fitug.de>]
9185 *) Add support for native Solaris shared libraries. Configure
9186 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9187 if someone would make that last step automatic.
9188 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9190 *) ctx_size was not built with the right compiler during "make links". Fixed.
9193 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9194 except NULL ciphers". This means the default cipher list will no longer
9195 enable NULL ciphers. They need to be specifically enabled e.g. with
9196 the string "DEFAULT:eNULL".
9199 *) Fix to RSA private encryption routines: if p < q then it would
9200 occasionally produce an invalid result. This will only happen with
9201 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9204 *) Be less restrictive and allow also `perl util/perlpath.pl
9205 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9206 because this way one can also use an interpreter named `perl5' (which is
9207 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9208 installed as `perl').
9209 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9211 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9212 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9214 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9215 advapi32.lib to Win32 build and change the pem test comparision
9216 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9217 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9218 and crypto/des/ede_cbcm_enc.c.
9221 *) DES quad checksum was broken on big-endian architectures. Fixed.
9224 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9225 Win32 test batch file so it (might) work again. The Win32 test batch file
9226 is horrible: I feel ill....
9229 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9230 in e_os.h. Audit of header files to check ANSI and non ANSI
9231 sections: 10 functions were absent from non ANSI section and not exported
9232 from Windows DLLs. Fixed up libeay.num for new functions.
9235 *) Make `openssl version' output lines consistent.
9236 [Ralf S. Engelschall]
9238 *) Fix Win32 symbol export lists for BIO functions: Added
9239 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9240 to ms/libeay{16,32}.def.
9241 [Ralf S. Engelschall]
9243 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9244 fine under Unix and passes some trivial tests I've now added. But the
9245 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9246 added to make sure no one expects that this stuff really works in the
9247 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9248 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9250 [Ralf S. Engelschall]
9252 *) Fix the generation of two part addresses in perl.
9253 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9255 *) Add config entry for Linux on MIPS.
9256 [John Tobey <jtobey@channel1.com>]
9258 *) Make links whenever Configure is run, unless we are on Windoze.
9261 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9262 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9266 *) Add a useful kludge to allow package maintainers to specify compiler and
9267 other platforms details on the command line without having to patch the
9268 Configure script everytime: One now can use ``perl Configure
9269 <id>:<details>'', i.e. platform ids are allowed to have details appended
9270 to them (seperated by colons). This is treated as there would be a static
9271 pre-configured entry in Configure's %table under key <id> with value
9272 <details> and ``perl Configure <id>'' is called. So, when you want to
9273 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9274 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9275 now, which overrides the FreeBSD-elf entry on-the-fly.
9276 [Ralf S. Engelschall]
9278 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9281 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9282 on the `perl Configure ...' command line. This way one can compile
9283 OpenSSL libraries with Position Independent Code (PIC) which is needed
9284 for linking it into DSOs.
9285 [Ralf S. Engelschall]
9287 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9291 *) Cleaned up the LICENSE document: The official contact for any license
9292 questions now is the OpenSSL core team under openssl-core@openssl.org.
9293 And add a paragraph about the dual-license situation to make sure people
9294 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9295 to the OpenSSL toolkit.
9296 [Ralf S. Engelschall]
9298 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9299 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9300 Additonally cleaned up the `make links' target: Remove unnecessary
9301 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9302 to speed processing and no longer clutter the display with confusing
9303 stuff. Instead only the actually done links are displayed.
9304 [Ralf S. Engelschall]
9306 *) Permit null encryption ciphersuites, used for authentication only. It used
9307 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9308 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9312 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9313 signed attributes when verifying signatures (this would break them),
9314 the detached data encoding was wrong and public keys obtained using
9315 X509_get_pubkey() weren't freed.
9318 *) Add text documentation for the BUFFER functions. Also added a work around
9319 to a Win95 console bug. This was triggered by the password read stuff: the
9320 last character typed gets carried over to the next fread(). If you were
9321 generating a new cert request using 'req' for example then the last
9322 character of the passphrase would be CR which would then enter the first
9326 *) Added the new `Includes OpenSSL Cryptography Software' button as
9327 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9328 button and can be used by applications based on OpenSSL to show the
9329 relationship to the OpenSSL project.
9330 [Ralf S. Engelschall]
9332 *) Remove confusing variables in function signatures in files
9333 ssl/ssl_lib.c and ssl/ssl.h.
9334 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9336 *) Don't install bss_file.c under PREFIX/include/
9337 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9339 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9340 functions that return function pointers and has support for NT specific
9341 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9342 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9343 unsigned to signed types: this was killing the Win32 compile.
9346 *) Add new certificate file to stack functions,
9347 SSL_add_dir_cert_subjects_to_stack() and
9348 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9349 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9350 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9351 This means that Apache-SSL and similar packages don't have to mess around
9352 to add as many CAs as they want to the preferred list.
9355 *) Experiment with doxygen documentation. Currently only partially applied to
9357 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9358 openssl.doxy as the configuration file.
9361 *) Get rid of remaining C++-style comments which strict C compilers hate.
9362 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9364 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9365 compiled in by default: it has problems with large keys.
9368 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9369 DH private keys and/or callback functions which directly correspond to
9370 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9371 is needed for applications which have to configure certificates on a
9372 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9374 For the RSA certificate situation is makes no difference, but
9375 for the DSA certificate situation this fixes the "no shared cipher"
9376 problem where the OpenSSL cipher selection procedure failed because the
9377 temporary keys were not overtaken from the context and the API provided
9378 no way to reconfigure them.
9379 The new functions now let applications reconfigure the stuff and they
9380 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9381 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9382 non-public-API function ssl_cert_instantiate() is used as a helper
9383 function and also to reduce code redundancy inside ssl_rsa.c.
9384 [Ralf S. Engelschall]
9386 *) Move s_server -dcert and -dkey options out of the undocumented feature
9387 area because they are useful for the DSA situation and should be
9388 recognized by the users.
9389 [Ralf S. Engelschall]
9391 *) Fix the cipher decision scheme for export ciphers: the export bits are
9392 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9393 SSL_EXP_MASK. So, the original variable has to be used instead of the
9394 already masked variable.
9395 [Richard Levitte <levitte@stacken.kth.se>]
9397 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9398 [Richard Levitte <levitte@stacken.kth.se>]
9400 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9401 from `int' to `unsigned int' because it's a length and initialized by
9402 EVP_DigestFinal() which expects an `unsigned int *'.
9403 [Richard Levitte <levitte@stacken.kth.se>]
9405 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9406 script. Instead use the usual Shell->Perl transition trick.
9407 [Ralf S. Engelschall]
9409 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9410 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9411 -noout -modulus' as it's already the case for `openssl rsa -noout
9412 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9413 currently the public key is printed (a decision which was already done by
9414 `openssl dsa -modulus' in the past) which serves a similar purpose.
9415 Additionally the NO_RSA no longer completely removes the whole -modulus
9416 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9418 [Ralf S. Engelschall]
9420 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9421 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9422 [Arne Ansper <arne@ats.cyber.ee>]
9424 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9425 to be added. Now both 'req' and 'ca' can use new objects defined in the
9429 *) Add cool BIO that does syslog (or event log on NT).
9430 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9432 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9433 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9434 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9435 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9438 *) Add preliminary config info for new extension code.
9441 *) Make RSA_NO_PADDING really use no padding.
9442 [Ulf Moeller <ulf@fitug.de>]
9444 *) Generate errors when private/public key check is done.
9447 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9448 for some CRL extensions and new objects added.
9451 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9452 key usage extension and fuller support for authority key id.
9455 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9456 padding method for RSA, which is recommended for new applications in PKCS
9457 #1 v2.0 (RFC 2437, October 1998).
9458 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9459 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9460 against Bleichbacher's attack on RSA.
9461 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9464 *) Updates to the new SSL compression code
9465 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9467 *) Fix so that the version number in the master secret, when passed
9468 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9469 (because the server will not accept higher), that the version number
9470 is 0x03,0x01, not 0x03,0x00
9471 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9473 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9474 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9475 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9478 *) Support for RAW extensions where an arbitrary extension can be
9479 created by including its DER encoding. See apps/openssl.cnf for
9483 *) Make sure latest Perl versions don't interpret some generated C array
9484 code as Perl array code in the crypto/err/err_genc.pl script.
9485 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9487 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9488 not many people have the assembler. Various Win32 compilation fixes and
9489 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9493 *) Modify configure script 'Configure' to automatically create crypto/date.h
9494 file under Win32 and also build pem.h from pem.org. New script
9495 util/mkfiles.pl to create the MINFO file on environments that can't do a
9496 'make files': perl util/mkfiles.pl >MINFO should work.
9499 *) Major rework of DES function declarations, in the pursuit of correctness
9500 and purity. As a result, many evil casts evaporated, and some weirdness,
9501 too. You may find this causes warnings in your code. Zapping your evil
9502 casts will probably fix them. Mostly.
9505 *) Fix for a typo in asn1.h. Bug fix to object creation script
9506 obj_dat.pl. It considered a zero in an object definition to mean
9507 "end of object": none of the objects in objects.h have any zeros
9508 so it wasn't spotted.
9509 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9511 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9512 Masking (CBCM). In the absence of test vectors, the best I have been able
9513 to do is check that the decrypt undoes the encrypt, so far. Send me test
9514 vectors if you have them.
9517 *) Correct calculation of key length for export ciphers (too much space was
9518 allocated for null ciphers). This has not been tested!
9521 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9522 message is now correct (it understands "crypto" and "ssl" on its
9523 command line). There is also now an "update" option. This will update
9524 the util/ssleay.num and util/libeay.num files with any new functions.
9526 perl util/mkdef.pl crypto ssl update
9527 it will update them.
9530 *) Overhauled the Perl interface (perl/*):
9531 - ported BN stuff to OpenSSL's different BN library
9532 - made the perl/ source tree CVS-aware
9533 - renamed the package from SSLeay to OpenSSL (the files still contain
9534 their history because I've copied them in the repository)
9535 - removed obsolete files (the test scripts will be replaced
9536 by better Test::Harness variants in the future)
9537 [Ralf S. Engelschall]
9539 *) First cut for a very conservative source tree cleanup:
9540 1. merge various obsolete readme texts into doc/ssleay.txt
9541 where we collect the old documents and readme texts.
9542 2. remove the first part of files where I'm already sure that we no
9543 longer need them because of three reasons: either they are just temporary
9544 files which were left by Eric or they are preserved original files where
9545 I've verified that the diff is also available in the CVS via "cvs diff
9546 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9547 the crypto/md/ stuff).
9548 [Ralf S. Engelschall]
9550 *) More extension code. Incomplete support for subject and issuer alt
9551 name, issuer and authority key id. Change the i2v function parameters
9552 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9553 what that's for :-) Fix to ASN1 macro which messed up
9554 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9557 *) Preliminary support for ENUMERATED type. This is largely copied from the
9561 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9562 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9564 *) Make sure `make rehash' target really finds the `openssl' program.
9565 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9567 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9568 like to hear about it if this slows down other processors.
9571 *) Add CygWin32 platform information to Configure script.
9572 [Alan Batie <batie@aahz.jf.intel.com>]
9574 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9575 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9577 *) New program nseq to manipulate netscape certificate sequences
9580 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9584 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9585 but the BN code had some problems that would cause failures when
9586 doing certificate verification and some other functions.
9587 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9589 *) Add ASN1 and PEM code to support netscape certificate sequences.
9592 *) Add ASN1 and PEM code to support netscape certificate sequences.
9595 *) Add several PKIX and private extended key usage OIDs.
9598 *) Modify the 'ca' program to handle the new extension code. Modify
9599 openssl.cnf for new extension format, add comments.
9602 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9603 and add a sample to openssl.cnf so req -x509 now adds appropriate
9607 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9608 error code, add initial support to X509_print() and x509 application.
9611 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9612 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9613 stuff is currently isolated and isn't even compiled yet.
9616 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9617 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9618 Removed the versions check from X509 routines when loading extensions:
9619 this allows certain broken certificates that don't set the version
9620 properly to be processed.
9623 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9624 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9625 can still be regenerated with "make depend".
9628 *) Spelling mistake in C version of CAST-128.
9629 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9631 *) Changes to the error generation code. The perl script err-code.pl
9632 now reads in the old error codes and retains the old numbers, only
9633 adding new ones if necessary. It also only changes the .err files if new
9634 codes are added. The makefiles have been modified to only insert errors
9635 when needed (to avoid needlessly modifying header files). This is done
9636 by only inserting errors if the .err file is newer than the auto generated
9637 C file. To rebuild all the error codes from scratch (the old behaviour)
9638 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9639 or delete all the .err files.
9642 *) CAST-128 was incorrectly implemented for short keys. The C version has
9643 been fixed, but is untested. The assembler versions are also fixed, but
9644 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9645 to regenerate it if needed.
9646 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9647 Hagino <itojun@kame.net>]
9649 *) File was opened incorrectly in randfile.c.
9650 [Ulf Möller <ulf@fitug.de>]
9652 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9653 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9654 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9655 al: it's just almost always a UTCTime. Note this patch adds new error
9656 codes so do a "make errors" if there are problems.
9659 *) Correct Linux 1 recognition in config.
9660 [Ulf Möller <ulf@fitug.de>]
9662 *) Remove pointless MD5 hash when using DSA keys in ca.
9663 [Anonymous <nobody@replay.com>]
9665 *) Generate an error if given an empty string as a cert directory. Also
9666 generate an error if handed NULL (previously returned 0 to indicate an
9667 error, but didn't set one).
9668 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9670 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9673 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9674 parameters. This was causing a warning which killed off the Win32 compile.
9677 *) Remove C++ style comments from crypto/bn/bn_local.h.
9678 [Neil Costigan <neil.costigan@celocom.com>]
9680 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9681 based on a text string, looking up short and long names and finally
9682 "dot" format. The "dot" format stuff didn't work. Added new function
9683 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9684 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9685 OID is not part of the table.
9688 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9689 X509_LOOKUP_by_alias().
9692 *) Sort openssl functions by name.
9695 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9696 encryption from sample DSA keys (in case anyone is interested the password
9700 *) Make _all_ *_free functions accept a NULL pointer.
9701 [Frans Heymans <fheymans@isaserver.be>]
9703 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9705 [Anonymous <nobody@replay.com>]
9707 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9708 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9710 *) Don't blow it for numeric -newkey arguments to apps/req.
9711 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9713 *) Temp key "for export" tests were wrong in s3_srvr.c.
9714 [Anonymous <nobody@replay.com>]
9716 *) Add prototype for temp key callback functions
9717 SSL_CTX_set_tmp_{rsa,dh}_callback().
9720 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9721 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9724 *) X509_name_add_entry() freed the wrong thing after an error.
9725 [Arne Ansper <arne@ats.cyber.ee>]
9727 *) rsa_eay.c would attempt to free a NULL context.
9728 [Arne Ansper <arne@ats.cyber.ee>]
9730 *) BIO_s_socket() had a broken should_retry() on Windoze.
9731 [Arne Ansper <arne@ats.cyber.ee>]
9733 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9734 [Arne Ansper <arne@ats.cyber.ee>]
9736 *) Make sure the already existing X509_STORE->depth variable is initialized
9737 in X509_STORE_new(), but document the fact that this variable is still
9738 unused in the certificate verification process.
9739 [Ralf S. Engelschall]
9741 *) Fix the various library and apps files to free up pkeys obtained from
9742 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9745 *) Fix reference counting in X509_PUBKEY_get(). This makes
9746 demos/maurice/example2.c work, amongst others, probably.
9747 [Steve Henson and Ben Laurie]
9749 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9750 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9751 are no longer created. This way we have a single and consistent command
9752 line interface `openssl <command>', similar to `cvs <command>'.
9753 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9755 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9756 BIT STRING wrapper always have zero unused bits.
9759 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9762 *) Make the top-level INSTALL documentation easier to understand.
9765 *) Makefiles updated to exit if an error occurs in a sub-directory
9766 make (including if user presses ^C) [Paul Sutton]
9768 *) Make Montgomery context stuff explicit in RSA data structure.
9771 *) Fix build order of pem and err to allow for generated pem.h.
9774 *) Fix renumbering bug in X509_NAME_delete_entry().
9777 *) Enhanced the err-ins.pl script so it makes the error library number
9778 global and can add a library name. This is needed for external ASN1 and
9779 other error libraries.
9782 *) Fixed sk_insert which never worked properly.
9785 *) Fix ASN1 macros so they can handle indefinite length construted
9786 EXPLICIT tags. Some non standard certificates use these: they can now
9790 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9791 into a single doc/ssleay.txt bundle. This way the information is still
9792 preserved but no longer messes up this directory. Now it's new room for
9793 the new set of documenation files.
9794 [Ralf S. Engelschall]
9796 *) SETs were incorrectly DER encoded. This was a major pain, because they
9797 shared code with SEQUENCEs, which aren't coded the same. This means that
9798 almost everything to do with SETs or SEQUENCEs has either changed name or
9799 number of arguments.
9800 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9802 *) Fix test data to work with the above.
9805 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9806 was already fixed by Eric for 0.9.1 it seems.
9807 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9809 *) Autodetect FreeBSD3.
9812 *) Fix various bugs in Configure. This affects the following platforms:
9816 unixware-2.0-pentium
9820 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9821 before they are needed.
9824 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9828 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9830 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9831 changed SSLeay to OpenSSL in version strings.
9832 [Ralf S. Engelschall]
9834 *) Some fixups to the top-level documents.
9837 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9838 because the symlink to include/ was missing.
9839 [Ralf S. Engelschall]
9841 *) Incorporated the popular no-RSA/DSA-only patches
9842 which allow to compile a RSA-free SSLeay.
9843 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9845 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9846 when "ssleay" is still not found.
9847 [Ralf S. Engelschall]
9849 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9850 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9852 *) Updated the README file.
9853 [Ralf S. Engelschall]
9855 *) Added various .cvsignore files in the CVS repository subdirs
9856 to make a "cvs update" really silent.
9857 [Ralf S. Engelschall]
9859 *) Recompiled the error-definition header files and added
9860 missing symbols to the Win32 linker tables.
9861 [Ralf S. Engelschall]
9863 *) Cleaned up the top-level documents;
9864 o new files: CHANGES and LICENSE
9865 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9866 o merged COPYRIGHT into LICENSE
9867 o removed obsolete TODO file
9868 o renamed MICROSOFT to INSTALL.W32
9869 [Ralf S. Engelschall]
9871 *) Removed dummy files from the 0.9.1b source tree:
9872 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9873 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9874 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9875 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9876 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9877 [Ralf S. Engelschall]
9879 *) Added various platform portability fixes.
9882 *) The Genesis of the OpenSSL rpject:
9883 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9884 Young and Tim J. Hudson created while they were working for C2Net until
9886 [The OpenSSL Project]
9889 Changes between 0.9.0b and 0.9.1b [not released]
9891 *) Updated a few CA certificates under certs/
9894 *) Changed some BIGNUM api stuff.
9897 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9898 DGUX x86, Linux Alpha, etc.
9901 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9902 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9906 *) Add -strparse option to asn1pars program which parses nested
9908 [Dr Stephen Henson <shenson@bigfoot.com>]
9910 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9913 *) DSA fix for "ca" program.
9916 *) Added "-genkey" option to "dsaparam" program.
9919 *) Added RIPE MD160 (rmd160) message digest.
9922 *) Added -a (all) option to "ssleay version" command.
9925 *) Added PLATFORM define which is the id given to Configure.
9928 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9931 *) Extended the ASN.1 parser routines.
9934 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9937 *) Added a BN_CTX to the BN library.
9940 *) Fixed the weak key values in DES library
9943 *) Changed API in EVP library for cipher aliases.
9946 *) Added support for RC2/64bit cipher.
9949 *) Converted the lhash library to the crypto/mem.c functions.
9952 *) Added more recognized ASN.1 object ids.
9955 *) Added more RSA padding checks for SSL/TLS.
9958 *) Added BIO proxy/filter functionality.
9961 *) Added extra_certs to SSL_CTX which can be used
9962 send extra CA certificates to the client in the CA cert chain sending
9963 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9966 *) Now Fortezza is denied in the authentication phase because
9967 this is key exchange mechanism is not supported by SSLeay at all.
9970 *) Additional PKCS1 checks.
9973 *) Support the string "TLSv1" for all TLS v1 ciphers.
9976 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9977 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9980 *) Fixed a few memory leaks.
9983 *) Fixed various code and comment typos.
9986 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9987 bytes sent in the client random.
9988 [Edward Bishop <ebishop@spyglass.com>]