5 Changes between 1.0.0n and 1.0.0o [xx XXX xxxx]
7 *) Add support for TLS_FALLBACK_SCSV.
8 Client applications doing fallback retries should call
9 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
11 [Adam Langley, Bodo Moeller]
13 *) Add additional DigestInfo checks.
15 Reencode DigestInto in DER and check against the original when
16 verifying RSA signature: this will reject any improperly encoded
17 DigestInfo structures.
19 Note: this is a precautionary measure and no attacks are currently known.
23 Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
25 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
26 to a denial of service attack. A malicious server can crash the client
27 with a null pointer dereference (read) by specifying an anonymous (EC)DH
28 ciphersuite and sending carefully crafted handshake messages.
30 Thanks to Felix Gröbert (Google) for discovering and researching this
35 *) By sending carefully crafted DTLS packets an attacker could cause openssl
36 to leak memory. This can be exploited through a Denial of Service attack.
37 Thanks to Adam Langley for discovering and researching this issue.
41 *) An attacker can force openssl to consume large amounts of memory whilst
42 processing DTLS handshake messages. This can be exploited through a
43 Denial of Service attack.
44 Thanks to Adam Langley for discovering and researching this issue.
48 *) An attacker can force an error condition which causes openssl to crash
49 whilst processing DTLS packets due to memory being freed twice. This
50 can be exploited through a Denial of Service attack.
51 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
56 *) If a multithreaded client connects to a malicious server using a resumed
57 session and the server sends an ec point format extension it could write
58 up to 255 bytes to freed memory.
60 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
65 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
66 X509_name_oneline, X509_name_print_ex et al. to leak some information
67 from the stack. Applications may be affected if they echo pretty printing
68 output to the attacker.
70 Thanks to Ivan Fratric (Google) for discovering this issue.
72 [Emilia Käsper, and Steve Henson]
74 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
75 for corner cases. (Certain input points at infinity could lead to
76 bogus results, with non-infinity inputs mapped to infinity too.)
79 Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
81 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
82 handshake can force the use of weak keying material in OpenSSL
83 SSL/TLS clients and servers.
85 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
86 researching this issue. (CVE-2014-0224)
87 [KIKUCHI Masashi, Steve Henson]
89 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
90 OpenSSL DTLS client the code can be made to recurse eventually crashing
93 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
95 [Imre Rad, Steve Henson]
97 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
98 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
99 client or server. This is potentially exploitable to run arbitrary
100 code on a vulnerable client or server.
102 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
103 [Jüri Aedla, Steve Henson]
105 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
106 are subject to a denial of service attack.
108 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
109 this issue. (CVE-2014-3470)
110 [Felix Gröbert, Ivan Fratric, Steve Henson]
112 *) Harmonize version and its documentation. -f flag is used to display
114 [mancha <mancha1@zoho.com>]
116 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
118 [mancha <mancha1@zoho.com>]
120 *) Fix some double frees. These are not thought to be exploitable.
121 [mancha <mancha1@zoho.com>]
123 *) Fix for the attack described in the paper "Recovering OpenSSL
124 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
125 by Yuval Yarom and Naomi Benger. Details can be obtained from:
126 http://eprint.iacr.org/2014/140
128 Thanks to Yuval Yarom and Naomi Benger for discovering this
129 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
130 [Yuval Yarom and Naomi Benger]
132 Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
134 *) Keep original DTLS digest and encryption contexts in retransmission
135 structures so we can use the previous session parameters if they need
136 to be resent. (CVE-2013-6450)
139 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
140 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
141 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
142 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
143 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
144 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
145 [Rob Stradling, Adam Langley]
147 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
149 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
151 This addresses the flaw in CBC record processing discovered by
152 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
153 at: http://www.isg.rhul.ac.uk/tls/
155 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
156 Security Group at Royal Holloway, University of London
157 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
158 Emilia Käsper for the initial patch.
160 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
162 *) Return an error when checking OCSP signatures when key is NULL.
163 This fixes a DoS attack. (CVE-2013-0166)
166 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
167 the right response is stapled. Also change SSL_get_certificate()
168 so it returns the certificate actually sent.
169 See http://rt.openssl.org/Ticket/Display.html?id=2836.
171 [Rob Stradling <rob.stradling@comodo.com>]
173 *) Fix possible deadlock when decoding public keys.
176 Changes between 1.0.0i and 1.0.0j [10 May 2012]
178 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
181 *) Sanity check record length before skipping explicit IV in DTLS
184 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
185 fuzzing as a service testing platform.
189 *) Initialise tkeylen properly when encrypting CMS messages.
190 Thanks to Solar Designer of Openwall for reporting this issue.
193 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
195 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
196 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
197 in CRYPTO_realloc_clean.
199 Thanks to Tavis Ormandy, Google Security Team, for discovering this
200 issue and to Adam Langley <agl@chromium.org> for fixing it.
202 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
204 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
206 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
207 in CMS and PKCS7 code. When RSA decryption fails use a random key for
208 content decryption and always return the same error. Note: this attack
209 needs on average 2^20 messages so it only affects automated senders. The
210 old behaviour can be reenabled in the CMS code by setting the
211 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
212 an MMA defence is not necessary.
213 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
214 this issue. (CVE-2012-0884)
217 *) Fix CVE-2011-4619: make sure we really are receiving a
218 client hello before rejecting multiple SGC restarts. Thanks to
219 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
222 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
224 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
225 Thanks to Antonio Martin, Enterprise Secure Access Research and
226 Development, Cisco Systems, Inc. for discovering this bug and
227 preparing a fix. (CVE-2012-0050)
230 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
232 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
233 of the Vaudenay padding oracle attack on CBC mode encryption
234 which enables an efficient plaintext recovery attack against
235 the OpenSSL implementation of DTLS. Their attack exploits timing
236 differences arising during decryption processing. A research
237 paper describing this attack can be found at:
238 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
239 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
240 Security Group at Royal Holloway, University of London
241 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
242 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
243 for preparing the fix. (CVE-2011-4108)
244 [Robin Seggelmann, Michael Tuexen]
246 *) Clear bytes used for block padding of SSL 3.0 records.
248 [Adam Langley (Google)]
250 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
251 Kadianakis <desnacked@gmail.com> for discovering this issue and
252 Adam Langley for preparing the fix. (CVE-2011-4619)
253 [Adam Langley (Google)]
255 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
256 [Andrey Kulikov <amdeich@gmail.com>]
258 *) Prevent malformed RFC3779 data triggering an assertion failure.
259 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
260 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
261 [Rob Austein <sra@hactrn.net>]
263 *) Improved PRNG seeding for VOS.
264 [Paul Green <Paul.Green@stratus.com>]
266 *) Fix ssl_ciph.c set-up race.
267 [Adam Langley (Google)]
269 *) Fix spurious failures in ecdsatest.c.
270 [Emilia Käsper (Google)]
272 *) Fix the BIO_f_buffer() implementation (which was mixing different
273 interpretations of the '..._len' fields).
274 [Adam Langley (Google)]
276 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
277 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
278 threads won't reuse the same blinding coefficients.
280 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
281 lock to call BN_BLINDING_invert_ex, and avoids one use of
282 BN_BLINDING_update for each BN_BLINDING structure (previously,
283 the last update always remained unused).
284 [Emilia Käsper (Google)]
286 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
287 [Bob Buckholz (Google)]
289 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
291 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
292 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
293 [Kaspar Brand <ossl@velox.ch>]
295 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
296 for multi-threaded use of ECDH. (CVE-2011-3210)
297 [Adam Langley (Google)]
299 *) Fix x509_name_ex_d2i memory leak on bad inputs.
302 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
303 signature public key algorithm by using OID xref utilities instead.
304 Before this you could only use some ECC ciphersuites with SHA1 only.
307 *) Add protection against ECDSA timing attacks as mentioned in the paper
308 by Billy Bob Brumley and Nicola Tuveri, see:
310 http://eprint.iacr.org/2011/232.pdf
312 [Billy Bob Brumley and Nicola Tuveri]
314 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
316 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
317 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
319 *) Fix bug in string printing code: if *any* escaping is enabled we must
320 escape the escape character (backslash) or the resulting string is
324 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
326 *) Disable code workaround for ancient and obsolete Netscape browsers
327 and servers: an attacker can use it in a ciphersuite downgrade attack.
328 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
331 *) Fixed J-PAKE implementation error, originally discovered by
332 Sebastien Martini, further info and confirmation from Stefan
333 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
336 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
338 *) Fix extension code to avoid race conditions which can result in a buffer
339 overrun vulnerability: resumed sessions must not be modified as they can
340 be shared by multiple threads. CVE-2010-3864
343 *) Fix WIN32 build system to correctly link an ENGINE directory into
347 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
349 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
351 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
353 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
355 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
356 context. The operation can be customised via the ctrl mechanism in
357 case ENGINEs want to include additional functionality.
360 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
363 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
364 output hashes compatible with older versions of OpenSSL.
365 [Willy Weisz <weisz@vcpc.univie.ac.at>]
367 *) Fix compression algorithm handling: if resuming a session use the
368 compression algorithm of the resumed session instead of determining
369 it from client hello again. Don't allow server to change algorithm.
372 *) Add load_crls() function to apps tidying load_certs() too. Add option
373 to verify utility to allow additional CRLs to be included.
376 *) Update OCSP request code to permit adding custom headers to the request:
377 some responders need this.
380 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
382 [Julia Lawall <julia@diku.dk>]
384 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
385 needlessly dereferenced structures, used obsolete functions and
386 didn't handle all updated verify codes correctly.
389 *) Disable MD2 in the default configuration.
392 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
393 indicate the initial BIO being pushed or popped. This makes it possible
394 to determine whether the BIO is the one explicitly called or as a result
395 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
396 it handles reference counts correctly and doesn't zero out the I/O bio
397 when it is not being explicitly popped. WARNING: applications which
398 included workarounds for the old buggy behaviour will need to be modified
399 or they could free up already freed BIOs.
402 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
403 renaming to all platforms (within the 0.9.8 branch, this was
404 done conditionally on Netware platforms to avoid a name clash).
405 [Guenter <lists@gknw.net>]
407 *) Add ECDHE and PSK support to DTLS.
408 [Michael Tuexen <tuexen@fh-muenster.de>]
410 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
414 *) Add "missing" function EVP_MD_flags() (without this the only way to
415 retrieve a digest flags is by accessing the structure directly. Update
416 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
417 or cipher is registered as in the "from" argument. Print out all
418 registered digests in the dgst usage message instead of manually
419 attempting to work them out.
422 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
423 this allows the use of compression and extensions. Change default cipher
424 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
425 by default unless an application cipher string requests it.
428 *) Alter match criteria in PKCS12_parse(). It used to try to use local
429 key ids to find matching certificates and keys but some PKCS#12 files
430 don't follow the (somewhat unwritten) rules and this strategy fails.
431 Now just gather all certificates together and the first private key
432 then look for the first certificate that matches the key.
435 *) Support use of registered digest and cipher names for dgst and cipher
436 commands instead of having to add each one as a special case. So now
443 openssl dgst -sha256 foo
445 and this works for ENGINE based algorithms too.
449 *) Update Gost ENGINE to support parameter files.
450 [Victor B. Wagner <vitus@cryptocom.ru>]
452 *) Support GeneralizedTime in ca utility.
453 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
455 *) Enhance the hash format used for certificate directory links. The new
456 form uses the canonical encoding (meaning equivalent names will work
457 even if they aren't identical) and uses SHA1 instead of MD5. This form
458 is incompatible with the older format and as a result c_rehash should
459 be used to rebuild symbolic links.
462 *) Make PKCS#8 the default write format for private keys, replacing the
463 traditional format. This form is standardised, more secure and doesn't
464 include an implicit MD5 dependency.
467 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
468 committed to OpenSSL should pass this lot as a minimum.
471 *) Add session ticket override functionality for use by EAP-FAST.
472 [Jouni Malinen <j@w1.fi>]
474 *) Modify HMAC functions to return a value. Since these can be implemented
475 in an ENGINE errors can occur.
478 *) Type-checked OBJ_bsearch_ex.
481 *) Type-checked OBJ_bsearch. Also some constification necessitated
482 by type-checking. Still to come: TXT_DB, bsearch(?),
483 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
487 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
488 seconds to a tm structure directly, instead of going through OS
489 specific date routines. This avoids any issues with OS routines such
490 as the year 2038 bug. New *_adj() functions for ASN1 time structures
491 and X509_time_adj_ex() to cover the extended range. The existing
492 X509_time_adj() is still usable and will no longer have any date issues.
495 *) Delta CRL support. New use deltas option which will attempt to locate
496 and search any appropriate delta CRLs available.
498 This work was sponsored by Google.
501 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
502 code and add additional score elements. Validate alternate CRL paths
503 as part of the CRL checking and indicate a new error "CRL path validation
504 error" in this case. Applications wanting additional details can use
505 the verify callback and check the new "parent" field. If this is not
506 NULL CRL path validation is taking place. Existing applications wont
507 see this because it requires extended CRL support which is off by
510 This work was sponsored by Google.
513 *) Support for freshest CRL extension.
515 This work was sponsored by Google.
518 *) Initial indirect CRL support. Currently only supported in the CRLs
519 passed directly and not via lookup. Process certificate issuer
520 CRL entry extension and lookup CRL entries by bother issuer name
521 and serial number. Check and process CRL issuer entry in IDP extension.
523 This work was sponsored by Google.
526 *) Add support for distinct certificate and CRL paths. The CRL issuer
527 certificate is validated separately in this case. Only enabled if
528 an extended CRL support flag is set: this flag will enable additional
529 CRL functionality in future.
531 This work was sponsored by Google.
534 *) Add support for policy mappings extension.
536 This work was sponsored by Google.
539 *) Fixes to pathlength constraint, self issued certificate handling,
540 policy processing to align with RFC3280 and PKITS tests.
542 This work was sponsored by Google.
545 *) Support for name constraints certificate extension. DN, email, DNS
546 and URI types are currently supported.
548 This work was sponsored by Google.
551 *) To cater for systems that provide a pointer-based thread ID rather
552 than numeric, deprecate the current numeric thread ID mechanism and
553 replace it with a structure and associated callback type. This
554 mechanism allows a numeric "hash" to be extracted from a thread ID in
555 either case, and on platforms where pointers are larger than 'long',
556 mixing is done to help ensure the numeric 'hash' is usable even if it
557 can't be guaranteed unique. The default mechanism is to use "&errno"
558 as a pointer-based thread ID to distinguish between threads.
560 Applications that want to provide their own thread IDs should now use
561 CRYPTO_THREADID_set_callback() to register a callback that will call
562 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
564 Note that ERR_remove_state() is now deprecated, because it is tied
565 to the assumption that thread IDs are numeric. ERR_remove_state(0)
566 to free the current thread's error state should be replaced by
567 ERR_remove_thread_state(NULL).
569 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
570 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
571 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
572 application was previously providing a numeric thread callback that
573 was inappropriate for distinguishing threads, then uniqueness might
574 have been obtained with &errno that happened immediately in the
575 intermediate development versions of OpenSSL; this is no longer the
576 case, the numeric thread callback will now override the automatic use
578 [Geoff Thorpe, with help from Bodo Moeller]
580 *) Initial support for different CRL issuing certificates. This covers a
581 simple case where the self issued certificates in the chain exist and
582 the real CRL issuer is higher in the existing chain.
584 This work was sponsored by Google.
587 *) Removed effectively defunct crypto/store from the build.
590 *) Revamp of STACK to provide stronger type-checking. Still to come:
591 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
592 ASN1_STRING, CONF_VALUE.
595 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
596 RAM on SSL connections. This option can save about 34k per idle SSL.
599 *) Revamp of LHASH to provide stronger type-checking. Still to come:
600 STACK, TXT_DB, bsearch, qsort.
603 *) Initial support for Cryptographic Message Syntax (aka CMS) based
604 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
605 support for data, signedData, compressedData, digestedData and
606 encryptedData, envelopedData types included. Scripts to check against
607 RFC4134 examples draft and interop and consistency checks of many
608 content types and variants.
611 *) Add options to enc utility to support use of zlib compression BIO.
614 *) Extend mk1mf to support importing of options and assembly language
615 files from Configure script, currently only included in VC-WIN32.
616 The assembly language rules can now optionally generate the source
617 files from the associated perl scripts.
620 *) Implement remaining functionality needed to support GOST ciphersuites.
621 Interop testing has been performed using CryptoPro implementations.
622 [Victor B. Wagner <vitus@cryptocom.ru>]
624 *) s390x assembler pack.
627 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
631 *) Implement Opaque PRF Input TLS extension as specified in
632 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
633 official specification yet and no extension type assignment by
634 IANA exists, this extension (for now) will have to be explicitly
635 enabled when building OpenSSL by providing the extension number
636 to use. For example, specify an option
638 -DTLSEXT_TYPE_opaque_prf_input=0x9527
640 to the "config" or "Configure" script to enable the extension,
641 assuming extension number 0x9527 (which is a completely arbitrary
642 and unofficial assignment based on the MD5 hash of the Internet
643 Draft). Note that by doing so, you potentially lose
644 interoperability with other TLS implementations since these might
645 be using the same extension number for other purposes.
647 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
648 opaque PRF input value to use in the handshake. This will create
649 an interal copy of the length-'len' string at 'src', and will
650 return non-zero for success.
652 To get more control and flexibility, provide a callback function
655 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
656 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
660 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
663 Callback function 'cb' will be called in handshakes, and is
664 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
665 Argument 'arg' is for application purposes (the value as given to
666 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
667 be provided to the callback function). The callback function
668 has to return non-zero to report success: usually 1 to use opaque
669 PRF input just if possible, or 2 to enforce use of the opaque PRF
670 input. In the latter case, the library will abort the handshake
671 if opaque PRF input is not successfully negotiated.
673 Arguments 'peerinput' and 'len' given to the callback function
674 will always be NULL and 0 in the case of a client. A server will
675 see the client's opaque PRF input through these variables if
676 available (NULL and 0 otherwise). Note that if the server
677 provides an opaque PRF input, the length must be the same as the
678 length of the client's opaque PRF input.
680 Note that the callback function will only be called when creating
681 a new session (session resumption can resume whatever was
682 previously negotiated), and will not be called in SSL 2.0
683 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
684 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
685 for applications that need to enforce opaque PRF input.
689 *) Update ssl code to support digests other than SHA1+MD5 for handshake
692 [Victor B. Wagner <vitus@cryptocom.ru>]
694 *) Add RFC4507 support to OpenSSL. This includes the corrections in
695 RFC4507bis. The encrypted ticket format is an encrypted encoded
696 SSL_SESSION structure, that way new session features are automatically
699 If a client application caches session in an SSL_SESSION structure
700 support is transparent because tickets are now stored in the encoded
703 The SSL_CTX structure automatically generates keys for ticket
704 protection in servers so again support should be possible
705 with no application modification.
707 If a client or server wishes to disable RFC4507 support then the option
708 SSL_OP_NO_TICKET can be set.
710 Add a TLS extension debugging callback to allow the contents of any client
711 or server extensions to be examined.
713 This work was sponsored by Google.
716 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
717 OpenSSL should now compile cleanly on gcc 4.2
718 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
720 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
721 support including streaming MAC support: this is required for GOST
723 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
725 *) Add option -stream to use PKCS#7 streaming in smime utility. New
726 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
727 to output in BER and PEM format.
730 *) Experimental support for use of HMAC via EVP_PKEY interface. This
731 allows HMAC to be handled via the EVP_DigestSign*() interface. The
732 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
733 ENGINE support for HMAC keys which are unextractable. New -mac and
734 -macopt options to dgst utility.
737 *) New option -sigopt to dgst utility. Update dgst to use
738 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
739 alternative signing paramaters such as X9.31 or PSS in the dgst
743 *) Change ssl_cipher_apply_rule(), the internal function that does
744 the work each time a ciphersuite string requests enabling
745 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
746 removing ("!foo+bar") a class of ciphersuites: Now it maintains
747 the order of disabled ciphersuites such that those ciphersuites
748 that most recently went from enabled to disabled not only stay
749 in order with respect to each other, but also have higher priority
750 than other disabled ciphersuites the next time ciphersuites are
753 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
754 the same ciphersuites as with "HIGH" alone, but in a specific
755 order where the PSK ciphersuites come first (since they are the
756 most recently disabled ciphersuites when "HIGH" is parsed).
758 Also, change ssl_create_cipher_list() (using this new
759 funcionality) such that between otherwise identical
760 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
764 *) Change ssl_create_cipher_list() so that it automatically
765 arranges the ciphersuites in reasonable order before starting
766 to process the rule string. Thus, the definition for "DEFAULT"
767 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
768 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
769 This makes it much easier to arrive at a reasonable default order
770 in applications for which anonymous ciphers are OK (meaning
771 that you can't actually use DEFAULT).
772 [Bodo Moeller; suggested by Victor Duchovni]
774 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
775 processing) into multiple integers instead of setting
776 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
777 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
778 (These masks as well as the individual bit definitions are hidden
779 away into the non-exported interface ssl/ssl_locl.h, so this
780 change to the definition of the SSL_CIPHER structure shouldn't
781 affect applications.) This give us more bits for each of these
782 categories, so there is no longer a need to coagulate AES128 and
783 AES256 into a single algorithm bit, and to coagulate Camellia128
784 and Camellia256 into a single algorithm bit, which has led to all
787 Thus, among other things, the kludge introduced in 0.9.7m and
788 0.9.8e for masking out AES256 independently of AES128 or masking
789 out Camellia256 independently of AES256 is not needed here in 0.9.9.
791 With the change, we also introduce new ciphersuite aliases that
792 so far were missing: "AES128", "AES256", "CAMELLIA128", and
796 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
797 Use the leftmost N bytes of the signature input if the input is
798 larger than the prime q (with N being the size in bytes of q).
801 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
802 it yet and it is largely untested.
805 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
808 *) Initial incomplete changes to avoid need for function casts in OpenSSL
809 some compilers (gcc 4.2 and later) reject their use. Safestack is
810 reimplemented. Update ASN1 to avoid use of legacy functions.
813 *) Win32/64 targets are linked with Winsock2.
816 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
817 to external functions. This can be used to increase CRL handling
818 efficiency especially when CRLs are very large by (for example) storing
819 the CRL revoked certificates in a database.
822 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
823 new CRLs added to a directory can be used. New command line option
824 -verify_return_error to s_client and s_server. This causes real errors
825 to be returned by the verify callback instead of carrying on no matter
826 what. This reflects the way a "real world" verify callback would behave.
829 *) GOST engine, supporting several GOST algorithms and public key formats.
830 Kindly donated by Cryptocom.
833 *) Partial support for Issuing Distribution Point CRL extension. CRLs
834 partitioned by DP are handled but no indirect CRL or reason partitioning
835 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
836 selected via a scoring technique which handles IDP and AKID in CRLs.
839 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
840 will ultimately be used for all verify operations: this will remove the
841 X509_STORE dependency on certificate verification and allow alternative
842 lookup methods. X509_STORE based implementations of these two callbacks.
845 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
846 Modify get_crl() to find a valid (unexpired) CRL if possible.
849 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
850 this would be called X509_CRL_cmp() but that name is already used by
851 a function that just compares CRL issuer names. Cache several CRL
852 extensions in X509_CRL structure and cache CRLDP in X509.
855 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
856 this maps equivalent X509_NAME structures into a consistent structure.
857 Name comparison can then be performed rapidly using memcmp().
860 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
864 *) Allow digests to supply their own micalg string for S/MIME type using
865 the ctrl EVP_MD_CTRL_MICALG.
868 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
869 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
870 ctrl. It can then customise the structure before and/or after signing
874 *) New function OBJ_add_sigid() to allow application defined signature OIDs
875 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
876 to free up any added signature OIDs.
879 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
880 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
881 digest and cipher tables. New options added to openssl utility:
882 list-message-digest-algorithms and list-cipher-algorithms.
885 *) Change the array representation of binary polynomials: the list
886 of degrees of non-zero coefficients is now terminated with -1.
887 Previously it was terminated with 0, which was also part of the
888 value; thus, the array representation was not applicable to
889 polynomials where t^0 has coefficient zero. This change makes
890 the array representation useful in a more general context.
893 *) Various modifications and fixes to SSL/TLS cipher string
894 handling. For ECC, the code now distinguishes between fixed ECDH
895 with RSA certificates on the one hand and with ECDSA certificates
896 on the other hand, since these are separate ciphersuites. The
897 unused code for Fortezza ciphersuites has been removed.
899 For consistency with EDH, ephemeral ECDH is now called "EECDH"
900 (not "ECDHE"). For consistency with the code for DH
901 certificates, use of ECDH certificates is now considered ECDH
902 authentication, not RSA or ECDSA authentication (the latter is
903 merely the CA's signing algorithm and not actively used in the
906 The temporary ciphersuite alias "ECCdraft" is no longer
907 available, and ECC ciphersuites are no longer excluded from "ALL"
908 and "DEFAULT". The following aliases now exist for RFC 4492
909 ciphersuites, most of these by analogy with the DH case:
911 kECDHr - ECDH cert, signed with RSA
912 kECDHe - ECDH cert, signed with ECDSA
913 kECDH - ECDH cert (signed with either RSA or ECDSA)
914 kEECDH - ephemeral ECDH
915 ECDH - ECDH cert or ephemeral ECDH
921 AECDH - anonymous ECDH
922 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
926 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
927 Use correct micalg parameters depending on digest(s) in signed message.
930 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
931 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
934 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
935 an engine to register a method. Add ENGINE lookups for methods and
936 functional reference processing.
939 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
940 EVP_{Sign,Verify}* which allow an application to customise the signature
944 *) New -resign option to smime utility. This adds one or more signers
945 to an existing PKCS#7 signedData structure. Also -md option to use an
946 alternative message digest algorithm for signing.
949 *) Tidy up PKCS#7 routines and add new functions to make it easier to
950 create PKCS7 structures containing multiple signers. Update smime
951 application to support multiple signers.
954 *) New -macalg option to pkcs12 utility to allow setting of an alternative
958 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
959 Reorganize PBE internals to lookup from a static table using NIDs,
960 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
961 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
962 PRF which will be automatically used with PBES2.
965 *) Replace the algorithm specific calls to generate keys in "req" with the
969 *) Update PKCS#7 enveloped data routines to use new API. This is now
970 supported by any public key method supporting the encrypt operation. A
971 ctrl is added to allow the public key algorithm to examine or modify
972 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
976 *) Add a ctrl to asn1 method to allow a public key algorithm to express
977 a default digest type to use. In most cases this will be SHA1 but some
978 algorithms (such as GOST) need to specify an alternative digest. The
979 return value indicates how strong the prefernce is 1 means optional and
980 2 is mandatory (that is it is the only supported type). Modify
981 ASN1_item_sign() to accept a NULL digest argument to indicate it should
982 use the default md. Update openssl utilities to use the default digest
983 type for signing if it is not explicitly indicated.
986 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
987 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
988 signing method from the key type. This effectively removes the link
989 between digests and public key types.
992 *) Add an OID cross reference table and utility functions. Its purpose is to
993 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
994 rsaEncryption. This will allow some of the algorithm specific hackery
995 needed to use the correct OID to be removed.
998 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
999 structures for PKCS7_sign(). They are now set up by the relevant public
1003 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1006 *) Add support for key derivation (agreement) in the API, DH method and
1010 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1011 public and private key formats. As a side effect these add additional
1012 command line functionality not previously available: DSA signatures can be
1013 generated and verified using pkeyutl and DH key support and generation in
1018 [Oliver Tappe <zooey@hirschkaefer.de>]
1020 *) New make target "install_html_docs" installs HTML renditions of the
1022 [Oliver Tappe <zooey@hirschkaefer.de>]
1024 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1025 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1026 support key and parameter generation and add initial key generation
1027 functionality for RSA.
1030 *) Add functions for main EVP_PKEY_method operations. The undocumented
1031 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1032 EVP_PKEY_{encrypt,decrypt}_old.
1035 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1036 key API, doesn't do much yet.
1039 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1040 public key algorithms. New option to openssl utility:
1041 "list-public-key-algorithms" to print out info.
1044 *) Implement the Supported Elliptic Curves Extension for
1045 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1048 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1049 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1052 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1053 utilities such as rsa, dsa, dsaparam etc except they process any key
1057 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1058 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1059 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1063 *) Initial support for pluggable public key ASN1.
1064 De-spaghettify the public key ASN1 handling. Move public and private
1065 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1066 algorithm specific handling to a single module within the relevant
1067 algorithm directory. Add functions to allow (near) opaque processing
1068 of public and private key structures.
1071 *) Implement the Supported Point Formats Extension for
1072 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1075 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1076 for the psk identity [hint] and the psk callback functions to the
1077 SSL_SESSION, SSL and SSL_CTX structure.
1080 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1084 SSL_CTX_use_psk_identity_hint
1085 SSL_get_psk_identity_hint
1086 SSL_get_psk_identity
1087 SSL_use_psk_identity_hint
1089 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1091 *) Add RFC 3161 compliant time stamp request creation, response generation
1092 and response verification functionality.
1093 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1095 *) Add initial support for TLS extensions, specifically for the server_name
1096 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1097 have new members for a host name. The SSL data structure has an
1098 additional member SSL_CTX *initial_ctx so that new sessions can be
1099 stored in that context to allow for session resumption, even after the
1100 SSL has been switched to a new SSL_CTX in reaction to a client's
1101 server_name extension.
1103 New functions (subject to change):
1105 SSL_get_servername()
1106 SSL_get_servername_type()
1109 New CTRL codes and macros (subject to change):
1111 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1112 - SSL_CTX_set_tlsext_servername_callback()
1113 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1114 - SSL_CTX_set_tlsext_servername_arg()
1115 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1117 openssl s_client has a new '-servername ...' option.
1119 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1120 '-key2 ...', '-servername_fatal' (subject to change). This allows
1121 testing the HostName extension for a specific single host name ('-cert'
1122 and '-key' remain fallbacks for handshakes without HostName
1123 negotiation). If the unrecogninzed_name alert has to be sent, this by
1124 default is a warning; it becomes fatal with the '-servername_fatal'
1127 [Peter Sylvester, Remy Allais, Christophe Renou]
1129 *) Whirlpool hash implementation is added.
1132 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1133 bn(64,32). Because of instruction set limitations it doesn't have
1134 any negative impact on performance. This was done mostly in order
1135 to make it possible to share assembler modules, such as bn_mul_mont
1136 implementations, between 32- and 64-bit builds without hassle.
1139 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1140 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1144 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1145 dedicated Montgomery multiplication procedure, is introduced.
1146 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1147 "64-bit" performance on certain 32-bit targets.
1150 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1151 in SSL structures. New SSL ctrl to set maximum send fragment size.
1152 Save memory by seeting the I/O buffer sizes dynamically instead of
1153 using the maximum available value.
1156 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1157 in addition to the text details.
1160 *) Very, very preliminary EXPERIMENTAL support for printing of general
1161 ASN1 structures. This currently produces rather ugly output and doesn't
1162 handle several customised structures at all.
1165 *) Integrated support for PVK file format and some related formats such
1166 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1167 these in the 'rsa' and 'dsa' utilities.
1170 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1173 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1174 place for the (very old) "NETSCAPE" format certificates which are now
1175 handled using new ASN1 code equivalents.
1178 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1179 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1180 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1183 *) Modify CRL distribution points extension code to print out previously
1184 unsupported fields. Enhance extension setting code to allow setting of
1188 *) Add print and set support for Issuing Distribution Point CRL extension.
1191 *) Change 'Configure' script to enable Camellia by default.
1194 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
1196 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1198 This addresses the flaw in CBC record processing discovered by
1199 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1200 at: http://www.isg.rhul.ac.uk/tls/
1202 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1203 Security Group at Royal Holloway, University of London
1204 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1205 Emilia Käsper for the initial patch.
1207 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1209 *) Return an error when checking OCSP signatures when key is NULL.
1210 This fixes a DoS attack. (CVE-2013-0166)
1213 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1214 the right response is stapled. Also change SSL_get_certificate()
1215 so it returns the certificate actually sent.
1216 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1217 (This is a backport)
1218 [Rob Stradling <rob.stradling@comodo.com>]
1220 *) Fix possible deadlock when decoding public keys.
1223 Changes between 0.9.8w and 0.9.8x [10 May 2012]
1225 *) Sanity check record length before skipping explicit IV in DTLS
1228 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1229 fuzzing as a service testing platform.
1233 *) Initialise tkeylen properly when encrypting CMS messages.
1234 Thanks to Solar Designer of Openwall for reporting this issue.
1237 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
1239 *) The fix for CVE-2012-2110 did not take into account that the
1240 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
1241 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
1242 rejecting negative len parameter. (CVE-2012-2131)
1243 [Tomas Hoger <thoger@redhat.com>]
1245 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
1247 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1248 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1249 in CRYPTO_realloc_clean.
1251 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1252 issue and to Adam Langley <agl@chromium.org> for fixing it.
1254 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1256 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
1258 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1259 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1260 content decryption and always return the same error. Note: this attack
1261 needs on average 2^20 messages so it only affects automated senders. The
1262 old behaviour can be reenabled in the CMS code by setting the
1263 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1264 an MMA defence is not necessary.
1265 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1266 this issue. (CVE-2012-0884)
1269 *) Fix CVE-2011-4619: make sure we really are receiving a
1270 client hello before rejecting multiple SGC restarts. Thanks to
1271 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1274 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1276 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1277 Thanks to Antonio Martin, Enterprise Secure Access Research and
1278 Development, Cisco Systems, Inc. for discovering this bug and
1279 preparing a fix. (CVE-2012-0050)
1282 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1284 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1285 of the Vaudenay padding oracle attack on CBC mode encryption
1286 which enables an efficient plaintext recovery attack against
1287 the OpenSSL implementation of DTLS. Their attack exploits timing
1288 differences arising during decryption processing. A research
1289 paper describing this attack can be found at:
1290 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1291 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1292 Security Group at Royal Holloway, University of London
1293 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1294 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1295 for preparing the fix. (CVE-2011-4108)
1296 [Robin Seggelmann, Michael Tuexen]
1298 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1299 [Ben Laurie, Kasper <ekasper@google.com>]
1301 *) Clear bytes used for block padding of SSL 3.0 records.
1303 [Adam Langley (Google)]
1305 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1306 Kadianakis <desnacked@gmail.com> for discovering this issue and
1307 Adam Langley for preparing the fix. (CVE-2011-4619)
1308 [Adam Langley (Google)]
1310 *) Prevent malformed RFC3779 data triggering an assertion failure.
1311 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1312 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1313 [Rob Austein <sra@hactrn.net>]
1315 *) Fix ssl_ciph.c set-up race.
1316 [Adam Langley (Google)]
1318 *) Fix spurious failures in ecdsatest.c.
1319 [Emilia Käsper (Google)]
1321 *) Fix the BIO_f_buffer() implementation (which was mixing different
1322 interpretations of the '..._len' fields).
1323 [Adam Langley (Google)]
1325 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1326 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1327 threads won't reuse the same blinding coefficients.
1329 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1330 lock to call BN_BLINDING_invert_ex, and avoids one use of
1331 BN_BLINDING_update for each BN_BLINDING structure (previously,
1332 the last update always remained unused).
1333 [Emilia Käsper (Google)]
1335 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1336 for multi-threaded use of ECDH.
1337 [Adam Langley (Google)]
1339 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1342 *) Add protection against ECDSA timing attacks as mentioned in the paper
1343 by Billy Bob Brumley and Nicola Tuveri, see:
1345 http://eprint.iacr.org/2011/232.pdf
1347 [Billy Bob Brumley and Nicola Tuveri]
1349 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1351 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1352 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1354 *) Fix bug in string printing code: if *any* escaping is enabled we must
1355 escape the escape character (backslash) or the resulting string is
1359 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1361 *) Disable code workaround for ancient and obsolete Netscape browsers
1362 and servers: an attacker can use it in a ciphersuite downgrade attack.
1363 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1366 *) Fixed J-PAKE implementation error, originally discovered by
1367 Sebastien Martini, further info and confirmation from Stefan
1368 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1371 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1373 *) Fix extension code to avoid race conditions which can result in a buffer
1374 overrun vulnerability: resumed sessions must not be modified as they can
1375 be shared by multiple threads. CVE-2010-3864
1378 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1381 *) Don't reencode certificate when calculating signature: cache and use
1382 the original encoding instead. This makes signature verification of
1383 some broken encodings work correctly.
1386 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1387 is also one of the inputs.
1388 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1390 *) Don't repeatedly append PBE algorithms to table if they already exist.
1391 Sort table on each new add. This effectively makes the table read only
1392 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1396 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1398 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1401 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1402 access or freeing data twice (CVE-2010-0742)
1403 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1405 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1406 common in certificates and some applications which only call
1407 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1411 Reduce copying into .apps and .test in makevms.com
1412 Don't try to use blank CA certificate in CA.com
1413 Allow use of C files from original directories in maketests.com
1414 [Steven M. Schweda" <sms@antinode.info>]
1416 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1418 *) When rejecting SSL/TLS records due to an incorrect version number, never
1419 update s->server with a new major version number. As of
1420 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1421 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1422 the previous behavior could result in a read attempt at NULL when
1423 receiving specific incorrect SSL/TLS records once record payload
1424 protection is active. (CVE-2010-0740)
1425 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1427 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1428 could be crashed if the relevant tables were not present (e.g. chrooted).
1429 [Tomas Hoger <thoger@redhat.com>]
1431 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1433 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1434 [Martin Olsson, Neel Mehta]
1436 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1437 accommodate for stack sorting, always a write lock!).
1440 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1441 excessive delays in the RAND_poll(): over a minute. As a workaround
1442 include a time check in the inner Heap32Next loop too.
1445 *) The code that handled flushing of data in SSL/TLS originally used the
1446 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1447 the problem outlined in PR#1949. The fix suggested there however can
1448 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1449 of Apache). So instead simplify the code to flush unconditionally.
1450 This should be fine since flushing with no data to flush is a no op.
1453 *) Handle TLS versions 2.0 and later properly and correctly use the
1454 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1455 off ancient servers have a habit of sticking around for a while...
1458 *) Modify compression code so it frees up structures without using the
1459 ex_data callbacks. This works around a problem where some applications
1460 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1461 restarting) then use compression (e.g. SSL with compression) later.
1462 This results in significant per-connection memory leaks and
1463 has caused some security issues including CVE-2008-1678 and
1467 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1468 change when encrypting or decrypting.
1471 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1472 connect and renegotiate with servers which do not support RI.
1473 Until RI is more widely deployed this option is enabled by default.
1476 *) Add "missing" ssl ctrls to clear options and mode.
1479 *) If client attempts to renegotiate and doesn't support RI respond with
1480 a no_renegotiation alert as required by RFC5746. Some renegotiating
1481 TLS clients will continue a connection gracefully when they receive
1482 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1483 waiting for a server hello which it will never receive. Now we treat a
1484 received no_renegotiation alert as a fatal error. This is because
1485 applications requesting a renegotiation might well expect it to succeed
1486 and would have no code in place to handle the server denying it so the
1487 only safe thing to do is to terminate the connection.
1490 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1491 peer supports secure renegotiation and 0 otherwise. Print out peer
1492 renegotiation support in s_client/s_server.
1495 *) Replace the highly broken and deprecated SPKAC certification method with
1496 the updated NID creation version. This should correctly handle UTF8.
1499 *) Implement RFC5746. Re-enable renegotiation but require the extension
1500 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1501 turns out to be a bad idea. It has been replaced by
1502 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1503 SSL_CTX_set_options(). This is really not recommended unless you
1504 know what you are doing.
1505 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1507 *) Fixes to stateless session resumption handling. Use initial_ctx when
1508 issuing and attempting to decrypt tickets in case it has changed during
1509 servername handling. Use a non-zero length session ID when attempting
1510 stateless session resumption: this makes it possible to determine if
1511 a resumption has occurred immediately after receiving server hello
1512 (several places in OpenSSL subtly assume this) instead of later in
1516 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1517 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1518 fixes for a few places where the return code is not checked
1520 [Julia Lawall <julia@diku.dk>]
1522 *) Add --strict-warnings option to Configure script to include devteam
1523 warnings in other configurations.
1526 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1527 makes it possible to install openssl libraries in locations which
1528 have names other than "lib", for example "/usr/lib64" which some
1530 [Steve Henson, based on patch from Jeremy Utley]
1532 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1533 X690 8.9.12 and can produce some misleading textual output of OIDs.
1534 [Steve Henson, reported by Dan Kaminsky]
1536 *) Delete MD2 from algorithm tables. This follows the recommendation in
1537 several standards that it is not used in new applications due to
1538 several cryptographic weaknesses. For binary compatibility reasons
1539 the MD2 API is still compiled in by default.
1542 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1546 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1547 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1549 [Guenter <lists@gknw.net>]
1551 *) Fix the server certificate chain building code to use X509_verify_cert(),
1552 it used to have an ad-hoc builder which was unable to cope with anything
1553 other than a simple chain.
1554 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1556 *) Don't check self signed certificate signatures in X509_verify_cert()
1557 by default (a flag can override this): it just wastes time without
1558 adding any security. As a useful side effect self signed root CAs
1559 with non-FIPS digests are now usable in FIPS mode.
1562 *) In dtls1_process_out_of_seq_message() the check if the current message
1563 is already buffered was missing. For every new message was memory
1564 allocated, allowing an attacker to perform an denial of service attack
1565 with sending out of seq handshake messages until there is no memory
1566 left. Additionally every future messege was buffered, even if the
1567 sequence number made no sense and would be part of another handshake.
1568 So only messages with sequence numbers less than 10 in advance will be
1569 buffered. (CVE-2009-1378)
1570 [Robin Seggelmann, discovered by Daniel Mentz]
1572 *) Records are buffered if they arrive with a future epoch to be
1573 processed after finishing the corresponding handshake. There is
1574 currently no limitation to this buffer allowing an attacker to perform
1575 a DOS attack with sending records with future epochs until there is no
1576 memory left. This patch adds the pqueue_size() function to detemine
1577 the size of a buffer and limits the record buffer to 100 entries.
1579 [Robin Seggelmann, discovered by Daniel Mentz]
1581 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1582 parent structure is freed. (CVE-2009-1379)
1585 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1586 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1589 [Ilya O. <vrghost@gmail.com>]
1591 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1593 *) Disable renegotiation completely - this fixes a severe security
1594 problem (CVE-2009-3555) at the cost of breaking all
1595 renegotiation. Renegotiation can be re-enabled by setting
1596 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1597 run-time. This is really not recommended unless you know what
1601 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1603 *) Don't set val to NULL when freeing up structures, it is freed up by
1604 underlying code. If sizeof(void *) > sizeof(long) this can result in
1605 zeroing past the valid field. (CVE-2009-0789)
1606 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1608 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1609 checked correctly. This would allow some invalid signed attributes to
1610 appear to verify correctly. (CVE-2009-0591)
1611 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1613 *) Reject UniversalString and BMPString types with invalid lengths. This
1614 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1615 a legal length. (CVE-2009-0590)
1618 *) Set S/MIME signing as the default purpose rather than setting it
1619 unconditionally. This allows applications to override it at the store
1623 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1624 to handle some structures.
1627 *) Improve efficiency of mem_gets: don't search whole buffer each time
1629 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1631 *) New -hex option for openssl rand.
1634 *) Print out UTF8String and NumericString when parsing ASN1.
1637 *) Support NumericString type for name components.
1640 *) Allow CC in the environment to override the automatically chosen
1641 compiler. Note that nothing is done to ensure flags work with the
1645 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1647 *) Properly check EVP_VerifyFinal() and similar return values
1649 [Ben Laurie, Bodo Moeller, Google Security Team]
1651 *) Enable TLS extensions by default.
1654 *) Allow the CHIL engine to be loaded, whether the application is
1655 multithreaded or not. (This does not release the developer from the
1656 obligation to set up the dynamic locking callbacks.)
1657 [Sander Temme <sander@temme.net>]
1659 *) Use correct exit code if there is an error in dgst command.
1660 [Steve Henson; problem pointed out by Roland Dirlewanger]
1662 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1663 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1666 *) Add experimental JPAKE support, including demo authentication in
1667 s_client and s_server.
1670 *) Set the comparison function in v3_addr_canonize().
1671 [Rob Austein <sra@hactrn.net>]
1673 *) Add support for XMPP STARTTLS in s_client.
1674 [Philip Paeps <philip@freebsd.org>]
1676 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1677 to ensure that even with this option, only ciphersuites in the
1678 server's preference list will be accepted. (Note that the option
1679 applies only when resuming a session, so the earlier behavior was
1680 just about the algorithm choice for symmetric cryptography.)
1683 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1685 *) Fix NULL pointer dereference if a DTLS server received
1686 ChangeCipherSpec as first record (CVE-2009-1386).
1689 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1690 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1693 *) The fix in 0.9.8c that supposedly got rid of unsafe
1694 double-checked locking was incomplete for RSA blinding,
1695 addressing just one layer of what turns out to have been
1696 doubly unsafe triple-checked locking.
1698 So now fix this for real by retiring the MONT_HELPER macro
1699 in crypto/rsa/rsa_eay.c.
1701 [Bodo Moeller; problem pointed out by Marius Schilder]
1703 *) Various precautionary measures:
1705 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1707 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1708 (NB: This would require knowledge of the secret session ticket key
1709 to exploit, in which case you'd be SOL either way.)
1711 - Change bn_nist.c so that it will properly handle input BIGNUMs
1712 outside the expected range.
1714 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1717 [Neel Mehta, Bodo Moeller]
1719 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1720 the load fails. Useful for distros.
1721 [Ben Laurie and the FreeBSD team]
1723 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1726 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1729 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1731 This work was sponsored by Logica.
1734 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1735 keystores. Support for SSL/TLS client authentication too.
1736 Not compiled unless enable-capieng specified to Configure.
1738 This work was sponsored by Logica.
1741 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1742 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1743 attribute creation routines such as certifcate requests and PKCS#12
1747 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1749 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1750 handshake which could lead to a cilent crash as found using the
1751 Codenomicon TLS test suite (CVE-2008-1672)
1752 [Steve Henson, Mark Cox]
1754 *) Fix double free in TLS server name extensions which could lead to
1755 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1758 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1760 Clear the error queue to ensure that error entries left from
1761 older function calls do not interfere with the correct operation.
1762 [Lutz Jaenicke, Erik de Castro Lopo]
1764 *) Remove root CA certificates of commercial CAs:
1766 The OpenSSL project does not recommend any specific CA and does not
1767 have any policy with respect to including or excluding any CA.
1768 Therefore it does not make any sense to ship an arbitrary selection
1769 of root CA certificates with the OpenSSL software.
1772 *) RSA OAEP patches to fix two separate invalid memory reads.
1773 The first one involves inputs when 'lzero' is greater than
1774 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1775 before the beginning of from). The second one involves inputs where
1776 the 'db' section contains nothing but zeroes (there is a one-byte
1777 invalid read after the end of 'db').
1778 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1780 *) Partial backport from 0.9.9-dev:
1782 Introduce bn_mul_mont (dedicated Montgomery multiplication
1783 procedure) as a candidate for BIGNUM assembler implementation.
1784 While 0.9.9-dev uses assembler for various architectures, only
1785 x86_64 is available by default here in the 0.9.8 branch, and
1786 32-bit x86 is available through a compile-time setting.
1788 To try the 32-bit x86 assembler implementation, use Configure
1789 option "enable-montasm" (which exists only for this backport).
1791 As "enable-montasm" for 32-bit x86 disclaims code stability
1792 anyway, in this constellation we activate additional code
1793 backported from 0.9.9-dev for further performance improvements,
1794 namely BN_from_montgomery_word. (To enable this otherwise,
1795 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1797 [Andy Polyakov (backport partially by Bodo Moeller)]
1799 *) Add TLS session ticket callback. This allows an application to set
1800 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1801 values. This is useful for key rollover for example where several key
1802 sets may exist with different names.
1805 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1806 This was broken until now in 0.9.8 releases, such that the only way
1807 a registered ENGINE could be used (assuming it initialises
1808 successfully on the host) was to explicitly set it as the default
1809 for the relevant algorithms. This is in contradiction with 0.9.7
1810 behaviour and the documentation. With this fix, when an ENGINE is
1811 registered into a given algorithm's table of implementations, the
1812 'uptodate' flag is reset so that auto-discovery will be used next
1813 time a new context for that algorithm attempts to select an
1815 [Ian Lister (tweaked by Geoff Thorpe)]
1817 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1818 implemention in the following ways:
1820 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1823 Lack of BER streaming support means one pass streaming processing is
1824 only supported if data is detached: setting the streaming flag is
1825 ignored for embedded content.
1827 CMS support is disabled by default and must be explicitly enabled
1828 with the enable-cms configuration option.
1831 *) Update the GMP engine glue to do direct copies between BIGNUM and
1832 mpz_t when openssl and GMP use the same limb size. Otherwise the
1833 existing "conversion via a text string export" trick is still used.
1834 [Paul Sheer <paulsheer@gmail.com>]
1836 *) Zlib compression BIO. This is a filter BIO which compressed and
1837 uncompresses any data passed through it.
1840 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1841 RFC3394 compatible AES key wrapping.
1844 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1845 sets string data without copying. X509_ALGOR_set0() and
1846 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1847 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1848 from an X509_ATTRIBUTE structure optionally checking it occurs only
1849 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1853 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1854 to get the expected BN_FLG_CONSTTIME behavior.
1855 [Bodo Moeller (Google)]
1859 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1860 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1861 - added some more tests to do_tests.pl
1862 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1863 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1864 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1865 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1866 - various changes to netware.pl to enable gcc-cross builds on Win32
1868 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1869 - various changes to fix missing prototype warnings
1870 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1871 - added AES, WHIRLPOOL and CPUID assembler code to build files
1872 - added missing AES assembler make rules to mk1mf.pl
1873 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1874 [Guenter Knauf <eflash@gmx.net>]
1876 *) Implement certificate status request TLS extension defined in RFC3546.
1877 A client can set the appropriate parameters and receive the encoded
1878 OCSP response via a callback. A server can query the supplied parameters
1879 and set the encoded OCSP response in the callback. Add simplified examples
1880 to s_client and s_server.
1883 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1885 *) Fix various bugs:
1886 + Binary incompatibility of ssl_ctx_st structure
1887 + DTLS interoperation with non-compliant servers
1888 + Don't call get_session_cb() without proposed session
1889 + Fix ia64 assembler code
1890 [Andy Polyakov, Steve Henson]
1892 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1894 *) DTLS Handshake overhaul. There were longstanding issues with
1895 OpenSSL DTLS implementation, which were making it impossible for
1896 RFC 4347 compliant client to communicate with OpenSSL server.
1897 Unfortunately just fixing these incompatibilities would "cut off"
1898 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1899 server keeps tolerating non RFC compliant syntax. The opposite is
1900 not true, 0.9.8f client can not communicate with earlier server.
1901 This update even addresses CVE-2007-4995.
1904 *) Changes to avoid need for function casts in OpenSSL: some compilers
1905 (gcc 4.2 and later) reject their use.
1906 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1909 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1910 RFC4507bis. The encrypted ticket format is an encrypted encoded
1911 SSL_SESSION structure, that way new session features are automatically
1914 If a client application caches session in an SSL_SESSION structure
1915 support is transparent because tickets are now stored in the encoded
1918 The SSL_CTX structure automatically generates keys for ticket
1919 protection in servers so again support should be possible
1920 with no application modification.
1922 If a client or server wishes to disable RFC4507 support then the option
1923 SSL_OP_NO_TICKET can be set.
1925 Add a TLS extension debugging callback to allow the contents of any client
1926 or server extensions to be examined.
1928 This work was sponsored by Google.
1931 *) Add initial support for TLS extensions, specifically for the server_name
1932 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1933 have new members for a host name. The SSL data structure has an
1934 additional member SSL_CTX *initial_ctx so that new sessions can be
1935 stored in that context to allow for session resumption, even after the
1936 SSL has been switched to a new SSL_CTX in reaction to a client's
1937 server_name extension.
1939 New functions (subject to change):
1941 SSL_get_servername()
1942 SSL_get_servername_type()
1945 New CTRL codes and macros (subject to change):
1947 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1948 - SSL_CTX_set_tlsext_servername_callback()
1949 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1950 - SSL_CTX_set_tlsext_servername_arg()
1951 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1953 openssl s_client has a new '-servername ...' option.
1955 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1956 '-key2 ...', '-servername_fatal' (subject to change). This allows
1957 testing the HostName extension for a specific single host name ('-cert'
1958 and '-key' remain fallbacks for handshakes without HostName
1959 negotiation). If the unrecogninzed_name alert has to be sent, this by
1960 default is a warning; it becomes fatal with the '-servername_fatal'
1963 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1965 *) Add AES and SSE2 assembly language support to VC++ build.
1968 *) Mitigate attack on final subtraction in Montgomery reduction.
1971 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1972 (which previously caused an internal error).
1975 *) Squeeze another 10% out of IGE mode when in != out.
1978 *) AES IGE mode speedup.
1979 [Dean Gaudet (Google)]
1981 *) Add the Korean symmetric 128-bit cipher SEED (see
1982 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1983 add SEED ciphersuites from RFC 4162:
1985 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1986 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1987 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1988 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1990 To minimize changes between patchlevels in the OpenSSL 0.9.8
1991 series, SEED remains excluded from compilation unless OpenSSL
1992 is configured with 'enable-seed'.
1993 [KISA, Bodo Moeller]
1995 *) Mitigate branch prediction attacks, which can be practical if a
1996 single processor is shared, allowing a spy process to extract
1997 information. For detailed background information, see
1998 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1999 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2000 and Necessary Software Countermeasures"). The core of the change
2001 are new versions BN_div_no_branch() and
2002 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2003 respectively, which are slower, but avoid the security-relevant
2004 conditional branches. These are automatically called by BN_div()
2005 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2006 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2007 remove a conditional branch.
2009 BN_FLG_CONSTTIME is the new name for the previous
2010 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2011 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2012 in the exponent causes BN_mod_exp_mont() to use the alternative
2013 implementation in BN_mod_exp_mont_consttime().) The old name
2014 remains as a deprecated alias.
2016 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2017 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2018 constant-time implementations for more than just exponentiation.
2019 Here too the old name is kept as a deprecated alias.
2021 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2022 the BN_BLINDING structure gets an independent copy of the
2023 modulus. This means that the previous "BIGNUM *m" argument to
2024 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2025 essentially becomes "const BIGNUM *m", although we can't actually
2026 change this in the header file before 0.9.9. It allows
2027 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2028 enable BN_FLG_CONSTTIME.
2030 [Matthew D Wood (Intel Corp)]
2032 *) In the SSL/TLS server implementation, be strict about session ID
2033 context matching (which matters if an application uses a single
2034 external cache for different purposes). Previously,
2035 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2036 set. This did ensure strict client verification, but meant that,
2037 with applications using a single external cache for quite
2038 different requirements, clients could circumvent ciphersuite
2039 restrictions for a given session ID context by starting a session
2040 in a different context.
2043 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2044 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2045 authentication-only ciphersuites.
2048 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2049 not complete and could lead to a possible single byte overflow
2050 (CVE-2007-5135) [Ben Laurie]
2052 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2054 *) Since AES128 and AES256 (and similarly Camellia128 and
2055 Camellia256) share a single mask bit in the logic of
2056 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2057 kludge to work properly if AES128 is available and AES256 isn't
2058 (or if Camellia128 is available and Camellia256 isn't).
2061 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2062 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2063 When a point or a seed is encoded in a BIT STRING, we need to
2064 prevent the removal of trailing zero bits to get the proper DER
2065 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2066 of a NamedBitList, for which trailing 0 bits need to be removed.)
2069 *) Have SSL/TLS server implementation tolerate "mismatched" record
2070 protocol version while receiving ClientHello even if the
2071 ClientHello is fragmented. (The server can't insist on the
2072 particular protocol version it has chosen before the ServerHello
2073 message has informed the client about his choice.)
2076 *) Add RFC 3779 support.
2077 [Rob Austein for ARIN, Ben Laurie]
2079 *) Load error codes if they are not already present instead of using a
2080 static variable. This allows them to be cleanly unloaded and reloaded.
2081 Improve header file function name parsing.
2084 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2085 or CAPABILITY handshake as required by RFCs.
2088 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2090 *) Introduce limits to prevent malicious keys being able to
2091 cause a denial of service. (CVE-2006-2940)
2092 [Steve Henson, Bodo Moeller]
2094 *) Fix ASN.1 parsing of certain invalid structures that can result
2095 in a denial of service. (CVE-2006-2937) [Steve Henson]
2097 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2098 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2100 *) Fix SSL client code which could crash if connecting to a
2101 malicious SSLv2 server. (CVE-2006-4343)
2102 [Tavis Ormandy and Will Drewry, Google Security Team]
2104 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2105 match only those. Before that, "AES256-SHA" would be interpreted
2106 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2107 the same strength classification in 0.9.7h) as we currently only
2108 have a single AES bit in the ciphersuite description bitmap.
2109 That change, however, also applied to ciphersuite strings such as
2110 "RC4-MD5" that intentionally matched multiple ciphersuites --
2111 namely, SSL 2.0 ciphersuites in addition to the more common ones
2112 from SSL 3.0/TLS 1.0.
2114 So we change the selection algorithm again: Naming an explicit
2115 ciphersuite selects this one ciphersuite, and any other similar
2116 ciphersuite (same bitmap) from *other* protocol versions.
2117 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2118 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2120 Since SSL 2.0 does not have any ciphersuites for which the
2121 128/256 bit distinction would be relevant, this works for now.
2122 The proper fix will be to use different bits for AES128 and
2123 AES256, which would have avoided the problems from the beginning;
2124 however, bits are scarce, so we can only do this in a new release
2125 (not just a patchlevel) when we can change the SSL_CIPHER
2126 definition to split the single 'unsigned long mask' bitmap into
2127 multiple values to extend the available space.
2131 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2133 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2134 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2136 *) Add AES IGE and biIGE modes.
2139 *) Change the Unix randomness entropy gathering to use poll() when
2140 possible instead of select(), since the latter has some
2141 undesirable limitations.
2142 [Darryl Miles via Richard Levitte and Bodo Moeller]
2144 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2145 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2146 cannot be implicitly activated as part of, e.g., the "AES" alias.
2147 However, please upgrade to OpenSSL 0.9.9[-dev] for
2148 non-experimental use of the ECC ciphersuites to get TLS extension
2149 support, which is required for curve and point format negotiation
2150 to avoid potential handshake problems.
2153 *) Disable rogue ciphersuites:
2155 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2156 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2157 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2159 The latter two were purportedly from
2160 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2163 Also deactivate the remaining ciphersuites from
2164 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2165 unofficial, and the ID has long expired.
2168 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2169 dual-core machines) and other potential thread-safety issues.
2172 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2173 versions), which is now available for royalty-free use
2174 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2175 Also, add Camellia TLS ciphersuites from RFC 4132.
2177 To minimize changes between patchlevels in the OpenSSL 0.9.8
2178 series, Camellia remains excluded from compilation unless OpenSSL
2179 is configured with 'enable-camellia'.
2182 *) Disable the padding bug check when compression is in use. The padding
2183 bug check assumes the first packet is of even length, this is not
2184 necessarily true if compresssion is enabled and can result in false
2185 positives causing handshake failure. The actual bug test is ancient
2186 code so it is hoped that implementations will either have fixed it by
2187 now or any which still have the bug do not support compression.
2190 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2192 *) When applying a cipher rule check to see if string match is an explicit
2193 cipher suite and only match that one cipher suite if it is.
2196 *) Link in manifests for VC++ if needed.
2197 [Austin Ziegler <halostatue@gmail.com>]
2199 *) Update support for ECC-based TLS ciphersuites according to
2200 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2201 TLS extensions, which are supported starting with the 0.9.9
2202 branch, not in the OpenSSL 0.9.8 branch).
2205 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2206 opaque EVP_CIPHER_CTX handling.
2209 *) Fixes and enhancements to zlib compression code. We now only use
2210 "zlib1.dll" and use the default __cdecl calling convention on Win32
2211 to conform with the standards mentioned here:
2212 http://www.zlib.net/DLL_FAQ.txt
2213 Static zlib linking now works on Windows and the new --with-zlib-include
2214 --with-zlib-lib options to Configure can be used to supply the location
2215 of the headers and library. Gracefully handle case where zlib library
2219 *) Several fixes and enhancements to the OID generation code. The old code
2220 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2221 handle numbers larger than ULONG_MAX, truncated printing and had a
2222 non standard OBJ_obj2txt() behaviour.
2225 *) Add support for building of engines under engine/ as shared libraries
2226 under VC++ build system.
2229 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2230 Hopefully, we will not see any false combination of paths any more.
2233 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2235 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2236 (part of SSL_OP_ALL). This option used to disable the
2237 countermeasure against man-in-the-middle protocol-version
2238 rollback in the SSL 2.0 server implementation, which is a bad
2239 idea. (CVE-2005-2969)
2241 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2242 for Information Security, National Institute of Advanced Industrial
2243 Science and Technology [AIST], Japan)]
2245 *) Add two function to clear and return the verify parameter flags.
2248 *) Keep cipherlists sorted in the source instead of sorting them at
2249 runtime, thus removing the need for a lock.
2252 *) Avoid some small subgroup attacks in Diffie-Hellman.
2253 [Nick Mathewson and Ben Laurie]
2255 *) Add functions for well-known primes.
2258 *) Extended Windows CE support.
2259 [Satoshi Nakamura and Andy Polyakov]
2261 *) Initialize SSL_METHOD structures at compile time instead of during
2262 runtime, thus removing the need for a lock.
2265 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2266 attempting to decrypt each encrypted key in turn. Add support to
2270 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2272 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2275 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2278 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2279 key into the same file any more.
2282 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2285 *) Add -utf8 command line and config file option to 'ca'.
2286 [Stefan <stf@udoma.org]
2288 *) Removed the macro des_crypt(), as it seems to conflict with some
2289 libraries. Use DES_crypt().
2292 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2293 involves renaming the source and generated shared-libs for
2294 both. The engines will accept the corrected or legacy ids
2295 ('ncipher' and '4758_cca' respectively) when binding. NB,
2296 this only applies when building 'shared'.
2297 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2299 *) Add attribute functions to EVP_PKEY structure. Modify
2300 PKCS12_create() to recognize a CSP name attribute and
2301 use it. Make -CSP option work again in pkcs12 utility.
2304 *) Add new functionality to the bn blinding code:
2305 - automatic re-creation of the BN_BLINDING parameters after
2306 a fixed number of uses (currently 32)
2307 - add new function for parameter creation
2308 - introduce flags to control the update behaviour of the
2309 BN_BLINDING parameters
2310 - hide BN_BLINDING structure
2311 Add a second BN_BLINDING slot to the RSA structure to improve
2312 performance when a single RSA object is shared among several
2316 *) Add support for DTLS.
2317 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2319 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2320 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2323 *) Remove buggy and incompletet DH cert support from
2324 ssl/ssl_rsa.c and ssl/s3_both.c
2327 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2328 the apps/openssl applications.
2331 *) Compile clean with "-Wall -Wmissing-prototypes
2332 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2333 DEBUG_SAFESTACK must also be set.
2336 *) Change ./Configure so that certain algorithms can be disabled by default.
2337 The new counterpiece to "no-xxx" is "enable-xxx".
2339 The patented RC5 and MDC2 algorithms will now be disabled unless
2340 "enable-rc5" and "enable-mdc2", respectively, are specified.
2342 (IDEA remains enabled despite being patented. This is because IDEA
2343 is frequently required for interoperability, and there is no license
2344 fee for non-commercial use. As before, "no-idea" can be used to
2345 avoid this algorithm.)
2349 *) Add processing of proxy certificates (see RFC 3820). This work was
2350 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2351 EGEE (Enabling Grids for E-science in Europe).
2354 *) RC4 performance overhaul on modern architectures/implementations, such
2355 as Intel P4, IA-64 and AMD64.
2358 *) New utility extract-section.pl. This can be used specify an alternative
2359 section number in a pod file instead of having to treat each file as
2360 a separate case in Makefile. This can be done by adding two lines to the
2363 =for comment openssl_section:XXX
2365 The blank line is mandatory.
2369 *) New arguments -certform, -keyform and -pass for s_client and s_server
2370 to allow alternative format key and certificate files and passphrase
2374 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2375 update associated structures and add various utility functions.
2377 Add new policy related verify parameters, include policy checking in
2378 standard verify code. Enhance 'smime' application with extra parameters
2379 to support policy checking and print out.
2382 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2383 Nehemiah processors. These extensions support AES encryption in hardware
2384 as well as RNG (though RNG support is currently disabled).
2385 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2387 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2390 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2391 [Andy Polyakov and a number of other people]
2393 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2394 implementation contributed by IBM.
2395 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2397 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2398 exponent rather than 'unsigned long'. There is a corresponding change to
2399 the new 'rsa_keygen' element of the RSA_METHOD structure.
2400 [Jelte Jansen, Geoff Thorpe]
2402 *) Functionality for creating the initial serial number file is now
2403 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2405 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2406 number file to 1, which is bound to cause problems. To avoid
2407 the problems while respecting compatibility between different 0.9.7
2408 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2409 CA.pl for serial number initialization. With the new release 0.9.8,
2410 we can fix the problem directly in the 'ca' utility.)
2413 *) Reduced header interdepencies by declaring more opaque objects in
2414 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2415 give fewer recursive includes, which could break lazy source code - so
2416 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2417 developers should define this symbol when building and using openssl to
2418 ensure they track the recommended behaviour, interfaces, [etc], but
2419 backwards-compatible behaviour prevails when this isn't defined.
2422 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2425 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2426 This will generate a random key of the appropriate length based on the
2427 cipher context. The EVP_CIPHER can provide its own random key generation
2428 routine to support keys of a specific form. This is used in the des and
2429 3des routines to generate a key of the correct parity. Update S/MIME
2430 code to use new functions and hence generate correct parity DES keys.
2431 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2432 valid (weak or incorrect parity).
2435 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2436 as looking them up. This is useful when the verified structure may contain
2437 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2438 present unless the new PKCS7_NO_CRL flag is asserted.
2441 *) Extend ASN1 oid configuration module. It now additionally accepts the
2444 shortName = some long name, 1.2.3.4
2447 *) Reimplemented the BN_CTX implementation. There is now no more static
2448 limitation on the number of variables it can handle nor the depth of the
2449 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2450 information can now expand as required, and rather than having a single
2451 static array of bignums, BN_CTX now uses a linked-list of such arrays
2452 allowing it to expand on demand whilst maintaining the usefulness of
2453 BN_CTX's "bundling".
2456 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2457 to allow all RSA operations to function using a single BN_CTX.
2460 *) Preliminary support for certificate policy evaluation and checking. This
2461 is initially intended to pass the tests outlined in "Conformance Testing
2462 of Relying Party Client Certificate Path Processing Logic" v1.07.
2465 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2466 remained unused and not that useful. A variety of other little bignum
2467 tweaks and fixes have also been made continuing on from the audit (see
2471 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2472 associated ASN1, EVP and SSL functions and old ASN1 macros.
2475 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2476 and this should never fail. So the return value from the use of
2477 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2478 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2481 *) BN_CTX_get() should return zero-valued bignums, providing the same
2482 initialised value as BN_new().
2483 [Geoff Thorpe, suggested by Ulf Möller]
2485 *) Support for inhibitAnyPolicy certificate extension.
2488 *) An audit of the BIGNUM code is underway, for which debugging code is
2489 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2490 is considered valid when processing BIGNUMs, and causes execution to
2491 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2492 further steps are taken to deliberately pollute unused data in BIGNUM
2493 structures to try and expose faulty code further on. For now, openssl will
2494 (in its default mode of operation) continue to tolerate the inconsistent
2495 forms that it has tolerated in the past, but authors and packagers should
2496 consider trying openssl and their own applications when compiled with
2497 these debugging symbols defined. It will help highlight potential bugs in
2498 their own code, and will improve the test coverage for OpenSSL itself. At
2499 some point, these tighter rules will become openssl's default to improve
2500 maintainability, though the assert()s and other overheads will remain only
2501 in debugging configurations. See bn.h for more details.
2502 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2504 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2505 that can only be obtained through BN_CTX_new() (which implicitly
2506 initialises it). The presence of this function only made it possible
2507 to overwrite an existing structure (and cause memory leaks).
2510 *) Because of the callback-based approach for implementing LHASH as a
2511 template type, lh_insert() adds opaque objects to hash-tables and
2512 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2513 to clean up those corresponding objects before destroying the hash table
2514 (and losing the object pointers). So some over-zealous constifications in
2515 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2516 objects as "const" and the lh_doall[_arg] callback wrappers are not
2517 prototyped to have "const" restrictions on the object pointers they are
2518 given (and so aren't required to cast them away any more).
2521 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2522 (speed) prefers to use its own implementation. The two implementations
2523 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2524 its object type properly exposed (MS_TM) instead of casting to/from "char
2525 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2526 aren't necessarily the greatest nomenclatures - but this is what was used
2527 internally to the implementation so I've used that for now.
2530 *) Ensure that deprecated functions do not get compiled when
2531 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2532 the self-tests were still using deprecated key-generation functions so
2533 these have been updated also.
2536 *) Reorganise PKCS#7 code to separate the digest location functionality
2537 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2538 New function PKCS7_set_digest() to set the digest type for PKCS#7
2539 digestedData type. Add additional code to correctly generate the
2540 digestedData type and add support for this type in PKCS7 initialization
2544 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2545 structure of type "other".
2548 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2549 sure the loop does correctly stop and breaking ("division by zero")
2550 modulus operations are not performed. The (pre-generated) prime
2551 table crypto/bn/bn_prime.h was already correct, but it could not be
2552 re-generated on some platforms because of the "division by zero"
2553 situation in the script.
2554 [Ralf S. Engelschall]
2556 *) Update support for ECC-based TLS ciphersuites according to
2557 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2558 SHA-1 now is only used for "small" curves (where the
2559 representation of a field element takes up to 24 bytes); for
2560 larger curves, the field element resulting from ECDH is directly
2561 used as premaster secret.
2562 [Douglas Stebila (Sun Microsystems Laboratories)]
2564 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2565 curve secp160r1 to the tests.
2566 [Douglas Stebila (Sun Microsystems Laboratories)]
2568 *) Add the possibility to load symbols globally with DSO.
2569 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2571 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2572 control of the error stack.
2575 *) Add support for STORE in ENGINE.
2578 *) Add the STORE type. The intention is to provide a common interface
2579 to certificate and key stores, be they simple file-based stores, or
2580 HSM-type store, or LDAP stores, or...
2581 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2584 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2585 pass a list of arguments to any function as well as provide a way
2586 for a function to pass data back to the caller.
2589 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2590 works like BUF_strdup() but can be used to duplicate a portion of
2591 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2595 *) Add the function sk_find_ex() which works like sk_find(), but will
2596 return an index to an element even if an exact match couldn't be
2597 found. The index is guaranteed to point at the element where the
2598 searched-for key would be inserted to preserve sorting order.
2601 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2602 takes an extra flags argument for optional functionality. Currently,
2603 the following flags are defined:
2605 OBJ_BSEARCH_VALUE_ON_NOMATCH
2606 This one gets OBJ_bsearch_ex() to return a pointer to the first
2607 element where the comparing function returns a negative or zero
2610 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2611 This one gets OBJ_bsearch_ex() to return a pointer to the first
2612 element where the comparing function returns zero. This is useful
2613 if there are more than one element where the comparing function
2617 *) Make it possible to create self-signed certificates with 'openssl ca'
2618 in such a way that the self-signed certificate becomes part of the
2619 CA database and uses the same mechanisms for serial number generation
2620 as all other certificate signing. The new flag '-selfsign' enables
2621 this functionality. Adapt CA.sh and CA.pl.in.
2624 *) Add functionality to check the public key of a certificate request
2625 against a given private. This is useful to check that a certificate
2626 request can be signed by that key (self-signing).
2629 *) Make it possible to have multiple active certificates with the same
2630 subject in the CA index file. This is done only if the keyword
2631 'unique_subject' is set to 'no' in the main CA section (default
2632 if 'CA_default') of the configuration file. The value is saved
2633 with the database itself in a separate index attribute file,
2634 named like the index file with '.attr' appended to the name.
2637 *) Generate muti valued AVAs using '+' notation in config files for
2641 *) Support for nameConstraints certificate extension.
2644 *) Support for policyConstraints certificate extension.
2647 *) Support for policyMappings certificate extension.
2650 *) Make sure the default DSA_METHOD implementation only uses its
2651 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2652 and change its own handlers to be NULL so as to remove unnecessary
2653 indirection. This lets alternative implementations fallback to the
2654 default implementation more easily.
2657 *) Support for directoryName in GeneralName related extensions
2661 *) Make it possible to link applications using Makefile.shared.
2662 Make that possible even when linking against static libraries!
2665 *) Support for single pass processing for S/MIME signing. This now
2666 means that S/MIME signing can be done from a pipe, in addition
2667 cleartext signing (multipart/signed type) is effectively streaming
2668 and the signed data does not need to be all held in memory.
2670 This is done with a new flag PKCS7_STREAM. When this flag is set
2671 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2672 is done after the data is output (and digests calculated) in
2673 SMIME_write_PKCS7().
2676 *) Add full support for -rpath/-R, both in shared libraries and
2677 applications, at least on the platforms where it's known how
2681 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2682 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2683 will now compute a table of multiples of the generator that
2684 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2685 faster (notably in the case of a single point multiplication,
2686 scalar * generator).
2687 [Nils Larsch, Bodo Moeller]
2689 *) IPv6 support for certificate extensions. The various extensions
2690 which use the IP:a.b.c.d can now take IPv6 addresses using the
2691 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2695 *) Added an ENGINE that implements RSA by performing private key
2696 exponentiations with the GMP library. The conversions to and from
2697 GMP's mpz_t format aren't optimised nor are any montgomery forms
2698 cached, and on x86 it appears OpenSSL's own performance has caught up.
2699 However there are likely to be other architectures where GMP could
2700 provide a boost. This ENGINE is not built in by default, but it can be
2701 specified at Configure time and should be accompanied by the necessary
2702 linker additions, eg;
2703 ./config -DOPENSSL_USE_GMP -lgmp
2706 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2707 testing availability of engines with "-t" - the old behaviour is
2708 produced by increasing the feature's verbosity with "-tt".
2711 *) ECDSA routines: under certain error conditions uninitialized BN objects
2712 could be freed. Solution: make sure initialization is performed early
2713 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2717 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2718 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2719 software implementations. For DSA and DH, parameter generation can
2720 also be overriden by providing the appropriate method callbacks.
2723 *) Change the "progress" mechanism used in key-generation and
2724 primality testing to functions that take a new BN_GENCB pointer in
2725 place of callback/argument pairs. The new API functions have "_ex"
2726 postfixes and the older functions are reimplemented as wrappers for
2727 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2728 declarations of the old functions to help (graceful) attempts to
2729 migrate to the new functions. Also, the new key-generation API
2730 functions operate on a caller-supplied key-structure and return
2731 success/failure rather than returning a key or NULL - this is to
2732 help make "keygen" another member function of RSA_METHOD etc.
2734 Example for using the new callback interface:
2736 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2740 BN_GENCB_set(&my_cb, my_callback, my_arg);
2742 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2743 /* For the meaning of a, b in calls to my_callback(), see the
2744 * documentation of the function that calls the callback.
2745 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2746 * my_callback should return 1 if it wants BN_is_prime_ex()
2747 * to continue, or 0 to stop.
2752 *) Change the ZLIB compression method to be stateful, and make it
2753 available to TLS with the number defined in
2754 draft-ietf-tls-compression-04.txt.
2757 *) Add the ASN.1 structures and functions for CertificatePair, which
2758 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2760 CertificatePair ::= SEQUENCE {
2761 forward [0] Certificate OPTIONAL,
2762 reverse [1] Certificate OPTIONAL,
2763 -- at least one of the pair shall be present -- }
2765 Also implement the PEM functions to read and write certificate
2766 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2768 This needed to be defined, mostly for the sake of the LDAP
2769 attribute crossCertificatePair, but may prove useful elsewhere as
2773 *) Make it possible to inhibit symlinking of shared libraries in
2774 Makefile.shared, for Cygwin's sake.
2777 *) Extend the BIGNUM API by creating a function
2778 void BN_set_negative(BIGNUM *a, int neg);
2779 and a macro that behave like
2780 int BN_is_negative(const BIGNUM *a);
2782 to avoid the need to access 'a->neg' directly in applications.
2785 *) Implement fast modular reduction for pseudo-Mersenne primes
2786 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2787 EC_GROUP_new_curve_GFp() will now automatically use this
2789 [Nils Larsch <nla@trustcenter.de>]
2791 *) Add new lock type (CRYPTO_LOCK_BN).
2794 *) Change the ENGINE framework to automatically load engines
2795 dynamically from specific directories unless they could be
2796 found to already be built in or loaded. Move all the
2797 current engines except for the cryptodev one to a new
2799 The engines in engines/ are built as shared libraries if
2800 the "shared" options was given to ./Configure or ./config.
2801 Otherwise, they are inserted in libcrypto.a.
2802 /usr/local/ssl/engines is the default directory for dynamic
2803 engines, but that can be overriden at configure time through
2804 the usual use of --prefix and/or --openssldir, and at run
2805 time with the environment variable OPENSSL_ENGINES.
2806 [Geoff Thorpe and Richard Levitte]
2808 *) Add Makefile.shared, a helper makefile to build shared
2809 libraries. Addapt Makefile.org.
2812 *) Add version info to Win32 DLLs.
2813 [Peter 'Luna' Runestig" <peter@runestig.com>]
2815 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2816 can be added using this API to created arbitrary PKCS#12
2817 files while avoiding the low level API.
2819 New options to PKCS12_create(), key or cert can be NULL and
2820 will then be omitted from the output file. The encryption
2821 algorithm NIDs can be set to -1 for no encryption, the mac
2822 iteration count can be set to 0 to omit the mac.
2824 Enhance pkcs12 utility by making the -nokeys and -nocerts
2825 options work when creating a PKCS#12 file. New option -nomac
2826 to omit the mac, NONE can be set for an encryption algorithm.
2827 New code is modified to use the enhanced PKCS12_create()
2828 instead of the low level API.
2831 *) Extend ASN1 encoder to support indefinite length constructed
2832 encoding. This can output sequences tags and octet strings in
2833 this form. Modify pk7_asn1.c to support indefinite length
2834 encoding. This is experimental and needs additional code to
2835 be useful, such as an ASN1 bio and some enhanced streaming
2838 Extend template encode functionality so that tagging is passed
2839 down to the template encoder.
2842 *) Let 'openssl req' fail if an argument to '-newkey' is not
2843 recognized instead of using RSA as a default.
2846 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2847 As these are not official, they are not included in "ALL";
2848 the "ECCdraft" ciphersuite group alias can be used to select them.
2849 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2851 *) Add ECDH engine support.
2852 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2854 *) Add ECDH in new directory crypto/ecdh/.
2855 [Douglas Stebila (Sun Microsystems Laboratories)]
2857 *) Let BN_rand_range() abort with an error after 100 iterations
2858 without success (which indicates a broken PRNG).
2861 *) Change BN_mod_sqrt() so that it verifies that the input value
2862 is really the square of the return value. (Previously,
2863 BN_mod_sqrt would show GIGO behaviour.)
2866 *) Add named elliptic curves over binary fields from X9.62, SECG,
2867 and WAP/WTLS; add OIDs that were still missing.
2869 [Sheueling Chang Shantz and Douglas Stebila
2870 (Sun Microsystems Laboratories)]
2872 *) Extend the EC library for elliptic curves over binary fields
2873 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2876 EC_GF2m_simple_method
2880 EC_GROUP_new_curve_GF2m
2881 EC_GROUP_set_curve_GF2m
2882 EC_GROUP_get_curve_GF2m
2883 EC_POINT_set_affine_coordinates_GF2m
2884 EC_POINT_get_affine_coordinates_GF2m
2885 EC_POINT_set_compressed_coordinates_GF2m
2887 Point compression for binary fields is disabled by default for
2888 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2891 As binary polynomials are represented as BIGNUMs, various members
2892 of the EC_GROUP and EC_POINT data structures can be shared
2893 between the implementations for prime fields and binary fields;
2894 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2895 are essentially identical to their ..._GFp counterparts.
2896 (For simplicity, the '..._GFp' prefix has been dropped from
2897 various internal method names.)
2899 An internal 'field_div' method (similar to 'field_mul' and
2900 'field_sqr') has been added; this is used only for binary fields.
2902 [Sheueling Chang Shantz and Douglas Stebila
2903 (Sun Microsystems Laboratories)]
2905 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2906 through methods ('mul', 'precompute_mult').
2908 The generic implementations (now internally called 'ec_wNAF_mul'
2909 and 'ec_wNAF_precomputed_mult') remain the default if these
2910 methods are undefined.
2912 [Sheueling Chang Shantz and Douglas Stebila
2913 (Sun Microsystems Laboratories)]
2915 *) New function EC_GROUP_get_degree, which is defined through
2916 EC_METHOD. For curves over prime fields, this returns the bit
2917 length of the modulus.
2919 [Sheueling Chang Shantz and Douglas Stebila
2920 (Sun Microsystems Laboratories)]
2922 *) New functions EC_GROUP_dup, EC_POINT_dup.
2923 (These simply call ..._new and ..._copy).
2925 [Sheueling Chang Shantz and Douglas Stebila
2926 (Sun Microsystems Laboratories)]
2928 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2929 Polynomials are represented as BIGNUMs (where the sign bit is not
2930 used) in the following functions [macros]:
2933 BN_GF2m_sub [= BN_GF2m_add]
2934 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2935 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2936 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2938 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2939 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2940 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2941 BN_GF2m_cmp [= BN_ucmp]
2943 (Note that only the 'mod' functions are actually for fields GF(2^m).
2944 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2946 For some functions, an the irreducible polynomial defining a
2947 field can be given as an 'unsigned int[]' with strictly
2948 decreasing elements giving the indices of those bits that are set;
2949 i.e., p[] represents the polynomial
2950 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2952 p[0] > p[1] > ... > p[k] = 0.
2953 This applies to the following functions:
2958 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2959 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2961 BN_GF2m_mod_sqrt_arr
2962 BN_GF2m_mod_solve_quad_arr
2966 Conversion can be performed by the following functions:
2971 bntest.c has additional tests for binary polynomial arithmetic.
2973 Two implementations for BN_GF2m_mod_div() are available.
2974 The default algorithm simply uses BN_GF2m_mod_inv() and
2975 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2976 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2977 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2979 [Sheueling Chang Shantz and Douglas Stebila
2980 (Sun Microsystems Laboratories)]
2982 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2983 functionality is disabled at compile-time.
2984 [Douglas Stebila <douglas.stebila@sun.com>]
2986 *) Change default behaviour of 'openssl asn1parse' so that more
2987 information is visible when viewing, e.g., a certificate:
2989 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2990 mode the content of non-printable OCTET STRINGs is output in a
2991 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2992 avoid the appearance of a printable string.
2993 [Nils Larsch <nla@trustcenter.de>]
2995 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2997 EC_GROUP_set_asn1_flag()
2998 EC_GROUP_get_asn1_flag()
2999 EC_GROUP_set_point_conversion_form()
3000 EC_GROUP_get_point_conversion_form()
3001 These control ASN1 encoding details:
3002 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3003 has been set to OPENSSL_EC_NAMED_CURVE.
3004 - Points are encoded in uncompressed form by default; options for
3005 asn1_for are as for point2oct, namely
3006 POINT_CONVERSION_COMPRESSED
3007 POINT_CONVERSION_UNCOMPRESSED
3008 POINT_CONVERSION_HYBRID
3010 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3013 EC_GROUP_get0_seed()
3014 EC_GROUP_get_seed_len()
3015 This is used only for ASN1 purposes (so far).
3016 [Nils Larsch <nla@trustcenter.de>]
3018 *) Add 'field_type' member to EC_METHOD, which holds the NID
3019 of the appropriate field type OID. The new function
3020 EC_METHOD_get_field_type() returns this value.
3021 [Nils Larsch <nla@trustcenter.de>]
3026 EC_POINT_point2hex()
3027 EC_POINT_hex2point()
3028 providing useful interfaces to EC_POINT_point2oct() and
3029 EC_POINT_oct2point().
3030 [Nils Larsch <nla@trustcenter.de>]
3032 *) Change internals of the EC library so that the functions
3033 EC_GROUP_set_generator()
3034 EC_GROUP_get_generator()
3035 EC_GROUP_get_order()
3036 EC_GROUP_get_cofactor()
3037 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3038 to methods, which would lead to unnecessary code duplication when
3039 adding different types of curves.
3040 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3042 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3043 arithmetic, and such that modified wNAFs are generated
3044 (which avoid length expansion in many cases).
3047 *) Add a function EC_GROUP_check_discriminant() (defined via
3048 EC_METHOD) that verifies that the curve discriminant is non-zero.
3050 Add a function EC_GROUP_check() that makes some sanity tests
3051 on a EC_GROUP, its generator and order. This includes
3052 EC_GROUP_check_discriminant().
3053 [Nils Larsch <nla@trustcenter.de>]
3055 *) Add ECDSA in new directory crypto/ecdsa/.
3057 Add applications 'openssl ecparam' and 'openssl ecdsa'
3058 (these are based on 'openssl dsaparam' and 'openssl dsa').
3060 ECDSA support is also included in various other files across the
3061 library. Most notably,
3062 - 'openssl req' now has a '-newkey ecdsa:file' option;
3063 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3064 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3065 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3066 them suitable for ECDSA where domain parameters must be
3067 extracted before the specific public key;
3068 - ECDSA engine support has been added.
3069 [Nils Larsch <nla@trustcenter.de>]
3071 *) Include some named elliptic curves, and add OIDs from X9.62,
3072 SECG, and WAP/WTLS. Each curve can be obtained from the new
3074 EC_GROUP_new_by_curve_name(),
3075 and the list of available named curves can be obtained with
3076 EC_get_builtin_curves().
3077 Also add a 'curve_name' member to EC_GROUP objects, which can be
3079 EC_GROUP_set_curve_name()
3080 EC_GROUP_get_curve_name()
3081 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3083 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3084 was actually never needed) and in BN_mul(). The removal in BN_mul()
3085 required a small change in bn_mul_part_recursive() and the addition
3086 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3087 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3088 bn_sub_words() and bn_add_words() except they take arrays with
3092 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3094 *) Cleanse PEM buffers before freeing them since they may contain
3096 [Benjamin Bennett <ben@psc.edu>]
3098 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3099 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3100 authentication-only ciphersuites.
3103 *) Since AES128 and AES256 share a single mask bit in the logic of
3104 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3105 kludge to work properly if AES128 is available and AES256 isn't.
3108 *) Expand security boundary to match 1.1.1 module.
3111 *) Remove redundant features: hash file source, editing of test vectors
3112 modify fipsld to use external fips_premain.c signature.
3115 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3116 run algorithm test programs.
3119 *) Make algorithm test programs more tolerant of whitespace.
3122 *) Have SSL/TLS server implementation tolerate "mismatched" record
3123 protocol version while receiving ClientHello even if the
3124 ClientHello is fragmented. (The server can't insist on the
3125 particular protocol version it has chosen before the ServerHello
3126 message has informed the client about his choice.)
3129 *) Load error codes if they are not already present instead of using a
3130 static variable. This allows them to be cleanly unloaded and reloaded.
3133 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3135 *) Introduce limits to prevent malicious keys being able to
3136 cause a denial of service. (CVE-2006-2940)
3137 [Steve Henson, Bodo Moeller]
3139 *) Fix ASN.1 parsing of certain invalid structures that can result
3140 in a denial of service. (CVE-2006-2937) [Steve Henson]
3142 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3143 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3145 *) Fix SSL client code which could crash if connecting to a
3146 malicious SSLv2 server. (CVE-2006-4343)
3147 [Tavis Ormandy and Will Drewry, Google Security Team]
3149 *) Change ciphersuite string processing so that an explicit
3150 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3151 will no longer include "AES128-SHA"), and any other similar
3152 ciphersuite (same bitmap) from *other* protocol versions (so that
3153 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3154 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3155 changes from 0.9.8b and 0.9.8d.
3158 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3160 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3161 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3163 *) Change the Unix randomness entropy gathering to use poll() when
3164 possible instead of select(), since the latter has some
3165 undesirable limitations.
3166 [Darryl Miles via Richard Levitte and Bodo Moeller]
3168 *) Disable rogue ciphersuites:
3170 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3171 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3172 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3174 The latter two were purportedly from
3175 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3178 Also deactive the remaining ciphersuites from
3179 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3180 unofficial, and the ID has long expired.
3183 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3184 dual-core machines) and other potential thread-safety issues.
3187 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3189 *) Adapt fipsld and the build system to link against the validated FIPS
3190 module in FIPS mode.
3193 *) Fixes for VC++ 2005 build under Windows.
3196 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3197 from a Windows bash shell such as MSYS. It is autodetected from the
3198 "config" script when run from a VC++ environment. Modify standard VC++
3199 build to use fipscanister.o from the GNU make build.
3202 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3204 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3205 The value now differs depending on if you build for FIPS or not.
3206 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3207 safely run with a non-FIPSed libcrypto, as it may crash because of
3208 the difference induced by this change.
3211 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3213 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3214 (part of SSL_OP_ALL). This option used to disable the
3215 countermeasure against man-in-the-middle protocol-version
3216 rollback in the SSL 2.0 server implementation, which is a bad
3217 idea. (CVE-2005-2969)
3219 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3220 for Information Security, National Institute of Advanced Industrial
3221 Science and Technology [AIST], Japan)]
3223 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3224 mainly for FIPS compliance and not fully integrated at this stage.
3227 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3228 the exponentiation using a fixed-length exponent. (Otherwise,
3229 the information leaked through timing could expose the secret key
3230 after many signatures; cf. Bleichenbacher's attack on DSA with
3234 *) Make a new fixed-window mod_exp implementation the default for
3235 RSA, DSA, and DH private-key operations so that the sequence of
3236 squares and multiplies and the memory access pattern are
3237 independent of the particular secret key. This will mitigate
3238 cache-timing and potential related attacks.
3240 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3241 and this is automatically used by BN_mod_exp_mont() if the new flag
3242 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3243 will use this BN flag for private exponents unless the flag
3244 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3245 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3247 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3249 *) Change the client implementation for SSLv23_method() and
3250 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3251 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3252 (Previously, the SSL 2.0 backwards compatible Client Hello
3253 message format would be used even with SSL_OP_NO_SSLv2.)
3256 *) Add support for smime-type MIME parameter in S/MIME messages which some
3260 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3261 a threadsafe manner. Modify rsa code to use new function and add calls
3262 to dsa and dh code (which had race conditions before).
3265 *) Include the fixed error library code in the C error file definitions
3266 instead of fixing them up at runtime. This keeps the error code
3267 structures constant.
3270 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3272 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3275 *) Fixes for newer kerberos headers. NB: the casts are needed because
3276 the 'length' field is signed on one version and unsigned on another
3277 with no (?) obvious way to tell the difference, without these VC++
3278 complains. Also the "definition" of FAR (blank) is no longer included
3279 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3280 some needed definitions.
3283 *) Undo Cygwin change.
3286 *) Added support for proxy certificates according to RFC 3820.
3287 Because they may be a security thread to unaware applications,
3288 they must be explicitely allowed in run-time. See
3289 docs/HOWTO/proxy_certificates.txt for further information.
3292 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3294 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3295 server and client random values. Previously
3296 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3297 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3299 This change has negligible security impact because:
3301 1. Server and client random values still have 24 bytes of pseudo random
3304 2. Server and client random values are sent in the clear in the initial
3307 3. The master secret is derived using the premaster secret (48 bytes in
3308 size for static RSA ciphersuites) as well as client server and random
3311 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3314 [Stephen Henson, reported by UK NISCC]
3316 *) Use Windows randomness collection on Cygwin.
3319 *) Fix hang in EGD/PRNGD query when communication socket is closed
3320 prematurely by EGD/PRNGD.
3321 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3323 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3326 *) Back-port of selected performance improvements from development
3327 branch, as well as improved support for PowerPC platforms.
3330 *) Add lots of checks for memory allocation failure, error codes to indicate
3331 failure and freeing up memory if a failure occurs.
3332 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3334 *) Add new -passin argument to dgst.
3337 *) Perform some character comparisons of different types in X509_NAME_cmp:
3338 this is needed for some certificates that reencode DNs into UTF8Strings
3339 (in violation of RFC3280) and can't or wont issue name rollover
3343 *) Make an explicit check during certificate validation to see that
3344 the CA setting in each certificate on the chain is correct. As a
3345 side effect always do the following basic checks on extensions,
3346 not just when there's an associated purpose to the check:
3348 - if there is an unhandled critical extension (unless the user
3349 has chosen to ignore this fault)
3350 - if the path length has been exceeded (if one is set at all)
3351 - that certain extensions fit the associated purpose (if one has
3355 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3357 *) Avoid a race condition when CRLs are checked in a multi threaded
3358 environment. This would happen due to the reordering of the revoked
3359 entries during signature checking and serial number lookup. Now the
3360 encoding is cached and the serial number sort performed under a lock.
3361 Add new STACK function sk_is_sorted().
3364 *) Add Delta CRL to the extension code.
3367 *) Various fixes to s3_pkt.c so alerts are sent properly.
3368 [David Holmes <d.holmes@f5.com>]
3370 *) Reduce the chances of duplicate issuer name and serial numbers (in
3371 violation of RFC3280) using the OpenSSL certificate creation utilities.
3372 This is done by creating a random 64 bit value for the initial serial
3373 number when a serial number file is created or when a self signed
3374 certificate is created using 'openssl req -x509'. The initial serial
3375 number file is created using 'openssl x509 -next_serial' in CA.pl
3376 rather than being initialized to 1.
3379 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3381 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3382 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3383 [Joe Orton, Steve Henson]
3385 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3387 [Joe Orton, Steve Henson]
3389 *) Make it possible to have multiple active certificates with the same
3390 subject in the CA index file. This is done only if the keyword
3391 'unique_subject' is set to 'no' in the main CA section (default
3392 if 'CA_default') of the configuration file. The value is saved
3393 with the database itself in a separate index attribute file,
3394 named like the index file with '.attr' appended to the name.
3397 *) X509 verify fixes. Disable broken certificate workarounds when
3398 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3399 keyUsage extension present. Don't accept CRLs with unhandled critical
3400 extensions: since verify currently doesn't process CRL extensions this
3401 rejects a CRL with *any* critical extensions. Add new verify error codes
3405 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3406 A clarification of RFC2560 will require the use of OCTET STRINGs and
3407 some implementations cannot handle the current raw format. Since OpenSSL
3408 copies and compares OCSP nonces as opaque blobs without any attempt at
3409 parsing them this should not create any compatibility issues.
3412 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3413 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3414 this HMAC (and other) operations are several times slower than OpenSSL
3418 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3419 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3421 *) Use the correct content when signing type "other".
3424 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3426 *) Fix various bugs revealed by running the NISCC test suite:
3428 Stop out of bounds reads in the ASN1 code when presented with
3429 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3431 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3433 If verify callback ignores invalid public key errors don't try to check
3434 certificate signature with the NULL public key.
3438 *) New -ignore_err option in ocsp application to stop the server
3439 exiting on the first error in a request.
3442 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3443 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3447 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3448 extra data after the compression methods not only for TLS 1.0
3449 but also for SSL 3.0 (as required by the specification).
3450 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3452 *) Change X509_certificate_type() to mark the key as exported/exportable
3453 when it's 512 *bits* long, not 512 bytes.
3456 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3457 blocks during encryption.
3460 *) Various fixes to base64 BIO and non blocking I/O. On write
3461 flushes were not handled properly if the BIO retried. On read
3462 data was not being buffered properly and had various logic bugs.
3463 This also affects blocking I/O when the data being decoded is a
3467 *) Various S/MIME bugfixes and compatibility changes:
3468 output correct application/pkcs7 MIME type if
3469 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3470 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3471 of files as .eml work). Correctly handle very long lines in MIME
3475 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3477 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3478 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3479 a protocol version number mismatch like a decryption error
3480 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3483 *) Turn on RSA blinding by default in the default implementation
3484 to avoid a timing attack. Applications that don't want it can call
3485 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3486 They would be ill-advised to do so in most cases.
3487 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3489 *) Change RSA blinding code so that it works when the PRNG is not
3490 seeded (in this case, the secret RSA exponent is abused as
3491 an unpredictable seed -- if it is not unpredictable, there
3492 is no point in blinding anyway). Make RSA blinding thread-safe
3493 by remembering the creator's thread ID in rsa->blinding and
3494 having all other threads use local one-time blinding factors
3495 (this requires more computation than sharing rsa->blinding, but
3496 avoids excessive locking; and if an RSA object is not shared
3497 between threads, blinding will still be very fast).
3500 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3501 ENGINE as defaults for all supported algorithms irrespective of
3502 the 'flags' parameter. 'flags' is now honoured, so applications
3503 should make sure they are passing it correctly.
3506 *) Target "mingw" now allows native Windows code to be generated in
3507 the Cygwin environment as well as with the MinGW compiler.
3510 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3512 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3513 via timing by performing a MAC computation even if incorrrect
3514 block cipher padding has been found. This is a countermeasure
3515 against active attacks where the attacker has to distinguish
3516 between bad padding and a MAC verification error. (CVE-2003-0078)
3518 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3519 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3520 Martin Vuagnoux (EPFL, Ilion)]
3522 *) Make the no-err option work as intended. The intention with no-err
3523 is not to have the whole error stack handling routines removed from
3524 libcrypto, it's only intended to remove all the function name and
3525 reason texts, thereby removing some of the footprint that may not
3526 be interesting if those errors aren't displayed anyway.
3528 NOTE: it's still possible for any application or module to have it's
3529 own set of error texts inserted. The routines are there, just not
3530 used by default when no-err is given.
3533 *) Add support for FreeBSD on IA64.
3534 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3536 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3537 Kerberos function mit_des_cbc_cksum(). Before this change,
3538 the value returned by DES_cbc_cksum() was like the one from
3539 mit_des_cbc_cksum(), except the bytes were swapped.
3540 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3542 *) Allow an application to disable the automatic SSL chain building.
3543 Before this a rather primitive chain build was always performed in
3544 ssl3_output_cert_chain(): an application had no way to send the
3545 correct chain if the automatic operation produced an incorrect result.
3547 Now the chain builder is disabled if either:
3549 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3551 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3553 The reasoning behind this is that an application would not want the
3554 auto chain building to take place if extra chain certificates are
3555 present and it might also want a means of sending no additional
3556 certificates (for example the chain has two certificates and the
3560 *) Add the possibility to build without the ENGINE framework.
3561 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3563 *) Under Win32 gmtime() can return NULL: check return value in
3564 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3567 *) DSA routines: under certain error conditions uninitialized BN objects
3568 could be freed. Solution: make sure initialization is performed early
3569 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3570 Nils Larsch <nla@trustcenter.de> via PR#459)
3573 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3574 checked on reconnect on the client side, therefore session resumption
3575 could still fail with a "ssl session id is different" error. This
3576 behaviour is masked when SSL_OP_ALL is used due to
3577 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3578 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3579 followup to PR #377.
3582 *) IA-32 assembler support enhancements: unified ELF targets, support
3583 for SCO/Caldera platforms, fix for Cygwin shared build.
3586 *) Add support for FreeBSD on sparc64. As a consequence, support for
3587 FreeBSD on non-x86 processors is separate from x86 processors on
3588 the config script, much like the NetBSD support.
3589 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3591 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3593 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3596 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3597 code (06) was taken as the first octet of the session ID and the last
3598 octet was ignored consequently. As a result SSLv2 client side session
3599 caching could not have worked due to the session ID mismatch between
3601 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3605 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3606 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3610 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3611 seems that in spite of existing for more than a year, many application
3612 author have done nothing to provide the necessary callbacks, which
3613 means that this particular engine will not work properly anywhere.
3614 This is a very unfortunate situation which forces us, in the name
3615 of usability, to give the hw_ncipher.c a static lock, which is part
3617 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3618 appear in 0.9.8 or later. We EXPECT application authors to have
3619 dealt properly with this when 0.9.8 is released (unless we actually
3620 make such changes in the libcrypto locking code that changes will
3621 have to be made anyway).
3624 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3625 octets have been read, EOF or an error occurs. Without this change
3626 some truncated ASN1 structures will not produce an error.
3629 *) Disable Heimdal support, since it hasn't been fully implemented.
3630 Still give the possibility to force the use of Heimdal, but with
3631 warnings and a request that patches get sent to openssl-dev.
3634 *) Add the VC-CE target, introduce the WINCE sysname, and add
3635 INSTALL.WCE and appropriate conditionals to make it build.
3636 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3638 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3639 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3640 edit numbers of the version.
3641 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3643 *) Introduce safe string copy and catenation functions
3644 (BUF_strlcpy() and BUF_strlcat()).
3645 [Ben Laurie (CHATS) and Richard Levitte]
3647 *) Avoid using fixed-size buffers for one-line DNs.
3648 [Ben Laurie (CHATS)]
3650 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3651 resizing buffers containing secrets, and use where appropriate.
3652 [Ben Laurie (CHATS)]
3654 *) Avoid using fixed size buffers for configuration file location.
3655 [Ben Laurie (CHATS)]
3657 *) Avoid filename truncation for various CA files.
3658 [Ben Laurie (CHATS)]
3660 *) Use sizeof in preference to magic numbers.
3661 [Ben Laurie (CHATS)]
3663 *) Avoid filename truncation in cert requests.
3664 [Ben Laurie (CHATS)]
3666 *) Add assertions to check for (supposedly impossible) buffer
3668 [Ben Laurie (CHATS)]
3670 *) Don't cache truncated DNS entries in the local cache (this could
3671 potentially lead to a spoofing attack).
3672 [Ben Laurie (CHATS)]
3674 *) Fix various buffers to be large enough for hex/decimal
3675 representations in a platform independent manner.
3676 [Ben Laurie (CHATS)]
3678 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3679 resizing buffers containing secrets, and use where appropriate.
3680 [Ben Laurie (CHATS)]
3682 *) Add BIO_indent() to avoid much slightly worrying code to do
3684 [Ben Laurie (CHATS)]
3686 *) Convert sprintf()/BIO_puts() to BIO_printf().
3687 [Ben Laurie (CHATS)]
3689 *) buffer_gets() could terminate with the buffer only half
3691 [Ben Laurie (CHATS)]
3693 *) Add assertions to prevent user-supplied crypto functions from
3694 overflowing internal buffers by having large block sizes, etc.
3695 [Ben Laurie (CHATS)]
3697 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3699 [Ben Laurie (CHATS)]
3701 *) Eliminate unused copy of key in RC4.
3702 [Ben Laurie (CHATS)]
3704 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3705 [Ben Laurie (CHATS)]
3707 *) Fix off-by-one error in EGD path.
3708 [Ben Laurie (CHATS)]
3710 *) If RANDFILE path is too long, ignore instead of truncating.
3711 [Ben Laurie (CHATS)]
3713 *) Eliminate unused and incorrectly sized X.509 structure
3715 [Ben Laurie (CHATS)]
3717 *) Eliminate unused and dangerous function knumber().
3718 [Ben Laurie (CHATS)]
3720 *) Eliminate unused and dangerous structure, KSSL_ERR.
3721 [Ben Laurie (CHATS)]
3723 *) Protect against overlong session ID context length in an encoded
3724 session object. Since these are local, this does not appear to be
3726 [Ben Laurie (CHATS)]
3728 *) Change from security patch (see 0.9.6e below) that did not affect
3729 the 0.9.6 release series:
3731 Remote buffer overflow in SSL3 protocol - an attacker could
3732 supply an oversized master key in Kerberos-enabled versions.
3734 [Ben Laurie (CHATS)]
3736 *) Change the SSL kerb5 codes to match RFC 2712.
3739 *) Make -nameopt work fully for req and add -reqopt switch.
3740 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3742 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3743 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3745 *) Make sure tests can be performed even if the corresponding algorithms
3746 have been removed entirely. This was also the last step to make
3747 OpenSSL compilable with DJGPP under all reasonable conditions.
3748 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3750 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3751 to allow version independent disabling of normally unselected ciphers,
3752 which may be activated as a side-effect of selecting a single cipher.
3754 (E.g., cipher list string "RSA" enables ciphersuites that are left
3755 out of "ALL" because they do not provide symmetric encryption.
3756 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3757 [Lutz Jaenicke, Bodo Moeller]
3759 *) Add appropriate support for separate platform-dependent build
3760 directories. The recommended way to make a platform-dependent
3761 build directory is the following (tested on Linux), maybe with
3764 # Place yourself outside of the OpenSSL source tree. In
3765 # this example, the environment variable OPENSSL_SOURCE
3766 # is assumed to contain the absolute OpenSSL source directory.
3767 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3768 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3769 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3770 mkdir -p `dirname $F`
3771 ln -s $OPENSSL_SOURCE/$F $F
3774 To be absolutely sure not to disturb the source tree, a "make clean"
3775 is a good thing. If it isn't successfull, don't worry about it,
3776 it probably means the source directory is very clean.
3779 *) Make sure any ENGINE control commands make local copies of string
3780 pointers passed to them whenever necessary. Otherwise it is possible
3781 the caller may have overwritten (or deallocated) the original string
3782 data when a later ENGINE operation tries to use the stored values.
3783 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3785 *) Improve diagnostics in file reading and command-line digests.
3786 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3788 *) Add AES modes CFB and OFB to the object database. Correct an
3789 error in AES-CFB decryption.
3792 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3793 allows existing EVP_CIPHER_CTX structures to be reused after
3794 calling EVP_*Final(). This behaviour is used by encryption
3795 BIOs and some applications. This has the side effect that
3796 applications must explicitly clean up cipher contexts with
3797 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3800 *) Check the values of dna and dnb in bn_mul_recursive before calling
3801 bn_mul_comba (a non zero value means the a or b arrays do not contain
3802 n2 elements) and fallback to bn_mul_normal if either is not zero.
3805 *) Fix escaping of non-ASCII characters when using the -subj option
3806 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3809 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3810 form for "surname", serialNumber has no short form.
3811 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3812 therefore remove "mail" short name for "internet 7".
3813 The OID for unique identifiers in X509 certificates is
3814 x500UniqueIdentifier, not uniqueIdentifier.
3815 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3818 *) Add an "init" command to the ENGINE config module and auto initialize
3819 ENGINEs. Without any "init" command the ENGINE will be initialized
3820 after all ctrl commands have been executed on it. If init=1 the
3821 ENGINE is initailized at that point (ctrls before that point are run
3822 on the uninitialized ENGINE and after on the initialized one). If
3823 init=0 then the ENGINE will not be iniatialized at all.
3826 *) Fix the 'app_verify_callback' interface so that the user-defined
3827 argument is actually passed to the callback: In the
3828 SSL_CTX_set_cert_verify_callback() prototype, the callback
3829 declaration has been changed from
3832 int (*cb)(X509_STORE_CTX *,void *);
3833 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3834 i=s->ctx->app_verify_callback(&ctx)
3835 has been changed into
3836 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3838 To update applications using SSL_CTX_set_cert_verify_callback(),
3839 a dummy argument can be added to their callback functions.
3840 [D. K. Smetters <smetters@parc.xerox.com>]
3842 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3843 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3845 *) Add and OPENSSL_LOAD_CONF define which will cause
3846 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3847 This allows older applications to transparently support certain
3848 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3849 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3850 load the config file and OPENSSL_add_all_algorithms_conf() which will
3851 always load it have also been added.
3854 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3855 Adjust NIDs and EVP layer.
3856 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3858 *) Config modules support in openssl utility.
3860 Most commands now load modules from the config file,
3861 though in a few (such as version) this isn't done
3862 because it couldn't be used for anything.
3864 In the case of ca and req the config file used is
3865 the same as the utility itself: that is the -config
3866 command line option can be used to specify an
3870 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3871 use "openssl_conf" if filename is NULL use default openssl config file.
3874 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3875 config section name. Add a new flag to tolerate a missing config file
3876 and move code to CONF_modules_load_file().
3879 *) Support for crypto accelerator cards from Accelerated Encryption
3880 Processing, www.aep.ie. (Use engine 'aep')
3881 The support was copied from 0.9.6c [engine] and adapted/corrected
3882 to work with the new engine framework.
3883 [AEP Inc. and Richard Levitte]
3885 *) Support for SureWare crypto accelerator cards from Baltimore
3886 Technologies. (Use engine 'sureware')
3887 The support was copied from 0.9.6c [engine] and adapted
3888 to work with the new engine framework.
3891 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3892 make the newer ENGINE framework commands for the CHIL engine work.
3893 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3895 *) Make it possible to produce shared libraries on ReliantUNIX.
3896 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3898 *) Add the configuration target debug-linux-ppro.
3899 Make 'openssl rsa' use the general key loading routines
3900 implemented in apps.c, and make those routines able to
3901 handle the key format FORMAT_NETSCAPE and the variant
3903 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3905 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3906 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3908 *) Add -keyform to rsautl, and document -engine.
3909 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3911 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3912 BIO_R_NO_SUCH_FILE error code rather than the generic
3913 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3916 *) Add new functions
3918 ERR_peek_last_error_line
3919 ERR_peek_last_error_line_data.
3920 These are similar to
3923 ERR_peek_error_line_data,
3924 but report on the latest error recorded rather than the first one
3925 still in the error queue.
3926 [Ben Laurie, Bodo Moeller]
3928 *) default_algorithms option in ENGINE config module. This allows things
3930 default_algorithms = ALL
3931 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3934 *) Prelminary ENGINE config module.
3937 *) New experimental application configuration code.
3940 *) Change the AES code to follow the same name structure as all other
3941 symmetric ciphers, and behave the same way. Move everything to
3942 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3943 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3945 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3946 [Ben Laurie and Theo de Raadt]
3948 *) Add option to output public keys in req command.
3949 [Massimiliano Pala madwolf@openca.org]
3951 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3952 (up to about 10% better than before for P-192 and P-224).
3955 *) New functions/macros
3957 SSL_CTX_set_msg_callback(ctx, cb)
3958 SSL_CTX_set_msg_callback_arg(ctx, arg)
3959 SSL_set_msg_callback(ssl, cb)
3960 SSL_set_msg_callback_arg(ssl, arg)
3962 to request calling a callback function
3964 void cb(int write_p, int version, int content_type,
3965 const void *buf, size_t len, SSL *ssl, void *arg)
3967 whenever a protocol message has been completely received
3968 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3969 protocol version according to which the SSL library interprets
3970 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3971 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3972 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3973 specification (change_cipher_spec(20), alert(21), handshake(22)).
3974 'buf' and 'len' point to the actual message, 'ssl' to the
3975 SSL object, and 'arg' is the application-defined value set by
3976 SSL[_CTX]_set_msg_callback_arg().
3978 'openssl s_client' and 'openssl s_server' have new '-msg' options
3979 to enable a callback that displays all protocol messages.
3982 *) Change the shared library support so shared libraries are built as
3983 soon as the corresponding static library is finished, and thereby get
3984 openssl and the test programs linked against the shared library.
3985 This still only happens when the keyword "shard" has been given to
3986 the configuration scripts.
3988 NOTE: shared library support is still an experimental thing, and
3989 backward binary compatibility is still not guaranteed.
3990 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3992 *) Add support for Subject Information Access extension.
3993 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3995 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3996 additional bytes when new memory had to be allocated, not just
3997 when reusing an existing buffer.
4000 *) New command line and configuration option 'utf8' for the req command.
4001 This allows field values to be specified as UTF8 strings.
4004 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4005 runs for the former and machine-readable output for the latter.
4008 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4009 of the e-mail address in the DN (i.e., it will go into a certificate
4010 extension only). The new configuration file option 'email_in_dn = no'
4011 has the same effect.
4012 [Massimiliano Pala madwolf@openca.org]
4014 *) Change all functions with names starting with des_ to be starting
4015 with DES_ instead. Add wrappers that are compatible with libdes,
4016 but are named _ossl_old_des_*. Finally, add macros that map the
4017 des_* symbols to the corresponding _ossl_old_des_* if libdes
4018 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4019 desired, the des_* symbols will be mapped to DES_*, with one
4022 Since we provide two compatibility mappings, the user needs to
4023 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4024 compatibility is desired. The default (i.e., when that macro
4025 isn't defined) is OpenSSL 0.9.6c compatibility.
4027 There are also macros that enable and disable the support of old
4028 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4029 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4030 are defined, the default will apply: to support the old des routines.
4032 In either case, one must include openssl/des.h to get the correct
4033 definitions. Do not try to just include openssl/des_old.h, that
4036 NOTE: This is a major break of an old API into a new one. Software
4037 authors are encouraged to switch to the DES_ style functions. Some
4038 time in the future, des_old.h and the libdes compatibility functions
4039 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4040 default), and then completely removed.
4043 *) Test for certificates which contain unsupported critical extensions.
4044 If such a certificate is found during a verify operation it is
4045 rejected by default: this behaviour can be overridden by either
4046 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4047 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4048 X509_supported_extension() has also been added which returns 1 if a
4049 particular extension is supported.
4052 *) Modify the behaviour of EVP cipher functions in similar way to digests
4053 to retain compatibility with existing code.
4056 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4057 compatibility with existing code. In particular the 'ctx' parameter does
4058 not have to be to be initialized before the call to EVP_DigestInit() and
4059 it is tidied up after a call to EVP_DigestFinal(). New function
4060 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4061 EVP_MD_CTX_copy() changed to not require the destination to be
4062 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4063 requires the destination to be valid.
4065 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4066 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4069 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4070 so that complete 'Handshake' protocol structures are kept in memory
4071 instead of overwriting 'msg_type' and 'length' with 'body' data.
4074 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4075 [Massimo Santin via Richard Levitte]
4077 *) Major restructuring to the underlying ENGINE code. This includes
4078 reduction of linker bloat, separation of pure "ENGINE" manipulation
4079 (initialisation, etc) from functionality dealing with implementations
4080 of specific crypto iterfaces. This change also introduces integrated
4081 support for symmetric ciphers and digest implementations - so ENGINEs
4082 can now accelerate these by providing EVP_CIPHER and EVP_MD
4083 implementations of their own. This is detailed in crypto/engine/README
4084 as it couldn't be adequately described here. However, there are a few
4085 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4086 were changed in the original introduction of ENGINE code have now
4087 reverted back - the hooking from this code to ENGINE is now a good
4088 deal more passive and at run-time, operations deal directly with
4089 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4090 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4091 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4092 they were not being used by the framework as there is no concept of a
4093 BIGNUM_METHOD and they could not be generalised to the new
4094 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4095 ENGINE_cpy() has been removed as it cannot be consistently defined in
4099 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4102 *) Change mkdef.pl to sort symbols that get the same entry number,
4103 and make sure the automatically generated functions ERR_load_*
4104 become part of libeay.num as well.
4107 *) New function SSL_renegotiate_pending(). This returns true once
4108 renegotiation has been requested (either SSL_renegotiate() call
4109 or HelloRequest/ClientHello receveived from the peer) and becomes
4110 false once a handshake has been completed.
4111 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4112 sends a HelloRequest, but does not ensure that a handshake takes
4113 place. SSL_renegotiate_pending() is useful for checking if the
4114 client has followed the request.)
4117 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4118 By default, clients may request session resumption even during
4119 renegotiation (if session ID contexts permit); with this option,
4120 session resumption is possible only in the first handshake.
4122 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4123 more bits available for options that should not be part of
4124 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4127 *) Add some demos for certificate and certificate request creation.
4130 *) Make maximum certificate chain size accepted from the peer application
4131 settable (SSL*_get/set_max_cert_list()), as proposed by
4132 "Douglas E. Engert" <deengert@anl.gov>.
4135 *) Add support for shared libraries for Unixware-7
4136 (Boyd Lynn Gerber <gerberb@zenez.com>).
4139 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4140 be done prior to destruction. Use this to unload error strings from
4141 ENGINEs that load their own error strings. NB: This adds two new API
4142 functions to "get" and "set" this destroy handler in an ENGINE.
4145 *) Alter all existing ENGINE implementations (except "openssl" and
4146 "openbsd") to dynamically instantiate their own error strings. This
4147 makes them more flexible to be built both as statically-linked ENGINEs
4148 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4149 Also, add stub code to each that makes building them as self-contained
4150 shared-libraries easier (see README.ENGINE).
4153 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4154 implementations into applications that are completely implemented in
4155 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4156 commands that can be used to configure what shared-library to load and
4157 to control aspects of the way it is handled. Also, made an update to
4158 the README.ENGINE file that brings its information up-to-date and
4159 provides some information and instructions on the "dynamic" ENGINE
4160 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4163 *) Make it possible to unload ranges of ERR strings with a new
4164 "ERR_unload_strings" function.
4167 *) Add a copy() function to EVP_MD.
4170 *) Make EVP_MD routines take a context pointer instead of just the
4171 md_data void pointer.
4174 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4175 that the digest can only process a single chunk of data
4176 (typically because it is provided by a piece of
4177 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4178 is only going to provide a single chunk of data, and hence the
4179 framework needn't accumulate the data for oneshot drivers.
4182 *) As with "ERR", make it possible to replace the underlying "ex_data"
4183 functions. This change also alters the storage and management of global
4184 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4185 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4186 index counters. The API functions that use this state have been changed
4187 to take a "class_index" rather than pointers to the class's local STACK
4188 and counter, and there is now an API function to dynamically create new
4189 classes. This centralisation allows us to (a) plug a lot of the
4190 thread-safety problems that existed, and (b) makes it possible to clean
4191 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4192 such data would previously have always leaked in application code and
4193 workarounds were in place to make the memory debugging turn a blind eye
4194 to it. Application code that doesn't use this new function will still
4195 leak as before, but their memory debugging output will announce it now
4196 rather than letting it slide.
4198 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4199 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4200 has a return value to indicate success or failure.
4203 *) Make it possible to replace the underlying "ERR" functions such that the
4204 global state (2 LHASH tables and 2 locks) is only used by the "default"
4205 implementation. This change also adds two functions to "get" and "set"
4206 the implementation prior to it being automatically set the first time
4207 any other ERR function takes place. Ie. an application can call "get",
4208 pass the return value to a module it has just loaded, and that module
4209 can call its own "set" function using that value. This means the
4210 module's "ERR" operations will use (and modify) the error state in the
4211 application and not in its own statically linked copy of OpenSSL code.
4214 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4215 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4216 the operation, and provides a more encapsulated way for external code
4217 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4218 to use these functions rather than manually incrementing the counts.
4220 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4223 *) Add EVP test program.
4226 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4229 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4230 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4231 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4232 These allow a CRL to be built without having to access X509_CRL fields
4233 directly. Modify 'ca' application to use new functions.
4236 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4237 bug workarounds. Rollback attack detection is a security feature.
4238 The problem will only arise on OpenSSL servers when TLSv1 is not
4239 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4240 Software authors not wanting to support TLSv1 will have special reasons
4241 for their choice and can explicitly enable this option.
4242 [Bodo Moeller, Lutz Jaenicke]
4244 *) Rationalise EVP so it can be extended: don't include a union of
4245 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4246 (similar to those existing for EVP_CIPHER_CTX).
4251 EVP_MD_CTX_init(&md); /* new function call */
4252 EVP_DigestInit(&md, EVP_sha1());
4253 EVP_DigestUpdate(&md, in, len);
4254 EVP_DigestFinal(&md, out, NULL);
4255 EVP_MD_CTX_cleanup(&md); /* new function call */
4259 *) Make DES key schedule conform to the usual scheme, as well as
4260 correcting its structure. This means that calls to DES functions
4261 now have to pass a pointer to a des_key_schedule instead of a
4262 plain des_key_schedule (which was actually always a pointer
4265 des_key_schedule ks;
4267 des_set_key_checked(..., &ks);
4268 des_ncbc_encrypt(..., &ks, ...);
4270 (Note that a later change renames 'des_...' into 'DES_...'.)
4273 *) Initial reduction of linker bloat: the use of some functions, such as
4274 PEM causes large amounts of unused functions to be linked in due to
4275 poor organisation. For example pem_all.c contains every PEM function
4276 which has a knock on effect of linking in large amounts of (unused)
4277 ASN1 code. Grouping together similar functions and splitting unrelated
4278 functions prevents this.
4281 *) Cleanup of EVP macros.
4284 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4285 correct _ecb suffix.
4288 *) Add initial OCSP responder support to ocsp application. The
4289 revocation information is handled using the text based index
4290 use by the ca application. The responder can either handle
4291 requests generated internally, supplied in files (for example
4292 via a CGI script) or using an internal minimal server.
4295 *) Add configuration choices to get zlib compression for TLS.
4298 *) Changes to Kerberos SSL for RFC 2712 compliance:
4299 1. Implemented real KerberosWrapper, instead of just using
4300 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4301 2. Implemented optional authenticator field of KerberosWrapper.
4303 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4304 and authenticator structs; see crypto/krb5/.
4306 Generalized Kerberos calls to support multiple Kerberos libraries.
4307 [Vern Staats <staatsvr@asc.hpc.mil>,
4308 Jeffrey Altman <jaltman@columbia.edu>
4309 via Richard Levitte]
4311 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4312 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4313 values for each of the key sizes rather than having just
4314 parameters (and 'speed' generating keys each time).
4317 *) Speed up EVP routines.
4320 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4321 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4322 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4323 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4325 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4326 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4327 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4330 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4332 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4335 *) Added the OS2-EMX target.
4336 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4338 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4339 to support NCONF routines in extension code. New function CONF_set_nconf()
4340 to allow functions which take an NCONF to also handle the old LHASH
4341 structure: this means that the old CONF compatible routines can be
4342 retained (in particular wrt extensions) without having to duplicate the
4343 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4346 *) Enhance the general user interface with mechanisms for inner control
4347 and with possibilities to have yes/no kind of prompts.
4350 *) Change all calls to low level digest routines in the library and
4351 applications to use EVP. Add missing calls to HMAC_cleanup() and
4352 don't assume HMAC_CTX can be copied using memcpy().
4353 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4355 *) Add the possibility to control engines through control names but with
4356 arbitrary arguments instead of just a string.
4357 Change the key loaders to take a UI_METHOD instead of a callback
4358 function pointer. NOTE: this breaks binary compatibility with earlier
4359 versions of OpenSSL [engine].
4360 Adapt the nCipher code for these new conditions and add a card insertion
4364 *) Enhance the general user interface with mechanisms to better support
4365 dialog box interfaces, application-defined prompts, the possibility
4366 to use defaults (for example default passwords from somewhere else)
4367 and interrupts/cancellations.
4370 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4371 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4374 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4375 tidy up some unnecessarily weird code in 'sk_new()').
4376 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4378 *) Change the key loading routines for ENGINEs to use the same kind
4379 callback (pem_password_cb) as all other routines that need this
4383 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4384 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4385 than this minimum value is recommended.
4388 *) New random seeder for OpenVMS, using the system process statistics
4389 that are easily reachable.
4392 *) Windows apparently can't transparently handle global
4393 variables defined in DLLs. Initialisations such as:
4395 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4397 wont compile. This is used by the any applications that need to
4398 declare their own ASN1 modules. This was fixed by adding the option
4399 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4400 needed for static libraries under Win32.
4403 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4404 setting of purpose and trust fields. New X509_STORE trust and
4405 purpose functions and tidy up setting in other SSL functions.
4408 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4409 structure. These are inherited by X509_STORE_CTX when it is
4410 initialised. This allows various defaults to be set in the
4411 X509_STORE structure (such as flags for CRL checking and custom
4412 purpose or trust settings) for functions which only use X509_STORE_CTX
4413 internally such as S/MIME.
4415 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4416 trust settings if they are not set in X509_STORE. This allows X509_STORE
4417 purposes and trust (in S/MIME for example) to override any set by default.
4419 Add command line options for CRL checking to smime, s_client and s_server
4423 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4424 are set then the CRL is looked up in the X509_STORE structure and
4425 its validity and signature checked, then if the certificate is found
4426 in the CRL the verify fails with a revoked error.
4428 Various new CRL related callbacks added to X509_STORE_CTX structure.
4430 Command line options added to 'verify' application to support this.
4432 This needs some additional work, such as being able to handle multiple
4433 CRLs with different times, extension based lookup (rather than just
4434 by subject name) and ultimately more complete V2 CRL extension
4438 *) Add a general user interface API (crypto/ui/). This is designed
4439 to replace things like des_read_password and friends (backward
4440 compatibility functions using this new API are provided).
4441 The purpose is to remove prompting functions from the DES code
4442 section as well as provide for prompting through dialog boxes in
4443 a window system and the like.
4446 *) Add "ex_data" support to ENGINE so implementations can add state at a
4447 per-structure level rather than having to store it globally.
4450 *) Make it possible for ENGINE structures to be copied when retrieved by
4451 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4452 This causes the "original" ENGINE structure to act like a template,
4453 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4454 operational state can be localised to each ENGINE structure, despite the
4455 fact they all share the same "methods". New ENGINE structures returned in
4456 this case have no functional references and the return value is the single
4457 structural reference. This matches the single structural reference returned
4458 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4462 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4463 needs to match any other type at all we need to manually clear the
4467 *) Changes to the "openssl engine" utility to include;
4468 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4469 about an ENGINE's available control commands.
4470 - executing control commands from command line arguments using the
4471 '-pre' and '-post' switches. '-post' is only used if '-t' is
4472 specified and the ENGINE is successfully initialised. The syntax for
4473 the individual commands are colon-separated, for example;
4474 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4477 *) New dynamic control command support for ENGINEs. ENGINEs can now
4478 declare their own commands (numbers), names (strings), descriptions,
4479 and input types for run-time discovery by calling applications. A
4480 subset of these commands are implicitly classed as "executable"
4481 depending on their input type, and only these can be invoked through
4482 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4483 can be based on user input, config files, etc). The distinction is
4484 that "executable" commands cannot return anything other than a boolean
4485 result and can only support numeric or string input, whereas some
4486 discoverable commands may only be for direct use through
4487 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4488 pointers, or other custom uses. The "executable" commands are to
4489 support parameterisations of ENGINE behaviour that can be
4490 unambiguously defined by ENGINEs and used consistently across any
4491 OpenSSL-based application. Commands have been added to all the
4492 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4493 control over shared-library paths without source code alterations.
4496 *) Changed all ENGINE implementations to dynamically allocate their
4497 ENGINEs rather than declaring them statically. Apart from this being
4498 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4499 this also allows the implementations to compile without using the
4500 internal engine_int.h header.
4503 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4504 'const' value. Any code that should be able to modify a RAND_METHOD
4505 should already have non-const pointers to it (ie. they should only
4506 modify their own ones).
4509 *) Made a variety of little tweaks to the ENGINE code.
4510 - "atalla" and "ubsec" string definitions were moved from header files
4511 to C code. "nuron" string definitions were placed in variables
4512 rather than hard-coded - allowing parameterisation of these values
4513 later on via ctrl() commands.
4514 - Removed unused "#if 0"'d code.
4515 - Fixed engine list iteration code so it uses ENGINE_free() to release
4516 structural references.
4517 - Constified the RAND_METHOD element of ENGINE structures.
4518 - Constified various get/set functions as appropriate and added
4519 missing functions (including a catch-all ENGINE_cpy that duplicates
4520 all ENGINE values onto a new ENGINE except reference counts/state).
4521 - Removed NULL parameter checks in get/set functions. Setting a method
4522 or function to NULL is a way of cancelling out a previously set
4523 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4524 and doesn't justify the extra error symbols and code.
4525 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4526 flags from engine_int.h to engine.h.
4527 - Changed prototypes for ENGINE handler functions (init(), finish(),
4528 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4531 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4532 to the algorithm using long division. The binary algorithm can be
4533 used only if the modulus is odd. On 32-bit systems, it is faster
4534 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4535 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4536 up to 450 bits. In 64-bit environments, the binary algorithm
4537 appears to be advantageous for much longer moduli; here we use it
4538 for moduli up to 2048 bits.
4541 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4542 could not support the combine flag in choice fields.
4545 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4546 extensions from a certificate request to the certificate.
4549 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4550 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4551 file: this allows the display of the certificate about to be
4552 signed to be customised, to allow certain fields to be included
4553 or excluded and extension details. The old system didn't display
4554 multicharacter strings properly, omitted fields not in the policy
4555 and couldn't display additional details such as extensions.
4558 *) Function EC_POINTs_mul for multiple scalar multiplication
4559 of an arbitrary number of elliptic curve points
4560 \sum scalars[i]*points[i],
4561 optionally including the generator defined for the EC_GROUP:
4562 scalar*generator + \sum scalars[i]*points[i].
4564 EC_POINT_mul is a simple wrapper function for the typical case
4565 that the point list has just one item (besides the optional
4569 *) First EC_METHODs for curves over GF(p):
4571 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4572 operations and provides various method functions that can also
4573 operate with faster implementations of modular arithmetic.
4575 EC_GFp_mont_method() reuses most functions that are part of
4576 EC_GFp_simple_method, but uses Montgomery arithmetic.
4578 [Bodo Moeller; point addition and point doubling
4579 implementation directly derived from source code provided by
4580 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4582 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4583 crypto/ec/ec_lib.c):
4585 Curves are EC_GROUP objects (with an optional group generator)
4586 based on EC_METHODs that are built into the library.
4588 Points are EC_POINT objects based on EC_GROUP objects.
4590 Most of the framework would be able to handle curves over arbitrary
4591 finite fields, but as there are no obvious types for fields other
4592 than GF(p), some functions are limited to that for now.
4595 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4596 that the file contains a complete HTTP response.
4599 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4600 change the def and num file printf format specifier from "%-40sXXX"
4601 to "%-39s XXX". The latter will always guarantee a space after the
4602 field while the former will cause them to run together if the field
4603 is 40 of more characters long.
4606 *) Constify the cipher and digest 'method' functions and structures
4607 and modify related functions to take constant EVP_MD and EVP_CIPHER
4611 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4612 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4615 *) Modify EVP_Digest*() routines so they now return values. Although the
4616 internal software routines can never fail additional hardware versions
4620 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4622 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4623 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4626 ERR_R_NESTED_ASN1_ERROR
4628 ERR_R_MISSING_ASN1_EOS
4629 were 4 .. 9, conflicting with
4630 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4632 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4633 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4635 Add new error code 'ERR_R_INTERNAL_ERROR'.
4638 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4642 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4643 sets the subject name for a new request or supersedes the
4644 subject name in a given request. Formats that can be parsed are
4645 'CN=Some Name, OU=myOU, C=IT'
4647 'CN=Some Name/OU=myOU/C=IT'.
4649 Add options '-batch' and '-verbose' to 'openssl req'.
4650 [Massimiliano Pala <madwolf@hackmasters.net>]
4652 *) Introduce the possibility to access global variables through
4653 functions on platform were that's the best way to handle exporting
4654 global variables in shared libraries. To enable this functionality,
4655 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4656 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4657 is normally done by Configure or something similar).
4659 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4660 in the source file (foo.c) like this:
4662 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4663 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4665 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4666 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4668 OPENSSL_DECLARE_GLOBAL(int,foo);
4669 #define foo OPENSSL_GLOBAL_REF(foo)
4670 OPENSSL_DECLARE_GLOBAL(double,bar);
4671 #define bar OPENSSL_GLOBAL_REF(bar)
4673 The #defines are very important, and therefore so is including the
4674 header file everywhere where the defined globals are used.
4676 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4677 of ASN.1 items, but that structure is a bit different.
4679 The largest change is in util/mkdef.pl which has been enhanced with
4680 better and easier to understand logic to choose which symbols should
4681 go into the Windows .def files as well as a number of fixes and code
4682 cleanup (among others, algorithm keywords are now sorted
4683 lexicographically to avoid constant rewrites).
4686 *) In BN_div() keep a copy of the sign of 'num' before writing the
4687 result to 'rm' because if rm==num the value will be overwritten
4688 and produce the wrong result if 'num' is negative: this caused
4689 problems with BN_mod() and BN_nnmod().
4692 *) Function OCSP_request_verify(). This checks the signature on an
4693 OCSP request and verifies the signer certificate. The signer
4694 certificate is just checked for a generic purpose and OCSP request
4698 *) Add OCSP_check_validity() function to check the validity of OCSP
4699 responses. OCSP responses are prepared in real time and may only
4700 be a few seconds old. Simply checking that the current time lies
4701 between thisUpdate and nextUpdate max reject otherwise valid responses
4702 caused by either OCSP responder or client clock inaccuracy. Instead
4703 we allow thisUpdate and nextUpdate to fall within a certain period of
4704 the current time. The age of the response can also optionally be
4705 checked. Two new options -validity_period and -status_age added to
4709 *) If signature or public key algorithm is unrecognized print out its
4710 OID rather that just UNKNOWN.
4713 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4714 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4715 ID to be generated from the issuer certificate alone which can then be
4716 passed to OCSP_id_issuer_cmp().
4719 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4720 ASN1 modules to export functions returning ASN1_ITEM pointers
4721 instead of the ASN1_ITEM structures themselves. This adds several
4722 new macros which allow the underlying ASN1 function/structure to
4723 be accessed transparently. As a result code should not use ASN1_ITEM
4724 references directly (such as &X509_it) but instead use the relevant
4725 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4726 use of the new ASN1 code on platforms where exporting structures
4727 is problematical (for example in shared libraries) but exporting
4728 functions returning pointers to structures is not.
4731 *) Add support for overriding the generation of SSL/TLS session IDs.
4732 These callbacks can be registered either in an SSL_CTX or per SSL.
4733 The purpose of this is to allow applications to control, if they wish,
4734 the arbitrary values chosen for use as session IDs, particularly as it
4735 can be useful for session caching in multiple-server environments. A
4736 command-line switch for testing this (and any client code that wishes
4737 to use such a feature) has been added to "s_server".
4738 [Geoff Thorpe, Lutz Jaenicke]
4740 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4741 of the form '#if defined(...) || defined(...) || ...' and
4742 '#if !defined(...) && !defined(...) && ...'. This also avoids
4743 the growing number of special cases it was previously handling.
4746 *) Make all configuration macros available for application by making
4747 sure they are available in opensslconf.h, by giving them names starting
4748 with "OPENSSL_" to avoid conflicts with other packages and by making
4749 sure e_os2.h will cover all platform-specific cases together with
4751 Additionally, it is now possible to define configuration/platform-
4752 specific names (called "system identities"). In the C code, these
4753 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4754 macro with the name beginning with "OPENSSL_SYS_", which is determined
4755 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4759 *) New option -set_serial to 'req' and 'x509' this allows the serial
4760 number to use to be specified on the command line. Previously self
4761 signed certificates were hard coded with serial number 0 and the
4762 CA options of 'x509' had to use a serial number in a file which was
4766 *) New options to 'ca' utility to support V2 CRL entry extensions.
4767 Currently CRL reason, invalidity date and hold instruction are
4768 supported. Add new CRL extensions to V3 code and some new objects.
4771 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4772 disable standard block padding (aka PKCS#5 padding) in the EVP
4773 API, which was previously mandatory. This means that the data is
4774 not padded in any way and so the total length much be a multiple
4775 of the block size, otherwise an error occurs.
4778 *) Initial (incomplete) OCSP SSL support.
4781 *) New function OCSP_parse_url(). This splits up a URL into its host,
4782 port and path components: primarily to parse OCSP URLs. New -url
4783 option to ocsp utility.
4786 *) New nonce behavior. The return value of OCSP_check_nonce() now
4787 reflects the various checks performed. Applications can decide
4788 whether to tolerate certain situations such as an absent nonce
4789 in a response when one was present in a request: the ocsp application
4790 just prints out a warning. New function OCSP_add1_basic_nonce()
4791 this is to allow responders to include a nonce in a response even if
4792 the request is nonce-less.
4795 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4796 skipped when using openssl x509 multiple times on a single input file,
4797 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4800 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4801 set string type: to handle setting ASN1_TIME structures. Fix ca
4802 utility to correctly initialize revocation date of CRLs.
4805 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4806 the clients preferred ciphersuites and rather use its own preferences.
4807 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4808 Internet Explorer by ensuring unchanged hash method during stepup.
4809 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4812 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4813 to aes and add a new 'exist' option to print out symbols that don't
4817 *) Additional options to ocsp utility to allow flags to be set and
4818 additional certificates supplied.
4821 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4822 OCSP client a number of certificate to only verify the response
4826 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4827 handle the new API. Currently only ECB, CBC modes supported. Add new
4830 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4831 Encryption Standard (AES) Ciphersuites for Transport Layer
4832 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4833 not enabled by default and were not part of the "ALL" ciphersuite
4834 alias because they were not yet official; they could be
4835 explicitly requested by specifying the "AESdraft" ciphersuite
4836 group alias. In the final release of OpenSSL 0.9.7, the group
4837 alias is called "AES" and is part of "ALL".)
4838 [Ben Laurie, Steve Henson, Bodo Moeller]
4840 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4841 request to response.
4844 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4845 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4846 extract information from a certificate request. OCSP_response_create()
4847 creates a response and optionally adds a basic response structure.
4848 OCSP_basic_add1_status() adds a complete single response to a basic
4849 response and returns the OCSP_SINGLERESP structure just added (to allow
4850 extensions to be included for example). OCSP_basic_add1_cert() adds a
4851 certificate to a basic response and OCSP_basic_sign() signs a basic
4852 response with various flags. New helper functions ASN1_TIME_check()
4853 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4854 (converts ASN1_TIME to GeneralizedTime).
4857 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4858 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4859 structure from a certificate. X509_pubkey_digest() digests the public_key
4860 contents: this is used in various key identifiers.
4863 *) Make sk_sort() tolerate a NULL argument.
4864 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4866 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4867 passed by the function are trusted implicitly. If any of them signed the
4868 response then it is assumed to be valid and is not verified.
4871 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4872 to data. This was previously part of the PKCS7 ASN1 code. This
4873 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4874 [Steve Henson, reported by Kenneth R. Robinette
4875 <support@securenetterm.com>]
4877 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4878 routines: without these tracing memory leaks is very painful.
4879 Fix leaks in PKCS12 and PKCS7 routines.
4882 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4883 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4884 effectively meant GeneralizedTime would never be used. Now it
4885 is initialised to -1 but X509_time_adj() now has to check the value
4886 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4887 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4888 [Steve Henson, reported by Kenneth R. Robinette
4889 <support@securenetterm.com>]
4891 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4892 result in a zero length in the ASN1_INTEGER structure which was
4893 not consistent with the structure when d2i_ASN1_INTEGER() was used
4894 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4895 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4896 where it did not print out a minus for negative ASN1_INTEGER.
4899 *) Add summary printout to ocsp utility. The various functions which
4900 convert status values to strings have been renamed to:
4901 OCSP_response_status_str(), OCSP_cert_status_str() and
4902 OCSP_crl_reason_str() and are no longer static. New options
4903 to verify nonce values and to disable verification. OCSP response
4904 printout format cleaned up.
4907 *) Add additional OCSP certificate checks. These are those specified
4908 in RFC2560. This consists of two separate checks: the CA of the
4909 certificate being checked must either be the OCSP signer certificate
4910 or the issuer of the OCSP signer certificate. In the latter case the
4911 OCSP signer certificate must contain the OCSP signing extended key
4912 usage. This check is performed by attempting to match the OCSP
4913 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4914 in the OCSP_CERTID structures of the response.
4917 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4918 and related routines. This uses the standard OpenSSL certificate
4919 verify routines to perform initial checks (just CA validity) and
4920 to obtain the certificate chain. Then additional checks will be
4921 performed on the chain. Currently the root CA is checked to see
4922 if it is explicitly trusted for OCSP signing. This is used to set
4923 a root CA as a global signing root: that is any certificate that
4924 chains to that CA is an acceptable OCSP signing certificate.
4927 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4928 extensions from a separate configuration file.
4929 As when reading extensions from the main configuration file,
4930 the '-extensions ...' option may be used for specifying the
4932 [Massimiliano Pala <madwolf@comune.modena.it>]
4934 *) New OCSP utility. Allows OCSP requests to be generated or
4935 read. The request can be sent to a responder and the output
4936 parsed, outputed or printed in text form. Not complete yet:
4937 still needs to check the OCSP response validity.
4940 *) New subcommands for 'openssl ca':
4941 'openssl ca -status <serial>' prints the status of the cert with
4942 the given serial number (according to the index file).
4943 'openssl ca -updatedb' updates the expiry status of certificates
4945 [Massimiliano Pala <madwolf@comune.modena.it>]
4947 *) New '-newreq-nodes' command option to CA.pl. This is like
4948 '-newreq', but calls 'openssl req' with the '-nodes' option
4949 so that the resulting key is not encrypted.
4950 [Damien Miller <djm@mindrot.org>]
4952 *) New configuration for the GNU Hurd.
4953 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4955 *) Initial code to implement OCSP basic response verify. This
4956 is currently incomplete. Currently just finds the signer's
4957 certificate and verifies the signature on the response.
4960 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4961 value of OPENSSLDIR. This is available via the new '-d' option
4962 to 'openssl version', and is also included in 'openssl version -a'.
4965 *) Allowing defining memory allocation callbacks that will be given
4966 file name and line number information in additional arguments
4967 (a const char* and an int). The basic functionality remains, as
4968 well as the original possibility to just replace malloc(),
4969 realloc() and free() by functions that do not know about these
4970 additional arguments. To register and find out the current
4971 settings for extended allocation functions, the following
4972 functions are provided:
4974 CRYPTO_set_mem_ex_functions
4975 CRYPTO_set_locked_mem_ex_functions
4976 CRYPTO_get_mem_ex_functions
4977 CRYPTO_get_locked_mem_ex_functions
4979 These work the same way as CRYPTO_set_mem_functions and friends.
4980 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4981 extended allocation function is enabled.
4982 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4983 a conventional allocation function is enabled.
4984 [Richard Levitte, Bodo Moeller]
4986 *) Finish off removing the remaining LHASH function pointer casts.
4987 There should no longer be any prototype-casting required when using
4988 the LHASH abstraction, and any casts that remain are "bugs". See
4989 the callback types and macros at the head of lhash.h for details
4990 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4993 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4994 If /dev/[u]random devices are not available or do not return enough
4995 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4997 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4998 /etc/entropy will be queried once each in this sequence, quering stops
4999 when enough entropy was collected without querying more sockets.
5002 *) Change the Unix RAND_poll() variant to be able to poll several
5003 random devices, as specified by DEVRANDOM, until a sufficient amount
5004 of data has been collected. We spend at most 10 ms on each file
5005 (select timeout) and read in non-blocking mode. DEVRANDOM now
5006 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5007 (previously it was just the string "/dev/urandom"), so on typical
5008 platforms the 10 ms delay will never occur.
5009 Also separate out the Unix variant to its own file, rand_unix.c.
5010 For VMS, there's a currently-empty rand_vms.c.
5013 *) Move OCSP client related routines to ocsp_cl.c. These
5014 provide utility functions which an application needing
5015 to issue a request to an OCSP responder and analyse the
5016 response will typically need: as opposed to those which an
5017 OCSP responder itself would need which will be added later.
5019 OCSP_request_sign() signs an OCSP request with an API similar
5020 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5021 response. OCSP_response_get1_basic() extracts basic response
5022 from response. OCSP_resp_find_status(): finds and extracts status
5023 information from an OCSP_CERTID structure (which will be created
5024 when the request structure is built). These are built from lower
5025 level functions which work on OCSP_SINGLERESP structures but
5026 wont normally be used unless the application wishes to examine
5027 extensions in the OCSP response for example.
5029 Replace nonce routines with a pair of functions.
5030 OCSP_request_add1_nonce() adds a nonce value and optionally
5031 generates a random value. OCSP_check_nonce() checks the
5032 validity of the nonce in an OCSP response.
5035 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5036 This doesn't copy the supplied OCSP_CERTID and avoids the
5037 need to free up the newly created id. Change return type
5038 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5039 This can then be used to add extensions to the request.
5040 Deleted OCSP_request_new(), since most of its functionality
5041 is now in OCSP_REQUEST_new() (and the case insensitive name
5042 clash) apart from the ability to set the request name which
5043 will be added elsewhere.
5046 *) Update OCSP API. Remove obsolete extensions argument from
5047 various functions. Extensions are now handled using the new
5048 OCSP extension code. New simple OCSP HTTP function which
5049 can be used to send requests and parse the response.
5052 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5053 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5054 uses the special reorder version of SET OF to sort the attributes
5055 and reorder them to match the encoded order. This resolves a long
5056 standing problem: a verify on a PKCS7 structure just after signing
5057 it used to fail because the attribute order did not match the
5058 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5059 it uses the received order. This is necessary to tolerate some broken
5060 software that does not order SET OF. This is handled by encoding
5061 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5062 to produce the required SET OF.
5065 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5066 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5067 files to get correct declarations of the ASN.1 item variables.
5070 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5071 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5072 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5073 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5074 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5075 ASN1_ITEM and no wrapper functions.
5078 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5079 replace the old function pointer based I/O routines. Change most of
5080 the *_d2i_bio() and *_d2i_fp() functions to use these.
5083 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5084 lines, recognice more "algorithms" that can be deselected, and make
5085 it complain about algorithm deselection that isn't recognised.
5088 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5089 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5090 to use new functions. Add NO_ASN1_OLD which can be set to remove
5091 some old style ASN1 functions: this can be used to determine if old
5092 code will still work when these eventually go away.
5095 *) New extension functions for OCSP structures, these follow the
5096 same conventions as certificates and CRLs.
5099 *) New function X509V3_add1_i2d(). This automatically encodes and
5100 adds an extension. Its behaviour can be customised with various
5101 flags to append, replace or delete. Various wrappers added for
5102 certifcates and CRLs.
5105 *) Fix to avoid calling the underlying ASN1 print routine when
5106 an extension cannot be parsed. Correct a typo in the
5107 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5110 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5111 entries for variables.
5114 *) Add functionality to apps/openssl.c for detecting locking
5115 problems: As the program is single-threaded, all we have
5116 to do is register a locking callback using an array for
5117 storing which locks are currently held by the program.
5120 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5121 SSL_get_ex_data_X509_STORE_idx(), which is used in
5122 ssl_verify_cert_chain() and thus can be called at any time
5123 during TLS/SSL handshakes so that thread-safety is essential.
5124 Unfortunately, the ex_data design is not at all suited
5125 for multi-threaded use, so it probably should be abolished.
5128 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5129 [Broadcom, tweaked and integrated by Geoff Thorpe]
5131 *) Move common extension printing code to new function
5132 X509V3_print_extensions(). Reorganise OCSP print routines and
5133 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5136 *) New function X509_signature_print() to remove duplication in some
5140 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5141 set (this was treated exactly the same as SET OF previously). This
5142 is used to reorder the STACK representing the structure to match the
5143 encoding. This will be used to get round a problem where a PKCS7
5144 structure which was signed could not be verified because the STACK
5145 order did not reflect the encoded order.
5148 *) Reimplement the OCSP ASN1 module using the new code.
5151 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5152 for its ASN1 operations. The old style function pointers still exist
5153 for now but they will eventually go away.
5156 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5157 completely replaces the old ASN1 functionality with a table driven
5158 encoder and decoder which interprets an ASN1_ITEM structure describing
5159 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5160 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5161 has also been converted to the new form.
5164 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5165 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5166 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5167 for negative moduli.
5170 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5171 of not touching the result's sign bit.
5174 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5178 *) Changed the LHASH code to use prototypes for callbacks, and created
5179 macros to declare and implement thin (optionally static) functions
5180 that provide type-safety and avoid function pointer casting for the
5181 type-specific callbacks.
5184 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5186 [Veers Staats <staatsvr@asc.hpc.mil>,
5187 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5189 *) Reformat the FAQ so the different questions and answers can be divided
5190 in sections depending on the subject.
5193 *) Have the zlib compression code load ZLIB.DLL dynamically under
5197 *) New function BN_mod_sqrt for computing square roots modulo a prime
5198 (using the probabilistic Tonelli-Shanks algorithm unless
5199 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5200 be handled deterministically).
5201 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5203 *) Make BN_mod_inverse faster by explicitly handling small quotients
5204 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5205 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5208 *) New function BN_kronecker.
5211 *) Fix BN_gcd so that it works on negative inputs; the result is
5212 positive unless both parameters are zero.
5213 Previously something reasonably close to an infinite loop was
5214 possible because numbers could be growing instead of shrinking
5215 in the implementation of Euclid's algorithm.
5218 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5219 sign of the number in question.
5221 Fix BN_is_word(a,w) to work correctly for w == 0.
5223 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5224 because its test if the absolute value of 'a' equals 'w'.
5225 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5226 it exists mostly for use in the implementations of BN_is_zero(),
5227 BN_is_one(), and BN_is_word().
5230 *) New function BN_swap.
5233 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5234 the exponentiation functions are more likely to produce reasonable
5235 results on negative inputs.
5238 *) Change BN_mod_mul so that the result is always non-negative.
5239 Previously, it could be negative if one of the factors was negative;
5240 I don't think anyone really wanted that behaviour.
5243 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5244 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5245 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5246 and add new functions:
5255 BN_mod_lshift1_quick
5259 These functions always generate non-negative results.
5261 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5262 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5264 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5265 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5266 be reduced modulo m.
5267 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5270 The following entry accidentily appeared in the CHANGES file
5271 distributed with OpenSSL 0.9.7. The modifications described in
5272 it do *not* apply to OpenSSL 0.9.7.
5274 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5275 was actually never needed) and in BN_mul(). The removal in BN_mul()
5276 required a small change in bn_mul_part_recursive() and the addition
5277 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5278 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5279 bn_sub_words() and bn_add_words() except they take arrays with
5284 *) In 'openssl passwd', verify passwords read from the terminal
5285 unless the '-salt' option is used (which usually means that
5286 verification would just waste user's time since the resulting
5287 hash is going to be compared with some given password hash)
5288 or the new '-noverify' option is used.
5290 This is an incompatible change, but it does not affect
5291 non-interactive use of 'openssl passwd' (passwords on the command
5292 line, '-stdin' option, '-in ...' option) and thus should not
5296 *) Remove all references to RSAref, since there's no more need for it.
5299 *) Make DSO load along a path given through an environment variable
5300 (SHLIB_PATH) with shl_load().
5303 *) Constify the ENGINE code as a result of BIGNUM constification.
5304 Also constify the RSA code and most things related to it. In a
5305 few places, most notable in the depth of the ASN.1 code, ugly
5306 casts back to non-const were required (to be solved at a later
5310 *) Make it so the openssl application has all engines loaded by default.
5313 *) Constify the BIGNUM routines a little more.
5316 *) Add the following functions:
5318 ENGINE_load_cswift()
5320 ENGINE_load_atalla()
5322 ENGINE_load_builtin_engines()
5324 That way, an application can itself choose if external engines that
5325 are built-in in OpenSSL shall ever be used or not. The benefit is
5326 that applications won't have to be linked with libdl or other dso
5327 libraries unless it's really needed.
5329 Changed 'openssl engine' to load all engines on demand.
5330 Changed the engine header files to avoid the duplication of some
5331 declarations (they differed!).
5334 *) 'openssl engine' can now list capabilities.
5337 *) Better error reporting in 'openssl engine'.
5340 *) Never call load_dh_param(NULL) in s_server.
5343 *) Add engine application. It can currently list engines by name and
5344 identity, and test if they are actually available.
5347 *) Improve RPM specification file by forcing symbolic linking and making
5348 sure the installed documentation is also owned by root.root.
5349 [Damien Miller <djm@mindrot.org>]
5351 *) Give the OpenSSL applications more possibilities to make use of
5352 keys (public as well as private) handled by engines.
5355 *) Add OCSP code that comes from CertCo.
5358 *) Add VMS support for the Rijndael code.
5361 *) Added untested support for Nuron crypto accelerator.
5364 *) Add support for external cryptographic devices. This code was
5365 previously distributed separately as the "engine" branch.
5366 [Geoff Thorpe, Richard Levitte]
5368 *) Rework the filename-translation in the DSO code. It is now possible to
5369 have far greater control over how a "name" is turned into a filename
5370 depending on the operating environment and any oddities about the
5371 different shared library filenames on each system.
5374 *) Support threads on FreeBSD-elf in Configure.
5377 *) Fix for SHA1 assembly problem with MASM: it produces
5378 warnings about corrupt line number information when assembling
5379 with debugging information. This is caused by the overlapping
5381 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5384 NCONF_get_number() has no error checking at all. As a replacement,
5385 NCONF_get_number_e() is defined (_e for "error checking") and is
5386 promoted strongly. The old NCONF_get_number is kept around for
5387 binary backward compatibility.
5388 Make it possible for methods to load from something other than a BIO,
5389 by providing a function pointer that is given a name instead of a BIO.
5390 For example, this could be used to load configuration data from an
5394 *) Fix for non blocking accept BIOs. Added new I/O special reason
5395 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5396 with non blocking I/O was not possible because no retry code was
5397 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5401 *) Added the beginnings of Rijndael support.
5404 *) Fix for bug in DirectoryString mask setting. Add support for
5405 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5406 to allow certificate printing to more controllable, additional
5407 'certopt' option to 'x509' to allow new printing options to be
5411 *) Clean old EAY MD5 hack from e_os.h.
5414 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5416 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5417 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5418 [Joe Orton, Steve Henson]
5420 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5422 *) Fix additional bug revealed by the NISCC test suite:
5424 Stop bug triggering large recursion when presented with
5425 certain ASN.1 tags (CVE-2003-0851)
5428 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5430 *) Fix various bugs revealed by running the NISCC test suite:
5432 Stop out of bounds reads in the ASN1 code when presented with
5433 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5435 If verify callback ignores invalid public key errors don't try to check
5436 certificate signature with the NULL public key.
5440 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5441 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5445 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5446 extra data after the compression methods not only for TLS 1.0
5447 but also for SSL 3.0 (as required by the specification).
5448 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5450 *) Change X509_certificate_type() to mark the key as exported/exportable
5451 when it's 512 *bits* long, not 512 bytes.
5454 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5456 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5457 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5458 a protocol version number mismatch like a decryption error
5459 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5462 *) Turn on RSA blinding by default in the default implementation
5463 to avoid a timing attack. Applications that don't want it can call
5464 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5465 They would be ill-advised to do so in most cases.
5466 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5468 *) Change RSA blinding code so that it works when the PRNG is not
5469 seeded (in this case, the secret RSA exponent is abused as
5470 an unpredictable seed -- if it is not unpredictable, there
5471 is no point in blinding anyway). Make RSA blinding thread-safe
5472 by remembering the creator's thread ID in rsa->blinding and
5473 having all other threads use local one-time blinding factors
5474 (this requires more computation than sharing rsa->blinding, but
5475 avoids excessive locking; and if an RSA object is not shared
5476 between threads, blinding will still be very fast).
5479 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5481 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5482 via timing by performing a MAC computation even if incorrrect
5483 block cipher padding has been found. This is a countermeasure
5484 against active attacks where the attacker has to distinguish
5485 between bad padding and a MAC verification error. (CVE-2003-0078)
5487 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5488 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5489 Martin Vuagnoux (EPFL, Ilion)]
5491 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5493 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5494 memory from it's contents. This is done with a counter that will
5495 place alternating values in each byte. This can be used to solve
5496 two issues: 1) the removal of calls to memset() by highly optimizing
5497 compilers, and 2) cleansing with other values than 0, since those can
5498 be read through on certain media, for example a swap space on disk.
5501 *) Bugfix: client side session caching did not work with external caching,
5502 because the session->cipher setting was not restored when reloading
5503 from the external cache. This problem was masked, when
5504 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5505 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5508 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5509 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5510 [Zeev Lieber <zeev-l@yahoo.com>]
5512 *) Undo an undocumented change introduced in 0.9.6e which caused
5513 repeated calls to OpenSSL_add_all_ciphers() and
5514 OpenSSL_add_all_digests() to be ignored, even after calling
5518 *) Change the default configuration reader to deal with last line not
5519 being properly terminated.
5522 *) Change X509_NAME_cmp() so it applies the special rules on handling
5523 DN values that are of type PrintableString, as well as RDNs of type
5524 emailAddress where the value has the type ia5String.
5525 [stefank@valicert.com via Richard Levitte]
5527 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5528 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5529 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5530 the bitwise-OR of the two for use by the majority of applications
5531 wanting this behaviour, and update the docs. The documented
5532 behaviour and actual behaviour were inconsistent and had been
5533 changing anyway, so this is more a bug-fix than a behavioural
5535 [Geoff Thorpe, diagnosed by Nadav Har'El]
5537 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5538 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5541 *) Fix initialization code race conditions in
5542 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5543 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5544 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5545 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5546 ssl2_get_cipher_by_char(),
5547 ssl3_get_cipher_by_char().
5548 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5550 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5551 the cached sessions are flushed, as the remove_cb() might use ex_data
5552 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5553 (see [openssl.org #212]).
5554 [Geoff Thorpe, Lutz Jaenicke]
5556 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5557 length, instead of the encoding length to d2i_ASN1_OBJECT.
5560 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5562 *) [In 0.9.6g-engine release:]
5563 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5564 [Lynn Gazis <lgazis@rainbow.com>]
5566 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5568 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5569 and get fix the header length calculation.
5570 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5571 Alon Kantor <alonk@checkpoint.com> (and others),
5574 *) Use proper error handling instead of 'assertions' in buffer
5575 overflow checks added in 0.9.6e. This prevents DoS (the
5576 assertions could call abort()).
5577 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5579 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5581 *) Add various sanity checks to asn1_get_length() to reject
5582 the ASN1 length bytes if they exceed sizeof(long), will appear
5583 negative or the content length exceeds the length of the
5585 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5587 *) Fix cipher selection routines: ciphers without encryption had no flags
5588 for the cipher strength set and where therefore not handled correctly
5589 by the selection routines (PR #130).
5592 *) Fix EVP_dsa_sha macro.
5596 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5597 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5598 that was added in OpenSSL 0.9.6d.
5600 As the countermeasure turned out to be incompatible with some
5601 broken SSL implementations, the new option is part of SSL_OP_ALL.
5602 SSL_OP_ALL is usually employed when compatibility with weird SSL
5603 implementations is desired (e.g. '-bugs' option to 's_client' and
5604 's_server'), so the new option is automatically set in many
5608 *) Changes in security patch:
5610 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5611 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5612 Air Force Materiel Command, USAF, under agreement number
5615 *) Add various sanity checks to asn1_get_length() to reject
5616 the ASN1 length bytes if they exceed sizeof(long), will appear
5617 negative or the content length exceeds the length of the
5618 supplied buffer. (CVE-2002-0659)
5619 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5621 *) Assertions for various potential buffer overflows, not known to
5623 [Ben Laurie (CHATS)]
5625 *) Various temporary buffers to hold ASCII versions of integers were
5626 too small for 64 bit platforms. (CVE-2002-0655)
5627 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5629 *) Remote buffer overflow in SSL3 protocol - an attacker could
5630 supply an oversized session ID to a client. (CVE-2002-0656)
5631 [Ben Laurie (CHATS)]
5633 *) Remote buffer overflow in SSL2 protocol - an attacker could
5634 supply an oversized client master key. (CVE-2002-0656)
5635 [Ben Laurie (CHATS)]
5637 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5639 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5640 encoded as NULL) with id-dsa-with-sha1.
5641 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5643 *) Check various X509_...() return values in apps/req.c.
5644 [Nils Larsch <nla@trustcenter.de>]
5646 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5647 an end-of-file condition would erronously be flagged, when the CRLF
5648 was just at the end of a processed block. The bug was discovered when
5649 processing data through a buffering memory BIO handing the data to a
5650 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5651 <ptsekov@syntrex.com> and Nedelcho Stanev.
5654 *) Implement a countermeasure against a vulnerability recently found
5655 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5656 before application data chunks to avoid the use of known IVs
5657 with data potentially chosen by the attacker.
5660 *) Fix length checks in ssl3_get_client_hello().
5663 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5664 to prevent ssl3_read_internal() from incorrectly assuming that
5665 ssl3_read_bytes() found application data while handshake
5666 processing was enabled when in fact s->s3->in_read_app_data was
5667 merely automatically cleared during the initial handshake.
5668 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5670 *) Fix object definitions for Private and Enterprise: they were not
5671 recognized in their shortname (=lowercase) representation. Extend
5672 obj_dat.pl to issue an error when using undefined keywords instead
5673 of silently ignoring the problem (Svenning Sorensen
5674 <sss@sss.dnsalias.net>).
5677 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5678 generators, i.e. generators other than 2 and 5. (Previously, the
5679 code did not properly initialise the 'add' and 'rem' values to
5680 BN_generate_prime().)
5682 In the new general case, we do not insist that 'generator' is
5683 actually a primitive root: This requirement is rather pointless;
5684 a generator of the order-q subgroup is just as good, if not
5688 *) Map new X509 verification errors to alerts. Discovered and submitted by
5689 Tom Wu <tom@arcot.com>.
5692 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5693 returning non-zero before the data has been completely received
5694 when using non-blocking I/O.
5695 [Bodo Moeller; problem pointed out by John Hughes]
5697 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5698 [Ben Laurie, Lutz Jaenicke]
5700 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5701 Yoram Zahavi <YoramZ@gilian.com>).
5704 *) Add information about CygWin 1.3 and on, and preserve proper
5705 configuration for the versions before that.
5706 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5708 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5709 check whether we deal with a copy of a session and do not delete from
5710 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5711 <izhar@checkpoint.com>.
5714 *) Do not store session data into the internal session cache, if it
5715 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5716 flag is set). Proposed by Aslam <aslam@funk.com>.
5719 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5723 *) [In 0.9.6d-engine release:]
5724 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5725 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5727 *) Add the configuration target linux-s390x.
5728 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5730 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5731 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5732 variable as an indication that a ClientHello message has been
5733 received. As the flag value will be lost between multiple
5734 invocations of ssl3_accept when using non-blocking I/O, the
5735 function may not be aware that a handshake has actually taken
5736 place, thus preventing a new session from being added to the
5739 To avoid this problem, we now set s->new_session to 2 instead of
5740 using a local variable.
5741 [Lutz Jaenicke, Bodo Moeller]
5743 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5744 if the SSL_R_LENGTH_MISMATCH error is detected.
5745 [Geoff Thorpe, Bodo Moeller]
5747 *) New 'shared_ldflag' column in Configure platform table.
5750 *) Fix EVP_CIPHER_mode macro.
5751 ["Dan S. Camper" <dan@bti.net>]
5753 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5754 type, we must throw them away by setting rr->length to 0.
5755 [D P Chang <dpc@qualys.com>]
5757 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5759 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5760 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5761 worked incorrectly for those cases where range = 10..._2 and
5762 3*range is two bits longer than range.)
5765 *) Only add signing time to PKCS7 structures if it is not already
5769 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5770 OBJ_ld_ce should be OBJ_id_ce.
5771 Also some ip-pda OIDs in crypto/objects/objects.txt were
5772 incorrect (cf. RFC 3039).
5773 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5775 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5776 returns early because it has nothing to do.
5777 [Andy Schneider <andy.schneider@bjss.co.uk>]
5779 *) [In 0.9.6c-engine release:]
5780 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5781 [Andy Schneider <andy.schneider@bjss.co.uk>]
5783 *) [In 0.9.6c-engine release:]
5784 Add support for Cryptographic Appliance's keyserver technology.
5785 (Use engine 'keyclient')
5786 [Cryptographic Appliances and Geoff Thorpe]
5788 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5789 is called via tools/c89.sh because arguments have to be
5790 rearranged (all '-L' options must appear before the first object
5792 [Richard Shapiro <rshapiro@abinitio.com>]
5794 *) [In 0.9.6c-engine release:]
5795 Add support for Broadcom crypto accelerator cards, backported
5797 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5799 *) [In 0.9.6c-engine release:]
5800 Add support for SureWare crypto accelerator cards from
5801 Baltimore Technologies. (Use engine 'sureware')
5802 [Baltimore Technologies and Mark Cox]
5804 *) [In 0.9.6c-engine release:]
5805 Add support for crypto accelerator cards from Accelerated
5806 Encryption Processing, www.aep.ie. (Use engine 'aep')
5807 [AEP Inc. and Mark Cox]
5809 *) Add a configuration entry for gcc on UnixWare.
5810 [Gary Benson <gbenson@redhat.com>]
5812 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5813 messages are stored in a single piece (fixed-length part and
5814 variable-length part combined) and fix various bugs found on the way.
5817 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5818 instead. BIO_gethostbyname() does not know what timeouts are
5819 appropriate, so entries would stay in cache even when they have
5821 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5823 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5824 faced with a pathologically small ClientHello fragment that does
5825 not contain client_version: Instead of aborting with an error,
5826 simply choose the highest available protocol version (i.e.,
5827 TLS 1.0 unless it is disabled). In practice, ClientHello
5828 messages are never sent like this, but this change gives us
5829 strictly correct behaviour at least for TLS.
5832 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5833 never resets s->method to s->ctx->method when called from within
5834 one of the SSL handshake functions.
5835 [Bodo Moeller; problem pointed out by Niko Baric]
5837 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5838 (sent using the client's version number) if client_version is
5839 smaller than the protocol version in use. Also change
5840 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5841 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5842 the client will at least see that alert.
5845 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5849 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5850 client receives HelloRequest while in a handshake.
5851 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5853 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5854 should end in 'break', not 'goto end' which circuments various
5855 cleanups done in state SSL_ST_OK. But session related stuff
5856 must be disabled for SSL_ST_OK in the case that we just sent a
5859 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5860 before just sending a HelloRequest.
5861 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5863 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5864 reveal whether illegal block cipher padding was found or a MAC
5865 verification error occured. (Neither SSLerr() codes nor alerts
5866 are directly visible to potential attackers, but the information
5867 may leak via logfiles.)
5869 Similar changes are not required for the SSL 2.0 implementation
5870 because the number of padding bytes is sent in clear for SSL 2.0,
5871 and the extra bytes are just ignored. However ssl/s2_pkt.c
5872 failed to verify that the purported number of padding bytes is in
5876 *) Add OpenUNIX-8 support including shared libraries
5877 (Boyd Lynn Gerber <gerberb@zenez.com>).
5880 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5881 'wristwatch attack' using huge encoding parameters (cf.
5882 James H. Manger's CRYPTO 2001 paper). Note that the
5883 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5884 encoding parameters and hence was not vulnerable.
5887 *) BN_sqr() bug fix.
5888 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5890 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5891 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5892 followed by modular reduction.
5893 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5895 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5896 equivalent based on BN_pseudo_rand() instead of BN_rand().
5899 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5900 This function was broken, as the check for a new client hello message
5901 to handle SGC did not allow these large messages.
5902 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5905 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5908 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5909 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5912 *) Rework the configuration and shared library support for Tru64 Unix.
5913 The configuration part makes use of modern compiler features and
5914 still retains old compiler behavior for those that run older versions
5915 of the OS. The shared library support part includes a variant that
5916 uses the RPATH feature, and is available through the special
5917 configuration target "alpha-cc-rpath", which will never be selected
5919 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5921 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5922 with the same message size as in ssl3_get_certificate_request().
5923 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5924 messages might inadvertently be reject as too long.
5925 [Petr Lampa <lampa@fee.vutbr.cz>]
5927 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5930 *) Modified SSL library such that the verify_callback that has been set
5931 specificly for an SSL object with SSL_set_verify() is actually being
5932 used. Before the change, a verify_callback set with this function was
5933 ignored and the verify_callback() set in the SSL_CTX at the time of
5934 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5935 to allow the necessary settings.
5938 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5939 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5940 done automatically (in contradiction to the requirements of the C
5941 standard). This made problems when used from OpenSSH.
5944 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5945 dh->length and always used
5947 BN_rand_range(priv_key, dh->p).
5949 BN_rand_range() is not necessary for Diffie-Hellman, and this
5950 specific range makes Diffie-Hellman unnecessarily inefficient if
5951 dh->length (recommended exponent length) is much smaller than the
5952 length of dh->p. We could use BN_rand_range() if the order of
5953 the subgroup was stored in the DH structure, but we only have
5958 BN_rand(priv_key, l, ...)
5960 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5966 RSA_eay_public_encrypt
5967 RSA_eay_private_decrypt
5968 RSA_eay_private_encrypt (signing)
5969 RSA_eay_public_decrypt (signature verification)
5971 (default implementations for RSA_public_encrypt,
5972 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5973 always reject numbers >= n.
5976 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5977 to synchronize access to 'locking_thread'. This is necessary on
5978 systems where access to 'locking_thread' (an 'unsigned long'
5979 variable) is not atomic.
5982 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5983 *before* setting the 'crypto_lock_rand' flag. The previous code had
5984 a race condition if 0 is a valid thread ID.
5985 [Travis Vitek <vitek@roguewave.com>]
5987 *) Add support for shared libraries under Irix.
5988 [Albert Chin-A-Young <china@thewrittenword.com>]
5990 *) Add configuration option to build on Linux on both big-endian and
5992 [Ralf Baechle <ralf@uni-koblenz.de>]
5994 *) Add the possibility to create shared libraries on HP-UX.
5997 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5999 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6000 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6001 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6002 PRNG state recovery was possible based on the output of
6003 one PRNG request appropriately sized to gain knowledge on
6004 'md' followed by enough consecutive 1-byte PRNG requests
6005 to traverse all of 'state'.
6007 1. When updating 'md_local' (the current thread's copy of 'md')
6008 during PRNG output generation, hash all of the previous
6009 'md_local' value, not just the half used for PRNG output.
6011 2. Make the number of bytes from 'state' included into the hash
6012 independent from the number of PRNG bytes requested.
6014 The first measure alone would be sufficient to avoid
6015 Markku-Juhani's attack. (Actually it had never occurred
6016 to me that the half of 'md_local' used for chaining was the
6017 half from which PRNG output bytes were taken -- I had always
6018 assumed that the secret half would be used.) The second
6019 measure makes sure that additional data from 'state' is never
6020 mixed into 'md_local' in small portions; this heuristically
6021 further strengthens the PRNG.
6024 *) Fix crypto/bn/asm/mips3.s.
6027 *) When only the key is given to "enc", the IV is undefined. Print out
6028 an error message in this case.
6031 *) Handle special case when X509_NAME is empty in X509 printing routines.
6034 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6035 positive and less than q.
6038 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6039 used: it isn't thread safe and the add_lock_callback should handle
6041 [Paul Rose <Paul.Rose@bridge.com>]
6043 *) Verify that incoming data obeys the block size in
6044 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6048 [Ulf Möller, Bodo Möller]
6050 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6051 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6052 when fixing the server behaviour for backwards-compatible 'client
6053 hello' messages. (Note that the attack is impractical against
6054 SSL 3.0 and TLS 1.0 anyway because length and version checking
6055 means that the probability of guessing a valid ciphertext is
6056 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6059 Before 0.9.5, the countermeasure (hide the error by generating a
6060 random 'decryption result') did not work properly because
6061 ERR_clear_error() was missing, meaning that SSL_get_error() would
6062 detect the supposedly ignored error.
6064 Both problems are now fixed.
6067 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6068 (previously it was 1024).
6071 *) Fix for compatibility mode trust settings: ignore trust settings
6072 unless some valid trust or reject settings are present.
6075 *) Fix for blowfish EVP: its a variable length cipher.
6078 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6079 parameters in DSA public key structures and return an error in the
6080 DSA routines if parameters are absent.
6083 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6084 in the current directory if neither $RANDFILE nor $HOME was set.
6085 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6086 caused some confusion to Windows users who haven't defined $HOME.
6087 Thus RAND_file_name() is changed again: e_os.h can define a
6088 DEFAULT_HOME, which will be used if $HOME is not set.
6089 For Windows, we use "C:"; on other platforms, we still require
6090 environment variables.
6092 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6093 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6094 having multiple threads call RAND_poll() concurrently.
6097 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6098 combination of a flag and a thread ID variable.
6099 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6100 flag), *other* threads can enter ssleay_add_bytes without obeying
6101 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6102 that they do not hold after the first thread unsets add_do_not_lock).
6105 *) Change bctest again: '-x' expressions are not available in all
6109 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6111 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6112 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6114 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6115 the default extension for executables, if any. Also, make the perl
6116 scripts that use symlink() to test if it really exists and use "cp"
6117 if it doesn't. All this made OpenSSL compilable and installable in
6121 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6122 If SEQUENCE is length is indefinite just set c->slen to the total
6123 amount of data available.
6124 [Steve Henson, reported by shige@FreeBSD.org]
6125 [This change does not apply to 0.9.7.]
6127 *) Change bctest to avoid here-documents inside command substitution
6128 (workaround for FreeBSD /bin/sh bug).
6129 For compatibility with Ultrix, avoid shell functions (introduced
6130 in the bctest version that searches along $PATH).
6133 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6134 with des_encrypt() defined on some operating systems, like Solaris
6138 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6139 On the Importance of Eliminating Errors in Cryptographic
6140 Computations, J. Cryptology 14 (2001) 2, 101-119,
6141 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6144 *) MIPS assembler BIGNUM division bug fix.
6147 *) Disabled incorrect Alpha assembler code.
6150 *) Fix PKCS#7 decode routines so they correctly update the length
6151 after reading an EOC for the EXPLICIT tag.
6153 [This change does not apply to 0.9.7.]
6155 *) Fix bug in PKCS#12 key generation routines. This was triggered
6156 if a 3DES key was generated with a 0 initial byte. Include
6157 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6158 (but broken) behaviour.
6161 *) Enhance bctest to search for a working bc along $PATH and print
6163 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6165 *) Fix memory leaks in err.c: free err_data string if necessary;
6166 don't write to the wrong index in ERR_set_error_data.
6169 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6173 *) Replace rdtsc with _emit statements for VC++ version 5.
6174 [Jeremy Cooper <jeremy@baymoo.org>]
6176 *) Make it possible to reuse SSLv2 sessions.
6179 *) In copy_email() check for >= 0 as a return value for
6180 X509_NAME_get_index_by_NID() since 0 is a valid index.
6181 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6183 *) Avoid coredump with unsupported or invalid public keys by checking if
6184 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6185 PKCS7_verify() fails with non detached data.
6188 *) Don't use getenv in library functions when run as setuid/setgid.
6189 New function OPENSSL_issetugid().
6192 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6193 due to incorrect handling of multi-threading:
6195 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6197 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6199 3. Count how many times MemCheck_off() has been called so that
6200 nested use can be treated correctly. This also avoids
6201 inband-signalling in the previous code (which relied on the
6202 assumption that thread ID 0 is impossible).
6205 *) Add "-rand" option also to s_client and s_server.
6208 *) Fix CPU detection on Irix 6.x.
6209 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6210 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6212 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6215 [This change does not apply to 0.9.7.]
6217 *) Use the cached encoding of an X509_NAME structure rather than
6218 copying it. This is apparently the reason for the libsafe "errors"
6219 but the code is actually correct.
6222 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6223 Bleichenbacher's DSA attack.
6224 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6225 to be set and top=0 forces the highest bit to be set; top=-1 is new
6226 and leaves the highest bit random.
6227 [Ulf Moeller, Bodo Moeller]
6229 *) In the NCONF_...-based implementations for CONF_... queries
6230 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6231 a temporary CONF structure with the data component set to NULL
6232 (which gives segmentation faults in lh_retrieve).
6233 Instead, use NULL for the CONF pointer in CONF_get_string and
6234 CONF_get_number (which may use environment variables) and directly
6235 return NULL from CONF_get_section.
6238 *) Fix potential buffer overrun for EBCDIC.
6241 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6242 keyUsage if basicConstraints absent for a CA.
6245 *) Make SMIME_write_PKCS7() write mail header values with a format that
6246 is more generally accepted (no spaces before the semicolon), since
6247 some programs can't parse those values properly otherwise. Also make
6248 sure BIO's that break lines after each write do not create invalid
6252 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6253 macros previously used would not encode an empty SEQUENCE OF
6254 and break the signature.
6256 [This change does not apply to 0.9.7.]
6258 *) Zero the premaster secret after deriving the master secret in
6262 *) Add some EVP_add_digest_alias registrations (as found in
6263 OpenSSL_add_all_digests()) to SSL_library_init()
6264 aka OpenSSL_add_ssl_algorithms(). This provides improved
6265 compatibility with peers using X.509 certificates
6266 with unconventional AlgorithmIdentifier OIDs.
6269 *) Fix for Irix with NO_ASM.
6270 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6272 *) ./config script fixes.
6273 [Ulf Moeller, Richard Levitte]
6275 *) Fix 'openssl passwd -1'.
6278 *) Change PKCS12_key_gen_asc() so it can cope with non null
6279 terminated strings whose length is passed in the passlen
6280 parameter, for example from PEM callbacks. This was done
6281 by adding an extra length parameter to asc2uni().
6282 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6284 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6285 call failed, free the DSA structure.
6288 *) Fix to uni2asc() to cope with zero length Unicode strings.
6289 These are present in some PKCS#12 files.
6292 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6293 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6294 when writing a 32767 byte record.
6295 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6297 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6298 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6300 (RSA objects have a reference count access to which is protected
6301 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6302 so they are meant to be shared between threads.)
6303 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6304 "Reddie, Steven" <Steven.Reddie@ca.com>]
6306 *) Fix a deadlock in CRYPTO_mem_leaks().
6309 *) Use better test patterns in bntest.
6312 *) rand_win.c fix for Borland C.
6315 *) BN_rshift bugfix for n == 0.
6318 *) Add a 'bctest' script that checks for some known 'bc' bugs
6319 so that 'make test' does not abort just because 'bc' is broken.
6322 *) Store verify_result within SSL_SESSION also for client side to
6323 avoid potential security hole. (Re-used sessions on the client side
6324 always resulted in verify_result==X509_V_OK, not using the original
6325 result of the server certificate verification.)
6328 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6329 SSL3_RT_APPLICATION_DATA, return 0.
6330 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6334 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6335 releases, have been re-implemented by renaming the previous
6336 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6337 and ssl3_read_internal, respectively, and adding 'peek' parameters
6338 to them. The new ssl[23]_{read,peek} functions are calls to
6339 ssl[23]_read_internal with the 'peek' flag set appropriately.
6340 A 'peek' parameter has also been added to ssl3_read_bytes, which
6341 does the actual work for ssl3_read_internal.
6344 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6345 the method-specific "init()" handler. Also clean up ex_data after
6346 calling the method-specific "finish()" handler. Previously, this was
6347 happening the other way round.
6350 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6351 The previous value, 12, was not always sufficient for BN_mod_exp().
6354 *) Make sure that shared libraries get the internal name engine with
6355 the full version number and not just 0. This should mark the
6356 shared libraries as not backward compatible. Of course, this should
6357 be changed again when we can guarantee backward binary compatibility.
6360 *) Fix typo in get_cert_by_subject() in by_dir.c
6361 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6363 *) Rework the system to generate shared libraries:
6365 - Make note of the expected extension for the shared libraries and
6366 if there is a need for symbolic links from for example libcrypto.so.0
6367 to libcrypto.so.0.9.7. There is extended info in Configure for
6370 - Make as few rebuilds of the shared libraries as possible.
6372 - Still avoid linking the OpenSSL programs with the shared libraries.
6374 - When installing, install the shared libraries separately from the
6378 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6380 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6381 and not in SSL_clear because the latter is also used by the
6382 accept/connect functions; previously, the settings made by
6383 SSL_set_read_ahead would be lost during the handshake.
6384 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6386 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6387 Previously, it would create entries for disableed algorithms no
6391 *) Added several new manual pages for SSL_* function.
6394 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6396 *) In ssl23_get_client_hello, generate an error message when faced
6397 with an initial SSL 3.0/TLS record that is too small to contain the
6398 first two bytes of the ClientHello message, i.e. client_version.
6399 (Note that this is a pathologic case that probably has never happened
6400 in real life.) The previous approach was to use the version number
6401 from the record header as a substitute; but our protocol choice
6402 should not depend on that one because it is not authenticated
6403 by the Finished messages.
6406 *) More robust randomness gathering functions for Windows.
6407 [Jeffrey Altman <jaltman@columbia.edu>]
6409 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6410 not set then we don't setup the error code for issuer check errors
6411 to avoid possibly overwriting other errors which the callback does
6412 handle. If an application does set the flag then we assume it knows
6413 what it is doing and can handle the new informational codes
6417 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6418 a general "ANY" type, as such it should be able to decode anything
6419 including tagged types. However it didn't check the class so it would
6420 wrongly interpret tagged types in the same way as their universal
6421 counterpart and unknown types were just rejected. Changed so that the
6422 tagged and unknown types are handled in the same way as a SEQUENCE:
6423 that is the encoding is stored intact. There is also a new type
6424 "V_ASN1_OTHER" which is used when the class is not universal, in this
6425 case we have no idea what the actual type is so we just lump them all
6429 *) On VMS, stdout may very well lead to a file that is written to
6430 in a record-oriented fashion. That means that every write() will
6431 write a separate record, which will be read separately by the
6432 programs trying to read from it. This can be very confusing.
6434 The solution is to put a BIO filter in the way that will buffer
6435 text until a linefeed is reached, and then write everything a
6436 line at a time, so every record written will be an actual line,
6437 not chunks of lines and not (usually doesn't happen, but I've
6438 seen it once) several lines in one record. BIO_f_linebuffer() is
6441 Currently, it's a VMS-only method, because that's where it has
6442 been tested well enough.
6445 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6446 it can return incorrect results.
6447 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6448 but it was in 0.9.6-beta[12].)
6451 *) Disable the check for content being present when verifying detached
6452 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6453 include zero length content when signing messages.
6456 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6457 BIO_ctrl (for BIO pairs).
6460 *) Add DSO method for VMS.
6463 *) Bug fix: Montgomery multiplication could produce results with the
6467 *) Add RPM specification openssl.spec and modify it to build three
6468 packages. The default package contains applications, application
6469 documentation and run-time libraries. The devel package contains
6470 include files, static libraries and function documentation. The
6471 doc package contains the contents of the doc directory. The original
6472 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6475 *) Add a large number of documentation files for many SSL routines.
6476 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6478 *) Add a configuration entry for Sony News 4.
6479 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6481 *) Don't set the two most significant bits to one when generating a
6482 random number < q in the DSA library.
6485 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6486 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6487 the underlying transport is blocking) if a handshake took place.
6488 (The default behaviour is needed by applications such as s_client
6489 and s_server that use select() to determine when to use SSL_read;
6490 but for applications that know in advance when to expect data, it
6491 just makes things more complicated.)
6494 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6498 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6499 work better on such systems.
6500 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6502 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6503 Update PKCS12_parse() so it copies the friendlyName and the
6504 keyid to the certificates aux info.
6507 *) Fix bug in PKCS7_verify() which caused an infinite loop
6508 if there was more than one signature.
6509 [Sven Uszpelkat <su@celocom.de>]
6511 *) Major change in util/mkdef.pl to include extra information
6512 about each symbol, as well as presentig variables as well
6513 as functions. This change means that there's n more need
6514 to rebuild the .num files when some algorithms are excluded.
6517 *) Allow the verify time to be set by an application,
6518 rather than always using the current time.
6521 *) Phase 2 verify code reorganisation. The certificate
6522 verify code now looks up an issuer certificate by a
6523 number of criteria: subject name, authority key id
6524 and key usage. It also verifies self signed certificates
6525 by the same criteria. The main comparison function is
6526 X509_check_issued() which performs these checks.
6528 Lot of changes were necessary in order to support this
6529 without completely rewriting the lookup code.
6531 Authority and subject key identifier are now cached.
6533 The LHASH 'certs' is X509_STORE has now been replaced
6534 by a STACK_OF(X509_OBJECT). This is mainly because an
6535 LHASH can't store or retrieve multiple objects with
6536 the same hash value.
6538 As a result various functions (which were all internal
6539 use only) have changed to handle the new X509_STORE
6540 structure. This will break anything that messed round
6541 with X509_STORE internally.
6543 The functions X509_STORE_add_cert() now checks for an
6544 exact match, rather than just subject name.
6546 The X509_STORE API doesn't directly support the retrieval
6547 of multiple certificates matching a given criteria, however
6548 this can be worked round by performing a lookup first
6549 (which will fill the cache with candidate certificates)
6550 and then examining the cache for matches. This is probably
6551 the best we can do without throwing out X509_LOOKUP
6552 entirely (maybe later...).
6554 The X509_VERIFY_CTX structure has been enhanced considerably.
6556 All certificate lookup operations now go via a get_issuer()
6557 callback. Although this currently uses an X509_STORE it
6558 can be replaced by custom lookups. This is a simple way
6559 to bypass the X509_STORE hackery necessary to make this
6560 work and makes it possible to use more efficient techniques
6561 in future. A very simple version which uses a simple
6562 STACK for its trusted certificate store is also provided
6563 using X509_STORE_CTX_trusted_stack().
6565 The verify_cb() and verify() callbacks now have equivalents
6566 in the X509_STORE_CTX structure.
6568 X509_STORE_CTX also has a 'flags' field which can be used
6569 to customise the verify behaviour.
6572 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6573 excludes S/MIME capabilities.
6576 *) When a certificate request is read in keep a copy of the
6577 original encoding of the signed data and use it when outputing
6578 again. Signatures then use the original encoding rather than
6579 a decoded, encoded version which may cause problems if the
6580 request is improperly encoded.
6583 *) For consistency with other BIO_puts implementations, call
6584 buffer_write(b, ...) directly in buffer_puts instead of calling
6587 In BIO_puts, increment b->num_write as in BIO_write.
6588 [Peter.Sylvester@EdelWeb.fr]
6590 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6591 BN_zero, we may not return a BIGNUM with an array consisting of
6595 *) Avoid calling abort() from within the library when problems are
6596 detected, except if preprocessor symbols have been defined
6597 (such as REF_CHECK, BN_DEBUG etc.).
6600 *) New openssl application 'rsautl'. This utility can be
6601 used for low level RSA operations. DER public key
6602 BIO/fp routines also added.
6605 *) New Configure entry and patches for compiling on QNX 4.
6606 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6608 *) A demo state-machine implementation was sponsored by
6609 Nuron (http://www.nuron.com/) and is now available in
6610 demos/state_machine.
6613 *) New options added to the 'dgst' utility for signature
6614 generation and verification.
6617 *) Unrecognized PKCS#7 content types are now handled via a
6618 catch all ASN1_TYPE structure. This allows unsupported
6619 types to be stored as a "blob" and an application can
6620 encode and decode it manually.
6623 *) Fix various signed/unsigned issues to make a_strex.c
6625 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6627 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6628 length if passed a buffer. ASN1_INTEGER_to_BN failed
6629 if passed a NULL BN and its argument was negative.
6630 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6632 *) Modification to PKCS#7 encoding routines to output definite
6633 length encoding. Since currently the whole structures are in
6634 memory there's not real point in using indefinite length
6635 constructed encoding. However if OpenSSL is compiled with
6636 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6639 *) Added BIO_vprintf() and BIO_vsnprintf().
6642 *) Added more prefixes to parse for in the the strings written
6643 through a logging bio, to cover all the levels that are available
6644 through syslog. The prefixes are now:
6646 PANIC, EMERG, EMR => LOG_EMERG
6647 ALERT, ALR => LOG_ALERT
6648 CRIT, CRI => LOG_CRIT
6649 ERROR, ERR => LOG_ERR
6650 WARNING, WARN, WAR => LOG_WARNING
6651 NOTICE, NOTE, NOT => LOG_NOTICE
6652 INFO, INF => LOG_INFO
6653 DEBUG, DBG => LOG_DEBUG
6655 and as before, if none of those prefixes are present at the
6656 beginning of the string, LOG_ERR is chosen.
6658 On Win32, the LOG_* levels are mapped according to this:
6660 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6661 LOG_WARNING => EVENTLOG_WARNING_TYPE
6662 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6666 *) Made it possible to reconfigure with just the configuration
6667 argument "reconf" or "reconfigure". The command line arguments
6668 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6669 and are retrieved from there when reconfiguring.
6673 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6675 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6678 *) The obj_dat.pl script was messing up the sorting of object
6679 names. The reason was that it compared the quoted version
6680 of strings as a result "OCSP" > "OCSP Signing" because
6681 " > SPACE. Changed script to store unquoted versions of
6682 names and add quotes on output. It was also omitting some
6683 names from the lookup table if they were given a default
6684 value (that is if SN is missing it is given the same
6685 value as LN and vice versa), these are now added on the
6686 grounds that if an object has a name we should be able to
6687 look it up. Finally added warning output when duplicate
6688 short or long names are found.
6691 *) Changes needed for Tandem NSK.
6692 [Scott Uroff <scott@xypro.com>]
6694 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6695 RSA_padding_check_SSLv23(), special padding was never detected
6696 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6697 version rollback attacks was not effective.
6699 In s23_clnt.c, don't use special rollback-attack detection padding
6700 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6701 client; similarly, in s23_srvr.c, don't do the rollback check if
6702 SSL 2.0 is the only protocol enabled in the server.
6705 *) Make it possible to get hexdumps of unprintable data with 'openssl
6706 asn1parse'. By implication, the functions ASN1_parse_dump() and
6707 BIO_dump_indent() are added.
6710 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6711 these print out strings and name structures based on various
6712 flags including RFC2253 support and proper handling of
6713 multibyte characters. Added options to the 'x509' utility
6714 to allow the various flags to be set.
6717 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6718 Also change the functions X509_cmp_current_time() and
6719 X509_gmtime_adj() work with an ASN1_TIME structure,
6720 this will enable certificates using GeneralizedTime in validity
6721 dates to be checked.
6724 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6725 negative public key encodings) on by default,
6726 NO_NEG_PUBKEY_BUG can be set to disable it.
6729 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6730 content octets. An i2c_ASN1_OBJECT is unnecessary because
6731 the encoding can be trivially obtained from the structure.
6734 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6735 not read locks (CRYPTO_r_[un]lock).
6738 *) A first attempt at creating official support for shared
6739 libraries through configuration. I've kept it so the
6740 default is static libraries only, and the OpenSSL programs
6741 are always statically linked for now, but there are
6742 preparations for dynamic linking in place.
6743 This has been tested on Linux and Tru64.
6746 *) Randomness polling function for Win9x, as described in:
6747 Peter Gutmann, Software Generation of Practically Strong
6751 *) Fix so PRNG is seeded in req if using an already existing
6755 *) New options to smime application. -inform and -outform
6756 allow alternative formats for the S/MIME message including
6757 PEM and DER. The -content option allows the content to be
6758 specified separately. This should allow things like Netscape
6759 form signing output easier to verify.
6762 *) Fix the ASN1 encoding of tags using the 'long form'.
6765 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6766 STRING types. These convert content octets to and from the
6767 underlying type. The actual tag and length octets are
6768 already assumed to have been read in and checked. These
6769 are needed because all other string types have virtually
6770 identical handling apart from the tag. By having versions
6771 of the ASN1 functions that just operate on content octets
6772 IMPLICIT tagging can be handled properly. It also allows
6773 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6774 and ASN1_INTEGER are identical apart from the tag.
6777 *) Change the handling of OID objects as follows:
6779 - New object identifiers are inserted in objects.txt, following
6780 the syntax given in objects.README.
6781 - objects.pl is used to process obj_mac.num and create a new
6783 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6786 This is currently kind of a hack, and the perl code in objects.pl
6787 isn't very elegant, but it works as I intended. The simplest way
6788 to check that it worked correctly is to look in obj_dat.h and
6789 check the array nid_objs and make sure the objects haven't moved
6790 around (this is important!). Additions are OK, as well as
6791 consistent name changes.
6794 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6797 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6798 The given file adds to whatever has already been seeded into the
6799 random pool through the RANDFILE configuration file option or
6800 environment variable, or the default random state file.
6803 *) mkstack.pl now sorts each macro group into lexical order.
6804 Previously the output order depended on the order the files
6805 appeared in the directory, resulting in needless rewriting
6809 *) Patches to make OpenSSL compile under Win32 again. Mostly
6810 work arounds for the VC++ problem that it treats func() as
6811 func(void). Also stripped out the parts of mkdef.pl that
6812 added extra typesafe functions: these no longer exist.
6815 *) Reorganisation of the stack code. The macros are now all
6816 collected in safestack.h . Each macro is defined in terms of
6817 a "stack macro" of the form SKM_<name>(type, a, b). The
6818 DEBUG_SAFESTACK is now handled in terms of function casts,
6819 this has the advantage of retaining type safety without the
6820 use of additional functions. If DEBUG_SAFESTACK is not defined
6821 then the non typesafe macros are used instead. Also modified the
6822 mkstack.pl script to handle the new form. Needs testing to see
6823 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6824 the default if no major problems. Similar behaviour for ASN1_SET_OF
6825 and PKCS12_STACK_OF.
6828 *) When some versions of IIS use the 'NET' form of private key the
6829 key derivation algorithm is different. Normally MD5(password) is
6830 used as a 128 bit RC4 key. In the modified case
6831 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6832 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6833 as the old Netscape_RSA functions except they have an additional
6834 'sgckey' parameter which uses the modified algorithm. Also added
6835 an -sgckey command line option to the rsa utility. Thanks to
6836 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6837 algorithm to openssl-dev.
6840 *) The evp_local.h macros were using 'c.##kname' which resulted in
6841 invalid expansion on some systems (SCO 5.0.5 for example).
6842 Corrected to 'c.kname'.
6843 [Phillip Porch <root@theporch.com>]
6845 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6846 a STACK of email addresses from a certificate or request, these look
6847 in the subject name and the subject alternative name extensions and
6848 omit any duplicate addresses.
6851 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6852 This makes DSA verification about 2 % faster.
6855 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6856 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6857 plus overhead for 1024 bit moduli).
6858 This makes exponentiations about 0.5 % faster for 1024 bit
6859 exponents (as measured by "openssl speed rsa2048").
6862 *) Rename memory handling macros to avoid conflicts with other
6864 Malloc => OPENSSL_malloc
6865 Malloc_locked => OPENSSL_malloc_locked
6866 Realloc => OPENSSL_realloc
6867 Free => OPENSSL_free
6870 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6871 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6874 *) CygWin32 support.
6875 [John Jarvie <jjarvie@newsguy.com>]
6877 *) The type-safe stack code has been rejigged. It is now only compiled
6878 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6879 by default all type-specific stack functions are "#define"d back to
6880 standard stack functions. This results in more streamlined output
6881 but retains the type-safety checking possibilities of the original
6885 *) The STACK code has been cleaned up, and certain type declarations
6886 that didn't make a lot of sense have been brought in line. This has
6887 also involved a cleanup of sorts in safestack.h to more correctly
6888 map type-safe stack functions onto their plain stack counterparts.
6889 This work has also resulted in a variety of "const"ifications of
6890 lots of the code, especially "_cmp" operations which should normally
6891 be prototyped with "const" parameters anyway.
6894 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6895 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6896 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6897 where all of 'md' is used each time the PRNG is used, but 'state'
6898 is used only indexed by a cyclic counter. As entropy may not be
6899 well distributed from the beginning, 'md' is important as a
6900 chaining variable. However, the output function chains only half
6901 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6902 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6903 in all of 'state' being rewritten, with the new values depending
6904 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6907 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6908 the handshake is continued after ssl_verify_cert_chain();
6909 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6910 can lead to 'unexplainable' connection aborts later.
6911 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6913 *) Major EVP API cipher revision.
6914 Add hooks for extra EVP features. This allows various cipher
6915 parameters to be set in the EVP interface. Support added for variable
6916 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6917 setting of RC2 and RC5 parameters.
6919 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6922 Remove lots of duplicated code from the EVP library. For example *every*
6923 cipher init() function handles the 'iv' in the same way according to the
6924 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6925 for CFB and OFB modes they zero ctx->num.
6927 New functionality allows removal of S/MIME code RC2 hack.
6929 Most of the routines have the same form and so can be declared in terms
6932 By shifting this to the top level EVP_CipherInit() it can be removed from
6933 all individual ciphers. If the cipher wants to handle IVs or keys
6934 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6937 Change lots of functions like EVP_EncryptUpdate() to now return a
6938 value: although software versions of the algorithms cannot fail
6939 any installed hardware versions can.
6942 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6943 this option is set, tolerate broken clients that send the negotiated
6944 protocol version number instead of the requested protocol version
6948 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6949 i.e. non-zero for export ciphersuites, zero otherwise.
6950 Previous versions had this flag inverted, inconsistent with
6951 rsa_tmp_cb (..._TMP_RSA_CB).
6952 [Bodo Moeller; problem reported by Amit Chopra]
6954 *) Add missing DSA library text string. Work around for some IIS
6955 key files with invalid SEQUENCE encoding.
6958 *) Add a document (doc/standards.txt) that list all kinds of standards
6959 and so on that are implemented in OpenSSL.
6962 *) Enhance c_rehash script. Old version would mishandle certificates
6963 with the same subject name hash and wouldn't handle CRLs at all.
6964 Added -fingerprint option to crl utility, to support new c_rehash
6968 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6971 *) Fix for SSL server purpose checking. Server checking was
6972 rejecting certificates which had extended key usage present
6973 but no ssl client purpose.
6974 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6976 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6977 is a little unclear about how a blank password is handled.
6978 Since the password in encoded as a BMPString with terminating
6979 double NULL a zero length password would end up as just the
6980 double NULL. However no password at all is different and is
6981 handled differently in the PKCS#12 key generation code. NS
6982 treats a blank password as zero length. MSIE treats it as no
6983 password on export: but it will try both on import. We now do
6984 the same: PKCS12_parse() tries zero length and no password if
6985 the password is set to "" or NULL (NULL is now a valid password:
6986 it wasn't before) as does the pkcs12 application.
6989 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6990 perror when PEM_read_bio_X509_REQ fails, the error message must
6991 be obtained from the error queue.
6994 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6995 it in ERR_remove_state if appropriate, and change ERR_get_state
6996 accordingly to avoid race conditions (this is necessary because
6997 thread_hash is no longer constant once set).
7000 *) Bugfix for linux-elf makefile.one.
7003 *) RSA_get_default_method() will now cause a default
7004 RSA_METHOD to be chosen if one doesn't exist already.
7005 Previously this was only set during a call to RSA_new()
7006 or RSA_new_method(NULL) meaning it was possible for
7007 RSA_get_default_method() to return NULL.
7010 *) Added native name translation to the existing DSO code
7011 that will convert (if the flag to do so is set) filenames
7012 that are sufficiently small and have no path information
7013 into a canonical native form. Eg. "blah" converted to
7014 "libblah.so" or "blah.dll" etc.
7017 *) New function ERR_error_string_n(e, buf, len) which is like
7018 ERR_error_string(e, buf), but writes at most 'len' bytes
7019 including the 0 terminator. For ERR_error_string_n, 'buf'
7021 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7023 *) CONF library reworked to become more general. A new CONF
7024 configuration file reader "class" is implemented as well as a
7025 new functions (NCONF_*, for "New CONF") to handle it. The now
7026 old CONF_* functions are still there, but are reimplemented to
7027 work in terms of the new functions. Also, a set of functions
7028 to handle the internal storage of the configuration data is
7029 provided to make it easier to write new configuration file
7030 reader "classes" (I can definitely see something reading a
7031 configuration file in XML format, for example), called _CONF_*,
7032 or "the configuration storage API"...
7034 The new configuration file reading functions are:
7036 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7037 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7039 NCONF_default, NCONF_WIN32
7041 NCONF_dump_fp, NCONF_dump_bio
7043 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7044 NCONF_new creates a new CONF object. This works in the same way
7045 as other interfaces in OpenSSL, like the BIO interface.
7046 NCONF_dump_* dump the internal storage of the configuration file,
7047 which is useful for debugging. All other functions take the same
7048 arguments as the old CONF_* functions wth the exception of the
7049 first that must be a `CONF *' instead of a `LHASH *'.
7051 To make it easer to use the new classes with the old CONF_* functions,
7052 the function CONF_set_default_method is provided.
7055 *) Add '-tls1' option to 'openssl ciphers', which was already
7056 mentioned in the documentation but had not been implemented.
7057 (This option is not yet really useful because even the additional
7058 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7061 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7062 OpenSSL-based applications) load shared libraries and bind to
7063 them in a portable way.
7064 [Geoff Thorpe, with contributions from Richard Levitte]
7066 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7068 *) Make sure _lrotl and _lrotr are only used with MSVC.
7070 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7071 (the default implementation of RAND_status).
7073 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7074 to '-clrext' (= clear extensions), as intended and documented.
7075 [Bodo Moeller; inconsistency pointed out by Michael Attili
7078 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7079 was larger than the MD block size.
7080 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7082 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7083 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7084 using the passed key: if the passed key was a private key the result
7085 of X509_print(), for example, would be to print out all the private key
7089 *) des_quad_cksum() byte order bug fix.
7090 [Ulf Möller, using the problem description in krb4-0.9.7, where
7091 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7093 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7095 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7097 *) For easily testing in shell scripts whether some command
7098 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7099 returns with exit code 0 iff no command of the given name is available.
7100 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7101 the output goes to stdout and nothing is printed to stderr.
7102 Additional arguments are always ignored.
7104 Since for each cipher there is a command of the same name,
7105 the 'no-cipher' compilation switches can be tested this way.
7107 ('openssl no-XXX' is not able to detect pseudo-commands such
7108 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7111 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7114 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7115 is set; it will be thrown away anyway because each handshake creates
7117 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7118 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7119 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7120 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7123 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7124 'Q' and 'R' lose their special meanings (quit/renegotiate).
7125 This is part of what -quiet does; unlike -quiet, -ign_eof
7126 does not suppress any output.
7129 *) Add compatibility options to the purpose and trust code. The
7130 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7131 accepts a certificate or CA, this was the previous behaviour,
7132 with all the associated security issues.
7134 X509_TRUST_COMPAT is the old trust behaviour: only and
7135 automatically trust self signed roots in certificate store. A
7136 new trust setting X509_TRUST_DEFAULT is used to specify that
7137 a purpose has no associated trust setting and it should instead
7138 use the value in the default purpose.
7141 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7142 and fix a memory leak.
7145 *) In util/mkerr.pl (which implements 'make errors'), preserve
7146 reason strings from the previous version of the .c file, as
7147 the default to have only downcase letters (and digits) in
7148 automatically generated reasons codes is not always appropriate.
7151 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7152 using strerror. Previously, ERR_reason_error_string() returned
7153 library names as reason strings for SYSerr; but SYSerr is a special
7154 case where small numbers are errno values, not library numbers.
7157 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7158 converts DSA parameters into DH parameters. (When creating parameters,
7159 DSA_generate_parameters is used.)
7162 *) Include 'length' (recommended exponent length) in C code generated
7163 by 'openssl dhparam -C'.
7166 *) The second argument to set_label in perlasm was already being used
7167 so couldn't be used as a "file scope" flag. Moved to third argument
7171 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7172 instead of RAND_bytes for encryption IVs and salts.
7175 *) Include RAND_status() into RAND_METHOD instead of implementing
7176 it only for md_rand.c Otherwise replacing the PRNG by calling
7177 RAND_set_rand_method would be impossible.
7180 *) Don't let DSA_generate_key() enter an infinite loop if the random
7181 number generation fails.
7184 *) New 'rand' application for creating pseudo-random output.
7187 *) Added configuration support for Linux/IA64
7188 [Rolf Haberrecker <rolf@suse.de>]
7190 *) Assembler module support for Mingw32.
7193 *) Shared library support for HPUX (in shlib/).
7194 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7196 *) Shared library support for Solaris gcc.
7197 [Lutz Behnke <behnke@trustcenter.de>]
7199 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7201 *) PKCS7_encrypt() was adding text MIME headers twice because they
7202 were added manually and by SMIME_crlf_copy().
7205 *) In bntest.c don't call BN_rand with zero bits argument.
7206 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7208 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7209 case was implemented. This caused BN_div_recp() to fail occasionally.
7212 *) Add an optional second argument to the set_label() in the perl
7213 assembly language builder. If this argument exists and is set
7214 to 1 it signals that the assembler should use a symbol whose
7215 scope is the entire file, not just the current function. This
7216 is needed with MASM which uses the format label:: for this scope.
7217 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7219 *) Change the ASN1 types so they are typedefs by default. Before
7220 almost all types were #define'd to ASN1_STRING which was causing
7221 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7225 *) Change names of new functions to the new get1/get0 naming
7226 convention: After 'get1', the caller owns a reference count
7227 and has to call ..._free; 'get0' returns a pointer to some
7228 data structure without incrementing reference counters.
7229 (Some of the existing 'get' functions increment a reference
7230 counter, some don't.)
7231 Similarly, 'set1' and 'add1' functions increase reference
7232 counters or duplicate objects.
7235 *) Allow for the possibility of temp RSA key generation failure:
7236 the code used to assume it always worked and crashed on failure.
7239 *) Fix potential buffer overrun problem in BIO_printf().
7240 [Ulf Möller, using public domain code by Patrick Powell; problem
7241 pointed out by David Sacerdote <das33@cornell.edu>]
7243 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7244 RAND_egd() and RAND_status(). In the command line application,
7245 the EGD socket can be specified like a seed file using RANDFILE
7249 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7250 Some CAs (e.g. Verisign) distribute certificates in this form.
7253 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7254 list to exclude them. This means that no special compilation option
7255 is needed to use anonymous DH: it just needs to be included in the
7259 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7260 EVP_MD_type. The old functionality is available in a new macro called
7261 EVP_MD_md(). Change code that uses it and update docs.
7264 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7265 where the 'void *' argument is replaced by a function pointer argument.
7266 Previously 'void *' was abused to point to functions, which works on
7267 many platforms, but is not correct. As these functions are usually
7268 called by macros defined in OpenSSL header files, most source code
7269 should work without changes.
7272 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7273 sections with information on -D... compiler switches used for
7274 compiling the library so that applications can see them. To enable
7275 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7276 must be defined. E.g.,
7277 #define OPENSSL_ALGORITHM_DEFINES
7278 #include <openssl/opensslconf.h>
7279 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7280 [Richard Levitte, Ulf and Bodo Möller]
7282 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7286 *) Change the 'other' type in certificate aux info to a STACK_OF
7287 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7288 the required ASN1 format: arbitrary types determined by an OID.
7291 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7292 argument to 'req'. This is not because the function is newer or
7293 better than others it just uses the work 'NEW' in the certificate
7294 request header lines. Some software needs this.
7297 *) Reorganise password command line arguments: now passwords can be
7298 obtained from various sources. Delete the PEM_cb function and make
7299 it the default behaviour: i.e. if the callback is NULL and the
7300 usrdata argument is not NULL interpret it as a null terminated pass
7301 phrase. If usrdata and the callback are NULL then the pass phrase
7302 is prompted for as usual.
7305 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7306 the support is automatically enabled. The resulting binaries will
7307 autodetect the card and use it if present.
7308 [Ben Laurie and Compaq Inc.]
7310 *) Work around for Netscape hang bug. This sends certificate request
7311 and server done in one record. Since this is perfectly legal in the
7312 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7313 the bugs/SSLv3 entry for more info.
7316 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7319 *) Add -rand argument to smime and pkcs12 applications and read/write
7323 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7326 *) Add command line password options to the remaining applications.
7329 *) Bug fix for BN_div_recp() for numerators with an even number of
7333 *) More tests in bntest.c, and changed test_bn output.
7336 *) ./config recognizes MacOS X now.
7339 *) Bug fix for BN_div() when the first words of num and divsor are
7340 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7343 *) Add support for various broken PKCS#8 formats, and command line
7344 options to produce them.
7347 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7348 get temporary BIGNUMs from a BN_CTX.
7351 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7355 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7356 include a #define from the old name to the new. The original intent
7357 was that statically linked binaries could for example just call
7358 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7359 link with digests. This never worked becayse SSLeay_add_all_digests()
7360 and SSLeay_add_all_ciphers() were in the same source file so calling
7361 one would link with the other. They are now in separate source files.
7364 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7367 *) Use a less unusual form of the Miller-Rabin primality test (it used
7368 a binary algorithm for exponentiation integrated into the Miller-Rabin
7369 loop, our standard modexp algorithms are faster).
7372 *) Support for the EBCDIC character set completed.
7373 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7375 *) Source code cleanups: use const where appropriate, eliminate casts,
7376 use void * instead of char * in lhash.
7379 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7380 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7381 this the server could overwrite ephemeral keys that the client
7385 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7386 using 50 iterations of the Rabin-Miller test.
7388 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7389 iterations of the Rabin-Miller test as required by the appendix
7390 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7391 As BN_is_prime_fasttest includes trial division, DSA parameter
7392 generation becomes much faster.
7394 This implies a change for the callback functions in DSA_is_prime
7395 and DSA_generate_parameters: The callback function is called once
7396 for each positive witness in the Rabin-Miller test, not just
7397 occasionally in the inner loop; and the parameters to the
7398 callback function now provide an iteration count for the outer
7399 loop rather than for the current invocation of the inner loop.
7400 DSA_generate_parameters additionally can call the callback
7401 function with an 'iteration count' of -1, meaning that a
7402 candidate has passed the trial division test (when q is generated
7403 from an application-provided seed, trial division is skipped).
7406 *) New function BN_is_prime_fasttest that optionally does trial
7407 division before starting the Rabin-Miller test and has
7408 an additional BN_CTX * argument (whereas BN_is_prime always
7409 has to allocate at least one BN_CTX).
7410 'callback(1, -1, cb_arg)' is called when a number has passed the
7411 trial division stage.
7414 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7418 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7421 *) New function BN_pseudo_rand().
7424 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7425 bignum version of BN_from_montgomery() with the working code from
7426 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7430 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7431 made it impossible to use the same SSL_SESSION data structure in
7432 SSL2 clients in multiple threads.
7435 *) The return value of RAND_load_file() no longer counts bytes obtained
7436 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7437 to seed the PRNG (previously an explicit byte count was required).
7438 [Ulf Möller, Bodo Möller]
7440 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7441 used (char *) instead of (void *) and had casts all over the place.
7444 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7447 *) Retain source code compatibility for BN_prime_checks macro:
7448 BN_is_prime(..., BN_prime_checks, ...) now uses
7449 BN_prime_checks_for_size to determine the appropriate number of
7450 Rabin-Miller iterations.
7453 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7454 DH_CHECK_P_NOT_SAFE_PRIME.
7455 (Check if this is true? OpenPGP calls them "strong".)
7458 *) Merge the functionality of "dh" and "gendh" programs into a new program
7459 "dhparam". The old programs are retained for now but will handle DH keys
7460 (instead of parameters) in future.
7463 *) Make the ciphers, s_server and s_client programs check the return values
7464 when a new cipher list is set.
7467 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7468 ciphers. Before when the 56bit ciphers were enabled the sorting was
7471 The syntax for the cipher sorting has been extended to support sorting by
7472 cipher-strength (using the strength_bits hard coded in the tables).
7473 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7475 Fix a bug in the cipher-command parser: when supplying a cipher command
7476 string with an "undefined" symbol (neither command nor alphanumeric
7477 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7478 an error is flagged.
7480 Due to the strength-sorting extension, the code of the
7481 ssl_create_cipher_list() function was completely rearranged. I hope that
7482 the readability was also increased :-)
7483 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7485 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7486 for the first serial number and places 2 in the serial number file. This
7487 avoids problems when the root CA is created with serial number zero and
7488 the first user certificate has the same issuer name and serial number
7492 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7493 the new code. Add documentation for this stuff.
7496 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7497 X509_*() to X509at_*() on the grounds that they don't handle X509
7498 structures and behave in an analagous way to the X509v3 functions:
7499 they shouldn't be called directly but wrapper functions should be used
7502 So we also now have some wrapper functions that call the X509at functions
7503 when passed certificate requests. (TO DO: similar things can be done with
7504 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7505 things. Some of these need some d2i or i2d and print functionality
7506 because they handle more complex structures.)
7509 *) Add missing #ifndefs that caused missing symbols when building libssl
7510 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7511 NO_RSA in ssl/s2*.c.
7512 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7514 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7515 has a return value which indicates the quality of the random data
7516 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7517 error queue. New function RAND_pseudo_bytes() generates output that is
7518 guaranteed to be unique but not unpredictable. RAND_add is like
7519 RAND_seed, but takes an extra argument for an entropy estimate
7520 (RAND_seed always assumes full entropy).
7523 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7524 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7525 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7526 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7527 false-positive rate of at most 2^-80 for random input.
7530 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7533 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7534 in the 0.9.5 release), this returns the chain
7535 from an X509_CTX structure with a dup of the stack and all
7536 the X509 reference counts upped: so the stack will exist
7537 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7540 Also make SSL_SESSION_print() print out the verify return
7544 *) Add manpage for the pkcs12 command. Also change the default
7545 behaviour so MAC iteration counts are used unless the new
7546 -nomaciter option is used. This improves file security and
7547 only older versions of MSIE (4.0 for example) need it.
7550 *) Honor the no-xxx Configure options when creating .DEF files.
7553 *) Add PKCS#10 attributes to field table: challengePassword,
7554 unstructuredName and unstructuredAddress. These are taken from
7555 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7556 international characters are used.
7558 More changes to X509_ATTRIBUTE code: allow the setting of types
7559 based on strings. Remove the 'loc' parameter when adding
7560 attributes because these will be a SET OF encoding which is sorted
7564 *) Initial changes to the 'req' utility to allow request generation
7565 automation. This will allow an application to just generate a template
7566 file containing all the field values and have req construct the
7569 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7570 used all over the place including certificate requests and PKCS#7
7571 structures. They are currently handled manually where necessary with
7572 some primitive wrappers for PKCS#7. The new functions behave in a
7573 manner analogous to the X509 extension functions: they allow
7574 attributes to be looked up by NID and added.
7576 Later something similar to the X509V3 code would be desirable to
7577 automatically handle the encoding, decoding and printing of the
7578 more complex types. The string types like challengePassword can
7579 be handled by the string table functions.
7581 Also modified the multi byte string table handling. Now there is
7582 a 'global mask' which masks out certain types. The table itself
7583 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7584 is useful when for example there is only one permissible type
7585 (as in countryName) and using the mask might result in no valid
7589 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7590 SSL_get_peer_finished to allow applications to obtain the latest
7591 Finished messages sent to the peer or expected from the peer,
7592 respectively. (SSL_get_peer_finished is usually the Finished message
7593 actually received from the peer, otherwise the protocol will be aborted.)
7595 As the Finished message are message digests of the complete handshake
7596 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7597 be used for external authentication procedures when the authentication
7598 provided by SSL/TLS is not desired or is not enough.
7601 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7602 the host supports BWX extension and if Compaq C is present on the
7603 $PATH. Just exploiting of the BWX extension results in 20-30%
7604 performance kick for some algorithms, e.g. DES and RC4 to mention
7605 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7609 *) Add support for MS "fast SGC". This is arguably a violation of the
7610 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7611 weak crypto and after checking the certificate is SGC a second one
7612 with strong crypto. MS SGC stops the first handshake after receiving
7613 the server certificate message and sends a second client hello. Since
7614 a server will typically do all the time consuming operations before
7615 expecting any further messages from the client (server key exchange
7616 is the most expensive) there is little difference between the two.
7618 To get OpenSSL to support MS SGC we have to permit a second client
7619 hello message after we have sent server done. In addition we have to
7620 reset the MAC if we do get this second client hello.
7623 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7624 if a DER encoded private key is RSA or DSA traditional format. Changed
7625 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7626 format DER encoded private key. Newer code should use PKCS#8 format which
7627 has the key type encoded in the ASN1 structure. Added DER private key
7628 support to pkcs8 application.
7631 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7632 ciphersuites has been selected (as required by the SSL 3/TLS 1
7633 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7634 is set, we interpret this as a request to violate the specification
7635 (the worst that can happen is a handshake failure, and 'correct'
7636 behaviour would result in a handshake failure anyway).
7639 *) In SSL_CTX_add_session, take into account that there might be multiple
7640 SSL_SESSION structures with the same session ID (e.g. when two threads
7641 concurrently obtain them from an external cache).
7642 The internal cache can handle only one SSL_SESSION with a given ID,
7643 so if there's a conflict, we now throw out the old one to achieve
7647 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7648 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7649 some routines that use cipher OIDs: some ciphers do not have OIDs
7650 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7654 *) Simplify the trust setting structure and code. Now we just have
7655 two sequences of OIDs for trusted and rejected settings. These will
7656 typically have values the same as the extended key usage extension
7657 and any application specific purposes.
7659 The trust checking code now has a default behaviour: it will just
7660 check for an object with the same NID as the passed id. Functions can
7661 be provided to override either the default behaviour or the behaviour
7662 for a given id. SSL client, server and email already have functions
7663 in place for compatibility: they check the NID and also return "trusted"
7664 if the certificate is self signed.
7667 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7668 traditional format into an EVP_PKEY structure.
7671 *) Add a password callback function PEM_cb() which either prompts for
7672 a password if usr_data is NULL or otherwise assumes it is a null
7673 terminated password. Allow passwords to be passed on command line
7674 environment or config files in a few more utilities.
7677 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7678 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7679 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7680 Update documentation.
7683 *) Support for ASN1 "NULL" type. This could be handled before by using
7684 ASN1_TYPE but there wasn't any function that would try to read a NULL
7685 and produce an error if it couldn't. For compatibility we also have
7686 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7687 don't allocate anything because they don't need to.
7690 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7692 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7694 *) Rebuild of the memory allocation routines used by OpenSSL code and
7695 possibly others as well. The purpose is to make an interface that
7696 provide hooks so anyone can build a separate set of allocation and
7697 deallocation routines to be used by OpenSSL, for example memory
7698 pool implementations, or something else, which was previously hard
7699 since Malloc(), Realloc() and Free() were defined as macros having
7700 the values malloc, realloc and free, respectively (except for Win32
7701 compilations). The same is provided for memory debugging code.
7702 OpenSSL already comes with functionality to find memory leaks, but
7703 this gives people a chance to debug other memory problems.
7705 With these changes, a new set of functions and macros have appeared:
7707 CRYPTO_set_mem_debug_functions() [F]
7708 CRYPTO_get_mem_debug_functions() [F]
7709 CRYPTO_dbg_set_options() [F]
7710 CRYPTO_dbg_get_options() [F]
7711 CRYPTO_malloc_debug_init() [M]
7713 The memory debug functions are NULL by default, unless the library
7714 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7715 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7716 gives the standard debugging functions that come with OpenSSL) or
7717 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7718 provided by the library user) must be used. When the standard
7719 debugging functions are used, CRYPTO_dbg_set_options can be used to
7720 request additional information:
7721 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7722 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7724 Also, things like CRYPTO_set_mem_functions will always give the
7725 expected result (the new set of functions is used for allocation
7726 and deallocation) at all times, regardless of platform and compiler
7729 To finish it up, some functions that were never use in any other
7730 way than through macros have a new API and new semantic:
7733 CRYPTO_dbg_realloc()
7736 All macros of value have retained their old syntax.
7737 [Richard Levitte and Bodo Moeller]
7739 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7740 ordering of SMIMECapabilities wasn't in "strength order" and there
7741 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7745 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7746 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7747 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7749 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7750 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7751 functionality to handle multipart/signed properly) and a utility
7752 called 'smime' to call all this stuff. This is based on code I
7753 originally wrote for Celo who have kindly allowed it to be
7754 included in OpenSSL.
7757 *) Add variants des_set_key_checked and des_set_key_unchecked of
7758 des_set_key (aka des_key_sched). Global variable des_check_key
7759 decides which of these is called by des_set_key; this way
7760 des_check_key behaves as it always did, but applications and
7761 the library itself, which was buggy for des_check_key == 1,
7762 have a cleaner way to pick the version they need.
7765 *) New function PKCS12_newpass() which changes the password of a
7769 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7770 dynamic mix. In both cases the ids can be used as an index into the
7771 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7772 functions so they accept a list of the field values and the
7773 application doesn't need to directly manipulate the X509_TRUST
7777 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7781 *) Modify the way the V3 extension code looks up extensions. This now
7782 works in a similar way to the object code: we have some "standard"
7783 extensions in a static table which is searched with OBJ_bsearch()
7784 and the application can add dynamic ones if needed. The file
7785 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7786 updated whenever a new extension is added to the core code and kept
7787 in ext_nid order. There is a simple program 'tabtest.c' which checks
7788 this. New extensions are not added too often so this file can readily
7789 be maintained manually.
7791 There are two big advantages in doing things this way. The extensions
7792 can be looked up immediately and no longer need to be "added" using
7793 X509V3_add_standard_extensions(): this function now does nothing.
7794 [Side note: I get *lots* of email saying the extension code doesn't
7795 work because people forget to call this function]
7796 Also no dynamic allocation is done unless new extensions are added:
7797 so if we don't add custom extensions there is no need to call
7798 X509V3_EXT_cleanup().
7801 *) Modify enc utility's salting as follows: make salting the default. Add a
7802 magic header, so unsalted files fail gracefully instead of just decrypting
7803 to garbage. This is because not salting is a big security hole, so people
7804 should be discouraged from doing it.
7807 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7808 digest to be passed on the command line but it only used this
7809 parameter when signing a certificate. Modified so all relevant
7810 operations are affected by the digest parameter including the
7811 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7812 DSA key was used because it didn't fix the digest.
7815 *) Initial certificate chain verify code. Currently tests the untrusted
7816 certificates for consistency with the verify purpose (which is set
7817 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7819 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7820 this is because it will reject chains with invalid extensions whereas
7821 every previous version of OpenSSL and SSLeay made no checks at all.
7823 Trust code: checks the root CA for the relevant trust settings. Trust
7824 settings have an initial value consistent with the verify purpose: e.g.
7825 if the verify purpose is for SSL client use it expects the CA to be
7826 trusted for SSL client use. However the default value can be changed to
7827 permit custom trust settings: one example of this would be to only trust
7828 certificates from a specific "secure" set of CAs.
7830 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7831 which should be used for version portability: especially since the
7832 verify structure is likely to change more often now.
7834 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7835 to set them. If not set then assume SSL clients will verify SSL servers
7838 Two new options to the verify program: -untrusted allows a set of
7839 untrusted certificates to be passed in and -purpose which sets the
7840 intended purpose of the certificate. If a purpose is set then the
7841 new chain verify code is used to check extension consistency.
7844 *) Support for the authority information access extension.
7847 *) Modify RSA and DSA PEM read routines to transparently handle
7848 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7849 public keys in a format compatible with certificate
7850 SubjectPublicKeyInfo structures. Unfortunately there were already
7851 functions called *_PublicKey_* which used various odd formats so
7852 these are retained for compatibility: however the DSA variants were
7853 never in a public release so they have been deleted. Changed dsa/rsa
7854 utilities to handle the new format: note no releases ever handled public
7855 keys so we should be OK.
7857 The primary motivation for this change is to avoid the same fiasco
7858 that dogs private keys: there are several incompatible private key
7859 formats some of which are standard and some OpenSSL specific and
7860 require various evil hacks to allow partial transparent handling and
7861 even then it doesn't work with DER formats. Given the option anything
7862 other than PKCS#8 should be dumped: but the other formats have to
7863 stay in the name of compatibility.
7865 With public keys and the benefit of hindsight one standard format
7866 is used which works with EVP_PKEY, RSA or DSA structures: though
7867 it clearly returns an error if you try to read the wrong kind of key.
7869 Added a -pubkey option to the 'x509' utility to output the public key.
7870 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7871 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7872 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7873 that do the same as the EVP_PKEY_assign_*() except they up the
7874 reference count of the added key (they don't "swallow" the
7878 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7879 CRLs would fail if the file contained no certificates or no CRLs:
7880 added a new function to read in both types and return the number
7881 read: this means that if none are read it will be an error. The
7882 DER versions of the certificate and CRL reader would always fail
7883 because it isn't possible to mix certificates and CRLs in DER format
7884 without choking one or the other routine. Changed this to just read
7885 a certificate: this is the best we can do. Also modified the code
7886 in apps/verify.c to take notice of return codes: it was previously
7887 attempting to read in certificates from NULL pointers and ignoring
7888 any errors: this is one reason why the cert and CRL reader seemed
7889 to work. It doesn't check return codes from the default certificate
7890 routines: these may well fail if the certificates aren't installed.
7893 *) Code to support otherName option in GeneralName.
7896 *) First update to verify code. Change the verify utility
7897 so it warns if it is passed a self signed certificate:
7898 for consistency with the normal behaviour. X509_verify
7899 has been modified to it will now verify a self signed
7900 certificate if *exactly* the same certificate appears
7901 in the store: it was previously impossible to trust a
7902 single self signed certificate. This means that:
7903 openssl verify ss.pem
7904 now gives a warning about a self signed certificate but
7905 openssl verify -CAfile ss.pem ss.pem
7909 *) For servers, store verify_result in SSL_SESSION data structure
7910 (and add it to external session representation).
7911 This is needed when client certificate verifications fails,
7912 but an application-provided verification callback (set by
7913 SSL_CTX_set_cert_verify_callback) allows accepting the session
7914 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7915 but returns 1): When the session is reused, we have to set
7916 ssl->verify_result to the appropriate error code to avoid
7918 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7920 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7921 case in PKCS7_dataInit() where the signed PKCS7 structure
7922 didn't contain any existing data because it was being created.
7923 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7925 *) Add a salt to the key derivation routines in enc.c. This
7926 forms the first 8 bytes of the encrypted file. Also add a
7927 -S option to allow a salt to be input on the command line.
7930 *) New function X509_cmp(). Oddly enough there wasn't a function
7931 to compare two certificates. We do this by working out the SHA1
7932 hash and comparing that. X509_cmp() will be needed by the trust
7936 *) SSL_get1_session() is like SSL_get_session(), but increments
7937 the reference count in the SSL_SESSION returned.
7938 [Geoff Thorpe <geoff@eu.c2.net>]
7940 *) Fix for 'req': it was adding a null to request attributes.
7941 Also change the X509_LOOKUP and X509_INFO code to handle
7942 certificate auxiliary information.
7945 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7949 *) Add the possibility to add extra information to the memory leak
7950 detecting output, to form tracebacks, showing from where each
7951 allocation was originated: CRYPTO_push_info("constant string") adds
7952 the string plus current file name and line number to a per-thread
7953 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7954 is like calling CYRPTO_pop_info() until the stack is empty.
7955 Also updated memory leak detection code to be multi-thread-safe.
7958 *) Add options -text and -noout to pkcs7 utility and delete the
7959 encryption options which never did anything. Update docs.
7962 *) Add options to some of the utilities to allow the pass phrase
7963 to be included on either the command line (not recommended on
7964 OSes like Unix) or read from the environment. Update the
7965 manpages and fix a few bugs.
7968 *) Add a few manpages for some of the openssl commands.
7971 *) Fix the -revoke option in ca. It was freeing up memory twice,
7972 leaking and not finding already revoked certificates.
7975 *) Extensive changes to support certificate auxiliary information.
7976 This involves the use of X509_CERT_AUX structure and X509_AUX
7977 functions. An X509_AUX function such as PEM_read_X509_AUX()
7978 can still read in a certificate file in the usual way but it
7979 will also read in any additional "auxiliary information". By
7980 doing things this way a fair degree of compatibility can be
7981 retained: existing certificates can have this information added
7982 using the new 'x509' options.
7984 Current auxiliary information includes an "alias" and some trust
7985 settings. The trust settings will ultimately be used in enhanced
7986 certificate chain verification routines: currently a certificate
7987 can only be trusted if it is self signed and then it is trusted
7991 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7992 The problem was that one of the replacement routines had not been working
7993 since SSLeay releases. For now the offending routine has been replaced
7994 with non-optimised assembler. Even so, this now gives around 95%
7995 performance improvement for 1024 bit RSA signs.
7998 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7999 handling. Most clients have the effective key size in bits equal to
8000 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8001 A few however don't do this and instead use the size of the decrypted key
8002 to determine the RC2 key length and the AlgorithmIdentifier to determine
8003 the effective key length. In this case the effective key length can still
8004 be 40 bits but the key length can be 168 bits for example. This is fixed
8005 by manually forcing an RC2 key into the EVP_PKEY structure because the
8006 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8007 the key length and effective key length are equal.
8010 *) Add a bunch of functions that should simplify the creation of
8011 X509_NAME structures. Now you should be able to do:
8012 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8013 and have it automatically work out the correct field type and fill in
8014 the structures. The more adventurous can try:
8015 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8016 and it will (hopefully) work out the correct multibyte encoding.
8019 *) Change the 'req' utility to use the new field handling and multibyte
8020 copy routines. Before the DN field creation was handled in an ad hoc
8021 way in req, ca, and x509 which was rather broken and didn't support
8022 BMPStrings or UTF8Strings. Since some software doesn't implement
8023 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8024 using the dirstring_type option. See the new comment in the default
8025 openssl.cnf for more info.
8028 *) Make crypto/rand/md_rand.c more robust:
8029 - Assure unique random numbers after fork().
8030 - Make sure that concurrent threads access the global counter and
8031 md serializably so that we never lose entropy in them
8032 or use exactly the same state in multiple threads.
8033 Access to the large state is not always serializable because
8034 the additional locking could be a performance killer, and
8035 md should be large enough anyway.
8038 *) New file apps/app_rand.c with commonly needed functionality
8039 for handling the random seed file.
8041 Use the random seed file in some applications that previously did not:
8043 dsaparam -genkey (which also ignored its '-rand' option),
8046 x509 (when signing).
8047 Except on systems with /dev/urandom, it is crucial to have a random
8048 seed file at least for key creation, DSA signing, and for DH exchanges;
8049 for RSA signatures we could do without one.
8051 gendh and gendsa (unlike genrsa) used to read only the first byte
8052 of each file listed in the '-rand' option. The function as previously
8053 found in genrsa is now in app_rand.c and is used by all programs
8054 that support '-rand'.
8057 *) In RAND_write_file, use mode 0600 for creating files;
8058 don't just chmod when it may be too late.
8061 *) Report an error from X509_STORE_load_locations
8062 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8065 *) New function ASN1_mbstring_copy() this copies a string in either
8066 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8067 into an ASN1_STRING type. A mask of permissible types is passed
8068 and it chooses the "minimal" type to use or an error if not type
8072 *) Add function equivalents to the various macros in asn1.h. The old
8073 macros are retained with an M_ prefix. Code inside the library can
8074 use the M_ macros. External code (including the openssl utility)
8075 should *NOT* in order to be "shared library friendly".
8078 *) Add various functions that can check a certificate's extensions
8079 to see if it usable for various purposes such as SSL client,
8080 server or S/MIME and CAs of these types. This is currently
8081 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8082 verification. Also added a -purpose flag to x509 utility to
8083 print out all the purposes.
8086 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8090 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8091 for, obtain and decode and extension and obtain its critical flag.
8092 This allows all the necessary extension code to be handled in a
8093 single function call.
8096 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8097 platforms. See crypto/rc4/rc4_enc.c for further details.
8100 *) New -noout option to asn1parse. This causes no output to be produced
8101 its main use is when combined with -strparse and -out to extract data
8102 from a file (which may not be in ASN.1 format).
8105 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8106 when producing the local key id.
8107 [Richard Levitte <levitte@stacken.kth.se>]
8109 *) New option -dhparam in s_server. This allows a DH parameter file to be
8110 stated explicitly. If it is not stated then it tries the first server
8111 certificate file. The previous behaviour hard coded the filename
8115 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8116 a public key to be input or output. For example:
8117 openssl rsa -in key.pem -pubout -out pubkey.pem
8118 Also added necessary DSA public key functions to handle this.
8121 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8122 in the message. This was handled by allowing
8123 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8124 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8126 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8127 to the end of the strings whereas this didn't. This would cause problems
8128 if strings read with d2i_ASN1_bytes() were later modified.
8129 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8131 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8132 data and it contains EOF it will end up returning an error. This is
8133 caused by input 46 bytes long. The cause is due to the way base64
8134 BIOs find the start of base64 encoded data. They do this by trying a
8135 trial decode on each line until they find one that works. When they
8136 do a flag is set and it starts again knowing it can pass all the
8137 data directly through the decoder. Unfortunately it doesn't reset
8138 the context it uses. This means that if EOF is reached an attempt
8139 is made to pass two EOFs through the context and this causes the
8140 resulting error. This can also cause other problems as well. As is
8141 usual with these problems it takes *ages* to find and the fix is
8142 trivial: move one line.
8143 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8145 *) Ugly workaround to get s_client and s_server working under Windows. The
8146 old code wouldn't work because it needed to select() on sockets and the
8147 tty (for keypresses and to see if data could be written). Win32 only
8148 supports select() on sockets so we select() with a 1s timeout on the
8149 sockets and then see if any characters are waiting to be read, if none
8150 are present then we retry, we also assume we can always write data to
8151 the tty. This isn't nice because the code then blocks until we've
8152 received a complete line of data and it is effectively polling the
8153 keyboard at 1s intervals: however it's quite a bit better than not
8154 working at all :-) A dedicated Windows application might handle this
8155 with an event loop for example.
8158 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8159 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8160 will be called when RSA_sign() and RSA_verify() are used. This is useful
8161 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8162 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8163 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8164 This necessitated the support of an extra signature type NID_md5_sha1
8165 for SSL signatures and modifications to the SSL library to use it instead
8166 of calling RSA_public_decrypt() and RSA_private_encrypt().
8169 *) Add new -verify -CAfile and -CApath options to the crl program, these
8170 will lookup a CRL issuers certificate and verify the signature in a
8171 similar way to the verify program. Tidy up the crl program so it
8172 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8173 less strict. It will now permit CRL extensions even if it is not
8174 a V2 CRL: this will allow it to tolerate some broken CRLs.
8177 *) Initialize all non-automatic variables each time one of the openssl
8178 sub-programs is started (this is necessary as they may be started
8179 multiple times from the "OpenSSL>" prompt).
8180 [Lennart Bang, Bodo Moeller]
8182 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8183 removing all other RSA functionality (this is what NO_RSA does). This
8184 is so (for example) those in the US can disable those operations covered
8185 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8189 *) Non-copying interface to BIO pairs.
8190 (still largely untested)
8193 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8194 ASCII string. This was handled independently in various places before.
8197 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8198 UTF8 strings a character at a time.
8201 *) Use client_version from client hello to select the protocol
8202 (s23_srvr.c) and for RSA client key exchange verification
8203 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8206 *) Add various utility functions to handle SPKACs, these were previously
8207 handled by poking round in the structure internals. Added new function
8208 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8209 print, verify and generate SPKACs. Based on an original idea from
8210 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8213 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8216 *) Allow the config file extension section to be overwritten on the
8217 command line. Based on an original idea from Massimiliano Pala
8218 <madwolf@comune.modena.it>. The new option is called -extensions
8219 and can be applied to ca, req and x509. Also -reqexts to override
8220 the request extensions in req and -crlexts to override the crl extensions
8224 *) Add new feature to the SPKAC handling in ca. Now you can include
8225 the same field multiple times by preceding it by "XXXX." for example:
8228 this is the same syntax as used in the req config file.
8231 *) Allow certificate extensions to be added to certificate requests. These
8232 are specified in a 'req_extensions' option of the req section of the
8233 config file. They can be printed out with the -text option to req but
8234 are otherwise ignored at present.
8237 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8238 data read consists of only the final block it would not decrypted because
8239 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8240 A misplaced 'break' also meant the decrypted final block might not be
8241 copied until the next read.
8244 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8245 a few extra parameters to the DH structure: these will be useful if
8246 for example we want the value of 'q' or implement X9.42 DH.
8249 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8250 provides hooks that allow the default DSA functions or functions on a
8251 "per key" basis to be replaced. This allows hardware acceleration and
8252 hardware key storage to be handled without major modification to the
8253 library. Also added low level modexp hooks and CRYPTO_EX structure and
8254 associated functions.
8257 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8258 as "read only": it can't be written to and the buffer it points to will
8259 not be freed. Reading from a read only BIO is much more efficient than
8260 a normal memory BIO. This was added because there are several times when
8261 an area of memory needs to be read from a BIO. The previous method was
8262 to create a memory BIO and write the data to it, this results in two
8263 copies of the data and an O(n^2) reading algorithm. There is a new
8264 function BIO_new_mem_buf() which creates a read only memory BIO from
8265 an area of memory. Also modified the PKCS#7 routines to use read only
8269 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8270 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8271 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8272 but a retry condition occured while trying to read the rest.
8275 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8276 NID_pkcs7_encrypted by default: this was wrong since this should almost
8277 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8278 the encrypted data type: this is a more sensible place to put it and it
8279 allows the PKCS#12 code to be tidied up that duplicated this
8283 *) Changed obj_dat.pl script so it takes its input and output files on
8284 the command line. This should avoid shell escape redirection problems
8288 *) Initial support for certificate extension requests, these are included
8289 in things like Xenroll certificate requests. Included functions to allow
8290 extensions to be obtained and added.
8293 *) -crlf option to s_client and s_server for sending newlines as
8294 CRLF (as required by many protocols).
8297 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8299 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8300 [Ralf S. Engelschall]
8302 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8303 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8305 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8309 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8310 DH parameters/keys (q is lost during that conversion, but the resulting
8311 DH parameters contain its length).
8313 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8314 much faster than DH_generate_parameters (which creates parameters
8315 where p = 2*q + 1), and also the smaller q makes DH computations
8316 much more efficient (160-bit exponentiation instead of 1024-bit
8317 exponentiation); so this provides a convenient way to support DHE
8318 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8319 utter importance to use
8320 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8322 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8323 when such DH parameters are used, because otherwise small subgroup
8324 attacks may become possible!
8327 *) Avoid memory leak in i2d_DHparams.
8330 *) Allow the -k option to be used more than once in the enc program:
8331 this allows the same encrypted message to be read by multiple recipients.
8334 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8335 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8336 it will always use the numerical form of the OID, even if it has a short
8340 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8341 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8342 otherwise bn_mod_exp was called. In the case of hardware keys for example
8343 no private key components need be present and it might store extra data
8344 in the RSA structure, which cannot be accessed from bn_mod_exp.
8345 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8346 private key operations.
8349 *) Added support for SPARC Linux.
8352 *) pem_password_cb function type incompatibly changed from
8353 typedef int pem_password_cb(char *buf, int size, int rwflag);
8355 ....(char *buf, int size, int rwflag, void *userdata);
8356 so that applications can pass data to their callbacks:
8357 The PEM[_ASN1]_{read,write}... functions and macros now take an
8358 additional void * argument, which is just handed through whenever
8359 the password callback is called.
8360 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8362 New function SSL_CTX_set_default_passwd_cb_userdata.
8364 Compatibility note: As many C implementations push function arguments
8365 onto the stack in reverse order, the new library version is likely to
8366 interoperate with programs that have been compiled with the old
8367 pem_password_cb definition (PEM_whatever takes some data that
8368 happens to be on the stack as its last argument, and the callback
8369 just ignores this garbage); but there is no guarantee whatsoever that
8372 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8373 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8374 problems not only on Windows, but also on some Unix platforms.
8375 To avoid problematic command lines, these definitions are now in an
8376 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8377 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8380 *) MIPS III/IV assembler module is reimplemented.
8383 *) More DES library cleanups: remove references to srand/rand and
8384 delete an unused file.
8387 *) Add support for the the free Netwide assembler (NASM) under Win32,
8388 since not many people have MASM (ml) and it can be hard to obtain.
8389 This is currently experimental but it seems to work OK and pass all
8390 the tests. Check out INSTALL.W32 for info.
8393 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8394 without temporary keys kept an extra copy of the server key,
8395 and connections with temporary keys did not free everything in case
8399 *) New function RSA_check_key and new openssl rsa option -check
8400 for verifying the consistency of RSA keys.
8401 [Ulf Moeller, Bodo Moeller]
8403 *) Various changes to make Win32 compile work:
8404 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8405 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8406 comparison" warnings.
8407 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8410 *) Add a debugging option to PKCS#5 v2 key generation function: when
8411 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8412 derived keys are printed to stderr.
8415 *) Copy the flags in ASN1_STRING_dup().
8416 [Roman E. Pavlov <pre@mo.msk.ru>]
8418 *) The x509 application mishandled signing requests containing DSA
8419 keys when the signing key was also DSA and the parameters didn't match.
8421 It was supposed to omit the parameters when they matched the signing key:
8422 the verifying software was then supposed to automatically use the CA's
8423 parameters if they were absent from the end user certificate.
8425 Omitting parameters is no longer recommended. The test was also
8426 the wrong way round! This was probably due to unusual behaviour in
8427 EVP_cmp_parameters() which returns 1 if the parameters match.
8428 This meant that parameters were omitted when they *didn't* match and
8429 the certificate was useless. Certificates signed with 'ca' didn't have
8431 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8433 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8434 The interface is as follows:
8435 Applications can use
8436 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8437 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8438 "off" is now the default.
8439 The library internally uses
8440 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8441 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8442 to disable memory-checking temporarily.
8444 Some inconsistent states that previously were possible (and were
8445 even the default) are now avoided.
8447 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8448 with each memory chunk allocated; this is occasionally more helpful
8449 than just having a counter.
8451 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8453 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8457 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8458 which largely parallels "options", but is for changing API behaviour,
8459 whereas "options" are about protocol behaviour.
8460 Initial "mode" flags are:
8462 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8463 a single record has been written.
8464 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8465 retries use the same buffer location.
8466 (But all of the contents must be
8470 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8473 *) Fix problems with no-hmac etc.
8474 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8476 *) New functions RSA_get_default_method(), RSA_set_method() and
8477 RSA_get_method(). These allows replacement of RSA_METHODs without having
8478 to mess around with the internals of an RSA structure.
8481 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8482 Also really enable memory leak checks in openssl.c and in some
8484 [Chad C. Mulligan, Bodo Moeller]
8486 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8487 up the length of negative integers. This has now been simplified to just
8488 store the length when it is first determined and use it later, rather
8489 than trying to keep track of where data is copied and updating it to
8491 [Steve Henson, reported by Brien Wheeler
8492 <bwheeler@authentica-security.com>]
8494 *) Add a new function PKCS7_signatureVerify. This allows the verification
8495 of a PKCS#7 signature but with the signing certificate passed to the
8496 function itself. This contrasts with PKCS7_dataVerify which assumes the
8497 certificate is present in the PKCS#7 structure. This isn't always the
8498 case: certificates can be omitted from a PKCS#7 structure and be
8499 distributed by "out of band" means (such as a certificate database).
8502 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8503 function prototypes in pem.h, also change util/mkdef.pl to add the
8504 necessary function names.
8507 *) mk1mf.pl (used by Windows builds) did not properly read the
8508 options set by Configure in the top level Makefile, and Configure
8509 was not even able to write more than one option correctly.
8510 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8513 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8514 file to be loaded from a BIO or FILE pointer. The BIO version will
8515 for example allow memory BIOs to contain config info.
8518 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8519 Whoever hopes to achieve shared-library compatibility across versions
8520 must use this, not the compile-time macro.
8521 (Exercise 0.9.4: Which is the minimum library version required by
8523 Note: All this applies only to multi-threaded programs, others don't
8527 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8528 through a BIO pair triggered the default case, i.e.
8529 SSLerr(...,SSL_R_UNKNOWN_STATE).
8532 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8533 can use the SSL library even if none of the specific BIOs is
8537 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8538 for the encoded length.
8539 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8541 *) Add initial documentation of the X509V3 functions.
8544 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8545 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8546 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8547 secure PKCS#8 private key format with a high iteration count.
8550 *) Fix determination of Perl interpreter: A perl or perl5
8551 _directory_ in $PATH was also accepted as the interpreter.
8552 [Ralf S. Engelschall]
8554 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8555 wrong with it but it was very old and did things like calling
8556 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8560 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8561 to use the new extension code.
8564 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8565 with macros. This should make it easier to change their form, add extra
8566 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8570 *) Add to configuration table a new entry that can specify an alternative
8571 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8572 according to Mark Crispin <MRC@Panda.COM>.
8576 *) DES CBC did not update the IV. Weird.
8579 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8580 Changing the behaviour of the former might break existing programs --
8581 where IV updating is needed, des_ncbc_encrypt can be used.
8584 *) When bntest is run from "make test" it drives bc to check its
8585 calculations, as well as internally checking them. If an internal check
8586 fails, it needs to cause bc to give a non-zero result or make test carries
8587 on without noticing the failure. Fixed.
8590 *) DES library cleanups.
8593 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8594 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8595 ciphers. NOTE: although the key derivation function has been verified
8596 against some published test vectors it has not been extensively tested
8597 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8601 *) Instead of "mkdir -p", which is not fully portable, use new
8602 Perl script "util/mkdir-p.pl".
8605 *) Rewrite the way password based encryption (PBE) is handled. It used to
8606 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8607 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8608 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8609 the 'parameter' field of the AlgorithmIdentifier is passed to the
8610 underlying key generation function so it must do its own ASN1 parsing.
8611 This has also changed the EVP_PBE_CipherInit() function which now has a
8612 'parameter' argument instead of literal salt and iteration count values
8613 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8616 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8617 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8618 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8619 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8620 value was just used as a "magic string" and not used directly its
8621 value doesn't matter.
8624 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8628 *) "linux-sparc64" configuration (ultrapenguin).
8629 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8630 "linux-sparc" configuration.
8631 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8633 *) config now generates no-xxx options for missing ciphers.
8636 *) Support the EBCDIC character set (work in progress).
8637 File ebcdic.c not yet included because it has a different license.
8638 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8640 *) Support BS2000/OSD-POSIX.
8641 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8643 *) Make callbacks for key generation use void * instead of char *.
8646 *) Make S/MIME samples compile (not yet tested).
8649 *) Additional typesafe stacks.
8652 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8656 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8658 *) New configuration variant "sco5-gcc".
8660 *) Updated some demos.
8661 [Sean O Riordain, Wade Scholine]
8663 *) Add missing BIO_free at exit of pkcs12 application.
8666 *) Fix memory leak in conf.c.
8669 *) Updates for Win32 to assembler version of MD5.
8672 *) Set #! path to perl in apps/der_chop to where we found it
8673 instead of using a fixed path.
8676 *) SHA library changes for irix64-mips4-cc.
8679 *) Improvements for VMS support.
8683 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8685 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8686 This also avoids the problems with SC4.2 and unpatched SC5.
8687 [Andy Polyakov <appro@fy.chalmers.se>]
8689 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8690 These are required because of the typesafe stack would otherwise break
8691 existing code. If old code used a structure member which used to be STACK
8692 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8693 sk_num or sk_value it would produce an error because the num, data members
8694 are not present in STACK_OF. Now it just produces a warning. sk_set
8695 replaces the old method of assigning a value to sk_value
8696 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8697 that does this will no longer work (and should use sk_set instead) but
8698 this could be regarded as a "questionable" behaviour anyway.
8701 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8702 correctly handle encrypted S/MIME data.
8705 *) Change type of various DES function arguments from des_cblock
8706 (which means, in function argument declarations, pointer to char)
8707 to des_cblock * (meaning pointer to array with 8 char elements),
8708 which allows the compiler to do more typechecking; it was like
8709 that back in SSLeay, but with lots of ugly casts.
8711 Introduce new type const_des_cblock.
8714 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8715 problems: find RecipientInfo structure that matches recipient certificate
8716 and initialise the ASN1 structures properly based on passed cipher.
8719 *) Belatedly make the BN tests actually check the results.
8722 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8723 to and from BNs: it was completely broken. New compilation option
8724 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8725 key elements as negative integers.
8728 *) Reorganize and speed up MD5.
8729 [Andy Polyakov <appro@fy.chalmers.se>]
8732 [Richard Levitte <richard@levitte.org>]
8734 *) New option -out to asn1parse to allow the parsed structure to be
8735 output to a file. This is most useful when combined with the -strparse
8736 option to examine the output of things like OCTET STRINGS.
8739 *) Make SSL library a little more fool-proof by not requiring any longer
8740 that SSL_set_{accept,connect}_state be called before
8741 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8742 in many applications because usually everything *appeared* to work as
8743 intended anyway -- now it really works as intended).
8746 *) Move openssl.cnf out of lib/.
8749 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8750 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8751 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8752 [Ralf S. Engelschall]
8754 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8755 handle PKCS#7 enveloped data properly.
8756 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8758 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8759 copying pointers. The cert_st handling is changed by this in
8760 various ways (and thus what used to be known as ctx->default_cert
8761 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8762 any longer when s->cert does not give us what we need).
8763 ssl_cert_instantiate becomes obsolete by this change.
8764 As soon as we've got the new code right (possibly it already is?),
8765 we have solved a couple of bugs of the earlier code where s->cert
8766 was used as if it could not have been shared with other SSL structures.
8768 Note that using the SSL API in certain dirty ways now will result
8769 in different behaviour than observed with earlier library versions:
8770 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8771 does not influence s as it used to.
8773 In order to clean up things more thoroughly, inside SSL_SESSION
8774 we don't use CERT any longer, but a new structure SESS_CERT
8775 that holds per-session data (if available); currently, this is
8776 the peer's certificate chain and, for clients, the server's certificate
8777 and temporary key. CERT holds only those values that can have
8778 meaningful defaults in an SSL_CTX.
8781 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8782 from the internal representation. Various PKCS#7 fixes: remove some
8783 evil casts and set the enc_dig_alg field properly based on the signing
8787 *) Allow PKCS#12 password to be set from the command line or the
8788 environment. Let 'ca' get its config file name from the environment
8789 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8793 *) Allow certificate policies extension to use an IA5STRING for the
8794 organization field. This is contrary to the PKIX definition but
8795 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8799 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8800 without disallowing inline assembler and the like for non-pedantic builds.
8803 *) Support Borland C++ builder.
8804 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8809 *) SHA-1 cleanups and performance enhancements.
8810 [Andy Polyakov <appro@fy.chalmers.se>]
8812 *) Sparc v8plus assembler for the bignum library.
8813 [Andy Polyakov <appro@fy.chalmers.se>]
8815 *) Accept any -xxx and +xxx compiler options in Configure.
8818 *) Update HPUX configuration.
8821 *) Add missing sk_<type>_unshift() function to safestack.h
8822 [Ralf S. Engelschall]
8824 *) New function SSL_CTX_use_certificate_chain_file that sets the
8825 "extra_cert"s in addition to the certificate. (This makes sense
8826 only for "PEM" format files, as chains as a whole are not
8830 *) Support verify_depth from the SSL API.
8831 x509_vfy.c had what can be considered an off-by-one-error:
8832 Its depth (which was not part of the external interface)
8833 was actually counting the number of certificates in a chain;
8834 now it really counts the depth.
8837 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8838 instead of X509err, which often resulted in confusing error
8839 messages since the error codes are not globally unique
8840 (e.g. an alleged error in ssl3_accept when a certificate
8841 didn't match the private key).
8843 *) New function SSL_CTX_set_session_id_context that allows to set a default
8844 value (so that you don't need SSL_set_session_id_context for each
8845 connection using the SSL_CTX).
8848 *) OAEP decoding bug fix.
8851 *) Support INSTALL_PREFIX for package builders, as proposed by
8855 *) New Configure options "threads" and "no-threads". For systems
8856 where the proper compiler options are known (currently Solaris
8857 and Linux), "threads" is the default.
8860 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8863 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8864 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8865 such as /usr/local/bin.
8868 *) "make linux-shared" to build shared libraries.
8869 [Niels Poppe <niels@netbox.org>]
8871 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8874 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8875 extension adding in x509 utility.
8878 *) Remove NOPROTO sections and error code comments.
8881 *) Partial rewrite of the DEF file generator to now parse the ANSI
8885 *) New Configure options --prefix=DIR and --openssldir=DIR.
8888 *) Complete rewrite of the error code script(s). It is all now handled
8889 by one script at the top level which handles error code gathering,
8890 header rewriting and C source file generation. It should be much better
8891 than the old method: it now uses a modified version of Ulf's parser to
8892 read the ANSI prototypes in all header files (thus the old K&R definitions
8893 aren't needed for error creation any more) and do a better job of
8894 translating function codes into names. The old 'ASN1 error code imbedded
8895 in a comment' is no longer necessary and it doesn't use .err files which
8896 have now been deleted. Also the error code call doesn't have to appear all
8897 on one line (which resulted in some large lines...).
8900 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8903 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8904 0 (which usually indicates a closed connection), but continue reading.
8907 *) Fix some race conditions.
8910 *) Add support for CRL distribution points extension. Add Certificate
8911 Policies and CRL distribution points documentation.
8914 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8917 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8918 8 of keying material. Merlin has also confirmed interop with this fix
8919 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8920 [Merlin Hughes <merlin@baltimore.ie>]
8922 *) Fix lots of warnings.
8923 [Richard Levitte <levitte@stacken.kth.se>]
8925 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8926 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8927 [Richard Levitte <levitte@stacken.kth.se>]
8929 *) Fix problems with sizeof(long) == 8.
8930 [Andy Polyakov <appro@fy.chalmers.se>]
8932 *) Change functions to ANSI C.
8935 *) Fix typos in error codes.
8936 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8938 *) Remove defunct assembler files from Configure.
8941 *) SPARC v8 assembler BIGNUM implementation.
8942 [Andy Polyakov <appro@fy.chalmers.se>]
8944 *) Support for Certificate Policies extension: both print and set.
8945 Various additions to support the r2i method this uses.
8948 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8949 return a const string when you are expecting an allocated buffer.
8952 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8953 types DirectoryString and DisplayText.
8956 *) Add code to allow r2i extensions to access the configuration database,
8957 add an LHASH database driver and add several ctx helper functions.
8960 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8961 fail when they extended the size of a BIGNUM.
8964 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8965 support typesafe stack.
8968 *) Fix typo in SSL_[gs]et_options().
8969 [Nils Frostberg <nils@medcom.se>]
8971 *) Delete various functions and files that belonged to the (now obsolete)
8972 old X509V3 handling code.
8975 *) New Configure option "rsaref".
8978 *) Don't auto-generate pem.h.
8981 *) Introduce type-safe ASN.1 SETs.
8984 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8985 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8987 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8988 that links with OpenSSL (well at least cause lots of warnings), but fear
8989 not: the conversion is trivial, and it eliminates loads of evil casts. A
8990 few STACKed things have been converted already. Feel free to convert more.
8991 In the fullness of time, I'll do away with the STACK type altogether.
8994 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8995 specified in <certfile> by updating the entry in the index.txt file.
8996 This way one no longer has to edit the index.txt file manually for
8997 revoking a certificate. The -revoke option does the gory details now.
8998 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9000 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9001 `-text' option at all and this way the `-noout -text' combination was
9002 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9003 [Ralf S. Engelschall]
9005 *) Make sure a corresponding plain text error message exists for the
9006 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9007 verify callback function determined that a certificate was revoked.
9008 [Ralf S. Engelschall]
9010 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9011 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9012 all available cipers including rc5, which was forgotten until now.
9013 In order to let the testing shell script know which algorithms
9014 are available, a new (up to now undocumented) command
9015 "openssl list-cipher-commands" is used.
9018 *) Bugfix: s_client occasionally would sleep in select() when
9019 it should have checked SSL_pending() first.
9022 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9023 the raw DSA values prior to ASN.1 encoding.
9026 *) Tweaks to Configure
9027 [Niels Poppe <niels@netbox.org>]
9029 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9033 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9036 *) New config option to avoid instructions that are illegal on the 80386.
9037 The default code is faster, but requires at least a 486.
9040 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9041 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9042 same as SSL2_VERSION anyway.
9045 *) New "-showcerts" option for s_client.
9048 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9049 application. Various cleanups and fixes.
9052 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9053 modify error routines to work internally. Add error codes and PBE init
9054 to library startup routines.
9057 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9058 packing functions to asn1 and evp. Changed function names and error
9059 codes along the way.
9062 *) PKCS12 integration: and so it begins... First of several patches to
9063 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9064 objects to objects.h
9067 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9068 and display support for Thawte strong extranet extension.
9071 *) Add LinuxPPC support.
9072 [Jeff Dubrule <igor@pobox.org>]
9074 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9075 bn_div_words in alpha.s.
9076 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9078 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9079 OAEP isn't supported when OpenSSL is built with RSAref.
9080 [Ulf Moeller <ulf@fitug.de>]
9082 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9083 so they no longer are missing under -DNOPROTO.
9084 [Soren S. Jorvang <soren@t.dk>]
9087 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9089 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9090 doesn't work when the session is reused. Coming soon!
9093 *) Fix a security hole, that allows sessions to be reused in the wrong
9094 context thus bypassing client cert protection! All software that uses
9095 client certs and session caches in multiple contexts NEEDS PATCHING to
9096 allow session reuse! A fuller solution is in the works.
9097 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9099 *) Some more source tree cleanups (removed obsolete files
9100 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9101 permission on "config" script to be executable) and a fix for the INSTALL
9103 [Ulf Moeller <ulf@fitug.de>]
9105 *) Remove some legacy and erroneous uses of malloc, free instead of
9107 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9109 *) Make rsa_oaep_test return non-zero on error.
9110 [Ulf Moeller <ulf@fitug.de>]
9112 *) Add support for native Solaris shared libraries. Configure
9113 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9114 if someone would make that last step automatic.
9115 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9117 *) ctx_size was not built with the right compiler during "make links". Fixed.
9120 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9121 except NULL ciphers". This means the default cipher list will no longer
9122 enable NULL ciphers. They need to be specifically enabled e.g. with
9123 the string "DEFAULT:eNULL".
9126 *) Fix to RSA private encryption routines: if p < q then it would
9127 occasionally produce an invalid result. This will only happen with
9128 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9131 *) Be less restrictive and allow also `perl util/perlpath.pl
9132 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9133 because this way one can also use an interpreter named `perl5' (which is
9134 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9135 installed as `perl').
9136 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9138 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9139 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9141 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9142 advapi32.lib to Win32 build and change the pem test comparision
9143 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9144 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9145 and crypto/des/ede_cbcm_enc.c.
9148 *) DES quad checksum was broken on big-endian architectures. Fixed.
9151 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9152 Win32 test batch file so it (might) work again. The Win32 test batch file
9153 is horrible: I feel ill....
9156 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9157 in e_os.h. Audit of header files to check ANSI and non ANSI
9158 sections: 10 functions were absent from non ANSI section and not exported
9159 from Windows DLLs. Fixed up libeay.num for new functions.
9162 *) Make `openssl version' output lines consistent.
9163 [Ralf S. Engelschall]
9165 *) Fix Win32 symbol export lists for BIO functions: Added
9166 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9167 to ms/libeay{16,32}.def.
9168 [Ralf S. Engelschall]
9170 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9171 fine under Unix and passes some trivial tests I've now added. But the
9172 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9173 added to make sure no one expects that this stuff really works in the
9174 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9175 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9177 [Ralf S. Engelschall]
9179 *) Fix the generation of two part addresses in perl.
9180 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9182 *) Add config entry for Linux on MIPS.
9183 [John Tobey <jtobey@channel1.com>]
9185 *) Make links whenever Configure is run, unless we are on Windoze.
9188 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9189 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9193 *) Add a useful kludge to allow package maintainers to specify compiler and
9194 other platforms details on the command line without having to patch the
9195 Configure script everytime: One now can use ``perl Configure
9196 <id>:<details>'', i.e. platform ids are allowed to have details appended
9197 to them (seperated by colons). This is treated as there would be a static
9198 pre-configured entry in Configure's %table under key <id> with value
9199 <details> and ``perl Configure <id>'' is called. So, when you want to
9200 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9201 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9202 now, which overrides the FreeBSD-elf entry on-the-fly.
9203 [Ralf S. Engelschall]
9205 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9208 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9209 on the `perl Configure ...' command line. This way one can compile
9210 OpenSSL libraries with Position Independent Code (PIC) which is needed
9211 for linking it into DSOs.
9212 [Ralf S. Engelschall]
9214 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9218 *) Cleaned up the LICENSE document: The official contact for any license
9219 questions now is the OpenSSL core team under openssl-core@openssl.org.
9220 And add a paragraph about the dual-license situation to make sure people
9221 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9222 to the OpenSSL toolkit.
9223 [Ralf S. Engelschall]
9225 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9226 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9227 Additonally cleaned up the `make links' target: Remove unnecessary
9228 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9229 to speed processing and no longer clutter the display with confusing
9230 stuff. Instead only the actually done links are displayed.
9231 [Ralf S. Engelschall]
9233 *) Permit null encryption ciphersuites, used for authentication only. It used
9234 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9235 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9239 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9240 signed attributes when verifying signatures (this would break them),
9241 the detached data encoding was wrong and public keys obtained using
9242 X509_get_pubkey() weren't freed.
9245 *) Add text documentation for the BUFFER functions. Also added a work around
9246 to a Win95 console bug. This was triggered by the password read stuff: the
9247 last character typed gets carried over to the next fread(). If you were
9248 generating a new cert request using 'req' for example then the last
9249 character of the passphrase would be CR which would then enter the first
9253 *) Added the new `Includes OpenSSL Cryptography Software' button as
9254 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9255 button and can be used by applications based on OpenSSL to show the
9256 relationship to the OpenSSL project.
9257 [Ralf S. Engelschall]
9259 *) Remove confusing variables in function signatures in files
9260 ssl/ssl_lib.c and ssl/ssl.h.
9261 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9263 *) Don't install bss_file.c under PREFIX/include/
9264 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9266 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9267 functions that return function pointers and has support for NT specific
9268 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9269 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9270 unsigned to signed types: this was killing the Win32 compile.
9273 *) Add new certificate file to stack functions,
9274 SSL_add_dir_cert_subjects_to_stack() and
9275 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9276 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9277 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9278 This means that Apache-SSL and similar packages don't have to mess around
9279 to add as many CAs as they want to the preferred list.
9282 *) Experiment with doxygen documentation. Currently only partially applied to
9284 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9285 openssl.doxy as the configuration file.
9288 *) Get rid of remaining C++-style comments which strict C compilers hate.
9289 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9291 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9292 compiled in by default: it has problems with large keys.
9295 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9296 DH private keys and/or callback functions which directly correspond to
9297 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9298 is needed for applications which have to configure certificates on a
9299 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9301 For the RSA certificate situation is makes no difference, but
9302 for the DSA certificate situation this fixes the "no shared cipher"
9303 problem where the OpenSSL cipher selection procedure failed because the
9304 temporary keys were not overtaken from the context and the API provided
9305 no way to reconfigure them.
9306 The new functions now let applications reconfigure the stuff and they
9307 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9308 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9309 non-public-API function ssl_cert_instantiate() is used as a helper
9310 function and also to reduce code redundancy inside ssl_rsa.c.
9311 [Ralf S. Engelschall]
9313 *) Move s_server -dcert and -dkey options out of the undocumented feature
9314 area because they are useful for the DSA situation and should be
9315 recognized by the users.
9316 [Ralf S. Engelschall]
9318 *) Fix the cipher decision scheme for export ciphers: the export bits are
9319 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9320 SSL_EXP_MASK. So, the original variable has to be used instead of the
9321 already masked variable.
9322 [Richard Levitte <levitte@stacken.kth.se>]
9324 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9325 [Richard Levitte <levitte@stacken.kth.se>]
9327 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9328 from `int' to `unsigned int' because it's a length and initialized by
9329 EVP_DigestFinal() which expects an `unsigned int *'.
9330 [Richard Levitte <levitte@stacken.kth.se>]
9332 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9333 script. Instead use the usual Shell->Perl transition trick.
9334 [Ralf S. Engelschall]
9336 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9337 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9338 -noout -modulus' as it's already the case for `openssl rsa -noout
9339 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9340 currently the public key is printed (a decision which was already done by
9341 `openssl dsa -modulus' in the past) which serves a similar purpose.
9342 Additionally the NO_RSA no longer completely removes the whole -modulus
9343 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9345 [Ralf S. Engelschall]
9347 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9348 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9349 [Arne Ansper <arne@ats.cyber.ee>]
9351 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9352 to be added. Now both 'req' and 'ca' can use new objects defined in the
9356 *) Add cool BIO that does syslog (or event log on NT).
9357 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9359 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9360 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9361 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9362 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9365 *) Add preliminary config info for new extension code.
9368 *) Make RSA_NO_PADDING really use no padding.
9369 [Ulf Moeller <ulf@fitug.de>]
9371 *) Generate errors when private/public key check is done.
9374 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9375 for some CRL extensions and new objects added.
9378 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9379 key usage extension and fuller support for authority key id.
9382 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9383 padding method for RSA, which is recommended for new applications in PKCS
9384 #1 v2.0 (RFC 2437, October 1998).
9385 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9386 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9387 against Bleichbacher's attack on RSA.
9388 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9391 *) Updates to the new SSL compression code
9392 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9394 *) Fix so that the version number in the master secret, when passed
9395 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9396 (because the server will not accept higher), that the version number
9397 is 0x03,0x01, not 0x03,0x00
9398 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9400 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9401 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9402 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9405 *) Support for RAW extensions where an arbitrary extension can be
9406 created by including its DER encoding. See apps/openssl.cnf for
9410 *) Make sure latest Perl versions don't interpret some generated C array
9411 code as Perl array code in the crypto/err/err_genc.pl script.
9412 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9414 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9415 not many people have the assembler. Various Win32 compilation fixes and
9416 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9420 *) Modify configure script 'Configure' to automatically create crypto/date.h
9421 file under Win32 and also build pem.h from pem.org. New script
9422 util/mkfiles.pl to create the MINFO file on environments that can't do a
9423 'make files': perl util/mkfiles.pl >MINFO should work.
9426 *) Major rework of DES function declarations, in the pursuit of correctness
9427 and purity. As a result, many evil casts evaporated, and some weirdness,
9428 too. You may find this causes warnings in your code. Zapping your evil
9429 casts will probably fix them. Mostly.
9432 *) Fix for a typo in asn1.h. Bug fix to object creation script
9433 obj_dat.pl. It considered a zero in an object definition to mean
9434 "end of object": none of the objects in objects.h have any zeros
9435 so it wasn't spotted.
9436 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9438 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9439 Masking (CBCM). In the absence of test vectors, the best I have been able
9440 to do is check that the decrypt undoes the encrypt, so far. Send me test
9441 vectors if you have them.
9444 *) Correct calculation of key length for export ciphers (too much space was
9445 allocated for null ciphers). This has not been tested!
9448 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9449 message is now correct (it understands "crypto" and "ssl" on its
9450 command line). There is also now an "update" option. This will update
9451 the util/ssleay.num and util/libeay.num files with any new functions.
9453 perl util/mkdef.pl crypto ssl update
9454 it will update them.
9457 *) Overhauled the Perl interface (perl/*):
9458 - ported BN stuff to OpenSSL's different BN library
9459 - made the perl/ source tree CVS-aware
9460 - renamed the package from SSLeay to OpenSSL (the files still contain
9461 their history because I've copied them in the repository)
9462 - removed obsolete files (the test scripts will be replaced
9463 by better Test::Harness variants in the future)
9464 [Ralf S. Engelschall]
9466 *) First cut for a very conservative source tree cleanup:
9467 1. merge various obsolete readme texts into doc/ssleay.txt
9468 where we collect the old documents and readme texts.
9469 2. remove the first part of files where I'm already sure that we no
9470 longer need them because of three reasons: either they are just temporary
9471 files which were left by Eric or they are preserved original files where
9472 I've verified that the diff is also available in the CVS via "cvs diff
9473 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9474 the crypto/md/ stuff).
9475 [Ralf S. Engelschall]
9477 *) More extension code. Incomplete support for subject and issuer alt
9478 name, issuer and authority key id. Change the i2v function parameters
9479 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9480 what that's for :-) Fix to ASN1 macro which messed up
9481 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9484 *) Preliminary support for ENUMERATED type. This is largely copied from the
9488 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9489 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9491 *) Make sure `make rehash' target really finds the `openssl' program.
9492 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9494 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9495 like to hear about it if this slows down other processors.
9498 *) Add CygWin32 platform information to Configure script.
9499 [Alan Batie <batie@aahz.jf.intel.com>]
9501 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9502 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9504 *) New program nseq to manipulate netscape certificate sequences
9507 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9511 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9512 but the BN code had some problems that would cause failures when
9513 doing certificate verification and some other functions.
9514 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9516 *) Add ASN1 and PEM code to support netscape certificate sequences.
9519 *) Add ASN1 and PEM code to support netscape certificate sequences.
9522 *) Add several PKIX and private extended key usage OIDs.
9525 *) Modify the 'ca' program to handle the new extension code. Modify
9526 openssl.cnf for new extension format, add comments.
9529 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9530 and add a sample to openssl.cnf so req -x509 now adds appropriate
9534 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9535 error code, add initial support to X509_print() and x509 application.
9538 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9539 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9540 stuff is currently isolated and isn't even compiled yet.
9543 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9544 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9545 Removed the versions check from X509 routines when loading extensions:
9546 this allows certain broken certificates that don't set the version
9547 properly to be processed.
9550 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9551 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9552 can still be regenerated with "make depend".
9555 *) Spelling mistake in C version of CAST-128.
9556 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9558 *) Changes to the error generation code. The perl script err-code.pl
9559 now reads in the old error codes and retains the old numbers, only
9560 adding new ones if necessary. It also only changes the .err files if new
9561 codes are added. The makefiles have been modified to only insert errors
9562 when needed (to avoid needlessly modifying header files). This is done
9563 by only inserting errors if the .err file is newer than the auto generated
9564 C file. To rebuild all the error codes from scratch (the old behaviour)
9565 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9566 or delete all the .err files.
9569 *) CAST-128 was incorrectly implemented for short keys. The C version has
9570 been fixed, but is untested. The assembler versions are also fixed, but
9571 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9572 to regenerate it if needed.
9573 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9574 Hagino <itojun@kame.net>]
9576 *) File was opened incorrectly in randfile.c.
9577 [Ulf Möller <ulf@fitug.de>]
9579 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9580 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9581 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9582 al: it's just almost always a UTCTime. Note this patch adds new error
9583 codes so do a "make errors" if there are problems.
9586 *) Correct Linux 1 recognition in config.
9587 [Ulf Möller <ulf@fitug.de>]
9589 *) Remove pointless MD5 hash when using DSA keys in ca.
9590 [Anonymous <nobody@replay.com>]
9592 *) Generate an error if given an empty string as a cert directory. Also
9593 generate an error if handed NULL (previously returned 0 to indicate an
9594 error, but didn't set one).
9595 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9597 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9600 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9601 parameters. This was causing a warning which killed off the Win32 compile.
9604 *) Remove C++ style comments from crypto/bn/bn_local.h.
9605 [Neil Costigan <neil.costigan@celocom.com>]
9607 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9608 based on a text string, looking up short and long names and finally
9609 "dot" format. The "dot" format stuff didn't work. Added new function
9610 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9611 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9612 OID is not part of the table.
9615 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9616 X509_LOOKUP_by_alias().
9619 *) Sort openssl functions by name.
9622 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9623 encryption from sample DSA keys (in case anyone is interested the password
9627 *) Make _all_ *_free functions accept a NULL pointer.
9628 [Frans Heymans <fheymans@isaserver.be>]
9630 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9632 [Anonymous <nobody@replay.com>]
9634 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9635 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9637 *) Don't blow it for numeric -newkey arguments to apps/req.
9638 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9640 *) Temp key "for export" tests were wrong in s3_srvr.c.
9641 [Anonymous <nobody@replay.com>]
9643 *) Add prototype for temp key callback functions
9644 SSL_CTX_set_tmp_{rsa,dh}_callback().
9647 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9648 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9651 *) X509_name_add_entry() freed the wrong thing after an error.
9652 [Arne Ansper <arne@ats.cyber.ee>]
9654 *) rsa_eay.c would attempt to free a NULL context.
9655 [Arne Ansper <arne@ats.cyber.ee>]
9657 *) BIO_s_socket() had a broken should_retry() on Windoze.
9658 [Arne Ansper <arne@ats.cyber.ee>]
9660 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9661 [Arne Ansper <arne@ats.cyber.ee>]
9663 *) Make sure the already existing X509_STORE->depth variable is initialized
9664 in X509_STORE_new(), but document the fact that this variable is still
9665 unused in the certificate verification process.
9666 [Ralf S. Engelschall]
9668 *) Fix the various library and apps files to free up pkeys obtained from
9669 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9672 *) Fix reference counting in X509_PUBKEY_get(). This makes
9673 demos/maurice/example2.c work, amongst others, probably.
9674 [Steve Henson and Ben Laurie]
9676 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9677 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9678 are no longer created. This way we have a single and consistent command
9679 line interface `openssl <command>', similar to `cvs <command>'.
9680 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9682 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9683 BIT STRING wrapper always have zero unused bits.
9686 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9689 *) Make the top-level INSTALL documentation easier to understand.
9692 *) Makefiles updated to exit if an error occurs in a sub-directory
9693 make (including if user presses ^C) [Paul Sutton]
9695 *) Make Montgomery context stuff explicit in RSA data structure.
9698 *) Fix build order of pem and err to allow for generated pem.h.
9701 *) Fix renumbering bug in X509_NAME_delete_entry().
9704 *) Enhanced the err-ins.pl script so it makes the error library number
9705 global and can add a library name. This is needed for external ASN1 and
9706 other error libraries.
9709 *) Fixed sk_insert which never worked properly.
9712 *) Fix ASN1 macros so they can handle indefinite length construted
9713 EXPLICIT tags. Some non standard certificates use these: they can now
9717 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9718 into a single doc/ssleay.txt bundle. This way the information is still
9719 preserved but no longer messes up this directory. Now it's new room for
9720 the new set of documenation files.
9721 [Ralf S. Engelschall]
9723 *) SETs were incorrectly DER encoded. This was a major pain, because they
9724 shared code with SEQUENCEs, which aren't coded the same. This means that
9725 almost everything to do with SETs or SEQUENCEs has either changed name or
9726 number of arguments.
9727 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9729 *) Fix test data to work with the above.
9732 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9733 was already fixed by Eric for 0.9.1 it seems.
9734 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9736 *) Autodetect FreeBSD3.
9739 *) Fix various bugs in Configure. This affects the following platforms:
9743 unixware-2.0-pentium
9747 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9748 before they are needed.
9751 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9755 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9757 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9758 changed SSLeay to OpenSSL in version strings.
9759 [Ralf S. Engelschall]
9761 *) Some fixups to the top-level documents.
9764 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9765 because the symlink to include/ was missing.
9766 [Ralf S. Engelschall]
9768 *) Incorporated the popular no-RSA/DSA-only patches
9769 which allow to compile a RSA-free SSLeay.
9770 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9772 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9773 when "ssleay" is still not found.
9774 [Ralf S. Engelschall]
9776 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9777 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9779 *) Updated the README file.
9780 [Ralf S. Engelschall]
9782 *) Added various .cvsignore files in the CVS repository subdirs
9783 to make a "cvs update" really silent.
9784 [Ralf S. Engelschall]
9786 *) Recompiled the error-definition header files and added
9787 missing symbols to the Win32 linker tables.
9788 [Ralf S. Engelschall]
9790 *) Cleaned up the top-level documents;
9791 o new files: CHANGES and LICENSE
9792 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9793 o merged COPYRIGHT into LICENSE
9794 o removed obsolete TODO file
9795 o renamed MICROSOFT to INSTALL.W32
9796 [Ralf S. Engelschall]
9798 *) Removed dummy files from the 0.9.1b source tree:
9799 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9800 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9801 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9802 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9803 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9804 [Ralf S. Engelschall]
9806 *) Added various platform portability fixes.
9809 *) The Genesis of the OpenSSL rpject:
9810 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9811 Young and Tim J. Hudson created while they were working for C2Net until
9813 [The OpenSSL Project]
9816 Changes between 0.9.0b and 0.9.1b [not released]
9818 *) Updated a few CA certificates under certs/
9821 *) Changed some BIGNUM api stuff.
9824 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9825 DGUX x86, Linux Alpha, etc.
9828 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9829 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9833 *) Add -strparse option to asn1pars program which parses nested
9835 [Dr Stephen Henson <shenson@bigfoot.com>]
9837 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9840 *) DSA fix for "ca" program.
9843 *) Added "-genkey" option to "dsaparam" program.
9846 *) Added RIPE MD160 (rmd160) message digest.
9849 *) Added -a (all) option to "ssleay version" command.
9852 *) Added PLATFORM define which is the id given to Configure.
9855 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9858 *) Extended the ASN.1 parser routines.
9861 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9864 *) Added a BN_CTX to the BN library.
9867 *) Fixed the weak key values in DES library
9870 *) Changed API in EVP library for cipher aliases.
9873 *) Added support for RC2/64bit cipher.
9876 *) Converted the lhash library to the crypto/mem.c functions.
9879 *) Added more recognized ASN.1 object ids.
9882 *) Added more RSA padding checks for SSL/TLS.
9885 *) Added BIO proxy/filter functionality.
9888 *) Added extra_certs to SSL_CTX which can be used
9889 send extra CA certificates to the client in the CA cert chain sending
9890 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9893 *) Now Fortezza is denied in the authentication phase because
9894 this is key exchange mechanism is not supported by SSLeay at all.
9897 *) Additional PKCS1 checks.
9900 *) Support the string "TLSv1" for all TLS v1 ciphers.
9903 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9904 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9907 *) Fixed a few memory leaks.
9910 *) Fixed various code and comment typos.
9913 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9914 bytes sent in the client random.
9915 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob