5 Changes between 1.0.0n and 1.0.0o [xx XXX xxxx]
7 *) Add additional DigestInfo checks.
9 Reencode DigestInto in DER and check against the original when
10 verifying RSA signature: this will reject any improperly encoded
11 DigestInfo structures.
13 Note: this is a precautionary measure and no attacks are currently known.
17 Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
19 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
20 to a denial of service attack. A malicious server can crash the client
21 with a null pointer dereference (read) by specifying an anonymous (EC)DH
22 ciphersuite and sending carefully crafted handshake messages.
24 Thanks to Felix Gröbert (Google) for discovering and researching this
29 *) By sending carefully crafted DTLS packets an attacker could cause openssl
30 to leak memory. This can be exploited through a Denial of Service attack.
31 Thanks to Adam Langley for discovering and researching this issue.
35 *) An attacker can force openssl to consume large amounts of memory whilst
36 processing DTLS handshake messages. This can be exploited through a
37 Denial of Service attack.
38 Thanks to Adam Langley for discovering and researching this issue.
42 *) An attacker can force an error condition which causes openssl to crash
43 whilst processing DTLS packets due to memory being freed twice. This
44 can be exploited through a Denial of Service attack.
45 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
50 *) If a multithreaded client connects to a malicious server using a resumed
51 session and the server sends an ec point format extension it could write
52 up to 255 bytes to freed memory.
54 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
59 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
60 X509_name_oneline, X509_name_print_ex et al. to leak some information
61 from the stack. Applications may be affected if they echo pretty printing
62 output to the attacker.
64 Thanks to Ivan Fratric (Google) for discovering this issue.
66 [Emilia Käsper, and Steve Henson]
68 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
69 for corner cases. (Certain input points at infinity could lead to
70 bogus results, with non-infinity inputs mapped to infinity too.)
73 Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
75 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
76 handshake can force the use of weak keying material in OpenSSL
77 SSL/TLS clients and servers.
79 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
80 researching this issue. (CVE-2014-0224)
81 [KIKUCHI Masashi, Steve Henson]
83 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
84 OpenSSL DTLS client the code can be made to recurse eventually crashing
87 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
89 [Imre Rad, Steve Henson]
91 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
92 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
93 client or server. This is potentially exploitable to run arbitrary
94 code on a vulnerable client or server.
96 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
97 [Jüri Aedla, Steve Henson]
99 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
100 are subject to a denial of service attack.
102 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
103 this issue. (CVE-2014-3470)
104 [Felix Gröbert, Ivan Fratric, Steve Henson]
106 *) Harmonize version and its documentation. -f flag is used to display
108 [mancha <mancha1@zoho.com>]
110 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
112 [mancha <mancha1@zoho.com>]
114 *) Fix some double frees. These are not thought to be exploitable.
115 [mancha <mancha1@zoho.com>]
117 *) Fix for the attack described in the paper "Recovering OpenSSL
118 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
119 by Yuval Yarom and Naomi Benger. Details can be obtained from:
120 http://eprint.iacr.org/2014/140
122 Thanks to Yuval Yarom and Naomi Benger for discovering this
123 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
124 [Yuval Yarom and Naomi Benger]
126 Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
128 *) Keep original DTLS digest and encryption contexts in retransmission
129 structures so we can use the previous session parameters if they need
130 to be resent. (CVE-2013-6450)
133 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
134 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
135 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
136 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
137 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
138 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
139 [Rob Stradling, Adam Langley]
141 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
143 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
145 This addresses the flaw in CBC record processing discovered by
146 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
147 at: http://www.isg.rhul.ac.uk/tls/
149 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
150 Security Group at Royal Holloway, University of London
151 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
152 Emilia Käsper for the initial patch.
154 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
156 *) Return an error when checking OCSP signatures when key is NULL.
157 This fixes a DoS attack. (CVE-2013-0166)
160 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
161 the right response is stapled. Also change SSL_get_certificate()
162 so it returns the certificate actually sent.
163 See http://rt.openssl.org/Ticket/Display.html?id=2836.
165 [Rob Stradling <rob.stradling@comodo.com>]
167 *) Fix possible deadlock when decoding public keys.
170 Changes between 1.0.0i and 1.0.0j [10 May 2012]
172 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
175 *) Sanity check record length before skipping explicit IV in DTLS
178 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
179 fuzzing as a service testing platform.
183 *) Initialise tkeylen properly when encrypting CMS messages.
184 Thanks to Solar Designer of Openwall for reporting this issue.
187 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
189 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
190 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
191 in CRYPTO_realloc_clean.
193 Thanks to Tavis Ormandy, Google Security Team, for discovering this
194 issue and to Adam Langley <agl@chromium.org> for fixing it.
196 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
198 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
200 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
201 in CMS and PKCS7 code. When RSA decryption fails use a random key for
202 content decryption and always return the same error. Note: this attack
203 needs on average 2^20 messages so it only affects automated senders. The
204 old behaviour can be reenabled in the CMS code by setting the
205 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
206 an MMA defence is not necessary.
207 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
208 this issue. (CVE-2012-0884)
211 *) Fix CVE-2011-4619: make sure we really are receiving a
212 client hello before rejecting multiple SGC restarts. Thanks to
213 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
216 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
218 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
219 Thanks to Antonio Martin, Enterprise Secure Access Research and
220 Development, Cisco Systems, Inc. for discovering this bug and
221 preparing a fix. (CVE-2012-0050)
224 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
226 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
227 of the Vaudenay padding oracle attack on CBC mode encryption
228 which enables an efficient plaintext recovery attack against
229 the OpenSSL implementation of DTLS. Their attack exploits timing
230 differences arising during decryption processing. A research
231 paper describing this attack can be found at:
232 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
233 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
234 Security Group at Royal Holloway, University of London
235 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
236 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
237 for preparing the fix. (CVE-2011-4108)
238 [Robin Seggelmann, Michael Tuexen]
240 *) Clear bytes used for block padding of SSL 3.0 records.
242 [Adam Langley (Google)]
244 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
245 Kadianakis <desnacked@gmail.com> for discovering this issue and
246 Adam Langley for preparing the fix. (CVE-2011-4619)
247 [Adam Langley (Google)]
249 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
250 [Andrey Kulikov <amdeich@gmail.com>]
252 *) Prevent malformed RFC3779 data triggering an assertion failure.
253 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
254 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
255 [Rob Austein <sra@hactrn.net>]
257 *) Improved PRNG seeding for VOS.
258 [Paul Green <Paul.Green@stratus.com>]
260 *) Fix ssl_ciph.c set-up race.
261 [Adam Langley (Google)]
263 *) Fix spurious failures in ecdsatest.c.
264 [Emilia Käsper (Google)]
266 *) Fix the BIO_f_buffer() implementation (which was mixing different
267 interpretations of the '..._len' fields).
268 [Adam Langley (Google)]
270 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
271 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
272 threads won't reuse the same blinding coefficients.
274 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
275 lock to call BN_BLINDING_invert_ex, and avoids one use of
276 BN_BLINDING_update for each BN_BLINDING structure (previously,
277 the last update always remained unused).
278 [Emilia Käsper (Google)]
280 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
281 [Bob Buckholz (Google)]
283 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
285 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
286 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
287 [Kaspar Brand <ossl@velox.ch>]
289 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
290 for multi-threaded use of ECDH. (CVE-2011-3210)
291 [Adam Langley (Google)]
293 *) Fix x509_name_ex_d2i memory leak on bad inputs.
296 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
297 signature public key algorithm by using OID xref utilities instead.
298 Before this you could only use some ECC ciphersuites with SHA1 only.
301 *) Add protection against ECDSA timing attacks as mentioned in the paper
302 by Billy Bob Brumley and Nicola Tuveri, see:
304 http://eprint.iacr.org/2011/232.pdf
306 [Billy Bob Brumley and Nicola Tuveri]
308 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
310 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
311 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
313 *) Fix bug in string printing code: if *any* escaping is enabled we must
314 escape the escape character (backslash) or the resulting string is
318 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
320 *) Disable code workaround for ancient and obsolete Netscape browsers
321 and servers: an attacker can use it in a ciphersuite downgrade attack.
322 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
325 *) Fixed J-PAKE implementation error, originally discovered by
326 Sebastien Martini, further info and confirmation from Stefan
327 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
330 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
332 *) Fix extension code to avoid race conditions which can result in a buffer
333 overrun vulnerability: resumed sessions must not be modified as they can
334 be shared by multiple threads. CVE-2010-3864
337 *) Fix WIN32 build system to correctly link an ENGINE directory into
341 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
343 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
345 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
347 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
349 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
350 context. The operation can be customised via the ctrl mechanism in
351 case ENGINEs want to include additional functionality.
354 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
357 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
358 output hashes compatible with older versions of OpenSSL.
359 [Willy Weisz <weisz@vcpc.univie.ac.at>]
361 *) Fix compression algorithm handling: if resuming a session use the
362 compression algorithm of the resumed session instead of determining
363 it from client hello again. Don't allow server to change algorithm.
366 *) Add load_crls() function to apps tidying load_certs() too. Add option
367 to verify utility to allow additional CRLs to be included.
370 *) Update OCSP request code to permit adding custom headers to the request:
371 some responders need this.
374 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
376 [Julia Lawall <julia@diku.dk>]
378 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
379 needlessly dereferenced structures, used obsolete functions and
380 didn't handle all updated verify codes correctly.
383 *) Disable MD2 in the default configuration.
386 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
387 indicate the initial BIO being pushed or popped. This makes it possible
388 to determine whether the BIO is the one explicitly called or as a result
389 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
390 it handles reference counts correctly and doesn't zero out the I/O bio
391 when it is not being explicitly popped. WARNING: applications which
392 included workarounds for the old buggy behaviour will need to be modified
393 or they could free up already freed BIOs.
396 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
397 renaming to all platforms (within the 0.9.8 branch, this was
398 done conditionally on Netware platforms to avoid a name clash).
399 [Guenter <lists@gknw.net>]
401 *) Add ECDHE and PSK support to DTLS.
402 [Michael Tuexen <tuexen@fh-muenster.de>]
404 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
408 *) Add "missing" function EVP_MD_flags() (without this the only way to
409 retrieve a digest flags is by accessing the structure directly. Update
410 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
411 or cipher is registered as in the "from" argument. Print out all
412 registered digests in the dgst usage message instead of manually
413 attempting to work them out.
416 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
417 this allows the use of compression and extensions. Change default cipher
418 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
419 by default unless an application cipher string requests it.
422 *) Alter match criteria in PKCS12_parse(). It used to try to use local
423 key ids to find matching certificates and keys but some PKCS#12 files
424 don't follow the (somewhat unwritten) rules and this strategy fails.
425 Now just gather all certificates together and the first private key
426 then look for the first certificate that matches the key.
429 *) Support use of registered digest and cipher names for dgst and cipher
430 commands instead of having to add each one as a special case. So now
437 openssl dgst -sha256 foo
439 and this works for ENGINE based algorithms too.
443 *) Update Gost ENGINE to support parameter files.
444 [Victor B. Wagner <vitus@cryptocom.ru>]
446 *) Support GeneralizedTime in ca utility.
447 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
449 *) Enhance the hash format used for certificate directory links. The new
450 form uses the canonical encoding (meaning equivalent names will work
451 even if they aren't identical) and uses SHA1 instead of MD5. This form
452 is incompatible with the older format and as a result c_rehash should
453 be used to rebuild symbolic links.
456 *) Make PKCS#8 the default write format for private keys, replacing the
457 traditional format. This form is standardised, more secure and doesn't
458 include an implicit MD5 dependency.
461 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
462 committed to OpenSSL should pass this lot as a minimum.
465 *) Add session ticket override functionality for use by EAP-FAST.
466 [Jouni Malinen <j@w1.fi>]
468 *) Modify HMAC functions to return a value. Since these can be implemented
469 in an ENGINE errors can occur.
472 *) Type-checked OBJ_bsearch_ex.
475 *) Type-checked OBJ_bsearch. Also some constification necessitated
476 by type-checking. Still to come: TXT_DB, bsearch(?),
477 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
481 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
482 seconds to a tm structure directly, instead of going through OS
483 specific date routines. This avoids any issues with OS routines such
484 as the year 2038 bug. New *_adj() functions for ASN1 time structures
485 and X509_time_adj_ex() to cover the extended range. The existing
486 X509_time_adj() is still usable and will no longer have any date issues.
489 *) Delta CRL support. New use deltas option which will attempt to locate
490 and search any appropriate delta CRLs available.
492 This work was sponsored by Google.
495 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
496 code and add additional score elements. Validate alternate CRL paths
497 as part of the CRL checking and indicate a new error "CRL path validation
498 error" in this case. Applications wanting additional details can use
499 the verify callback and check the new "parent" field. If this is not
500 NULL CRL path validation is taking place. Existing applications wont
501 see this because it requires extended CRL support which is off by
504 This work was sponsored by Google.
507 *) Support for freshest CRL extension.
509 This work was sponsored by Google.
512 *) Initial indirect CRL support. Currently only supported in the CRLs
513 passed directly and not via lookup. Process certificate issuer
514 CRL entry extension and lookup CRL entries by bother issuer name
515 and serial number. Check and process CRL issuer entry in IDP extension.
517 This work was sponsored by Google.
520 *) Add support for distinct certificate and CRL paths. The CRL issuer
521 certificate is validated separately in this case. Only enabled if
522 an extended CRL support flag is set: this flag will enable additional
523 CRL functionality in future.
525 This work was sponsored by Google.
528 *) Add support for policy mappings extension.
530 This work was sponsored by Google.
533 *) Fixes to pathlength constraint, self issued certificate handling,
534 policy processing to align with RFC3280 and PKITS tests.
536 This work was sponsored by Google.
539 *) Support for name constraints certificate extension. DN, email, DNS
540 and URI types are currently supported.
542 This work was sponsored by Google.
545 *) To cater for systems that provide a pointer-based thread ID rather
546 than numeric, deprecate the current numeric thread ID mechanism and
547 replace it with a structure and associated callback type. This
548 mechanism allows a numeric "hash" to be extracted from a thread ID in
549 either case, and on platforms where pointers are larger than 'long',
550 mixing is done to help ensure the numeric 'hash' is usable even if it
551 can't be guaranteed unique. The default mechanism is to use "&errno"
552 as a pointer-based thread ID to distinguish between threads.
554 Applications that want to provide their own thread IDs should now use
555 CRYPTO_THREADID_set_callback() to register a callback that will call
556 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
558 Note that ERR_remove_state() is now deprecated, because it is tied
559 to the assumption that thread IDs are numeric. ERR_remove_state(0)
560 to free the current thread's error state should be replaced by
561 ERR_remove_thread_state(NULL).
563 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
564 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
565 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
566 application was previously providing a numeric thread callback that
567 was inappropriate for distinguishing threads, then uniqueness might
568 have been obtained with &errno that happened immediately in the
569 intermediate development versions of OpenSSL; this is no longer the
570 case, the numeric thread callback will now override the automatic use
572 [Geoff Thorpe, with help from Bodo Moeller]
574 *) Initial support for different CRL issuing certificates. This covers a
575 simple case where the self issued certificates in the chain exist and
576 the real CRL issuer is higher in the existing chain.
578 This work was sponsored by Google.
581 *) Removed effectively defunct crypto/store from the build.
584 *) Revamp of STACK to provide stronger type-checking. Still to come:
585 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
586 ASN1_STRING, CONF_VALUE.
589 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
590 RAM on SSL connections. This option can save about 34k per idle SSL.
593 *) Revamp of LHASH to provide stronger type-checking. Still to come:
594 STACK, TXT_DB, bsearch, qsort.
597 *) Initial support for Cryptographic Message Syntax (aka CMS) based
598 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
599 support for data, signedData, compressedData, digestedData and
600 encryptedData, envelopedData types included. Scripts to check against
601 RFC4134 examples draft and interop and consistency checks of many
602 content types and variants.
605 *) Add options to enc utility to support use of zlib compression BIO.
608 *) Extend mk1mf to support importing of options and assembly language
609 files from Configure script, currently only included in VC-WIN32.
610 The assembly language rules can now optionally generate the source
611 files from the associated perl scripts.
614 *) Implement remaining functionality needed to support GOST ciphersuites.
615 Interop testing has been performed using CryptoPro implementations.
616 [Victor B. Wagner <vitus@cryptocom.ru>]
618 *) s390x assembler pack.
621 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
625 *) Implement Opaque PRF Input TLS extension as specified in
626 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
627 official specification yet and no extension type assignment by
628 IANA exists, this extension (for now) will have to be explicitly
629 enabled when building OpenSSL by providing the extension number
630 to use. For example, specify an option
632 -DTLSEXT_TYPE_opaque_prf_input=0x9527
634 to the "config" or "Configure" script to enable the extension,
635 assuming extension number 0x9527 (which is a completely arbitrary
636 and unofficial assignment based on the MD5 hash of the Internet
637 Draft). Note that by doing so, you potentially lose
638 interoperability with other TLS implementations since these might
639 be using the same extension number for other purposes.
641 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
642 opaque PRF input value to use in the handshake. This will create
643 an interal copy of the length-'len' string at 'src', and will
644 return non-zero for success.
646 To get more control and flexibility, provide a callback function
649 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
650 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
654 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
657 Callback function 'cb' will be called in handshakes, and is
658 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
659 Argument 'arg' is for application purposes (the value as given to
660 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
661 be provided to the callback function). The callback function
662 has to return non-zero to report success: usually 1 to use opaque
663 PRF input just if possible, or 2 to enforce use of the opaque PRF
664 input. In the latter case, the library will abort the handshake
665 if opaque PRF input is not successfully negotiated.
667 Arguments 'peerinput' and 'len' given to the callback function
668 will always be NULL and 0 in the case of a client. A server will
669 see the client's opaque PRF input through these variables if
670 available (NULL and 0 otherwise). Note that if the server
671 provides an opaque PRF input, the length must be the same as the
672 length of the client's opaque PRF input.
674 Note that the callback function will only be called when creating
675 a new session (session resumption can resume whatever was
676 previously negotiated), and will not be called in SSL 2.0
677 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
678 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
679 for applications that need to enforce opaque PRF input.
683 *) Update ssl code to support digests other than SHA1+MD5 for handshake
686 [Victor B. Wagner <vitus@cryptocom.ru>]
688 *) Add RFC4507 support to OpenSSL. This includes the corrections in
689 RFC4507bis. The encrypted ticket format is an encrypted encoded
690 SSL_SESSION structure, that way new session features are automatically
693 If a client application caches session in an SSL_SESSION structure
694 support is transparent because tickets are now stored in the encoded
697 The SSL_CTX structure automatically generates keys for ticket
698 protection in servers so again support should be possible
699 with no application modification.
701 If a client or server wishes to disable RFC4507 support then the option
702 SSL_OP_NO_TICKET can be set.
704 Add a TLS extension debugging callback to allow the contents of any client
705 or server extensions to be examined.
707 This work was sponsored by Google.
710 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
711 OpenSSL should now compile cleanly on gcc 4.2
712 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
714 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
715 support including streaming MAC support: this is required for GOST
717 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
719 *) Add option -stream to use PKCS#7 streaming in smime utility. New
720 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
721 to output in BER and PEM format.
724 *) Experimental support for use of HMAC via EVP_PKEY interface. This
725 allows HMAC to be handled via the EVP_DigestSign*() interface. The
726 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
727 ENGINE support for HMAC keys which are unextractable. New -mac and
728 -macopt options to dgst utility.
731 *) New option -sigopt to dgst utility. Update dgst to use
732 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
733 alternative signing paramaters such as X9.31 or PSS in the dgst
737 *) Change ssl_cipher_apply_rule(), the internal function that does
738 the work each time a ciphersuite string requests enabling
739 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
740 removing ("!foo+bar") a class of ciphersuites: Now it maintains
741 the order of disabled ciphersuites such that those ciphersuites
742 that most recently went from enabled to disabled not only stay
743 in order with respect to each other, but also have higher priority
744 than other disabled ciphersuites the next time ciphersuites are
747 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
748 the same ciphersuites as with "HIGH" alone, but in a specific
749 order where the PSK ciphersuites come first (since they are the
750 most recently disabled ciphersuites when "HIGH" is parsed).
752 Also, change ssl_create_cipher_list() (using this new
753 funcionality) such that between otherwise identical
754 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
758 *) Change ssl_create_cipher_list() so that it automatically
759 arranges the ciphersuites in reasonable order before starting
760 to process the rule string. Thus, the definition for "DEFAULT"
761 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
762 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
763 This makes it much easier to arrive at a reasonable default order
764 in applications for which anonymous ciphers are OK (meaning
765 that you can't actually use DEFAULT).
766 [Bodo Moeller; suggested by Victor Duchovni]
768 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
769 processing) into multiple integers instead of setting
770 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
771 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
772 (These masks as well as the individual bit definitions are hidden
773 away into the non-exported interface ssl/ssl_locl.h, so this
774 change to the definition of the SSL_CIPHER structure shouldn't
775 affect applications.) This give us more bits for each of these
776 categories, so there is no longer a need to coagulate AES128 and
777 AES256 into a single algorithm bit, and to coagulate Camellia128
778 and Camellia256 into a single algorithm bit, which has led to all
781 Thus, among other things, the kludge introduced in 0.9.7m and
782 0.9.8e for masking out AES256 independently of AES128 or masking
783 out Camellia256 independently of AES256 is not needed here in 0.9.9.
785 With the change, we also introduce new ciphersuite aliases that
786 so far were missing: "AES128", "AES256", "CAMELLIA128", and
790 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
791 Use the leftmost N bytes of the signature input if the input is
792 larger than the prime q (with N being the size in bytes of q).
795 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
796 it yet and it is largely untested.
799 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
802 *) Initial incomplete changes to avoid need for function casts in OpenSSL
803 some compilers (gcc 4.2 and later) reject their use. Safestack is
804 reimplemented. Update ASN1 to avoid use of legacy functions.
807 *) Win32/64 targets are linked with Winsock2.
810 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
811 to external functions. This can be used to increase CRL handling
812 efficiency especially when CRLs are very large by (for example) storing
813 the CRL revoked certificates in a database.
816 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
817 new CRLs added to a directory can be used. New command line option
818 -verify_return_error to s_client and s_server. This causes real errors
819 to be returned by the verify callback instead of carrying on no matter
820 what. This reflects the way a "real world" verify callback would behave.
823 *) GOST engine, supporting several GOST algorithms and public key formats.
824 Kindly donated by Cryptocom.
827 *) Partial support for Issuing Distribution Point CRL extension. CRLs
828 partitioned by DP are handled but no indirect CRL or reason partitioning
829 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
830 selected via a scoring technique which handles IDP and AKID in CRLs.
833 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
834 will ultimately be used for all verify operations: this will remove the
835 X509_STORE dependency on certificate verification and allow alternative
836 lookup methods. X509_STORE based implementations of these two callbacks.
839 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
840 Modify get_crl() to find a valid (unexpired) CRL if possible.
843 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
844 this would be called X509_CRL_cmp() but that name is already used by
845 a function that just compares CRL issuer names. Cache several CRL
846 extensions in X509_CRL structure and cache CRLDP in X509.
849 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
850 this maps equivalent X509_NAME structures into a consistent structure.
851 Name comparison can then be performed rapidly using memcmp().
854 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
858 *) Allow digests to supply their own micalg string for S/MIME type using
859 the ctrl EVP_MD_CTRL_MICALG.
862 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
863 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
864 ctrl. It can then customise the structure before and/or after signing
868 *) New function OBJ_add_sigid() to allow application defined signature OIDs
869 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
870 to free up any added signature OIDs.
873 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
874 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
875 digest and cipher tables. New options added to openssl utility:
876 list-message-digest-algorithms and list-cipher-algorithms.
879 *) Change the array representation of binary polynomials: the list
880 of degrees of non-zero coefficients is now terminated with -1.
881 Previously it was terminated with 0, which was also part of the
882 value; thus, the array representation was not applicable to
883 polynomials where t^0 has coefficient zero. This change makes
884 the array representation useful in a more general context.
887 *) Various modifications and fixes to SSL/TLS cipher string
888 handling. For ECC, the code now distinguishes between fixed ECDH
889 with RSA certificates on the one hand and with ECDSA certificates
890 on the other hand, since these are separate ciphersuites. The
891 unused code for Fortezza ciphersuites has been removed.
893 For consistency with EDH, ephemeral ECDH is now called "EECDH"
894 (not "ECDHE"). For consistency with the code for DH
895 certificates, use of ECDH certificates is now considered ECDH
896 authentication, not RSA or ECDSA authentication (the latter is
897 merely the CA's signing algorithm and not actively used in the
900 The temporary ciphersuite alias "ECCdraft" is no longer
901 available, and ECC ciphersuites are no longer excluded from "ALL"
902 and "DEFAULT". The following aliases now exist for RFC 4492
903 ciphersuites, most of these by analogy with the DH case:
905 kECDHr - ECDH cert, signed with RSA
906 kECDHe - ECDH cert, signed with ECDSA
907 kECDH - ECDH cert (signed with either RSA or ECDSA)
908 kEECDH - ephemeral ECDH
909 ECDH - ECDH cert or ephemeral ECDH
915 AECDH - anonymous ECDH
916 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
920 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
921 Use correct micalg parameters depending on digest(s) in signed message.
924 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
925 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
928 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
929 an engine to register a method. Add ENGINE lookups for methods and
930 functional reference processing.
933 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
934 EVP_{Sign,Verify}* which allow an application to customise the signature
938 *) New -resign option to smime utility. This adds one or more signers
939 to an existing PKCS#7 signedData structure. Also -md option to use an
940 alternative message digest algorithm for signing.
943 *) Tidy up PKCS#7 routines and add new functions to make it easier to
944 create PKCS7 structures containing multiple signers. Update smime
945 application to support multiple signers.
948 *) New -macalg option to pkcs12 utility to allow setting of an alternative
952 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
953 Reorganize PBE internals to lookup from a static table using NIDs,
954 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
955 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
956 PRF which will be automatically used with PBES2.
959 *) Replace the algorithm specific calls to generate keys in "req" with the
963 *) Update PKCS#7 enveloped data routines to use new API. This is now
964 supported by any public key method supporting the encrypt operation. A
965 ctrl is added to allow the public key algorithm to examine or modify
966 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
970 *) Add a ctrl to asn1 method to allow a public key algorithm to express
971 a default digest type to use. In most cases this will be SHA1 but some
972 algorithms (such as GOST) need to specify an alternative digest. The
973 return value indicates how strong the prefernce is 1 means optional and
974 2 is mandatory (that is it is the only supported type). Modify
975 ASN1_item_sign() to accept a NULL digest argument to indicate it should
976 use the default md. Update openssl utilities to use the default digest
977 type for signing if it is not explicitly indicated.
980 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
981 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
982 signing method from the key type. This effectively removes the link
983 between digests and public key types.
986 *) Add an OID cross reference table and utility functions. Its purpose is to
987 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
988 rsaEncryption. This will allow some of the algorithm specific hackery
989 needed to use the correct OID to be removed.
992 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
993 structures for PKCS7_sign(). They are now set up by the relevant public
997 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1000 *) Add support for key derivation (agreement) in the API, DH method and
1004 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1005 public and private key formats. As a side effect these add additional
1006 command line functionality not previously available: DSA signatures can be
1007 generated and verified using pkeyutl and DH key support and generation in
1012 [Oliver Tappe <zooey@hirschkaefer.de>]
1014 *) New make target "install_html_docs" installs HTML renditions of the
1016 [Oliver Tappe <zooey@hirschkaefer.de>]
1018 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1019 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1020 support key and parameter generation and add initial key generation
1021 functionality for RSA.
1024 *) Add functions for main EVP_PKEY_method operations. The undocumented
1025 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1026 EVP_PKEY_{encrypt,decrypt}_old.
1029 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1030 key API, doesn't do much yet.
1033 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1034 public key algorithms. New option to openssl utility:
1035 "list-public-key-algorithms" to print out info.
1038 *) Implement the Supported Elliptic Curves Extension for
1039 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1042 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1043 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1046 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1047 utilities such as rsa, dsa, dsaparam etc except they process any key
1051 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1052 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1053 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1057 *) Initial support for pluggable public key ASN1.
1058 De-spaghettify the public key ASN1 handling. Move public and private
1059 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1060 algorithm specific handling to a single module within the relevant
1061 algorithm directory. Add functions to allow (near) opaque processing
1062 of public and private key structures.
1065 *) Implement the Supported Point Formats Extension for
1066 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1069 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1070 for the psk identity [hint] and the psk callback functions to the
1071 SSL_SESSION, SSL and SSL_CTX structure.
1074 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1078 SSL_CTX_use_psk_identity_hint
1079 SSL_get_psk_identity_hint
1080 SSL_get_psk_identity
1081 SSL_use_psk_identity_hint
1083 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1085 *) Add RFC 3161 compliant time stamp request creation, response generation
1086 and response verification functionality.
1087 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1089 *) Add initial support for TLS extensions, specifically for the server_name
1090 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1091 have new members for a host name. The SSL data structure has an
1092 additional member SSL_CTX *initial_ctx so that new sessions can be
1093 stored in that context to allow for session resumption, even after the
1094 SSL has been switched to a new SSL_CTX in reaction to a client's
1095 server_name extension.
1097 New functions (subject to change):
1099 SSL_get_servername()
1100 SSL_get_servername_type()
1103 New CTRL codes and macros (subject to change):
1105 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1106 - SSL_CTX_set_tlsext_servername_callback()
1107 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1108 - SSL_CTX_set_tlsext_servername_arg()
1109 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1111 openssl s_client has a new '-servername ...' option.
1113 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1114 '-key2 ...', '-servername_fatal' (subject to change). This allows
1115 testing the HostName extension for a specific single host name ('-cert'
1116 and '-key' remain fallbacks for handshakes without HostName
1117 negotiation). If the unrecogninzed_name alert has to be sent, this by
1118 default is a warning; it becomes fatal with the '-servername_fatal'
1121 [Peter Sylvester, Remy Allais, Christophe Renou]
1123 *) Whirlpool hash implementation is added.
1126 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1127 bn(64,32). Because of instruction set limitations it doesn't have
1128 any negative impact on performance. This was done mostly in order
1129 to make it possible to share assembler modules, such as bn_mul_mont
1130 implementations, between 32- and 64-bit builds without hassle.
1133 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1134 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1138 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1139 dedicated Montgomery multiplication procedure, is introduced.
1140 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1141 "64-bit" performance on certain 32-bit targets.
1144 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1145 in SSL structures. New SSL ctrl to set maximum send fragment size.
1146 Save memory by seeting the I/O buffer sizes dynamically instead of
1147 using the maximum available value.
1150 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1151 in addition to the text details.
1154 *) Very, very preliminary EXPERIMENTAL support for printing of general
1155 ASN1 structures. This currently produces rather ugly output and doesn't
1156 handle several customised structures at all.
1159 *) Integrated support for PVK file format and some related formats such
1160 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1161 these in the 'rsa' and 'dsa' utilities.
1164 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1167 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1168 place for the (very old) "NETSCAPE" format certificates which are now
1169 handled using new ASN1 code equivalents.
1172 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1173 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1174 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1177 *) Modify CRL distribution points extension code to print out previously
1178 unsupported fields. Enhance extension setting code to allow setting of
1182 *) Add print and set support for Issuing Distribution Point CRL extension.
1185 *) Change 'Configure' script to enable Camellia by default.
1188 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
1190 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1192 This addresses the flaw in CBC record processing discovered by
1193 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1194 at: http://www.isg.rhul.ac.uk/tls/
1196 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1197 Security Group at Royal Holloway, University of London
1198 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1199 Emilia Käsper for the initial patch.
1201 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1203 *) Return an error when checking OCSP signatures when key is NULL.
1204 This fixes a DoS attack. (CVE-2013-0166)
1207 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1208 the right response is stapled. Also change SSL_get_certificate()
1209 so it returns the certificate actually sent.
1210 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1211 (This is a backport)
1212 [Rob Stradling <rob.stradling@comodo.com>]
1214 *) Fix possible deadlock when decoding public keys.
1217 Changes between 0.9.8w and 0.9.8x [10 May 2012]
1219 *) Sanity check record length before skipping explicit IV in DTLS
1222 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1223 fuzzing as a service testing platform.
1227 *) Initialise tkeylen properly when encrypting CMS messages.
1228 Thanks to Solar Designer of Openwall for reporting this issue.
1231 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
1233 *) The fix for CVE-2012-2110 did not take into account that the
1234 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
1235 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
1236 rejecting negative len parameter. (CVE-2012-2131)
1237 [Tomas Hoger <thoger@redhat.com>]
1239 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
1241 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1242 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1243 in CRYPTO_realloc_clean.
1245 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1246 issue and to Adam Langley <agl@chromium.org> for fixing it.
1248 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1250 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
1252 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1253 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1254 content decryption and always return the same error. Note: this attack
1255 needs on average 2^20 messages so it only affects automated senders. The
1256 old behaviour can be reenabled in the CMS code by setting the
1257 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1258 an MMA defence is not necessary.
1259 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1260 this issue. (CVE-2012-0884)
1263 *) Fix CVE-2011-4619: make sure we really are receiving a
1264 client hello before rejecting multiple SGC restarts. Thanks to
1265 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1268 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1270 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1271 Thanks to Antonio Martin, Enterprise Secure Access Research and
1272 Development, Cisco Systems, Inc. for discovering this bug and
1273 preparing a fix. (CVE-2012-0050)
1276 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1278 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1279 of the Vaudenay padding oracle attack on CBC mode encryption
1280 which enables an efficient plaintext recovery attack against
1281 the OpenSSL implementation of DTLS. Their attack exploits timing
1282 differences arising during decryption processing. A research
1283 paper describing this attack can be found at:
1284 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1285 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1286 Security Group at Royal Holloway, University of London
1287 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1288 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1289 for preparing the fix. (CVE-2011-4108)
1290 [Robin Seggelmann, Michael Tuexen]
1292 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1293 [Ben Laurie, Kasper <ekasper@google.com>]
1295 *) Clear bytes used for block padding of SSL 3.0 records.
1297 [Adam Langley (Google)]
1299 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1300 Kadianakis <desnacked@gmail.com> for discovering this issue and
1301 Adam Langley for preparing the fix. (CVE-2011-4619)
1302 [Adam Langley (Google)]
1304 *) Prevent malformed RFC3779 data triggering an assertion failure.
1305 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1306 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1307 [Rob Austein <sra@hactrn.net>]
1309 *) Fix ssl_ciph.c set-up race.
1310 [Adam Langley (Google)]
1312 *) Fix spurious failures in ecdsatest.c.
1313 [Emilia Käsper (Google)]
1315 *) Fix the BIO_f_buffer() implementation (which was mixing different
1316 interpretations of the '..._len' fields).
1317 [Adam Langley (Google)]
1319 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1320 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1321 threads won't reuse the same blinding coefficients.
1323 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1324 lock to call BN_BLINDING_invert_ex, and avoids one use of
1325 BN_BLINDING_update for each BN_BLINDING structure (previously,
1326 the last update always remained unused).
1327 [Emilia Käsper (Google)]
1329 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1330 for multi-threaded use of ECDH.
1331 [Adam Langley (Google)]
1333 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1336 *) Add protection against ECDSA timing attacks as mentioned in the paper
1337 by Billy Bob Brumley and Nicola Tuveri, see:
1339 http://eprint.iacr.org/2011/232.pdf
1341 [Billy Bob Brumley and Nicola Tuveri]
1343 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1345 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1346 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1348 *) Fix bug in string printing code: if *any* escaping is enabled we must
1349 escape the escape character (backslash) or the resulting string is
1353 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1355 *) Disable code workaround for ancient and obsolete Netscape browsers
1356 and servers: an attacker can use it in a ciphersuite downgrade attack.
1357 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1360 *) Fixed J-PAKE implementation error, originally discovered by
1361 Sebastien Martini, further info and confirmation from Stefan
1362 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1365 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1367 *) Fix extension code to avoid race conditions which can result in a buffer
1368 overrun vulnerability: resumed sessions must not be modified as they can
1369 be shared by multiple threads. CVE-2010-3864
1372 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1375 *) Don't reencode certificate when calculating signature: cache and use
1376 the original encoding instead. This makes signature verification of
1377 some broken encodings work correctly.
1380 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1381 is also one of the inputs.
1382 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1384 *) Don't repeatedly append PBE algorithms to table if they already exist.
1385 Sort table on each new add. This effectively makes the table read only
1386 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1390 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1392 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1395 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1396 access or freeing data twice (CVE-2010-0742)
1397 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1399 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1400 common in certificates and some applications which only call
1401 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1405 Reduce copying into .apps and .test in makevms.com
1406 Don't try to use blank CA certificate in CA.com
1407 Allow use of C files from original directories in maketests.com
1408 [Steven M. Schweda" <sms@antinode.info>]
1410 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1412 *) When rejecting SSL/TLS records due to an incorrect version number, never
1413 update s->server with a new major version number. As of
1414 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1415 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1416 the previous behavior could result in a read attempt at NULL when
1417 receiving specific incorrect SSL/TLS records once record payload
1418 protection is active. (CVE-2010-0740)
1419 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1421 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1422 could be crashed if the relevant tables were not present (e.g. chrooted).
1423 [Tomas Hoger <thoger@redhat.com>]
1425 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1427 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1428 [Martin Olsson, Neel Mehta]
1430 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1431 accommodate for stack sorting, always a write lock!).
1434 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1435 excessive delays in the RAND_poll(): over a minute. As a workaround
1436 include a time check in the inner Heap32Next loop too.
1439 *) The code that handled flushing of data in SSL/TLS originally used the
1440 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1441 the problem outlined in PR#1949. The fix suggested there however can
1442 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1443 of Apache). So instead simplify the code to flush unconditionally.
1444 This should be fine since flushing with no data to flush is a no op.
1447 *) Handle TLS versions 2.0 and later properly and correctly use the
1448 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1449 off ancient servers have a habit of sticking around for a while...
1452 *) Modify compression code so it frees up structures without using the
1453 ex_data callbacks. This works around a problem where some applications
1454 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1455 restarting) then use compression (e.g. SSL with compression) later.
1456 This results in significant per-connection memory leaks and
1457 has caused some security issues including CVE-2008-1678 and
1461 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1462 change when encrypting or decrypting.
1465 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1466 connect and renegotiate with servers which do not support RI.
1467 Until RI is more widely deployed this option is enabled by default.
1470 *) Add "missing" ssl ctrls to clear options and mode.
1473 *) If client attempts to renegotiate and doesn't support RI respond with
1474 a no_renegotiation alert as required by RFC5746. Some renegotiating
1475 TLS clients will continue a connection gracefully when they receive
1476 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1477 waiting for a server hello which it will never receive. Now we treat a
1478 received no_renegotiation alert as a fatal error. This is because
1479 applications requesting a renegotiation might well expect it to succeed
1480 and would have no code in place to handle the server denying it so the
1481 only safe thing to do is to terminate the connection.
1484 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1485 peer supports secure renegotiation and 0 otherwise. Print out peer
1486 renegotiation support in s_client/s_server.
1489 *) Replace the highly broken and deprecated SPKAC certification method with
1490 the updated NID creation version. This should correctly handle UTF8.
1493 *) Implement RFC5746. Re-enable renegotiation but require the extension
1494 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1495 turns out to be a bad idea. It has been replaced by
1496 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1497 SSL_CTX_set_options(). This is really not recommended unless you
1498 know what you are doing.
1499 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1501 *) Fixes to stateless session resumption handling. Use initial_ctx when
1502 issuing and attempting to decrypt tickets in case it has changed during
1503 servername handling. Use a non-zero length session ID when attempting
1504 stateless session resumption: this makes it possible to determine if
1505 a resumption has occurred immediately after receiving server hello
1506 (several places in OpenSSL subtly assume this) instead of later in
1510 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1511 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1512 fixes for a few places where the return code is not checked
1514 [Julia Lawall <julia@diku.dk>]
1516 *) Add --strict-warnings option to Configure script to include devteam
1517 warnings in other configurations.
1520 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1521 makes it possible to install openssl libraries in locations which
1522 have names other than "lib", for example "/usr/lib64" which some
1524 [Steve Henson, based on patch from Jeremy Utley]
1526 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1527 X690 8.9.12 and can produce some misleading textual output of OIDs.
1528 [Steve Henson, reported by Dan Kaminsky]
1530 *) Delete MD2 from algorithm tables. This follows the recommendation in
1531 several standards that it is not used in new applications due to
1532 several cryptographic weaknesses. For binary compatibility reasons
1533 the MD2 API is still compiled in by default.
1536 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1540 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1541 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1543 [Guenter <lists@gknw.net>]
1545 *) Fix the server certificate chain building code to use X509_verify_cert(),
1546 it used to have an ad-hoc builder which was unable to cope with anything
1547 other than a simple chain.
1548 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1550 *) Don't check self signed certificate signatures in X509_verify_cert()
1551 by default (a flag can override this): it just wastes time without
1552 adding any security. As a useful side effect self signed root CAs
1553 with non-FIPS digests are now usable in FIPS mode.
1556 *) In dtls1_process_out_of_seq_message() the check if the current message
1557 is already buffered was missing. For every new message was memory
1558 allocated, allowing an attacker to perform an denial of service attack
1559 with sending out of seq handshake messages until there is no memory
1560 left. Additionally every future messege was buffered, even if the
1561 sequence number made no sense and would be part of another handshake.
1562 So only messages with sequence numbers less than 10 in advance will be
1563 buffered. (CVE-2009-1378)
1564 [Robin Seggelmann, discovered by Daniel Mentz]
1566 *) Records are buffered if they arrive with a future epoch to be
1567 processed after finishing the corresponding handshake. There is
1568 currently no limitation to this buffer allowing an attacker to perform
1569 a DOS attack with sending records with future epochs until there is no
1570 memory left. This patch adds the pqueue_size() function to detemine
1571 the size of a buffer and limits the record buffer to 100 entries.
1573 [Robin Seggelmann, discovered by Daniel Mentz]
1575 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1576 parent structure is freed. (CVE-2009-1379)
1579 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1580 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1583 [Ilya O. <vrghost@gmail.com>]
1585 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1587 *) Disable renegotiation completely - this fixes a severe security
1588 problem (CVE-2009-3555) at the cost of breaking all
1589 renegotiation. Renegotiation can be re-enabled by setting
1590 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1591 run-time. This is really not recommended unless you know what
1595 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1597 *) Don't set val to NULL when freeing up structures, it is freed up by
1598 underlying code. If sizeof(void *) > sizeof(long) this can result in
1599 zeroing past the valid field. (CVE-2009-0789)
1600 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1602 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1603 checked correctly. This would allow some invalid signed attributes to
1604 appear to verify correctly. (CVE-2009-0591)
1605 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1607 *) Reject UniversalString and BMPString types with invalid lengths. This
1608 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1609 a legal length. (CVE-2009-0590)
1612 *) Set S/MIME signing as the default purpose rather than setting it
1613 unconditionally. This allows applications to override it at the store
1617 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1618 to handle some structures.
1621 *) Improve efficiency of mem_gets: don't search whole buffer each time
1623 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1625 *) New -hex option for openssl rand.
1628 *) Print out UTF8String and NumericString when parsing ASN1.
1631 *) Support NumericString type for name components.
1634 *) Allow CC in the environment to override the automatically chosen
1635 compiler. Note that nothing is done to ensure flags work with the
1639 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1641 *) Properly check EVP_VerifyFinal() and similar return values
1643 [Ben Laurie, Bodo Moeller, Google Security Team]
1645 *) Enable TLS extensions by default.
1648 *) Allow the CHIL engine to be loaded, whether the application is
1649 multithreaded or not. (This does not release the developer from the
1650 obligation to set up the dynamic locking callbacks.)
1651 [Sander Temme <sander@temme.net>]
1653 *) Use correct exit code if there is an error in dgst command.
1654 [Steve Henson; problem pointed out by Roland Dirlewanger]
1656 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1657 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1660 *) Add experimental JPAKE support, including demo authentication in
1661 s_client and s_server.
1664 *) Set the comparison function in v3_addr_canonize().
1665 [Rob Austein <sra@hactrn.net>]
1667 *) Add support for XMPP STARTTLS in s_client.
1668 [Philip Paeps <philip@freebsd.org>]
1670 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1671 to ensure that even with this option, only ciphersuites in the
1672 server's preference list will be accepted. (Note that the option
1673 applies only when resuming a session, so the earlier behavior was
1674 just about the algorithm choice for symmetric cryptography.)
1677 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1679 *) Fix NULL pointer dereference if a DTLS server received
1680 ChangeCipherSpec as first record (CVE-2009-1386).
1683 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1684 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1687 *) The fix in 0.9.8c that supposedly got rid of unsafe
1688 double-checked locking was incomplete for RSA blinding,
1689 addressing just one layer of what turns out to have been
1690 doubly unsafe triple-checked locking.
1692 So now fix this for real by retiring the MONT_HELPER macro
1693 in crypto/rsa/rsa_eay.c.
1695 [Bodo Moeller; problem pointed out by Marius Schilder]
1697 *) Various precautionary measures:
1699 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1701 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1702 (NB: This would require knowledge of the secret session ticket key
1703 to exploit, in which case you'd be SOL either way.)
1705 - Change bn_nist.c so that it will properly handle input BIGNUMs
1706 outside the expected range.
1708 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1711 [Neel Mehta, Bodo Moeller]
1713 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1714 the load fails. Useful for distros.
1715 [Ben Laurie and the FreeBSD team]
1717 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1720 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1723 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1725 This work was sponsored by Logica.
1728 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1729 keystores. Support for SSL/TLS client authentication too.
1730 Not compiled unless enable-capieng specified to Configure.
1732 This work was sponsored by Logica.
1735 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1736 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1737 attribute creation routines such as certifcate requests and PKCS#12
1741 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1743 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1744 handshake which could lead to a cilent crash as found using the
1745 Codenomicon TLS test suite (CVE-2008-1672)
1746 [Steve Henson, Mark Cox]
1748 *) Fix double free in TLS server name extensions which could lead to
1749 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1752 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1754 Clear the error queue to ensure that error entries left from
1755 older function calls do not interfere with the correct operation.
1756 [Lutz Jaenicke, Erik de Castro Lopo]
1758 *) Remove root CA certificates of commercial CAs:
1760 The OpenSSL project does not recommend any specific CA and does not
1761 have any policy with respect to including or excluding any CA.
1762 Therefore it does not make any sense to ship an arbitrary selection
1763 of root CA certificates with the OpenSSL software.
1766 *) RSA OAEP patches to fix two separate invalid memory reads.
1767 The first one involves inputs when 'lzero' is greater than
1768 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1769 before the beginning of from). The second one involves inputs where
1770 the 'db' section contains nothing but zeroes (there is a one-byte
1771 invalid read after the end of 'db').
1772 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1774 *) Partial backport from 0.9.9-dev:
1776 Introduce bn_mul_mont (dedicated Montgomery multiplication
1777 procedure) as a candidate for BIGNUM assembler implementation.
1778 While 0.9.9-dev uses assembler for various architectures, only
1779 x86_64 is available by default here in the 0.9.8 branch, and
1780 32-bit x86 is available through a compile-time setting.
1782 To try the 32-bit x86 assembler implementation, use Configure
1783 option "enable-montasm" (which exists only for this backport).
1785 As "enable-montasm" for 32-bit x86 disclaims code stability
1786 anyway, in this constellation we activate additional code
1787 backported from 0.9.9-dev for further performance improvements,
1788 namely BN_from_montgomery_word. (To enable this otherwise,
1789 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1791 [Andy Polyakov (backport partially by Bodo Moeller)]
1793 *) Add TLS session ticket callback. This allows an application to set
1794 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1795 values. This is useful for key rollover for example where several key
1796 sets may exist with different names.
1799 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1800 This was broken until now in 0.9.8 releases, such that the only way
1801 a registered ENGINE could be used (assuming it initialises
1802 successfully on the host) was to explicitly set it as the default
1803 for the relevant algorithms. This is in contradiction with 0.9.7
1804 behaviour and the documentation. With this fix, when an ENGINE is
1805 registered into a given algorithm's table of implementations, the
1806 'uptodate' flag is reset so that auto-discovery will be used next
1807 time a new context for that algorithm attempts to select an
1809 [Ian Lister (tweaked by Geoff Thorpe)]
1811 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1812 implemention in the following ways:
1814 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1817 Lack of BER streaming support means one pass streaming processing is
1818 only supported if data is detached: setting the streaming flag is
1819 ignored for embedded content.
1821 CMS support is disabled by default and must be explicitly enabled
1822 with the enable-cms configuration option.
1825 *) Update the GMP engine glue to do direct copies between BIGNUM and
1826 mpz_t when openssl and GMP use the same limb size. Otherwise the
1827 existing "conversion via a text string export" trick is still used.
1828 [Paul Sheer <paulsheer@gmail.com>]
1830 *) Zlib compression BIO. This is a filter BIO which compressed and
1831 uncompresses any data passed through it.
1834 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1835 RFC3394 compatible AES key wrapping.
1838 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1839 sets string data without copying. X509_ALGOR_set0() and
1840 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1841 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1842 from an X509_ATTRIBUTE structure optionally checking it occurs only
1843 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1847 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1848 to get the expected BN_FLG_CONSTTIME behavior.
1849 [Bodo Moeller (Google)]
1853 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1854 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1855 - added some more tests to do_tests.pl
1856 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1857 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1858 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1859 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1860 - various changes to netware.pl to enable gcc-cross builds on Win32
1862 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1863 - various changes to fix missing prototype warnings
1864 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1865 - added AES, WHIRLPOOL and CPUID assembler code to build files
1866 - added missing AES assembler make rules to mk1mf.pl
1867 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1868 [Guenter Knauf <eflash@gmx.net>]
1870 *) Implement certificate status request TLS extension defined in RFC3546.
1871 A client can set the appropriate parameters and receive the encoded
1872 OCSP response via a callback. A server can query the supplied parameters
1873 and set the encoded OCSP response in the callback. Add simplified examples
1874 to s_client and s_server.
1877 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1879 *) Fix various bugs:
1880 + Binary incompatibility of ssl_ctx_st structure
1881 + DTLS interoperation with non-compliant servers
1882 + Don't call get_session_cb() without proposed session
1883 + Fix ia64 assembler code
1884 [Andy Polyakov, Steve Henson]
1886 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1888 *) DTLS Handshake overhaul. There were longstanding issues with
1889 OpenSSL DTLS implementation, which were making it impossible for
1890 RFC 4347 compliant client to communicate with OpenSSL server.
1891 Unfortunately just fixing these incompatibilities would "cut off"
1892 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1893 server keeps tolerating non RFC compliant syntax. The opposite is
1894 not true, 0.9.8f client can not communicate with earlier server.
1895 This update even addresses CVE-2007-4995.
1898 *) Changes to avoid need for function casts in OpenSSL: some compilers
1899 (gcc 4.2 and later) reject their use.
1900 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1903 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1904 RFC4507bis. The encrypted ticket format is an encrypted encoded
1905 SSL_SESSION structure, that way new session features are automatically
1908 If a client application caches session in an SSL_SESSION structure
1909 support is transparent because tickets are now stored in the encoded
1912 The SSL_CTX structure automatically generates keys for ticket
1913 protection in servers so again support should be possible
1914 with no application modification.
1916 If a client or server wishes to disable RFC4507 support then the option
1917 SSL_OP_NO_TICKET can be set.
1919 Add a TLS extension debugging callback to allow the contents of any client
1920 or server extensions to be examined.
1922 This work was sponsored by Google.
1925 *) Add initial support for TLS extensions, specifically for the server_name
1926 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1927 have new members for a host name. The SSL data structure has an
1928 additional member SSL_CTX *initial_ctx so that new sessions can be
1929 stored in that context to allow for session resumption, even after the
1930 SSL has been switched to a new SSL_CTX in reaction to a client's
1931 server_name extension.
1933 New functions (subject to change):
1935 SSL_get_servername()
1936 SSL_get_servername_type()
1939 New CTRL codes and macros (subject to change):
1941 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1942 - SSL_CTX_set_tlsext_servername_callback()
1943 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1944 - SSL_CTX_set_tlsext_servername_arg()
1945 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1947 openssl s_client has a new '-servername ...' option.
1949 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1950 '-key2 ...', '-servername_fatal' (subject to change). This allows
1951 testing the HostName extension for a specific single host name ('-cert'
1952 and '-key' remain fallbacks for handshakes without HostName
1953 negotiation). If the unrecogninzed_name alert has to be sent, this by
1954 default is a warning; it becomes fatal with the '-servername_fatal'
1957 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1959 *) Add AES and SSE2 assembly language support to VC++ build.
1962 *) Mitigate attack on final subtraction in Montgomery reduction.
1965 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1966 (which previously caused an internal error).
1969 *) Squeeze another 10% out of IGE mode when in != out.
1972 *) AES IGE mode speedup.
1973 [Dean Gaudet (Google)]
1975 *) Add the Korean symmetric 128-bit cipher SEED (see
1976 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1977 add SEED ciphersuites from RFC 4162:
1979 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1980 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1981 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1982 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1984 To minimize changes between patchlevels in the OpenSSL 0.9.8
1985 series, SEED remains excluded from compilation unless OpenSSL
1986 is configured with 'enable-seed'.
1987 [KISA, Bodo Moeller]
1989 *) Mitigate branch prediction attacks, which can be practical if a
1990 single processor is shared, allowing a spy process to extract
1991 information. For detailed background information, see
1992 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1993 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
1994 and Necessary Software Countermeasures"). The core of the change
1995 are new versions BN_div_no_branch() and
1996 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
1997 respectively, which are slower, but avoid the security-relevant
1998 conditional branches. These are automatically called by BN_div()
1999 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2000 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2001 remove a conditional branch.
2003 BN_FLG_CONSTTIME is the new name for the previous
2004 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2005 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2006 in the exponent causes BN_mod_exp_mont() to use the alternative
2007 implementation in BN_mod_exp_mont_consttime().) The old name
2008 remains as a deprecated alias.
2010 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2011 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2012 constant-time implementations for more than just exponentiation.
2013 Here too the old name is kept as a deprecated alias.
2015 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2016 the BN_BLINDING structure gets an independent copy of the
2017 modulus. This means that the previous "BIGNUM *m" argument to
2018 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2019 essentially becomes "const BIGNUM *m", although we can't actually
2020 change this in the header file before 0.9.9. It allows
2021 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2022 enable BN_FLG_CONSTTIME.
2024 [Matthew D Wood (Intel Corp)]
2026 *) In the SSL/TLS server implementation, be strict about session ID
2027 context matching (which matters if an application uses a single
2028 external cache for different purposes). Previously,
2029 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2030 set. This did ensure strict client verification, but meant that,
2031 with applications using a single external cache for quite
2032 different requirements, clients could circumvent ciphersuite
2033 restrictions for a given session ID context by starting a session
2034 in a different context.
2037 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2038 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2039 authentication-only ciphersuites.
2042 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2043 not complete and could lead to a possible single byte overflow
2044 (CVE-2007-5135) [Ben Laurie]
2046 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2048 *) Since AES128 and AES256 (and similarly Camellia128 and
2049 Camellia256) share a single mask bit in the logic of
2050 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2051 kludge to work properly if AES128 is available and AES256 isn't
2052 (or if Camellia128 is available and Camellia256 isn't).
2055 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2056 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2057 When a point or a seed is encoded in a BIT STRING, we need to
2058 prevent the removal of trailing zero bits to get the proper DER
2059 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2060 of a NamedBitList, for which trailing 0 bits need to be removed.)
2063 *) Have SSL/TLS server implementation tolerate "mismatched" record
2064 protocol version while receiving ClientHello even if the
2065 ClientHello is fragmented. (The server can't insist on the
2066 particular protocol version it has chosen before the ServerHello
2067 message has informed the client about his choice.)
2070 *) Add RFC 3779 support.
2071 [Rob Austein for ARIN, Ben Laurie]
2073 *) Load error codes if they are not already present instead of using a
2074 static variable. This allows them to be cleanly unloaded and reloaded.
2075 Improve header file function name parsing.
2078 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2079 or CAPABILITY handshake as required by RFCs.
2082 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2084 *) Introduce limits to prevent malicious keys being able to
2085 cause a denial of service. (CVE-2006-2940)
2086 [Steve Henson, Bodo Moeller]
2088 *) Fix ASN.1 parsing of certain invalid structures that can result
2089 in a denial of service. (CVE-2006-2937) [Steve Henson]
2091 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2092 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2094 *) Fix SSL client code which could crash if connecting to a
2095 malicious SSLv2 server. (CVE-2006-4343)
2096 [Tavis Ormandy and Will Drewry, Google Security Team]
2098 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2099 match only those. Before that, "AES256-SHA" would be interpreted
2100 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2101 the same strength classification in 0.9.7h) as we currently only
2102 have a single AES bit in the ciphersuite description bitmap.
2103 That change, however, also applied to ciphersuite strings such as
2104 "RC4-MD5" that intentionally matched multiple ciphersuites --
2105 namely, SSL 2.0 ciphersuites in addition to the more common ones
2106 from SSL 3.0/TLS 1.0.
2108 So we change the selection algorithm again: Naming an explicit
2109 ciphersuite selects this one ciphersuite, and any other similar
2110 ciphersuite (same bitmap) from *other* protocol versions.
2111 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2112 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2114 Since SSL 2.0 does not have any ciphersuites for which the
2115 128/256 bit distinction would be relevant, this works for now.
2116 The proper fix will be to use different bits for AES128 and
2117 AES256, which would have avoided the problems from the beginning;
2118 however, bits are scarce, so we can only do this in a new release
2119 (not just a patchlevel) when we can change the SSL_CIPHER
2120 definition to split the single 'unsigned long mask' bitmap into
2121 multiple values to extend the available space.
2125 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2127 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2128 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2130 *) Add AES IGE and biIGE modes.
2133 *) Change the Unix randomness entropy gathering to use poll() when
2134 possible instead of select(), since the latter has some
2135 undesirable limitations.
2136 [Darryl Miles via Richard Levitte and Bodo Moeller]
2138 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2139 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2140 cannot be implicitly activated as part of, e.g., the "AES" alias.
2141 However, please upgrade to OpenSSL 0.9.9[-dev] for
2142 non-experimental use of the ECC ciphersuites to get TLS extension
2143 support, which is required for curve and point format negotiation
2144 to avoid potential handshake problems.
2147 *) Disable rogue ciphersuites:
2149 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2150 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2151 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2153 The latter two were purportedly from
2154 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2157 Also deactivate the remaining ciphersuites from
2158 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2159 unofficial, and the ID has long expired.
2162 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2163 dual-core machines) and other potential thread-safety issues.
2166 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2167 versions), which is now available for royalty-free use
2168 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2169 Also, add Camellia TLS ciphersuites from RFC 4132.
2171 To minimize changes between patchlevels in the OpenSSL 0.9.8
2172 series, Camellia remains excluded from compilation unless OpenSSL
2173 is configured with 'enable-camellia'.
2176 *) Disable the padding bug check when compression is in use. The padding
2177 bug check assumes the first packet is of even length, this is not
2178 necessarily true if compresssion is enabled and can result in false
2179 positives causing handshake failure. The actual bug test is ancient
2180 code so it is hoped that implementations will either have fixed it by
2181 now or any which still have the bug do not support compression.
2184 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2186 *) When applying a cipher rule check to see if string match is an explicit
2187 cipher suite and only match that one cipher suite if it is.
2190 *) Link in manifests for VC++ if needed.
2191 [Austin Ziegler <halostatue@gmail.com>]
2193 *) Update support for ECC-based TLS ciphersuites according to
2194 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2195 TLS extensions, which are supported starting with the 0.9.9
2196 branch, not in the OpenSSL 0.9.8 branch).
2199 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2200 opaque EVP_CIPHER_CTX handling.
2203 *) Fixes and enhancements to zlib compression code. We now only use
2204 "zlib1.dll" and use the default __cdecl calling convention on Win32
2205 to conform with the standards mentioned here:
2206 http://www.zlib.net/DLL_FAQ.txt
2207 Static zlib linking now works on Windows and the new --with-zlib-include
2208 --with-zlib-lib options to Configure can be used to supply the location
2209 of the headers and library. Gracefully handle case where zlib library
2213 *) Several fixes and enhancements to the OID generation code. The old code
2214 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2215 handle numbers larger than ULONG_MAX, truncated printing and had a
2216 non standard OBJ_obj2txt() behaviour.
2219 *) Add support for building of engines under engine/ as shared libraries
2220 under VC++ build system.
2223 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2224 Hopefully, we will not see any false combination of paths any more.
2227 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2229 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2230 (part of SSL_OP_ALL). This option used to disable the
2231 countermeasure against man-in-the-middle protocol-version
2232 rollback in the SSL 2.0 server implementation, which is a bad
2233 idea. (CVE-2005-2969)
2235 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2236 for Information Security, National Institute of Advanced Industrial
2237 Science and Technology [AIST], Japan)]
2239 *) Add two function to clear and return the verify parameter flags.
2242 *) Keep cipherlists sorted in the source instead of sorting them at
2243 runtime, thus removing the need for a lock.
2246 *) Avoid some small subgroup attacks in Diffie-Hellman.
2247 [Nick Mathewson and Ben Laurie]
2249 *) Add functions for well-known primes.
2252 *) Extended Windows CE support.
2253 [Satoshi Nakamura and Andy Polyakov]
2255 *) Initialize SSL_METHOD structures at compile time instead of during
2256 runtime, thus removing the need for a lock.
2259 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2260 attempting to decrypt each encrypted key in turn. Add support to
2264 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2266 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2269 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2272 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2273 key into the same file any more.
2276 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2279 *) Add -utf8 command line and config file option to 'ca'.
2280 [Stefan <stf@udoma.org]
2282 *) Removed the macro des_crypt(), as it seems to conflict with some
2283 libraries. Use DES_crypt().
2286 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2287 involves renaming the source and generated shared-libs for
2288 both. The engines will accept the corrected or legacy ids
2289 ('ncipher' and '4758_cca' respectively) when binding. NB,
2290 this only applies when building 'shared'.
2291 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2293 *) Add attribute functions to EVP_PKEY structure. Modify
2294 PKCS12_create() to recognize a CSP name attribute and
2295 use it. Make -CSP option work again in pkcs12 utility.
2298 *) Add new functionality to the bn blinding code:
2299 - automatic re-creation of the BN_BLINDING parameters after
2300 a fixed number of uses (currently 32)
2301 - add new function for parameter creation
2302 - introduce flags to control the update behaviour of the
2303 BN_BLINDING parameters
2304 - hide BN_BLINDING structure
2305 Add a second BN_BLINDING slot to the RSA structure to improve
2306 performance when a single RSA object is shared among several
2310 *) Add support for DTLS.
2311 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2313 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2314 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2317 *) Remove buggy and incompletet DH cert support from
2318 ssl/ssl_rsa.c and ssl/s3_both.c
2321 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2322 the apps/openssl applications.
2325 *) Compile clean with "-Wall -Wmissing-prototypes
2326 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2327 DEBUG_SAFESTACK must also be set.
2330 *) Change ./Configure so that certain algorithms can be disabled by default.
2331 The new counterpiece to "no-xxx" is "enable-xxx".
2333 The patented RC5 and MDC2 algorithms will now be disabled unless
2334 "enable-rc5" and "enable-mdc2", respectively, are specified.
2336 (IDEA remains enabled despite being patented. This is because IDEA
2337 is frequently required for interoperability, and there is no license
2338 fee for non-commercial use. As before, "no-idea" can be used to
2339 avoid this algorithm.)
2343 *) Add processing of proxy certificates (see RFC 3820). This work was
2344 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2345 EGEE (Enabling Grids for E-science in Europe).
2348 *) RC4 performance overhaul on modern architectures/implementations, such
2349 as Intel P4, IA-64 and AMD64.
2352 *) New utility extract-section.pl. This can be used specify an alternative
2353 section number in a pod file instead of having to treat each file as
2354 a separate case in Makefile. This can be done by adding two lines to the
2357 =for comment openssl_section:XXX
2359 The blank line is mandatory.
2363 *) New arguments -certform, -keyform and -pass for s_client and s_server
2364 to allow alternative format key and certificate files and passphrase
2368 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2369 update associated structures and add various utility functions.
2371 Add new policy related verify parameters, include policy checking in
2372 standard verify code. Enhance 'smime' application with extra parameters
2373 to support policy checking and print out.
2376 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2377 Nehemiah processors. These extensions support AES encryption in hardware
2378 as well as RNG (though RNG support is currently disabled).
2379 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2381 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2384 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2385 [Andy Polyakov and a number of other people]
2387 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2388 implementation contributed by IBM.
2389 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2391 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2392 exponent rather than 'unsigned long'. There is a corresponding change to
2393 the new 'rsa_keygen' element of the RSA_METHOD structure.
2394 [Jelte Jansen, Geoff Thorpe]
2396 *) Functionality for creating the initial serial number file is now
2397 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2399 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2400 number file to 1, which is bound to cause problems. To avoid
2401 the problems while respecting compatibility between different 0.9.7
2402 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2403 CA.pl for serial number initialization. With the new release 0.9.8,
2404 we can fix the problem directly in the 'ca' utility.)
2407 *) Reduced header interdepencies by declaring more opaque objects in
2408 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2409 give fewer recursive includes, which could break lazy source code - so
2410 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2411 developers should define this symbol when building and using openssl to
2412 ensure they track the recommended behaviour, interfaces, [etc], but
2413 backwards-compatible behaviour prevails when this isn't defined.
2416 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2419 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2420 This will generate a random key of the appropriate length based on the
2421 cipher context. The EVP_CIPHER can provide its own random key generation
2422 routine to support keys of a specific form. This is used in the des and
2423 3des routines to generate a key of the correct parity. Update S/MIME
2424 code to use new functions and hence generate correct parity DES keys.
2425 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2426 valid (weak or incorrect parity).
2429 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2430 as looking them up. This is useful when the verified structure may contain
2431 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2432 present unless the new PKCS7_NO_CRL flag is asserted.
2435 *) Extend ASN1 oid configuration module. It now additionally accepts the
2438 shortName = some long name, 1.2.3.4
2441 *) Reimplemented the BN_CTX implementation. There is now no more static
2442 limitation on the number of variables it can handle nor the depth of the
2443 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2444 information can now expand as required, and rather than having a single
2445 static array of bignums, BN_CTX now uses a linked-list of such arrays
2446 allowing it to expand on demand whilst maintaining the usefulness of
2447 BN_CTX's "bundling".
2450 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2451 to allow all RSA operations to function using a single BN_CTX.
2454 *) Preliminary support for certificate policy evaluation and checking. This
2455 is initially intended to pass the tests outlined in "Conformance Testing
2456 of Relying Party Client Certificate Path Processing Logic" v1.07.
2459 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2460 remained unused and not that useful. A variety of other little bignum
2461 tweaks and fixes have also been made continuing on from the audit (see
2465 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2466 associated ASN1, EVP and SSL functions and old ASN1 macros.
2469 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2470 and this should never fail. So the return value from the use of
2471 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2472 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2475 *) BN_CTX_get() should return zero-valued bignums, providing the same
2476 initialised value as BN_new().
2477 [Geoff Thorpe, suggested by Ulf Möller]
2479 *) Support for inhibitAnyPolicy certificate extension.
2482 *) An audit of the BIGNUM code is underway, for which debugging code is
2483 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2484 is considered valid when processing BIGNUMs, and causes execution to
2485 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2486 further steps are taken to deliberately pollute unused data in BIGNUM
2487 structures to try and expose faulty code further on. For now, openssl will
2488 (in its default mode of operation) continue to tolerate the inconsistent
2489 forms that it has tolerated in the past, but authors and packagers should
2490 consider trying openssl and their own applications when compiled with
2491 these debugging symbols defined. It will help highlight potential bugs in
2492 their own code, and will improve the test coverage for OpenSSL itself. At
2493 some point, these tighter rules will become openssl's default to improve
2494 maintainability, though the assert()s and other overheads will remain only
2495 in debugging configurations. See bn.h for more details.
2496 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2498 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2499 that can only be obtained through BN_CTX_new() (which implicitly
2500 initialises it). The presence of this function only made it possible
2501 to overwrite an existing structure (and cause memory leaks).
2504 *) Because of the callback-based approach for implementing LHASH as a
2505 template type, lh_insert() adds opaque objects to hash-tables and
2506 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2507 to clean up those corresponding objects before destroying the hash table
2508 (and losing the object pointers). So some over-zealous constifications in
2509 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2510 objects as "const" and the lh_doall[_arg] callback wrappers are not
2511 prototyped to have "const" restrictions on the object pointers they are
2512 given (and so aren't required to cast them away any more).
2515 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2516 (speed) prefers to use its own implementation. The two implementations
2517 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2518 its object type properly exposed (MS_TM) instead of casting to/from "char
2519 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2520 aren't necessarily the greatest nomenclatures - but this is what was used
2521 internally to the implementation so I've used that for now.
2524 *) Ensure that deprecated functions do not get compiled when
2525 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2526 the self-tests were still using deprecated key-generation functions so
2527 these have been updated also.
2530 *) Reorganise PKCS#7 code to separate the digest location functionality
2531 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2532 New function PKCS7_set_digest() to set the digest type for PKCS#7
2533 digestedData type. Add additional code to correctly generate the
2534 digestedData type and add support for this type in PKCS7 initialization
2538 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2539 structure of type "other".
2542 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2543 sure the loop does correctly stop and breaking ("division by zero")
2544 modulus operations are not performed. The (pre-generated) prime
2545 table crypto/bn/bn_prime.h was already correct, but it could not be
2546 re-generated on some platforms because of the "division by zero"
2547 situation in the script.
2548 [Ralf S. Engelschall]
2550 *) Update support for ECC-based TLS ciphersuites according to
2551 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2552 SHA-1 now is only used for "small" curves (where the
2553 representation of a field element takes up to 24 bytes); for
2554 larger curves, the field element resulting from ECDH is directly
2555 used as premaster secret.
2556 [Douglas Stebila (Sun Microsystems Laboratories)]
2558 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2559 curve secp160r1 to the tests.
2560 [Douglas Stebila (Sun Microsystems Laboratories)]
2562 *) Add the possibility to load symbols globally with DSO.
2563 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2565 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2566 control of the error stack.
2569 *) Add support for STORE in ENGINE.
2572 *) Add the STORE type. The intention is to provide a common interface
2573 to certificate and key stores, be they simple file-based stores, or
2574 HSM-type store, or LDAP stores, or...
2575 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2578 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2579 pass a list of arguments to any function as well as provide a way
2580 for a function to pass data back to the caller.
2583 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2584 works like BUF_strdup() but can be used to duplicate a portion of
2585 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2589 *) Add the function sk_find_ex() which works like sk_find(), but will
2590 return an index to an element even if an exact match couldn't be
2591 found. The index is guaranteed to point at the element where the
2592 searched-for key would be inserted to preserve sorting order.
2595 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2596 takes an extra flags argument for optional functionality. Currently,
2597 the following flags are defined:
2599 OBJ_BSEARCH_VALUE_ON_NOMATCH
2600 This one gets OBJ_bsearch_ex() to return a pointer to the first
2601 element where the comparing function returns a negative or zero
2604 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2605 This one gets OBJ_bsearch_ex() to return a pointer to the first
2606 element where the comparing function returns zero. This is useful
2607 if there are more than one element where the comparing function
2611 *) Make it possible to create self-signed certificates with 'openssl ca'
2612 in such a way that the self-signed certificate becomes part of the
2613 CA database and uses the same mechanisms for serial number generation
2614 as all other certificate signing. The new flag '-selfsign' enables
2615 this functionality. Adapt CA.sh and CA.pl.in.
2618 *) Add functionality to check the public key of a certificate request
2619 against a given private. This is useful to check that a certificate
2620 request can be signed by that key (self-signing).
2623 *) Make it possible to have multiple active certificates with the same
2624 subject in the CA index file. This is done only if the keyword
2625 'unique_subject' is set to 'no' in the main CA section (default
2626 if 'CA_default') of the configuration file. The value is saved
2627 with the database itself in a separate index attribute file,
2628 named like the index file with '.attr' appended to the name.
2631 *) Generate muti valued AVAs using '+' notation in config files for
2635 *) Support for nameConstraints certificate extension.
2638 *) Support for policyConstraints certificate extension.
2641 *) Support for policyMappings certificate extension.
2644 *) Make sure the default DSA_METHOD implementation only uses its
2645 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2646 and change its own handlers to be NULL so as to remove unnecessary
2647 indirection. This lets alternative implementations fallback to the
2648 default implementation more easily.
2651 *) Support for directoryName in GeneralName related extensions
2655 *) Make it possible to link applications using Makefile.shared.
2656 Make that possible even when linking against static libraries!
2659 *) Support for single pass processing for S/MIME signing. This now
2660 means that S/MIME signing can be done from a pipe, in addition
2661 cleartext signing (multipart/signed type) is effectively streaming
2662 and the signed data does not need to be all held in memory.
2664 This is done with a new flag PKCS7_STREAM. When this flag is set
2665 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2666 is done after the data is output (and digests calculated) in
2667 SMIME_write_PKCS7().
2670 *) Add full support for -rpath/-R, both in shared libraries and
2671 applications, at least on the platforms where it's known how
2675 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2676 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2677 will now compute a table of multiples of the generator that
2678 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2679 faster (notably in the case of a single point multiplication,
2680 scalar * generator).
2681 [Nils Larsch, Bodo Moeller]
2683 *) IPv6 support for certificate extensions. The various extensions
2684 which use the IP:a.b.c.d can now take IPv6 addresses using the
2685 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2689 *) Added an ENGINE that implements RSA by performing private key
2690 exponentiations with the GMP library. The conversions to and from
2691 GMP's mpz_t format aren't optimised nor are any montgomery forms
2692 cached, and on x86 it appears OpenSSL's own performance has caught up.
2693 However there are likely to be other architectures where GMP could
2694 provide a boost. This ENGINE is not built in by default, but it can be
2695 specified at Configure time and should be accompanied by the necessary
2696 linker additions, eg;
2697 ./config -DOPENSSL_USE_GMP -lgmp
2700 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2701 testing availability of engines with "-t" - the old behaviour is
2702 produced by increasing the feature's verbosity with "-tt".
2705 *) ECDSA routines: under certain error conditions uninitialized BN objects
2706 could be freed. Solution: make sure initialization is performed early
2707 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2711 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2712 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2713 software implementations. For DSA and DH, parameter generation can
2714 also be overriden by providing the appropriate method callbacks.
2717 *) Change the "progress" mechanism used in key-generation and
2718 primality testing to functions that take a new BN_GENCB pointer in
2719 place of callback/argument pairs. The new API functions have "_ex"
2720 postfixes and the older functions are reimplemented as wrappers for
2721 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2722 declarations of the old functions to help (graceful) attempts to
2723 migrate to the new functions. Also, the new key-generation API
2724 functions operate on a caller-supplied key-structure and return
2725 success/failure rather than returning a key or NULL - this is to
2726 help make "keygen" another member function of RSA_METHOD etc.
2728 Example for using the new callback interface:
2730 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2734 BN_GENCB_set(&my_cb, my_callback, my_arg);
2736 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2737 /* For the meaning of a, b in calls to my_callback(), see the
2738 * documentation of the function that calls the callback.
2739 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2740 * my_callback should return 1 if it wants BN_is_prime_ex()
2741 * to continue, or 0 to stop.
2746 *) Change the ZLIB compression method to be stateful, and make it
2747 available to TLS with the number defined in
2748 draft-ietf-tls-compression-04.txt.
2751 *) Add the ASN.1 structures and functions for CertificatePair, which
2752 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2754 CertificatePair ::= SEQUENCE {
2755 forward [0] Certificate OPTIONAL,
2756 reverse [1] Certificate OPTIONAL,
2757 -- at least one of the pair shall be present -- }
2759 Also implement the PEM functions to read and write certificate
2760 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2762 This needed to be defined, mostly for the sake of the LDAP
2763 attribute crossCertificatePair, but may prove useful elsewhere as
2767 *) Make it possible to inhibit symlinking of shared libraries in
2768 Makefile.shared, for Cygwin's sake.
2771 *) Extend the BIGNUM API by creating a function
2772 void BN_set_negative(BIGNUM *a, int neg);
2773 and a macro that behave like
2774 int BN_is_negative(const BIGNUM *a);
2776 to avoid the need to access 'a->neg' directly in applications.
2779 *) Implement fast modular reduction for pseudo-Mersenne primes
2780 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2781 EC_GROUP_new_curve_GFp() will now automatically use this
2783 [Nils Larsch <nla@trustcenter.de>]
2785 *) Add new lock type (CRYPTO_LOCK_BN).
2788 *) Change the ENGINE framework to automatically load engines
2789 dynamically from specific directories unless they could be
2790 found to already be built in or loaded. Move all the
2791 current engines except for the cryptodev one to a new
2793 The engines in engines/ are built as shared libraries if
2794 the "shared" options was given to ./Configure or ./config.
2795 Otherwise, they are inserted in libcrypto.a.
2796 /usr/local/ssl/engines is the default directory for dynamic
2797 engines, but that can be overriden at configure time through
2798 the usual use of --prefix and/or --openssldir, and at run
2799 time with the environment variable OPENSSL_ENGINES.
2800 [Geoff Thorpe and Richard Levitte]
2802 *) Add Makefile.shared, a helper makefile to build shared
2803 libraries. Addapt Makefile.org.
2806 *) Add version info to Win32 DLLs.
2807 [Peter 'Luna' Runestig" <peter@runestig.com>]
2809 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2810 can be added using this API to created arbitrary PKCS#12
2811 files while avoiding the low level API.
2813 New options to PKCS12_create(), key or cert can be NULL and
2814 will then be omitted from the output file. The encryption
2815 algorithm NIDs can be set to -1 for no encryption, the mac
2816 iteration count can be set to 0 to omit the mac.
2818 Enhance pkcs12 utility by making the -nokeys and -nocerts
2819 options work when creating a PKCS#12 file. New option -nomac
2820 to omit the mac, NONE can be set for an encryption algorithm.
2821 New code is modified to use the enhanced PKCS12_create()
2822 instead of the low level API.
2825 *) Extend ASN1 encoder to support indefinite length constructed
2826 encoding. This can output sequences tags and octet strings in
2827 this form. Modify pk7_asn1.c to support indefinite length
2828 encoding. This is experimental and needs additional code to
2829 be useful, such as an ASN1 bio and some enhanced streaming
2832 Extend template encode functionality so that tagging is passed
2833 down to the template encoder.
2836 *) Let 'openssl req' fail if an argument to '-newkey' is not
2837 recognized instead of using RSA as a default.
2840 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2841 As these are not official, they are not included in "ALL";
2842 the "ECCdraft" ciphersuite group alias can be used to select them.
2843 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2845 *) Add ECDH engine support.
2846 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2848 *) Add ECDH in new directory crypto/ecdh/.
2849 [Douglas Stebila (Sun Microsystems Laboratories)]
2851 *) Let BN_rand_range() abort with an error after 100 iterations
2852 without success (which indicates a broken PRNG).
2855 *) Change BN_mod_sqrt() so that it verifies that the input value
2856 is really the square of the return value. (Previously,
2857 BN_mod_sqrt would show GIGO behaviour.)
2860 *) Add named elliptic curves over binary fields from X9.62, SECG,
2861 and WAP/WTLS; add OIDs that were still missing.
2863 [Sheueling Chang Shantz and Douglas Stebila
2864 (Sun Microsystems Laboratories)]
2866 *) Extend the EC library for elliptic curves over binary fields
2867 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2870 EC_GF2m_simple_method
2874 EC_GROUP_new_curve_GF2m
2875 EC_GROUP_set_curve_GF2m
2876 EC_GROUP_get_curve_GF2m
2877 EC_POINT_set_affine_coordinates_GF2m
2878 EC_POINT_get_affine_coordinates_GF2m
2879 EC_POINT_set_compressed_coordinates_GF2m
2881 Point compression for binary fields is disabled by default for
2882 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2885 As binary polynomials are represented as BIGNUMs, various members
2886 of the EC_GROUP and EC_POINT data structures can be shared
2887 between the implementations for prime fields and binary fields;
2888 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2889 are essentially identical to their ..._GFp counterparts.
2890 (For simplicity, the '..._GFp' prefix has been dropped from
2891 various internal method names.)
2893 An internal 'field_div' method (similar to 'field_mul' and
2894 'field_sqr') has been added; this is used only for binary fields.
2896 [Sheueling Chang Shantz and Douglas Stebila
2897 (Sun Microsystems Laboratories)]
2899 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2900 through methods ('mul', 'precompute_mult').
2902 The generic implementations (now internally called 'ec_wNAF_mul'
2903 and 'ec_wNAF_precomputed_mult') remain the default if these
2904 methods are undefined.
2906 [Sheueling Chang Shantz and Douglas Stebila
2907 (Sun Microsystems Laboratories)]
2909 *) New function EC_GROUP_get_degree, which is defined through
2910 EC_METHOD. For curves over prime fields, this returns the bit
2911 length of the modulus.
2913 [Sheueling Chang Shantz and Douglas Stebila
2914 (Sun Microsystems Laboratories)]
2916 *) New functions EC_GROUP_dup, EC_POINT_dup.
2917 (These simply call ..._new and ..._copy).
2919 [Sheueling Chang Shantz and Douglas Stebila
2920 (Sun Microsystems Laboratories)]
2922 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2923 Polynomials are represented as BIGNUMs (where the sign bit is not
2924 used) in the following functions [macros]:
2927 BN_GF2m_sub [= BN_GF2m_add]
2928 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2929 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2930 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2932 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2933 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2934 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2935 BN_GF2m_cmp [= BN_ucmp]
2937 (Note that only the 'mod' functions are actually for fields GF(2^m).
2938 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2940 For some functions, an the irreducible polynomial defining a
2941 field can be given as an 'unsigned int[]' with strictly
2942 decreasing elements giving the indices of those bits that are set;
2943 i.e., p[] represents the polynomial
2944 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2946 p[0] > p[1] > ... > p[k] = 0.
2947 This applies to the following functions:
2952 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2953 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2955 BN_GF2m_mod_sqrt_arr
2956 BN_GF2m_mod_solve_quad_arr
2960 Conversion can be performed by the following functions:
2965 bntest.c has additional tests for binary polynomial arithmetic.
2967 Two implementations for BN_GF2m_mod_div() are available.
2968 The default algorithm simply uses BN_GF2m_mod_inv() and
2969 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2970 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2971 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2973 [Sheueling Chang Shantz and Douglas Stebila
2974 (Sun Microsystems Laboratories)]
2976 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2977 functionality is disabled at compile-time.
2978 [Douglas Stebila <douglas.stebila@sun.com>]
2980 *) Change default behaviour of 'openssl asn1parse' so that more
2981 information is visible when viewing, e.g., a certificate:
2983 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2984 mode the content of non-printable OCTET STRINGs is output in a
2985 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2986 avoid the appearance of a printable string.
2987 [Nils Larsch <nla@trustcenter.de>]
2989 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2991 EC_GROUP_set_asn1_flag()
2992 EC_GROUP_get_asn1_flag()
2993 EC_GROUP_set_point_conversion_form()
2994 EC_GROUP_get_point_conversion_form()
2995 These control ASN1 encoding details:
2996 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
2997 has been set to OPENSSL_EC_NAMED_CURVE.
2998 - Points are encoded in uncompressed form by default; options for
2999 asn1_for are as for point2oct, namely
3000 POINT_CONVERSION_COMPRESSED
3001 POINT_CONVERSION_UNCOMPRESSED
3002 POINT_CONVERSION_HYBRID
3004 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3007 EC_GROUP_get0_seed()
3008 EC_GROUP_get_seed_len()
3009 This is used only for ASN1 purposes (so far).
3010 [Nils Larsch <nla@trustcenter.de>]
3012 *) Add 'field_type' member to EC_METHOD, which holds the NID
3013 of the appropriate field type OID. The new function
3014 EC_METHOD_get_field_type() returns this value.
3015 [Nils Larsch <nla@trustcenter.de>]
3020 EC_POINT_point2hex()
3021 EC_POINT_hex2point()
3022 providing useful interfaces to EC_POINT_point2oct() and
3023 EC_POINT_oct2point().
3024 [Nils Larsch <nla@trustcenter.de>]
3026 *) Change internals of the EC library so that the functions
3027 EC_GROUP_set_generator()
3028 EC_GROUP_get_generator()
3029 EC_GROUP_get_order()
3030 EC_GROUP_get_cofactor()
3031 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3032 to methods, which would lead to unnecessary code duplication when
3033 adding different types of curves.
3034 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3036 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3037 arithmetic, and such that modified wNAFs are generated
3038 (which avoid length expansion in many cases).
3041 *) Add a function EC_GROUP_check_discriminant() (defined via
3042 EC_METHOD) that verifies that the curve discriminant is non-zero.
3044 Add a function EC_GROUP_check() that makes some sanity tests
3045 on a EC_GROUP, its generator and order. This includes
3046 EC_GROUP_check_discriminant().
3047 [Nils Larsch <nla@trustcenter.de>]
3049 *) Add ECDSA in new directory crypto/ecdsa/.
3051 Add applications 'openssl ecparam' and 'openssl ecdsa'
3052 (these are based on 'openssl dsaparam' and 'openssl dsa').
3054 ECDSA support is also included in various other files across the
3055 library. Most notably,
3056 - 'openssl req' now has a '-newkey ecdsa:file' option;
3057 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3058 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3059 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3060 them suitable for ECDSA where domain parameters must be
3061 extracted before the specific public key;
3062 - ECDSA engine support has been added.
3063 [Nils Larsch <nla@trustcenter.de>]
3065 *) Include some named elliptic curves, and add OIDs from X9.62,
3066 SECG, and WAP/WTLS. Each curve can be obtained from the new
3068 EC_GROUP_new_by_curve_name(),
3069 and the list of available named curves can be obtained with
3070 EC_get_builtin_curves().
3071 Also add a 'curve_name' member to EC_GROUP objects, which can be
3073 EC_GROUP_set_curve_name()
3074 EC_GROUP_get_curve_name()
3075 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3077 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3078 was actually never needed) and in BN_mul(). The removal in BN_mul()
3079 required a small change in bn_mul_part_recursive() and the addition
3080 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3081 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3082 bn_sub_words() and bn_add_words() except they take arrays with
3086 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3088 *) Cleanse PEM buffers before freeing them since they may contain
3090 [Benjamin Bennett <ben@psc.edu>]
3092 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3093 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3094 authentication-only ciphersuites.
3097 *) Since AES128 and AES256 share a single mask bit in the logic of
3098 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3099 kludge to work properly if AES128 is available and AES256 isn't.
3102 *) Expand security boundary to match 1.1.1 module.
3105 *) Remove redundant features: hash file source, editing of test vectors
3106 modify fipsld to use external fips_premain.c signature.
3109 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3110 run algorithm test programs.
3113 *) Make algorithm test programs more tolerant of whitespace.
3116 *) Have SSL/TLS server implementation tolerate "mismatched" record
3117 protocol version while receiving ClientHello even if the
3118 ClientHello is fragmented. (The server can't insist on the
3119 particular protocol version it has chosen before the ServerHello
3120 message has informed the client about his choice.)
3123 *) Load error codes if they are not already present instead of using a
3124 static variable. This allows them to be cleanly unloaded and reloaded.
3127 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3129 *) Introduce limits to prevent malicious keys being able to
3130 cause a denial of service. (CVE-2006-2940)
3131 [Steve Henson, Bodo Moeller]
3133 *) Fix ASN.1 parsing of certain invalid structures that can result
3134 in a denial of service. (CVE-2006-2937) [Steve Henson]
3136 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3137 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3139 *) Fix SSL client code which could crash if connecting to a
3140 malicious SSLv2 server. (CVE-2006-4343)
3141 [Tavis Ormandy and Will Drewry, Google Security Team]
3143 *) Change ciphersuite string processing so that an explicit
3144 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3145 will no longer include "AES128-SHA"), and any other similar
3146 ciphersuite (same bitmap) from *other* protocol versions (so that
3147 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3148 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3149 changes from 0.9.8b and 0.9.8d.
3152 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3154 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3155 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3157 *) Change the Unix randomness entropy gathering to use poll() when
3158 possible instead of select(), since the latter has some
3159 undesirable limitations.
3160 [Darryl Miles via Richard Levitte and Bodo Moeller]
3162 *) Disable rogue ciphersuites:
3164 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3165 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3166 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3168 The latter two were purportedly from
3169 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3172 Also deactive the remaining ciphersuites from
3173 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3174 unofficial, and the ID has long expired.
3177 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3178 dual-core machines) and other potential thread-safety issues.
3181 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3183 *) Adapt fipsld and the build system to link against the validated FIPS
3184 module in FIPS mode.
3187 *) Fixes for VC++ 2005 build under Windows.
3190 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3191 from a Windows bash shell such as MSYS. It is autodetected from the
3192 "config" script when run from a VC++ environment. Modify standard VC++
3193 build to use fipscanister.o from the GNU make build.
3196 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3198 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3199 The value now differs depending on if you build for FIPS or not.
3200 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3201 safely run with a non-FIPSed libcrypto, as it may crash because of
3202 the difference induced by this change.
3205 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3207 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3208 (part of SSL_OP_ALL). This option used to disable the
3209 countermeasure against man-in-the-middle protocol-version
3210 rollback in the SSL 2.0 server implementation, which is a bad
3211 idea. (CVE-2005-2969)
3213 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3214 for Information Security, National Institute of Advanced Industrial
3215 Science and Technology [AIST], Japan)]
3217 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3218 mainly for FIPS compliance and not fully integrated at this stage.
3221 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3222 the exponentiation using a fixed-length exponent. (Otherwise,
3223 the information leaked through timing could expose the secret key
3224 after many signatures; cf. Bleichenbacher's attack on DSA with
3228 *) Make a new fixed-window mod_exp implementation the default for
3229 RSA, DSA, and DH private-key operations so that the sequence of
3230 squares and multiplies and the memory access pattern are
3231 independent of the particular secret key. This will mitigate
3232 cache-timing and potential related attacks.
3234 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3235 and this is automatically used by BN_mod_exp_mont() if the new flag
3236 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3237 will use this BN flag for private exponents unless the flag
3238 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3239 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3241 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3243 *) Change the client implementation for SSLv23_method() and
3244 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3245 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3246 (Previously, the SSL 2.0 backwards compatible Client Hello
3247 message format would be used even with SSL_OP_NO_SSLv2.)
3250 *) Add support for smime-type MIME parameter in S/MIME messages which some
3254 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3255 a threadsafe manner. Modify rsa code to use new function and add calls
3256 to dsa and dh code (which had race conditions before).
3259 *) Include the fixed error library code in the C error file definitions
3260 instead of fixing them up at runtime. This keeps the error code
3261 structures constant.
3264 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3266 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3269 *) Fixes for newer kerberos headers. NB: the casts are needed because
3270 the 'length' field is signed on one version and unsigned on another
3271 with no (?) obvious way to tell the difference, without these VC++
3272 complains. Also the "definition" of FAR (blank) is no longer included
3273 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3274 some needed definitions.
3277 *) Undo Cygwin change.
3280 *) Added support for proxy certificates according to RFC 3820.
3281 Because they may be a security thread to unaware applications,
3282 they must be explicitely allowed in run-time. See
3283 docs/HOWTO/proxy_certificates.txt for further information.
3286 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3288 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3289 server and client random values. Previously
3290 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3291 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3293 This change has negligible security impact because:
3295 1. Server and client random values still have 24 bytes of pseudo random
3298 2. Server and client random values are sent in the clear in the initial
3301 3. The master secret is derived using the premaster secret (48 bytes in
3302 size for static RSA ciphersuites) as well as client server and random
3305 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3308 [Stephen Henson, reported by UK NISCC]
3310 *) Use Windows randomness collection on Cygwin.
3313 *) Fix hang in EGD/PRNGD query when communication socket is closed
3314 prematurely by EGD/PRNGD.
3315 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3317 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3320 *) Back-port of selected performance improvements from development
3321 branch, as well as improved support for PowerPC platforms.
3324 *) Add lots of checks for memory allocation failure, error codes to indicate
3325 failure and freeing up memory if a failure occurs.
3326 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3328 *) Add new -passin argument to dgst.
3331 *) Perform some character comparisons of different types in X509_NAME_cmp:
3332 this is needed for some certificates that reencode DNs into UTF8Strings
3333 (in violation of RFC3280) and can't or wont issue name rollover
3337 *) Make an explicit check during certificate validation to see that
3338 the CA setting in each certificate on the chain is correct. As a
3339 side effect always do the following basic checks on extensions,
3340 not just when there's an associated purpose to the check:
3342 - if there is an unhandled critical extension (unless the user
3343 has chosen to ignore this fault)
3344 - if the path length has been exceeded (if one is set at all)
3345 - that certain extensions fit the associated purpose (if one has
3349 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3351 *) Avoid a race condition when CRLs are checked in a multi threaded
3352 environment. This would happen due to the reordering of the revoked
3353 entries during signature checking and serial number lookup. Now the
3354 encoding is cached and the serial number sort performed under a lock.
3355 Add new STACK function sk_is_sorted().
3358 *) Add Delta CRL to the extension code.
3361 *) Various fixes to s3_pkt.c so alerts are sent properly.
3362 [David Holmes <d.holmes@f5.com>]
3364 *) Reduce the chances of duplicate issuer name and serial numbers (in
3365 violation of RFC3280) using the OpenSSL certificate creation utilities.
3366 This is done by creating a random 64 bit value for the initial serial
3367 number when a serial number file is created or when a self signed
3368 certificate is created using 'openssl req -x509'. The initial serial
3369 number file is created using 'openssl x509 -next_serial' in CA.pl
3370 rather than being initialized to 1.
3373 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3375 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3376 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3377 [Joe Orton, Steve Henson]
3379 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3381 [Joe Orton, Steve Henson]
3383 *) Make it possible to have multiple active certificates with the same
3384 subject in the CA index file. This is done only if the keyword
3385 'unique_subject' is set to 'no' in the main CA section (default
3386 if 'CA_default') of the configuration file. The value is saved
3387 with the database itself in a separate index attribute file,
3388 named like the index file with '.attr' appended to the name.
3391 *) X509 verify fixes. Disable broken certificate workarounds when
3392 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3393 keyUsage extension present. Don't accept CRLs with unhandled critical
3394 extensions: since verify currently doesn't process CRL extensions this
3395 rejects a CRL with *any* critical extensions. Add new verify error codes
3399 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3400 A clarification of RFC2560 will require the use of OCTET STRINGs and
3401 some implementations cannot handle the current raw format. Since OpenSSL
3402 copies and compares OCSP nonces as opaque blobs without any attempt at
3403 parsing them this should not create any compatibility issues.
3406 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3407 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3408 this HMAC (and other) operations are several times slower than OpenSSL
3412 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3413 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3415 *) Use the correct content when signing type "other".
3418 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3420 *) Fix various bugs revealed by running the NISCC test suite:
3422 Stop out of bounds reads in the ASN1 code when presented with
3423 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3425 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3427 If verify callback ignores invalid public key errors don't try to check
3428 certificate signature with the NULL public key.
3432 *) New -ignore_err option in ocsp application to stop the server
3433 exiting on the first error in a request.
3436 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3437 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3441 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3442 extra data after the compression methods not only for TLS 1.0
3443 but also for SSL 3.0 (as required by the specification).
3444 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3446 *) Change X509_certificate_type() to mark the key as exported/exportable
3447 when it's 512 *bits* long, not 512 bytes.
3450 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3451 blocks during encryption.
3454 *) Various fixes to base64 BIO and non blocking I/O. On write
3455 flushes were not handled properly if the BIO retried. On read
3456 data was not being buffered properly and had various logic bugs.
3457 This also affects blocking I/O when the data being decoded is a
3461 *) Various S/MIME bugfixes and compatibility changes:
3462 output correct application/pkcs7 MIME type if
3463 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3464 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3465 of files as .eml work). Correctly handle very long lines in MIME
3469 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3471 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3472 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3473 a protocol version number mismatch like a decryption error
3474 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3477 *) Turn on RSA blinding by default in the default implementation
3478 to avoid a timing attack. Applications that don't want it can call
3479 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3480 They would be ill-advised to do so in most cases.
3481 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3483 *) Change RSA blinding code so that it works when the PRNG is not
3484 seeded (in this case, the secret RSA exponent is abused as
3485 an unpredictable seed -- if it is not unpredictable, there
3486 is no point in blinding anyway). Make RSA blinding thread-safe
3487 by remembering the creator's thread ID in rsa->blinding and
3488 having all other threads use local one-time blinding factors
3489 (this requires more computation than sharing rsa->blinding, but
3490 avoids excessive locking; and if an RSA object is not shared
3491 between threads, blinding will still be very fast).
3494 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3495 ENGINE as defaults for all supported algorithms irrespective of
3496 the 'flags' parameter. 'flags' is now honoured, so applications
3497 should make sure they are passing it correctly.
3500 *) Target "mingw" now allows native Windows code to be generated in
3501 the Cygwin environment as well as with the MinGW compiler.
3504 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3506 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3507 via timing by performing a MAC computation even if incorrrect
3508 block cipher padding has been found. This is a countermeasure
3509 against active attacks where the attacker has to distinguish
3510 between bad padding and a MAC verification error. (CVE-2003-0078)
3512 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3513 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3514 Martin Vuagnoux (EPFL, Ilion)]
3516 *) Make the no-err option work as intended. The intention with no-err
3517 is not to have the whole error stack handling routines removed from
3518 libcrypto, it's only intended to remove all the function name and
3519 reason texts, thereby removing some of the footprint that may not
3520 be interesting if those errors aren't displayed anyway.
3522 NOTE: it's still possible for any application or module to have it's
3523 own set of error texts inserted. The routines are there, just not
3524 used by default when no-err is given.
3527 *) Add support for FreeBSD on IA64.
3528 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3530 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3531 Kerberos function mit_des_cbc_cksum(). Before this change,
3532 the value returned by DES_cbc_cksum() was like the one from
3533 mit_des_cbc_cksum(), except the bytes were swapped.
3534 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3536 *) Allow an application to disable the automatic SSL chain building.
3537 Before this a rather primitive chain build was always performed in
3538 ssl3_output_cert_chain(): an application had no way to send the
3539 correct chain if the automatic operation produced an incorrect result.
3541 Now the chain builder is disabled if either:
3543 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3545 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3547 The reasoning behind this is that an application would not want the
3548 auto chain building to take place if extra chain certificates are
3549 present and it might also want a means of sending no additional
3550 certificates (for example the chain has two certificates and the
3554 *) Add the possibility to build without the ENGINE framework.
3555 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3557 *) Under Win32 gmtime() can return NULL: check return value in
3558 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3561 *) DSA routines: under certain error conditions uninitialized BN objects
3562 could be freed. Solution: make sure initialization is performed early
3563 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3564 Nils Larsch <nla@trustcenter.de> via PR#459)
3567 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3568 checked on reconnect on the client side, therefore session resumption
3569 could still fail with a "ssl session id is different" error. This
3570 behaviour is masked when SSL_OP_ALL is used due to
3571 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3572 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3573 followup to PR #377.
3576 *) IA-32 assembler support enhancements: unified ELF targets, support
3577 for SCO/Caldera platforms, fix for Cygwin shared build.
3580 *) Add support for FreeBSD on sparc64. As a consequence, support for
3581 FreeBSD on non-x86 processors is separate from x86 processors on
3582 the config script, much like the NetBSD support.
3583 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3585 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3587 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3590 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3591 code (06) was taken as the first octet of the session ID and the last
3592 octet was ignored consequently. As a result SSLv2 client side session
3593 caching could not have worked due to the session ID mismatch between
3595 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3599 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3600 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3604 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3605 seems that in spite of existing for more than a year, many application
3606 author have done nothing to provide the necessary callbacks, which
3607 means that this particular engine will not work properly anywhere.
3608 This is a very unfortunate situation which forces us, in the name
3609 of usability, to give the hw_ncipher.c a static lock, which is part
3611 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3612 appear in 0.9.8 or later. We EXPECT application authors to have
3613 dealt properly with this when 0.9.8 is released (unless we actually
3614 make such changes in the libcrypto locking code that changes will
3615 have to be made anyway).
3618 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3619 octets have been read, EOF or an error occurs. Without this change
3620 some truncated ASN1 structures will not produce an error.
3623 *) Disable Heimdal support, since it hasn't been fully implemented.
3624 Still give the possibility to force the use of Heimdal, but with
3625 warnings and a request that patches get sent to openssl-dev.
3628 *) Add the VC-CE target, introduce the WINCE sysname, and add
3629 INSTALL.WCE and appropriate conditionals to make it build.
3630 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3632 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3633 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3634 edit numbers of the version.
3635 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3637 *) Introduce safe string copy and catenation functions
3638 (BUF_strlcpy() and BUF_strlcat()).
3639 [Ben Laurie (CHATS) and Richard Levitte]
3641 *) Avoid using fixed-size buffers for one-line DNs.
3642 [Ben Laurie (CHATS)]
3644 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3645 resizing buffers containing secrets, and use where appropriate.
3646 [Ben Laurie (CHATS)]
3648 *) Avoid using fixed size buffers for configuration file location.
3649 [Ben Laurie (CHATS)]
3651 *) Avoid filename truncation for various CA files.
3652 [Ben Laurie (CHATS)]
3654 *) Use sizeof in preference to magic numbers.
3655 [Ben Laurie (CHATS)]
3657 *) Avoid filename truncation in cert requests.
3658 [Ben Laurie (CHATS)]
3660 *) Add assertions to check for (supposedly impossible) buffer
3662 [Ben Laurie (CHATS)]
3664 *) Don't cache truncated DNS entries in the local cache (this could
3665 potentially lead to a spoofing attack).
3666 [Ben Laurie (CHATS)]
3668 *) Fix various buffers to be large enough for hex/decimal
3669 representations in a platform independent manner.
3670 [Ben Laurie (CHATS)]
3672 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3673 resizing buffers containing secrets, and use where appropriate.
3674 [Ben Laurie (CHATS)]
3676 *) Add BIO_indent() to avoid much slightly worrying code to do
3678 [Ben Laurie (CHATS)]
3680 *) Convert sprintf()/BIO_puts() to BIO_printf().
3681 [Ben Laurie (CHATS)]
3683 *) buffer_gets() could terminate with the buffer only half
3685 [Ben Laurie (CHATS)]
3687 *) Add assertions to prevent user-supplied crypto functions from
3688 overflowing internal buffers by having large block sizes, etc.
3689 [Ben Laurie (CHATS)]
3691 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3693 [Ben Laurie (CHATS)]
3695 *) Eliminate unused copy of key in RC4.
3696 [Ben Laurie (CHATS)]
3698 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3699 [Ben Laurie (CHATS)]
3701 *) Fix off-by-one error in EGD path.
3702 [Ben Laurie (CHATS)]
3704 *) If RANDFILE path is too long, ignore instead of truncating.
3705 [Ben Laurie (CHATS)]
3707 *) Eliminate unused and incorrectly sized X.509 structure
3709 [Ben Laurie (CHATS)]
3711 *) Eliminate unused and dangerous function knumber().
3712 [Ben Laurie (CHATS)]
3714 *) Eliminate unused and dangerous structure, KSSL_ERR.
3715 [Ben Laurie (CHATS)]
3717 *) Protect against overlong session ID context length in an encoded
3718 session object. Since these are local, this does not appear to be
3720 [Ben Laurie (CHATS)]
3722 *) Change from security patch (see 0.9.6e below) that did not affect
3723 the 0.9.6 release series:
3725 Remote buffer overflow in SSL3 protocol - an attacker could
3726 supply an oversized master key in Kerberos-enabled versions.
3728 [Ben Laurie (CHATS)]
3730 *) Change the SSL kerb5 codes to match RFC 2712.
3733 *) Make -nameopt work fully for req and add -reqopt switch.
3734 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3736 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3737 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3739 *) Make sure tests can be performed even if the corresponding algorithms
3740 have been removed entirely. This was also the last step to make
3741 OpenSSL compilable with DJGPP under all reasonable conditions.
3742 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3744 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3745 to allow version independent disabling of normally unselected ciphers,
3746 which may be activated as a side-effect of selecting a single cipher.
3748 (E.g., cipher list string "RSA" enables ciphersuites that are left
3749 out of "ALL" because they do not provide symmetric encryption.
3750 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3751 [Lutz Jaenicke, Bodo Moeller]
3753 *) Add appropriate support for separate platform-dependent build
3754 directories. The recommended way to make a platform-dependent
3755 build directory is the following (tested on Linux), maybe with
3758 # Place yourself outside of the OpenSSL source tree. In
3759 # this example, the environment variable OPENSSL_SOURCE
3760 # is assumed to contain the absolute OpenSSL source directory.
3761 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3762 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3763 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3764 mkdir -p `dirname $F`
3765 ln -s $OPENSSL_SOURCE/$F $F
3768 To be absolutely sure not to disturb the source tree, a "make clean"
3769 is a good thing. If it isn't successfull, don't worry about it,
3770 it probably means the source directory is very clean.
3773 *) Make sure any ENGINE control commands make local copies of string
3774 pointers passed to them whenever necessary. Otherwise it is possible
3775 the caller may have overwritten (or deallocated) the original string
3776 data when a later ENGINE operation tries to use the stored values.
3777 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3779 *) Improve diagnostics in file reading and command-line digests.
3780 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3782 *) Add AES modes CFB and OFB to the object database. Correct an
3783 error in AES-CFB decryption.
3786 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3787 allows existing EVP_CIPHER_CTX structures to be reused after
3788 calling EVP_*Final(). This behaviour is used by encryption
3789 BIOs and some applications. This has the side effect that
3790 applications must explicitly clean up cipher contexts with
3791 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3794 *) Check the values of dna and dnb in bn_mul_recursive before calling
3795 bn_mul_comba (a non zero value means the a or b arrays do not contain
3796 n2 elements) and fallback to bn_mul_normal if either is not zero.
3799 *) Fix escaping of non-ASCII characters when using the -subj option
3800 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3803 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3804 form for "surname", serialNumber has no short form.
3805 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3806 therefore remove "mail" short name for "internet 7".
3807 The OID for unique identifiers in X509 certificates is
3808 x500UniqueIdentifier, not uniqueIdentifier.
3809 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3812 *) Add an "init" command to the ENGINE config module and auto initialize
3813 ENGINEs. Without any "init" command the ENGINE will be initialized
3814 after all ctrl commands have been executed on it. If init=1 the
3815 ENGINE is initailized at that point (ctrls before that point are run
3816 on the uninitialized ENGINE and after on the initialized one). If
3817 init=0 then the ENGINE will not be iniatialized at all.
3820 *) Fix the 'app_verify_callback' interface so that the user-defined
3821 argument is actually passed to the callback: In the
3822 SSL_CTX_set_cert_verify_callback() prototype, the callback
3823 declaration has been changed from
3826 int (*cb)(X509_STORE_CTX *,void *);
3827 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3828 i=s->ctx->app_verify_callback(&ctx)
3829 has been changed into
3830 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3832 To update applications using SSL_CTX_set_cert_verify_callback(),
3833 a dummy argument can be added to their callback functions.
3834 [D. K. Smetters <smetters@parc.xerox.com>]
3836 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3837 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3839 *) Add and OPENSSL_LOAD_CONF define which will cause
3840 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3841 This allows older applications to transparently support certain
3842 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3843 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3844 load the config file and OPENSSL_add_all_algorithms_conf() which will
3845 always load it have also been added.
3848 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3849 Adjust NIDs and EVP layer.
3850 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3852 *) Config modules support in openssl utility.
3854 Most commands now load modules from the config file,
3855 though in a few (such as version) this isn't done
3856 because it couldn't be used for anything.
3858 In the case of ca and req the config file used is
3859 the same as the utility itself: that is the -config
3860 command line option can be used to specify an
3864 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3865 use "openssl_conf" if filename is NULL use default openssl config file.
3868 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3869 config section name. Add a new flag to tolerate a missing config file
3870 and move code to CONF_modules_load_file().
3873 *) Support for crypto accelerator cards from Accelerated Encryption
3874 Processing, www.aep.ie. (Use engine 'aep')
3875 The support was copied from 0.9.6c [engine] and adapted/corrected
3876 to work with the new engine framework.
3877 [AEP Inc. and Richard Levitte]
3879 *) Support for SureWare crypto accelerator cards from Baltimore
3880 Technologies. (Use engine 'sureware')
3881 The support was copied from 0.9.6c [engine] and adapted
3882 to work with the new engine framework.
3885 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3886 make the newer ENGINE framework commands for the CHIL engine work.
3887 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3889 *) Make it possible to produce shared libraries on ReliantUNIX.
3890 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3892 *) Add the configuration target debug-linux-ppro.
3893 Make 'openssl rsa' use the general key loading routines
3894 implemented in apps.c, and make those routines able to
3895 handle the key format FORMAT_NETSCAPE and the variant
3897 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3899 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3900 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3902 *) Add -keyform to rsautl, and document -engine.
3903 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3905 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3906 BIO_R_NO_SUCH_FILE error code rather than the generic
3907 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3910 *) Add new functions
3912 ERR_peek_last_error_line
3913 ERR_peek_last_error_line_data.
3914 These are similar to
3917 ERR_peek_error_line_data,
3918 but report on the latest error recorded rather than the first one
3919 still in the error queue.
3920 [Ben Laurie, Bodo Moeller]
3922 *) default_algorithms option in ENGINE config module. This allows things
3924 default_algorithms = ALL
3925 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3928 *) Prelminary ENGINE config module.
3931 *) New experimental application configuration code.
3934 *) Change the AES code to follow the same name structure as all other
3935 symmetric ciphers, and behave the same way. Move everything to
3936 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3937 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3939 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3940 [Ben Laurie and Theo de Raadt]
3942 *) Add option to output public keys in req command.
3943 [Massimiliano Pala madwolf@openca.org]
3945 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3946 (up to about 10% better than before for P-192 and P-224).
3949 *) New functions/macros
3951 SSL_CTX_set_msg_callback(ctx, cb)
3952 SSL_CTX_set_msg_callback_arg(ctx, arg)
3953 SSL_set_msg_callback(ssl, cb)
3954 SSL_set_msg_callback_arg(ssl, arg)
3956 to request calling a callback function
3958 void cb(int write_p, int version, int content_type,
3959 const void *buf, size_t len, SSL *ssl, void *arg)
3961 whenever a protocol message has been completely received
3962 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3963 protocol version according to which the SSL library interprets
3964 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3965 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3966 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3967 specification (change_cipher_spec(20), alert(21), handshake(22)).
3968 'buf' and 'len' point to the actual message, 'ssl' to the
3969 SSL object, and 'arg' is the application-defined value set by
3970 SSL[_CTX]_set_msg_callback_arg().
3972 'openssl s_client' and 'openssl s_server' have new '-msg' options
3973 to enable a callback that displays all protocol messages.
3976 *) Change the shared library support so shared libraries are built as
3977 soon as the corresponding static library is finished, and thereby get
3978 openssl and the test programs linked against the shared library.
3979 This still only happens when the keyword "shard" has been given to
3980 the configuration scripts.
3982 NOTE: shared library support is still an experimental thing, and
3983 backward binary compatibility is still not guaranteed.
3984 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3986 *) Add support for Subject Information Access extension.
3987 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3989 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3990 additional bytes when new memory had to be allocated, not just
3991 when reusing an existing buffer.
3994 *) New command line and configuration option 'utf8' for the req command.
3995 This allows field values to be specified as UTF8 strings.
3998 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
3999 runs for the former and machine-readable output for the latter.
4002 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4003 of the e-mail address in the DN (i.e., it will go into a certificate
4004 extension only). The new configuration file option 'email_in_dn = no'
4005 has the same effect.
4006 [Massimiliano Pala madwolf@openca.org]
4008 *) Change all functions with names starting with des_ to be starting
4009 with DES_ instead. Add wrappers that are compatible with libdes,
4010 but are named _ossl_old_des_*. Finally, add macros that map the
4011 des_* symbols to the corresponding _ossl_old_des_* if libdes
4012 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4013 desired, the des_* symbols will be mapped to DES_*, with one
4016 Since we provide two compatibility mappings, the user needs to
4017 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4018 compatibility is desired. The default (i.e., when that macro
4019 isn't defined) is OpenSSL 0.9.6c compatibility.
4021 There are also macros that enable and disable the support of old
4022 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4023 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4024 are defined, the default will apply: to support the old des routines.
4026 In either case, one must include openssl/des.h to get the correct
4027 definitions. Do not try to just include openssl/des_old.h, that
4030 NOTE: This is a major break of an old API into a new one. Software
4031 authors are encouraged to switch to the DES_ style functions. Some
4032 time in the future, des_old.h and the libdes compatibility functions
4033 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4034 default), and then completely removed.
4037 *) Test for certificates which contain unsupported critical extensions.
4038 If such a certificate is found during a verify operation it is
4039 rejected by default: this behaviour can be overridden by either
4040 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4041 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4042 X509_supported_extension() has also been added which returns 1 if a
4043 particular extension is supported.
4046 *) Modify the behaviour of EVP cipher functions in similar way to digests
4047 to retain compatibility with existing code.
4050 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4051 compatibility with existing code. In particular the 'ctx' parameter does
4052 not have to be to be initialized before the call to EVP_DigestInit() and
4053 it is tidied up after a call to EVP_DigestFinal(). New function
4054 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4055 EVP_MD_CTX_copy() changed to not require the destination to be
4056 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4057 requires the destination to be valid.
4059 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4060 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4063 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4064 so that complete 'Handshake' protocol structures are kept in memory
4065 instead of overwriting 'msg_type' and 'length' with 'body' data.
4068 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4069 [Massimo Santin via Richard Levitte]
4071 *) Major restructuring to the underlying ENGINE code. This includes
4072 reduction of linker bloat, separation of pure "ENGINE" manipulation
4073 (initialisation, etc) from functionality dealing with implementations
4074 of specific crypto iterfaces. This change also introduces integrated
4075 support for symmetric ciphers and digest implementations - so ENGINEs
4076 can now accelerate these by providing EVP_CIPHER and EVP_MD
4077 implementations of their own. This is detailed in crypto/engine/README
4078 as it couldn't be adequately described here. However, there are a few
4079 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4080 were changed in the original introduction of ENGINE code have now
4081 reverted back - the hooking from this code to ENGINE is now a good
4082 deal more passive and at run-time, operations deal directly with
4083 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4084 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4085 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4086 they were not being used by the framework as there is no concept of a
4087 BIGNUM_METHOD and they could not be generalised to the new
4088 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4089 ENGINE_cpy() has been removed as it cannot be consistently defined in
4093 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4096 *) Change mkdef.pl to sort symbols that get the same entry number,
4097 and make sure the automatically generated functions ERR_load_*
4098 become part of libeay.num as well.
4101 *) New function SSL_renegotiate_pending(). This returns true once
4102 renegotiation has been requested (either SSL_renegotiate() call
4103 or HelloRequest/ClientHello receveived from the peer) and becomes
4104 false once a handshake has been completed.
4105 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4106 sends a HelloRequest, but does not ensure that a handshake takes
4107 place. SSL_renegotiate_pending() is useful for checking if the
4108 client has followed the request.)
4111 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4112 By default, clients may request session resumption even during
4113 renegotiation (if session ID contexts permit); with this option,
4114 session resumption is possible only in the first handshake.
4116 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4117 more bits available for options that should not be part of
4118 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4121 *) Add some demos for certificate and certificate request creation.
4124 *) Make maximum certificate chain size accepted from the peer application
4125 settable (SSL*_get/set_max_cert_list()), as proposed by
4126 "Douglas E. Engert" <deengert@anl.gov>.
4129 *) Add support for shared libraries for Unixware-7
4130 (Boyd Lynn Gerber <gerberb@zenez.com>).
4133 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4134 be done prior to destruction. Use this to unload error strings from
4135 ENGINEs that load their own error strings. NB: This adds two new API
4136 functions to "get" and "set" this destroy handler in an ENGINE.
4139 *) Alter all existing ENGINE implementations (except "openssl" and
4140 "openbsd") to dynamically instantiate their own error strings. This
4141 makes them more flexible to be built both as statically-linked ENGINEs
4142 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4143 Also, add stub code to each that makes building them as self-contained
4144 shared-libraries easier (see README.ENGINE).
4147 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4148 implementations into applications that are completely implemented in
4149 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4150 commands that can be used to configure what shared-library to load and
4151 to control aspects of the way it is handled. Also, made an update to
4152 the README.ENGINE file that brings its information up-to-date and
4153 provides some information and instructions on the "dynamic" ENGINE
4154 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4157 *) Make it possible to unload ranges of ERR strings with a new
4158 "ERR_unload_strings" function.
4161 *) Add a copy() function to EVP_MD.
4164 *) Make EVP_MD routines take a context pointer instead of just the
4165 md_data void pointer.
4168 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4169 that the digest can only process a single chunk of data
4170 (typically because it is provided by a piece of
4171 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4172 is only going to provide a single chunk of data, and hence the
4173 framework needn't accumulate the data for oneshot drivers.
4176 *) As with "ERR", make it possible to replace the underlying "ex_data"
4177 functions. This change also alters the storage and management of global
4178 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4179 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4180 index counters. The API functions that use this state have been changed
4181 to take a "class_index" rather than pointers to the class's local STACK
4182 and counter, and there is now an API function to dynamically create new
4183 classes. This centralisation allows us to (a) plug a lot of the
4184 thread-safety problems that existed, and (b) makes it possible to clean
4185 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4186 such data would previously have always leaked in application code and
4187 workarounds were in place to make the memory debugging turn a blind eye
4188 to it. Application code that doesn't use this new function will still
4189 leak as before, but their memory debugging output will announce it now
4190 rather than letting it slide.
4192 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4193 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4194 has a return value to indicate success or failure.
4197 *) Make it possible to replace the underlying "ERR" functions such that the
4198 global state (2 LHASH tables and 2 locks) is only used by the "default"
4199 implementation. This change also adds two functions to "get" and "set"
4200 the implementation prior to it being automatically set the first time
4201 any other ERR function takes place. Ie. an application can call "get",
4202 pass the return value to a module it has just loaded, and that module
4203 can call its own "set" function using that value. This means the
4204 module's "ERR" operations will use (and modify) the error state in the
4205 application and not in its own statically linked copy of OpenSSL code.
4208 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4209 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4210 the operation, and provides a more encapsulated way for external code
4211 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4212 to use these functions rather than manually incrementing the counts.
4214 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4217 *) Add EVP test program.
4220 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4223 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4224 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4225 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4226 These allow a CRL to be built without having to access X509_CRL fields
4227 directly. Modify 'ca' application to use new functions.
4230 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4231 bug workarounds. Rollback attack detection is a security feature.
4232 The problem will only arise on OpenSSL servers when TLSv1 is not
4233 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4234 Software authors not wanting to support TLSv1 will have special reasons
4235 for their choice and can explicitly enable this option.
4236 [Bodo Moeller, Lutz Jaenicke]
4238 *) Rationalise EVP so it can be extended: don't include a union of
4239 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4240 (similar to those existing for EVP_CIPHER_CTX).
4245 EVP_MD_CTX_init(&md); /* new function call */
4246 EVP_DigestInit(&md, EVP_sha1());
4247 EVP_DigestUpdate(&md, in, len);
4248 EVP_DigestFinal(&md, out, NULL);
4249 EVP_MD_CTX_cleanup(&md); /* new function call */
4253 *) Make DES key schedule conform to the usual scheme, as well as
4254 correcting its structure. This means that calls to DES functions
4255 now have to pass a pointer to a des_key_schedule instead of a
4256 plain des_key_schedule (which was actually always a pointer
4259 des_key_schedule ks;
4261 des_set_key_checked(..., &ks);
4262 des_ncbc_encrypt(..., &ks, ...);
4264 (Note that a later change renames 'des_...' into 'DES_...'.)
4267 *) Initial reduction of linker bloat: the use of some functions, such as
4268 PEM causes large amounts of unused functions to be linked in due to
4269 poor organisation. For example pem_all.c contains every PEM function
4270 which has a knock on effect of linking in large amounts of (unused)
4271 ASN1 code. Grouping together similar functions and splitting unrelated
4272 functions prevents this.
4275 *) Cleanup of EVP macros.
4278 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4279 correct _ecb suffix.
4282 *) Add initial OCSP responder support to ocsp application. The
4283 revocation information is handled using the text based index
4284 use by the ca application. The responder can either handle
4285 requests generated internally, supplied in files (for example
4286 via a CGI script) or using an internal minimal server.
4289 *) Add configuration choices to get zlib compression for TLS.
4292 *) Changes to Kerberos SSL for RFC 2712 compliance:
4293 1. Implemented real KerberosWrapper, instead of just using
4294 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4295 2. Implemented optional authenticator field of KerberosWrapper.
4297 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4298 and authenticator structs; see crypto/krb5/.
4300 Generalized Kerberos calls to support multiple Kerberos libraries.
4301 [Vern Staats <staatsvr@asc.hpc.mil>,
4302 Jeffrey Altman <jaltman@columbia.edu>
4303 via Richard Levitte]
4305 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4306 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4307 values for each of the key sizes rather than having just
4308 parameters (and 'speed' generating keys each time).
4311 *) Speed up EVP routines.
4314 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4315 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4316 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4317 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4319 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4320 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4321 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4324 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4326 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4329 *) Added the OS2-EMX target.
4330 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4332 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4333 to support NCONF routines in extension code. New function CONF_set_nconf()
4334 to allow functions which take an NCONF to also handle the old LHASH
4335 structure: this means that the old CONF compatible routines can be
4336 retained (in particular wrt extensions) without having to duplicate the
4337 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4340 *) Enhance the general user interface with mechanisms for inner control
4341 and with possibilities to have yes/no kind of prompts.
4344 *) Change all calls to low level digest routines in the library and
4345 applications to use EVP. Add missing calls to HMAC_cleanup() and
4346 don't assume HMAC_CTX can be copied using memcpy().
4347 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4349 *) Add the possibility to control engines through control names but with
4350 arbitrary arguments instead of just a string.
4351 Change the key loaders to take a UI_METHOD instead of a callback
4352 function pointer. NOTE: this breaks binary compatibility with earlier
4353 versions of OpenSSL [engine].
4354 Adapt the nCipher code for these new conditions and add a card insertion
4358 *) Enhance the general user interface with mechanisms to better support
4359 dialog box interfaces, application-defined prompts, the possibility
4360 to use defaults (for example default passwords from somewhere else)
4361 and interrupts/cancellations.
4364 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4365 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4368 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4369 tidy up some unnecessarily weird code in 'sk_new()').
4370 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4372 *) Change the key loading routines for ENGINEs to use the same kind
4373 callback (pem_password_cb) as all other routines that need this
4377 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4378 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4379 than this minimum value is recommended.
4382 *) New random seeder for OpenVMS, using the system process statistics
4383 that are easily reachable.
4386 *) Windows apparently can't transparently handle global
4387 variables defined in DLLs. Initialisations such as:
4389 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4391 wont compile. This is used by the any applications that need to
4392 declare their own ASN1 modules. This was fixed by adding the option
4393 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4394 needed for static libraries under Win32.
4397 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4398 setting of purpose and trust fields. New X509_STORE trust and
4399 purpose functions and tidy up setting in other SSL functions.
4402 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4403 structure. These are inherited by X509_STORE_CTX when it is
4404 initialised. This allows various defaults to be set in the
4405 X509_STORE structure (such as flags for CRL checking and custom
4406 purpose or trust settings) for functions which only use X509_STORE_CTX
4407 internally such as S/MIME.
4409 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4410 trust settings if they are not set in X509_STORE. This allows X509_STORE
4411 purposes and trust (in S/MIME for example) to override any set by default.
4413 Add command line options for CRL checking to smime, s_client and s_server
4417 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4418 are set then the CRL is looked up in the X509_STORE structure and
4419 its validity and signature checked, then if the certificate is found
4420 in the CRL the verify fails with a revoked error.
4422 Various new CRL related callbacks added to X509_STORE_CTX structure.
4424 Command line options added to 'verify' application to support this.
4426 This needs some additional work, such as being able to handle multiple
4427 CRLs with different times, extension based lookup (rather than just
4428 by subject name) and ultimately more complete V2 CRL extension
4432 *) Add a general user interface API (crypto/ui/). This is designed
4433 to replace things like des_read_password and friends (backward
4434 compatibility functions using this new API are provided).
4435 The purpose is to remove prompting functions from the DES code
4436 section as well as provide for prompting through dialog boxes in
4437 a window system and the like.
4440 *) Add "ex_data" support to ENGINE so implementations can add state at a
4441 per-structure level rather than having to store it globally.
4444 *) Make it possible for ENGINE structures to be copied when retrieved by
4445 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4446 This causes the "original" ENGINE structure to act like a template,
4447 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4448 operational state can be localised to each ENGINE structure, despite the
4449 fact they all share the same "methods". New ENGINE structures returned in
4450 this case have no functional references and the return value is the single
4451 structural reference. This matches the single structural reference returned
4452 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4456 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4457 needs to match any other type at all we need to manually clear the
4461 *) Changes to the "openssl engine" utility to include;
4462 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4463 about an ENGINE's available control commands.
4464 - executing control commands from command line arguments using the
4465 '-pre' and '-post' switches. '-post' is only used if '-t' is
4466 specified and the ENGINE is successfully initialised. The syntax for
4467 the individual commands are colon-separated, for example;
4468 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4471 *) New dynamic control command support for ENGINEs. ENGINEs can now
4472 declare their own commands (numbers), names (strings), descriptions,
4473 and input types for run-time discovery by calling applications. A
4474 subset of these commands are implicitly classed as "executable"
4475 depending on their input type, and only these can be invoked through
4476 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4477 can be based on user input, config files, etc). The distinction is
4478 that "executable" commands cannot return anything other than a boolean
4479 result and can only support numeric or string input, whereas some
4480 discoverable commands may only be for direct use through
4481 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4482 pointers, or other custom uses. The "executable" commands are to
4483 support parameterisations of ENGINE behaviour that can be
4484 unambiguously defined by ENGINEs and used consistently across any
4485 OpenSSL-based application. Commands have been added to all the
4486 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4487 control over shared-library paths without source code alterations.
4490 *) Changed all ENGINE implementations to dynamically allocate their
4491 ENGINEs rather than declaring them statically. Apart from this being
4492 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4493 this also allows the implementations to compile without using the
4494 internal engine_int.h header.
4497 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4498 'const' value. Any code that should be able to modify a RAND_METHOD
4499 should already have non-const pointers to it (ie. they should only
4500 modify their own ones).
4503 *) Made a variety of little tweaks to the ENGINE code.
4504 - "atalla" and "ubsec" string definitions were moved from header files
4505 to C code. "nuron" string definitions were placed in variables
4506 rather than hard-coded - allowing parameterisation of these values
4507 later on via ctrl() commands.
4508 - Removed unused "#if 0"'d code.
4509 - Fixed engine list iteration code so it uses ENGINE_free() to release
4510 structural references.
4511 - Constified the RAND_METHOD element of ENGINE structures.
4512 - Constified various get/set functions as appropriate and added
4513 missing functions (including a catch-all ENGINE_cpy that duplicates
4514 all ENGINE values onto a new ENGINE except reference counts/state).
4515 - Removed NULL parameter checks in get/set functions. Setting a method
4516 or function to NULL is a way of cancelling out a previously set
4517 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4518 and doesn't justify the extra error symbols and code.
4519 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4520 flags from engine_int.h to engine.h.
4521 - Changed prototypes for ENGINE handler functions (init(), finish(),
4522 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4525 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4526 to the algorithm using long division. The binary algorithm can be
4527 used only if the modulus is odd. On 32-bit systems, it is faster
4528 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4529 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4530 up to 450 bits. In 64-bit environments, the binary algorithm
4531 appears to be advantageous for much longer moduli; here we use it
4532 for moduli up to 2048 bits.
4535 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4536 could not support the combine flag in choice fields.
4539 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4540 extensions from a certificate request to the certificate.
4543 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4544 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4545 file: this allows the display of the certificate about to be
4546 signed to be customised, to allow certain fields to be included
4547 or excluded and extension details. The old system didn't display
4548 multicharacter strings properly, omitted fields not in the policy
4549 and couldn't display additional details such as extensions.
4552 *) Function EC_POINTs_mul for multiple scalar multiplication
4553 of an arbitrary number of elliptic curve points
4554 \sum scalars[i]*points[i],
4555 optionally including the generator defined for the EC_GROUP:
4556 scalar*generator + \sum scalars[i]*points[i].
4558 EC_POINT_mul is a simple wrapper function for the typical case
4559 that the point list has just one item (besides the optional
4563 *) First EC_METHODs for curves over GF(p):
4565 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4566 operations and provides various method functions that can also
4567 operate with faster implementations of modular arithmetic.
4569 EC_GFp_mont_method() reuses most functions that are part of
4570 EC_GFp_simple_method, but uses Montgomery arithmetic.
4572 [Bodo Moeller; point addition and point doubling
4573 implementation directly derived from source code provided by
4574 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4576 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4577 crypto/ec/ec_lib.c):
4579 Curves are EC_GROUP objects (with an optional group generator)
4580 based on EC_METHODs that are built into the library.
4582 Points are EC_POINT objects based on EC_GROUP objects.
4584 Most of the framework would be able to handle curves over arbitrary
4585 finite fields, but as there are no obvious types for fields other
4586 than GF(p), some functions are limited to that for now.
4589 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4590 that the file contains a complete HTTP response.
4593 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4594 change the def and num file printf format specifier from "%-40sXXX"
4595 to "%-39s XXX". The latter will always guarantee a space after the
4596 field while the former will cause them to run together if the field
4597 is 40 of more characters long.
4600 *) Constify the cipher and digest 'method' functions and structures
4601 and modify related functions to take constant EVP_MD and EVP_CIPHER
4605 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4606 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4609 *) Modify EVP_Digest*() routines so they now return values. Although the
4610 internal software routines can never fail additional hardware versions
4614 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4616 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4617 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4620 ERR_R_NESTED_ASN1_ERROR
4622 ERR_R_MISSING_ASN1_EOS
4623 were 4 .. 9, conflicting with
4624 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4626 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4627 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4629 Add new error code 'ERR_R_INTERNAL_ERROR'.
4632 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4636 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4637 sets the subject name for a new request or supersedes the
4638 subject name in a given request. Formats that can be parsed are
4639 'CN=Some Name, OU=myOU, C=IT'
4641 'CN=Some Name/OU=myOU/C=IT'.
4643 Add options '-batch' and '-verbose' to 'openssl req'.
4644 [Massimiliano Pala <madwolf@hackmasters.net>]
4646 *) Introduce the possibility to access global variables through
4647 functions on platform were that's the best way to handle exporting
4648 global variables in shared libraries. To enable this functionality,
4649 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4650 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4651 is normally done by Configure or something similar).
4653 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4654 in the source file (foo.c) like this:
4656 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4657 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4659 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4660 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4662 OPENSSL_DECLARE_GLOBAL(int,foo);
4663 #define foo OPENSSL_GLOBAL_REF(foo)
4664 OPENSSL_DECLARE_GLOBAL(double,bar);
4665 #define bar OPENSSL_GLOBAL_REF(bar)
4667 The #defines are very important, and therefore so is including the
4668 header file everywhere where the defined globals are used.
4670 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4671 of ASN.1 items, but that structure is a bit different.
4673 The largest change is in util/mkdef.pl which has been enhanced with
4674 better and easier to understand logic to choose which symbols should
4675 go into the Windows .def files as well as a number of fixes and code
4676 cleanup (among others, algorithm keywords are now sorted
4677 lexicographically to avoid constant rewrites).
4680 *) In BN_div() keep a copy of the sign of 'num' before writing the
4681 result to 'rm' because if rm==num the value will be overwritten
4682 and produce the wrong result if 'num' is negative: this caused
4683 problems with BN_mod() and BN_nnmod().
4686 *) Function OCSP_request_verify(). This checks the signature on an
4687 OCSP request and verifies the signer certificate. The signer
4688 certificate is just checked for a generic purpose and OCSP request
4692 *) Add OCSP_check_validity() function to check the validity of OCSP
4693 responses. OCSP responses are prepared in real time and may only
4694 be a few seconds old. Simply checking that the current time lies
4695 between thisUpdate and nextUpdate max reject otherwise valid responses
4696 caused by either OCSP responder or client clock inaccuracy. Instead
4697 we allow thisUpdate and nextUpdate to fall within a certain period of
4698 the current time. The age of the response can also optionally be
4699 checked. Two new options -validity_period and -status_age added to
4703 *) If signature or public key algorithm is unrecognized print out its
4704 OID rather that just UNKNOWN.
4707 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4708 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4709 ID to be generated from the issuer certificate alone which can then be
4710 passed to OCSP_id_issuer_cmp().
4713 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4714 ASN1 modules to export functions returning ASN1_ITEM pointers
4715 instead of the ASN1_ITEM structures themselves. This adds several
4716 new macros which allow the underlying ASN1 function/structure to
4717 be accessed transparently. As a result code should not use ASN1_ITEM
4718 references directly (such as &X509_it) but instead use the relevant
4719 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4720 use of the new ASN1 code on platforms where exporting structures
4721 is problematical (for example in shared libraries) but exporting
4722 functions returning pointers to structures is not.
4725 *) Add support for overriding the generation of SSL/TLS session IDs.
4726 These callbacks can be registered either in an SSL_CTX or per SSL.
4727 The purpose of this is to allow applications to control, if they wish,
4728 the arbitrary values chosen for use as session IDs, particularly as it
4729 can be useful for session caching in multiple-server environments. A
4730 command-line switch for testing this (and any client code that wishes
4731 to use such a feature) has been added to "s_server".
4732 [Geoff Thorpe, Lutz Jaenicke]
4734 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4735 of the form '#if defined(...) || defined(...) || ...' and
4736 '#if !defined(...) && !defined(...) && ...'. This also avoids
4737 the growing number of special cases it was previously handling.
4740 *) Make all configuration macros available for application by making
4741 sure they are available in opensslconf.h, by giving them names starting
4742 with "OPENSSL_" to avoid conflicts with other packages and by making
4743 sure e_os2.h will cover all platform-specific cases together with
4745 Additionally, it is now possible to define configuration/platform-
4746 specific names (called "system identities"). In the C code, these
4747 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4748 macro with the name beginning with "OPENSSL_SYS_", which is determined
4749 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4753 *) New option -set_serial to 'req' and 'x509' this allows the serial
4754 number to use to be specified on the command line. Previously self
4755 signed certificates were hard coded with serial number 0 and the
4756 CA options of 'x509' had to use a serial number in a file which was
4760 *) New options to 'ca' utility to support V2 CRL entry extensions.
4761 Currently CRL reason, invalidity date and hold instruction are
4762 supported. Add new CRL extensions to V3 code and some new objects.
4765 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4766 disable standard block padding (aka PKCS#5 padding) in the EVP
4767 API, which was previously mandatory. This means that the data is
4768 not padded in any way and so the total length much be a multiple
4769 of the block size, otherwise an error occurs.
4772 *) Initial (incomplete) OCSP SSL support.
4775 *) New function OCSP_parse_url(). This splits up a URL into its host,
4776 port and path components: primarily to parse OCSP URLs. New -url
4777 option to ocsp utility.
4780 *) New nonce behavior. The return value of OCSP_check_nonce() now
4781 reflects the various checks performed. Applications can decide
4782 whether to tolerate certain situations such as an absent nonce
4783 in a response when one was present in a request: the ocsp application
4784 just prints out a warning. New function OCSP_add1_basic_nonce()
4785 this is to allow responders to include a nonce in a response even if
4786 the request is nonce-less.
4789 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4790 skipped when using openssl x509 multiple times on a single input file,
4791 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4794 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4795 set string type: to handle setting ASN1_TIME structures. Fix ca
4796 utility to correctly initialize revocation date of CRLs.
4799 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4800 the clients preferred ciphersuites and rather use its own preferences.
4801 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4802 Internet Explorer by ensuring unchanged hash method during stepup.
4803 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4806 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4807 to aes and add a new 'exist' option to print out symbols that don't
4811 *) Additional options to ocsp utility to allow flags to be set and
4812 additional certificates supplied.
4815 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4816 OCSP client a number of certificate to only verify the response
4820 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4821 handle the new API. Currently only ECB, CBC modes supported. Add new
4824 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4825 Encryption Standard (AES) Ciphersuites for Transport Layer
4826 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4827 not enabled by default and were not part of the "ALL" ciphersuite
4828 alias because they were not yet official; they could be
4829 explicitly requested by specifying the "AESdraft" ciphersuite
4830 group alias. In the final release of OpenSSL 0.9.7, the group
4831 alias is called "AES" and is part of "ALL".)
4832 [Ben Laurie, Steve Henson, Bodo Moeller]
4834 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4835 request to response.
4838 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4839 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4840 extract information from a certificate request. OCSP_response_create()
4841 creates a response and optionally adds a basic response structure.
4842 OCSP_basic_add1_status() adds a complete single response to a basic
4843 response and returns the OCSP_SINGLERESP structure just added (to allow
4844 extensions to be included for example). OCSP_basic_add1_cert() adds a
4845 certificate to a basic response and OCSP_basic_sign() signs a basic
4846 response with various flags. New helper functions ASN1_TIME_check()
4847 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4848 (converts ASN1_TIME to GeneralizedTime).
4851 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4852 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4853 structure from a certificate. X509_pubkey_digest() digests the public_key
4854 contents: this is used in various key identifiers.
4857 *) Make sk_sort() tolerate a NULL argument.
4858 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4860 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4861 passed by the function are trusted implicitly. If any of them signed the
4862 response then it is assumed to be valid and is not verified.
4865 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4866 to data. This was previously part of the PKCS7 ASN1 code. This
4867 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4868 [Steve Henson, reported by Kenneth R. Robinette
4869 <support@securenetterm.com>]
4871 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4872 routines: without these tracing memory leaks is very painful.
4873 Fix leaks in PKCS12 and PKCS7 routines.
4876 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4877 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4878 effectively meant GeneralizedTime would never be used. Now it
4879 is initialised to -1 but X509_time_adj() now has to check the value
4880 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4881 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4882 [Steve Henson, reported by Kenneth R. Robinette
4883 <support@securenetterm.com>]
4885 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4886 result in a zero length in the ASN1_INTEGER structure which was
4887 not consistent with the structure when d2i_ASN1_INTEGER() was used
4888 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4889 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4890 where it did not print out a minus for negative ASN1_INTEGER.
4893 *) Add summary printout to ocsp utility. The various functions which
4894 convert status values to strings have been renamed to:
4895 OCSP_response_status_str(), OCSP_cert_status_str() and
4896 OCSP_crl_reason_str() and are no longer static. New options
4897 to verify nonce values and to disable verification. OCSP response
4898 printout format cleaned up.
4901 *) Add additional OCSP certificate checks. These are those specified
4902 in RFC2560. This consists of two separate checks: the CA of the
4903 certificate being checked must either be the OCSP signer certificate
4904 or the issuer of the OCSP signer certificate. In the latter case the
4905 OCSP signer certificate must contain the OCSP signing extended key
4906 usage. This check is performed by attempting to match the OCSP
4907 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4908 in the OCSP_CERTID structures of the response.
4911 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4912 and related routines. This uses the standard OpenSSL certificate
4913 verify routines to perform initial checks (just CA validity) and
4914 to obtain the certificate chain. Then additional checks will be
4915 performed on the chain. Currently the root CA is checked to see
4916 if it is explicitly trusted for OCSP signing. This is used to set
4917 a root CA as a global signing root: that is any certificate that
4918 chains to that CA is an acceptable OCSP signing certificate.
4921 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4922 extensions from a separate configuration file.
4923 As when reading extensions from the main configuration file,
4924 the '-extensions ...' option may be used for specifying the
4926 [Massimiliano Pala <madwolf@comune.modena.it>]
4928 *) New OCSP utility. Allows OCSP requests to be generated or
4929 read. The request can be sent to a responder and the output
4930 parsed, outputed or printed in text form. Not complete yet:
4931 still needs to check the OCSP response validity.
4934 *) New subcommands for 'openssl ca':
4935 'openssl ca -status <serial>' prints the status of the cert with
4936 the given serial number (according to the index file).
4937 'openssl ca -updatedb' updates the expiry status of certificates
4939 [Massimiliano Pala <madwolf@comune.modena.it>]
4941 *) New '-newreq-nodes' command option to CA.pl. This is like
4942 '-newreq', but calls 'openssl req' with the '-nodes' option
4943 so that the resulting key is not encrypted.
4944 [Damien Miller <djm@mindrot.org>]
4946 *) New configuration for the GNU Hurd.
4947 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4949 *) Initial code to implement OCSP basic response verify. This
4950 is currently incomplete. Currently just finds the signer's
4951 certificate and verifies the signature on the response.
4954 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4955 value of OPENSSLDIR. This is available via the new '-d' option
4956 to 'openssl version', and is also included in 'openssl version -a'.
4959 *) Allowing defining memory allocation callbacks that will be given
4960 file name and line number information in additional arguments
4961 (a const char* and an int). The basic functionality remains, as
4962 well as the original possibility to just replace malloc(),
4963 realloc() and free() by functions that do not know about these
4964 additional arguments. To register and find out the current
4965 settings for extended allocation functions, the following
4966 functions are provided:
4968 CRYPTO_set_mem_ex_functions
4969 CRYPTO_set_locked_mem_ex_functions
4970 CRYPTO_get_mem_ex_functions
4971 CRYPTO_get_locked_mem_ex_functions
4973 These work the same way as CRYPTO_set_mem_functions and friends.
4974 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4975 extended allocation function is enabled.
4976 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4977 a conventional allocation function is enabled.
4978 [Richard Levitte, Bodo Moeller]
4980 *) Finish off removing the remaining LHASH function pointer casts.
4981 There should no longer be any prototype-casting required when using
4982 the LHASH abstraction, and any casts that remain are "bugs". See
4983 the callback types and macros at the head of lhash.h for details
4984 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4987 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4988 If /dev/[u]random devices are not available or do not return enough
4989 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4991 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4992 /etc/entropy will be queried once each in this sequence, quering stops
4993 when enough entropy was collected without querying more sockets.
4996 *) Change the Unix RAND_poll() variant to be able to poll several
4997 random devices, as specified by DEVRANDOM, until a sufficient amount
4998 of data has been collected. We spend at most 10 ms on each file
4999 (select timeout) and read in non-blocking mode. DEVRANDOM now
5000 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5001 (previously it was just the string "/dev/urandom"), so on typical
5002 platforms the 10 ms delay will never occur.
5003 Also separate out the Unix variant to its own file, rand_unix.c.
5004 For VMS, there's a currently-empty rand_vms.c.
5007 *) Move OCSP client related routines to ocsp_cl.c. These
5008 provide utility functions which an application needing
5009 to issue a request to an OCSP responder and analyse the
5010 response will typically need: as opposed to those which an
5011 OCSP responder itself would need which will be added later.
5013 OCSP_request_sign() signs an OCSP request with an API similar
5014 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5015 response. OCSP_response_get1_basic() extracts basic response
5016 from response. OCSP_resp_find_status(): finds and extracts status
5017 information from an OCSP_CERTID structure (which will be created
5018 when the request structure is built). These are built from lower
5019 level functions which work on OCSP_SINGLERESP structures but
5020 wont normally be used unless the application wishes to examine
5021 extensions in the OCSP response for example.
5023 Replace nonce routines with a pair of functions.
5024 OCSP_request_add1_nonce() adds a nonce value and optionally
5025 generates a random value. OCSP_check_nonce() checks the
5026 validity of the nonce in an OCSP response.
5029 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5030 This doesn't copy the supplied OCSP_CERTID and avoids the
5031 need to free up the newly created id. Change return type
5032 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5033 This can then be used to add extensions to the request.
5034 Deleted OCSP_request_new(), since most of its functionality
5035 is now in OCSP_REQUEST_new() (and the case insensitive name
5036 clash) apart from the ability to set the request name which
5037 will be added elsewhere.
5040 *) Update OCSP API. Remove obsolete extensions argument from
5041 various functions. Extensions are now handled using the new
5042 OCSP extension code. New simple OCSP HTTP function which
5043 can be used to send requests and parse the response.
5046 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5047 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5048 uses the special reorder version of SET OF to sort the attributes
5049 and reorder them to match the encoded order. This resolves a long
5050 standing problem: a verify on a PKCS7 structure just after signing
5051 it used to fail because the attribute order did not match the
5052 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5053 it uses the received order. This is necessary to tolerate some broken
5054 software that does not order SET OF. This is handled by encoding
5055 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5056 to produce the required SET OF.
5059 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5060 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5061 files to get correct declarations of the ASN.1 item variables.
5064 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5065 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5066 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5067 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5068 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5069 ASN1_ITEM and no wrapper functions.
5072 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5073 replace the old function pointer based I/O routines. Change most of
5074 the *_d2i_bio() and *_d2i_fp() functions to use these.
5077 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5078 lines, recognice more "algorithms" that can be deselected, and make
5079 it complain about algorithm deselection that isn't recognised.
5082 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5083 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5084 to use new functions. Add NO_ASN1_OLD which can be set to remove
5085 some old style ASN1 functions: this can be used to determine if old
5086 code will still work when these eventually go away.
5089 *) New extension functions for OCSP structures, these follow the
5090 same conventions as certificates and CRLs.
5093 *) New function X509V3_add1_i2d(). This automatically encodes and
5094 adds an extension. Its behaviour can be customised with various
5095 flags to append, replace or delete. Various wrappers added for
5096 certifcates and CRLs.
5099 *) Fix to avoid calling the underlying ASN1 print routine when
5100 an extension cannot be parsed. Correct a typo in the
5101 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5104 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5105 entries for variables.
5108 *) Add functionality to apps/openssl.c for detecting locking
5109 problems: As the program is single-threaded, all we have
5110 to do is register a locking callback using an array for
5111 storing which locks are currently held by the program.
5114 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5115 SSL_get_ex_data_X509_STORE_idx(), which is used in
5116 ssl_verify_cert_chain() and thus can be called at any time
5117 during TLS/SSL handshakes so that thread-safety is essential.
5118 Unfortunately, the ex_data design is not at all suited
5119 for multi-threaded use, so it probably should be abolished.
5122 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5123 [Broadcom, tweaked and integrated by Geoff Thorpe]
5125 *) Move common extension printing code to new function
5126 X509V3_print_extensions(). Reorganise OCSP print routines and
5127 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5130 *) New function X509_signature_print() to remove duplication in some
5134 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5135 set (this was treated exactly the same as SET OF previously). This
5136 is used to reorder the STACK representing the structure to match the
5137 encoding. This will be used to get round a problem where a PKCS7
5138 structure which was signed could not be verified because the STACK
5139 order did not reflect the encoded order.
5142 *) Reimplement the OCSP ASN1 module using the new code.
5145 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5146 for its ASN1 operations. The old style function pointers still exist
5147 for now but they will eventually go away.
5150 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5151 completely replaces the old ASN1 functionality with a table driven
5152 encoder and decoder which interprets an ASN1_ITEM structure describing
5153 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5154 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5155 has also been converted to the new form.
5158 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5159 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5160 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5161 for negative moduli.
5164 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5165 of not touching the result's sign bit.
5168 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5172 *) Changed the LHASH code to use prototypes for callbacks, and created
5173 macros to declare and implement thin (optionally static) functions
5174 that provide type-safety and avoid function pointer casting for the
5175 type-specific callbacks.
5178 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5180 [Veers Staats <staatsvr@asc.hpc.mil>,
5181 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5183 *) Reformat the FAQ so the different questions and answers can be divided
5184 in sections depending on the subject.
5187 *) Have the zlib compression code load ZLIB.DLL dynamically under
5191 *) New function BN_mod_sqrt for computing square roots modulo a prime
5192 (using the probabilistic Tonelli-Shanks algorithm unless
5193 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5194 be handled deterministically).
5195 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5197 *) Make BN_mod_inverse faster by explicitly handling small quotients
5198 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5199 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5202 *) New function BN_kronecker.
5205 *) Fix BN_gcd so that it works on negative inputs; the result is
5206 positive unless both parameters are zero.
5207 Previously something reasonably close to an infinite loop was
5208 possible because numbers could be growing instead of shrinking
5209 in the implementation of Euclid's algorithm.
5212 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5213 sign of the number in question.
5215 Fix BN_is_word(a,w) to work correctly for w == 0.
5217 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5218 because its test if the absolute value of 'a' equals 'w'.
5219 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5220 it exists mostly for use in the implementations of BN_is_zero(),
5221 BN_is_one(), and BN_is_word().
5224 *) New function BN_swap.
5227 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5228 the exponentiation functions are more likely to produce reasonable
5229 results on negative inputs.
5232 *) Change BN_mod_mul so that the result is always non-negative.
5233 Previously, it could be negative if one of the factors was negative;
5234 I don't think anyone really wanted that behaviour.
5237 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5238 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5239 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5240 and add new functions:
5249 BN_mod_lshift1_quick
5253 These functions always generate non-negative results.
5255 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5256 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5258 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5259 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5260 be reduced modulo m.
5261 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5264 The following entry accidentily appeared in the CHANGES file
5265 distributed with OpenSSL 0.9.7. The modifications described in
5266 it do *not* apply to OpenSSL 0.9.7.
5268 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5269 was actually never needed) and in BN_mul(). The removal in BN_mul()
5270 required a small change in bn_mul_part_recursive() and the addition
5271 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5272 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5273 bn_sub_words() and bn_add_words() except they take arrays with
5278 *) In 'openssl passwd', verify passwords read from the terminal
5279 unless the '-salt' option is used (which usually means that
5280 verification would just waste user's time since the resulting
5281 hash is going to be compared with some given password hash)
5282 or the new '-noverify' option is used.
5284 This is an incompatible change, but it does not affect
5285 non-interactive use of 'openssl passwd' (passwords on the command
5286 line, '-stdin' option, '-in ...' option) and thus should not
5290 *) Remove all references to RSAref, since there's no more need for it.
5293 *) Make DSO load along a path given through an environment variable
5294 (SHLIB_PATH) with shl_load().
5297 *) Constify the ENGINE code as a result of BIGNUM constification.
5298 Also constify the RSA code and most things related to it. In a
5299 few places, most notable in the depth of the ASN.1 code, ugly
5300 casts back to non-const were required (to be solved at a later
5304 *) Make it so the openssl application has all engines loaded by default.
5307 *) Constify the BIGNUM routines a little more.
5310 *) Add the following functions:
5312 ENGINE_load_cswift()
5314 ENGINE_load_atalla()
5316 ENGINE_load_builtin_engines()
5318 That way, an application can itself choose if external engines that
5319 are built-in in OpenSSL shall ever be used or not. The benefit is
5320 that applications won't have to be linked with libdl or other dso
5321 libraries unless it's really needed.
5323 Changed 'openssl engine' to load all engines on demand.
5324 Changed the engine header files to avoid the duplication of some
5325 declarations (they differed!).
5328 *) 'openssl engine' can now list capabilities.
5331 *) Better error reporting in 'openssl engine'.
5334 *) Never call load_dh_param(NULL) in s_server.
5337 *) Add engine application. It can currently list engines by name and
5338 identity, and test if they are actually available.
5341 *) Improve RPM specification file by forcing symbolic linking and making
5342 sure the installed documentation is also owned by root.root.
5343 [Damien Miller <djm@mindrot.org>]
5345 *) Give the OpenSSL applications more possibilities to make use of
5346 keys (public as well as private) handled by engines.
5349 *) Add OCSP code that comes from CertCo.
5352 *) Add VMS support for the Rijndael code.
5355 *) Added untested support for Nuron crypto accelerator.
5358 *) Add support for external cryptographic devices. This code was
5359 previously distributed separately as the "engine" branch.
5360 [Geoff Thorpe, Richard Levitte]
5362 *) Rework the filename-translation in the DSO code. It is now possible to
5363 have far greater control over how a "name" is turned into a filename
5364 depending on the operating environment and any oddities about the
5365 different shared library filenames on each system.
5368 *) Support threads on FreeBSD-elf in Configure.
5371 *) Fix for SHA1 assembly problem with MASM: it produces
5372 warnings about corrupt line number information when assembling
5373 with debugging information. This is caused by the overlapping
5375 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5378 NCONF_get_number() has no error checking at all. As a replacement,
5379 NCONF_get_number_e() is defined (_e for "error checking") and is
5380 promoted strongly. The old NCONF_get_number is kept around for
5381 binary backward compatibility.
5382 Make it possible for methods to load from something other than a BIO,
5383 by providing a function pointer that is given a name instead of a BIO.
5384 For example, this could be used to load configuration data from an
5388 *) Fix for non blocking accept BIOs. Added new I/O special reason
5389 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5390 with non blocking I/O was not possible because no retry code was
5391 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5395 *) Added the beginnings of Rijndael support.
5398 *) Fix for bug in DirectoryString mask setting. Add support for
5399 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5400 to allow certificate printing to more controllable, additional
5401 'certopt' option to 'x509' to allow new printing options to be
5405 *) Clean old EAY MD5 hack from e_os.h.
5408 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5410 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5411 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5412 [Joe Orton, Steve Henson]
5414 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5416 *) Fix additional bug revealed by the NISCC test suite:
5418 Stop bug triggering large recursion when presented with
5419 certain ASN.1 tags (CVE-2003-0851)
5422 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5424 *) Fix various bugs revealed by running the NISCC test suite:
5426 Stop out of bounds reads in the ASN1 code when presented with
5427 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5429 If verify callback ignores invalid public key errors don't try to check
5430 certificate signature with the NULL public key.
5434 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5435 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5439 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5440 extra data after the compression methods not only for TLS 1.0
5441 but also for SSL 3.0 (as required by the specification).
5442 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5444 *) Change X509_certificate_type() to mark the key as exported/exportable
5445 when it's 512 *bits* long, not 512 bytes.
5448 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5450 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5451 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5452 a protocol version number mismatch like a decryption error
5453 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5456 *) Turn on RSA blinding by default in the default implementation
5457 to avoid a timing attack. Applications that don't want it can call
5458 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5459 They would be ill-advised to do so in most cases.
5460 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5462 *) Change RSA blinding code so that it works when the PRNG is not
5463 seeded (in this case, the secret RSA exponent is abused as
5464 an unpredictable seed -- if it is not unpredictable, there
5465 is no point in blinding anyway). Make RSA blinding thread-safe
5466 by remembering the creator's thread ID in rsa->blinding and
5467 having all other threads use local one-time blinding factors
5468 (this requires more computation than sharing rsa->blinding, but
5469 avoids excessive locking; and if an RSA object is not shared
5470 between threads, blinding will still be very fast).
5473 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5475 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5476 via timing by performing a MAC computation even if incorrrect
5477 block cipher padding has been found. This is a countermeasure
5478 against active attacks where the attacker has to distinguish
5479 between bad padding and a MAC verification error. (CVE-2003-0078)
5481 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5482 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5483 Martin Vuagnoux (EPFL, Ilion)]
5485 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5487 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5488 memory from it's contents. This is done with a counter that will
5489 place alternating values in each byte. This can be used to solve
5490 two issues: 1) the removal of calls to memset() by highly optimizing
5491 compilers, and 2) cleansing with other values than 0, since those can
5492 be read through on certain media, for example a swap space on disk.
5495 *) Bugfix: client side session caching did not work with external caching,
5496 because the session->cipher setting was not restored when reloading
5497 from the external cache. This problem was masked, when
5498 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5499 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5502 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5503 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5504 [Zeev Lieber <zeev-l@yahoo.com>]
5506 *) Undo an undocumented change introduced in 0.9.6e which caused
5507 repeated calls to OpenSSL_add_all_ciphers() and
5508 OpenSSL_add_all_digests() to be ignored, even after calling
5512 *) Change the default configuration reader to deal with last line not
5513 being properly terminated.
5516 *) Change X509_NAME_cmp() so it applies the special rules on handling
5517 DN values that are of type PrintableString, as well as RDNs of type
5518 emailAddress where the value has the type ia5String.
5519 [stefank@valicert.com via Richard Levitte]
5521 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5522 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5523 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5524 the bitwise-OR of the two for use by the majority of applications
5525 wanting this behaviour, and update the docs. The documented
5526 behaviour and actual behaviour were inconsistent and had been
5527 changing anyway, so this is more a bug-fix than a behavioural
5529 [Geoff Thorpe, diagnosed by Nadav Har'El]
5531 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5532 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5535 *) Fix initialization code race conditions in
5536 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5537 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5538 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5539 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5540 ssl2_get_cipher_by_char(),
5541 ssl3_get_cipher_by_char().
5542 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5544 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5545 the cached sessions are flushed, as the remove_cb() might use ex_data
5546 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5547 (see [openssl.org #212]).
5548 [Geoff Thorpe, Lutz Jaenicke]
5550 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5551 length, instead of the encoding length to d2i_ASN1_OBJECT.
5554 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5556 *) [In 0.9.6g-engine release:]
5557 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5558 [Lynn Gazis <lgazis@rainbow.com>]
5560 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5562 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5563 and get fix the header length calculation.
5564 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5565 Alon Kantor <alonk@checkpoint.com> (and others),
5568 *) Use proper error handling instead of 'assertions' in buffer
5569 overflow checks added in 0.9.6e. This prevents DoS (the
5570 assertions could call abort()).
5571 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5573 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5575 *) Add various sanity checks to asn1_get_length() to reject
5576 the ASN1 length bytes if they exceed sizeof(long), will appear
5577 negative or the content length exceeds the length of the
5579 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5581 *) Fix cipher selection routines: ciphers without encryption had no flags
5582 for the cipher strength set and where therefore not handled correctly
5583 by the selection routines (PR #130).
5586 *) Fix EVP_dsa_sha macro.
5590 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5591 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5592 that was added in OpenSSL 0.9.6d.
5594 As the countermeasure turned out to be incompatible with some
5595 broken SSL implementations, the new option is part of SSL_OP_ALL.
5596 SSL_OP_ALL is usually employed when compatibility with weird SSL
5597 implementations is desired (e.g. '-bugs' option to 's_client' and
5598 's_server'), so the new option is automatically set in many
5602 *) Changes in security patch:
5604 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5605 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5606 Air Force Materiel Command, USAF, under agreement number
5609 *) Add various sanity checks to asn1_get_length() to reject
5610 the ASN1 length bytes if they exceed sizeof(long), will appear
5611 negative or the content length exceeds the length of the
5612 supplied buffer. (CVE-2002-0659)
5613 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5615 *) Assertions for various potential buffer overflows, not known to
5617 [Ben Laurie (CHATS)]
5619 *) Various temporary buffers to hold ASCII versions of integers were
5620 too small for 64 bit platforms. (CVE-2002-0655)
5621 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5623 *) Remote buffer overflow in SSL3 protocol - an attacker could
5624 supply an oversized session ID to a client. (CVE-2002-0656)
5625 [Ben Laurie (CHATS)]
5627 *) Remote buffer overflow in SSL2 protocol - an attacker could
5628 supply an oversized client master key. (CVE-2002-0656)
5629 [Ben Laurie (CHATS)]
5631 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5633 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5634 encoded as NULL) with id-dsa-with-sha1.
5635 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5637 *) Check various X509_...() return values in apps/req.c.
5638 [Nils Larsch <nla@trustcenter.de>]
5640 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5641 an end-of-file condition would erronously be flagged, when the CRLF
5642 was just at the end of a processed block. The bug was discovered when
5643 processing data through a buffering memory BIO handing the data to a
5644 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5645 <ptsekov@syntrex.com> and Nedelcho Stanev.
5648 *) Implement a countermeasure against a vulnerability recently found
5649 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5650 before application data chunks to avoid the use of known IVs
5651 with data potentially chosen by the attacker.
5654 *) Fix length checks in ssl3_get_client_hello().
5657 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5658 to prevent ssl3_read_internal() from incorrectly assuming that
5659 ssl3_read_bytes() found application data while handshake
5660 processing was enabled when in fact s->s3->in_read_app_data was
5661 merely automatically cleared during the initial handshake.
5662 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5664 *) Fix object definitions for Private and Enterprise: they were not
5665 recognized in their shortname (=lowercase) representation. Extend
5666 obj_dat.pl to issue an error when using undefined keywords instead
5667 of silently ignoring the problem (Svenning Sorensen
5668 <sss@sss.dnsalias.net>).
5671 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5672 generators, i.e. generators other than 2 and 5. (Previously, the
5673 code did not properly initialise the 'add' and 'rem' values to
5674 BN_generate_prime().)
5676 In the new general case, we do not insist that 'generator' is
5677 actually a primitive root: This requirement is rather pointless;
5678 a generator of the order-q subgroup is just as good, if not
5682 *) Map new X509 verification errors to alerts. Discovered and submitted by
5683 Tom Wu <tom@arcot.com>.
5686 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5687 returning non-zero before the data has been completely received
5688 when using non-blocking I/O.
5689 [Bodo Moeller; problem pointed out by John Hughes]
5691 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5692 [Ben Laurie, Lutz Jaenicke]
5694 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5695 Yoram Zahavi <YoramZ@gilian.com>).
5698 *) Add information about CygWin 1.3 and on, and preserve proper
5699 configuration for the versions before that.
5700 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5702 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5703 check whether we deal with a copy of a session and do not delete from
5704 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5705 <izhar@checkpoint.com>.
5708 *) Do not store session data into the internal session cache, if it
5709 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5710 flag is set). Proposed by Aslam <aslam@funk.com>.
5713 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5717 *) [In 0.9.6d-engine release:]
5718 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5719 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5721 *) Add the configuration target linux-s390x.
5722 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5724 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5725 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5726 variable as an indication that a ClientHello message has been
5727 received. As the flag value will be lost between multiple
5728 invocations of ssl3_accept when using non-blocking I/O, the
5729 function may not be aware that a handshake has actually taken
5730 place, thus preventing a new session from being added to the
5733 To avoid this problem, we now set s->new_session to 2 instead of
5734 using a local variable.
5735 [Lutz Jaenicke, Bodo Moeller]
5737 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5738 if the SSL_R_LENGTH_MISMATCH error is detected.
5739 [Geoff Thorpe, Bodo Moeller]
5741 *) New 'shared_ldflag' column in Configure platform table.
5744 *) Fix EVP_CIPHER_mode macro.
5745 ["Dan S. Camper" <dan@bti.net>]
5747 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5748 type, we must throw them away by setting rr->length to 0.
5749 [D P Chang <dpc@qualys.com>]
5751 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5753 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5754 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5755 worked incorrectly for those cases where range = 10..._2 and
5756 3*range is two bits longer than range.)
5759 *) Only add signing time to PKCS7 structures if it is not already
5763 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5764 OBJ_ld_ce should be OBJ_id_ce.
5765 Also some ip-pda OIDs in crypto/objects/objects.txt were
5766 incorrect (cf. RFC 3039).
5767 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5769 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5770 returns early because it has nothing to do.
5771 [Andy Schneider <andy.schneider@bjss.co.uk>]
5773 *) [In 0.9.6c-engine release:]
5774 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5775 [Andy Schneider <andy.schneider@bjss.co.uk>]
5777 *) [In 0.9.6c-engine release:]
5778 Add support for Cryptographic Appliance's keyserver technology.
5779 (Use engine 'keyclient')
5780 [Cryptographic Appliances and Geoff Thorpe]
5782 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5783 is called via tools/c89.sh because arguments have to be
5784 rearranged (all '-L' options must appear before the first object
5786 [Richard Shapiro <rshapiro@abinitio.com>]
5788 *) [In 0.9.6c-engine release:]
5789 Add support for Broadcom crypto accelerator cards, backported
5791 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5793 *) [In 0.9.6c-engine release:]
5794 Add support for SureWare crypto accelerator cards from
5795 Baltimore Technologies. (Use engine 'sureware')
5796 [Baltimore Technologies and Mark Cox]
5798 *) [In 0.9.6c-engine release:]
5799 Add support for crypto accelerator cards from Accelerated
5800 Encryption Processing, www.aep.ie. (Use engine 'aep')
5801 [AEP Inc. and Mark Cox]
5803 *) Add a configuration entry for gcc on UnixWare.
5804 [Gary Benson <gbenson@redhat.com>]
5806 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5807 messages are stored in a single piece (fixed-length part and
5808 variable-length part combined) and fix various bugs found on the way.
5811 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5812 instead. BIO_gethostbyname() does not know what timeouts are
5813 appropriate, so entries would stay in cache even when they have
5815 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5817 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5818 faced with a pathologically small ClientHello fragment that does
5819 not contain client_version: Instead of aborting with an error,
5820 simply choose the highest available protocol version (i.e.,
5821 TLS 1.0 unless it is disabled). In practice, ClientHello
5822 messages are never sent like this, but this change gives us
5823 strictly correct behaviour at least for TLS.
5826 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5827 never resets s->method to s->ctx->method when called from within
5828 one of the SSL handshake functions.
5829 [Bodo Moeller; problem pointed out by Niko Baric]
5831 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5832 (sent using the client's version number) if client_version is
5833 smaller than the protocol version in use. Also change
5834 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5835 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5836 the client will at least see that alert.
5839 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5843 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5844 client receives HelloRequest while in a handshake.
5845 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5847 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5848 should end in 'break', not 'goto end' which circuments various
5849 cleanups done in state SSL_ST_OK. But session related stuff
5850 must be disabled for SSL_ST_OK in the case that we just sent a
5853 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5854 before just sending a HelloRequest.
5855 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5857 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5858 reveal whether illegal block cipher padding was found or a MAC
5859 verification error occured. (Neither SSLerr() codes nor alerts
5860 are directly visible to potential attackers, but the information
5861 may leak via logfiles.)
5863 Similar changes are not required for the SSL 2.0 implementation
5864 because the number of padding bytes is sent in clear for SSL 2.0,
5865 and the extra bytes are just ignored. However ssl/s2_pkt.c
5866 failed to verify that the purported number of padding bytes is in
5870 *) Add OpenUNIX-8 support including shared libraries
5871 (Boyd Lynn Gerber <gerberb@zenez.com>).
5874 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5875 'wristwatch attack' using huge encoding parameters (cf.
5876 James H. Manger's CRYPTO 2001 paper). Note that the
5877 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5878 encoding parameters and hence was not vulnerable.
5881 *) BN_sqr() bug fix.
5882 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5884 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5885 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5886 followed by modular reduction.
5887 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5889 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5890 equivalent based on BN_pseudo_rand() instead of BN_rand().
5893 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5894 This function was broken, as the check for a new client hello message
5895 to handle SGC did not allow these large messages.
5896 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5899 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5902 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5903 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5906 *) Rework the configuration and shared library support for Tru64 Unix.
5907 The configuration part makes use of modern compiler features and
5908 still retains old compiler behavior for those that run older versions
5909 of the OS. The shared library support part includes a variant that
5910 uses the RPATH feature, and is available through the special
5911 configuration target "alpha-cc-rpath", which will never be selected
5913 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5915 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5916 with the same message size as in ssl3_get_certificate_request().
5917 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5918 messages might inadvertently be reject as too long.
5919 [Petr Lampa <lampa@fee.vutbr.cz>]
5921 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5924 *) Modified SSL library such that the verify_callback that has been set
5925 specificly for an SSL object with SSL_set_verify() is actually being
5926 used. Before the change, a verify_callback set with this function was
5927 ignored and the verify_callback() set in the SSL_CTX at the time of
5928 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5929 to allow the necessary settings.
5932 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5933 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5934 done automatically (in contradiction to the requirements of the C
5935 standard). This made problems when used from OpenSSH.
5938 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5939 dh->length and always used
5941 BN_rand_range(priv_key, dh->p).
5943 BN_rand_range() is not necessary for Diffie-Hellman, and this
5944 specific range makes Diffie-Hellman unnecessarily inefficient if
5945 dh->length (recommended exponent length) is much smaller than the
5946 length of dh->p. We could use BN_rand_range() if the order of
5947 the subgroup was stored in the DH structure, but we only have
5952 BN_rand(priv_key, l, ...)
5954 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5960 RSA_eay_public_encrypt
5961 RSA_eay_private_decrypt
5962 RSA_eay_private_encrypt (signing)
5963 RSA_eay_public_decrypt (signature verification)
5965 (default implementations for RSA_public_encrypt,
5966 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5967 always reject numbers >= n.
5970 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5971 to synchronize access to 'locking_thread'. This is necessary on
5972 systems where access to 'locking_thread' (an 'unsigned long'
5973 variable) is not atomic.
5976 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5977 *before* setting the 'crypto_lock_rand' flag. The previous code had
5978 a race condition if 0 is a valid thread ID.
5979 [Travis Vitek <vitek@roguewave.com>]
5981 *) Add support for shared libraries under Irix.
5982 [Albert Chin-A-Young <china@thewrittenword.com>]
5984 *) Add configuration option to build on Linux on both big-endian and
5986 [Ralf Baechle <ralf@uni-koblenz.de>]
5988 *) Add the possibility to create shared libraries on HP-UX.
5991 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5993 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
5994 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
5995 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
5996 PRNG state recovery was possible based on the output of
5997 one PRNG request appropriately sized to gain knowledge on
5998 'md' followed by enough consecutive 1-byte PRNG requests
5999 to traverse all of 'state'.
6001 1. When updating 'md_local' (the current thread's copy of 'md')
6002 during PRNG output generation, hash all of the previous
6003 'md_local' value, not just the half used for PRNG output.
6005 2. Make the number of bytes from 'state' included into the hash
6006 independent from the number of PRNG bytes requested.
6008 The first measure alone would be sufficient to avoid
6009 Markku-Juhani's attack. (Actually it had never occurred
6010 to me that the half of 'md_local' used for chaining was the
6011 half from which PRNG output bytes were taken -- I had always
6012 assumed that the secret half would be used.) The second
6013 measure makes sure that additional data from 'state' is never
6014 mixed into 'md_local' in small portions; this heuristically
6015 further strengthens the PRNG.
6018 *) Fix crypto/bn/asm/mips3.s.
6021 *) When only the key is given to "enc", the IV is undefined. Print out
6022 an error message in this case.
6025 *) Handle special case when X509_NAME is empty in X509 printing routines.
6028 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6029 positive and less than q.
6032 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6033 used: it isn't thread safe and the add_lock_callback should handle
6035 [Paul Rose <Paul.Rose@bridge.com>]
6037 *) Verify that incoming data obeys the block size in
6038 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6042 [Ulf Möller, Bodo Möller]
6044 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6045 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6046 when fixing the server behaviour for backwards-compatible 'client
6047 hello' messages. (Note that the attack is impractical against
6048 SSL 3.0 and TLS 1.0 anyway because length and version checking
6049 means that the probability of guessing a valid ciphertext is
6050 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6053 Before 0.9.5, the countermeasure (hide the error by generating a
6054 random 'decryption result') did not work properly because
6055 ERR_clear_error() was missing, meaning that SSL_get_error() would
6056 detect the supposedly ignored error.
6058 Both problems are now fixed.
6061 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6062 (previously it was 1024).
6065 *) Fix for compatibility mode trust settings: ignore trust settings
6066 unless some valid trust or reject settings are present.
6069 *) Fix for blowfish EVP: its a variable length cipher.
6072 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6073 parameters in DSA public key structures and return an error in the
6074 DSA routines if parameters are absent.
6077 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6078 in the current directory if neither $RANDFILE nor $HOME was set.
6079 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6080 caused some confusion to Windows users who haven't defined $HOME.
6081 Thus RAND_file_name() is changed again: e_os.h can define a
6082 DEFAULT_HOME, which will be used if $HOME is not set.
6083 For Windows, we use "C:"; on other platforms, we still require
6084 environment variables.
6086 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6087 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6088 having multiple threads call RAND_poll() concurrently.
6091 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6092 combination of a flag and a thread ID variable.
6093 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6094 flag), *other* threads can enter ssleay_add_bytes without obeying
6095 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6096 that they do not hold after the first thread unsets add_do_not_lock).
6099 *) Change bctest again: '-x' expressions are not available in all
6103 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6105 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6106 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6108 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6109 the default extension for executables, if any. Also, make the perl
6110 scripts that use symlink() to test if it really exists and use "cp"
6111 if it doesn't. All this made OpenSSL compilable and installable in
6115 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6116 If SEQUENCE is length is indefinite just set c->slen to the total
6117 amount of data available.
6118 [Steve Henson, reported by shige@FreeBSD.org]
6119 [This change does not apply to 0.9.7.]
6121 *) Change bctest to avoid here-documents inside command substitution
6122 (workaround for FreeBSD /bin/sh bug).
6123 For compatibility with Ultrix, avoid shell functions (introduced
6124 in the bctest version that searches along $PATH).
6127 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6128 with des_encrypt() defined on some operating systems, like Solaris
6132 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6133 On the Importance of Eliminating Errors in Cryptographic
6134 Computations, J. Cryptology 14 (2001) 2, 101-119,
6135 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6138 *) MIPS assembler BIGNUM division bug fix.
6141 *) Disabled incorrect Alpha assembler code.
6144 *) Fix PKCS#7 decode routines so they correctly update the length
6145 after reading an EOC for the EXPLICIT tag.
6147 [This change does not apply to 0.9.7.]
6149 *) Fix bug in PKCS#12 key generation routines. This was triggered
6150 if a 3DES key was generated with a 0 initial byte. Include
6151 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6152 (but broken) behaviour.
6155 *) Enhance bctest to search for a working bc along $PATH and print
6157 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6159 *) Fix memory leaks in err.c: free err_data string if necessary;
6160 don't write to the wrong index in ERR_set_error_data.
6163 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6167 *) Replace rdtsc with _emit statements for VC++ version 5.
6168 [Jeremy Cooper <jeremy@baymoo.org>]
6170 *) Make it possible to reuse SSLv2 sessions.
6173 *) In copy_email() check for >= 0 as a return value for
6174 X509_NAME_get_index_by_NID() since 0 is a valid index.
6175 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6177 *) Avoid coredump with unsupported or invalid public keys by checking if
6178 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6179 PKCS7_verify() fails with non detached data.
6182 *) Don't use getenv in library functions when run as setuid/setgid.
6183 New function OPENSSL_issetugid().
6186 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6187 due to incorrect handling of multi-threading:
6189 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6191 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6193 3. Count how many times MemCheck_off() has been called so that
6194 nested use can be treated correctly. This also avoids
6195 inband-signalling in the previous code (which relied on the
6196 assumption that thread ID 0 is impossible).
6199 *) Add "-rand" option also to s_client and s_server.
6202 *) Fix CPU detection on Irix 6.x.
6203 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6204 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6206 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6209 [This change does not apply to 0.9.7.]
6211 *) Use the cached encoding of an X509_NAME structure rather than
6212 copying it. This is apparently the reason for the libsafe "errors"
6213 but the code is actually correct.
6216 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6217 Bleichenbacher's DSA attack.
6218 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6219 to be set and top=0 forces the highest bit to be set; top=-1 is new
6220 and leaves the highest bit random.
6221 [Ulf Moeller, Bodo Moeller]
6223 *) In the NCONF_...-based implementations for CONF_... queries
6224 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6225 a temporary CONF structure with the data component set to NULL
6226 (which gives segmentation faults in lh_retrieve).
6227 Instead, use NULL for the CONF pointer in CONF_get_string and
6228 CONF_get_number (which may use environment variables) and directly
6229 return NULL from CONF_get_section.
6232 *) Fix potential buffer overrun for EBCDIC.
6235 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6236 keyUsage if basicConstraints absent for a CA.
6239 *) Make SMIME_write_PKCS7() write mail header values with a format that
6240 is more generally accepted (no spaces before the semicolon), since
6241 some programs can't parse those values properly otherwise. Also make
6242 sure BIO's that break lines after each write do not create invalid
6246 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6247 macros previously used would not encode an empty SEQUENCE OF
6248 and break the signature.
6250 [This change does not apply to 0.9.7.]
6252 *) Zero the premaster secret after deriving the master secret in
6256 *) Add some EVP_add_digest_alias registrations (as found in
6257 OpenSSL_add_all_digests()) to SSL_library_init()
6258 aka OpenSSL_add_ssl_algorithms(). This provides improved
6259 compatibility with peers using X.509 certificates
6260 with unconventional AlgorithmIdentifier OIDs.
6263 *) Fix for Irix with NO_ASM.
6264 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6266 *) ./config script fixes.
6267 [Ulf Moeller, Richard Levitte]
6269 *) Fix 'openssl passwd -1'.
6272 *) Change PKCS12_key_gen_asc() so it can cope with non null
6273 terminated strings whose length is passed in the passlen
6274 parameter, for example from PEM callbacks. This was done
6275 by adding an extra length parameter to asc2uni().
6276 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6278 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6279 call failed, free the DSA structure.
6282 *) Fix to uni2asc() to cope with zero length Unicode strings.
6283 These are present in some PKCS#12 files.
6286 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6287 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6288 when writing a 32767 byte record.
6289 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6291 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6292 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6294 (RSA objects have a reference count access to which is protected
6295 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6296 so they are meant to be shared between threads.)
6297 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6298 "Reddie, Steven" <Steven.Reddie@ca.com>]
6300 *) Fix a deadlock in CRYPTO_mem_leaks().
6303 *) Use better test patterns in bntest.
6306 *) rand_win.c fix for Borland C.
6309 *) BN_rshift bugfix for n == 0.
6312 *) Add a 'bctest' script that checks for some known 'bc' bugs
6313 so that 'make test' does not abort just because 'bc' is broken.
6316 *) Store verify_result within SSL_SESSION also for client side to
6317 avoid potential security hole. (Re-used sessions on the client side
6318 always resulted in verify_result==X509_V_OK, not using the original
6319 result of the server certificate verification.)
6322 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6323 SSL3_RT_APPLICATION_DATA, return 0.
6324 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6328 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6329 releases, have been re-implemented by renaming the previous
6330 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6331 and ssl3_read_internal, respectively, and adding 'peek' parameters
6332 to them. The new ssl[23]_{read,peek} functions are calls to
6333 ssl[23]_read_internal with the 'peek' flag set appropriately.
6334 A 'peek' parameter has also been added to ssl3_read_bytes, which
6335 does the actual work for ssl3_read_internal.
6338 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6339 the method-specific "init()" handler. Also clean up ex_data after
6340 calling the method-specific "finish()" handler. Previously, this was
6341 happening the other way round.
6344 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6345 The previous value, 12, was not always sufficient for BN_mod_exp().
6348 *) Make sure that shared libraries get the internal name engine with
6349 the full version number and not just 0. This should mark the
6350 shared libraries as not backward compatible. Of course, this should
6351 be changed again when we can guarantee backward binary compatibility.
6354 *) Fix typo in get_cert_by_subject() in by_dir.c
6355 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6357 *) Rework the system to generate shared libraries:
6359 - Make note of the expected extension for the shared libraries and
6360 if there is a need for symbolic links from for example libcrypto.so.0
6361 to libcrypto.so.0.9.7. There is extended info in Configure for
6364 - Make as few rebuilds of the shared libraries as possible.
6366 - Still avoid linking the OpenSSL programs with the shared libraries.
6368 - When installing, install the shared libraries separately from the
6372 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6374 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6375 and not in SSL_clear because the latter is also used by the
6376 accept/connect functions; previously, the settings made by
6377 SSL_set_read_ahead would be lost during the handshake.
6378 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6380 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6381 Previously, it would create entries for disableed algorithms no
6385 *) Added several new manual pages for SSL_* function.
6388 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6390 *) In ssl23_get_client_hello, generate an error message when faced
6391 with an initial SSL 3.0/TLS record that is too small to contain the
6392 first two bytes of the ClientHello message, i.e. client_version.
6393 (Note that this is a pathologic case that probably has never happened
6394 in real life.) The previous approach was to use the version number
6395 from the record header as a substitute; but our protocol choice
6396 should not depend on that one because it is not authenticated
6397 by the Finished messages.
6400 *) More robust randomness gathering functions for Windows.
6401 [Jeffrey Altman <jaltman@columbia.edu>]
6403 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6404 not set then we don't setup the error code for issuer check errors
6405 to avoid possibly overwriting other errors which the callback does
6406 handle. If an application does set the flag then we assume it knows
6407 what it is doing and can handle the new informational codes
6411 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6412 a general "ANY" type, as such it should be able to decode anything
6413 including tagged types. However it didn't check the class so it would
6414 wrongly interpret tagged types in the same way as their universal
6415 counterpart and unknown types were just rejected. Changed so that the
6416 tagged and unknown types are handled in the same way as a SEQUENCE:
6417 that is the encoding is stored intact. There is also a new type
6418 "V_ASN1_OTHER" which is used when the class is not universal, in this
6419 case we have no idea what the actual type is so we just lump them all
6423 *) On VMS, stdout may very well lead to a file that is written to
6424 in a record-oriented fashion. That means that every write() will
6425 write a separate record, which will be read separately by the
6426 programs trying to read from it. This can be very confusing.
6428 The solution is to put a BIO filter in the way that will buffer
6429 text until a linefeed is reached, and then write everything a
6430 line at a time, so every record written will be an actual line,
6431 not chunks of lines and not (usually doesn't happen, but I've
6432 seen it once) several lines in one record. BIO_f_linebuffer() is
6435 Currently, it's a VMS-only method, because that's where it has
6436 been tested well enough.
6439 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6440 it can return incorrect results.
6441 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6442 but it was in 0.9.6-beta[12].)
6445 *) Disable the check for content being present when verifying detached
6446 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6447 include zero length content when signing messages.
6450 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6451 BIO_ctrl (for BIO pairs).
6454 *) Add DSO method for VMS.
6457 *) Bug fix: Montgomery multiplication could produce results with the
6461 *) Add RPM specification openssl.spec and modify it to build three
6462 packages. The default package contains applications, application
6463 documentation and run-time libraries. The devel package contains
6464 include files, static libraries and function documentation. The
6465 doc package contains the contents of the doc directory. The original
6466 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6469 *) Add a large number of documentation files for many SSL routines.
6470 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6472 *) Add a configuration entry for Sony News 4.
6473 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6475 *) Don't set the two most significant bits to one when generating a
6476 random number < q in the DSA library.
6479 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6480 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6481 the underlying transport is blocking) if a handshake took place.
6482 (The default behaviour is needed by applications such as s_client
6483 and s_server that use select() to determine when to use SSL_read;
6484 but for applications that know in advance when to expect data, it
6485 just makes things more complicated.)
6488 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6492 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6493 work better on such systems.
6494 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6496 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6497 Update PKCS12_parse() so it copies the friendlyName and the
6498 keyid to the certificates aux info.
6501 *) Fix bug in PKCS7_verify() which caused an infinite loop
6502 if there was more than one signature.
6503 [Sven Uszpelkat <su@celocom.de>]
6505 *) Major change in util/mkdef.pl to include extra information
6506 about each symbol, as well as presentig variables as well
6507 as functions. This change means that there's n more need
6508 to rebuild the .num files when some algorithms are excluded.
6511 *) Allow the verify time to be set by an application,
6512 rather than always using the current time.
6515 *) Phase 2 verify code reorganisation. The certificate
6516 verify code now looks up an issuer certificate by a
6517 number of criteria: subject name, authority key id
6518 and key usage. It also verifies self signed certificates
6519 by the same criteria. The main comparison function is
6520 X509_check_issued() which performs these checks.
6522 Lot of changes were necessary in order to support this
6523 without completely rewriting the lookup code.
6525 Authority and subject key identifier are now cached.
6527 The LHASH 'certs' is X509_STORE has now been replaced
6528 by a STACK_OF(X509_OBJECT). This is mainly because an
6529 LHASH can't store or retrieve multiple objects with
6530 the same hash value.
6532 As a result various functions (which were all internal
6533 use only) have changed to handle the new X509_STORE
6534 structure. This will break anything that messed round
6535 with X509_STORE internally.
6537 The functions X509_STORE_add_cert() now checks for an
6538 exact match, rather than just subject name.
6540 The X509_STORE API doesn't directly support the retrieval
6541 of multiple certificates matching a given criteria, however
6542 this can be worked round by performing a lookup first
6543 (which will fill the cache with candidate certificates)
6544 and then examining the cache for matches. This is probably
6545 the best we can do without throwing out X509_LOOKUP
6546 entirely (maybe later...).
6548 The X509_VERIFY_CTX structure has been enhanced considerably.
6550 All certificate lookup operations now go via a get_issuer()
6551 callback. Although this currently uses an X509_STORE it
6552 can be replaced by custom lookups. This is a simple way
6553 to bypass the X509_STORE hackery necessary to make this
6554 work and makes it possible to use more efficient techniques
6555 in future. A very simple version which uses a simple
6556 STACK for its trusted certificate store is also provided
6557 using X509_STORE_CTX_trusted_stack().
6559 The verify_cb() and verify() callbacks now have equivalents
6560 in the X509_STORE_CTX structure.
6562 X509_STORE_CTX also has a 'flags' field which can be used
6563 to customise the verify behaviour.
6566 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6567 excludes S/MIME capabilities.
6570 *) When a certificate request is read in keep a copy of the
6571 original encoding of the signed data and use it when outputing
6572 again. Signatures then use the original encoding rather than
6573 a decoded, encoded version which may cause problems if the
6574 request is improperly encoded.
6577 *) For consistency with other BIO_puts implementations, call
6578 buffer_write(b, ...) directly in buffer_puts instead of calling
6581 In BIO_puts, increment b->num_write as in BIO_write.
6582 [Peter.Sylvester@EdelWeb.fr]
6584 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6585 BN_zero, we may not return a BIGNUM with an array consisting of
6589 *) Avoid calling abort() from within the library when problems are
6590 detected, except if preprocessor symbols have been defined
6591 (such as REF_CHECK, BN_DEBUG etc.).
6594 *) New openssl application 'rsautl'. This utility can be
6595 used for low level RSA operations. DER public key
6596 BIO/fp routines also added.
6599 *) New Configure entry and patches for compiling on QNX 4.
6600 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6602 *) A demo state-machine implementation was sponsored by
6603 Nuron (http://www.nuron.com/) and is now available in
6604 demos/state_machine.
6607 *) New options added to the 'dgst' utility for signature
6608 generation and verification.
6611 *) Unrecognized PKCS#7 content types are now handled via a
6612 catch all ASN1_TYPE structure. This allows unsupported
6613 types to be stored as a "blob" and an application can
6614 encode and decode it manually.
6617 *) Fix various signed/unsigned issues to make a_strex.c
6619 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6621 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6622 length if passed a buffer. ASN1_INTEGER_to_BN failed
6623 if passed a NULL BN and its argument was negative.
6624 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6626 *) Modification to PKCS#7 encoding routines to output definite
6627 length encoding. Since currently the whole structures are in
6628 memory there's not real point in using indefinite length
6629 constructed encoding. However if OpenSSL is compiled with
6630 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6633 *) Added BIO_vprintf() and BIO_vsnprintf().
6636 *) Added more prefixes to parse for in the the strings written
6637 through a logging bio, to cover all the levels that are available
6638 through syslog. The prefixes are now:
6640 PANIC, EMERG, EMR => LOG_EMERG
6641 ALERT, ALR => LOG_ALERT
6642 CRIT, CRI => LOG_CRIT
6643 ERROR, ERR => LOG_ERR
6644 WARNING, WARN, WAR => LOG_WARNING
6645 NOTICE, NOTE, NOT => LOG_NOTICE
6646 INFO, INF => LOG_INFO
6647 DEBUG, DBG => LOG_DEBUG
6649 and as before, if none of those prefixes are present at the
6650 beginning of the string, LOG_ERR is chosen.
6652 On Win32, the LOG_* levels are mapped according to this:
6654 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6655 LOG_WARNING => EVENTLOG_WARNING_TYPE
6656 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6660 *) Made it possible to reconfigure with just the configuration
6661 argument "reconf" or "reconfigure". The command line arguments
6662 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6663 and are retrieved from there when reconfiguring.
6667 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6669 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6672 *) The obj_dat.pl script was messing up the sorting of object
6673 names. The reason was that it compared the quoted version
6674 of strings as a result "OCSP" > "OCSP Signing" because
6675 " > SPACE. Changed script to store unquoted versions of
6676 names and add quotes on output. It was also omitting some
6677 names from the lookup table if they were given a default
6678 value (that is if SN is missing it is given the same
6679 value as LN and vice versa), these are now added on the
6680 grounds that if an object has a name we should be able to
6681 look it up. Finally added warning output when duplicate
6682 short or long names are found.
6685 *) Changes needed for Tandem NSK.
6686 [Scott Uroff <scott@xypro.com>]
6688 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6689 RSA_padding_check_SSLv23(), special padding was never detected
6690 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6691 version rollback attacks was not effective.
6693 In s23_clnt.c, don't use special rollback-attack detection padding
6694 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6695 client; similarly, in s23_srvr.c, don't do the rollback check if
6696 SSL 2.0 is the only protocol enabled in the server.
6699 *) Make it possible to get hexdumps of unprintable data with 'openssl
6700 asn1parse'. By implication, the functions ASN1_parse_dump() and
6701 BIO_dump_indent() are added.
6704 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6705 these print out strings and name structures based on various
6706 flags including RFC2253 support and proper handling of
6707 multibyte characters. Added options to the 'x509' utility
6708 to allow the various flags to be set.
6711 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6712 Also change the functions X509_cmp_current_time() and
6713 X509_gmtime_adj() work with an ASN1_TIME structure,
6714 this will enable certificates using GeneralizedTime in validity
6715 dates to be checked.
6718 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6719 negative public key encodings) on by default,
6720 NO_NEG_PUBKEY_BUG can be set to disable it.
6723 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6724 content octets. An i2c_ASN1_OBJECT is unnecessary because
6725 the encoding can be trivially obtained from the structure.
6728 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6729 not read locks (CRYPTO_r_[un]lock).
6732 *) A first attempt at creating official support for shared
6733 libraries through configuration. I've kept it so the
6734 default is static libraries only, and the OpenSSL programs
6735 are always statically linked for now, but there are
6736 preparations for dynamic linking in place.
6737 This has been tested on Linux and Tru64.
6740 *) Randomness polling function for Win9x, as described in:
6741 Peter Gutmann, Software Generation of Practically Strong
6745 *) Fix so PRNG is seeded in req if using an already existing
6749 *) New options to smime application. -inform and -outform
6750 allow alternative formats for the S/MIME message including
6751 PEM and DER. The -content option allows the content to be
6752 specified separately. This should allow things like Netscape
6753 form signing output easier to verify.
6756 *) Fix the ASN1 encoding of tags using the 'long form'.
6759 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6760 STRING types. These convert content octets to and from the
6761 underlying type. The actual tag and length octets are
6762 already assumed to have been read in and checked. These
6763 are needed because all other string types have virtually
6764 identical handling apart from the tag. By having versions
6765 of the ASN1 functions that just operate on content octets
6766 IMPLICIT tagging can be handled properly. It also allows
6767 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6768 and ASN1_INTEGER are identical apart from the tag.
6771 *) Change the handling of OID objects as follows:
6773 - New object identifiers are inserted in objects.txt, following
6774 the syntax given in objects.README.
6775 - objects.pl is used to process obj_mac.num and create a new
6777 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6780 This is currently kind of a hack, and the perl code in objects.pl
6781 isn't very elegant, but it works as I intended. The simplest way
6782 to check that it worked correctly is to look in obj_dat.h and
6783 check the array nid_objs and make sure the objects haven't moved
6784 around (this is important!). Additions are OK, as well as
6785 consistent name changes.
6788 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6791 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6792 The given file adds to whatever has already been seeded into the
6793 random pool through the RANDFILE configuration file option or
6794 environment variable, or the default random state file.
6797 *) mkstack.pl now sorts each macro group into lexical order.
6798 Previously the output order depended on the order the files
6799 appeared in the directory, resulting in needless rewriting
6803 *) Patches to make OpenSSL compile under Win32 again. Mostly
6804 work arounds for the VC++ problem that it treats func() as
6805 func(void). Also stripped out the parts of mkdef.pl that
6806 added extra typesafe functions: these no longer exist.
6809 *) Reorganisation of the stack code. The macros are now all
6810 collected in safestack.h . Each macro is defined in terms of
6811 a "stack macro" of the form SKM_<name>(type, a, b). The
6812 DEBUG_SAFESTACK is now handled in terms of function casts,
6813 this has the advantage of retaining type safety without the
6814 use of additional functions. If DEBUG_SAFESTACK is not defined
6815 then the non typesafe macros are used instead. Also modified the
6816 mkstack.pl script to handle the new form. Needs testing to see
6817 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6818 the default if no major problems. Similar behaviour for ASN1_SET_OF
6819 and PKCS12_STACK_OF.
6822 *) When some versions of IIS use the 'NET' form of private key the
6823 key derivation algorithm is different. Normally MD5(password) is
6824 used as a 128 bit RC4 key. In the modified case
6825 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6826 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6827 as the old Netscape_RSA functions except they have an additional
6828 'sgckey' parameter which uses the modified algorithm. Also added
6829 an -sgckey command line option to the rsa utility. Thanks to
6830 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6831 algorithm to openssl-dev.
6834 *) The evp_local.h macros were using 'c.##kname' which resulted in
6835 invalid expansion on some systems (SCO 5.0.5 for example).
6836 Corrected to 'c.kname'.
6837 [Phillip Porch <root@theporch.com>]
6839 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6840 a STACK of email addresses from a certificate or request, these look
6841 in the subject name and the subject alternative name extensions and
6842 omit any duplicate addresses.
6845 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6846 This makes DSA verification about 2 % faster.
6849 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6850 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6851 plus overhead for 1024 bit moduli).
6852 This makes exponentiations about 0.5 % faster for 1024 bit
6853 exponents (as measured by "openssl speed rsa2048").
6856 *) Rename memory handling macros to avoid conflicts with other
6858 Malloc => OPENSSL_malloc
6859 Malloc_locked => OPENSSL_malloc_locked
6860 Realloc => OPENSSL_realloc
6861 Free => OPENSSL_free
6864 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6865 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6868 *) CygWin32 support.
6869 [John Jarvie <jjarvie@newsguy.com>]
6871 *) The type-safe stack code has been rejigged. It is now only compiled
6872 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6873 by default all type-specific stack functions are "#define"d back to
6874 standard stack functions. This results in more streamlined output
6875 but retains the type-safety checking possibilities of the original
6879 *) The STACK code has been cleaned up, and certain type declarations
6880 that didn't make a lot of sense have been brought in line. This has
6881 also involved a cleanup of sorts in safestack.h to more correctly
6882 map type-safe stack functions onto their plain stack counterparts.
6883 This work has also resulted in a variety of "const"ifications of
6884 lots of the code, especially "_cmp" operations which should normally
6885 be prototyped with "const" parameters anyway.
6888 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6889 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6890 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6891 where all of 'md' is used each time the PRNG is used, but 'state'
6892 is used only indexed by a cyclic counter. As entropy may not be
6893 well distributed from the beginning, 'md' is important as a
6894 chaining variable. However, the output function chains only half
6895 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6896 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6897 in all of 'state' being rewritten, with the new values depending
6898 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6901 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6902 the handshake is continued after ssl_verify_cert_chain();
6903 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6904 can lead to 'unexplainable' connection aborts later.
6905 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6907 *) Major EVP API cipher revision.
6908 Add hooks for extra EVP features. This allows various cipher
6909 parameters to be set in the EVP interface. Support added for variable
6910 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6911 setting of RC2 and RC5 parameters.
6913 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6916 Remove lots of duplicated code from the EVP library. For example *every*
6917 cipher init() function handles the 'iv' in the same way according to the
6918 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6919 for CFB and OFB modes they zero ctx->num.
6921 New functionality allows removal of S/MIME code RC2 hack.
6923 Most of the routines have the same form and so can be declared in terms
6926 By shifting this to the top level EVP_CipherInit() it can be removed from
6927 all individual ciphers. If the cipher wants to handle IVs or keys
6928 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6931 Change lots of functions like EVP_EncryptUpdate() to now return a
6932 value: although software versions of the algorithms cannot fail
6933 any installed hardware versions can.
6936 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6937 this option is set, tolerate broken clients that send the negotiated
6938 protocol version number instead of the requested protocol version
6942 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6943 i.e. non-zero for export ciphersuites, zero otherwise.
6944 Previous versions had this flag inverted, inconsistent with
6945 rsa_tmp_cb (..._TMP_RSA_CB).
6946 [Bodo Moeller; problem reported by Amit Chopra]
6948 *) Add missing DSA library text string. Work around for some IIS
6949 key files with invalid SEQUENCE encoding.
6952 *) Add a document (doc/standards.txt) that list all kinds of standards
6953 and so on that are implemented in OpenSSL.
6956 *) Enhance c_rehash script. Old version would mishandle certificates
6957 with the same subject name hash and wouldn't handle CRLs at all.
6958 Added -fingerprint option to crl utility, to support new c_rehash
6962 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6965 *) Fix for SSL server purpose checking. Server checking was
6966 rejecting certificates which had extended key usage present
6967 but no ssl client purpose.
6968 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6970 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6971 is a little unclear about how a blank password is handled.
6972 Since the password in encoded as a BMPString with terminating
6973 double NULL a zero length password would end up as just the
6974 double NULL. However no password at all is different and is
6975 handled differently in the PKCS#12 key generation code. NS
6976 treats a blank password as zero length. MSIE treats it as no
6977 password on export: but it will try both on import. We now do
6978 the same: PKCS12_parse() tries zero length and no password if
6979 the password is set to "" or NULL (NULL is now a valid password:
6980 it wasn't before) as does the pkcs12 application.
6983 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6984 perror when PEM_read_bio_X509_REQ fails, the error message must
6985 be obtained from the error queue.
6988 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6989 it in ERR_remove_state if appropriate, and change ERR_get_state
6990 accordingly to avoid race conditions (this is necessary because
6991 thread_hash is no longer constant once set).
6994 *) Bugfix for linux-elf makefile.one.
6997 *) RSA_get_default_method() will now cause a default
6998 RSA_METHOD to be chosen if one doesn't exist already.
6999 Previously this was only set during a call to RSA_new()
7000 or RSA_new_method(NULL) meaning it was possible for
7001 RSA_get_default_method() to return NULL.
7004 *) Added native name translation to the existing DSO code
7005 that will convert (if the flag to do so is set) filenames
7006 that are sufficiently small and have no path information
7007 into a canonical native form. Eg. "blah" converted to
7008 "libblah.so" or "blah.dll" etc.
7011 *) New function ERR_error_string_n(e, buf, len) which is like
7012 ERR_error_string(e, buf), but writes at most 'len' bytes
7013 including the 0 terminator. For ERR_error_string_n, 'buf'
7015 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7017 *) CONF library reworked to become more general. A new CONF
7018 configuration file reader "class" is implemented as well as a
7019 new functions (NCONF_*, for "New CONF") to handle it. The now
7020 old CONF_* functions are still there, but are reimplemented to
7021 work in terms of the new functions. Also, a set of functions
7022 to handle the internal storage of the configuration data is
7023 provided to make it easier to write new configuration file
7024 reader "classes" (I can definitely see something reading a
7025 configuration file in XML format, for example), called _CONF_*,
7026 or "the configuration storage API"...
7028 The new configuration file reading functions are:
7030 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7031 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7033 NCONF_default, NCONF_WIN32
7035 NCONF_dump_fp, NCONF_dump_bio
7037 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7038 NCONF_new creates a new CONF object. This works in the same way
7039 as other interfaces in OpenSSL, like the BIO interface.
7040 NCONF_dump_* dump the internal storage of the configuration file,
7041 which is useful for debugging. All other functions take the same
7042 arguments as the old CONF_* functions wth the exception of the
7043 first that must be a `CONF *' instead of a `LHASH *'.
7045 To make it easer to use the new classes with the old CONF_* functions,
7046 the function CONF_set_default_method is provided.
7049 *) Add '-tls1' option to 'openssl ciphers', which was already
7050 mentioned in the documentation but had not been implemented.
7051 (This option is not yet really useful because even the additional
7052 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7055 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7056 OpenSSL-based applications) load shared libraries and bind to
7057 them in a portable way.
7058 [Geoff Thorpe, with contributions from Richard Levitte]
7060 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7062 *) Make sure _lrotl and _lrotr are only used with MSVC.
7064 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7065 (the default implementation of RAND_status).
7067 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7068 to '-clrext' (= clear extensions), as intended and documented.
7069 [Bodo Moeller; inconsistency pointed out by Michael Attili
7072 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7073 was larger than the MD block size.
7074 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7076 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7077 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7078 using the passed key: if the passed key was a private key the result
7079 of X509_print(), for example, would be to print out all the private key
7083 *) des_quad_cksum() byte order bug fix.
7084 [Ulf Möller, using the problem description in krb4-0.9.7, where
7085 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7087 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7089 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7091 *) For easily testing in shell scripts whether some command
7092 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7093 returns with exit code 0 iff no command of the given name is available.
7094 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7095 the output goes to stdout and nothing is printed to stderr.
7096 Additional arguments are always ignored.
7098 Since for each cipher there is a command of the same name,
7099 the 'no-cipher' compilation switches can be tested this way.
7101 ('openssl no-XXX' is not able to detect pseudo-commands such
7102 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7105 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7108 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7109 is set; it will be thrown away anyway because each handshake creates
7111 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7112 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7113 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7114 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7117 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7118 'Q' and 'R' lose their special meanings (quit/renegotiate).
7119 This is part of what -quiet does; unlike -quiet, -ign_eof
7120 does not suppress any output.
7123 *) Add compatibility options to the purpose and trust code. The
7124 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7125 accepts a certificate or CA, this was the previous behaviour,
7126 with all the associated security issues.
7128 X509_TRUST_COMPAT is the old trust behaviour: only and
7129 automatically trust self signed roots in certificate store. A
7130 new trust setting X509_TRUST_DEFAULT is used to specify that
7131 a purpose has no associated trust setting and it should instead
7132 use the value in the default purpose.
7135 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7136 and fix a memory leak.
7139 *) In util/mkerr.pl (which implements 'make errors'), preserve
7140 reason strings from the previous version of the .c file, as
7141 the default to have only downcase letters (and digits) in
7142 automatically generated reasons codes is not always appropriate.
7145 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7146 using strerror. Previously, ERR_reason_error_string() returned
7147 library names as reason strings for SYSerr; but SYSerr is a special
7148 case where small numbers are errno values, not library numbers.
7151 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7152 converts DSA parameters into DH parameters. (When creating parameters,
7153 DSA_generate_parameters is used.)
7156 *) Include 'length' (recommended exponent length) in C code generated
7157 by 'openssl dhparam -C'.
7160 *) The second argument to set_label in perlasm was already being used
7161 so couldn't be used as a "file scope" flag. Moved to third argument
7165 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7166 instead of RAND_bytes for encryption IVs and salts.
7169 *) Include RAND_status() into RAND_METHOD instead of implementing
7170 it only for md_rand.c Otherwise replacing the PRNG by calling
7171 RAND_set_rand_method would be impossible.
7174 *) Don't let DSA_generate_key() enter an infinite loop if the random
7175 number generation fails.
7178 *) New 'rand' application for creating pseudo-random output.
7181 *) Added configuration support for Linux/IA64
7182 [Rolf Haberrecker <rolf@suse.de>]
7184 *) Assembler module support for Mingw32.
7187 *) Shared library support for HPUX (in shlib/).
7188 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7190 *) Shared library support for Solaris gcc.
7191 [Lutz Behnke <behnke@trustcenter.de>]
7193 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7195 *) PKCS7_encrypt() was adding text MIME headers twice because they
7196 were added manually and by SMIME_crlf_copy().
7199 *) In bntest.c don't call BN_rand with zero bits argument.
7200 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7202 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7203 case was implemented. This caused BN_div_recp() to fail occasionally.
7206 *) Add an optional second argument to the set_label() in the perl
7207 assembly language builder. If this argument exists and is set
7208 to 1 it signals that the assembler should use a symbol whose
7209 scope is the entire file, not just the current function. This
7210 is needed with MASM which uses the format label:: for this scope.
7211 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7213 *) Change the ASN1 types so they are typedefs by default. Before
7214 almost all types were #define'd to ASN1_STRING which was causing
7215 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7219 *) Change names of new functions to the new get1/get0 naming
7220 convention: After 'get1', the caller owns a reference count
7221 and has to call ..._free; 'get0' returns a pointer to some
7222 data structure without incrementing reference counters.
7223 (Some of the existing 'get' functions increment a reference
7224 counter, some don't.)
7225 Similarly, 'set1' and 'add1' functions increase reference
7226 counters or duplicate objects.
7229 *) Allow for the possibility of temp RSA key generation failure:
7230 the code used to assume it always worked and crashed on failure.
7233 *) Fix potential buffer overrun problem in BIO_printf().
7234 [Ulf Möller, using public domain code by Patrick Powell; problem
7235 pointed out by David Sacerdote <das33@cornell.edu>]
7237 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7238 RAND_egd() and RAND_status(). In the command line application,
7239 the EGD socket can be specified like a seed file using RANDFILE
7243 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7244 Some CAs (e.g. Verisign) distribute certificates in this form.
7247 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7248 list to exclude them. This means that no special compilation option
7249 is needed to use anonymous DH: it just needs to be included in the
7253 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7254 EVP_MD_type. The old functionality is available in a new macro called
7255 EVP_MD_md(). Change code that uses it and update docs.
7258 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7259 where the 'void *' argument is replaced by a function pointer argument.
7260 Previously 'void *' was abused to point to functions, which works on
7261 many platforms, but is not correct. As these functions are usually
7262 called by macros defined in OpenSSL header files, most source code
7263 should work without changes.
7266 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7267 sections with information on -D... compiler switches used for
7268 compiling the library so that applications can see them. To enable
7269 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7270 must be defined. E.g.,
7271 #define OPENSSL_ALGORITHM_DEFINES
7272 #include <openssl/opensslconf.h>
7273 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7274 [Richard Levitte, Ulf and Bodo Möller]
7276 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7280 *) Change the 'other' type in certificate aux info to a STACK_OF
7281 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7282 the required ASN1 format: arbitrary types determined by an OID.
7285 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7286 argument to 'req'. This is not because the function is newer or
7287 better than others it just uses the work 'NEW' in the certificate
7288 request header lines. Some software needs this.
7291 *) Reorganise password command line arguments: now passwords can be
7292 obtained from various sources. Delete the PEM_cb function and make
7293 it the default behaviour: i.e. if the callback is NULL and the
7294 usrdata argument is not NULL interpret it as a null terminated pass
7295 phrase. If usrdata and the callback are NULL then the pass phrase
7296 is prompted for as usual.
7299 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7300 the support is automatically enabled. The resulting binaries will
7301 autodetect the card and use it if present.
7302 [Ben Laurie and Compaq Inc.]
7304 *) Work around for Netscape hang bug. This sends certificate request
7305 and server done in one record. Since this is perfectly legal in the
7306 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7307 the bugs/SSLv3 entry for more info.
7310 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7313 *) Add -rand argument to smime and pkcs12 applications and read/write
7317 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7320 *) Add command line password options to the remaining applications.
7323 *) Bug fix for BN_div_recp() for numerators with an even number of
7327 *) More tests in bntest.c, and changed test_bn output.
7330 *) ./config recognizes MacOS X now.
7333 *) Bug fix for BN_div() when the first words of num and divsor are
7334 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7337 *) Add support for various broken PKCS#8 formats, and command line
7338 options to produce them.
7341 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7342 get temporary BIGNUMs from a BN_CTX.
7345 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7349 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7350 include a #define from the old name to the new. The original intent
7351 was that statically linked binaries could for example just call
7352 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7353 link with digests. This never worked becayse SSLeay_add_all_digests()
7354 and SSLeay_add_all_ciphers() were in the same source file so calling
7355 one would link with the other. They are now in separate source files.
7358 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7361 *) Use a less unusual form of the Miller-Rabin primality test (it used
7362 a binary algorithm for exponentiation integrated into the Miller-Rabin
7363 loop, our standard modexp algorithms are faster).
7366 *) Support for the EBCDIC character set completed.
7367 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7369 *) Source code cleanups: use const where appropriate, eliminate casts,
7370 use void * instead of char * in lhash.
7373 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7374 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7375 this the server could overwrite ephemeral keys that the client
7379 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7380 using 50 iterations of the Rabin-Miller test.
7382 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7383 iterations of the Rabin-Miller test as required by the appendix
7384 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7385 As BN_is_prime_fasttest includes trial division, DSA parameter
7386 generation becomes much faster.
7388 This implies a change for the callback functions in DSA_is_prime
7389 and DSA_generate_parameters: The callback function is called once
7390 for each positive witness in the Rabin-Miller test, not just
7391 occasionally in the inner loop; and the parameters to the
7392 callback function now provide an iteration count for the outer
7393 loop rather than for the current invocation of the inner loop.
7394 DSA_generate_parameters additionally can call the callback
7395 function with an 'iteration count' of -1, meaning that a
7396 candidate has passed the trial division test (when q is generated
7397 from an application-provided seed, trial division is skipped).
7400 *) New function BN_is_prime_fasttest that optionally does trial
7401 division before starting the Rabin-Miller test and has
7402 an additional BN_CTX * argument (whereas BN_is_prime always
7403 has to allocate at least one BN_CTX).
7404 'callback(1, -1, cb_arg)' is called when a number has passed the
7405 trial division stage.
7408 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7412 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7415 *) New function BN_pseudo_rand().
7418 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7419 bignum version of BN_from_montgomery() with the working code from
7420 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7424 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7425 made it impossible to use the same SSL_SESSION data structure in
7426 SSL2 clients in multiple threads.
7429 *) The return value of RAND_load_file() no longer counts bytes obtained
7430 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7431 to seed the PRNG (previously an explicit byte count was required).
7432 [Ulf Möller, Bodo Möller]
7434 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7435 used (char *) instead of (void *) and had casts all over the place.
7438 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7441 *) Retain source code compatibility for BN_prime_checks macro:
7442 BN_is_prime(..., BN_prime_checks, ...) now uses
7443 BN_prime_checks_for_size to determine the appropriate number of
7444 Rabin-Miller iterations.
7447 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7448 DH_CHECK_P_NOT_SAFE_PRIME.
7449 (Check if this is true? OpenPGP calls them "strong".)
7452 *) Merge the functionality of "dh" and "gendh" programs into a new program
7453 "dhparam". The old programs are retained for now but will handle DH keys
7454 (instead of parameters) in future.
7457 *) Make the ciphers, s_server and s_client programs check the return values
7458 when a new cipher list is set.
7461 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7462 ciphers. Before when the 56bit ciphers were enabled the sorting was
7465 The syntax for the cipher sorting has been extended to support sorting by
7466 cipher-strength (using the strength_bits hard coded in the tables).
7467 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7469 Fix a bug in the cipher-command parser: when supplying a cipher command
7470 string with an "undefined" symbol (neither command nor alphanumeric
7471 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7472 an error is flagged.
7474 Due to the strength-sorting extension, the code of the
7475 ssl_create_cipher_list() function was completely rearranged. I hope that
7476 the readability was also increased :-)
7477 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7479 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7480 for the first serial number and places 2 in the serial number file. This
7481 avoids problems when the root CA is created with serial number zero and
7482 the first user certificate has the same issuer name and serial number
7486 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7487 the new code. Add documentation for this stuff.
7490 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7491 X509_*() to X509at_*() on the grounds that they don't handle X509
7492 structures and behave in an analagous way to the X509v3 functions:
7493 they shouldn't be called directly but wrapper functions should be used
7496 So we also now have some wrapper functions that call the X509at functions
7497 when passed certificate requests. (TO DO: similar things can be done with
7498 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7499 things. Some of these need some d2i or i2d and print functionality
7500 because they handle more complex structures.)
7503 *) Add missing #ifndefs that caused missing symbols when building libssl
7504 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7505 NO_RSA in ssl/s2*.c.
7506 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7508 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7509 has a return value which indicates the quality of the random data
7510 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7511 error queue. New function RAND_pseudo_bytes() generates output that is
7512 guaranteed to be unique but not unpredictable. RAND_add is like
7513 RAND_seed, but takes an extra argument for an entropy estimate
7514 (RAND_seed always assumes full entropy).
7517 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7518 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7519 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7520 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7521 false-positive rate of at most 2^-80 for random input.
7524 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7527 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7528 in the 0.9.5 release), this returns the chain
7529 from an X509_CTX structure with a dup of the stack and all
7530 the X509 reference counts upped: so the stack will exist
7531 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7534 Also make SSL_SESSION_print() print out the verify return
7538 *) Add manpage for the pkcs12 command. Also change the default
7539 behaviour so MAC iteration counts are used unless the new
7540 -nomaciter option is used. This improves file security and
7541 only older versions of MSIE (4.0 for example) need it.
7544 *) Honor the no-xxx Configure options when creating .DEF files.
7547 *) Add PKCS#10 attributes to field table: challengePassword,
7548 unstructuredName and unstructuredAddress. These are taken from
7549 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7550 international characters are used.
7552 More changes to X509_ATTRIBUTE code: allow the setting of types
7553 based on strings. Remove the 'loc' parameter when adding
7554 attributes because these will be a SET OF encoding which is sorted
7558 *) Initial changes to the 'req' utility to allow request generation
7559 automation. This will allow an application to just generate a template
7560 file containing all the field values and have req construct the
7563 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7564 used all over the place including certificate requests and PKCS#7
7565 structures. They are currently handled manually where necessary with
7566 some primitive wrappers for PKCS#7. The new functions behave in a
7567 manner analogous to the X509 extension functions: they allow
7568 attributes to be looked up by NID and added.
7570 Later something similar to the X509V3 code would be desirable to
7571 automatically handle the encoding, decoding and printing of the
7572 more complex types. The string types like challengePassword can
7573 be handled by the string table functions.
7575 Also modified the multi byte string table handling. Now there is
7576 a 'global mask' which masks out certain types. The table itself
7577 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7578 is useful when for example there is only one permissible type
7579 (as in countryName) and using the mask might result in no valid
7583 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7584 SSL_get_peer_finished to allow applications to obtain the latest
7585 Finished messages sent to the peer or expected from the peer,
7586 respectively. (SSL_get_peer_finished is usually the Finished message
7587 actually received from the peer, otherwise the protocol will be aborted.)
7589 As the Finished message are message digests of the complete handshake
7590 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7591 be used for external authentication procedures when the authentication
7592 provided by SSL/TLS is not desired or is not enough.
7595 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7596 the host supports BWX extension and if Compaq C is present on the
7597 $PATH. Just exploiting of the BWX extension results in 20-30%
7598 performance kick for some algorithms, e.g. DES and RC4 to mention
7599 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7603 *) Add support for MS "fast SGC". This is arguably a violation of the
7604 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7605 weak crypto and after checking the certificate is SGC a second one
7606 with strong crypto. MS SGC stops the first handshake after receiving
7607 the server certificate message and sends a second client hello. Since
7608 a server will typically do all the time consuming operations before
7609 expecting any further messages from the client (server key exchange
7610 is the most expensive) there is little difference between the two.
7612 To get OpenSSL to support MS SGC we have to permit a second client
7613 hello message after we have sent server done. In addition we have to
7614 reset the MAC if we do get this second client hello.
7617 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7618 if a DER encoded private key is RSA or DSA traditional format. Changed
7619 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7620 format DER encoded private key. Newer code should use PKCS#8 format which
7621 has the key type encoded in the ASN1 structure. Added DER private key
7622 support to pkcs8 application.
7625 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7626 ciphersuites has been selected (as required by the SSL 3/TLS 1
7627 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7628 is set, we interpret this as a request to violate the specification
7629 (the worst that can happen is a handshake failure, and 'correct'
7630 behaviour would result in a handshake failure anyway).
7633 *) In SSL_CTX_add_session, take into account that there might be multiple
7634 SSL_SESSION structures with the same session ID (e.g. when two threads
7635 concurrently obtain them from an external cache).
7636 The internal cache can handle only one SSL_SESSION with a given ID,
7637 so if there's a conflict, we now throw out the old one to achieve
7641 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7642 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7643 some routines that use cipher OIDs: some ciphers do not have OIDs
7644 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7648 *) Simplify the trust setting structure and code. Now we just have
7649 two sequences of OIDs for trusted and rejected settings. These will
7650 typically have values the same as the extended key usage extension
7651 and any application specific purposes.
7653 The trust checking code now has a default behaviour: it will just
7654 check for an object with the same NID as the passed id. Functions can
7655 be provided to override either the default behaviour or the behaviour
7656 for a given id. SSL client, server and email already have functions
7657 in place for compatibility: they check the NID and also return "trusted"
7658 if the certificate is self signed.
7661 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7662 traditional format into an EVP_PKEY structure.
7665 *) Add a password callback function PEM_cb() which either prompts for
7666 a password if usr_data is NULL or otherwise assumes it is a null
7667 terminated password. Allow passwords to be passed on command line
7668 environment or config files in a few more utilities.
7671 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7672 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7673 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7674 Update documentation.
7677 *) Support for ASN1 "NULL" type. This could be handled before by using
7678 ASN1_TYPE but there wasn't any function that would try to read a NULL
7679 and produce an error if it couldn't. For compatibility we also have
7680 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7681 don't allocate anything because they don't need to.
7684 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7686 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7688 *) Rebuild of the memory allocation routines used by OpenSSL code and
7689 possibly others as well. The purpose is to make an interface that
7690 provide hooks so anyone can build a separate set of allocation and
7691 deallocation routines to be used by OpenSSL, for example memory
7692 pool implementations, or something else, which was previously hard
7693 since Malloc(), Realloc() and Free() were defined as macros having
7694 the values malloc, realloc and free, respectively (except for Win32
7695 compilations). The same is provided for memory debugging code.
7696 OpenSSL already comes with functionality to find memory leaks, but
7697 this gives people a chance to debug other memory problems.
7699 With these changes, a new set of functions and macros have appeared:
7701 CRYPTO_set_mem_debug_functions() [F]
7702 CRYPTO_get_mem_debug_functions() [F]
7703 CRYPTO_dbg_set_options() [F]
7704 CRYPTO_dbg_get_options() [F]
7705 CRYPTO_malloc_debug_init() [M]
7707 The memory debug functions are NULL by default, unless the library
7708 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7709 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7710 gives the standard debugging functions that come with OpenSSL) or
7711 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7712 provided by the library user) must be used. When the standard
7713 debugging functions are used, CRYPTO_dbg_set_options can be used to
7714 request additional information:
7715 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7716 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7718 Also, things like CRYPTO_set_mem_functions will always give the
7719 expected result (the new set of functions is used for allocation
7720 and deallocation) at all times, regardless of platform and compiler
7723 To finish it up, some functions that were never use in any other
7724 way than through macros have a new API and new semantic:
7727 CRYPTO_dbg_realloc()
7730 All macros of value have retained their old syntax.
7731 [Richard Levitte and Bodo Moeller]
7733 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7734 ordering of SMIMECapabilities wasn't in "strength order" and there
7735 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7739 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7740 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7741 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7743 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7744 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7745 functionality to handle multipart/signed properly) and a utility
7746 called 'smime' to call all this stuff. This is based on code I
7747 originally wrote for Celo who have kindly allowed it to be
7748 included in OpenSSL.
7751 *) Add variants des_set_key_checked and des_set_key_unchecked of
7752 des_set_key (aka des_key_sched). Global variable des_check_key
7753 decides which of these is called by des_set_key; this way
7754 des_check_key behaves as it always did, but applications and
7755 the library itself, which was buggy for des_check_key == 1,
7756 have a cleaner way to pick the version they need.
7759 *) New function PKCS12_newpass() which changes the password of a
7763 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7764 dynamic mix. In both cases the ids can be used as an index into the
7765 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7766 functions so they accept a list of the field values and the
7767 application doesn't need to directly manipulate the X509_TRUST
7771 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7775 *) Modify the way the V3 extension code looks up extensions. This now
7776 works in a similar way to the object code: we have some "standard"
7777 extensions in a static table which is searched with OBJ_bsearch()
7778 and the application can add dynamic ones if needed. The file
7779 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7780 updated whenever a new extension is added to the core code and kept
7781 in ext_nid order. There is a simple program 'tabtest.c' which checks
7782 this. New extensions are not added too often so this file can readily
7783 be maintained manually.
7785 There are two big advantages in doing things this way. The extensions
7786 can be looked up immediately and no longer need to be "added" using
7787 X509V3_add_standard_extensions(): this function now does nothing.
7788 [Side note: I get *lots* of email saying the extension code doesn't
7789 work because people forget to call this function]
7790 Also no dynamic allocation is done unless new extensions are added:
7791 so if we don't add custom extensions there is no need to call
7792 X509V3_EXT_cleanup().
7795 *) Modify enc utility's salting as follows: make salting the default. Add a
7796 magic header, so unsalted files fail gracefully instead of just decrypting
7797 to garbage. This is because not salting is a big security hole, so people
7798 should be discouraged from doing it.
7801 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7802 digest to be passed on the command line but it only used this
7803 parameter when signing a certificate. Modified so all relevant
7804 operations are affected by the digest parameter including the
7805 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7806 DSA key was used because it didn't fix the digest.
7809 *) Initial certificate chain verify code. Currently tests the untrusted
7810 certificates for consistency with the verify purpose (which is set
7811 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7813 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7814 this is because it will reject chains with invalid extensions whereas
7815 every previous version of OpenSSL and SSLeay made no checks at all.
7817 Trust code: checks the root CA for the relevant trust settings. Trust
7818 settings have an initial value consistent with the verify purpose: e.g.
7819 if the verify purpose is for SSL client use it expects the CA to be
7820 trusted for SSL client use. However the default value can be changed to
7821 permit custom trust settings: one example of this would be to only trust
7822 certificates from a specific "secure" set of CAs.
7824 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7825 which should be used for version portability: especially since the
7826 verify structure is likely to change more often now.
7828 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7829 to set them. If not set then assume SSL clients will verify SSL servers
7832 Two new options to the verify program: -untrusted allows a set of
7833 untrusted certificates to be passed in and -purpose which sets the
7834 intended purpose of the certificate. If a purpose is set then the
7835 new chain verify code is used to check extension consistency.
7838 *) Support for the authority information access extension.
7841 *) Modify RSA and DSA PEM read routines to transparently handle
7842 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7843 public keys in a format compatible with certificate
7844 SubjectPublicKeyInfo structures. Unfortunately there were already
7845 functions called *_PublicKey_* which used various odd formats so
7846 these are retained for compatibility: however the DSA variants were
7847 never in a public release so they have been deleted. Changed dsa/rsa
7848 utilities to handle the new format: note no releases ever handled public
7849 keys so we should be OK.
7851 The primary motivation for this change is to avoid the same fiasco
7852 that dogs private keys: there are several incompatible private key
7853 formats some of which are standard and some OpenSSL specific and
7854 require various evil hacks to allow partial transparent handling and
7855 even then it doesn't work with DER formats. Given the option anything
7856 other than PKCS#8 should be dumped: but the other formats have to
7857 stay in the name of compatibility.
7859 With public keys and the benefit of hindsight one standard format
7860 is used which works with EVP_PKEY, RSA or DSA structures: though
7861 it clearly returns an error if you try to read the wrong kind of key.
7863 Added a -pubkey option to the 'x509' utility to output the public key.
7864 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7865 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7866 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7867 that do the same as the EVP_PKEY_assign_*() except they up the
7868 reference count of the added key (they don't "swallow" the
7872 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7873 CRLs would fail if the file contained no certificates or no CRLs:
7874 added a new function to read in both types and return the number
7875 read: this means that if none are read it will be an error. The
7876 DER versions of the certificate and CRL reader would always fail
7877 because it isn't possible to mix certificates and CRLs in DER format
7878 without choking one or the other routine. Changed this to just read
7879 a certificate: this is the best we can do. Also modified the code
7880 in apps/verify.c to take notice of return codes: it was previously
7881 attempting to read in certificates from NULL pointers and ignoring
7882 any errors: this is one reason why the cert and CRL reader seemed
7883 to work. It doesn't check return codes from the default certificate
7884 routines: these may well fail if the certificates aren't installed.
7887 *) Code to support otherName option in GeneralName.
7890 *) First update to verify code. Change the verify utility
7891 so it warns if it is passed a self signed certificate:
7892 for consistency with the normal behaviour. X509_verify
7893 has been modified to it will now verify a self signed
7894 certificate if *exactly* the same certificate appears
7895 in the store: it was previously impossible to trust a
7896 single self signed certificate. This means that:
7897 openssl verify ss.pem
7898 now gives a warning about a self signed certificate but
7899 openssl verify -CAfile ss.pem ss.pem
7903 *) For servers, store verify_result in SSL_SESSION data structure
7904 (and add it to external session representation).
7905 This is needed when client certificate verifications fails,
7906 but an application-provided verification callback (set by
7907 SSL_CTX_set_cert_verify_callback) allows accepting the session
7908 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7909 but returns 1): When the session is reused, we have to set
7910 ssl->verify_result to the appropriate error code to avoid
7912 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7914 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7915 case in PKCS7_dataInit() where the signed PKCS7 structure
7916 didn't contain any existing data because it was being created.
7917 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7919 *) Add a salt to the key derivation routines in enc.c. This
7920 forms the first 8 bytes of the encrypted file. Also add a
7921 -S option to allow a salt to be input on the command line.
7924 *) New function X509_cmp(). Oddly enough there wasn't a function
7925 to compare two certificates. We do this by working out the SHA1
7926 hash and comparing that. X509_cmp() will be needed by the trust
7930 *) SSL_get1_session() is like SSL_get_session(), but increments
7931 the reference count in the SSL_SESSION returned.
7932 [Geoff Thorpe <geoff@eu.c2.net>]
7934 *) Fix for 'req': it was adding a null to request attributes.
7935 Also change the X509_LOOKUP and X509_INFO code to handle
7936 certificate auxiliary information.
7939 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7943 *) Add the possibility to add extra information to the memory leak
7944 detecting output, to form tracebacks, showing from where each
7945 allocation was originated: CRYPTO_push_info("constant string") adds
7946 the string plus current file name and line number to a per-thread
7947 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7948 is like calling CYRPTO_pop_info() until the stack is empty.
7949 Also updated memory leak detection code to be multi-thread-safe.
7952 *) Add options -text and -noout to pkcs7 utility and delete the
7953 encryption options which never did anything. Update docs.
7956 *) Add options to some of the utilities to allow the pass phrase
7957 to be included on either the command line (not recommended on
7958 OSes like Unix) or read from the environment. Update the
7959 manpages and fix a few bugs.
7962 *) Add a few manpages for some of the openssl commands.
7965 *) Fix the -revoke option in ca. It was freeing up memory twice,
7966 leaking and not finding already revoked certificates.
7969 *) Extensive changes to support certificate auxiliary information.
7970 This involves the use of X509_CERT_AUX structure and X509_AUX
7971 functions. An X509_AUX function such as PEM_read_X509_AUX()
7972 can still read in a certificate file in the usual way but it
7973 will also read in any additional "auxiliary information". By
7974 doing things this way a fair degree of compatibility can be
7975 retained: existing certificates can have this information added
7976 using the new 'x509' options.
7978 Current auxiliary information includes an "alias" and some trust
7979 settings. The trust settings will ultimately be used in enhanced
7980 certificate chain verification routines: currently a certificate
7981 can only be trusted if it is self signed and then it is trusted
7985 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7986 The problem was that one of the replacement routines had not been working
7987 since SSLeay releases. For now the offending routine has been replaced
7988 with non-optimised assembler. Even so, this now gives around 95%
7989 performance improvement for 1024 bit RSA signs.
7992 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7993 handling. Most clients have the effective key size in bits equal to
7994 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
7995 A few however don't do this and instead use the size of the decrypted key
7996 to determine the RC2 key length and the AlgorithmIdentifier to determine
7997 the effective key length. In this case the effective key length can still
7998 be 40 bits but the key length can be 168 bits for example. This is fixed
7999 by manually forcing an RC2 key into the EVP_PKEY structure because the
8000 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8001 the key length and effective key length are equal.
8004 *) Add a bunch of functions that should simplify the creation of
8005 X509_NAME structures. Now you should be able to do:
8006 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8007 and have it automatically work out the correct field type and fill in
8008 the structures. The more adventurous can try:
8009 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8010 and it will (hopefully) work out the correct multibyte encoding.
8013 *) Change the 'req' utility to use the new field handling and multibyte
8014 copy routines. Before the DN field creation was handled in an ad hoc
8015 way in req, ca, and x509 which was rather broken and didn't support
8016 BMPStrings or UTF8Strings. Since some software doesn't implement
8017 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8018 using the dirstring_type option. See the new comment in the default
8019 openssl.cnf for more info.
8022 *) Make crypto/rand/md_rand.c more robust:
8023 - Assure unique random numbers after fork().
8024 - Make sure that concurrent threads access the global counter and
8025 md serializably so that we never lose entropy in them
8026 or use exactly the same state in multiple threads.
8027 Access to the large state is not always serializable because
8028 the additional locking could be a performance killer, and
8029 md should be large enough anyway.
8032 *) New file apps/app_rand.c with commonly needed functionality
8033 for handling the random seed file.
8035 Use the random seed file in some applications that previously did not:
8037 dsaparam -genkey (which also ignored its '-rand' option),
8040 x509 (when signing).
8041 Except on systems with /dev/urandom, it is crucial to have a random
8042 seed file at least for key creation, DSA signing, and for DH exchanges;
8043 for RSA signatures we could do without one.
8045 gendh and gendsa (unlike genrsa) used to read only the first byte
8046 of each file listed in the '-rand' option. The function as previously
8047 found in genrsa is now in app_rand.c and is used by all programs
8048 that support '-rand'.
8051 *) In RAND_write_file, use mode 0600 for creating files;
8052 don't just chmod when it may be too late.
8055 *) Report an error from X509_STORE_load_locations
8056 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8059 *) New function ASN1_mbstring_copy() this copies a string in either
8060 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8061 into an ASN1_STRING type. A mask of permissible types is passed
8062 and it chooses the "minimal" type to use or an error if not type
8066 *) Add function equivalents to the various macros in asn1.h. The old
8067 macros are retained with an M_ prefix. Code inside the library can
8068 use the M_ macros. External code (including the openssl utility)
8069 should *NOT* in order to be "shared library friendly".
8072 *) Add various functions that can check a certificate's extensions
8073 to see if it usable for various purposes such as SSL client,
8074 server or S/MIME and CAs of these types. This is currently
8075 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8076 verification. Also added a -purpose flag to x509 utility to
8077 print out all the purposes.
8080 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8084 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8085 for, obtain and decode and extension and obtain its critical flag.
8086 This allows all the necessary extension code to be handled in a
8087 single function call.
8090 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8091 platforms. See crypto/rc4/rc4_enc.c for further details.
8094 *) New -noout option to asn1parse. This causes no output to be produced
8095 its main use is when combined with -strparse and -out to extract data
8096 from a file (which may not be in ASN.1 format).
8099 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8100 when producing the local key id.
8101 [Richard Levitte <levitte@stacken.kth.se>]
8103 *) New option -dhparam in s_server. This allows a DH parameter file to be
8104 stated explicitly. If it is not stated then it tries the first server
8105 certificate file. The previous behaviour hard coded the filename
8109 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8110 a public key to be input or output. For example:
8111 openssl rsa -in key.pem -pubout -out pubkey.pem
8112 Also added necessary DSA public key functions to handle this.
8115 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8116 in the message. This was handled by allowing
8117 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8118 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8120 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8121 to the end of the strings whereas this didn't. This would cause problems
8122 if strings read with d2i_ASN1_bytes() were later modified.
8123 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8125 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8126 data and it contains EOF it will end up returning an error. This is
8127 caused by input 46 bytes long. The cause is due to the way base64
8128 BIOs find the start of base64 encoded data. They do this by trying a
8129 trial decode on each line until they find one that works. When they
8130 do a flag is set and it starts again knowing it can pass all the
8131 data directly through the decoder. Unfortunately it doesn't reset
8132 the context it uses. This means that if EOF is reached an attempt
8133 is made to pass two EOFs through the context and this causes the
8134 resulting error. This can also cause other problems as well. As is
8135 usual with these problems it takes *ages* to find and the fix is
8136 trivial: move one line.
8137 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8139 *) Ugly workaround to get s_client and s_server working under Windows. The
8140 old code wouldn't work because it needed to select() on sockets and the
8141 tty (for keypresses and to see if data could be written). Win32 only
8142 supports select() on sockets so we select() with a 1s timeout on the
8143 sockets and then see if any characters are waiting to be read, if none
8144 are present then we retry, we also assume we can always write data to
8145 the tty. This isn't nice because the code then blocks until we've
8146 received a complete line of data and it is effectively polling the
8147 keyboard at 1s intervals: however it's quite a bit better than not
8148 working at all :-) A dedicated Windows application might handle this
8149 with an event loop for example.
8152 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8153 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8154 will be called when RSA_sign() and RSA_verify() are used. This is useful
8155 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8156 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8157 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8158 This necessitated the support of an extra signature type NID_md5_sha1
8159 for SSL signatures and modifications to the SSL library to use it instead
8160 of calling RSA_public_decrypt() and RSA_private_encrypt().
8163 *) Add new -verify -CAfile and -CApath options to the crl program, these
8164 will lookup a CRL issuers certificate and verify the signature in a
8165 similar way to the verify program. Tidy up the crl program so it
8166 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8167 less strict. It will now permit CRL extensions even if it is not
8168 a V2 CRL: this will allow it to tolerate some broken CRLs.
8171 *) Initialize all non-automatic variables each time one of the openssl
8172 sub-programs is started (this is necessary as they may be started
8173 multiple times from the "OpenSSL>" prompt).
8174 [Lennart Bang, Bodo Moeller]
8176 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8177 removing all other RSA functionality (this is what NO_RSA does). This
8178 is so (for example) those in the US can disable those operations covered
8179 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8183 *) Non-copying interface to BIO pairs.
8184 (still largely untested)
8187 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8188 ASCII string. This was handled independently in various places before.
8191 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8192 UTF8 strings a character at a time.
8195 *) Use client_version from client hello to select the protocol
8196 (s23_srvr.c) and for RSA client key exchange verification
8197 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8200 *) Add various utility functions to handle SPKACs, these were previously
8201 handled by poking round in the structure internals. Added new function
8202 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8203 print, verify and generate SPKACs. Based on an original idea from
8204 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8207 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8210 *) Allow the config file extension section to be overwritten on the
8211 command line. Based on an original idea from Massimiliano Pala
8212 <madwolf@comune.modena.it>. The new option is called -extensions
8213 and can be applied to ca, req and x509. Also -reqexts to override
8214 the request extensions in req and -crlexts to override the crl extensions
8218 *) Add new feature to the SPKAC handling in ca. Now you can include
8219 the same field multiple times by preceding it by "XXXX." for example:
8222 this is the same syntax as used in the req config file.
8225 *) Allow certificate extensions to be added to certificate requests. These
8226 are specified in a 'req_extensions' option of the req section of the
8227 config file. They can be printed out with the -text option to req but
8228 are otherwise ignored at present.
8231 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8232 data read consists of only the final block it would not decrypted because
8233 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8234 A misplaced 'break' also meant the decrypted final block might not be
8235 copied until the next read.
8238 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8239 a few extra parameters to the DH structure: these will be useful if
8240 for example we want the value of 'q' or implement X9.42 DH.
8243 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8244 provides hooks that allow the default DSA functions or functions on a
8245 "per key" basis to be replaced. This allows hardware acceleration and
8246 hardware key storage to be handled without major modification to the
8247 library. Also added low level modexp hooks and CRYPTO_EX structure and
8248 associated functions.
8251 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8252 as "read only": it can't be written to and the buffer it points to will
8253 not be freed. Reading from a read only BIO is much more efficient than
8254 a normal memory BIO. This was added because there are several times when
8255 an area of memory needs to be read from a BIO. The previous method was
8256 to create a memory BIO and write the data to it, this results in two
8257 copies of the data and an O(n^2) reading algorithm. There is a new
8258 function BIO_new_mem_buf() which creates a read only memory BIO from
8259 an area of memory. Also modified the PKCS#7 routines to use read only
8263 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8264 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8265 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8266 but a retry condition occured while trying to read the rest.
8269 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8270 NID_pkcs7_encrypted by default: this was wrong since this should almost
8271 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8272 the encrypted data type: this is a more sensible place to put it and it
8273 allows the PKCS#12 code to be tidied up that duplicated this
8277 *) Changed obj_dat.pl script so it takes its input and output files on
8278 the command line. This should avoid shell escape redirection problems
8282 *) Initial support for certificate extension requests, these are included
8283 in things like Xenroll certificate requests. Included functions to allow
8284 extensions to be obtained and added.
8287 *) -crlf option to s_client and s_server for sending newlines as
8288 CRLF (as required by many protocols).
8291 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8293 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8294 [Ralf S. Engelschall]
8296 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8297 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8299 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8303 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8304 DH parameters/keys (q is lost during that conversion, but the resulting
8305 DH parameters contain its length).
8307 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8308 much faster than DH_generate_parameters (which creates parameters
8309 where p = 2*q + 1), and also the smaller q makes DH computations
8310 much more efficient (160-bit exponentiation instead of 1024-bit
8311 exponentiation); so this provides a convenient way to support DHE
8312 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8313 utter importance to use
8314 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8316 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8317 when such DH parameters are used, because otherwise small subgroup
8318 attacks may become possible!
8321 *) Avoid memory leak in i2d_DHparams.
8324 *) Allow the -k option to be used more than once in the enc program:
8325 this allows the same encrypted message to be read by multiple recipients.
8328 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8329 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8330 it will always use the numerical form of the OID, even if it has a short
8334 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8335 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8336 otherwise bn_mod_exp was called. In the case of hardware keys for example
8337 no private key components need be present and it might store extra data
8338 in the RSA structure, which cannot be accessed from bn_mod_exp.
8339 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8340 private key operations.
8343 *) Added support for SPARC Linux.
8346 *) pem_password_cb function type incompatibly changed from
8347 typedef int pem_password_cb(char *buf, int size, int rwflag);
8349 ....(char *buf, int size, int rwflag, void *userdata);
8350 so that applications can pass data to their callbacks:
8351 The PEM[_ASN1]_{read,write}... functions and macros now take an
8352 additional void * argument, which is just handed through whenever
8353 the password callback is called.
8354 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8356 New function SSL_CTX_set_default_passwd_cb_userdata.
8358 Compatibility note: As many C implementations push function arguments
8359 onto the stack in reverse order, the new library version is likely to
8360 interoperate with programs that have been compiled with the old
8361 pem_password_cb definition (PEM_whatever takes some data that
8362 happens to be on the stack as its last argument, and the callback
8363 just ignores this garbage); but there is no guarantee whatsoever that
8366 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8367 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8368 problems not only on Windows, but also on some Unix platforms.
8369 To avoid problematic command lines, these definitions are now in an
8370 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8371 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8374 *) MIPS III/IV assembler module is reimplemented.
8377 *) More DES library cleanups: remove references to srand/rand and
8378 delete an unused file.
8381 *) Add support for the the free Netwide assembler (NASM) under Win32,
8382 since not many people have MASM (ml) and it can be hard to obtain.
8383 This is currently experimental but it seems to work OK and pass all
8384 the tests. Check out INSTALL.W32 for info.
8387 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8388 without temporary keys kept an extra copy of the server key,
8389 and connections with temporary keys did not free everything in case
8393 *) New function RSA_check_key and new openssl rsa option -check
8394 for verifying the consistency of RSA keys.
8395 [Ulf Moeller, Bodo Moeller]
8397 *) Various changes to make Win32 compile work:
8398 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8399 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8400 comparison" warnings.
8401 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8404 *) Add a debugging option to PKCS#5 v2 key generation function: when
8405 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8406 derived keys are printed to stderr.
8409 *) Copy the flags in ASN1_STRING_dup().
8410 [Roman E. Pavlov <pre@mo.msk.ru>]
8412 *) The x509 application mishandled signing requests containing DSA
8413 keys when the signing key was also DSA and the parameters didn't match.
8415 It was supposed to omit the parameters when they matched the signing key:
8416 the verifying software was then supposed to automatically use the CA's
8417 parameters if they were absent from the end user certificate.
8419 Omitting parameters is no longer recommended. The test was also
8420 the wrong way round! This was probably due to unusual behaviour in
8421 EVP_cmp_parameters() which returns 1 if the parameters match.
8422 This meant that parameters were omitted when they *didn't* match and
8423 the certificate was useless. Certificates signed with 'ca' didn't have
8425 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8427 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8428 The interface is as follows:
8429 Applications can use
8430 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8431 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8432 "off" is now the default.
8433 The library internally uses
8434 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8435 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8436 to disable memory-checking temporarily.
8438 Some inconsistent states that previously were possible (and were
8439 even the default) are now avoided.
8441 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8442 with each memory chunk allocated; this is occasionally more helpful
8443 than just having a counter.
8445 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8447 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8451 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8452 which largely parallels "options", but is for changing API behaviour,
8453 whereas "options" are about protocol behaviour.
8454 Initial "mode" flags are:
8456 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8457 a single record has been written.
8458 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8459 retries use the same buffer location.
8460 (But all of the contents must be
8464 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8467 *) Fix problems with no-hmac etc.
8468 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8470 *) New functions RSA_get_default_method(), RSA_set_method() and
8471 RSA_get_method(). These allows replacement of RSA_METHODs without having
8472 to mess around with the internals of an RSA structure.
8475 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8476 Also really enable memory leak checks in openssl.c and in some
8478 [Chad C. Mulligan, Bodo Moeller]
8480 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8481 up the length of negative integers. This has now been simplified to just
8482 store the length when it is first determined and use it later, rather
8483 than trying to keep track of where data is copied and updating it to
8485 [Steve Henson, reported by Brien Wheeler
8486 <bwheeler@authentica-security.com>]
8488 *) Add a new function PKCS7_signatureVerify. This allows the verification
8489 of a PKCS#7 signature but with the signing certificate passed to the
8490 function itself. This contrasts with PKCS7_dataVerify which assumes the
8491 certificate is present in the PKCS#7 structure. This isn't always the
8492 case: certificates can be omitted from a PKCS#7 structure and be
8493 distributed by "out of band" means (such as a certificate database).
8496 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8497 function prototypes in pem.h, also change util/mkdef.pl to add the
8498 necessary function names.
8501 *) mk1mf.pl (used by Windows builds) did not properly read the
8502 options set by Configure in the top level Makefile, and Configure
8503 was not even able to write more than one option correctly.
8504 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8507 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8508 file to be loaded from a BIO or FILE pointer. The BIO version will
8509 for example allow memory BIOs to contain config info.
8512 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8513 Whoever hopes to achieve shared-library compatibility across versions
8514 must use this, not the compile-time macro.
8515 (Exercise 0.9.4: Which is the minimum library version required by
8517 Note: All this applies only to multi-threaded programs, others don't
8521 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8522 through a BIO pair triggered the default case, i.e.
8523 SSLerr(...,SSL_R_UNKNOWN_STATE).
8526 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8527 can use the SSL library even if none of the specific BIOs is
8531 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8532 for the encoded length.
8533 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8535 *) Add initial documentation of the X509V3 functions.
8538 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8539 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8540 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8541 secure PKCS#8 private key format with a high iteration count.
8544 *) Fix determination of Perl interpreter: A perl or perl5
8545 _directory_ in $PATH was also accepted as the interpreter.
8546 [Ralf S. Engelschall]
8548 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8549 wrong with it but it was very old and did things like calling
8550 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8554 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8555 to use the new extension code.
8558 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8559 with macros. This should make it easier to change their form, add extra
8560 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8564 *) Add to configuration table a new entry that can specify an alternative
8565 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8566 according to Mark Crispin <MRC@Panda.COM>.
8570 *) DES CBC did not update the IV. Weird.
8573 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8574 Changing the behaviour of the former might break existing programs --
8575 where IV updating is needed, des_ncbc_encrypt can be used.
8578 *) When bntest is run from "make test" it drives bc to check its
8579 calculations, as well as internally checking them. If an internal check
8580 fails, it needs to cause bc to give a non-zero result or make test carries
8581 on without noticing the failure. Fixed.
8584 *) DES library cleanups.
8587 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8588 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8589 ciphers. NOTE: although the key derivation function has been verified
8590 against some published test vectors it has not been extensively tested
8591 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8595 *) Instead of "mkdir -p", which is not fully portable, use new
8596 Perl script "util/mkdir-p.pl".
8599 *) Rewrite the way password based encryption (PBE) is handled. It used to
8600 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8601 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8602 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8603 the 'parameter' field of the AlgorithmIdentifier is passed to the
8604 underlying key generation function so it must do its own ASN1 parsing.
8605 This has also changed the EVP_PBE_CipherInit() function which now has a
8606 'parameter' argument instead of literal salt and iteration count values
8607 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8610 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8611 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8612 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8613 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8614 value was just used as a "magic string" and not used directly its
8615 value doesn't matter.
8618 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8622 *) "linux-sparc64" configuration (ultrapenguin).
8623 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8624 "linux-sparc" configuration.
8625 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8627 *) config now generates no-xxx options for missing ciphers.
8630 *) Support the EBCDIC character set (work in progress).
8631 File ebcdic.c not yet included because it has a different license.
8632 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8634 *) Support BS2000/OSD-POSIX.
8635 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8637 *) Make callbacks for key generation use void * instead of char *.
8640 *) Make S/MIME samples compile (not yet tested).
8643 *) Additional typesafe stacks.
8646 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8650 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8652 *) New configuration variant "sco5-gcc".
8654 *) Updated some demos.
8655 [Sean O Riordain, Wade Scholine]
8657 *) Add missing BIO_free at exit of pkcs12 application.
8660 *) Fix memory leak in conf.c.
8663 *) Updates for Win32 to assembler version of MD5.
8666 *) Set #! path to perl in apps/der_chop to where we found it
8667 instead of using a fixed path.
8670 *) SHA library changes for irix64-mips4-cc.
8673 *) Improvements for VMS support.
8677 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8679 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8680 This also avoids the problems with SC4.2 and unpatched SC5.
8681 [Andy Polyakov <appro@fy.chalmers.se>]
8683 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8684 These are required because of the typesafe stack would otherwise break
8685 existing code. If old code used a structure member which used to be STACK
8686 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8687 sk_num or sk_value it would produce an error because the num, data members
8688 are not present in STACK_OF. Now it just produces a warning. sk_set
8689 replaces the old method of assigning a value to sk_value
8690 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8691 that does this will no longer work (and should use sk_set instead) but
8692 this could be regarded as a "questionable" behaviour anyway.
8695 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8696 correctly handle encrypted S/MIME data.
8699 *) Change type of various DES function arguments from des_cblock
8700 (which means, in function argument declarations, pointer to char)
8701 to des_cblock * (meaning pointer to array with 8 char elements),
8702 which allows the compiler to do more typechecking; it was like
8703 that back in SSLeay, but with lots of ugly casts.
8705 Introduce new type const_des_cblock.
8708 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8709 problems: find RecipientInfo structure that matches recipient certificate
8710 and initialise the ASN1 structures properly based on passed cipher.
8713 *) Belatedly make the BN tests actually check the results.
8716 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8717 to and from BNs: it was completely broken. New compilation option
8718 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8719 key elements as negative integers.
8722 *) Reorganize and speed up MD5.
8723 [Andy Polyakov <appro@fy.chalmers.se>]
8726 [Richard Levitte <richard@levitte.org>]
8728 *) New option -out to asn1parse to allow the parsed structure to be
8729 output to a file. This is most useful when combined with the -strparse
8730 option to examine the output of things like OCTET STRINGS.
8733 *) Make SSL library a little more fool-proof by not requiring any longer
8734 that SSL_set_{accept,connect}_state be called before
8735 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8736 in many applications because usually everything *appeared* to work as
8737 intended anyway -- now it really works as intended).
8740 *) Move openssl.cnf out of lib/.
8743 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8744 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8745 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8746 [Ralf S. Engelschall]
8748 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8749 handle PKCS#7 enveloped data properly.
8750 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8752 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8753 copying pointers. The cert_st handling is changed by this in
8754 various ways (and thus what used to be known as ctx->default_cert
8755 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8756 any longer when s->cert does not give us what we need).
8757 ssl_cert_instantiate becomes obsolete by this change.
8758 As soon as we've got the new code right (possibly it already is?),
8759 we have solved a couple of bugs of the earlier code where s->cert
8760 was used as if it could not have been shared with other SSL structures.
8762 Note that using the SSL API in certain dirty ways now will result
8763 in different behaviour than observed with earlier library versions:
8764 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8765 does not influence s as it used to.
8767 In order to clean up things more thoroughly, inside SSL_SESSION
8768 we don't use CERT any longer, but a new structure SESS_CERT
8769 that holds per-session data (if available); currently, this is
8770 the peer's certificate chain and, for clients, the server's certificate
8771 and temporary key. CERT holds only those values that can have
8772 meaningful defaults in an SSL_CTX.
8775 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8776 from the internal representation. Various PKCS#7 fixes: remove some
8777 evil casts and set the enc_dig_alg field properly based on the signing
8781 *) Allow PKCS#12 password to be set from the command line or the
8782 environment. Let 'ca' get its config file name from the environment
8783 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8787 *) Allow certificate policies extension to use an IA5STRING for the
8788 organization field. This is contrary to the PKIX definition but
8789 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8793 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8794 without disallowing inline assembler and the like for non-pedantic builds.
8797 *) Support Borland C++ builder.
8798 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8803 *) SHA-1 cleanups and performance enhancements.
8804 [Andy Polyakov <appro@fy.chalmers.se>]
8806 *) Sparc v8plus assembler for the bignum library.
8807 [Andy Polyakov <appro@fy.chalmers.se>]
8809 *) Accept any -xxx and +xxx compiler options in Configure.
8812 *) Update HPUX configuration.
8815 *) Add missing sk_<type>_unshift() function to safestack.h
8816 [Ralf S. Engelschall]
8818 *) New function SSL_CTX_use_certificate_chain_file that sets the
8819 "extra_cert"s in addition to the certificate. (This makes sense
8820 only for "PEM" format files, as chains as a whole are not
8824 *) Support verify_depth from the SSL API.
8825 x509_vfy.c had what can be considered an off-by-one-error:
8826 Its depth (which was not part of the external interface)
8827 was actually counting the number of certificates in a chain;
8828 now it really counts the depth.
8831 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8832 instead of X509err, which often resulted in confusing error
8833 messages since the error codes are not globally unique
8834 (e.g. an alleged error in ssl3_accept when a certificate
8835 didn't match the private key).
8837 *) New function SSL_CTX_set_session_id_context that allows to set a default
8838 value (so that you don't need SSL_set_session_id_context for each
8839 connection using the SSL_CTX).
8842 *) OAEP decoding bug fix.
8845 *) Support INSTALL_PREFIX for package builders, as proposed by
8849 *) New Configure options "threads" and "no-threads". For systems
8850 where the proper compiler options are known (currently Solaris
8851 and Linux), "threads" is the default.
8854 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8857 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8858 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8859 such as /usr/local/bin.
8862 *) "make linux-shared" to build shared libraries.
8863 [Niels Poppe <niels@netbox.org>]
8865 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8868 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8869 extension adding in x509 utility.
8872 *) Remove NOPROTO sections and error code comments.
8875 *) Partial rewrite of the DEF file generator to now parse the ANSI
8879 *) New Configure options --prefix=DIR and --openssldir=DIR.
8882 *) Complete rewrite of the error code script(s). It is all now handled
8883 by one script at the top level which handles error code gathering,
8884 header rewriting and C source file generation. It should be much better
8885 than the old method: it now uses a modified version of Ulf's parser to
8886 read the ANSI prototypes in all header files (thus the old K&R definitions
8887 aren't needed for error creation any more) and do a better job of
8888 translating function codes into names. The old 'ASN1 error code imbedded
8889 in a comment' is no longer necessary and it doesn't use .err files which
8890 have now been deleted. Also the error code call doesn't have to appear all
8891 on one line (which resulted in some large lines...).
8894 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8897 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8898 0 (which usually indicates a closed connection), but continue reading.
8901 *) Fix some race conditions.
8904 *) Add support for CRL distribution points extension. Add Certificate
8905 Policies and CRL distribution points documentation.
8908 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8911 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8912 8 of keying material. Merlin has also confirmed interop with this fix
8913 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8914 [Merlin Hughes <merlin@baltimore.ie>]
8916 *) Fix lots of warnings.
8917 [Richard Levitte <levitte@stacken.kth.se>]
8919 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8920 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8921 [Richard Levitte <levitte@stacken.kth.se>]
8923 *) Fix problems with sizeof(long) == 8.
8924 [Andy Polyakov <appro@fy.chalmers.se>]
8926 *) Change functions to ANSI C.
8929 *) Fix typos in error codes.
8930 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8932 *) Remove defunct assembler files from Configure.
8935 *) SPARC v8 assembler BIGNUM implementation.
8936 [Andy Polyakov <appro@fy.chalmers.se>]
8938 *) Support for Certificate Policies extension: both print and set.
8939 Various additions to support the r2i method this uses.
8942 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8943 return a const string when you are expecting an allocated buffer.
8946 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8947 types DirectoryString and DisplayText.
8950 *) Add code to allow r2i extensions to access the configuration database,
8951 add an LHASH database driver and add several ctx helper functions.
8954 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8955 fail when they extended the size of a BIGNUM.
8958 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8959 support typesafe stack.
8962 *) Fix typo in SSL_[gs]et_options().
8963 [Nils Frostberg <nils@medcom.se>]
8965 *) Delete various functions and files that belonged to the (now obsolete)
8966 old X509V3 handling code.
8969 *) New Configure option "rsaref".
8972 *) Don't auto-generate pem.h.
8975 *) Introduce type-safe ASN.1 SETs.
8978 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8979 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8981 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8982 that links with OpenSSL (well at least cause lots of warnings), but fear
8983 not: the conversion is trivial, and it eliminates loads of evil casts. A
8984 few STACKed things have been converted already. Feel free to convert more.
8985 In the fullness of time, I'll do away with the STACK type altogether.
8988 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8989 specified in <certfile> by updating the entry in the index.txt file.
8990 This way one no longer has to edit the index.txt file manually for
8991 revoking a certificate. The -revoke option does the gory details now.
8992 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
8994 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
8995 `-text' option at all and this way the `-noout -text' combination was
8996 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
8997 [Ralf S. Engelschall]
8999 *) Make sure a corresponding plain text error message exists for the
9000 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9001 verify callback function determined that a certificate was revoked.
9002 [Ralf S. Engelschall]
9004 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9005 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9006 all available cipers including rc5, which was forgotten until now.
9007 In order to let the testing shell script know which algorithms
9008 are available, a new (up to now undocumented) command
9009 "openssl list-cipher-commands" is used.
9012 *) Bugfix: s_client occasionally would sleep in select() when
9013 it should have checked SSL_pending() first.
9016 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9017 the raw DSA values prior to ASN.1 encoding.
9020 *) Tweaks to Configure
9021 [Niels Poppe <niels@netbox.org>]
9023 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9027 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9030 *) New config option to avoid instructions that are illegal on the 80386.
9031 The default code is faster, but requires at least a 486.
9034 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9035 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9036 same as SSL2_VERSION anyway.
9039 *) New "-showcerts" option for s_client.
9042 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9043 application. Various cleanups and fixes.
9046 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9047 modify error routines to work internally. Add error codes and PBE init
9048 to library startup routines.
9051 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9052 packing functions to asn1 and evp. Changed function names and error
9053 codes along the way.
9056 *) PKCS12 integration: and so it begins... First of several patches to
9057 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9058 objects to objects.h
9061 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9062 and display support for Thawte strong extranet extension.
9065 *) Add LinuxPPC support.
9066 [Jeff Dubrule <igor@pobox.org>]
9068 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9069 bn_div_words in alpha.s.
9070 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9072 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9073 OAEP isn't supported when OpenSSL is built with RSAref.
9074 [Ulf Moeller <ulf@fitug.de>]
9076 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9077 so they no longer are missing under -DNOPROTO.
9078 [Soren S. Jorvang <soren@t.dk>]
9081 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9083 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9084 doesn't work when the session is reused. Coming soon!
9087 *) Fix a security hole, that allows sessions to be reused in the wrong
9088 context thus bypassing client cert protection! All software that uses
9089 client certs and session caches in multiple contexts NEEDS PATCHING to
9090 allow session reuse! A fuller solution is in the works.
9091 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9093 *) Some more source tree cleanups (removed obsolete files
9094 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9095 permission on "config" script to be executable) and a fix for the INSTALL
9097 [Ulf Moeller <ulf@fitug.de>]
9099 *) Remove some legacy and erroneous uses of malloc, free instead of
9101 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9103 *) Make rsa_oaep_test return non-zero on error.
9104 [Ulf Moeller <ulf@fitug.de>]
9106 *) Add support for native Solaris shared libraries. Configure
9107 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9108 if someone would make that last step automatic.
9109 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9111 *) ctx_size was not built with the right compiler during "make links". Fixed.
9114 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9115 except NULL ciphers". This means the default cipher list will no longer
9116 enable NULL ciphers. They need to be specifically enabled e.g. with
9117 the string "DEFAULT:eNULL".
9120 *) Fix to RSA private encryption routines: if p < q then it would
9121 occasionally produce an invalid result. This will only happen with
9122 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9125 *) Be less restrictive and allow also `perl util/perlpath.pl
9126 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9127 because this way one can also use an interpreter named `perl5' (which is
9128 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9129 installed as `perl').
9130 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9132 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9133 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9135 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9136 advapi32.lib to Win32 build and change the pem test comparision
9137 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9138 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9139 and crypto/des/ede_cbcm_enc.c.
9142 *) DES quad checksum was broken on big-endian architectures. Fixed.
9145 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9146 Win32 test batch file so it (might) work again. The Win32 test batch file
9147 is horrible: I feel ill....
9150 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9151 in e_os.h. Audit of header files to check ANSI and non ANSI
9152 sections: 10 functions were absent from non ANSI section and not exported
9153 from Windows DLLs. Fixed up libeay.num for new functions.
9156 *) Make `openssl version' output lines consistent.
9157 [Ralf S. Engelschall]
9159 *) Fix Win32 symbol export lists for BIO functions: Added
9160 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9161 to ms/libeay{16,32}.def.
9162 [Ralf S. Engelschall]
9164 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9165 fine under Unix and passes some trivial tests I've now added. But the
9166 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9167 added to make sure no one expects that this stuff really works in the
9168 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9169 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9171 [Ralf S. Engelschall]
9173 *) Fix the generation of two part addresses in perl.
9174 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9176 *) Add config entry for Linux on MIPS.
9177 [John Tobey <jtobey@channel1.com>]
9179 *) Make links whenever Configure is run, unless we are on Windoze.
9182 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9183 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9187 *) Add a useful kludge to allow package maintainers to specify compiler and
9188 other platforms details on the command line without having to patch the
9189 Configure script everytime: One now can use ``perl Configure
9190 <id>:<details>'', i.e. platform ids are allowed to have details appended
9191 to them (seperated by colons). This is treated as there would be a static
9192 pre-configured entry in Configure's %table under key <id> with value
9193 <details> and ``perl Configure <id>'' is called. So, when you want to
9194 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9195 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9196 now, which overrides the FreeBSD-elf entry on-the-fly.
9197 [Ralf S. Engelschall]
9199 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9202 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9203 on the `perl Configure ...' command line. This way one can compile
9204 OpenSSL libraries with Position Independent Code (PIC) which is needed
9205 for linking it into DSOs.
9206 [Ralf S. Engelschall]
9208 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9212 *) Cleaned up the LICENSE document: The official contact for any license
9213 questions now is the OpenSSL core team under openssl-core@openssl.org.
9214 And add a paragraph about the dual-license situation to make sure people
9215 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9216 to the OpenSSL toolkit.
9217 [Ralf S. Engelschall]
9219 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9220 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9221 Additonally cleaned up the `make links' target: Remove unnecessary
9222 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9223 to speed processing and no longer clutter the display with confusing
9224 stuff. Instead only the actually done links are displayed.
9225 [Ralf S. Engelschall]
9227 *) Permit null encryption ciphersuites, used for authentication only. It used
9228 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9229 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9233 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9234 signed attributes when verifying signatures (this would break them),
9235 the detached data encoding was wrong and public keys obtained using
9236 X509_get_pubkey() weren't freed.
9239 *) Add text documentation for the BUFFER functions. Also added a work around
9240 to a Win95 console bug. This was triggered by the password read stuff: the
9241 last character typed gets carried over to the next fread(). If you were
9242 generating a new cert request using 'req' for example then the last
9243 character of the passphrase would be CR which would then enter the first
9247 *) Added the new `Includes OpenSSL Cryptography Software' button as
9248 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9249 button and can be used by applications based on OpenSSL to show the
9250 relationship to the OpenSSL project.
9251 [Ralf S. Engelschall]
9253 *) Remove confusing variables in function signatures in files
9254 ssl/ssl_lib.c and ssl/ssl.h.
9255 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9257 *) Don't install bss_file.c under PREFIX/include/
9258 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9260 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9261 functions that return function pointers and has support for NT specific
9262 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9263 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9264 unsigned to signed types: this was killing the Win32 compile.
9267 *) Add new certificate file to stack functions,
9268 SSL_add_dir_cert_subjects_to_stack() and
9269 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9270 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9271 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9272 This means that Apache-SSL and similar packages don't have to mess around
9273 to add as many CAs as they want to the preferred list.
9276 *) Experiment with doxygen documentation. Currently only partially applied to
9278 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9279 openssl.doxy as the configuration file.
9282 *) Get rid of remaining C++-style comments which strict C compilers hate.
9283 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9285 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9286 compiled in by default: it has problems with large keys.
9289 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9290 DH private keys and/or callback functions which directly correspond to
9291 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9292 is needed for applications which have to configure certificates on a
9293 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9295 For the RSA certificate situation is makes no difference, but
9296 for the DSA certificate situation this fixes the "no shared cipher"
9297 problem where the OpenSSL cipher selection procedure failed because the
9298 temporary keys were not overtaken from the context and the API provided
9299 no way to reconfigure them.
9300 The new functions now let applications reconfigure the stuff and they
9301 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9302 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9303 non-public-API function ssl_cert_instantiate() is used as a helper
9304 function and also to reduce code redundancy inside ssl_rsa.c.
9305 [Ralf S. Engelschall]
9307 *) Move s_server -dcert and -dkey options out of the undocumented feature
9308 area because they are useful for the DSA situation and should be
9309 recognized by the users.
9310 [Ralf S. Engelschall]
9312 *) Fix the cipher decision scheme for export ciphers: the export bits are
9313 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9314 SSL_EXP_MASK. So, the original variable has to be used instead of the
9315 already masked variable.
9316 [Richard Levitte <levitte@stacken.kth.se>]
9318 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9319 [Richard Levitte <levitte@stacken.kth.se>]
9321 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9322 from `int' to `unsigned int' because it's a length and initialized by
9323 EVP_DigestFinal() which expects an `unsigned int *'.
9324 [Richard Levitte <levitte@stacken.kth.se>]
9326 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9327 script. Instead use the usual Shell->Perl transition trick.
9328 [Ralf S. Engelschall]
9330 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9331 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9332 -noout -modulus' as it's already the case for `openssl rsa -noout
9333 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9334 currently the public key is printed (a decision which was already done by
9335 `openssl dsa -modulus' in the past) which serves a similar purpose.
9336 Additionally the NO_RSA no longer completely removes the whole -modulus
9337 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9339 [Ralf S. Engelschall]
9341 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9342 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9343 [Arne Ansper <arne@ats.cyber.ee>]
9345 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9346 to be added. Now both 'req' and 'ca' can use new objects defined in the
9350 *) Add cool BIO that does syslog (or event log on NT).
9351 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9353 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9354 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9355 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9356 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9359 *) Add preliminary config info for new extension code.
9362 *) Make RSA_NO_PADDING really use no padding.
9363 [Ulf Moeller <ulf@fitug.de>]
9365 *) Generate errors when private/public key check is done.
9368 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9369 for some CRL extensions and new objects added.
9372 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9373 key usage extension and fuller support for authority key id.
9376 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9377 padding method for RSA, which is recommended for new applications in PKCS
9378 #1 v2.0 (RFC 2437, October 1998).
9379 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9380 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9381 against Bleichbacher's attack on RSA.
9382 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9385 *) Updates to the new SSL compression code
9386 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9388 *) Fix so that the version number in the master secret, when passed
9389 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9390 (because the server will not accept higher), that the version number
9391 is 0x03,0x01, not 0x03,0x00
9392 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9394 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9395 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9396 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9399 *) Support for RAW extensions where an arbitrary extension can be
9400 created by including its DER encoding. See apps/openssl.cnf for
9404 *) Make sure latest Perl versions don't interpret some generated C array
9405 code as Perl array code in the crypto/err/err_genc.pl script.
9406 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9408 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9409 not many people have the assembler. Various Win32 compilation fixes and
9410 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9414 *) Modify configure script 'Configure' to automatically create crypto/date.h
9415 file under Win32 and also build pem.h from pem.org. New script
9416 util/mkfiles.pl to create the MINFO file on environments that can't do a
9417 'make files': perl util/mkfiles.pl >MINFO should work.
9420 *) Major rework of DES function declarations, in the pursuit of correctness
9421 and purity. As a result, many evil casts evaporated, and some weirdness,
9422 too. You may find this causes warnings in your code. Zapping your evil
9423 casts will probably fix them. Mostly.
9426 *) Fix for a typo in asn1.h. Bug fix to object creation script
9427 obj_dat.pl. It considered a zero in an object definition to mean
9428 "end of object": none of the objects in objects.h have any zeros
9429 so it wasn't spotted.
9430 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9432 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9433 Masking (CBCM). In the absence of test vectors, the best I have been able
9434 to do is check that the decrypt undoes the encrypt, so far. Send me test
9435 vectors if you have them.
9438 *) Correct calculation of key length for export ciphers (too much space was
9439 allocated for null ciphers). This has not been tested!
9442 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9443 message is now correct (it understands "crypto" and "ssl" on its
9444 command line). There is also now an "update" option. This will update
9445 the util/ssleay.num and util/libeay.num files with any new functions.
9447 perl util/mkdef.pl crypto ssl update
9448 it will update them.
9451 *) Overhauled the Perl interface (perl/*):
9452 - ported BN stuff to OpenSSL's different BN library
9453 - made the perl/ source tree CVS-aware
9454 - renamed the package from SSLeay to OpenSSL (the files still contain
9455 their history because I've copied them in the repository)
9456 - removed obsolete files (the test scripts will be replaced
9457 by better Test::Harness variants in the future)
9458 [Ralf S. Engelschall]
9460 *) First cut for a very conservative source tree cleanup:
9461 1. merge various obsolete readme texts into doc/ssleay.txt
9462 where we collect the old documents and readme texts.
9463 2. remove the first part of files where I'm already sure that we no
9464 longer need them because of three reasons: either they are just temporary
9465 files which were left by Eric or they are preserved original files where
9466 I've verified that the diff is also available in the CVS via "cvs diff
9467 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9468 the crypto/md/ stuff).
9469 [Ralf S. Engelschall]
9471 *) More extension code. Incomplete support for subject and issuer alt
9472 name, issuer and authority key id. Change the i2v function parameters
9473 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9474 what that's for :-) Fix to ASN1 macro which messed up
9475 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9478 *) Preliminary support for ENUMERATED type. This is largely copied from the
9482 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9483 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9485 *) Make sure `make rehash' target really finds the `openssl' program.
9486 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9488 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9489 like to hear about it if this slows down other processors.
9492 *) Add CygWin32 platform information to Configure script.
9493 [Alan Batie <batie@aahz.jf.intel.com>]
9495 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9496 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9498 *) New program nseq to manipulate netscape certificate sequences
9501 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9505 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9506 but the BN code had some problems that would cause failures when
9507 doing certificate verification and some other functions.
9508 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9510 *) Add ASN1 and PEM code to support netscape certificate sequences.
9513 *) Add ASN1 and PEM code to support netscape certificate sequences.
9516 *) Add several PKIX and private extended key usage OIDs.
9519 *) Modify the 'ca' program to handle the new extension code. Modify
9520 openssl.cnf for new extension format, add comments.
9523 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9524 and add a sample to openssl.cnf so req -x509 now adds appropriate
9528 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9529 error code, add initial support to X509_print() and x509 application.
9532 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9533 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9534 stuff is currently isolated and isn't even compiled yet.
9537 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9538 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9539 Removed the versions check from X509 routines when loading extensions:
9540 this allows certain broken certificates that don't set the version
9541 properly to be processed.
9544 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9545 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9546 can still be regenerated with "make depend".
9549 *) Spelling mistake in C version of CAST-128.
9550 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9552 *) Changes to the error generation code. The perl script err-code.pl
9553 now reads in the old error codes and retains the old numbers, only
9554 adding new ones if necessary. It also only changes the .err files if new
9555 codes are added. The makefiles have been modified to only insert errors
9556 when needed (to avoid needlessly modifying header files). This is done
9557 by only inserting errors if the .err file is newer than the auto generated
9558 C file. To rebuild all the error codes from scratch (the old behaviour)
9559 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9560 or delete all the .err files.
9563 *) CAST-128 was incorrectly implemented for short keys. The C version has
9564 been fixed, but is untested. The assembler versions are also fixed, but
9565 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9566 to regenerate it if needed.
9567 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9568 Hagino <itojun@kame.net>]
9570 *) File was opened incorrectly in randfile.c.
9571 [Ulf Möller <ulf@fitug.de>]
9573 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9574 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9575 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9576 al: it's just almost always a UTCTime. Note this patch adds new error
9577 codes so do a "make errors" if there are problems.
9580 *) Correct Linux 1 recognition in config.
9581 [Ulf Möller <ulf@fitug.de>]
9583 *) Remove pointless MD5 hash when using DSA keys in ca.
9584 [Anonymous <nobody@replay.com>]
9586 *) Generate an error if given an empty string as a cert directory. Also
9587 generate an error if handed NULL (previously returned 0 to indicate an
9588 error, but didn't set one).
9589 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9591 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9594 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9595 parameters. This was causing a warning which killed off the Win32 compile.
9598 *) Remove C++ style comments from crypto/bn/bn_local.h.
9599 [Neil Costigan <neil.costigan@celocom.com>]
9601 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9602 based on a text string, looking up short and long names and finally
9603 "dot" format. The "dot" format stuff didn't work. Added new function
9604 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9605 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9606 OID is not part of the table.
9609 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9610 X509_LOOKUP_by_alias().
9613 *) Sort openssl functions by name.
9616 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9617 encryption from sample DSA keys (in case anyone is interested the password
9621 *) Make _all_ *_free functions accept a NULL pointer.
9622 [Frans Heymans <fheymans@isaserver.be>]
9624 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9626 [Anonymous <nobody@replay.com>]
9628 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9629 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9631 *) Don't blow it for numeric -newkey arguments to apps/req.
9632 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9634 *) Temp key "for export" tests were wrong in s3_srvr.c.
9635 [Anonymous <nobody@replay.com>]
9637 *) Add prototype for temp key callback functions
9638 SSL_CTX_set_tmp_{rsa,dh}_callback().
9641 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9642 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9645 *) X509_name_add_entry() freed the wrong thing after an error.
9646 [Arne Ansper <arne@ats.cyber.ee>]
9648 *) rsa_eay.c would attempt to free a NULL context.
9649 [Arne Ansper <arne@ats.cyber.ee>]
9651 *) BIO_s_socket() had a broken should_retry() on Windoze.
9652 [Arne Ansper <arne@ats.cyber.ee>]
9654 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9655 [Arne Ansper <arne@ats.cyber.ee>]
9657 *) Make sure the already existing X509_STORE->depth variable is initialized
9658 in X509_STORE_new(), but document the fact that this variable is still
9659 unused in the certificate verification process.
9660 [Ralf S. Engelschall]
9662 *) Fix the various library and apps files to free up pkeys obtained from
9663 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9666 *) Fix reference counting in X509_PUBKEY_get(). This makes
9667 demos/maurice/example2.c work, amongst others, probably.
9668 [Steve Henson and Ben Laurie]
9670 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9671 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9672 are no longer created. This way we have a single and consistent command
9673 line interface `openssl <command>', similar to `cvs <command>'.
9674 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9676 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9677 BIT STRING wrapper always have zero unused bits.
9680 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9683 *) Make the top-level INSTALL documentation easier to understand.
9686 *) Makefiles updated to exit if an error occurs in a sub-directory
9687 make (including if user presses ^C) [Paul Sutton]
9689 *) Make Montgomery context stuff explicit in RSA data structure.
9692 *) Fix build order of pem and err to allow for generated pem.h.
9695 *) Fix renumbering bug in X509_NAME_delete_entry().
9698 *) Enhanced the err-ins.pl script so it makes the error library number
9699 global and can add a library name. This is needed for external ASN1 and
9700 other error libraries.
9703 *) Fixed sk_insert which never worked properly.
9706 *) Fix ASN1 macros so they can handle indefinite length construted
9707 EXPLICIT tags. Some non standard certificates use these: they can now
9711 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9712 into a single doc/ssleay.txt bundle. This way the information is still
9713 preserved but no longer messes up this directory. Now it's new room for
9714 the new set of documenation files.
9715 [Ralf S. Engelschall]
9717 *) SETs were incorrectly DER encoded. This was a major pain, because they
9718 shared code with SEQUENCEs, which aren't coded the same. This means that
9719 almost everything to do with SETs or SEQUENCEs has either changed name or
9720 number of arguments.
9721 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9723 *) Fix test data to work with the above.
9726 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9727 was already fixed by Eric for 0.9.1 it seems.
9728 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9730 *) Autodetect FreeBSD3.
9733 *) Fix various bugs in Configure. This affects the following platforms:
9737 unixware-2.0-pentium
9741 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9742 before they are needed.
9745 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9749 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9751 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9752 changed SSLeay to OpenSSL in version strings.
9753 [Ralf S. Engelschall]
9755 *) Some fixups to the top-level documents.
9758 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9759 because the symlink to include/ was missing.
9760 [Ralf S. Engelschall]
9762 *) Incorporated the popular no-RSA/DSA-only patches
9763 which allow to compile a RSA-free SSLeay.
9764 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9766 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9767 when "ssleay" is still not found.
9768 [Ralf S. Engelschall]
9770 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9771 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9773 *) Updated the README file.
9774 [Ralf S. Engelschall]
9776 *) Added various .cvsignore files in the CVS repository subdirs
9777 to make a "cvs update" really silent.
9778 [Ralf S. Engelschall]
9780 *) Recompiled the error-definition header files and added
9781 missing symbols to the Win32 linker tables.
9782 [Ralf S. Engelschall]
9784 *) Cleaned up the top-level documents;
9785 o new files: CHANGES and LICENSE
9786 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9787 o merged COPYRIGHT into LICENSE
9788 o removed obsolete TODO file
9789 o renamed MICROSOFT to INSTALL.W32
9790 [Ralf S. Engelschall]
9792 *) Removed dummy files from the 0.9.1b source tree:
9793 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9794 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9795 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9796 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9797 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9798 [Ralf S. Engelschall]
9800 *) Added various platform portability fixes.
9803 *) The Genesis of the OpenSSL rpject:
9804 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9805 Young and Tim J. Hudson created while they were working for C2Net until
9807 [The OpenSSL Project]
9810 Changes between 0.9.0b and 0.9.1b [not released]
9812 *) Updated a few CA certificates under certs/
9815 *) Changed some BIGNUM api stuff.
9818 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9819 DGUX x86, Linux Alpha, etc.
9822 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9823 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9827 *) Add -strparse option to asn1pars program which parses nested
9829 [Dr Stephen Henson <shenson@bigfoot.com>]
9831 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9834 *) DSA fix for "ca" program.
9837 *) Added "-genkey" option to "dsaparam" program.
9840 *) Added RIPE MD160 (rmd160) message digest.
9843 *) Added -a (all) option to "ssleay version" command.
9846 *) Added PLATFORM define which is the id given to Configure.
9849 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9852 *) Extended the ASN.1 parser routines.
9855 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9858 *) Added a BN_CTX to the BN library.
9861 *) Fixed the weak key values in DES library
9864 *) Changed API in EVP library for cipher aliases.
9867 *) Added support for RC2/64bit cipher.
9870 *) Converted the lhash library to the crypto/mem.c functions.
9873 *) Added more recognized ASN.1 object ids.
9876 *) Added more RSA padding checks for SSL/TLS.
9879 *) Added BIO proxy/filter functionality.
9882 *) Added extra_certs to SSL_CTX which can be used
9883 send extra CA certificates to the client in the CA cert chain sending
9884 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9887 *) Now Fortezza is denied in the authentication phase because
9888 this is key exchange mechanism is not supported by SSLeay at all.
9891 *) Additional PKCS1 checks.
9894 *) Support the string "TLSv1" for all TLS v1 ciphers.
9897 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9898 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9901 *) Fixed a few memory leaks.
9904 *) Fixed various code and comment typos.
9907 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9908 bytes sent in the client random.
9909 [Edward Bishop <ebishop@spyglass.com>]