5 Changes between 1.0.0l and 1.0.0m [xx XXX xxxx]
9 Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
11 *) Keep original DTLS digest and encryption contexts in retransmission
12 structures so we can use the previous session parameters if they need
13 to be resent. (CVE-2013-6450)
16 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
17 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
18 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
19 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
20 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
21 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
22 [Rob Stradling, Adam Langley]
24 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
26 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
28 This addresses the flaw in CBC record processing discovered by
29 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
30 at: http://www.isg.rhul.ac.uk/tls/
32 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
33 Security Group at Royal Holloway, University of London
34 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
35 Emilia Käsper for the initial patch.
37 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
39 *) Return an error when checking OCSP signatures when key is NULL.
40 This fixes a DoS attack. (CVE-2013-0166)
43 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
44 the right response is stapled. Also change SSL_get_certificate()
45 so it returns the certificate actually sent.
46 See http://rt.openssl.org/Ticket/Display.html?id=2836.
48 [Rob Stradling <rob.stradling@comodo.com>]
50 *) Fix possible deadlock when decoding public keys.
53 Changes between 1.0.0i and 1.0.0j [10 May 2012]
55 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
58 *) Sanity check record length before skipping explicit IV in DTLS
61 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
62 fuzzing as a service testing platform.
66 *) Initialise tkeylen properly when encrypting CMS messages.
67 Thanks to Solar Designer of Openwall for reporting this issue.
70 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
72 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
73 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
74 in CRYPTO_realloc_clean.
76 Thanks to Tavis Ormandy, Google Security Team, for discovering this
77 issue and to Adam Langley <agl@chromium.org> for fixing it.
79 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
81 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
83 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
84 in CMS and PKCS7 code. When RSA decryption fails use a random key for
85 content decryption and always return the same error. Note: this attack
86 needs on average 2^20 messages so it only affects automated senders. The
87 old behaviour can be reenabled in the CMS code by setting the
88 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
89 an MMA defence is not necessary.
90 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
91 this issue. (CVE-2012-0884)
94 *) Fix CVE-2011-4619: make sure we really are receiving a
95 client hello before rejecting multiple SGC restarts. Thanks to
96 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
99 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
101 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
102 Thanks to Antonio Martin, Enterprise Secure Access Research and
103 Development, Cisco Systems, Inc. for discovering this bug and
104 preparing a fix. (CVE-2012-0050)
107 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
109 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
110 of the Vaudenay padding oracle attack on CBC mode encryption
111 which enables an efficient plaintext recovery attack against
112 the OpenSSL implementation of DTLS. Their attack exploits timing
113 differences arising during decryption processing. A research
114 paper describing this attack can be found at:
115 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
116 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
117 Security Group at Royal Holloway, University of London
118 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
119 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
120 for preparing the fix. (CVE-2011-4108)
121 [Robin Seggelmann, Michael Tuexen]
123 *) Clear bytes used for block padding of SSL 3.0 records.
125 [Adam Langley (Google)]
127 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
128 Kadianakis <desnacked@gmail.com> for discovering this issue and
129 Adam Langley for preparing the fix. (CVE-2011-4619)
130 [Adam Langley (Google)]
132 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
133 [Andrey Kulikov <amdeich@gmail.com>]
135 *) Prevent malformed RFC3779 data triggering an assertion failure.
136 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
137 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
138 [Rob Austein <sra@hactrn.net>]
140 *) Improved PRNG seeding for VOS.
141 [Paul Green <Paul.Green@stratus.com>]
143 *) Fix ssl_ciph.c set-up race.
144 [Adam Langley (Google)]
146 *) Fix spurious failures in ecdsatest.c.
147 [Emilia Käsper (Google)]
149 *) Fix the BIO_f_buffer() implementation (which was mixing different
150 interpretations of the '..._len' fields).
151 [Adam Langley (Google)]
153 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
154 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
155 threads won't reuse the same blinding coefficients.
157 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
158 lock to call BN_BLINDING_invert_ex, and avoids one use of
159 BN_BLINDING_update for each BN_BLINDING structure (previously,
160 the last update always remained unused).
161 [Emilia Käsper (Google)]
163 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
164 [Bob Buckholz (Google)]
166 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
168 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
169 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
170 [Kaspar Brand <ossl@velox.ch>]
172 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
173 for multi-threaded use of ECDH. (CVE-2011-3210)
174 [Adam Langley (Google)]
176 *) Fix x509_name_ex_d2i memory leak on bad inputs.
179 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
180 signature public key algorithm by using OID xref utilities instead.
181 Before this you could only use some ECC ciphersuites with SHA1 only.
184 *) Add protection against ECDSA timing attacks as mentioned in the paper
185 by Billy Bob Brumley and Nicola Tuveri, see:
187 http://eprint.iacr.org/2011/232.pdf
189 [Billy Bob Brumley and Nicola Tuveri]
191 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
193 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
194 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
196 *) Fix bug in string printing code: if *any* escaping is enabled we must
197 escape the escape character (backslash) or the resulting string is
201 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
203 *) Disable code workaround for ancient and obsolete Netscape browsers
204 and servers: an attacker can use it in a ciphersuite downgrade attack.
205 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
208 *) Fixed J-PAKE implementation error, originally discovered by
209 Sebastien Martini, further info and confirmation from Stefan
210 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
213 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
215 *) Fix extension code to avoid race conditions which can result in a buffer
216 overrun vulnerability: resumed sessions must not be modified as they can
217 be shared by multiple threads. CVE-2010-3864
220 *) Fix WIN32 build system to correctly link an ENGINE directory into
224 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
226 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
228 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
230 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
232 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
233 context. The operation can be customised via the ctrl mechanism in
234 case ENGINEs want to include additional functionality.
237 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
240 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
241 output hashes compatible with older versions of OpenSSL.
242 [Willy Weisz <weisz@vcpc.univie.ac.at>]
244 *) Fix compression algorithm handling: if resuming a session use the
245 compression algorithm of the resumed session instead of determining
246 it from client hello again. Don't allow server to change algorithm.
249 *) Add load_crls() function to apps tidying load_certs() too. Add option
250 to verify utility to allow additional CRLs to be included.
253 *) Update OCSP request code to permit adding custom headers to the request:
254 some responders need this.
257 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
259 [Julia Lawall <julia@diku.dk>]
261 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
262 needlessly dereferenced structures, used obsolete functions and
263 didn't handle all updated verify codes correctly.
266 *) Disable MD2 in the default configuration.
269 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
270 indicate the initial BIO being pushed or popped. This makes it possible
271 to determine whether the BIO is the one explicitly called or as a result
272 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
273 it handles reference counts correctly and doesn't zero out the I/O bio
274 when it is not being explicitly popped. WARNING: applications which
275 included workarounds for the old buggy behaviour will need to be modified
276 or they could free up already freed BIOs.
279 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
280 renaming to all platforms (within the 0.9.8 branch, this was
281 done conditionally on Netware platforms to avoid a name clash).
282 [Guenter <lists@gknw.net>]
284 *) Add ECDHE and PSK support to DTLS.
285 [Michael Tuexen <tuexen@fh-muenster.de>]
287 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
291 *) Add "missing" function EVP_MD_flags() (without this the only way to
292 retrieve a digest flags is by accessing the structure directly. Update
293 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
294 or cipher is registered as in the "from" argument. Print out all
295 registered digests in the dgst usage message instead of manually
296 attempting to work them out.
299 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
300 this allows the use of compression and extensions. Change default cipher
301 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
302 by default unless an application cipher string requests it.
305 *) Alter match criteria in PKCS12_parse(). It used to try to use local
306 key ids to find matching certificates and keys but some PKCS#12 files
307 don't follow the (somewhat unwritten) rules and this strategy fails.
308 Now just gather all certificates together and the first private key
309 then look for the first certificate that matches the key.
312 *) Support use of registered digest and cipher names for dgst and cipher
313 commands instead of having to add each one as a special case. So now
320 openssl dgst -sha256 foo
322 and this works for ENGINE based algorithms too.
326 *) Update Gost ENGINE to support parameter files.
327 [Victor B. Wagner <vitus@cryptocom.ru>]
329 *) Support GeneralizedTime in ca utility.
330 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
332 *) Enhance the hash format used for certificate directory links. The new
333 form uses the canonical encoding (meaning equivalent names will work
334 even if they aren't identical) and uses SHA1 instead of MD5. This form
335 is incompatible with the older format and as a result c_rehash should
336 be used to rebuild symbolic links.
339 *) Make PKCS#8 the default write format for private keys, replacing the
340 traditional format. This form is standardised, more secure and doesn't
341 include an implicit MD5 dependency.
344 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
345 committed to OpenSSL should pass this lot as a minimum.
348 *) Add session ticket override functionality for use by EAP-FAST.
349 [Jouni Malinen <j@w1.fi>]
351 *) Modify HMAC functions to return a value. Since these can be implemented
352 in an ENGINE errors can occur.
355 *) Type-checked OBJ_bsearch_ex.
358 *) Type-checked OBJ_bsearch. Also some constification necessitated
359 by type-checking. Still to come: TXT_DB, bsearch(?),
360 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
364 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
365 seconds to a tm structure directly, instead of going through OS
366 specific date routines. This avoids any issues with OS routines such
367 as the year 2038 bug. New *_adj() functions for ASN1 time structures
368 and X509_time_adj_ex() to cover the extended range. The existing
369 X509_time_adj() is still usable and will no longer have any date issues.
372 *) Delta CRL support. New use deltas option which will attempt to locate
373 and search any appropriate delta CRLs available.
375 This work was sponsored by Google.
378 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
379 code and add additional score elements. Validate alternate CRL paths
380 as part of the CRL checking and indicate a new error "CRL path validation
381 error" in this case. Applications wanting additional details can use
382 the verify callback and check the new "parent" field. If this is not
383 NULL CRL path validation is taking place. Existing applications wont
384 see this because it requires extended CRL support which is off by
387 This work was sponsored by Google.
390 *) Support for freshest CRL extension.
392 This work was sponsored by Google.
395 *) Initial indirect CRL support. Currently only supported in the CRLs
396 passed directly and not via lookup. Process certificate issuer
397 CRL entry extension and lookup CRL entries by bother issuer name
398 and serial number. Check and process CRL issuer entry in IDP extension.
400 This work was sponsored by Google.
403 *) Add support for distinct certificate and CRL paths. The CRL issuer
404 certificate is validated separately in this case. Only enabled if
405 an extended CRL support flag is set: this flag will enable additional
406 CRL functionality in future.
408 This work was sponsored by Google.
411 *) Add support for policy mappings extension.
413 This work was sponsored by Google.
416 *) Fixes to pathlength constraint, self issued certificate handling,
417 policy processing to align with RFC3280 and PKITS tests.
419 This work was sponsored by Google.
422 *) Support for name constraints certificate extension. DN, email, DNS
423 and URI types are currently supported.
425 This work was sponsored by Google.
428 *) To cater for systems that provide a pointer-based thread ID rather
429 than numeric, deprecate the current numeric thread ID mechanism and
430 replace it with a structure and associated callback type. This
431 mechanism allows a numeric "hash" to be extracted from a thread ID in
432 either case, and on platforms where pointers are larger than 'long',
433 mixing is done to help ensure the numeric 'hash' is usable even if it
434 can't be guaranteed unique. The default mechanism is to use "&errno"
435 as a pointer-based thread ID to distinguish between threads.
437 Applications that want to provide their own thread IDs should now use
438 CRYPTO_THREADID_set_callback() to register a callback that will call
439 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
441 Note that ERR_remove_state() is now deprecated, because it is tied
442 to the assumption that thread IDs are numeric. ERR_remove_state(0)
443 to free the current thread's error state should be replaced by
444 ERR_remove_thread_state(NULL).
446 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
447 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
448 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
449 application was previously providing a numeric thread callback that
450 was inappropriate for distinguishing threads, then uniqueness might
451 have been obtained with &errno that happened immediately in the
452 intermediate development versions of OpenSSL; this is no longer the
453 case, the numeric thread callback will now override the automatic use
455 [Geoff Thorpe, with help from Bodo Moeller]
457 *) Initial support for different CRL issuing certificates. This covers a
458 simple case where the self issued certificates in the chain exist and
459 the real CRL issuer is higher in the existing chain.
461 This work was sponsored by Google.
464 *) Removed effectively defunct crypto/store from the build.
467 *) Revamp of STACK to provide stronger type-checking. Still to come:
468 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
469 ASN1_STRING, CONF_VALUE.
472 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
473 RAM on SSL connections. This option can save about 34k per idle SSL.
476 *) Revamp of LHASH to provide stronger type-checking. Still to come:
477 STACK, TXT_DB, bsearch, qsort.
480 *) Initial support for Cryptographic Message Syntax (aka CMS) based
481 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
482 support for data, signedData, compressedData, digestedData and
483 encryptedData, envelopedData types included. Scripts to check against
484 RFC4134 examples draft and interop and consistency checks of many
485 content types and variants.
488 *) Add options to enc utility to support use of zlib compression BIO.
491 *) Extend mk1mf to support importing of options and assembly language
492 files from Configure script, currently only included in VC-WIN32.
493 The assembly language rules can now optionally generate the source
494 files from the associated perl scripts.
497 *) Implement remaining functionality needed to support GOST ciphersuites.
498 Interop testing has been performed using CryptoPro implementations.
499 [Victor B. Wagner <vitus@cryptocom.ru>]
501 *) s390x assembler pack.
504 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
508 *) Implement Opaque PRF Input TLS extension as specified in
509 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
510 official specification yet and no extension type assignment by
511 IANA exists, this extension (for now) will have to be explicitly
512 enabled when building OpenSSL by providing the extension number
513 to use. For example, specify an option
515 -DTLSEXT_TYPE_opaque_prf_input=0x9527
517 to the "config" or "Configure" script to enable the extension,
518 assuming extension number 0x9527 (which is a completely arbitrary
519 and unofficial assignment based on the MD5 hash of the Internet
520 Draft). Note that by doing so, you potentially lose
521 interoperability with other TLS implementations since these might
522 be using the same extension number for other purposes.
524 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
525 opaque PRF input value to use in the handshake. This will create
526 an interal copy of the length-'len' string at 'src', and will
527 return non-zero for success.
529 To get more control and flexibility, provide a callback function
532 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
533 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
537 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
540 Callback function 'cb' will be called in handshakes, and is
541 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
542 Argument 'arg' is for application purposes (the value as given to
543 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
544 be provided to the callback function). The callback function
545 has to return non-zero to report success: usually 1 to use opaque
546 PRF input just if possible, or 2 to enforce use of the opaque PRF
547 input. In the latter case, the library will abort the handshake
548 if opaque PRF input is not successfully negotiated.
550 Arguments 'peerinput' and 'len' given to the callback function
551 will always be NULL and 0 in the case of a client. A server will
552 see the client's opaque PRF input through these variables if
553 available (NULL and 0 otherwise). Note that if the server
554 provides an opaque PRF input, the length must be the same as the
555 length of the client's opaque PRF input.
557 Note that the callback function will only be called when creating
558 a new session (session resumption can resume whatever was
559 previously negotiated), and will not be called in SSL 2.0
560 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
561 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
562 for applications that need to enforce opaque PRF input.
566 *) Update ssl code to support digests other than SHA1+MD5 for handshake
569 [Victor B. Wagner <vitus@cryptocom.ru>]
571 *) Add RFC4507 support to OpenSSL. This includes the corrections in
572 RFC4507bis. The encrypted ticket format is an encrypted encoded
573 SSL_SESSION structure, that way new session features are automatically
576 If a client application caches session in an SSL_SESSION structure
577 support is transparent because tickets are now stored in the encoded
580 The SSL_CTX structure automatically generates keys for ticket
581 protection in servers so again support should be possible
582 with no application modification.
584 If a client or server wishes to disable RFC4507 support then the option
585 SSL_OP_NO_TICKET can be set.
587 Add a TLS extension debugging callback to allow the contents of any client
588 or server extensions to be examined.
590 This work was sponsored by Google.
593 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
594 OpenSSL should now compile cleanly on gcc 4.2
595 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
597 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
598 support including streaming MAC support: this is required for GOST
600 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
602 *) Add option -stream to use PKCS#7 streaming in smime utility. New
603 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
604 to output in BER and PEM format.
607 *) Experimental support for use of HMAC via EVP_PKEY interface. This
608 allows HMAC to be handled via the EVP_DigestSign*() interface. The
609 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
610 ENGINE support for HMAC keys which are unextractable. New -mac and
611 -macopt options to dgst utility.
614 *) New option -sigopt to dgst utility. Update dgst to use
615 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
616 alternative signing paramaters such as X9.31 or PSS in the dgst
620 *) Change ssl_cipher_apply_rule(), the internal function that does
621 the work each time a ciphersuite string requests enabling
622 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
623 removing ("!foo+bar") a class of ciphersuites: Now it maintains
624 the order of disabled ciphersuites such that those ciphersuites
625 that most recently went from enabled to disabled not only stay
626 in order with respect to each other, but also have higher priority
627 than other disabled ciphersuites the next time ciphersuites are
630 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
631 the same ciphersuites as with "HIGH" alone, but in a specific
632 order where the PSK ciphersuites come first (since they are the
633 most recently disabled ciphersuites when "HIGH" is parsed).
635 Also, change ssl_create_cipher_list() (using this new
636 funcionality) such that between otherwise identical
637 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
641 *) Change ssl_create_cipher_list() so that it automatically
642 arranges the ciphersuites in reasonable order before starting
643 to process the rule string. Thus, the definition for "DEFAULT"
644 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
645 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
646 This makes it much easier to arrive at a reasonable default order
647 in applications for which anonymous ciphers are OK (meaning
648 that you can't actually use DEFAULT).
649 [Bodo Moeller; suggested by Victor Duchovni]
651 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
652 processing) into multiple integers instead of setting
653 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
654 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
655 (These masks as well as the individual bit definitions are hidden
656 away into the non-exported interface ssl/ssl_locl.h, so this
657 change to the definition of the SSL_CIPHER structure shouldn't
658 affect applications.) This give us more bits for each of these
659 categories, so there is no longer a need to coagulate AES128 and
660 AES256 into a single algorithm bit, and to coagulate Camellia128
661 and Camellia256 into a single algorithm bit, which has led to all
664 Thus, among other things, the kludge introduced in 0.9.7m and
665 0.9.8e for masking out AES256 independently of AES128 or masking
666 out Camellia256 independently of AES256 is not needed here in 0.9.9.
668 With the change, we also introduce new ciphersuite aliases that
669 so far were missing: "AES128", "AES256", "CAMELLIA128", and
673 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
674 Use the leftmost N bytes of the signature input if the input is
675 larger than the prime q (with N being the size in bytes of q).
678 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
679 it yet and it is largely untested.
682 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
685 *) Initial incomplete changes to avoid need for function casts in OpenSSL
686 some compilers (gcc 4.2 and later) reject their use. Safestack is
687 reimplemented. Update ASN1 to avoid use of legacy functions.
690 *) Win32/64 targets are linked with Winsock2.
693 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
694 to external functions. This can be used to increase CRL handling
695 efficiency especially when CRLs are very large by (for example) storing
696 the CRL revoked certificates in a database.
699 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
700 new CRLs added to a directory can be used. New command line option
701 -verify_return_error to s_client and s_server. This causes real errors
702 to be returned by the verify callback instead of carrying on no matter
703 what. This reflects the way a "real world" verify callback would behave.
706 *) GOST engine, supporting several GOST algorithms and public key formats.
707 Kindly donated by Cryptocom.
710 *) Partial support for Issuing Distribution Point CRL extension. CRLs
711 partitioned by DP are handled but no indirect CRL or reason partitioning
712 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
713 selected via a scoring technique which handles IDP and AKID in CRLs.
716 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
717 will ultimately be used for all verify operations: this will remove the
718 X509_STORE dependency on certificate verification and allow alternative
719 lookup methods. X509_STORE based implementations of these two callbacks.
722 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
723 Modify get_crl() to find a valid (unexpired) CRL if possible.
726 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
727 this would be called X509_CRL_cmp() but that name is already used by
728 a function that just compares CRL issuer names. Cache several CRL
729 extensions in X509_CRL structure and cache CRLDP in X509.
732 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
733 this maps equivalent X509_NAME structures into a consistent structure.
734 Name comparison can then be performed rapidly using memcmp().
737 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
741 *) Allow digests to supply their own micalg string for S/MIME type using
742 the ctrl EVP_MD_CTRL_MICALG.
745 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
746 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
747 ctrl. It can then customise the structure before and/or after signing
751 *) New function OBJ_add_sigid() to allow application defined signature OIDs
752 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
753 to free up any added signature OIDs.
756 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
757 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
758 digest and cipher tables. New options added to openssl utility:
759 list-message-digest-algorithms and list-cipher-algorithms.
762 *) Change the array representation of binary polynomials: the list
763 of degrees of non-zero coefficients is now terminated with -1.
764 Previously it was terminated with 0, which was also part of the
765 value; thus, the array representation was not applicable to
766 polynomials where t^0 has coefficient zero. This change makes
767 the array representation useful in a more general context.
770 *) Various modifications and fixes to SSL/TLS cipher string
771 handling. For ECC, the code now distinguishes between fixed ECDH
772 with RSA certificates on the one hand and with ECDSA certificates
773 on the other hand, since these are separate ciphersuites. The
774 unused code for Fortezza ciphersuites has been removed.
776 For consistency with EDH, ephemeral ECDH is now called "EECDH"
777 (not "ECDHE"). For consistency with the code for DH
778 certificates, use of ECDH certificates is now considered ECDH
779 authentication, not RSA or ECDSA authentication (the latter is
780 merely the CA's signing algorithm and not actively used in the
783 The temporary ciphersuite alias "ECCdraft" is no longer
784 available, and ECC ciphersuites are no longer excluded from "ALL"
785 and "DEFAULT". The following aliases now exist for RFC 4492
786 ciphersuites, most of these by analogy with the DH case:
788 kECDHr - ECDH cert, signed with RSA
789 kECDHe - ECDH cert, signed with ECDSA
790 kECDH - ECDH cert (signed with either RSA or ECDSA)
791 kEECDH - ephemeral ECDH
792 ECDH - ECDH cert or ephemeral ECDH
798 AECDH - anonymous ECDH
799 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
803 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
804 Use correct micalg parameters depending on digest(s) in signed message.
807 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
808 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
811 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
812 an engine to register a method. Add ENGINE lookups for methods and
813 functional reference processing.
816 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
817 EVP_{Sign,Verify}* which allow an application to customise the signature
821 *) New -resign option to smime utility. This adds one or more signers
822 to an existing PKCS#7 signedData structure. Also -md option to use an
823 alternative message digest algorithm for signing.
826 *) Tidy up PKCS#7 routines and add new functions to make it easier to
827 create PKCS7 structures containing multiple signers. Update smime
828 application to support multiple signers.
831 *) New -macalg option to pkcs12 utility to allow setting of an alternative
835 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
836 Reorganize PBE internals to lookup from a static table using NIDs,
837 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
838 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
839 PRF which will be automatically used with PBES2.
842 *) Replace the algorithm specific calls to generate keys in "req" with the
846 *) Update PKCS#7 enveloped data routines to use new API. This is now
847 supported by any public key method supporting the encrypt operation. A
848 ctrl is added to allow the public key algorithm to examine or modify
849 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
853 *) Add a ctrl to asn1 method to allow a public key algorithm to express
854 a default digest type to use. In most cases this will be SHA1 but some
855 algorithms (such as GOST) need to specify an alternative digest. The
856 return value indicates how strong the prefernce is 1 means optional and
857 2 is mandatory (that is it is the only supported type). Modify
858 ASN1_item_sign() to accept a NULL digest argument to indicate it should
859 use the default md. Update openssl utilities to use the default digest
860 type for signing if it is not explicitly indicated.
863 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
864 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
865 signing method from the key type. This effectively removes the link
866 between digests and public key types.
869 *) Add an OID cross reference table and utility functions. Its purpose is to
870 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
871 rsaEncryption. This will allow some of the algorithm specific hackery
872 needed to use the correct OID to be removed.
875 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
876 structures for PKCS7_sign(). They are now set up by the relevant public
880 *) Add provisional EC pkey method with support for ECDSA and ECDH.
883 *) Add support for key derivation (agreement) in the API, DH method and
887 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
888 public and private key formats. As a side effect these add additional
889 command line functionality not previously available: DSA signatures can be
890 generated and verified using pkeyutl and DH key support and generation in
895 [Oliver Tappe <zooey@hirschkaefer.de>]
897 *) New make target "install_html_docs" installs HTML renditions of the
899 [Oliver Tappe <zooey@hirschkaefer.de>]
901 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
902 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
903 support key and parameter generation and add initial key generation
904 functionality for RSA.
907 *) Add functions for main EVP_PKEY_method operations. The undocumented
908 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
909 EVP_PKEY_{encrypt,decrypt}_old.
912 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
913 key API, doesn't do much yet.
916 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
917 public key algorithms. New option to openssl utility:
918 "list-public-key-algorithms" to print out info.
921 *) Implement the Supported Elliptic Curves Extension for
922 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
925 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
926 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
929 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
930 utilities such as rsa, dsa, dsaparam etc except they process any key
934 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
935 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
936 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
940 *) Initial support for pluggable public key ASN1.
941 De-spaghettify the public key ASN1 handling. Move public and private
942 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
943 algorithm specific handling to a single module within the relevant
944 algorithm directory. Add functions to allow (near) opaque processing
945 of public and private key structures.
948 *) Implement the Supported Point Formats Extension for
949 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
952 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
953 for the psk identity [hint] and the psk callback functions to the
954 SSL_SESSION, SSL and SSL_CTX structure.
957 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
961 SSL_CTX_use_psk_identity_hint
962 SSL_get_psk_identity_hint
964 SSL_use_psk_identity_hint
966 [Mika Kousa and Pasi Eronen of Nokia Corporation]
968 *) Add RFC 3161 compliant time stamp request creation, response generation
969 and response verification functionality.
970 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
972 *) Add initial support for TLS extensions, specifically for the server_name
973 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
974 have new members for a host name. The SSL data structure has an
975 additional member SSL_CTX *initial_ctx so that new sessions can be
976 stored in that context to allow for session resumption, even after the
977 SSL has been switched to a new SSL_CTX in reaction to a client's
978 server_name extension.
980 New functions (subject to change):
983 SSL_get_servername_type()
986 New CTRL codes and macros (subject to change):
988 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
989 - SSL_CTX_set_tlsext_servername_callback()
990 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
991 - SSL_CTX_set_tlsext_servername_arg()
992 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
994 openssl s_client has a new '-servername ...' option.
996 openssl s_server has new options '-servername_host ...', '-cert2 ...',
997 '-key2 ...', '-servername_fatal' (subject to change). This allows
998 testing the HostName extension for a specific single host name ('-cert'
999 and '-key' remain fallbacks for handshakes without HostName
1000 negotiation). If the unrecogninzed_name alert has to be sent, this by
1001 default is a warning; it becomes fatal with the '-servername_fatal'
1004 [Peter Sylvester, Remy Allais, Christophe Renou]
1006 *) Whirlpool hash implementation is added.
1009 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1010 bn(64,32). Because of instruction set limitations it doesn't have
1011 any negative impact on performance. This was done mostly in order
1012 to make it possible to share assembler modules, such as bn_mul_mont
1013 implementations, between 32- and 64-bit builds without hassle.
1016 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1017 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1021 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1022 dedicated Montgomery multiplication procedure, is introduced.
1023 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1024 "64-bit" performance on certain 32-bit targets.
1027 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1028 in SSL structures. New SSL ctrl to set maximum send fragment size.
1029 Save memory by seeting the I/O buffer sizes dynamically instead of
1030 using the maximum available value.
1033 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1034 in addition to the text details.
1037 *) Very, very preliminary EXPERIMENTAL support for printing of general
1038 ASN1 structures. This currently produces rather ugly output and doesn't
1039 handle several customised structures at all.
1042 *) Integrated support for PVK file format and some related formats such
1043 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1044 these in the 'rsa' and 'dsa' utilities.
1047 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1050 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1051 place for the (very old) "NETSCAPE" format certificates which are now
1052 handled using new ASN1 code equivalents.
1055 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1056 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1057 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1060 *) Modify CRL distribution points extension code to print out previously
1061 unsupported fields. Enhance extension setting code to allow setting of
1065 *) Add print and set support for Issuing Distribution Point CRL extension.
1068 *) Change 'Configure' script to enable Camellia by default.
1071 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
1073 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1075 This addresses the flaw in CBC record processing discovered by
1076 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1077 at: http://www.isg.rhul.ac.uk/tls/
1079 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1080 Security Group at Royal Holloway, University of London
1081 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1082 Emilia Käsper for the initial patch.
1084 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1086 *) Return an error when checking OCSP signatures when key is NULL.
1087 This fixes a DoS attack. (CVE-2013-0166)
1090 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1091 the right response is stapled. Also change SSL_get_certificate()
1092 so it returns the certificate actually sent.
1093 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1094 (This is a backport)
1095 [Rob Stradling <rob.stradling@comodo.com>]
1097 *) Fix possible deadlock when decoding public keys.
1100 Changes between 0.9.8w and 0.9.8x [10 May 2012]
1102 *) Sanity check record length before skipping explicit IV in DTLS
1105 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1106 fuzzing as a service testing platform.
1110 *) Initialise tkeylen properly when encrypting CMS messages.
1111 Thanks to Solar Designer of Openwall for reporting this issue.
1114 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
1116 *) The fix for CVE-2012-2110 did not take into account that the
1117 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
1118 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
1119 rejecting negative len parameter. (CVE-2012-2131)
1120 [Tomas Hoger <thoger@redhat.com>]
1122 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
1124 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1125 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1126 in CRYPTO_realloc_clean.
1128 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1129 issue and to Adam Langley <agl@chromium.org> for fixing it.
1131 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1133 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
1135 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1136 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1137 content decryption and always return the same error. Note: this attack
1138 needs on average 2^20 messages so it only affects automated senders. The
1139 old behaviour can be reenabled in the CMS code by setting the
1140 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1141 an MMA defence is not necessary.
1142 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1143 this issue. (CVE-2012-0884)
1146 *) Fix CVE-2011-4619: make sure we really are receiving a
1147 client hello before rejecting multiple SGC restarts. Thanks to
1148 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1151 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1153 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1154 Thanks to Antonio Martin, Enterprise Secure Access Research and
1155 Development, Cisco Systems, Inc. for discovering this bug and
1156 preparing a fix. (CVE-2012-0050)
1159 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1161 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1162 of the Vaudenay padding oracle attack on CBC mode encryption
1163 which enables an efficient plaintext recovery attack against
1164 the OpenSSL implementation of DTLS. Their attack exploits timing
1165 differences arising during decryption processing. A research
1166 paper describing this attack can be found at:
1167 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1168 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1169 Security Group at Royal Holloway, University of London
1170 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1171 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1172 for preparing the fix. (CVE-2011-4108)
1173 [Robin Seggelmann, Michael Tuexen]
1175 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1176 [Ben Laurie, Kasper <ekasper@google.com>]
1178 *) Clear bytes used for block padding of SSL 3.0 records.
1180 [Adam Langley (Google)]
1182 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1183 Kadianakis <desnacked@gmail.com> for discovering this issue and
1184 Adam Langley for preparing the fix. (CVE-2011-4619)
1185 [Adam Langley (Google)]
1187 *) Prevent malformed RFC3779 data triggering an assertion failure.
1188 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1189 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1190 [Rob Austein <sra@hactrn.net>]
1192 *) Fix ssl_ciph.c set-up race.
1193 [Adam Langley (Google)]
1195 *) Fix spurious failures in ecdsatest.c.
1196 [Emilia Käsper (Google)]
1198 *) Fix the BIO_f_buffer() implementation (which was mixing different
1199 interpretations of the '..._len' fields).
1200 [Adam Langley (Google)]
1202 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1203 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1204 threads won't reuse the same blinding coefficients.
1206 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1207 lock to call BN_BLINDING_invert_ex, and avoids one use of
1208 BN_BLINDING_update for each BN_BLINDING structure (previously,
1209 the last update always remained unused).
1210 [Emilia Käsper (Google)]
1212 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1213 for multi-threaded use of ECDH.
1214 [Adam Langley (Google)]
1216 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1219 *) Add protection against ECDSA timing attacks as mentioned in the paper
1220 by Billy Bob Brumley and Nicola Tuveri, see:
1222 http://eprint.iacr.org/2011/232.pdf
1224 [Billy Bob Brumley and Nicola Tuveri]
1226 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1228 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1229 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1231 *) Fix bug in string printing code: if *any* escaping is enabled we must
1232 escape the escape character (backslash) or the resulting string is
1236 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1238 *) Disable code workaround for ancient and obsolete Netscape browsers
1239 and servers: an attacker can use it in a ciphersuite downgrade attack.
1240 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1243 *) Fixed J-PAKE implementation error, originally discovered by
1244 Sebastien Martini, further info and confirmation from Stefan
1245 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1248 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1250 *) Fix extension code to avoid race conditions which can result in a buffer
1251 overrun vulnerability: resumed sessions must not be modified as they can
1252 be shared by multiple threads. CVE-2010-3864
1255 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1258 *) Don't reencode certificate when calculating signature: cache and use
1259 the original encoding instead. This makes signature verification of
1260 some broken encodings work correctly.
1263 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1264 is also one of the inputs.
1265 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1267 *) Don't repeatedly append PBE algorithms to table if they already exist.
1268 Sort table on each new add. This effectively makes the table read only
1269 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1273 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1275 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1278 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1279 access or freeing data twice (CVE-2010-0742)
1280 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1282 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1283 common in certificates and some applications which only call
1284 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1288 Reduce copying into .apps and .test in makevms.com
1289 Don't try to use blank CA certificate in CA.com
1290 Allow use of C files from original directories in maketests.com
1291 [Steven M. Schweda" <sms@antinode.info>]
1293 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1295 *) When rejecting SSL/TLS records due to an incorrect version number, never
1296 update s->server with a new major version number. As of
1297 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1298 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1299 the previous behavior could result in a read attempt at NULL when
1300 receiving specific incorrect SSL/TLS records once record payload
1301 protection is active. (CVE-2010-0740)
1302 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1304 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1305 could be crashed if the relevant tables were not present (e.g. chrooted).
1306 [Tomas Hoger <thoger@redhat.com>]
1308 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1310 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1311 [Martin Olsson, Neel Mehta]
1313 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1314 accommodate for stack sorting, always a write lock!).
1317 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1318 excessive delays in the RAND_poll(): over a minute. As a workaround
1319 include a time check in the inner Heap32Next loop too.
1322 *) The code that handled flushing of data in SSL/TLS originally used the
1323 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1324 the problem outlined in PR#1949. The fix suggested there however can
1325 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1326 of Apache). So instead simplify the code to flush unconditionally.
1327 This should be fine since flushing with no data to flush is a no op.
1330 *) Handle TLS versions 2.0 and later properly and correctly use the
1331 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1332 off ancient servers have a habit of sticking around for a while...
1335 *) Modify compression code so it frees up structures without using the
1336 ex_data callbacks. This works around a problem where some applications
1337 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1338 restarting) then use compression (e.g. SSL with compression) later.
1339 This results in significant per-connection memory leaks and
1340 has caused some security issues including CVE-2008-1678 and
1344 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1345 change when encrypting or decrypting.
1348 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1349 connect and renegotiate with servers which do not support RI.
1350 Until RI is more widely deployed this option is enabled by default.
1353 *) Add "missing" ssl ctrls to clear options and mode.
1356 *) If client attempts to renegotiate and doesn't support RI respond with
1357 a no_renegotiation alert as required by RFC5746. Some renegotiating
1358 TLS clients will continue a connection gracefully when they receive
1359 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1360 waiting for a server hello which it will never receive. Now we treat a
1361 received no_renegotiation alert as a fatal error. This is because
1362 applications requesting a renegotiation might well expect it to succeed
1363 and would have no code in place to handle the server denying it so the
1364 only safe thing to do is to terminate the connection.
1367 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1368 peer supports secure renegotiation and 0 otherwise. Print out peer
1369 renegotiation support in s_client/s_server.
1372 *) Replace the highly broken and deprecated SPKAC certification method with
1373 the updated NID creation version. This should correctly handle UTF8.
1376 *) Implement RFC5746. Re-enable renegotiation but require the extension
1377 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1378 turns out to be a bad idea. It has been replaced by
1379 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1380 SSL_CTX_set_options(). This is really not recommended unless you
1381 know what you are doing.
1382 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1384 *) Fixes to stateless session resumption handling. Use initial_ctx when
1385 issuing and attempting to decrypt tickets in case it has changed during
1386 servername handling. Use a non-zero length session ID when attempting
1387 stateless session resumption: this makes it possible to determine if
1388 a resumption has occurred immediately after receiving server hello
1389 (several places in OpenSSL subtly assume this) instead of later in
1393 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1394 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1395 fixes for a few places where the return code is not checked
1397 [Julia Lawall <julia@diku.dk>]
1399 *) Add --strict-warnings option to Configure script to include devteam
1400 warnings in other configurations.
1403 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1404 makes it possible to install openssl libraries in locations which
1405 have names other than "lib", for example "/usr/lib64" which some
1407 [Steve Henson, based on patch from Jeremy Utley]
1409 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1410 X690 8.9.12 and can produce some misleading textual output of OIDs.
1411 [Steve Henson, reported by Dan Kaminsky]
1413 *) Delete MD2 from algorithm tables. This follows the recommendation in
1414 several standards that it is not used in new applications due to
1415 several cryptographic weaknesses. For binary compatibility reasons
1416 the MD2 API is still compiled in by default.
1419 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1423 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1424 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1426 [Guenter <lists@gknw.net>]
1428 *) Fix the server certificate chain building code to use X509_verify_cert(),
1429 it used to have an ad-hoc builder which was unable to cope with anything
1430 other than a simple chain.
1431 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1433 *) Don't check self signed certificate signatures in X509_verify_cert()
1434 by default (a flag can override this): it just wastes time without
1435 adding any security. As a useful side effect self signed root CAs
1436 with non-FIPS digests are now usable in FIPS mode.
1439 *) In dtls1_process_out_of_seq_message() the check if the current message
1440 is already buffered was missing. For every new message was memory
1441 allocated, allowing an attacker to perform an denial of service attack
1442 with sending out of seq handshake messages until there is no memory
1443 left. Additionally every future messege was buffered, even if the
1444 sequence number made no sense and would be part of another handshake.
1445 So only messages with sequence numbers less than 10 in advance will be
1446 buffered. (CVE-2009-1378)
1447 [Robin Seggelmann, discovered by Daniel Mentz]
1449 *) Records are buffered if they arrive with a future epoch to be
1450 processed after finishing the corresponding handshake. There is
1451 currently no limitation to this buffer allowing an attacker to perform
1452 a DOS attack with sending records with future epochs until there is no
1453 memory left. This patch adds the pqueue_size() function to detemine
1454 the size of a buffer and limits the record buffer to 100 entries.
1456 [Robin Seggelmann, discovered by Daniel Mentz]
1458 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1459 parent structure is freed. (CVE-2009-1379)
1462 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1463 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1466 [Ilya O. <vrghost@gmail.com>]
1468 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1470 *) Disable renegotiation completely - this fixes a severe security
1471 problem (CVE-2009-3555) at the cost of breaking all
1472 renegotiation. Renegotiation can be re-enabled by setting
1473 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1474 run-time. This is really not recommended unless you know what
1478 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1480 *) Don't set val to NULL when freeing up structures, it is freed up by
1481 underlying code. If sizeof(void *) > sizeof(long) this can result in
1482 zeroing past the valid field. (CVE-2009-0789)
1483 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1485 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1486 checked correctly. This would allow some invalid signed attributes to
1487 appear to verify correctly. (CVE-2009-0591)
1488 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1490 *) Reject UniversalString and BMPString types with invalid lengths. This
1491 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1492 a legal length. (CVE-2009-0590)
1495 *) Set S/MIME signing as the default purpose rather than setting it
1496 unconditionally. This allows applications to override it at the store
1500 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1501 to handle some structures.
1504 *) Improve efficiency of mem_gets: don't search whole buffer each time
1506 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1508 *) New -hex option for openssl rand.
1511 *) Print out UTF8String and NumericString when parsing ASN1.
1514 *) Support NumericString type for name components.
1517 *) Allow CC in the environment to override the automatically chosen
1518 compiler. Note that nothing is done to ensure flags work with the
1522 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1524 *) Properly check EVP_VerifyFinal() and similar return values
1526 [Ben Laurie, Bodo Moeller, Google Security Team]
1528 *) Enable TLS extensions by default.
1531 *) Allow the CHIL engine to be loaded, whether the application is
1532 multithreaded or not. (This does not release the developer from the
1533 obligation to set up the dynamic locking callbacks.)
1534 [Sander Temme <sander@temme.net>]
1536 *) Use correct exit code if there is an error in dgst command.
1537 [Steve Henson; problem pointed out by Roland Dirlewanger]
1539 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1540 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1543 *) Add experimental JPAKE support, including demo authentication in
1544 s_client and s_server.
1547 *) Set the comparison function in v3_addr_canonize().
1548 [Rob Austein <sra@hactrn.net>]
1550 *) Add support for XMPP STARTTLS in s_client.
1551 [Philip Paeps <philip@freebsd.org>]
1553 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1554 to ensure that even with this option, only ciphersuites in the
1555 server's preference list will be accepted. (Note that the option
1556 applies only when resuming a session, so the earlier behavior was
1557 just about the algorithm choice for symmetric cryptography.)
1560 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1562 *) Fix NULL pointer dereference if a DTLS server received
1563 ChangeCipherSpec as first record (CVE-2009-1386).
1566 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1567 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1570 *) The fix in 0.9.8c that supposedly got rid of unsafe
1571 double-checked locking was incomplete for RSA blinding,
1572 addressing just one layer of what turns out to have been
1573 doubly unsafe triple-checked locking.
1575 So now fix this for real by retiring the MONT_HELPER macro
1576 in crypto/rsa/rsa_eay.c.
1578 [Bodo Moeller; problem pointed out by Marius Schilder]
1580 *) Various precautionary measures:
1582 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1584 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1585 (NB: This would require knowledge of the secret session ticket key
1586 to exploit, in which case you'd be SOL either way.)
1588 - Change bn_nist.c so that it will properly handle input BIGNUMs
1589 outside the expected range.
1591 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1594 [Neel Mehta, Bodo Moeller]
1596 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1597 the load fails. Useful for distros.
1598 [Ben Laurie and the FreeBSD team]
1600 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1603 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1606 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1608 This work was sponsored by Logica.
1611 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1612 keystores. Support for SSL/TLS client authentication too.
1613 Not compiled unless enable-capieng specified to Configure.
1615 This work was sponsored by Logica.
1618 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1619 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1620 attribute creation routines such as certifcate requests and PKCS#12
1624 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1626 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1627 handshake which could lead to a cilent crash as found using the
1628 Codenomicon TLS test suite (CVE-2008-1672)
1629 [Steve Henson, Mark Cox]
1631 *) Fix double free in TLS server name extensions which could lead to
1632 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1635 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1637 Clear the error queue to ensure that error entries left from
1638 older function calls do not interfere with the correct operation.
1639 [Lutz Jaenicke, Erik de Castro Lopo]
1641 *) Remove root CA certificates of commercial CAs:
1643 The OpenSSL project does not recommend any specific CA and does not
1644 have any policy with respect to including or excluding any CA.
1645 Therefore it does not make any sense to ship an arbitrary selection
1646 of root CA certificates with the OpenSSL software.
1649 *) RSA OAEP patches to fix two separate invalid memory reads.
1650 The first one involves inputs when 'lzero' is greater than
1651 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1652 before the beginning of from). The second one involves inputs where
1653 the 'db' section contains nothing but zeroes (there is a one-byte
1654 invalid read after the end of 'db').
1655 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1657 *) Partial backport from 0.9.9-dev:
1659 Introduce bn_mul_mont (dedicated Montgomery multiplication
1660 procedure) as a candidate for BIGNUM assembler implementation.
1661 While 0.9.9-dev uses assembler for various architectures, only
1662 x86_64 is available by default here in the 0.9.8 branch, and
1663 32-bit x86 is available through a compile-time setting.
1665 To try the 32-bit x86 assembler implementation, use Configure
1666 option "enable-montasm" (which exists only for this backport).
1668 As "enable-montasm" for 32-bit x86 disclaims code stability
1669 anyway, in this constellation we activate additional code
1670 backported from 0.9.9-dev for further performance improvements,
1671 namely BN_from_montgomery_word. (To enable this otherwise,
1672 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1674 [Andy Polyakov (backport partially by Bodo Moeller)]
1676 *) Add TLS session ticket callback. This allows an application to set
1677 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1678 values. This is useful for key rollover for example where several key
1679 sets may exist with different names.
1682 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1683 This was broken until now in 0.9.8 releases, such that the only way
1684 a registered ENGINE could be used (assuming it initialises
1685 successfully on the host) was to explicitly set it as the default
1686 for the relevant algorithms. This is in contradiction with 0.9.7
1687 behaviour and the documentation. With this fix, when an ENGINE is
1688 registered into a given algorithm's table of implementations, the
1689 'uptodate' flag is reset so that auto-discovery will be used next
1690 time a new context for that algorithm attempts to select an
1692 [Ian Lister (tweaked by Geoff Thorpe)]
1694 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1695 implemention in the following ways:
1697 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1700 Lack of BER streaming support means one pass streaming processing is
1701 only supported if data is detached: setting the streaming flag is
1702 ignored for embedded content.
1704 CMS support is disabled by default and must be explicitly enabled
1705 with the enable-cms configuration option.
1708 *) Update the GMP engine glue to do direct copies between BIGNUM and
1709 mpz_t when openssl and GMP use the same limb size. Otherwise the
1710 existing "conversion via a text string export" trick is still used.
1711 [Paul Sheer <paulsheer@gmail.com>]
1713 *) Zlib compression BIO. This is a filter BIO which compressed and
1714 uncompresses any data passed through it.
1717 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1718 RFC3394 compatible AES key wrapping.
1721 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1722 sets string data without copying. X509_ALGOR_set0() and
1723 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1724 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1725 from an X509_ATTRIBUTE structure optionally checking it occurs only
1726 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1730 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1731 to get the expected BN_FLG_CONSTTIME behavior.
1732 [Bodo Moeller (Google)]
1736 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1737 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1738 - added some more tests to do_tests.pl
1739 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1740 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1741 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1742 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1743 - various changes to netware.pl to enable gcc-cross builds on Win32
1745 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1746 - various changes to fix missing prototype warnings
1747 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1748 - added AES, WHIRLPOOL and CPUID assembler code to build files
1749 - added missing AES assembler make rules to mk1mf.pl
1750 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1751 [Guenter Knauf <eflash@gmx.net>]
1753 *) Implement certificate status request TLS extension defined in RFC3546.
1754 A client can set the appropriate parameters and receive the encoded
1755 OCSP response via a callback. A server can query the supplied parameters
1756 and set the encoded OCSP response in the callback. Add simplified examples
1757 to s_client and s_server.
1760 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1762 *) Fix various bugs:
1763 + Binary incompatibility of ssl_ctx_st structure
1764 + DTLS interoperation with non-compliant servers
1765 + Don't call get_session_cb() without proposed session
1766 + Fix ia64 assembler code
1767 [Andy Polyakov, Steve Henson]
1769 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1771 *) DTLS Handshake overhaul. There were longstanding issues with
1772 OpenSSL DTLS implementation, which were making it impossible for
1773 RFC 4347 compliant client to communicate with OpenSSL server.
1774 Unfortunately just fixing these incompatibilities would "cut off"
1775 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1776 server keeps tolerating non RFC compliant syntax. The opposite is
1777 not true, 0.9.8f client can not communicate with earlier server.
1778 This update even addresses CVE-2007-4995.
1781 *) Changes to avoid need for function casts in OpenSSL: some compilers
1782 (gcc 4.2 and later) reject their use.
1783 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1786 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1787 RFC4507bis. The encrypted ticket format is an encrypted encoded
1788 SSL_SESSION structure, that way new session features are automatically
1791 If a client application caches session in an SSL_SESSION structure
1792 support is transparent because tickets are now stored in the encoded
1795 The SSL_CTX structure automatically generates keys for ticket
1796 protection in servers so again support should be possible
1797 with no application modification.
1799 If a client or server wishes to disable RFC4507 support then the option
1800 SSL_OP_NO_TICKET can be set.
1802 Add a TLS extension debugging callback to allow the contents of any client
1803 or server extensions to be examined.
1805 This work was sponsored by Google.
1808 *) Add initial support for TLS extensions, specifically for the server_name
1809 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1810 have new members for a host name. The SSL data structure has an
1811 additional member SSL_CTX *initial_ctx so that new sessions can be
1812 stored in that context to allow for session resumption, even after the
1813 SSL has been switched to a new SSL_CTX in reaction to a client's
1814 server_name extension.
1816 New functions (subject to change):
1818 SSL_get_servername()
1819 SSL_get_servername_type()
1822 New CTRL codes and macros (subject to change):
1824 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1825 - SSL_CTX_set_tlsext_servername_callback()
1826 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1827 - SSL_CTX_set_tlsext_servername_arg()
1828 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1830 openssl s_client has a new '-servername ...' option.
1832 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1833 '-key2 ...', '-servername_fatal' (subject to change). This allows
1834 testing the HostName extension for a specific single host name ('-cert'
1835 and '-key' remain fallbacks for handshakes without HostName
1836 negotiation). If the unrecogninzed_name alert has to be sent, this by
1837 default is a warning; it becomes fatal with the '-servername_fatal'
1840 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1842 *) Add AES and SSE2 assembly language support to VC++ build.
1845 *) Mitigate attack on final subtraction in Montgomery reduction.
1848 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1849 (which previously caused an internal error).
1852 *) Squeeze another 10% out of IGE mode when in != out.
1855 *) AES IGE mode speedup.
1856 [Dean Gaudet (Google)]
1858 *) Add the Korean symmetric 128-bit cipher SEED (see
1859 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1860 add SEED ciphersuites from RFC 4162:
1862 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1863 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1864 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1865 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1867 To minimize changes between patchlevels in the OpenSSL 0.9.8
1868 series, SEED remains excluded from compilation unless OpenSSL
1869 is configured with 'enable-seed'.
1870 [KISA, Bodo Moeller]
1872 *) Mitigate branch prediction attacks, which can be practical if a
1873 single processor is shared, allowing a spy process to extract
1874 information. For detailed background information, see
1875 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1876 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
1877 and Necessary Software Countermeasures"). The core of the change
1878 are new versions BN_div_no_branch() and
1879 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
1880 respectively, which are slower, but avoid the security-relevant
1881 conditional branches. These are automatically called by BN_div()
1882 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
1883 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
1884 remove a conditional branch.
1886 BN_FLG_CONSTTIME is the new name for the previous
1887 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
1888 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
1889 in the exponent causes BN_mod_exp_mont() to use the alternative
1890 implementation in BN_mod_exp_mont_consttime().) The old name
1891 remains as a deprecated alias.
1893 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
1894 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
1895 constant-time implementations for more than just exponentiation.
1896 Here too the old name is kept as a deprecated alias.
1898 BN_BLINDING_new() will now use BN_dup() for the modulus so that
1899 the BN_BLINDING structure gets an independent copy of the
1900 modulus. This means that the previous "BIGNUM *m" argument to
1901 BN_BLINDING_new() and to BN_BLINDING_create_param() now
1902 essentially becomes "const BIGNUM *m", although we can't actually
1903 change this in the header file before 0.9.9. It allows
1904 RSA_setup_blinding() to use BN_with_flags() on the modulus to
1905 enable BN_FLG_CONSTTIME.
1907 [Matthew D Wood (Intel Corp)]
1909 *) In the SSL/TLS server implementation, be strict about session ID
1910 context matching (which matters if an application uses a single
1911 external cache for different purposes). Previously,
1912 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1913 set. This did ensure strict client verification, but meant that,
1914 with applications using a single external cache for quite
1915 different requirements, clients could circumvent ciphersuite
1916 restrictions for a given session ID context by starting a session
1917 in a different context.
1920 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1921 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1922 authentication-only ciphersuites.
1925 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
1926 not complete and could lead to a possible single byte overflow
1927 (CVE-2007-5135) [Ben Laurie]
1929 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
1931 *) Since AES128 and AES256 (and similarly Camellia128 and
1932 Camellia256) share a single mask bit in the logic of
1933 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1934 kludge to work properly if AES128 is available and AES256 isn't
1935 (or if Camellia128 is available and Camellia256 isn't).
1938 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
1939 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
1940 When a point or a seed is encoded in a BIT STRING, we need to
1941 prevent the removal of trailing zero bits to get the proper DER
1942 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
1943 of a NamedBitList, for which trailing 0 bits need to be removed.)
1946 *) Have SSL/TLS server implementation tolerate "mismatched" record
1947 protocol version while receiving ClientHello even if the
1948 ClientHello is fragmented. (The server can't insist on the
1949 particular protocol version it has chosen before the ServerHello
1950 message has informed the client about his choice.)
1953 *) Add RFC 3779 support.
1954 [Rob Austein for ARIN, Ben Laurie]
1956 *) Load error codes if they are not already present instead of using a
1957 static variable. This allows them to be cleanly unloaded and reloaded.
1958 Improve header file function name parsing.
1961 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
1962 or CAPABILITY handshake as required by RFCs.
1965 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
1967 *) Introduce limits to prevent malicious keys being able to
1968 cause a denial of service. (CVE-2006-2940)
1969 [Steve Henson, Bodo Moeller]
1971 *) Fix ASN.1 parsing of certain invalid structures that can result
1972 in a denial of service. (CVE-2006-2937) [Steve Henson]
1974 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1975 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1977 *) Fix SSL client code which could crash if connecting to a
1978 malicious SSLv2 server. (CVE-2006-4343)
1979 [Tavis Ormandy and Will Drewry, Google Security Team]
1981 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
1982 match only those. Before that, "AES256-SHA" would be interpreted
1983 as a pattern and match "AES128-SHA" too (since AES128-SHA got
1984 the same strength classification in 0.9.7h) as we currently only
1985 have a single AES bit in the ciphersuite description bitmap.
1986 That change, however, also applied to ciphersuite strings such as
1987 "RC4-MD5" that intentionally matched multiple ciphersuites --
1988 namely, SSL 2.0 ciphersuites in addition to the more common ones
1989 from SSL 3.0/TLS 1.0.
1991 So we change the selection algorithm again: Naming an explicit
1992 ciphersuite selects this one ciphersuite, and any other similar
1993 ciphersuite (same bitmap) from *other* protocol versions.
1994 Thus, "RC4-MD5" again will properly select both the SSL 2.0
1995 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
1997 Since SSL 2.0 does not have any ciphersuites for which the
1998 128/256 bit distinction would be relevant, this works for now.
1999 The proper fix will be to use different bits for AES128 and
2000 AES256, which would have avoided the problems from the beginning;
2001 however, bits are scarce, so we can only do this in a new release
2002 (not just a patchlevel) when we can change the SSL_CIPHER
2003 definition to split the single 'unsigned long mask' bitmap into
2004 multiple values to extend the available space.
2008 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2010 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2011 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2013 *) Add AES IGE and biIGE modes.
2016 *) Change the Unix randomness entropy gathering to use poll() when
2017 possible instead of select(), since the latter has some
2018 undesirable limitations.
2019 [Darryl Miles via Richard Levitte and Bodo Moeller]
2021 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2022 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2023 cannot be implicitly activated as part of, e.g., the "AES" alias.
2024 However, please upgrade to OpenSSL 0.9.9[-dev] for
2025 non-experimental use of the ECC ciphersuites to get TLS extension
2026 support, which is required for curve and point format negotiation
2027 to avoid potential handshake problems.
2030 *) Disable rogue ciphersuites:
2032 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2033 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2034 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2036 The latter two were purportedly from
2037 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2040 Also deactivate the remaining ciphersuites from
2041 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2042 unofficial, and the ID has long expired.
2045 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2046 dual-core machines) and other potential thread-safety issues.
2049 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2050 versions), which is now available for royalty-free use
2051 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2052 Also, add Camellia TLS ciphersuites from RFC 4132.
2054 To minimize changes between patchlevels in the OpenSSL 0.9.8
2055 series, Camellia remains excluded from compilation unless OpenSSL
2056 is configured with 'enable-camellia'.
2059 *) Disable the padding bug check when compression is in use. The padding
2060 bug check assumes the first packet is of even length, this is not
2061 necessarily true if compresssion is enabled and can result in false
2062 positives causing handshake failure. The actual bug test is ancient
2063 code so it is hoped that implementations will either have fixed it by
2064 now or any which still have the bug do not support compression.
2067 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2069 *) When applying a cipher rule check to see if string match is an explicit
2070 cipher suite and only match that one cipher suite if it is.
2073 *) Link in manifests for VC++ if needed.
2074 [Austin Ziegler <halostatue@gmail.com>]
2076 *) Update support for ECC-based TLS ciphersuites according to
2077 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2078 TLS extensions, which are supported starting with the 0.9.9
2079 branch, not in the OpenSSL 0.9.8 branch).
2082 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2083 opaque EVP_CIPHER_CTX handling.
2086 *) Fixes and enhancements to zlib compression code. We now only use
2087 "zlib1.dll" and use the default __cdecl calling convention on Win32
2088 to conform with the standards mentioned here:
2089 http://www.zlib.net/DLL_FAQ.txt
2090 Static zlib linking now works on Windows and the new --with-zlib-include
2091 --with-zlib-lib options to Configure can be used to supply the location
2092 of the headers and library. Gracefully handle case where zlib library
2096 *) Several fixes and enhancements to the OID generation code. The old code
2097 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2098 handle numbers larger than ULONG_MAX, truncated printing and had a
2099 non standard OBJ_obj2txt() behaviour.
2102 *) Add support for building of engines under engine/ as shared libraries
2103 under VC++ build system.
2106 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2107 Hopefully, we will not see any false combination of paths any more.
2110 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2112 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2113 (part of SSL_OP_ALL). This option used to disable the
2114 countermeasure against man-in-the-middle protocol-version
2115 rollback in the SSL 2.0 server implementation, which is a bad
2116 idea. (CVE-2005-2969)
2118 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2119 for Information Security, National Institute of Advanced Industrial
2120 Science and Technology [AIST], Japan)]
2122 *) Add two function to clear and return the verify parameter flags.
2125 *) Keep cipherlists sorted in the source instead of sorting them at
2126 runtime, thus removing the need for a lock.
2129 *) Avoid some small subgroup attacks in Diffie-Hellman.
2130 [Nick Mathewson and Ben Laurie]
2132 *) Add functions for well-known primes.
2135 *) Extended Windows CE support.
2136 [Satoshi Nakamura and Andy Polyakov]
2138 *) Initialize SSL_METHOD structures at compile time instead of during
2139 runtime, thus removing the need for a lock.
2142 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2143 attempting to decrypt each encrypted key in turn. Add support to
2147 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2149 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2152 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2155 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2156 key into the same file any more.
2159 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2162 *) Add -utf8 command line and config file option to 'ca'.
2163 [Stefan <stf@udoma.org]
2165 *) Removed the macro des_crypt(), as it seems to conflict with some
2166 libraries. Use DES_crypt().
2169 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2170 involves renaming the source and generated shared-libs for
2171 both. The engines will accept the corrected or legacy ids
2172 ('ncipher' and '4758_cca' respectively) when binding. NB,
2173 this only applies when building 'shared'.
2174 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2176 *) Add attribute functions to EVP_PKEY structure. Modify
2177 PKCS12_create() to recognize a CSP name attribute and
2178 use it. Make -CSP option work again in pkcs12 utility.
2181 *) Add new functionality to the bn blinding code:
2182 - automatic re-creation of the BN_BLINDING parameters after
2183 a fixed number of uses (currently 32)
2184 - add new function for parameter creation
2185 - introduce flags to control the update behaviour of the
2186 BN_BLINDING parameters
2187 - hide BN_BLINDING structure
2188 Add a second BN_BLINDING slot to the RSA structure to improve
2189 performance when a single RSA object is shared among several
2193 *) Add support for DTLS.
2194 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2196 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2197 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2200 *) Remove buggy and incompletet DH cert support from
2201 ssl/ssl_rsa.c and ssl/s3_both.c
2204 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2205 the apps/openssl applications.
2208 *) Compile clean with "-Wall -Wmissing-prototypes
2209 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2210 DEBUG_SAFESTACK must also be set.
2213 *) Change ./Configure so that certain algorithms can be disabled by default.
2214 The new counterpiece to "no-xxx" is "enable-xxx".
2216 The patented RC5 and MDC2 algorithms will now be disabled unless
2217 "enable-rc5" and "enable-mdc2", respectively, are specified.
2219 (IDEA remains enabled despite being patented. This is because IDEA
2220 is frequently required for interoperability, and there is no license
2221 fee for non-commercial use. As before, "no-idea" can be used to
2222 avoid this algorithm.)
2226 *) Add processing of proxy certificates (see RFC 3820). This work was
2227 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2228 EGEE (Enabling Grids for E-science in Europe).
2231 *) RC4 performance overhaul on modern architectures/implementations, such
2232 as Intel P4, IA-64 and AMD64.
2235 *) New utility extract-section.pl. This can be used specify an alternative
2236 section number in a pod file instead of having to treat each file as
2237 a separate case in Makefile. This can be done by adding two lines to the
2240 =for comment openssl_section:XXX
2242 The blank line is mandatory.
2246 *) New arguments -certform, -keyform and -pass for s_client and s_server
2247 to allow alternative format key and certificate files and passphrase
2251 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2252 update associated structures and add various utility functions.
2254 Add new policy related verify parameters, include policy checking in
2255 standard verify code. Enhance 'smime' application with extra parameters
2256 to support policy checking and print out.
2259 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2260 Nehemiah processors. These extensions support AES encryption in hardware
2261 as well as RNG (though RNG support is currently disabled).
2262 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2264 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2267 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2268 [Andy Polyakov and a number of other people]
2270 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2271 implementation contributed by IBM.
2272 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2274 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2275 exponent rather than 'unsigned long'. There is a corresponding change to
2276 the new 'rsa_keygen' element of the RSA_METHOD structure.
2277 [Jelte Jansen, Geoff Thorpe]
2279 *) Functionality for creating the initial serial number file is now
2280 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2282 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2283 number file to 1, which is bound to cause problems. To avoid
2284 the problems while respecting compatibility between different 0.9.7
2285 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2286 CA.pl for serial number initialization. With the new release 0.9.8,
2287 we can fix the problem directly in the 'ca' utility.)
2290 *) Reduced header interdepencies by declaring more opaque objects in
2291 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2292 give fewer recursive includes, which could break lazy source code - so
2293 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2294 developers should define this symbol when building and using openssl to
2295 ensure they track the recommended behaviour, interfaces, [etc], but
2296 backwards-compatible behaviour prevails when this isn't defined.
2299 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2302 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2303 This will generate a random key of the appropriate length based on the
2304 cipher context. The EVP_CIPHER can provide its own random key generation
2305 routine to support keys of a specific form. This is used in the des and
2306 3des routines to generate a key of the correct parity. Update S/MIME
2307 code to use new functions and hence generate correct parity DES keys.
2308 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2309 valid (weak or incorrect parity).
2312 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2313 as looking them up. This is useful when the verified structure may contain
2314 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2315 present unless the new PKCS7_NO_CRL flag is asserted.
2318 *) Extend ASN1 oid configuration module. It now additionally accepts the
2321 shortName = some long name, 1.2.3.4
2324 *) Reimplemented the BN_CTX implementation. There is now no more static
2325 limitation on the number of variables it can handle nor the depth of the
2326 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2327 information can now expand as required, and rather than having a single
2328 static array of bignums, BN_CTX now uses a linked-list of such arrays
2329 allowing it to expand on demand whilst maintaining the usefulness of
2330 BN_CTX's "bundling".
2333 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2334 to allow all RSA operations to function using a single BN_CTX.
2337 *) Preliminary support for certificate policy evaluation and checking. This
2338 is initially intended to pass the tests outlined in "Conformance Testing
2339 of Relying Party Client Certificate Path Processing Logic" v1.07.
2342 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2343 remained unused and not that useful. A variety of other little bignum
2344 tweaks and fixes have also been made continuing on from the audit (see
2348 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2349 associated ASN1, EVP and SSL functions and old ASN1 macros.
2352 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2353 and this should never fail. So the return value from the use of
2354 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2355 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2358 *) BN_CTX_get() should return zero-valued bignums, providing the same
2359 initialised value as BN_new().
2360 [Geoff Thorpe, suggested by Ulf Möller]
2362 *) Support for inhibitAnyPolicy certificate extension.
2365 *) An audit of the BIGNUM code is underway, for which debugging code is
2366 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2367 is considered valid when processing BIGNUMs, and causes execution to
2368 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2369 further steps are taken to deliberately pollute unused data in BIGNUM
2370 structures to try and expose faulty code further on. For now, openssl will
2371 (in its default mode of operation) continue to tolerate the inconsistent
2372 forms that it has tolerated in the past, but authors and packagers should
2373 consider trying openssl and their own applications when compiled with
2374 these debugging symbols defined. It will help highlight potential bugs in
2375 their own code, and will improve the test coverage for OpenSSL itself. At
2376 some point, these tighter rules will become openssl's default to improve
2377 maintainability, though the assert()s and other overheads will remain only
2378 in debugging configurations. See bn.h for more details.
2379 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2381 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2382 that can only be obtained through BN_CTX_new() (which implicitly
2383 initialises it). The presence of this function only made it possible
2384 to overwrite an existing structure (and cause memory leaks).
2387 *) Because of the callback-based approach for implementing LHASH as a
2388 template type, lh_insert() adds opaque objects to hash-tables and
2389 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2390 to clean up those corresponding objects before destroying the hash table
2391 (and losing the object pointers). So some over-zealous constifications in
2392 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2393 objects as "const" and the lh_doall[_arg] callback wrappers are not
2394 prototyped to have "const" restrictions on the object pointers they are
2395 given (and so aren't required to cast them away any more).
2398 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2399 (speed) prefers to use its own implementation. The two implementations
2400 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2401 its object type properly exposed (MS_TM) instead of casting to/from "char
2402 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2403 aren't necessarily the greatest nomenclatures - but this is what was used
2404 internally to the implementation so I've used that for now.
2407 *) Ensure that deprecated functions do not get compiled when
2408 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2409 the self-tests were still using deprecated key-generation functions so
2410 these have been updated also.
2413 *) Reorganise PKCS#7 code to separate the digest location functionality
2414 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2415 New function PKCS7_set_digest() to set the digest type for PKCS#7
2416 digestedData type. Add additional code to correctly generate the
2417 digestedData type and add support for this type in PKCS7 initialization
2421 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2422 structure of type "other".
2425 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2426 sure the loop does correctly stop and breaking ("division by zero")
2427 modulus operations are not performed. The (pre-generated) prime
2428 table crypto/bn/bn_prime.h was already correct, but it could not be
2429 re-generated on some platforms because of the "division by zero"
2430 situation in the script.
2431 [Ralf S. Engelschall]
2433 *) Update support for ECC-based TLS ciphersuites according to
2434 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2435 SHA-1 now is only used for "small" curves (where the
2436 representation of a field element takes up to 24 bytes); for
2437 larger curves, the field element resulting from ECDH is directly
2438 used as premaster secret.
2439 [Douglas Stebila (Sun Microsystems Laboratories)]
2441 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2442 curve secp160r1 to the tests.
2443 [Douglas Stebila (Sun Microsystems Laboratories)]
2445 *) Add the possibility to load symbols globally with DSO.
2446 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2448 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2449 control of the error stack.
2452 *) Add support for STORE in ENGINE.
2455 *) Add the STORE type. The intention is to provide a common interface
2456 to certificate and key stores, be they simple file-based stores, or
2457 HSM-type store, or LDAP stores, or...
2458 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2461 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2462 pass a list of arguments to any function as well as provide a way
2463 for a function to pass data back to the caller.
2466 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2467 works like BUF_strdup() but can be used to duplicate a portion of
2468 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2472 *) Add the function sk_find_ex() which works like sk_find(), but will
2473 return an index to an element even if an exact match couldn't be
2474 found. The index is guaranteed to point at the element where the
2475 searched-for key would be inserted to preserve sorting order.
2478 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2479 takes an extra flags argument for optional functionality. Currently,
2480 the following flags are defined:
2482 OBJ_BSEARCH_VALUE_ON_NOMATCH
2483 This one gets OBJ_bsearch_ex() to return a pointer to the first
2484 element where the comparing function returns a negative or zero
2487 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2488 This one gets OBJ_bsearch_ex() to return a pointer to the first
2489 element where the comparing function returns zero. This is useful
2490 if there are more than one element where the comparing function
2494 *) Make it possible to create self-signed certificates with 'openssl ca'
2495 in such a way that the self-signed certificate becomes part of the
2496 CA database and uses the same mechanisms for serial number generation
2497 as all other certificate signing. The new flag '-selfsign' enables
2498 this functionality. Adapt CA.sh and CA.pl.in.
2501 *) Add functionality to check the public key of a certificate request
2502 against a given private. This is useful to check that a certificate
2503 request can be signed by that key (self-signing).
2506 *) Make it possible to have multiple active certificates with the same
2507 subject in the CA index file. This is done only if the keyword
2508 'unique_subject' is set to 'no' in the main CA section (default
2509 if 'CA_default') of the configuration file. The value is saved
2510 with the database itself in a separate index attribute file,
2511 named like the index file with '.attr' appended to the name.
2514 *) Generate muti valued AVAs using '+' notation in config files for
2518 *) Support for nameConstraints certificate extension.
2521 *) Support for policyConstraints certificate extension.
2524 *) Support for policyMappings certificate extension.
2527 *) Make sure the default DSA_METHOD implementation only uses its
2528 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2529 and change its own handlers to be NULL so as to remove unnecessary
2530 indirection. This lets alternative implementations fallback to the
2531 default implementation more easily.
2534 *) Support for directoryName in GeneralName related extensions
2538 *) Make it possible to link applications using Makefile.shared.
2539 Make that possible even when linking against static libraries!
2542 *) Support for single pass processing for S/MIME signing. This now
2543 means that S/MIME signing can be done from a pipe, in addition
2544 cleartext signing (multipart/signed type) is effectively streaming
2545 and the signed data does not need to be all held in memory.
2547 This is done with a new flag PKCS7_STREAM. When this flag is set
2548 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2549 is done after the data is output (and digests calculated) in
2550 SMIME_write_PKCS7().
2553 *) Add full support for -rpath/-R, both in shared libraries and
2554 applications, at least on the platforms where it's known how
2558 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2559 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2560 will now compute a table of multiples of the generator that
2561 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2562 faster (notably in the case of a single point multiplication,
2563 scalar * generator).
2564 [Nils Larsch, Bodo Moeller]
2566 *) IPv6 support for certificate extensions. The various extensions
2567 which use the IP:a.b.c.d can now take IPv6 addresses using the
2568 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2572 *) Added an ENGINE that implements RSA by performing private key
2573 exponentiations with the GMP library. The conversions to and from
2574 GMP's mpz_t format aren't optimised nor are any montgomery forms
2575 cached, and on x86 it appears OpenSSL's own performance has caught up.
2576 However there are likely to be other architectures where GMP could
2577 provide a boost. This ENGINE is not built in by default, but it can be
2578 specified at Configure time and should be accompanied by the necessary
2579 linker additions, eg;
2580 ./config -DOPENSSL_USE_GMP -lgmp
2583 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2584 testing availability of engines with "-t" - the old behaviour is
2585 produced by increasing the feature's verbosity with "-tt".
2588 *) ECDSA routines: under certain error conditions uninitialized BN objects
2589 could be freed. Solution: make sure initialization is performed early
2590 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2594 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2595 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2596 software implementations. For DSA and DH, parameter generation can
2597 also be overriden by providing the appropriate method callbacks.
2600 *) Change the "progress" mechanism used in key-generation and
2601 primality testing to functions that take a new BN_GENCB pointer in
2602 place of callback/argument pairs. The new API functions have "_ex"
2603 postfixes and the older functions are reimplemented as wrappers for
2604 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2605 declarations of the old functions to help (graceful) attempts to
2606 migrate to the new functions. Also, the new key-generation API
2607 functions operate on a caller-supplied key-structure and return
2608 success/failure rather than returning a key or NULL - this is to
2609 help make "keygen" another member function of RSA_METHOD etc.
2611 Example for using the new callback interface:
2613 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2617 BN_GENCB_set(&my_cb, my_callback, my_arg);
2619 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2620 /* For the meaning of a, b in calls to my_callback(), see the
2621 * documentation of the function that calls the callback.
2622 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2623 * my_callback should return 1 if it wants BN_is_prime_ex()
2624 * to continue, or 0 to stop.
2629 *) Change the ZLIB compression method to be stateful, and make it
2630 available to TLS with the number defined in
2631 draft-ietf-tls-compression-04.txt.
2634 *) Add the ASN.1 structures and functions for CertificatePair, which
2635 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2637 CertificatePair ::= SEQUENCE {
2638 forward [0] Certificate OPTIONAL,
2639 reverse [1] Certificate OPTIONAL,
2640 -- at least one of the pair shall be present -- }
2642 Also implement the PEM functions to read and write certificate
2643 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2645 This needed to be defined, mostly for the sake of the LDAP
2646 attribute crossCertificatePair, but may prove useful elsewhere as
2650 *) Make it possible to inhibit symlinking of shared libraries in
2651 Makefile.shared, for Cygwin's sake.
2654 *) Extend the BIGNUM API by creating a function
2655 void BN_set_negative(BIGNUM *a, int neg);
2656 and a macro that behave like
2657 int BN_is_negative(const BIGNUM *a);
2659 to avoid the need to access 'a->neg' directly in applications.
2662 *) Implement fast modular reduction for pseudo-Mersenne primes
2663 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2664 EC_GROUP_new_curve_GFp() will now automatically use this
2666 [Nils Larsch <nla@trustcenter.de>]
2668 *) Add new lock type (CRYPTO_LOCK_BN).
2671 *) Change the ENGINE framework to automatically load engines
2672 dynamically from specific directories unless they could be
2673 found to already be built in or loaded. Move all the
2674 current engines except for the cryptodev one to a new
2676 The engines in engines/ are built as shared libraries if
2677 the "shared" options was given to ./Configure or ./config.
2678 Otherwise, they are inserted in libcrypto.a.
2679 /usr/local/ssl/engines is the default directory for dynamic
2680 engines, but that can be overriden at configure time through
2681 the usual use of --prefix and/or --openssldir, and at run
2682 time with the environment variable OPENSSL_ENGINES.
2683 [Geoff Thorpe and Richard Levitte]
2685 *) Add Makefile.shared, a helper makefile to build shared
2686 libraries. Addapt Makefile.org.
2689 *) Add version info to Win32 DLLs.
2690 [Peter 'Luna' Runestig" <peter@runestig.com>]
2692 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2693 can be added using this API to created arbitrary PKCS#12
2694 files while avoiding the low level API.
2696 New options to PKCS12_create(), key or cert can be NULL and
2697 will then be omitted from the output file. The encryption
2698 algorithm NIDs can be set to -1 for no encryption, the mac
2699 iteration count can be set to 0 to omit the mac.
2701 Enhance pkcs12 utility by making the -nokeys and -nocerts
2702 options work when creating a PKCS#12 file. New option -nomac
2703 to omit the mac, NONE can be set for an encryption algorithm.
2704 New code is modified to use the enhanced PKCS12_create()
2705 instead of the low level API.
2708 *) Extend ASN1 encoder to support indefinite length constructed
2709 encoding. This can output sequences tags and octet strings in
2710 this form. Modify pk7_asn1.c to support indefinite length
2711 encoding. This is experimental and needs additional code to
2712 be useful, such as an ASN1 bio and some enhanced streaming
2715 Extend template encode functionality so that tagging is passed
2716 down to the template encoder.
2719 *) Let 'openssl req' fail if an argument to '-newkey' is not
2720 recognized instead of using RSA as a default.
2723 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2724 As these are not official, they are not included in "ALL";
2725 the "ECCdraft" ciphersuite group alias can be used to select them.
2726 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2728 *) Add ECDH engine support.
2729 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2731 *) Add ECDH in new directory crypto/ecdh/.
2732 [Douglas Stebila (Sun Microsystems Laboratories)]
2734 *) Let BN_rand_range() abort with an error after 100 iterations
2735 without success (which indicates a broken PRNG).
2738 *) Change BN_mod_sqrt() so that it verifies that the input value
2739 is really the square of the return value. (Previously,
2740 BN_mod_sqrt would show GIGO behaviour.)
2743 *) Add named elliptic curves over binary fields from X9.62, SECG,
2744 and WAP/WTLS; add OIDs that were still missing.
2746 [Sheueling Chang Shantz and Douglas Stebila
2747 (Sun Microsystems Laboratories)]
2749 *) Extend the EC library for elliptic curves over binary fields
2750 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2753 EC_GF2m_simple_method
2757 EC_GROUP_new_curve_GF2m
2758 EC_GROUP_set_curve_GF2m
2759 EC_GROUP_get_curve_GF2m
2760 EC_POINT_set_affine_coordinates_GF2m
2761 EC_POINT_get_affine_coordinates_GF2m
2762 EC_POINT_set_compressed_coordinates_GF2m
2764 Point compression for binary fields is disabled by default for
2765 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2768 As binary polynomials are represented as BIGNUMs, various members
2769 of the EC_GROUP and EC_POINT data structures can be shared
2770 between the implementations for prime fields and binary fields;
2771 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2772 are essentially identical to their ..._GFp counterparts.
2773 (For simplicity, the '..._GFp' prefix has been dropped from
2774 various internal method names.)
2776 An internal 'field_div' method (similar to 'field_mul' and
2777 'field_sqr') has been added; this is used only for binary fields.
2779 [Sheueling Chang Shantz and Douglas Stebila
2780 (Sun Microsystems Laboratories)]
2782 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2783 through methods ('mul', 'precompute_mult').
2785 The generic implementations (now internally called 'ec_wNAF_mul'
2786 and 'ec_wNAF_precomputed_mult') remain the default if these
2787 methods are undefined.
2789 [Sheueling Chang Shantz and Douglas Stebila
2790 (Sun Microsystems Laboratories)]
2792 *) New function EC_GROUP_get_degree, which is defined through
2793 EC_METHOD. For curves over prime fields, this returns the bit
2794 length of the modulus.
2796 [Sheueling Chang Shantz and Douglas Stebila
2797 (Sun Microsystems Laboratories)]
2799 *) New functions EC_GROUP_dup, EC_POINT_dup.
2800 (These simply call ..._new and ..._copy).
2802 [Sheueling Chang Shantz and Douglas Stebila
2803 (Sun Microsystems Laboratories)]
2805 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2806 Polynomials are represented as BIGNUMs (where the sign bit is not
2807 used) in the following functions [macros]:
2810 BN_GF2m_sub [= BN_GF2m_add]
2811 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2812 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2813 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2815 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2816 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2817 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2818 BN_GF2m_cmp [= BN_ucmp]
2820 (Note that only the 'mod' functions are actually for fields GF(2^m).
2821 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2823 For some functions, an the irreducible polynomial defining a
2824 field can be given as an 'unsigned int[]' with strictly
2825 decreasing elements giving the indices of those bits that are set;
2826 i.e., p[] represents the polynomial
2827 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2829 p[0] > p[1] > ... > p[k] = 0.
2830 This applies to the following functions:
2835 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2836 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2838 BN_GF2m_mod_sqrt_arr
2839 BN_GF2m_mod_solve_quad_arr
2843 Conversion can be performed by the following functions:
2848 bntest.c has additional tests for binary polynomial arithmetic.
2850 Two implementations for BN_GF2m_mod_div() are available.
2851 The default algorithm simply uses BN_GF2m_mod_inv() and
2852 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2853 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2854 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2856 [Sheueling Chang Shantz and Douglas Stebila
2857 (Sun Microsystems Laboratories)]
2859 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2860 functionality is disabled at compile-time.
2861 [Douglas Stebila <douglas.stebila@sun.com>]
2863 *) Change default behaviour of 'openssl asn1parse' so that more
2864 information is visible when viewing, e.g., a certificate:
2866 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2867 mode the content of non-printable OCTET STRINGs is output in a
2868 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2869 avoid the appearance of a printable string.
2870 [Nils Larsch <nla@trustcenter.de>]
2872 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2874 EC_GROUP_set_asn1_flag()
2875 EC_GROUP_get_asn1_flag()
2876 EC_GROUP_set_point_conversion_form()
2877 EC_GROUP_get_point_conversion_form()
2878 These control ASN1 encoding details:
2879 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
2880 has been set to OPENSSL_EC_NAMED_CURVE.
2881 - Points are encoded in uncompressed form by default; options for
2882 asn1_for are as for point2oct, namely
2883 POINT_CONVERSION_COMPRESSED
2884 POINT_CONVERSION_UNCOMPRESSED
2885 POINT_CONVERSION_HYBRID
2887 Also add 'seed' and 'seed_len' members to EC_GROUP with access
2890 EC_GROUP_get0_seed()
2891 EC_GROUP_get_seed_len()
2892 This is used only for ASN1 purposes (so far).
2893 [Nils Larsch <nla@trustcenter.de>]
2895 *) Add 'field_type' member to EC_METHOD, which holds the NID
2896 of the appropriate field type OID. The new function
2897 EC_METHOD_get_field_type() returns this value.
2898 [Nils Larsch <nla@trustcenter.de>]
2903 EC_POINT_point2hex()
2904 EC_POINT_hex2point()
2905 providing useful interfaces to EC_POINT_point2oct() and
2906 EC_POINT_oct2point().
2907 [Nils Larsch <nla@trustcenter.de>]
2909 *) Change internals of the EC library so that the functions
2910 EC_GROUP_set_generator()
2911 EC_GROUP_get_generator()
2912 EC_GROUP_get_order()
2913 EC_GROUP_get_cofactor()
2914 are implemented directly in crypto/ec/ec_lib.c and not dispatched
2915 to methods, which would lead to unnecessary code duplication when
2916 adding different types of curves.
2917 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
2919 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
2920 arithmetic, and such that modified wNAFs are generated
2921 (which avoid length expansion in many cases).
2924 *) Add a function EC_GROUP_check_discriminant() (defined via
2925 EC_METHOD) that verifies that the curve discriminant is non-zero.
2927 Add a function EC_GROUP_check() that makes some sanity tests
2928 on a EC_GROUP, its generator and order. This includes
2929 EC_GROUP_check_discriminant().
2930 [Nils Larsch <nla@trustcenter.de>]
2932 *) Add ECDSA in new directory crypto/ecdsa/.
2934 Add applications 'openssl ecparam' and 'openssl ecdsa'
2935 (these are based on 'openssl dsaparam' and 'openssl dsa').
2937 ECDSA support is also included in various other files across the
2938 library. Most notably,
2939 - 'openssl req' now has a '-newkey ecdsa:file' option;
2940 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
2941 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
2942 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
2943 them suitable for ECDSA where domain parameters must be
2944 extracted before the specific public key;
2945 - ECDSA engine support has been added.
2946 [Nils Larsch <nla@trustcenter.de>]
2948 *) Include some named elliptic curves, and add OIDs from X9.62,
2949 SECG, and WAP/WTLS. Each curve can be obtained from the new
2951 EC_GROUP_new_by_curve_name(),
2952 and the list of available named curves can be obtained with
2953 EC_get_builtin_curves().
2954 Also add a 'curve_name' member to EC_GROUP objects, which can be
2956 EC_GROUP_set_curve_name()
2957 EC_GROUP_get_curve_name()
2958 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
2960 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2961 was actually never needed) and in BN_mul(). The removal in BN_mul()
2962 required a small change in bn_mul_part_recursive() and the addition
2963 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2964 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2965 bn_sub_words() and bn_add_words() except they take arrays with
2969 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
2971 *) Cleanse PEM buffers before freeing them since they may contain
2973 [Benjamin Bennett <ben@psc.edu>]
2975 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2976 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2977 authentication-only ciphersuites.
2980 *) Since AES128 and AES256 share a single mask bit in the logic of
2981 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2982 kludge to work properly if AES128 is available and AES256 isn't.
2985 *) Expand security boundary to match 1.1.1 module.
2988 *) Remove redundant features: hash file source, editing of test vectors
2989 modify fipsld to use external fips_premain.c signature.
2992 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
2993 run algorithm test programs.
2996 *) Make algorithm test programs more tolerant of whitespace.
2999 *) Have SSL/TLS server implementation tolerate "mismatched" record
3000 protocol version while receiving ClientHello even if the
3001 ClientHello is fragmented. (The server can't insist on the
3002 particular protocol version it has chosen before the ServerHello
3003 message has informed the client about his choice.)
3006 *) Load error codes if they are not already present instead of using a
3007 static variable. This allows them to be cleanly unloaded and reloaded.
3010 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3012 *) Introduce limits to prevent malicious keys being able to
3013 cause a denial of service. (CVE-2006-2940)
3014 [Steve Henson, Bodo Moeller]
3016 *) Fix ASN.1 parsing of certain invalid structures that can result
3017 in a denial of service. (CVE-2006-2937) [Steve Henson]
3019 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3020 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3022 *) Fix SSL client code which could crash if connecting to a
3023 malicious SSLv2 server. (CVE-2006-4343)
3024 [Tavis Ormandy and Will Drewry, Google Security Team]
3026 *) Change ciphersuite string processing so that an explicit
3027 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3028 will no longer include "AES128-SHA"), and any other similar
3029 ciphersuite (same bitmap) from *other* protocol versions (so that
3030 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3031 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3032 changes from 0.9.8b and 0.9.8d.
3035 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3037 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3038 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3040 *) Change the Unix randomness entropy gathering to use poll() when
3041 possible instead of select(), since the latter has some
3042 undesirable limitations.
3043 [Darryl Miles via Richard Levitte and Bodo Moeller]
3045 *) Disable rogue ciphersuites:
3047 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3048 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3049 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3051 The latter two were purportedly from
3052 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3055 Also deactive the remaining ciphersuites from
3056 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3057 unofficial, and the ID has long expired.
3060 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3061 dual-core machines) and other potential thread-safety issues.
3064 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3066 *) Adapt fipsld and the build system to link against the validated FIPS
3067 module in FIPS mode.
3070 *) Fixes for VC++ 2005 build under Windows.
3073 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3074 from a Windows bash shell such as MSYS. It is autodetected from the
3075 "config" script when run from a VC++ environment. Modify standard VC++
3076 build to use fipscanister.o from the GNU make build.
3079 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3081 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3082 The value now differs depending on if you build for FIPS or not.
3083 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3084 safely run with a non-FIPSed libcrypto, as it may crash because of
3085 the difference induced by this change.
3088 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3090 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3091 (part of SSL_OP_ALL). This option used to disable the
3092 countermeasure against man-in-the-middle protocol-version
3093 rollback in the SSL 2.0 server implementation, which is a bad
3094 idea. (CVE-2005-2969)
3096 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3097 for Information Security, National Institute of Advanced Industrial
3098 Science and Technology [AIST], Japan)]
3100 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3101 mainly for FIPS compliance and not fully integrated at this stage.
3104 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3105 the exponentiation using a fixed-length exponent. (Otherwise,
3106 the information leaked through timing could expose the secret key
3107 after many signatures; cf. Bleichenbacher's attack on DSA with
3111 *) Make a new fixed-window mod_exp implementation the default for
3112 RSA, DSA, and DH private-key operations so that the sequence of
3113 squares and multiplies and the memory access pattern are
3114 independent of the particular secret key. This will mitigate
3115 cache-timing and potential related attacks.
3117 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3118 and this is automatically used by BN_mod_exp_mont() if the new flag
3119 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3120 will use this BN flag for private exponents unless the flag
3121 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3122 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3124 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3126 *) Change the client implementation for SSLv23_method() and
3127 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3128 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3129 (Previously, the SSL 2.0 backwards compatible Client Hello
3130 message format would be used even with SSL_OP_NO_SSLv2.)
3133 *) Add support for smime-type MIME parameter in S/MIME messages which some
3137 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3138 a threadsafe manner. Modify rsa code to use new function and add calls
3139 to dsa and dh code (which had race conditions before).
3142 *) Include the fixed error library code in the C error file definitions
3143 instead of fixing them up at runtime. This keeps the error code
3144 structures constant.
3147 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3149 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3152 *) Fixes for newer kerberos headers. NB: the casts are needed because
3153 the 'length' field is signed on one version and unsigned on another
3154 with no (?) obvious way to tell the difference, without these VC++
3155 complains. Also the "definition" of FAR (blank) is no longer included
3156 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3157 some needed definitions.
3160 *) Undo Cygwin change.
3163 *) Added support for proxy certificates according to RFC 3820.
3164 Because they may be a security thread to unaware applications,
3165 they must be explicitely allowed in run-time. See
3166 docs/HOWTO/proxy_certificates.txt for further information.
3169 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3171 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3172 server and client random values. Previously
3173 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3174 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3176 This change has negligible security impact because:
3178 1. Server and client random values still have 24 bytes of pseudo random
3181 2. Server and client random values are sent in the clear in the initial
3184 3. The master secret is derived using the premaster secret (48 bytes in
3185 size for static RSA ciphersuites) as well as client server and random
3188 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3191 [Stephen Henson, reported by UK NISCC]
3193 *) Use Windows randomness collection on Cygwin.
3196 *) Fix hang in EGD/PRNGD query when communication socket is closed
3197 prematurely by EGD/PRNGD.
3198 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3200 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3203 *) Back-port of selected performance improvements from development
3204 branch, as well as improved support for PowerPC platforms.
3207 *) Add lots of checks for memory allocation failure, error codes to indicate
3208 failure and freeing up memory if a failure occurs.
3209 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3211 *) Add new -passin argument to dgst.
3214 *) Perform some character comparisons of different types in X509_NAME_cmp:
3215 this is needed for some certificates that reencode DNs into UTF8Strings
3216 (in violation of RFC3280) and can't or wont issue name rollover
3220 *) Make an explicit check during certificate validation to see that
3221 the CA setting in each certificate on the chain is correct. As a
3222 side effect always do the following basic checks on extensions,
3223 not just when there's an associated purpose to the check:
3225 - if there is an unhandled critical extension (unless the user
3226 has chosen to ignore this fault)
3227 - if the path length has been exceeded (if one is set at all)
3228 - that certain extensions fit the associated purpose (if one has
3232 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3234 *) Avoid a race condition when CRLs are checked in a multi threaded
3235 environment. This would happen due to the reordering of the revoked
3236 entries during signature checking and serial number lookup. Now the
3237 encoding is cached and the serial number sort performed under a lock.
3238 Add new STACK function sk_is_sorted().
3241 *) Add Delta CRL to the extension code.
3244 *) Various fixes to s3_pkt.c so alerts are sent properly.
3245 [David Holmes <d.holmes@f5.com>]
3247 *) Reduce the chances of duplicate issuer name and serial numbers (in
3248 violation of RFC3280) using the OpenSSL certificate creation utilities.
3249 This is done by creating a random 64 bit value for the initial serial
3250 number when a serial number file is created or when a self signed
3251 certificate is created using 'openssl req -x509'. The initial serial
3252 number file is created using 'openssl x509 -next_serial' in CA.pl
3253 rather than being initialized to 1.
3256 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3258 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3259 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3260 [Joe Orton, Steve Henson]
3262 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3264 [Joe Orton, Steve Henson]
3266 *) Make it possible to have multiple active certificates with the same
3267 subject in the CA index file. This is done only if the keyword
3268 'unique_subject' is set to 'no' in the main CA section (default
3269 if 'CA_default') of the configuration file. The value is saved
3270 with the database itself in a separate index attribute file,
3271 named like the index file with '.attr' appended to the name.
3274 *) X509 verify fixes. Disable broken certificate workarounds when
3275 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3276 keyUsage extension present. Don't accept CRLs with unhandled critical
3277 extensions: since verify currently doesn't process CRL extensions this
3278 rejects a CRL with *any* critical extensions. Add new verify error codes
3282 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3283 A clarification of RFC2560 will require the use of OCTET STRINGs and
3284 some implementations cannot handle the current raw format. Since OpenSSL
3285 copies and compares OCSP nonces as opaque blobs without any attempt at
3286 parsing them this should not create any compatibility issues.
3289 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3290 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3291 this HMAC (and other) operations are several times slower than OpenSSL
3295 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3296 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3298 *) Use the correct content when signing type "other".
3301 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3303 *) Fix various bugs revealed by running the NISCC test suite:
3305 Stop out of bounds reads in the ASN1 code when presented with
3306 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3308 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3310 If verify callback ignores invalid public key errors don't try to check
3311 certificate signature with the NULL public key.
3315 *) New -ignore_err option in ocsp application to stop the server
3316 exiting on the first error in a request.
3319 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3320 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3324 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3325 extra data after the compression methods not only for TLS 1.0
3326 but also for SSL 3.0 (as required by the specification).
3327 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3329 *) Change X509_certificate_type() to mark the key as exported/exportable
3330 when it's 512 *bits* long, not 512 bytes.
3333 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3334 blocks during encryption.
3337 *) Various fixes to base64 BIO and non blocking I/O. On write
3338 flushes were not handled properly if the BIO retried. On read
3339 data was not being buffered properly and had various logic bugs.
3340 This also affects blocking I/O when the data being decoded is a
3344 *) Various S/MIME bugfixes and compatibility changes:
3345 output correct application/pkcs7 MIME type if
3346 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3347 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3348 of files as .eml work). Correctly handle very long lines in MIME
3352 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3354 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3355 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3356 a protocol version number mismatch like a decryption error
3357 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3360 *) Turn on RSA blinding by default in the default implementation
3361 to avoid a timing attack. Applications that don't want it can call
3362 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3363 They would be ill-advised to do so in most cases.
3364 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3366 *) Change RSA blinding code so that it works when the PRNG is not
3367 seeded (in this case, the secret RSA exponent is abused as
3368 an unpredictable seed -- if it is not unpredictable, there
3369 is no point in blinding anyway). Make RSA blinding thread-safe
3370 by remembering the creator's thread ID in rsa->blinding and
3371 having all other threads use local one-time blinding factors
3372 (this requires more computation than sharing rsa->blinding, but
3373 avoids excessive locking; and if an RSA object is not shared
3374 between threads, blinding will still be very fast).
3377 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3378 ENGINE as defaults for all supported algorithms irrespective of
3379 the 'flags' parameter. 'flags' is now honoured, so applications
3380 should make sure they are passing it correctly.
3383 *) Target "mingw" now allows native Windows code to be generated in
3384 the Cygwin environment as well as with the MinGW compiler.
3387 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3389 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3390 via timing by performing a MAC computation even if incorrrect
3391 block cipher padding has been found. This is a countermeasure
3392 against active attacks where the attacker has to distinguish
3393 between bad padding and a MAC verification error. (CVE-2003-0078)
3395 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3396 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3397 Martin Vuagnoux (EPFL, Ilion)]
3399 *) Make the no-err option work as intended. The intention with no-err
3400 is not to have the whole error stack handling routines removed from
3401 libcrypto, it's only intended to remove all the function name and
3402 reason texts, thereby removing some of the footprint that may not
3403 be interesting if those errors aren't displayed anyway.
3405 NOTE: it's still possible for any application or module to have it's
3406 own set of error texts inserted. The routines are there, just not
3407 used by default when no-err is given.
3410 *) Add support for FreeBSD on IA64.
3411 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3413 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3414 Kerberos function mit_des_cbc_cksum(). Before this change,
3415 the value returned by DES_cbc_cksum() was like the one from
3416 mit_des_cbc_cksum(), except the bytes were swapped.
3417 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3419 *) Allow an application to disable the automatic SSL chain building.
3420 Before this a rather primitive chain build was always performed in
3421 ssl3_output_cert_chain(): an application had no way to send the
3422 correct chain if the automatic operation produced an incorrect result.
3424 Now the chain builder is disabled if either:
3426 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3428 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3430 The reasoning behind this is that an application would not want the
3431 auto chain building to take place if extra chain certificates are
3432 present and it might also want a means of sending no additional
3433 certificates (for example the chain has two certificates and the
3437 *) Add the possibility to build without the ENGINE framework.
3438 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3440 *) Under Win32 gmtime() can return NULL: check return value in
3441 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3444 *) DSA routines: under certain error conditions uninitialized BN objects
3445 could be freed. Solution: make sure initialization is performed early
3446 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3447 Nils Larsch <nla@trustcenter.de> via PR#459)
3450 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3451 checked on reconnect on the client side, therefore session resumption
3452 could still fail with a "ssl session id is different" error. This
3453 behaviour is masked when SSL_OP_ALL is used due to
3454 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3455 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3456 followup to PR #377.
3459 *) IA-32 assembler support enhancements: unified ELF targets, support
3460 for SCO/Caldera platforms, fix for Cygwin shared build.
3463 *) Add support for FreeBSD on sparc64. As a consequence, support for
3464 FreeBSD on non-x86 processors is separate from x86 processors on
3465 the config script, much like the NetBSD support.
3466 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3468 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3470 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3473 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3474 code (06) was taken as the first octet of the session ID and the last
3475 octet was ignored consequently. As a result SSLv2 client side session
3476 caching could not have worked due to the session ID mismatch between
3478 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3482 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3483 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3487 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3488 seems that in spite of existing for more than a year, many application
3489 author have done nothing to provide the necessary callbacks, which
3490 means that this particular engine will not work properly anywhere.
3491 This is a very unfortunate situation which forces us, in the name
3492 of usability, to give the hw_ncipher.c a static lock, which is part
3494 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3495 appear in 0.9.8 or later. We EXPECT application authors to have
3496 dealt properly with this when 0.9.8 is released (unless we actually
3497 make such changes in the libcrypto locking code that changes will
3498 have to be made anyway).
3501 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3502 octets have been read, EOF or an error occurs. Without this change
3503 some truncated ASN1 structures will not produce an error.
3506 *) Disable Heimdal support, since it hasn't been fully implemented.
3507 Still give the possibility to force the use of Heimdal, but with
3508 warnings and a request that patches get sent to openssl-dev.
3511 *) Add the VC-CE target, introduce the WINCE sysname, and add
3512 INSTALL.WCE and appropriate conditionals to make it build.
3513 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3515 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3516 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3517 edit numbers of the version.
3518 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3520 *) Introduce safe string copy and catenation functions
3521 (BUF_strlcpy() and BUF_strlcat()).
3522 [Ben Laurie (CHATS) and Richard Levitte]
3524 *) Avoid using fixed-size buffers for one-line DNs.
3525 [Ben Laurie (CHATS)]
3527 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3528 resizing buffers containing secrets, and use where appropriate.
3529 [Ben Laurie (CHATS)]
3531 *) Avoid using fixed size buffers for configuration file location.
3532 [Ben Laurie (CHATS)]
3534 *) Avoid filename truncation for various CA files.
3535 [Ben Laurie (CHATS)]
3537 *) Use sizeof in preference to magic numbers.
3538 [Ben Laurie (CHATS)]
3540 *) Avoid filename truncation in cert requests.
3541 [Ben Laurie (CHATS)]
3543 *) Add assertions to check for (supposedly impossible) buffer
3545 [Ben Laurie (CHATS)]
3547 *) Don't cache truncated DNS entries in the local cache (this could
3548 potentially lead to a spoofing attack).
3549 [Ben Laurie (CHATS)]
3551 *) Fix various buffers to be large enough for hex/decimal
3552 representations in a platform independent manner.
3553 [Ben Laurie (CHATS)]
3555 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3556 resizing buffers containing secrets, and use where appropriate.
3557 [Ben Laurie (CHATS)]
3559 *) Add BIO_indent() to avoid much slightly worrying code to do
3561 [Ben Laurie (CHATS)]
3563 *) Convert sprintf()/BIO_puts() to BIO_printf().
3564 [Ben Laurie (CHATS)]
3566 *) buffer_gets() could terminate with the buffer only half
3568 [Ben Laurie (CHATS)]
3570 *) Add assertions to prevent user-supplied crypto functions from
3571 overflowing internal buffers by having large block sizes, etc.
3572 [Ben Laurie (CHATS)]
3574 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3576 [Ben Laurie (CHATS)]
3578 *) Eliminate unused copy of key in RC4.
3579 [Ben Laurie (CHATS)]
3581 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3582 [Ben Laurie (CHATS)]
3584 *) Fix off-by-one error in EGD path.
3585 [Ben Laurie (CHATS)]
3587 *) If RANDFILE path is too long, ignore instead of truncating.
3588 [Ben Laurie (CHATS)]
3590 *) Eliminate unused and incorrectly sized X.509 structure
3592 [Ben Laurie (CHATS)]
3594 *) Eliminate unused and dangerous function knumber().
3595 [Ben Laurie (CHATS)]
3597 *) Eliminate unused and dangerous structure, KSSL_ERR.
3598 [Ben Laurie (CHATS)]
3600 *) Protect against overlong session ID context length in an encoded
3601 session object. Since these are local, this does not appear to be
3603 [Ben Laurie (CHATS)]
3605 *) Change from security patch (see 0.9.6e below) that did not affect
3606 the 0.9.6 release series:
3608 Remote buffer overflow in SSL3 protocol - an attacker could
3609 supply an oversized master key in Kerberos-enabled versions.
3611 [Ben Laurie (CHATS)]
3613 *) Change the SSL kerb5 codes to match RFC 2712.
3616 *) Make -nameopt work fully for req and add -reqopt switch.
3617 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3619 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3620 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3622 *) Make sure tests can be performed even if the corresponding algorithms
3623 have been removed entirely. This was also the last step to make
3624 OpenSSL compilable with DJGPP under all reasonable conditions.
3625 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3627 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3628 to allow version independent disabling of normally unselected ciphers,
3629 which may be activated as a side-effect of selecting a single cipher.
3631 (E.g., cipher list string "RSA" enables ciphersuites that are left
3632 out of "ALL" because they do not provide symmetric encryption.
3633 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3634 [Lutz Jaenicke, Bodo Moeller]
3636 *) Add appropriate support for separate platform-dependent build
3637 directories. The recommended way to make a platform-dependent
3638 build directory is the following (tested on Linux), maybe with
3641 # Place yourself outside of the OpenSSL source tree. In
3642 # this example, the environment variable OPENSSL_SOURCE
3643 # is assumed to contain the absolute OpenSSL source directory.
3644 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3645 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3646 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3647 mkdir -p `dirname $F`
3648 ln -s $OPENSSL_SOURCE/$F $F
3651 To be absolutely sure not to disturb the source tree, a "make clean"
3652 is a good thing. If it isn't successfull, don't worry about it,
3653 it probably means the source directory is very clean.
3656 *) Make sure any ENGINE control commands make local copies of string
3657 pointers passed to them whenever necessary. Otherwise it is possible
3658 the caller may have overwritten (or deallocated) the original string
3659 data when a later ENGINE operation tries to use the stored values.
3660 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3662 *) Improve diagnostics in file reading and command-line digests.
3663 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3665 *) Add AES modes CFB and OFB to the object database. Correct an
3666 error in AES-CFB decryption.
3669 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3670 allows existing EVP_CIPHER_CTX structures to be reused after
3671 calling EVP_*Final(). This behaviour is used by encryption
3672 BIOs and some applications. This has the side effect that
3673 applications must explicitly clean up cipher contexts with
3674 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3677 *) Check the values of dna and dnb in bn_mul_recursive before calling
3678 bn_mul_comba (a non zero value means the a or b arrays do not contain
3679 n2 elements) and fallback to bn_mul_normal if either is not zero.
3682 *) Fix escaping of non-ASCII characters when using the -subj option
3683 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3686 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3687 form for "surname", serialNumber has no short form.
3688 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3689 therefore remove "mail" short name for "internet 7".
3690 The OID for unique identifiers in X509 certificates is
3691 x500UniqueIdentifier, not uniqueIdentifier.
3692 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3695 *) Add an "init" command to the ENGINE config module and auto initialize
3696 ENGINEs. Without any "init" command the ENGINE will be initialized
3697 after all ctrl commands have been executed on it. If init=1 the
3698 ENGINE is initailized at that point (ctrls before that point are run
3699 on the uninitialized ENGINE and after on the initialized one). If
3700 init=0 then the ENGINE will not be iniatialized at all.
3703 *) Fix the 'app_verify_callback' interface so that the user-defined
3704 argument is actually passed to the callback: In the
3705 SSL_CTX_set_cert_verify_callback() prototype, the callback
3706 declaration has been changed from
3709 int (*cb)(X509_STORE_CTX *,void *);
3710 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3711 i=s->ctx->app_verify_callback(&ctx)
3712 has been changed into
3713 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3715 To update applications using SSL_CTX_set_cert_verify_callback(),
3716 a dummy argument can be added to their callback functions.
3717 [D. K. Smetters <smetters@parc.xerox.com>]
3719 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3720 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3722 *) Add and OPENSSL_LOAD_CONF define which will cause
3723 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3724 This allows older applications to transparently support certain
3725 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3726 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3727 load the config file and OPENSSL_add_all_algorithms_conf() which will
3728 always load it have also been added.
3731 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3732 Adjust NIDs and EVP layer.
3733 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3735 *) Config modules support in openssl utility.
3737 Most commands now load modules from the config file,
3738 though in a few (such as version) this isn't done
3739 because it couldn't be used for anything.
3741 In the case of ca and req the config file used is
3742 the same as the utility itself: that is the -config
3743 command line option can be used to specify an
3747 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3748 use "openssl_conf" if filename is NULL use default openssl config file.
3751 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3752 config section name. Add a new flag to tolerate a missing config file
3753 and move code to CONF_modules_load_file().
3756 *) Support for crypto accelerator cards from Accelerated Encryption
3757 Processing, www.aep.ie. (Use engine 'aep')
3758 The support was copied from 0.9.6c [engine] and adapted/corrected
3759 to work with the new engine framework.
3760 [AEP Inc. and Richard Levitte]
3762 *) Support for SureWare crypto accelerator cards from Baltimore
3763 Technologies. (Use engine 'sureware')
3764 The support was copied from 0.9.6c [engine] and adapted
3765 to work with the new engine framework.
3768 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3769 make the newer ENGINE framework commands for the CHIL engine work.
3770 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3772 *) Make it possible to produce shared libraries on ReliantUNIX.
3773 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3775 *) Add the configuration target debug-linux-ppro.
3776 Make 'openssl rsa' use the general key loading routines
3777 implemented in apps.c, and make those routines able to
3778 handle the key format FORMAT_NETSCAPE and the variant
3780 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3782 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3783 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3785 *) Add -keyform to rsautl, and document -engine.
3786 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3788 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3789 BIO_R_NO_SUCH_FILE error code rather than the generic
3790 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3793 *) Add new functions
3795 ERR_peek_last_error_line
3796 ERR_peek_last_error_line_data.
3797 These are similar to
3800 ERR_peek_error_line_data,
3801 but report on the latest error recorded rather than the first one
3802 still in the error queue.
3803 [Ben Laurie, Bodo Moeller]
3805 *) default_algorithms option in ENGINE config module. This allows things
3807 default_algorithms = ALL
3808 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3811 *) Prelminary ENGINE config module.
3814 *) New experimental application configuration code.
3817 *) Change the AES code to follow the same name structure as all other
3818 symmetric ciphers, and behave the same way. Move everything to
3819 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3820 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3822 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3823 [Ben Laurie and Theo de Raadt]
3825 *) Add option to output public keys in req command.
3826 [Massimiliano Pala madwolf@openca.org]
3828 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3829 (up to about 10% better than before for P-192 and P-224).
3832 *) New functions/macros
3834 SSL_CTX_set_msg_callback(ctx, cb)
3835 SSL_CTX_set_msg_callback_arg(ctx, arg)
3836 SSL_set_msg_callback(ssl, cb)
3837 SSL_set_msg_callback_arg(ssl, arg)
3839 to request calling a callback function
3841 void cb(int write_p, int version, int content_type,
3842 const void *buf, size_t len, SSL *ssl, void *arg)
3844 whenever a protocol message has been completely received
3845 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3846 protocol version according to which the SSL library interprets
3847 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3848 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3849 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3850 specification (change_cipher_spec(20), alert(21), handshake(22)).
3851 'buf' and 'len' point to the actual message, 'ssl' to the
3852 SSL object, and 'arg' is the application-defined value set by
3853 SSL[_CTX]_set_msg_callback_arg().
3855 'openssl s_client' and 'openssl s_server' have new '-msg' options
3856 to enable a callback that displays all protocol messages.
3859 *) Change the shared library support so shared libraries are built as
3860 soon as the corresponding static library is finished, and thereby get
3861 openssl and the test programs linked against the shared library.
3862 This still only happens when the keyword "shard" has been given to
3863 the configuration scripts.
3865 NOTE: shared library support is still an experimental thing, and
3866 backward binary compatibility is still not guaranteed.
3867 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3869 *) Add support for Subject Information Access extension.
3870 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3872 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3873 additional bytes when new memory had to be allocated, not just
3874 when reusing an existing buffer.
3877 *) New command line and configuration option 'utf8' for the req command.
3878 This allows field values to be specified as UTF8 strings.
3881 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
3882 runs for the former and machine-readable output for the latter.
3885 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
3886 of the e-mail address in the DN (i.e., it will go into a certificate
3887 extension only). The new configuration file option 'email_in_dn = no'
3888 has the same effect.
3889 [Massimiliano Pala madwolf@openca.org]
3891 *) Change all functions with names starting with des_ to be starting
3892 with DES_ instead. Add wrappers that are compatible with libdes,
3893 but are named _ossl_old_des_*. Finally, add macros that map the
3894 des_* symbols to the corresponding _ossl_old_des_* if libdes
3895 compatibility is desired. If OpenSSL 0.9.6c compatibility is
3896 desired, the des_* symbols will be mapped to DES_*, with one
3899 Since we provide two compatibility mappings, the user needs to
3900 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
3901 compatibility is desired. The default (i.e., when that macro
3902 isn't defined) is OpenSSL 0.9.6c compatibility.
3904 There are also macros that enable and disable the support of old
3905 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
3906 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
3907 are defined, the default will apply: to support the old des routines.
3909 In either case, one must include openssl/des.h to get the correct
3910 definitions. Do not try to just include openssl/des_old.h, that
3913 NOTE: This is a major break of an old API into a new one. Software
3914 authors are encouraged to switch to the DES_ style functions. Some
3915 time in the future, des_old.h and the libdes compatibility functions
3916 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
3917 default), and then completely removed.
3920 *) Test for certificates which contain unsupported critical extensions.
3921 If such a certificate is found during a verify operation it is
3922 rejected by default: this behaviour can be overridden by either
3923 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
3924 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
3925 X509_supported_extension() has also been added which returns 1 if a
3926 particular extension is supported.
3929 *) Modify the behaviour of EVP cipher functions in similar way to digests
3930 to retain compatibility with existing code.
3933 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
3934 compatibility with existing code. In particular the 'ctx' parameter does
3935 not have to be to be initialized before the call to EVP_DigestInit() and
3936 it is tidied up after a call to EVP_DigestFinal(). New function
3937 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
3938 EVP_MD_CTX_copy() changed to not require the destination to be
3939 initialized valid and new function EVP_MD_CTX_copy_ex() added which
3940 requires the destination to be valid.
3942 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
3943 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
3946 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
3947 so that complete 'Handshake' protocol structures are kept in memory
3948 instead of overwriting 'msg_type' and 'length' with 'body' data.
3951 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
3952 [Massimo Santin via Richard Levitte]
3954 *) Major restructuring to the underlying ENGINE code. This includes
3955 reduction of linker bloat, separation of pure "ENGINE" manipulation
3956 (initialisation, etc) from functionality dealing with implementations
3957 of specific crypto iterfaces. This change also introduces integrated
3958 support for symmetric ciphers and digest implementations - so ENGINEs
3959 can now accelerate these by providing EVP_CIPHER and EVP_MD
3960 implementations of their own. This is detailed in crypto/engine/README
3961 as it couldn't be adequately described here. However, there are a few
3962 API changes worth noting - some RSA, DSA, DH, and RAND functions that
3963 were changed in the original introduction of ENGINE code have now
3964 reverted back - the hooking from this code to ENGINE is now a good
3965 deal more passive and at run-time, operations deal directly with
3966 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
3967 dereferencing through an ENGINE pointer any more. Also, the ENGINE
3968 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
3969 they were not being used by the framework as there is no concept of a
3970 BIGNUM_METHOD and they could not be generalised to the new
3971 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
3972 ENGINE_cpy() has been removed as it cannot be consistently defined in
3976 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
3979 *) Change mkdef.pl to sort symbols that get the same entry number,
3980 and make sure the automatically generated functions ERR_load_*
3981 become part of libeay.num as well.
3984 *) New function SSL_renegotiate_pending(). This returns true once
3985 renegotiation has been requested (either SSL_renegotiate() call
3986 or HelloRequest/ClientHello receveived from the peer) and becomes
3987 false once a handshake has been completed.
3988 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
3989 sends a HelloRequest, but does not ensure that a handshake takes
3990 place. SSL_renegotiate_pending() is useful for checking if the
3991 client has followed the request.)
3994 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
3995 By default, clients may request session resumption even during
3996 renegotiation (if session ID contexts permit); with this option,
3997 session resumption is possible only in the first handshake.
3999 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4000 more bits available for options that should not be part of
4001 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4004 *) Add some demos for certificate and certificate request creation.
4007 *) Make maximum certificate chain size accepted from the peer application
4008 settable (SSL*_get/set_max_cert_list()), as proposed by
4009 "Douglas E. Engert" <deengert@anl.gov>.
4012 *) Add support for shared libraries for Unixware-7
4013 (Boyd Lynn Gerber <gerberb@zenez.com>).
4016 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4017 be done prior to destruction. Use this to unload error strings from
4018 ENGINEs that load their own error strings. NB: This adds two new API
4019 functions to "get" and "set" this destroy handler in an ENGINE.
4022 *) Alter all existing ENGINE implementations (except "openssl" and
4023 "openbsd") to dynamically instantiate their own error strings. This
4024 makes them more flexible to be built both as statically-linked ENGINEs
4025 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4026 Also, add stub code to each that makes building them as self-contained
4027 shared-libraries easier (see README.ENGINE).
4030 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4031 implementations into applications that are completely implemented in
4032 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4033 commands that can be used to configure what shared-library to load and
4034 to control aspects of the way it is handled. Also, made an update to
4035 the README.ENGINE file that brings its information up-to-date and
4036 provides some information and instructions on the "dynamic" ENGINE
4037 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4040 *) Make it possible to unload ranges of ERR strings with a new
4041 "ERR_unload_strings" function.
4044 *) Add a copy() function to EVP_MD.
4047 *) Make EVP_MD routines take a context pointer instead of just the
4048 md_data void pointer.
4051 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4052 that the digest can only process a single chunk of data
4053 (typically because it is provided by a piece of
4054 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4055 is only going to provide a single chunk of data, and hence the
4056 framework needn't accumulate the data for oneshot drivers.
4059 *) As with "ERR", make it possible to replace the underlying "ex_data"
4060 functions. This change also alters the storage and management of global
4061 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4062 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4063 index counters. The API functions that use this state have been changed
4064 to take a "class_index" rather than pointers to the class's local STACK
4065 and counter, and there is now an API function to dynamically create new
4066 classes. This centralisation allows us to (a) plug a lot of the
4067 thread-safety problems that existed, and (b) makes it possible to clean
4068 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4069 such data would previously have always leaked in application code and
4070 workarounds were in place to make the memory debugging turn a blind eye
4071 to it. Application code that doesn't use this new function will still
4072 leak as before, but their memory debugging output will announce it now
4073 rather than letting it slide.
4075 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4076 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4077 has a return value to indicate success or failure.
4080 *) Make it possible to replace the underlying "ERR" functions such that the
4081 global state (2 LHASH tables and 2 locks) is only used by the "default"
4082 implementation. This change also adds two functions to "get" and "set"
4083 the implementation prior to it being automatically set the first time
4084 any other ERR function takes place. Ie. an application can call "get",
4085 pass the return value to a module it has just loaded, and that module
4086 can call its own "set" function using that value. This means the
4087 module's "ERR" operations will use (and modify) the error state in the
4088 application and not in its own statically linked copy of OpenSSL code.
4091 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4092 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4093 the operation, and provides a more encapsulated way for external code
4094 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4095 to use these functions rather than manually incrementing the counts.
4097 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4100 *) Add EVP test program.
4103 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4106 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4107 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4108 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4109 These allow a CRL to be built without having to access X509_CRL fields
4110 directly. Modify 'ca' application to use new functions.
4113 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4114 bug workarounds. Rollback attack detection is a security feature.
4115 The problem will only arise on OpenSSL servers when TLSv1 is not
4116 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4117 Software authors not wanting to support TLSv1 will have special reasons
4118 for their choice and can explicitly enable this option.
4119 [Bodo Moeller, Lutz Jaenicke]
4121 *) Rationalise EVP so it can be extended: don't include a union of
4122 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4123 (similar to those existing for EVP_CIPHER_CTX).
4128 EVP_MD_CTX_init(&md); /* new function call */
4129 EVP_DigestInit(&md, EVP_sha1());
4130 EVP_DigestUpdate(&md, in, len);
4131 EVP_DigestFinal(&md, out, NULL);
4132 EVP_MD_CTX_cleanup(&md); /* new function call */
4136 *) Make DES key schedule conform to the usual scheme, as well as
4137 correcting its structure. This means that calls to DES functions
4138 now have to pass a pointer to a des_key_schedule instead of a
4139 plain des_key_schedule (which was actually always a pointer
4142 des_key_schedule ks;
4144 des_set_key_checked(..., &ks);
4145 des_ncbc_encrypt(..., &ks, ...);
4147 (Note that a later change renames 'des_...' into 'DES_...'.)
4150 *) Initial reduction of linker bloat: the use of some functions, such as
4151 PEM causes large amounts of unused functions to be linked in due to
4152 poor organisation. For example pem_all.c contains every PEM function
4153 which has a knock on effect of linking in large amounts of (unused)
4154 ASN1 code. Grouping together similar functions and splitting unrelated
4155 functions prevents this.
4158 *) Cleanup of EVP macros.
4161 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4162 correct _ecb suffix.
4165 *) Add initial OCSP responder support to ocsp application. The
4166 revocation information is handled using the text based index
4167 use by the ca application. The responder can either handle
4168 requests generated internally, supplied in files (for example
4169 via a CGI script) or using an internal minimal server.
4172 *) Add configuration choices to get zlib compression for TLS.
4175 *) Changes to Kerberos SSL for RFC 2712 compliance:
4176 1. Implemented real KerberosWrapper, instead of just using
4177 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4178 2. Implemented optional authenticator field of KerberosWrapper.
4180 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4181 and authenticator structs; see crypto/krb5/.
4183 Generalized Kerberos calls to support multiple Kerberos libraries.
4184 [Vern Staats <staatsvr@asc.hpc.mil>,
4185 Jeffrey Altman <jaltman@columbia.edu>
4186 via Richard Levitte]
4188 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4189 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4190 values for each of the key sizes rather than having just
4191 parameters (and 'speed' generating keys each time).
4194 *) Speed up EVP routines.
4197 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4198 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4199 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4200 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4202 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4203 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4204 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4207 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4209 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4212 *) Added the OS2-EMX target.
4213 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4215 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4216 to support NCONF routines in extension code. New function CONF_set_nconf()
4217 to allow functions which take an NCONF to also handle the old LHASH
4218 structure: this means that the old CONF compatible routines can be
4219 retained (in particular wrt extensions) without having to duplicate the
4220 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4223 *) Enhance the general user interface with mechanisms for inner control
4224 and with possibilities to have yes/no kind of prompts.
4227 *) Change all calls to low level digest routines in the library and
4228 applications to use EVP. Add missing calls to HMAC_cleanup() and
4229 don't assume HMAC_CTX can be copied using memcpy().
4230 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4232 *) Add the possibility to control engines through control names but with
4233 arbitrary arguments instead of just a string.
4234 Change the key loaders to take a UI_METHOD instead of a callback
4235 function pointer. NOTE: this breaks binary compatibility with earlier
4236 versions of OpenSSL [engine].
4237 Adapt the nCipher code for these new conditions and add a card insertion
4241 *) Enhance the general user interface with mechanisms to better support
4242 dialog box interfaces, application-defined prompts, the possibility
4243 to use defaults (for example default passwords from somewhere else)
4244 and interrupts/cancellations.
4247 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4248 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4251 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4252 tidy up some unnecessarily weird code in 'sk_new()').
4253 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4255 *) Change the key loading routines for ENGINEs to use the same kind
4256 callback (pem_password_cb) as all other routines that need this
4260 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4261 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4262 than this minimum value is recommended.
4265 *) New random seeder for OpenVMS, using the system process statistics
4266 that are easily reachable.
4269 *) Windows apparently can't transparently handle global
4270 variables defined in DLLs. Initialisations such as:
4272 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4274 wont compile. This is used by the any applications that need to
4275 declare their own ASN1 modules. This was fixed by adding the option
4276 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4277 needed for static libraries under Win32.
4280 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4281 setting of purpose and trust fields. New X509_STORE trust and
4282 purpose functions and tidy up setting in other SSL functions.
4285 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4286 structure. These are inherited by X509_STORE_CTX when it is
4287 initialised. This allows various defaults to be set in the
4288 X509_STORE structure (such as flags for CRL checking and custom
4289 purpose or trust settings) for functions which only use X509_STORE_CTX
4290 internally such as S/MIME.
4292 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4293 trust settings if they are not set in X509_STORE. This allows X509_STORE
4294 purposes and trust (in S/MIME for example) to override any set by default.
4296 Add command line options for CRL checking to smime, s_client and s_server
4300 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4301 are set then the CRL is looked up in the X509_STORE structure and
4302 its validity and signature checked, then if the certificate is found
4303 in the CRL the verify fails with a revoked error.
4305 Various new CRL related callbacks added to X509_STORE_CTX structure.
4307 Command line options added to 'verify' application to support this.
4309 This needs some additional work, such as being able to handle multiple
4310 CRLs with different times, extension based lookup (rather than just
4311 by subject name) and ultimately more complete V2 CRL extension
4315 *) Add a general user interface API (crypto/ui/). This is designed
4316 to replace things like des_read_password and friends (backward
4317 compatibility functions using this new API are provided).
4318 The purpose is to remove prompting functions from the DES code
4319 section as well as provide for prompting through dialog boxes in
4320 a window system and the like.
4323 *) Add "ex_data" support to ENGINE so implementations can add state at a
4324 per-structure level rather than having to store it globally.
4327 *) Make it possible for ENGINE structures to be copied when retrieved by
4328 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4329 This causes the "original" ENGINE structure to act like a template,
4330 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4331 operational state can be localised to each ENGINE structure, despite the
4332 fact they all share the same "methods". New ENGINE structures returned in
4333 this case have no functional references and the return value is the single
4334 structural reference. This matches the single structural reference returned
4335 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4339 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4340 needs to match any other type at all we need to manually clear the
4344 *) Changes to the "openssl engine" utility to include;
4345 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4346 about an ENGINE's available control commands.
4347 - executing control commands from command line arguments using the
4348 '-pre' and '-post' switches. '-post' is only used if '-t' is
4349 specified and the ENGINE is successfully initialised. The syntax for
4350 the individual commands are colon-separated, for example;
4351 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4354 *) New dynamic control command support for ENGINEs. ENGINEs can now
4355 declare their own commands (numbers), names (strings), descriptions,
4356 and input types for run-time discovery by calling applications. A
4357 subset of these commands are implicitly classed as "executable"
4358 depending on their input type, and only these can be invoked through
4359 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4360 can be based on user input, config files, etc). The distinction is
4361 that "executable" commands cannot return anything other than a boolean
4362 result and can only support numeric or string input, whereas some
4363 discoverable commands may only be for direct use through
4364 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4365 pointers, or other custom uses. The "executable" commands are to
4366 support parameterisations of ENGINE behaviour that can be
4367 unambiguously defined by ENGINEs and used consistently across any
4368 OpenSSL-based application. Commands have been added to all the
4369 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4370 control over shared-library paths without source code alterations.
4373 *) Changed all ENGINE implementations to dynamically allocate their
4374 ENGINEs rather than declaring them statically. Apart from this being
4375 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4376 this also allows the implementations to compile without using the
4377 internal engine_int.h header.
4380 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4381 'const' value. Any code that should be able to modify a RAND_METHOD
4382 should already have non-const pointers to it (ie. they should only
4383 modify their own ones).
4386 *) Made a variety of little tweaks to the ENGINE code.
4387 - "atalla" and "ubsec" string definitions were moved from header files
4388 to C code. "nuron" string definitions were placed in variables
4389 rather than hard-coded - allowing parameterisation of these values
4390 later on via ctrl() commands.
4391 - Removed unused "#if 0"'d code.
4392 - Fixed engine list iteration code so it uses ENGINE_free() to release
4393 structural references.
4394 - Constified the RAND_METHOD element of ENGINE structures.
4395 - Constified various get/set functions as appropriate and added
4396 missing functions (including a catch-all ENGINE_cpy that duplicates
4397 all ENGINE values onto a new ENGINE except reference counts/state).
4398 - Removed NULL parameter checks in get/set functions. Setting a method
4399 or function to NULL is a way of cancelling out a previously set
4400 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4401 and doesn't justify the extra error symbols and code.
4402 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4403 flags from engine_int.h to engine.h.
4404 - Changed prototypes for ENGINE handler functions (init(), finish(),
4405 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4408 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4409 to the algorithm using long division. The binary algorithm can be
4410 used only if the modulus is odd. On 32-bit systems, it is faster
4411 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4412 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4413 up to 450 bits. In 64-bit environments, the binary algorithm
4414 appears to be advantageous for much longer moduli; here we use it
4415 for moduli up to 2048 bits.
4418 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4419 could not support the combine flag in choice fields.
4422 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4423 extensions from a certificate request to the certificate.
4426 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4427 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4428 file: this allows the display of the certificate about to be
4429 signed to be customised, to allow certain fields to be included
4430 or excluded and extension details. The old system didn't display
4431 multicharacter strings properly, omitted fields not in the policy
4432 and couldn't display additional details such as extensions.
4435 *) Function EC_POINTs_mul for multiple scalar multiplication
4436 of an arbitrary number of elliptic curve points
4437 \sum scalars[i]*points[i],
4438 optionally including the generator defined for the EC_GROUP:
4439 scalar*generator + \sum scalars[i]*points[i].
4441 EC_POINT_mul is a simple wrapper function for the typical case
4442 that the point list has just one item (besides the optional
4446 *) First EC_METHODs for curves over GF(p):
4448 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4449 operations and provides various method functions that can also
4450 operate with faster implementations of modular arithmetic.
4452 EC_GFp_mont_method() reuses most functions that are part of
4453 EC_GFp_simple_method, but uses Montgomery arithmetic.
4455 [Bodo Moeller; point addition and point doubling
4456 implementation directly derived from source code provided by
4457 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4459 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4460 crypto/ec/ec_lib.c):
4462 Curves are EC_GROUP objects (with an optional group generator)
4463 based on EC_METHODs that are built into the library.
4465 Points are EC_POINT objects based on EC_GROUP objects.
4467 Most of the framework would be able to handle curves over arbitrary
4468 finite fields, but as there are no obvious types for fields other
4469 than GF(p), some functions are limited to that for now.
4472 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4473 that the file contains a complete HTTP response.
4476 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4477 change the def and num file printf format specifier from "%-40sXXX"
4478 to "%-39s XXX". The latter will always guarantee a space after the
4479 field while the former will cause them to run together if the field
4480 is 40 of more characters long.
4483 *) Constify the cipher and digest 'method' functions and structures
4484 and modify related functions to take constant EVP_MD and EVP_CIPHER
4488 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4489 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4492 *) Modify EVP_Digest*() routines so they now return values. Although the
4493 internal software routines can never fail additional hardware versions
4497 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4499 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4500 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4503 ERR_R_NESTED_ASN1_ERROR
4505 ERR_R_MISSING_ASN1_EOS
4506 were 4 .. 9, conflicting with
4507 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4509 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4510 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4512 Add new error code 'ERR_R_INTERNAL_ERROR'.
4515 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4519 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4520 sets the subject name for a new request or supersedes the
4521 subject name in a given request. Formats that can be parsed are
4522 'CN=Some Name, OU=myOU, C=IT'
4524 'CN=Some Name/OU=myOU/C=IT'.
4526 Add options '-batch' and '-verbose' to 'openssl req'.
4527 [Massimiliano Pala <madwolf@hackmasters.net>]
4529 *) Introduce the possibility to access global variables through
4530 functions on platform were that's the best way to handle exporting
4531 global variables in shared libraries. To enable this functionality,
4532 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4533 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4534 is normally done by Configure or something similar).
4536 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4537 in the source file (foo.c) like this:
4539 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4540 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4542 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4543 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4545 OPENSSL_DECLARE_GLOBAL(int,foo);
4546 #define foo OPENSSL_GLOBAL_REF(foo)
4547 OPENSSL_DECLARE_GLOBAL(double,bar);
4548 #define bar OPENSSL_GLOBAL_REF(bar)
4550 The #defines are very important, and therefore so is including the
4551 header file everywhere where the defined globals are used.
4553 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4554 of ASN.1 items, but that structure is a bit different.
4556 The largest change is in util/mkdef.pl which has been enhanced with
4557 better and easier to understand logic to choose which symbols should
4558 go into the Windows .def files as well as a number of fixes and code
4559 cleanup (among others, algorithm keywords are now sorted
4560 lexicographically to avoid constant rewrites).
4563 *) In BN_div() keep a copy of the sign of 'num' before writing the
4564 result to 'rm' because if rm==num the value will be overwritten
4565 and produce the wrong result if 'num' is negative: this caused
4566 problems with BN_mod() and BN_nnmod().
4569 *) Function OCSP_request_verify(). This checks the signature on an
4570 OCSP request and verifies the signer certificate. The signer
4571 certificate is just checked for a generic purpose and OCSP request
4575 *) Add OCSP_check_validity() function to check the validity of OCSP
4576 responses. OCSP responses are prepared in real time and may only
4577 be a few seconds old. Simply checking that the current time lies
4578 between thisUpdate and nextUpdate max reject otherwise valid responses
4579 caused by either OCSP responder or client clock inaccuracy. Instead
4580 we allow thisUpdate and nextUpdate to fall within a certain period of
4581 the current time. The age of the response can also optionally be
4582 checked. Two new options -validity_period and -status_age added to
4586 *) If signature or public key algorithm is unrecognized print out its
4587 OID rather that just UNKNOWN.
4590 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4591 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4592 ID to be generated from the issuer certificate alone which can then be
4593 passed to OCSP_id_issuer_cmp().
4596 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4597 ASN1 modules to export functions returning ASN1_ITEM pointers
4598 instead of the ASN1_ITEM structures themselves. This adds several
4599 new macros which allow the underlying ASN1 function/structure to
4600 be accessed transparently. As a result code should not use ASN1_ITEM
4601 references directly (such as &X509_it) but instead use the relevant
4602 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4603 use of the new ASN1 code on platforms where exporting structures
4604 is problematical (for example in shared libraries) but exporting
4605 functions returning pointers to structures is not.
4608 *) Add support for overriding the generation of SSL/TLS session IDs.
4609 These callbacks can be registered either in an SSL_CTX or per SSL.
4610 The purpose of this is to allow applications to control, if they wish,
4611 the arbitrary values chosen for use as session IDs, particularly as it
4612 can be useful for session caching in multiple-server environments. A
4613 command-line switch for testing this (and any client code that wishes
4614 to use such a feature) has been added to "s_server".
4615 [Geoff Thorpe, Lutz Jaenicke]
4617 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4618 of the form '#if defined(...) || defined(...) || ...' and
4619 '#if !defined(...) && !defined(...) && ...'. This also avoids
4620 the growing number of special cases it was previously handling.
4623 *) Make all configuration macros available for application by making
4624 sure they are available in opensslconf.h, by giving them names starting
4625 with "OPENSSL_" to avoid conflicts with other packages and by making
4626 sure e_os2.h will cover all platform-specific cases together with
4628 Additionally, it is now possible to define configuration/platform-
4629 specific names (called "system identities"). In the C code, these
4630 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4631 macro with the name beginning with "OPENSSL_SYS_", which is determined
4632 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4636 *) New option -set_serial to 'req' and 'x509' this allows the serial
4637 number to use to be specified on the command line. Previously self
4638 signed certificates were hard coded with serial number 0 and the
4639 CA options of 'x509' had to use a serial number in a file which was
4643 *) New options to 'ca' utility to support V2 CRL entry extensions.
4644 Currently CRL reason, invalidity date and hold instruction are
4645 supported. Add new CRL extensions to V3 code and some new objects.
4648 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4649 disable standard block padding (aka PKCS#5 padding) in the EVP
4650 API, which was previously mandatory. This means that the data is
4651 not padded in any way and so the total length much be a multiple
4652 of the block size, otherwise an error occurs.
4655 *) Initial (incomplete) OCSP SSL support.
4658 *) New function OCSP_parse_url(). This splits up a URL into its host,
4659 port and path components: primarily to parse OCSP URLs. New -url
4660 option to ocsp utility.
4663 *) New nonce behavior. The return value of OCSP_check_nonce() now
4664 reflects the various checks performed. Applications can decide
4665 whether to tolerate certain situations such as an absent nonce
4666 in a response when one was present in a request: the ocsp application
4667 just prints out a warning. New function OCSP_add1_basic_nonce()
4668 this is to allow responders to include a nonce in a response even if
4669 the request is nonce-less.
4672 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4673 skipped when using openssl x509 multiple times on a single input file,
4674 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4677 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4678 set string type: to handle setting ASN1_TIME structures. Fix ca
4679 utility to correctly initialize revocation date of CRLs.
4682 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4683 the clients preferred ciphersuites and rather use its own preferences.
4684 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4685 Internet Explorer by ensuring unchanged hash method during stepup.
4686 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4689 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4690 to aes and add a new 'exist' option to print out symbols that don't
4694 *) Additional options to ocsp utility to allow flags to be set and
4695 additional certificates supplied.
4698 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4699 OCSP client a number of certificate to only verify the response
4703 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4704 handle the new API. Currently only ECB, CBC modes supported. Add new
4707 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4708 Encryption Standard (AES) Ciphersuites for Transport Layer
4709 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4710 not enabled by default and were not part of the "ALL" ciphersuite
4711 alias because they were not yet official; they could be
4712 explicitly requested by specifying the "AESdraft" ciphersuite
4713 group alias. In the final release of OpenSSL 0.9.7, the group
4714 alias is called "AES" and is part of "ALL".)
4715 [Ben Laurie, Steve Henson, Bodo Moeller]
4717 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4718 request to response.
4721 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4722 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4723 extract information from a certificate request. OCSP_response_create()
4724 creates a response and optionally adds a basic response structure.
4725 OCSP_basic_add1_status() adds a complete single response to a basic
4726 response and returns the OCSP_SINGLERESP structure just added (to allow
4727 extensions to be included for example). OCSP_basic_add1_cert() adds a
4728 certificate to a basic response and OCSP_basic_sign() signs a basic
4729 response with various flags. New helper functions ASN1_TIME_check()
4730 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4731 (converts ASN1_TIME to GeneralizedTime).
4734 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4735 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4736 structure from a certificate. X509_pubkey_digest() digests the public_key
4737 contents: this is used in various key identifiers.
4740 *) Make sk_sort() tolerate a NULL argument.
4741 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4743 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4744 passed by the function are trusted implicitly. If any of them signed the
4745 response then it is assumed to be valid and is not verified.
4748 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4749 to data. This was previously part of the PKCS7 ASN1 code. This
4750 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4751 [Steve Henson, reported by Kenneth R. Robinette
4752 <support@securenetterm.com>]
4754 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4755 routines: without these tracing memory leaks is very painful.
4756 Fix leaks in PKCS12 and PKCS7 routines.
4759 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4760 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4761 effectively meant GeneralizedTime would never be used. Now it
4762 is initialised to -1 but X509_time_adj() now has to check the value
4763 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4764 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4765 [Steve Henson, reported by Kenneth R. Robinette
4766 <support@securenetterm.com>]
4768 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4769 result in a zero length in the ASN1_INTEGER structure which was
4770 not consistent with the structure when d2i_ASN1_INTEGER() was used
4771 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4772 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4773 where it did not print out a minus for negative ASN1_INTEGER.
4776 *) Add summary printout to ocsp utility. The various functions which
4777 convert status values to strings have been renamed to:
4778 OCSP_response_status_str(), OCSP_cert_status_str() and
4779 OCSP_crl_reason_str() and are no longer static. New options
4780 to verify nonce values and to disable verification. OCSP response
4781 printout format cleaned up.
4784 *) Add additional OCSP certificate checks. These are those specified
4785 in RFC2560. This consists of two separate checks: the CA of the
4786 certificate being checked must either be the OCSP signer certificate
4787 or the issuer of the OCSP signer certificate. In the latter case the
4788 OCSP signer certificate must contain the OCSP signing extended key
4789 usage. This check is performed by attempting to match the OCSP
4790 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4791 in the OCSP_CERTID structures of the response.
4794 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4795 and related routines. This uses the standard OpenSSL certificate
4796 verify routines to perform initial checks (just CA validity) and
4797 to obtain the certificate chain. Then additional checks will be
4798 performed on the chain. Currently the root CA is checked to see
4799 if it is explicitly trusted for OCSP signing. This is used to set
4800 a root CA as a global signing root: that is any certificate that
4801 chains to that CA is an acceptable OCSP signing certificate.
4804 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4805 extensions from a separate configuration file.
4806 As when reading extensions from the main configuration file,
4807 the '-extensions ...' option may be used for specifying the
4809 [Massimiliano Pala <madwolf@comune.modena.it>]
4811 *) New OCSP utility. Allows OCSP requests to be generated or
4812 read. The request can be sent to a responder and the output
4813 parsed, outputed or printed in text form. Not complete yet:
4814 still needs to check the OCSP response validity.
4817 *) New subcommands for 'openssl ca':
4818 'openssl ca -status <serial>' prints the status of the cert with
4819 the given serial number (according to the index file).
4820 'openssl ca -updatedb' updates the expiry status of certificates
4822 [Massimiliano Pala <madwolf@comune.modena.it>]
4824 *) New '-newreq-nodes' command option to CA.pl. This is like
4825 '-newreq', but calls 'openssl req' with the '-nodes' option
4826 so that the resulting key is not encrypted.
4827 [Damien Miller <djm@mindrot.org>]
4829 *) New configuration for the GNU Hurd.
4830 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4832 *) Initial code to implement OCSP basic response verify. This
4833 is currently incomplete. Currently just finds the signer's
4834 certificate and verifies the signature on the response.
4837 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4838 value of OPENSSLDIR. This is available via the new '-d' option
4839 to 'openssl version', and is also included in 'openssl version -a'.
4842 *) Allowing defining memory allocation callbacks that will be given
4843 file name and line number information in additional arguments
4844 (a const char* and an int). The basic functionality remains, as
4845 well as the original possibility to just replace malloc(),
4846 realloc() and free() by functions that do not know about these
4847 additional arguments. To register and find out the current
4848 settings for extended allocation functions, the following
4849 functions are provided:
4851 CRYPTO_set_mem_ex_functions
4852 CRYPTO_set_locked_mem_ex_functions
4853 CRYPTO_get_mem_ex_functions
4854 CRYPTO_get_locked_mem_ex_functions
4856 These work the same way as CRYPTO_set_mem_functions and friends.
4857 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4858 extended allocation function is enabled.
4859 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4860 a conventional allocation function is enabled.
4861 [Richard Levitte, Bodo Moeller]
4863 *) Finish off removing the remaining LHASH function pointer casts.
4864 There should no longer be any prototype-casting required when using
4865 the LHASH abstraction, and any casts that remain are "bugs". See
4866 the callback types and macros at the head of lhash.h for details
4867 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4870 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4871 If /dev/[u]random devices are not available or do not return enough
4872 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4874 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4875 /etc/entropy will be queried once each in this sequence, quering stops
4876 when enough entropy was collected without querying more sockets.
4879 *) Change the Unix RAND_poll() variant to be able to poll several
4880 random devices, as specified by DEVRANDOM, until a sufficient amount
4881 of data has been collected. We spend at most 10 ms on each file
4882 (select timeout) and read in non-blocking mode. DEVRANDOM now
4883 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
4884 (previously it was just the string "/dev/urandom"), so on typical
4885 platforms the 10 ms delay will never occur.
4886 Also separate out the Unix variant to its own file, rand_unix.c.
4887 For VMS, there's a currently-empty rand_vms.c.
4890 *) Move OCSP client related routines to ocsp_cl.c. These
4891 provide utility functions which an application needing
4892 to issue a request to an OCSP responder and analyse the
4893 response will typically need: as opposed to those which an
4894 OCSP responder itself would need which will be added later.
4896 OCSP_request_sign() signs an OCSP request with an API similar
4897 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
4898 response. OCSP_response_get1_basic() extracts basic response
4899 from response. OCSP_resp_find_status(): finds and extracts status
4900 information from an OCSP_CERTID structure (which will be created
4901 when the request structure is built). These are built from lower
4902 level functions which work on OCSP_SINGLERESP structures but
4903 wont normally be used unless the application wishes to examine
4904 extensions in the OCSP response for example.
4906 Replace nonce routines with a pair of functions.
4907 OCSP_request_add1_nonce() adds a nonce value and optionally
4908 generates a random value. OCSP_check_nonce() checks the
4909 validity of the nonce in an OCSP response.
4912 *) Change function OCSP_request_add() to OCSP_request_add0_id().
4913 This doesn't copy the supplied OCSP_CERTID and avoids the
4914 need to free up the newly created id. Change return type
4915 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
4916 This can then be used to add extensions to the request.
4917 Deleted OCSP_request_new(), since most of its functionality
4918 is now in OCSP_REQUEST_new() (and the case insensitive name
4919 clash) apart from the ability to set the request name which
4920 will be added elsewhere.
4923 *) Update OCSP API. Remove obsolete extensions argument from
4924 various functions. Extensions are now handled using the new
4925 OCSP extension code. New simple OCSP HTTP function which
4926 can be used to send requests and parse the response.
4929 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
4930 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
4931 uses the special reorder version of SET OF to sort the attributes
4932 and reorder them to match the encoded order. This resolves a long
4933 standing problem: a verify on a PKCS7 structure just after signing
4934 it used to fail because the attribute order did not match the
4935 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
4936 it uses the received order. This is necessary to tolerate some broken
4937 software that does not order SET OF. This is handled by encoding
4938 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
4939 to produce the required SET OF.
4942 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
4943 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
4944 files to get correct declarations of the ASN.1 item variables.
4947 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
4948 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
4949 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
4950 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
4951 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
4952 ASN1_ITEM and no wrapper functions.
4955 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
4956 replace the old function pointer based I/O routines. Change most of
4957 the *_d2i_bio() and *_d2i_fp() functions to use these.
4960 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
4961 lines, recognice more "algorithms" that can be deselected, and make
4962 it complain about algorithm deselection that isn't recognised.
4965 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
4966 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
4967 to use new functions. Add NO_ASN1_OLD which can be set to remove
4968 some old style ASN1 functions: this can be used to determine if old
4969 code will still work when these eventually go away.
4972 *) New extension functions for OCSP structures, these follow the
4973 same conventions as certificates and CRLs.
4976 *) New function X509V3_add1_i2d(). This automatically encodes and
4977 adds an extension. Its behaviour can be customised with various
4978 flags to append, replace or delete. Various wrappers added for
4979 certifcates and CRLs.
4982 *) Fix to avoid calling the underlying ASN1 print routine when
4983 an extension cannot be parsed. Correct a typo in the
4984 OCSP_SERVICELOC extension. Tidy up print OCSP format.
4987 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
4988 entries for variables.
4991 *) Add functionality to apps/openssl.c for detecting locking
4992 problems: As the program is single-threaded, all we have
4993 to do is register a locking callback using an array for
4994 storing which locks are currently held by the program.
4997 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
4998 SSL_get_ex_data_X509_STORE_idx(), which is used in
4999 ssl_verify_cert_chain() and thus can be called at any time
5000 during TLS/SSL handshakes so that thread-safety is essential.
5001 Unfortunately, the ex_data design is not at all suited
5002 for multi-threaded use, so it probably should be abolished.
5005 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5006 [Broadcom, tweaked and integrated by Geoff Thorpe]
5008 *) Move common extension printing code to new function
5009 X509V3_print_extensions(). Reorganise OCSP print routines and
5010 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5013 *) New function X509_signature_print() to remove duplication in some
5017 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5018 set (this was treated exactly the same as SET OF previously). This
5019 is used to reorder the STACK representing the structure to match the
5020 encoding. This will be used to get round a problem where a PKCS7
5021 structure which was signed could not be verified because the STACK
5022 order did not reflect the encoded order.
5025 *) Reimplement the OCSP ASN1 module using the new code.
5028 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5029 for its ASN1 operations. The old style function pointers still exist
5030 for now but they will eventually go away.
5033 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5034 completely replaces the old ASN1 functionality with a table driven
5035 encoder and decoder which interprets an ASN1_ITEM structure describing
5036 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5037 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5038 has also been converted to the new form.
5041 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5042 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5043 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5044 for negative moduli.
5047 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5048 of not touching the result's sign bit.
5051 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5055 *) Changed the LHASH code to use prototypes for callbacks, and created
5056 macros to declare and implement thin (optionally static) functions
5057 that provide type-safety and avoid function pointer casting for the
5058 type-specific callbacks.
5061 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5063 [Veers Staats <staatsvr@asc.hpc.mil>,
5064 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5066 *) Reformat the FAQ so the different questions and answers can be divided
5067 in sections depending on the subject.
5070 *) Have the zlib compression code load ZLIB.DLL dynamically under
5074 *) New function BN_mod_sqrt for computing square roots modulo a prime
5075 (using the probabilistic Tonelli-Shanks algorithm unless
5076 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5077 be handled deterministically).
5078 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5080 *) Make BN_mod_inverse faster by explicitly handling small quotients
5081 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5082 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5085 *) New function BN_kronecker.
5088 *) Fix BN_gcd so that it works on negative inputs; the result is
5089 positive unless both parameters are zero.
5090 Previously something reasonably close to an infinite loop was
5091 possible because numbers could be growing instead of shrinking
5092 in the implementation of Euclid's algorithm.
5095 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5096 sign of the number in question.
5098 Fix BN_is_word(a,w) to work correctly for w == 0.
5100 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5101 because its test if the absolute value of 'a' equals 'w'.
5102 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5103 it exists mostly for use in the implementations of BN_is_zero(),
5104 BN_is_one(), and BN_is_word().
5107 *) New function BN_swap.
5110 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5111 the exponentiation functions are more likely to produce reasonable
5112 results on negative inputs.
5115 *) Change BN_mod_mul so that the result is always non-negative.
5116 Previously, it could be negative if one of the factors was negative;
5117 I don't think anyone really wanted that behaviour.
5120 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5121 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5122 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5123 and add new functions:
5132 BN_mod_lshift1_quick
5136 These functions always generate non-negative results.
5138 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5139 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5141 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5142 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5143 be reduced modulo m.
5144 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5147 The following entry accidentily appeared in the CHANGES file
5148 distributed with OpenSSL 0.9.7. The modifications described in
5149 it do *not* apply to OpenSSL 0.9.7.
5151 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5152 was actually never needed) and in BN_mul(). The removal in BN_mul()
5153 required a small change in bn_mul_part_recursive() and the addition
5154 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5155 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5156 bn_sub_words() and bn_add_words() except they take arrays with
5161 *) In 'openssl passwd', verify passwords read from the terminal
5162 unless the '-salt' option is used (which usually means that
5163 verification would just waste user's time since the resulting
5164 hash is going to be compared with some given password hash)
5165 or the new '-noverify' option is used.
5167 This is an incompatible change, but it does not affect
5168 non-interactive use of 'openssl passwd' (passwords on the command
5169 line, '-stdin' option, '-in ...' option) and thus should not
5173 *) Remove all references to RSAref, since there's no more need for it.
5176 *) Make DSO load along a path given through an environment variable
5177 (SHLIB_PATH) with shl_load().
5180 *) Constify the ENGINE code as a result of BIGNUM constification.
5181 Also constify the RSA code and most things related to it. In a
5182 few places, most notable in the depth of the ASN.1 code, ugly
5183 casts back to non-const were required (to be solved at a later
5187 *) Make it so the openssl application has all engines loaded by default.
5190 *) Constify the BIGNUM routines a little more.
5193 *) Add the following functions:
5195 ENGINE_load_cswift()
5197 ENGINE_load_atalla()
5199 ENGINE_load_builtin_engines()
5201 That way, an application can itself choose if external engines that
5202 are built-in in OpenSSL shall ever be used or not. The benefit is
5203 that applications won't have to be linked with libdl or other dso
5204 libraries unless it's really needed.
5206 Changed 'openssl engine' to load all engines on demand.
5207 Changed the engine header files to avoid the duplication of some
5208 declarations (they differed!).
5211 *) 'openssl engine' can now list capabilities.
5214 *) Better error reporting in 'openssl engine'.
5217 *) Never call load_dh_param(NULL) in s_server.
5220 *) Add engine application. It can currently list engines by name and
5221 identity, and test if they are actually available.
5224 *) Improve RPM specification file by forcing symbolic linking and making
5225 sure the installed documentation is also owned by root.root.
5226 [Damien Miller <djm@mindrot.org>]
5228 *) Give the OpenSSL applications more possibilities to make use of
5229 keys (public as well as private) handled by engines.
5232 *) Add OCSP code that comes from CertCo.
5235 *) Add VMS support for the Rijndael code.
5238 *) Added untested support for Nuron crypto accelerator.
5241 *) Add support for external cryptographic devices. This code was
5242 previously distributed separately as the "engine" branch.
5243 [Geoff Thorpe, Richard Levitte]
5245 *) Rework the filename-translation in the DSO code. It is now possible to
5246 have far greater control over how a "name" is turned into a filename
5247 depending on the operating environment and any oddities about the
5248 different shared library filenames on each system.
5251 *) Support threads on FreeBSD-elf in Configure.
5254 *) Fix for SHA1 assembly problem with MASM: it produces
5255 warnings about corrupt line number information when assembling
5256 with debugging information. This is caused by the overlapping
5258 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5261 NCONF_get_number() has no error checking at all. As a replacement,
5262 NCONF_get_number_e() is defined (_e for "error checking") and is
5263 promoted strongly. The old NCONF_get_number is kept around for
5264 binary backward compatibility.
5265 Make it possible for methods to load from something other than a BIO,
5266 by providing a function pointer that is given a name instead of a BIO.
5267 For example, this could be used to load configuration data from an
5271 *) Fix for non blocking accept BIOs. Added new I/O special reason
5272 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5273 with non blocking I/O was not possible because no retry code was
5274 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5278 *) Added the beginnings of Rijndael support.
5281 *) Fix for bug in DirectoryString mask setting. Add support for
5282 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5283 to allow certificate printing to more controllable, additional
5284 'certopt' option to 'x509' to allow new printing options to be
5288 *) Clean old EAY MD5 hack from e_os.h.
5291 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5293 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5294 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5295 [Joe Orton, Steve Henson]
5297 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5299 *) Fix additional bug revealed by the NISCC test suite:
5301 Stop bug triggering large recursion when presented with
5302 certain ASN.1 tags (CVE-2003-0851)
5305 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5307 *) Fix various bugs revealed by running the NISCC test suite:
5309 Stop out of bounds reads in the ASN1 code when presented with
5310 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5312 If verify callback ignores invalid public key errors don't try to check
5313 certificate signature with the NULL public key.
5317 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5318 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5322 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5323 extra data after the compression methods not only for TLS 1.0
5324 but also for SSL 3.0 (as required by the specification).
5325 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5327 *) Change X509_certificate_type() to mark the key as exported/exportable
5328 when it's 512 *bits* long, not 512 bytes.
5331 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5333 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5334 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5335 a protocol version number mismatch like a decryption error
5336 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5339 *) Turn on RSA blinding by default in the default implementation
5340 to avoid a timing attack. Applications that don't want it can call
5341 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5342 They would be ill-advised to do so in most cases.
5343 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5345 *) Change RSA blinding code so that it works when the PRNG is not
5346 seeded (in this case, the secret RSA exponent is abused as
5347 an unpredictable seed -- if it is not unpredictable, there
5348 is no point in blinding anyway). Make RSA blinding thread-safe
5349 by remembering the creator's thread ID in rsa->blinding and
5350 having all other threads use local one-time blinding factors
5351 (this requires more computation than sharing rsa->blinding, but
5352 avoids excessive locking; and if an RSA object is not shared
5353 between threads, blinding will still be very fast).
5356 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5358 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5359 via timing by performing a MAC computation even if incorrrect
5360 block cipher padding has been found. This is a countermeasure
5361 against active attacks where the attacker has to distinguish
5362 between bad padding and a MAC verification error. (CVE-2003-0078)
5364 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5365 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5366 Martin Vuagnoux (EPFL, Ilion)]
5368 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5370 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5371 memory from it's contents. This is done with a counter that will
5372 place alternating values in each byte. This can be used to solve
5373 two issues: 1) the removal of calls to memset() by highly optimizing
5374 compilers, and 2) cleansing with other values than 0, since those can
5375 be read through on certain media, for example a swap space on disk.
5378 *) Bugfix: client side session caching did not work with external caching,
5379 because the session->cipher setting was not restored when reloading
5380 from the external cache. This problem was masked, when
5381 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5382 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5385 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5386 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5387 [Zeev Lieber <zeev-l@yahoo.com>]
5389 *) Undo an undocumented change introduced in 0.9.6e which caused
5390 repeated calls to OpenSSL_add_all_ciphers() and
5391 OpenSSL_add_all_digests() to be ignored, even after calling
5395 *) Change the default configuration reader to deal with last line not
5396 being properly terminated.
5399 *) Change X509_NAME_cmp() so it applies the special rules on handling
5400 DN values that are of type PrintableString, as well as RDNs of type
5401 emailAddress where the value has the type ia5String.
5402 [stefank@valicert.com via Richard Levitte]
5404 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5405 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5406 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5407 the bitwise-OR of the two for use by the majority of applications
5408 wanting this behaviour, and update the docs. The documented
5409 behaviour and actual behaviour were inconsistent and had been
5410 changing anyway, so this is more a bug-fix than a behavioural
5412 [Geoff Thorpe, diagnosed by Nadav Har'El]
5414 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5415 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5418 *) Fix initialization code race conditions in
5419 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5420 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5421 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5422 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5423 ssl2_get_cipher_by_char(),
5424 ssl3_get_cipher_by_char().
5425 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5427 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5428 the cached sessions are flushed, as the remove_cb() might use ex_data
5429 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5430 (see [openssl.org #212]).
5431 [Geoff Thorpe, Lutz Jaenicke]
5433 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5434 length, instead of the encoding length to d2i_ASN1_OBJECT.
5437 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5439 *) [In 0.9.6g-engine release:]
5440 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5441 [Lynn Gazis <lgazis@rainbow.com>]
5443 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5445 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5446 and get fix the header length calculation.
5447 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5448 Alon Kantor <alonk@checkpoint.com> (and others),
5451 *) Use proper error handling instead of 'assertions' in buffer
5452 overflow checks added in 0.9.6e. This prevents DoS (the
5453 assertions could call abort()).
5454 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5456 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5458 *) Add various sanity checks to asn1_get_length() to reject
5459 the ASN1 length bytes if they exceed sizeof(long), will appear
5460 negative or the content length exceeds the length of the
5462 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5464 *) Fix cipher selection routines: ciphers without encryption had no flags
5465 for the cipher strength set and where therefore not handled correctly
5466 by the selection routines (PR #130).
5469 *) Fix EVP_dsa_sha macro.
5473 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5474 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5475 that was added in OpenSSL 0.9.6d.
5477 As the countermeasure turned out to be incompatible with some
5478 broken SSL implementations, the new option is part of SSL_OP_ALL.
5479 SSL_OP_ALL is usually employed when compatibility with weird SSL
5480 implementations is desired (e.g. '-bugs' option to 's_client' and
5481 's_server'), so the new option is automatically set in many
5485 *) Changes in security patch:
5487 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5488 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5489 Air Force Materiel Command, USAF, under agreement number
5492 *) Add various sanity checks to asn1_get_length() to reject
5493 the ASN1 length bytes if they exceed sizeof(long), will appear
5494 negative or the content length exceeds the length of the
5495 supplied buffer. (CVE-2002-0659)
5496 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5498 *) Assertions for various potential buffer overflows, not known to
5500 [Ben Laurie (CHATS)]
5502 *) Various temporary buffers to hold ASCII versions of integers were
5503 too small for 64 bit platforms. (CVE-2002-0655)
5504 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5506 *) Remote buffer overflow in SSL3 protocol - an attacker could
5507 supply an oversized session ID to a client. (CVE-2002-0656)
5508 [Ben Laurie (CHATS)]
5510 *) Remote buffer overflow in SSL2 protocol - an attacker could
5511 supply an oversized client master key. (CVE-2002-0656)
5512 [Ben Laurie (CHATS)]
5514 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5516 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5517 encoded as NULL) with id-dsa-with-sha1.
5518 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5520 *) Check various X509_...() return values in apps/req.c.
5521 [Nils Larsch <nla@trustcenter.de>]
5523 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5524 an end-of-file condition would erronously be flagged, when the CRLF
5525 was just at the end of a processed block. The bug was discovered when
5526 processing data through a buffering memory BIO handing the data to a
5527 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5528 <ptsekov@syntrex.com> and Nedelcho Stanev.
5531 *) Implement a countermeasure against a vulnerability recently found
5532 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5533 before application data chunks to avoid the use of known IVs
5534 with data potentially chosen by the attacker.
5537 *) Fix length checks in ssl3_get_client_hello().
5540 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5541 to prevent ssl3_read_internal() from incorrectly assuming that
5542 ssl3_read_bytes() found application data while handshake
5543 processing was enabled when in fact s->s3->in_read_app_data was
5544 merely automatically cleared during the initial handshake.
5545 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5547 *) Fix object definitions for Private and Enterprise: they were not
5548 recognized in their shortname (=lowercase) representation. Extend
5549 obj_dat.pl to issue an error when using undefined keywords instead
5550 of silently ignoring the problem (Svenning Sorensen
5551 <sss@sss.dnsalias.net>).
5554 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5555 generators, i.e. generators other than 2 and 5. (Previously, the
5556 code did not properly initialise the 'add' and 'rem' values to
5557 BN_generate_prime().)
5559 In the new general case, we do not insist that 'generator' is
5560 actually a primitive root: This requirement is rather pointless;
5561 a generator of the order-q subgroup is just as good, if not
5565 *) Map new X509 verification errors to alerts. Discovered and submitted by
5566 Tom Wu <tom@arcot.com>.
5569 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5570 returning non-zero before the data has been completely received
5571 when using non-blocking I/O.
5572 [Bodo Moeller; problem pointed out by John Hughes]
5574 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5575 [Ben Laurie, Lutz Jaenicke]
5577 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5578 Yoram Zahavi <YoramZ@gilian.com>).
5581 *) Add information about CygWin 1.3 and on, and preserve proper
5582 configuration for the versions before that.
5583 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5585 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5586 check whether we deal with a copy of a session and do not delete from
5587 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5588 <izhar@checkpoint.com>.
5591 *) Do not store session data into the internal session cache, if it
5592 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5593 flag is set). Proposed by Aslam <aslam@funk.com>.
5596 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5600 *) [In 0.9.6d-engine release:]
5601 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5602 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5604 *) Add the configuration target linux-s390x.
5605 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5607 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5608 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5609 variable as an indication that a ClientHello message has been
5610 received. As the flag value will be lost between multiple
5611 invocations of ssl3_accept when using non-blocking I/O, the
5612 function may not be aware that a handshake has actually taken
5613 place, thus preventing a new session from being added to the
5616 To avoid this problem, we now set s->new_session to 2 instead of
5617 using a local variable.
5618 [Lutz Jaenicke, Bodo Moeller]
5620 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5621 if the SSL_R_LENGTH_MISMATCH error is detected.
5622 [Geoff Thorpe, Bodo Moeller]
5624 *) New 'shared_ldflag' column in Configure platform table.
5627 *) Fix EVP_CIPHER_mode macro.
5628 ["Dan S. Camper" <dan@bti.net>]
5630 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5631 type, we must throw them away by setting rr->length to 0.
5632 [D P Chang <dpc@qualys.com>]
5634 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5636 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5637 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5638 worked incorrectly for those cases where range = 10..._2 and
5639 3*range is two bits longer than range.)
5642 *) Only add signing time to PKCS7 structures if it is not already
5646 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5647 OBJ_ld_ce should be OBJ_id_ce.
5648 Also some ip-pda OIDs in crypto/objects/objects.txt were
5649 incorrect (cf. RFC 3039).
5650 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5652 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5653 returns early because it has nothing to do.
5654 [Andy Schneider <andy.schneider@bjss.co.uk>]
5656 *) [In 0.9.6c-engine release:]
5657 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5658 [Andy Schneider <andy.schneider@bjss.co.uk>]
5660 *) [In 0.9.6c-engine release:]
5661 Add support for Cryptographic Appliance's keyserver technology.
5662 (Use engine 'keyclient')
5663 [Cryptographic Appliances and Geoff Thorpe]
5665 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5666 is called via tools/c89.sh because arguments have to be
5667 rearranged (all '-L' options must appear before the first object
5669 [Richard Shapiro <rshapiro@abinitio.com>]
5671 *) [In 0.9.6c-engine release:]
5672 Add support for Broadcom crypto accelerator cards, backported
5674 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5676 *) [In 0.9.6c-engine release:]
5677 Add support for SureWare crypto accelerator cards from
5678 Baltimore Technologies. (Use engine 'sureware')
5679 [Baltimore Technologies and Mark Cox]
5681 *) [In 0.9.6c-engine release:]
5682 Add support for crypto accelerator cards from Accelerated
5683 Encryption Processing, www.aep.ie. (Use engine 'aep')
5684 [AEP Inc. and Mark Cox]
5686 *) Add a configuration entry for gcc on UnixWare.
5687 [Gary Benson <gbenson@redhat.com>]
5689 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5690 messages are stored in a single piece (fixed-length part and
5691 variable-length part combined) and fix various bugs found on the way.
5694 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5695 instead. BIO_gethostbyname() does not know what timeouts are
5696 appropriate, so entries would stay in cache even when they have
5698 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5700 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5701 faced with a pathologically small ClientHello fragment that does
5702 not contain client_version: Instead of aborting with an error,
5703 simply choose the highest available protocol version (i.e.,
5704 TLS 1.0 unless it is disabled). In practice, ClientHello
5705 messages are never sent like this, but this change gives us
5706 strictly correct behaviour at least for TLS.
5709 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5710 never resets s->method to s->ctx->method when called from within
5711 one of the SSL handshake functions.
5712 [Bodo Moeller; problem pointed out by Niko Baric]
5714 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5715 (sent using the client's version number) if client_version is
5716 smaller than the protocol version in use. Also change
5717 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5718 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5719 the client will at least see that alert.
5722 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5726 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5727 client receives HelloRequest while in a handshake.
5728 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5730 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5731 should end in 'break', not 'goto end' which circuments various
5732 cleanups done in state SSL_ST_OK. But session related stuff
5733 must be disabled for SSL_ST_OK in the case that we just sent a
5736 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5737 before just sending a HelloRequest.
5738 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5740 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5741 reveal whether illegal block cipher padding was found or a MAC
5742 verification error occured. (Neither SSLerr() codes nor alerts
5743 are directly visible to potential attackers, but the information
5744 may leak via logfiles.)
5746 Similar changes are not required for the SSL 2.0 implementation
5747 because the number of padding bytes is sent in clear for SSL 2.0,
5748 and the extra bytes are just ignored. However ssl/s2_pkt.c
5749 failed to verify that the purported number of padding bytes is in
5753 *) Add OpenUNIX-8 support including shared libraries
5754 (Boyd Lynn Gerber <gerberb@zenez.com>).
5757 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5758 'wristwatch attack' using huge encoding parameters (cf.
5759 James H. Manger's CRYPTO 2001 paper). Note that the
5760 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5761 encoding parameters and hence was not vulnerable.
5764 *) BN_sqr() bug fix.
5765 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5767 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5768 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5769 followed by modular reduction.
5770 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5772 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5773 equivalent based on BN_pseudo_rand() instead of BN_rand().
5776 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5777 This function was broken, as the check for a new client hello message
5778 to handle SGC did not allow these large messages.
5779 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5782 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5785 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5786 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5789 *) Rework the configuration and shared library support for Tru64 Unix.
5790 The configuration part makes use of modern compiler features and
5791 still retains old compiler behavior for those that run older versions
5792 of the OS. The shared library support part includes a variant that
5793 uses the RPATH feature, and is available through the special
5794 configuration target "alpha-cc-rpath", which will never be selected
5796 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5798 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5799 with the same message size as in ssl3_get_certificate_request().
5800 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5801 messages might inadvertently be reject as too long.
5802 [Petr Lampa <lampa@fee.vutbr.cz>]
5804 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5807 *) Modified SSL library such that the verify_callback that has been set
5808 specificly for an SSL object with SSL_set_verify() is actually being
5809 used. Before the change, a verify_callback set with this function was
5810 ignored and the verify_callback() set in the SSL_CTX at the time of
5811 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5812 to allow the necessary settings.
5815 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5816 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5817 done automatically (in contradiction to the requirements of the C
5818 standard). This made problems when used from OpenSSH.
5821 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5822 dh->length and always used
5824 BN_rand_range(priv_key, dh->p).
5826 BN_rand_range() is not necessary for Diffie-Hellman, and this
5827 specific range makes Diffie-Hellman unnecessarily inefficient if
5828 dh->length (recommended exponent length) is much smaller than the
5829 length of dh->p. We could use BN_rand_range() if the order of
5830 the subgroup was stored in the DH structure, but we only have
5835 BN_rand(priv_key, l, ...)
5837 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5843 RSA_eay_public_encrypt
5844 RSA_eay_private_decrypt
5845 RSA_eay_private_encrypt (signing)
5846 RSA_eay_public_decrypt (signature verification)
5848 (default implementations for RSA_public_encrypt,
5849 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5850 always reject numbers >= n.
5853 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5854 to synchronize access to 'locking_thread'. This is necessary on
5855 systems where access to 'locking_thread' (an 'unsigned long'
5856 variable) is not atomic.
5859 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5860 *before* setting the 'crypto_lock_rand' flag. The previous code had
5861 a race condition if 0 is a valid thread ID.
5862 [Travis Vitek <vitek@roguewave.com>]
5864 *) Add support for shared libraries under Irix.
5865 [Albert Chin-A-Young <china@thewrittenword.com>]
5867 *) Add configuration option to build on Linux on both big-endian and
5869 [Ralf Baechle <ralf@uni-koblenz.de>]
5871 *) Add the possibility to create shared libraries on HP-UX.
5874 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5876 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
5877 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
5878 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
5879 PRNG state recovery was possible based on the output of
5880 one PRNG request appropriately sized to gain knowledge on
5881 'md' followed by enough consecutive 1-byte PRNG requests
5882 to traverse all of 'state'.
5884 1. When updating 'md_local' (the current thread's copy of 'md')
5885 during PRNG output generation, hash all of the previous
5886 'md_local' value, not just the half used for PRNG output.
5888 2. Make the number of bytes from 'state' included into the hash
5889 independent from the number of PRNG bytes requested.
5891 The first measure alone would be sufficient to avoid
5892 Markku-Juhani's attack. (Actually it had never occurred
5893 to me that the half of 'md_local' used for chaining was the
5894 half from which PRNG output bytes were taken -- I had always
5895 assumed that the secret half would be used.) The second
5896 measure makes sure that additional data from 'state' is never
5897 mixed into 'md_local' in small portions; this heuristically
5898 further strengthens the PRNG.
5901 *) Fix crypto/bn/asm/mips3.s.
5904 *) When only the key is given to "enc", the IV is undefined. Print out
5905 an error message in this case.
5908 *) Handle special case when X509_NAME is empty in X509 printing routines.
5911 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
5912 positive and less than q.
5915 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
5916 used: it isn't thread safe and the add_lock_callback should handle
5918 [Paul Rose <Paul.Rose@bridge.com>]
5920 *) Verify that incoming data obeys the block size in
5921 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
5925 [Ulf Möller, Bodo Möller]
5927 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
5928 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
5929 when fixing the server behaviour for backwards-compatible 'client
5930 hello' messages. (Note that the attack is impractical against
5931 SSL 3.0 and TLS 1.0 anyway because length and version checking
5932 means that the probability of guessing a valid ciphertext is
5933 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
5936 Before 0.9.5, the countermeasure (hide the error by generating a
5937 random 'decryption result') did not work properly because
5938 ERR_clear_error() was missing, meaning that SSL_get_error() would
5939 detect the supposedly ignored error.
5941 Both problems are now fixed.
5944 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
5945 (previously it was 1024).
5948 *) Fix for compatibility mode trust settings: ignore trust settings
5949 unless some valid trust or reject settings are present.
5952 *) Fix for blowfish EVP: its a variable length cipher.
5955 *) Fix various bugs related to DSA S/MIME verification. Handle missing
5956 parameters in DSA public key structures and return an error in the
5957 DSA routines if parameters are absent.
5960 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
5961 in the current directory if neither $RANDFILE nor $HOME was set.
5962 RAND_file_name() in 0.9.6a returned NULL in this case. This has
5963 caused some confusion to Windows users who haven't defined $HOME.
5964 Thus RAND_file_name() is changed again: e_os.h can define a
5965 DEFAULT_HOME, which will be used if $HOME is not set.
5966 For Windows, we use "C:"; on other platforms, we still require
5967 environment variables.
5969 *) Move 'if (!initialized) RAND_poll()' into regions protected by
5970 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
5971 having multiple threads call RAND_poll() concurrently.
5974 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
5975 combination of a flag and a thread ID variable.
5976 Otherwise while one thread is in ssleay_rand_bytes (which sets the
5977 flag), *other* threads can enter ssleay_add_bytes without obeying
5978 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
5979 that they do not hold after the first thread unsets add_do_not_lock).
5982 *) Change bctest again: '-x' expressions are not available in all
5986 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
5988 *) Fix a couple of memory leaks in PKCS7_dataDecode()
5989 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
5991 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
5992 the default extension for executables, if any. Also, make the perl
5993 scripts that use symlink() to test if it really exists and use "cp"
5994 if it doesn't. All this made OpenSSL compilable and installable in
5998 *) Fix for asn1_GetSequence() for indefinite length constructed data.
5999 If SEQUENCE is length is indefinite just set c->slen to the total
6000 amount of data available.
6001 [Steve Henson, reported by shige@FreeBSD.org]
6002 [This change does not apply to 0.9.7.]
6004 *) Change bctest to avoid here-documents inside command substitution
6005 (workaround for FreeBSD /bin/sh bug).
6006 For compatibility with Ultrix, avoid shell functions (introduced
6007 in the bctest version that searches along $PATH).
6010 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6011 with des_encrypt() defined on some operating systems, like Solaris
6015 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6016 On the Importance of Eliminating Errors in Cryptographic
6017 Computations, J. Cryptology 14 (2001) 2, 101-119,
6018 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6021 *) MIPS assembler BIGNUM division bug fix.
6024 *) Disabled incorrect Alpha assembler code.
6027 *) Fix PKCS#7 decode routines so they correctly update the length
6028 after reading an EOC for the EXPLICIT tag.
6030 [This change does not apply to 0.9.7.]
6032 *) Fix bug in PKCS#12 key generation routines. This was triggered
6033 if a 3DES key was generated with a 0 initial byte. Include
6034 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6035 (but broken) behaviour.
6038 *) Enhance bctest to search for a working bc along $PATH and print
6040 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6042 *) Fix memory leaks in err.c: free err_data string if necessary;
6043 don't write to the wrong index in ERR_set_error_data.
6046 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6050 *) Replace rdtsc with _emit statements for VC++ version 5.
6051 [Jeremy Cooper <jeremy@baymoo.org>]
6053 *) Make it possible to reuse SSLv2 sessions.
6056 *) In copy_email() check for >= 0 as a return value for
6057 X509_NAME_get_index_by_NID() since 0 is a valid index.
6058 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6060 *) Avoid coredump with unsupported or invalid public keys by checking if
6061 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6062 PKCS7_verify() fails with non detached data.
6065 *) Don't use getenv in library functions when run as setuid/setgid.
6066 New function OPENSSL_issetugid().
6069 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6070 due to incorrect handling of multi-threading:
6072 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6074 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6076 3. Count how many times MemCheck_off() has been called so that
6077 nested use can be treated correctly. This also avoids
6078 inband-signalling in the previous code (which relied on the
6079 assumption that thread ID 0 is impossible).
6082 *) Add "-rand" option also to s_client and s_server.
6085 *) Fix CPU detection on Irix 6.x.
6086 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6087 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6089 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6092 [This change does not apply to 0.9.7.]
6094 *) Use the cached encoding of an X509_NAME structure rather than
6095 copying it. This is apparently the reason for the libsafe "errors"
6096 but the code is actually correct.
6099 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6100 Bleichenbacher's DSA attack.
6101 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6102 to be set and top=0 forces the highest bit to be set; top=-1 is new
6103 and leaves the highest bit random.
6104 [Ulf Moeller, Bodo Moeller]
6106 *) In the NCONF_...-based implementations for CONF_... queries
6107 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6108 a temporary CONF structure with the data component set to NULL
6109 (which gives segmentation faults in lh_retrieve).
6110 Instead, use NULL for the CONF pointer in CONF_get_string and
6111 CONF_get_number (which may use environment variables) and directly
6112 return NULL from CONF_get_section.
6115 *) Fix potential buffer overrun for EBCDIC.
6118 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6119 keyUsage if basicConstraints absent for a CA.
6122 *) Make SMIME_write_PKCS7() write mail header values with a format that
6123 is more generally accepted (no spaces before the semicolon), since
6124 some programs can't parse those values properly otherwise. Also make
6125 sure BIO's that break lines after each write do not create invalid
6129 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6130 macros previously used would not encode an empty SEQUENCE OF
6131 and break the signature.
6133 [This change does not apply to 0.9.7.]
6135 *) Zero the premaster secret after deriving the master secret in
6139 *) Add some EVP_add_digest_alias registrations (as found in
6140 OpenSSL_add_all_digests()) to SSL_library_init()
6141 aka OpenSSL_add_ssl_algorithms(). This provides improved
6142 compatibility with peers using X.509 certificates
6143 with unconventional AlgorithmIdentifier OIDs.
6146 *) Fix for Irix with NO_ASM.
6147 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6149 *) ./config script fixes.
6150 [Ulf Moeller, Richard Levitte]
6152 *) Fix 'openssl passwd -1'.
6155 *) Change PKCS12_key_gen_asc() so it can cope with non null
6156 terminated strings whose length is passed in the passlen
6157 parameter, for example from PEM callbacks. This was done
6158 by adding an extra length parameter to asc2uni().
6159 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6161 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6162 call failed, free the DSA structure.
6165 *) Fix to uni2asc() to cope with zero length Unicode strings.
6166 These are present in some PKCS#12 files.
6169 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6170 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6171 when writing a 32767 byte record.
6172 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6174 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6175 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6177 (RSA objects have a reference count access to which is protected
6178 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6179 so they are meant to be shared between threads.)
6180 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6181 "Reddie, Steven" <Steven.Reddie@ca.com>]
6183 *) Fix a deadlock in CRYPTO_mem_leaks().
6186 *) Use better test patterns in bntest.
6189 *) rand_win.c fix for Borland C.
6192 *) BN_rshift bugfix for n == 0.
6195 *) Add a 'bctest' script that checks for some known 'bc' bugs
6196 so that 'make test' does not abort just because 'bc' is broken.
6199 *) Store verify_result within SSL_SESSION also for client side to
6200 avoid potential security hole. (Re-used sessions on the client side
6201 always resulted in verify_result==X509_V_OK, not using the original
6202 result of the server certificate verification.)
6205 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6206 SSL3_RT_APPLICATION_DATA, return 0.
6207 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6211 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6212 releases, have been re-implemented by renaming the previous
6213 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6214 and ssl3_read_internal, respectively, and adding 'peek' parameters
6215 to them. The new ssl[23]_{read,peek} functions are calls to
6216 ssl[23]_read_internal with the 'peek' flag set appropriately.
6217 A 'peek' parameter has also been added to ssl3_read_bytes, which
6218 does the actual work for ssl3_read_internal.
6221 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6222 the method-specific "init()" handler. Also clean up ex_data after
6223 calling the method-specific "finish()" handler. Previously, this was
6224 happening the other way round.
6227 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6228 The previous value, 12, was not always sufficient for BN_mod_exp().
6231 *) Make sure that shared libraries get the internal name engine with
6232 the full version number and not just 0. This should mark the
6233 shared libraries as not backward compatible. Of course, this should
6234 be changed again when we can guarantee backward binary compatibility.
6237 *) Fix typo in get_cert_by_subject() in by_dir.c
6238 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6240 *) Rework the system to generate shared libraries:
6242 - Make note of the expected extension for the shared libraries and
6243 if there is a need for symbolic links from for example libcrypto.so.0
6244 to libcrypto.so.0.9.7. There is extended info in Configure for
6247 - Make as few rebuilds of the shared libraries as possible.
6249 - Still avoid linking the OpenSSL programs with the shared libraries.
6251 - When installing, install the shared libraries separately from the
6255 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6257 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6258 and not in SSL_clear because the latter is also used by the
6259 accept/connect functions; previously, the settings made by
6260 SSL_set_read_ahead would be lost during the handshake.
6261 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6263 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6264 Previously, it would create entries for disableed algorithms no
6268 *) Added several new manual pages for SSL_* function.
6271 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6273 *) In ssl23_get_client_hello, generate an error message when faced
6274 with an initial SSL 3.0/TLS record that is too small to contain the
6275 first two bytes of the ClientHello message, i.e. client_version.
6276 (Note that this is a pathologic case that probably has never happened
6277 in real life.) The previous approach was to use the version number
6278 from the record header as a substitute; but our protocol choice
6279 should not depend on that one because it is not authenticated
6280 by the Finished messages.
6283 *) More robust randomness gathering functions for Windows.
6284 [Jeffrey Altman <jaltman@columbia.edu>]
6286 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6287 not set then we don't setup the error code for issuer check errors
6288 to avoid possibly overwriting other errors which the callback does
6289 handle. If an application does set the flag then we assume it knows
6290 what it is doing and can handle the new informational codes
6294 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6295 a general "ANY" type, as such it should be able to decode anything
6296 including tagged types. However it didn't check the class so it would
6297 wrongly interpret tagged types in the same way as their universal
6298 counterpart and unknown types were just rejected. Changed so that the
6299 tagged and unknown types are handled in the same way as a SEQUENCE:
6300 that is the encoding is stored intact. There is also a new type
6301 "V_ASN1_OTHER" which is used when the class is not universal, in this
6302 case we have no idea what the actual type is so we just lump them all
6306 *) On VMS, stdout may very well lead to a file that is written to
6307 in a record-oriented fashion. That means that every write() will
6308 write a separate record, which will be read separately by the
6309 programs trying to read from it. This can be very confusing.
6311 The solution is to put a BIO filter in the way that will buffer
6312 text until a linefeed is reached, and then write everything a
6313 line at a time, so every record written will be an actual line,
6314 not chunks of lines and not (usually doesn't happen, but I've
6315 seen it once) several lines in one record. BIO_f_linebuffer() is
6318 Currently, it's a VMS-only method, because that's where it has
6319 been tested well enough.
6322 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6323 it can return incorrect results.
6324 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6325 but it was in 0.9.6-beta[12].)
6328 *) Disable the check for content being present when verifying detached
6329 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6330 include zero length content when signing messages.
6333 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6334 BIO_ctrl (for BIO pairs).
6337 *) Add DSO method for VMS.
6340 *) Bug fix: Montgomery multiplication could produce results with the
6344 *) Add RPM specification openssl.spec and modify it to build three
6345 packages. The default package contains applications, application
6346 documentation and run-time libraries. The devel package contains
6347 include files, static libraries and function documentation. The
6348 doc package contains the contents of the doc directory. The original
6349 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6352 *) Add a large number of documentation files for many SSL routines.
6353 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6355 *) Add a configuration entry for Sony News 4.
6356 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6358 *) Don't set the two most significant bits to one when generating a
6359 random number < q in the DSA library.
6362 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6363 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6364 the underlying transport is blocking) if a handshake took place.
6365 (The default behaviour is needed by applications such as s_client
6366 and s_server that use select() to determine when to use SSL_read;
6367 but for applications that know in advance when to expect data, it
6368 just makes things more complicated.)
6371 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6375 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6376 work better on such systems.
6377 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6379 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6380 Update PKCS12_parse() so it copies the friendlyName and the
6381 keyid to the certificates aux info.
6384 *) Fix bug in PKCS7_verify() which caused an infinite loop
6385 if there was more than one signature.
6386 [Sven Uszpelkat <su@celocom.de>]
6388 *) Major change in util/mkdef.pl to include extra information
6389 about each symbol, as well as presentig variables as well
6390 as functions. This change means that there's n more need
6391 to rebuild the .num files when some algorithms are excluded.
6394 *) Allow the verify time to be set by an application,
6395 rather than always using the current time.
6398 *) Phase 2 verify code reorganisation. The certificate
6399 verify code now looks up an issuer certificate by a
6400 number of criteria: subject name, authority key id
6401 and key usage. It also verifies self signed certificates
6402 by the same criteria. The main comparison function is
6403 X509_check_issued() which performs these checks.
6405 Lot of changes were necessary in order to support this
6406 without completely rewriting the lookup code.
6408 Authority and subject key identifier are now cached.
6410 The LHASH 'certs' is X509_STORE has now been replaced
6411 by a STACK_OF(X509_OBJECT). This is mainly because an
6412 LHASH can't store or retrieve multiple objects with
6413 the same hash value.
6415 As a result various functions (which were all internal
6416 use only) have changed to handle the new X509_STORE
6417 structure. This will break anything that messed round
6418 with X509_STORE internally.
6420 The functions X509_STORE_add_cert() now checks for an
6421 exact match, rather than just subject name.
6423 The X509_STORE API doesn't directly support the retrieval
6424 of multiple certificates matching a given criteria, however
6425 this can be worked round by performing a lookup first
6426 (which will fill the cache with candidate certificates)
6427 and then examining the cache for matches. This is probably
6428 the best we can do without throwing out X509_LOOKUP
6429 entirely (maybe later...).
6431 The X509_VERIFY_CTX structure has been enhanced considerably.
6433 All certificate lookup operations now go via a get_issuer()
6434 callback. Although this currently uses an X509_STORE it
6435 can be replaced by custom lookups. This is a simple way
6436 to bypass the X509_STORE hackery necessary to make this
6437 work and makes it possible to use more efficient techniques
6438 in future. A very simple version which uses a simple
6439 STACK for its trusted certificate store is also provided
6440 using X509_STORE_CTX_trusted_stack().
6442 The verify_cb() and verify() callbacks now have equivalents
6443 in the X509_STORE_CTX structure.
6445 X509_STORE_CTX also has a 'flags' field which can be used
6446 to customise the verify behaviour.
6449 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6450 excludes S/MIME capabilities.
6453 *) When a certificate request is read in keep a copy of the
6454 original encoding of the signed data and use it when outputing
6455 again. Signatures then use the original encoding rather than
6456 a decoded, encoded version which may cause problems if the
6457 request is improperly encoded.
6460 *) For consistency with other BIO_puts implementations, call
6461 buffer_write(b, ...) directly in buffer_puts instead of calling
6464 In BIO_puts, increment b->num_write as in BIO_write.
6465 [Peter.Sylvester@EdelWeb.fr]
6467 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6468 BN_zero, we may not return a BIGNUM with an array consisting of
6472 *) Avoid calling abort() from within the library when problems are
6473 detected, except if preprocessor symbols have been defined
6474 (such as REF_CHECK, BN_DEBUG etc.).
6477 *) New openssl application 'rsautl'. This utility can be
6478 used for low level RSA operations. DER public key
6479 BIO/fp routines also added.
6482 *) New Configure entry and patches for compiling on QNX 4.
6483 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6485 *) A demo state-machine implementation was sponsored by
6486 Nuron (http://www.nuron.com/) and is now available in
6487 demos/state_machine.
6490 *) New options added to the 'dgst' utility for signature
6491 generation and verification.
6494 *) Unrecognized PKCS#7 content types are now handled via a
6495 catch all ASN1_TYPE structure. This allows unsupported
6496 types to be stored as a "blob" and an application can
6497 encode and decode it manually.
6500 *) Fix various signed/unsigned issues to make a_strex.c
6502 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6504 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6505 length if passed a buffer. ASN1_INTEGER_to_BN failed
6506 if passed a NULL BN and its argument was negative.
6507 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6509 *) Modification to PKCS#7 encoding routines to output definite
6510 length encoding. Since currently the whole structures are in
6511 memory there's not real point in using indefinite length
6512 constructed encoding. However if OpenSSL is compiled with
6513 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6516 *) Added BIO_vprintf() and BIO_vsnprintf().
6519 *) Added more prefixes to parse for in the the strings written
6520 through a logging bio, to cover all the levels that are available
6521 through syslog. The prefixes are now:
6523 PANIC, EMERG, EMR => LOG_EMERG
6524 ALERT, ALR => LOG_ALERT
6525 CRIT, CRI => LOG_CRIT
6526 ERROR, ERR => LOG_ERR
6527 WARNING, WARN, WAR => LOG_WARNING
6528 NOTICE, NOTE, NOT => LOG_NOTICE
6529 INFO, INF => LOG_INFO
6530 DEBUG, DBG => LOG_DEBUG
6532 and as before, if none of those prefixes are present at the
6533 beginning of the string, LOG_ERR is chosen.
6535 On Win32, the LOG_* levels are mapped according to this:
6537 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6538 LOG_WARNING => EVENTLOG_WARNING_TYPE
6539 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6543 *) Made it possible to reconfigure with just the configuration
6544 argument "reconf" or "reconfigure". The command line arguments
6545 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6546 and are retrieved from there when reconfiguring.
6550 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6552 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6555 *) The obj_dat.pl script was messing up the sorting of object
6556 names. The reason was that it compared the quoted version
6557 of strings as a result "OCSP" > "OCSP Signing" because
6558 " > SPACE. Changed script to store unquoted versions of
6559 names and add quotes on output. It was also omitting some
6560 names from the lookup table if they were given a default
6561 value (that is if SN is missing it is given the same
6562 value as LN and vice versa), these are now added on the
6563 grounds that if an object has a name we should be able to
6564 look it up. Finally added warning output when duplicate
6565 short or long names are found.
6568 *) Changes needed for Tandem NSK.
6569 [Scott Uroff <scott@xypro.com>]
6571 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6572 RSA_padding_check_SSLv23(), special padding was never detected
6573 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6574 version rollback attacks was not effective.
6576 In s23_clnt.c, don't use special rollback-attack detection padding
6577 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6578 client; similarly, in s23_srvr.c, don't do the rollback check if
6579 SSL 2.0 is the only protocol enabled in the server.
6582 *) Make it possible to get hexdumps of unprintable data with 'openssl
6583 asn1parse'. By implication, the functions ASN1_parse_dump() and
6584 BIO_dump_indent() are added.
6587 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6588 these print out strings and name structures based on various
6589 flags including RFC2253 support and proper handling of
6590 multibyte characters. Added options to the 'x509' utility
6591 to allow the various flags to be set.
6594 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6595 Also change the functions X509_cmp_current_time() and
6596 X509_gmtime_adj() work with an ASN1_TIME structure,
6597 this will enable certificates using GeneralizedTime in validity
6598 dates to be checked.
6601 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6602 negative public key encodings) on by default,
6603 NO_NEG_PUBKEY_BUG can be set to disable it.
6606 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6607 content octets. An i2c_ASN1_OBJECT is unnecessary because
6608 the encoding can be trivially obtained from the structure.
6611 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6612 not read locks (CRYPTO_r_[un]lock).
6615 *) A first attempt at creating official support for shared
6616 libraries through configuration. I've kept it so the
6617 default is static libraries only, and the OpenSSL programs
6618 are always statically linked for now, but there are
6619 preparations for dynamic linking in place.
6620 This has been tested on Linux and Tru64.
6623 *) Randomness polling function for Win9x, as described in:
6624 Peter Gutmann, Software Generation of Practically Strong
6628 *) Fix so PRNG is seeded in req if using an already existing
6632 *) New options to smime application. -inform and -outform
6633 allow alternative formats for the S/MIME message including
6634 PEM and DER. The -content option allows the content to be
6635 specified separately. This should allow things like Netscape
6636 form signing output easier to verify.
6639 *) Fix the ASN1 encoding of tags using the 'long form'.
6642 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6643 STRING types. These convert content octets to and from the
6644 underlying type. The actual tag and length octets are
6645 already assumed to have been read in and checked. These
6646 are needed because all other string types have virtually
6647 identical handling apart from the tag. By having versions
6648 of the ASN1 functions that just operate on content octets
6649 IMPLICIT tagging can be handled properly. It also allows
6650 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6651 and ASN1_INTEGER are identical apart from the tag.
6654 *) Change the handling of OID objects as follows:
6656 - New object identifiers are inserted in objects.txt, following
6657 the syntax given in objects.README.
6658 - objects.pl is used to process obj_mac.num and create a new
6660 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6663 This is currently kind of a hack, and the perl code in objects.pl
6664 isn't very elegant, but it works as I intended. The simplest way
6665 to check that it worked correctly is to look in obj_dat.h and
6666 check the array nid_objs and make sure the objects haven't moved
6667 around (this is important!). Additions are OK, as well as
6668 consistent name changes.
6671 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6674 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6675 The given file adds to whatever has already been seeded into the
6676 random pool through the RANDFILE configuration file option or
6677 environment variable, or the default random state file.
6680 *) mkstack.pl now sorts each macro group into lexical order.
6681 Previously the output order depended on the order the files
6682 appeared in the directory, resulting in needless rewriting
6686 *) Patches to make OpenSSL compile under Win32 again. Mostly
6687 work arounds for the VC++ problem that it treats func() as
6688 func(void). Also stripped out the parts of mkdef.pl that
6689 added extra typesafe functions: these no longer exist.
6692 *) Reorganisation of the stack code. The macros are now all
6693 collected in safestack.h . Each macro is defined in terms of
6694 a "stack macro" of the form SKM_<name>(type, a, b). The
6695 DEBUG_SAFESTACK is now handled in terms of function casts,
6696 this has the advantage of retaining type safety without the
6697 use of additional functions. If DEBUG_SAFESTACK is not defined
6698 then the non typesafe macros are used instead. Also modified the
6699 mkstack.pl script to handle the new form. Needs testing to see
6700 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6701 the default if no major problems. Similar behaviour for ASN1_SET_OF
6702 and PKCS12_STACK_OF.
6705 *) When some versions of IIS use the 'NET' form of private key the
6706 key derivation algorithm is different. Normally MD5(password) is
6707 used as a 128 bit RC4 key. In the modified case
6708 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6709 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6710 as the old Netscape_RSA functions except they have an additional
6711 'sgckey' parameter which uses the modified algorithm. Also added
6712 an -sgckey command line option to the rsa utility. Thanks to
6713 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6714 algorithm to openssl-dev.
6717 *) The evp_local.h macros were using 'c.##kname' which resulted in
6718 invalid expansion on some systems (SCO 5.0.5 for example).
6719 Corrected to 'c.kname'.
6720 [Phillip Porch <root@theporch.com>]
6722 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6723 a STACK of email addresses from a certificate or request, these look
6724 in the subject name and the subject alternative name extensions and
6725 omit any duplicate addresses.
6728 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6729 This makes DSA verification about 2 % faster.
6732 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6733 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6734 plus overhead for 1024 bit moduli).
6735 This makes exponentiations about 0.5 % faster for 1024 bit
6736 exponents (as measured by "openssl speed rsa2048").
6739 *) Rename memory handling macros to avoid conflicts with other
6741 Malloc => OPENSSL_malloc
6742 Malloc_locked => OPENSSL_malloc_locked
6743 Realloc => OPENSSL_realloc
6744 Free => OPENSSL_free
6747 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6748 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6751 *) CygWin32 support.
6752 [John Jarvie <jjarvie@newsguy.com>]
6754 *) The type-safe stack code has been rejigged. It is now only compiled
6755 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6756 by default all type-specific stack functions are "#define"d back to
6757 standard stack functions. This results in more streamlined output
6758 but retains the type-safety checking possibilities of the original
6762 *) The STACK code has been cleaned up, and certain type declarations
6763 that didn't make a lot of sense have been brought in line. This has
6764 also involved a cleanup of sorts in safestack.h to more correctly
6765 map type-safe stack functions onto their plain stack counterparts.
6766 This work has also resulted in a variety of "const"ifications of
6767 lots of the code, especially "_cmp" operations which should normally
6768 be prototyped with "const" parameters anyway.
6771 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6772 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6773 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6774 where all of 'md' is used each time the PRNG is used, but 'state'
6775 is used only indexed by a cyclic counter. As entropy may not be
6776 well distributed from the beginning, 'md' is important as a
6777 chaining variable. However, the output function chains only half
6778 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6779 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6780 in all of 'state' being rewritten, with the new values depending
6781 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6784 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6785 the handshake is continued after ssl_verify_cert_chain();
6786 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6787 can lead to 'unexplainable' connection aborts later.
6788 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6790 *) Major EVP API cipher revision.
6791 Add hooks for extra EVP features. This allows various cipher
6792 parameters to be set in the EVP interface. Support added for variable
6793 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6794 setting of RC2 and RC5 parameters.
6796 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6799 Remove lots of duplicated code from the EVP library. For example *every*
6800 cipher init() function handles the 'iv' in the same way according to the
6801 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6802 for CFB and OFB modes they zero ctx->num.
6804 New functionality allows removal of S/MIME code RC2 hack.
6806 Most of the routines have the same form and so can be declared in terms
6809 By shifting this to the top level EVP_CipherInit() it can be removed from
6810 all individual ciphers. If the cipher wants to handle IVs or keys
6811 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6814 Change lots of functions like EVP_EncryptUpdate() to now return a
6815 value: although software versions of the algorithms cannot fail
6816 any installed hardware versions can.
6819 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6820 this option is set, tolerate broken clients that send the negotiated
6821 protocol version number instead of the requested protocol version
6825 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6826 i.e. non-zero for export ciphersuites, zero otherwise.
6827 Previous versions had this flag inverted, inconsistent with
6828 rsa_tmp_cb (..._TMP_RSA_CB).
6829 [Bodo Moeller; problem reported by Amit Chopra]
6831 *) Add missing DSA library text string. Work around for some IIS
6832 key files with invalid SEQUENCE encoding.
6835 *) Add a document (doc/standards.txt) that list all kinds of standards
6836 and so on that are implemented in OpenSSL.
6839 *) Enhance c_rehash script. Old version would mishandle certificates
6840 with the same subject name hash and wouldn't handle CRLs at all.
6841 Added -fingerprint option to crl utility, to support new c_rehash
6845 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6848 *) Fix for SSL server purpose checking. Server checking was
6849 rejecting certificates which had extended key usage present
6850 but no ssl client purpose.
6851 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6853 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6854 is a little unclear about how a blank password is handled.
6855 Since the password in encoded as a BMPString with terminating
6856 double NULL a zero length password would end up as just the
6857 double NULL. However no password at all is different and is
6858 handled differently in the PKCS#12 key generation code. NS
6859 treats a blank password as zero length. MSIE treats it as no
6860 password on export: but it will try both on import. We now do
6861 the same: PKCS12_parse() tries zero length and no password if
6862 the password is set to "" or NULL (NULL is now a valid password:
6863 it wasn't before) as does the pkcs12 application.
6866 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6867 perror when PEM_read_bio_X509_REQ fails, the error message must
6868 be obtained from the error queue.
6871 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6872 it in ERR_remove_state if appropriate, and change ERR_get_state
6873 accordingly to avoid race conditions (this is necessary because
6874 thread_hash is no longer constant once set).
6877 *) Bugfix for linux-elf makefile.one.
6880 *) RSA_get_default_method() will now cause a default
6881 RSA_METHOD to be chosen if one doesn't exist already.
6882 Previously this was only set during a call to RSA_new()
6883 or RSA_new_method(NULL) meaning it was possible for
6884 RSA_get_default_method() to return NULL.
6887 *) Added native name translation to the existing DSO code
6888 that will convert (if the flag to do so is set) filenames
6889 that are sufficiently small and have no path information
6890 into a canonical native form. Eg. "blah" converted to
6891 "libblah.so" or "blah.dll" etc.
6894 *) New function ERR_error_string_n(e, buf, len) which is like
6895 ERR_error_string(e, buf), but writes at most 'len' bytes
6896 including the 0 terminator. For ERR_error_string_n, 'buf'
6898 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
6900 *) CONF library reworked to become more general. A new CONF
6901 configuration file reader "class" is implemented as well as a
6902 new functions (NCONF_*, for "New CONF") to handle it. The now
6903 old CONF_* functions are still there, but are reimplemented to
6904 work in terms of the new functions. Also, a set of functions
6905 to handle the internal storage of the configuration data is
6906 provided to make it easier to write new configuration file
6907 reader "classes" (I can definitely see something reading a
6908 configuration file in XML format, for example), called _CONF_*,
6909 or "the configuration storage API"...
6911 The new configuration file reading functions are:
6913 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
6914 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
6916 NCONF_default, NCONF_WIN32
6918 NCONF_dump_fp, NCONF_dump_bio
6920 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
6921 NCONF_new creates a new CONF object. This works in the same way
6922 as other interfaces in OpenSSL, like the BIO interface.
6923 NCONF_dump_* dump the internal storage of the configuration file,
6924 which is useful for debugging. All other functions take the same
6925 arguments as the old CONF_* functions wth the exception of the
6926 first that must be a `CONF *' instead of a `LHASH *'.
6928 To make it easer to use the new classes with the old CONF_* functions,
6929 the function CONF_set_default_method is provided.
6932 *) Add '-tls1' option to 'openssl ciphers', which was already
6933 mentioned in the documentation but had not been implemented.
6934 (This option is not yet really useful because even the additional
6935 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
6938 *) Initial DSO code added into libcrypto for letting OpenSSL (and
6939 OpenSSL-based applications) load shared libraries and bind to
6940 them in a portable way.
6941 [Geoff Thorpe, with contributions from Richard Levitte]
6943 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
6945 *) Make sure _lrotl and _lrotr are only used with MSVC.
6947 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
6948 (the default implementation of RAND_status).
6950 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
6951 to '-clrext' (= clear extensions), as intended and documented.
6952 [Bodo Moeller; inconsistency pointed out by Michael Attili
6955 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
6956 was larger than the MD block size.
6957 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
6959 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
6960 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
6961 using the passed key: if the passed key was a private key the result
6962 of X509_print(), for example, would be to print out all the private key
6966 *) des_quad_cksum() byte order bug fix.
6967 [Ulf Möller, using the problem description in krb4-0.9.7, where
6968 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
6970 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
6972 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
6974 *) For easily testing in shell scripts whether some command
6975 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
6976 returns with exit code 0 iff no command of the given name is available.
6977 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
6978 the output goes to stdout and nothing is printed to stderr.
6979 Additional arguments are always ignored.
6981 Since for each cipher there is a command of the same name,
6982 the 'no-cipher' compilation switches can be tested this way.
6984 ('openssl no-XXX' is not able to detect pseudo-commands such
6985 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
6988 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
6991 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
6992 is set; it will be thrown away anyway because each handshake creates
6994 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
6995 to parameters -- in previous versions (since OpenSSL 0.9.3) the
6996 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
6997 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7000 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7001 'Q' and 'R' lose their special meanings (quit/renegotiate).
7002 This is part of what -quiet does; unlike -quiet, -ign_eof
7003 does not suppress any output.
7006 *) Add compatibility options to the purpose and trust code. The
7007 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7008 accepts a certificate or CA, this was the previous behaviour,
7009 with all the associated security issues.
7011 X509_TRUST_COMPAT is the old trust behaviour: only and
7012 automatically trust self signed roots in certificate store. A
7013 new trust setting X509_TRUST_DEFAULT is used to specify that
7014 a purpose has no associated trust setting and it should instead
7015 use the value in the default purpose.
7018 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7019 and fix a memory leak.
7022 *) In util/mkerr.pl (which implements 'make errors'), preserve
7023 reason strings from the previous version of the .c file, as
7024 the default to have only downcase letters (and digits) in
7025 automatically generated reasons codes is not always appropriate.
7028 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7029 using strerror. Previously, ERR_reason_error_string() returned
7030 library names as reason strings for SYSerr; but SYSerr is a special
7031 case where small numbers are errno values, not library numbers.
7034 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7035 converts DSA parameters into DH parameters. (When creating parameters,
7036 DSA_generate_parameters is used.)
7039 *) Include 'length' (recommended exponent length) in C code generated
7040 by 'openssl dhparam -C'.
7043 *) The second argument to set_label in perlasm was already being used
7044 so couldn't be used as a "file scope" flag. Moved to third argument
7048 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7049 instead of RAND_bytes for encryption IVs and salts.
7052 *) Include RAND_status() into RAND_METHOD instead of implementing
7053 it only for md_rand.c Otherwise replacing the PRNG by calling
7054 RAND_set_rand_method would be impossible.
7057 *) Don't let DSA_generate_key() enter an infinite loop if the random
7058 number generation fails.
7061 *) New 'rand' application for creating pseudo-random output.
7064 *) Added configuration support for Linux/IA64
7065 [Rolf Haberrecker <rolf@suse.de>]
7067 *) Assembler module support for Mingw32.
7070 *) Shared library support for HPUX (in shlib/).
7071 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7073 *) Shared library support for Solaris gcc.
7074 [Lutz Behnke <behnke@trustcenter.de>]
7076 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7078 *) PKCS7_encrypt() was adding text MIME headers twice because they
7079 were added manually and by SMIME_crlf_copy().
7082 *) In bntest.c don't call BN_rand with zero bits argument.
7083 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7085 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7086 case was implemented. This caused BN_div_recp() to fail occasionally.
7089 *) Add an optional second argument to the set_label() in the perl
7090 assembly language builder. If this argument exists and is set
7091 to 1 it signals that the assembler should use a symbol whose
7092 scope is the entire file, not just the current function. This
7093 is needed with MASM which uses the format label:: for this scope.
7094 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7096 *) Change the ASN1 types so they are typedefs by default. Before
7097 almost all types were #define'd to ASN1_STRING which was causing
7098 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7102 *) Change names of new functions to the new get1/get0 naming
7103 convention: After 'get1', the caller owns a reference count
7104 and has to call ..._free; 'get0' returns a pointer to some
7105 data structure without incrementing reference counters.
7106 (Some of the existing 'get' functions increment a reference
7107 counter, some don't.)
7108 Similarly, 'set1' and 'add1' functions increase reference
7109 counters or duplicate objects.
7112 *) Allow for the possibility of temp RSA key generation failure:
7113 the code used to assume it always worked and crashed on failure.
7116 *) Fix potential buffer overrun problem in BIO_printf().
7117 [Ulf Möller, using public domain code by Patrick Powell; problem
7118 pointed out by David Sacerdote <das33@cornell.edu>]
7120 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7121 RAND_egd() and RAND_status(). In the command line application,
7122 the EGD socket can be specified like a seed file using RANDFILE
7126 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7127 Some CAs (e.g. Verisign) distribute certificates in this form.
7130 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7131 list to exclude them. This means that no special compilation option
7132 is needed to use anonymous DH: it just needs to be included in the
7136 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7137 EVP_MD_type. The old functionality is available in a new macro called
7138 EVP_MD_md(). Change code that uses it and update docs.
7141 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7142 where the 'void *' argument is replaced by a function pointer argument.
7143 Previously 'void *' was abused to point to functions, which works on
7144 many platforms, but is not correct. As these functions are usually
7145 called by macros defined in OpenSSL header files, most source code
7146 should work without changes.
7149 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7150 sections with information on -D... compiler switches used for
7151 compiling the library so that applications can see them. To enable
7152 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7153 must be defined. E.g.,
7154 #define OPENSSL_ALGORITHM_DEFINES
7155 #include <openssl/opensslconf.h>
7156 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7157 [Richard Levitte, Ulf and Bodo Möller]
7159 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7163 *) Change the 'other' type in certificate aux info to a STACK_OF
7164 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7165 the required ASN1 format: arbitrary types determined by an OID.
7168 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7169 argument to 'req'. This is not because the function is newer or
7170 better than others it just uses the work 'NEW' in the certificate
7171 request header lines. Some software needs this.
7174 *) Reorganise password command line arguments: now passwords can be
7175 obtained from various sources. Delete the PEM_cb function and make
7176 it the default behaviour: i.e. if the callback is NULL and the
7177 usrdata argument is not NULL interpret it as a null terminated pass
7178 phrase. If usrdata and the callback are NULL then the pass phrase
7179 is prompted for as usual.
7182 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7183 the support is automatically enabled. The resulting binaries will
7184 autodetect the card and use it if present.
7185 [Ben Laurie and Compaq Inc.]
7187 *) Work around for Netscape hang bug. This sends certificate request
7188 and server done in one record. Since this is perfectly legal in the
7189 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7190 the bugs/SSLv3 entry for more info.
7193 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7196 *) Add -rand argument to smime and pkcs12 applications and read/write
7200 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7203 *) Add command line password options to the remaining applications.
7206 *) Bug fix for BN_div_recp() for numerators with an even number of
7210 *) More tests in bntest.c, and changed test_bn output.
7213 *) ./config recognizes MacOS X now.
7216 *) Bug fix for BN_div() when the first words of num and divsor are
7217 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7220 *) Add support for various broken PKCS#8 formats, and command line
7221 options to produce them.
7224 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7225 get temporary BIGNUMs from a BN_CTX.
7228 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7232 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7233 include a #define from the old name to the new. The original intent
7234 was that statically linked binaries could for example just call
7235 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7236 link with digests. This never worked becayse SSLeay_add_all_digests()
7237 and SSLeay_add_all_ciphers() were in the same source file so calling
7238 one would link with the other. They are now in separate source files.
7241 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7244 *) Use a less unusual form of the Miller-Rabin primality test (it used
7245 a binary algorithm for exponentiation integrated into the Miller-Rabin
7246 loop, our standard modexp algorithms are faster).
7249 *) Support for the EBCDIC character set completed.
7250 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7252 *) Source code cleanups: use const where appropriate, eliminate casts,
7253 use void * instead of char * in lhash.
7256 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7257 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7258 this the server could overwrite ephemeral keys that the client
7262 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7263 using 50 iterations of the Rabin-Miller test.
7265 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7266 iterations of the Rabin-Miller test as required by the appendix
7267 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7268 As BN_is_prime_fasttest includes trial division, DSA parameter
7269 generation becomes much faster.
7271 This implies a change for the callback functions in DSA_is_prime
7272 and DSA_generate_parameters: The callback function is called once
7273 for each positive witness in the Rabin-Miller test, not just
7274 occasionally in the inner loop; and the parameters to the
7275 callback function now provide an iteration count for the outer
7276 loop rather than for the current invocation of the inner loop.
7277 DSA_generate_parameters additionally can call the callback
7278 function with an 'iteration count' of -1, meaning that a
7279 candidate has passed the trial division test (when q is generated
7280 from an application-provided seed, trial division is skipped).
7283 *) New function BN_is_prime_fasttest that optionally does trial
7284 division before starting the Rabin-Miller test and has
7285 an additional BN_CTX * argument (whereas BN_is_prime always
7286 has to allocate at least one BN_CTX).
7287 'callback(1, -1, cb_arg)' is called when a number has passed the
7288 trial division stage.
7291 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7295 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7298 *) New function BN_pseudo_rand().
7301 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7302 bignum version of BN_from_montgomery() with the working code from
7303 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7307 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7308 made it impossible to use the same SSL_SESSION data structure in
7309 SSL2 clients in multiple threads.
7312 *) The return value of RAND_load_file() no longer counts bytes obtained
7313 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7314 to seed the PRNG (previously an explicit byte count was required).
7315 [Ulf Möller, Bodo Möller]
7317 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7318 used (char *) instead of (void *) and had casts all over the place.
7321 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7324 *) Retain source code compatibility for BN_prime_checks macro:
7325 BN_is_prime(..., BN_prime_checks, ...) now uses
7326 BN_prime_checks_for_size to determine the appropriate number of
7327 Rabin-Miller iterations.
7330 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7331 DH_CHECK_P_NOT_SAFE_PRIME.
7332 (Check if this is true? OpenPGP calls them "strong".)
7335 *) Merge the functionality of "dh" and "gendh" programs into a new program
7336 "dhparam". The old programs are retained for now but will handle DH keys
7337 (instead of parameters) in future.
7340 *) Make the ciphers, s_server and s_client programs check the return values
7341 when a new cipher list is set.
7344 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7345 ciphers. Before when the 56bit ciphers were enabled the sorting was
7348 The syntax for the cipher sorting has been extended to support sorting by
7349 cipher-strength (using the strength_bits hard coded in the tables).
7350 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7352 Fix a bug in the cipher-command parser: when supplying a cipher command
7353 string with an "undefined" symbol (neither command nor alphanumeric
7354 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7355 an error is flagged.
7357 Due to the strength-sorting extension, the code of the
7358 ssl_create_cipher_list() function was completely rearranged. I hope that
7359 the readability was also increased :-)
7360 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7362 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7363 for the first serial number and places 2 in the serial number file. This
7364 avoids problems when the root CA is created with serial number zero and
7365 the first user certificate has the same issuer name and serial number
7369 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7370 the new code. Add documentation for this stuff.
7373 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7374 X509_*() to X509at_*() on the grounds that they don't handle X509
7375 structures and behave in an analagous way to the X509v3 functions:
7376 they shouldn't be called directly but wrapper functions should be used
7379 So we also now have some wrapper functions that call the X509at functions
7380 when passed certificate requests. (TO DO: similar things can be done with
7381 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7382 things. Some of these need some d2i or i2d and print functionality
7383 because they handle more complex structures.)
7386 *) Add missing #ifndefs that caused missing symbols when building libssl
7387 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7388 NO_RSA in ssl/s2*.c.
7389 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7391 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7392 has a return value which indicates the quality of the random data
7393 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7394 error queue. New function RAND_pseudo_bytes() generates output that is
7395 guaranteed to be unique but not unpredictable. RAND_add is like
7396 RAND_seed, but takes an extra argument for an entropy estimate
7397 (RAND_seed always assumes full entropy).
7400 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7401 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7402 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7403 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7404 false-positive rate of at most 2^-80 for random input.
7407 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7410 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7411 in the 0.9.5 release), this returns the chain
7412 from an X509_CTX structure with a dup of the stack and all
7413 the X509 reference counts upped: so the stack will exist
7414 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7417 Also make SSL_SESSION_print() print out the verify return
7421 *) Add manpage for the pkcs12 command. Also change the default
7422 behaviour so MAC iteration counts are used unless the new
7423 -nomaciter option is used. This improves file security and
7424 only older versions of MSIE (4.0 for example) need it.
7427 *) Honor the no-xxx Configure options when creating .DEF files.
7430 *) Add PKCS#10 attributes to field table: challengePassword,
7431 unstructuredName and unstructuredAddress. These are taken from
7432 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7433 international characters are used.
7435 More changes to X509_ATTRIBUTE code: allow the setting of types
7436 based on strings. Remove the 'loc' parameter when adding
7437 attributes because these will be a SET OF encoding which is sorted
7441 *) Initial changes to the 'req' utility to allow request generation
7442 automation. This will allow an application to just generate a template
7443 file containing all the field values and have req construct the
7446 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7447 used all over the place including certificate requests and PKCS#7
7448 structures. They are currently handled manually where necessary with
7449 some primitive wrappers for PKCS#7. The new functions behave in a
7450 manner analogous to the X509 extension functions: they allow
7451 attributes to be looked up by NID and added.
7453 Later something similar to the X509V3 code would be desirable to
7454 automatically handle the encoding, decoding and printing of the
7455 more complex types. The string types like challengePassword can
7456 be handled by the string table functions.
7458 Also modified the multi byte string table handling. Now there is
7459 a 'global mask' which masks out certain types. The table itself
7460 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7461 is useful when for example there is only one permissible type
7462 (as in countryName) and using the mask might result in no valid
7466 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7467 SSL_get_peer_finished to allow applications to obtain the latest
7468 Finished messages sent to the peer or expected from the peer,
7469 respectively. (SSL_get_peer_finished is usually the Finished message
7470 actually received from the peer, otherwise the protocol will be aborted.)
7472 As the Finished message are message digests of the complete handshake
7473 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7474 be used for external authentication procedures when the authentication
7475 provided by SSL/TLS is not desired or is not enough.
7478 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7479 the host supports BWX extension and if Compaq C is present on the
7480 $PATH. Just exploiting of the BWX extension results in 20-30%
7481 performance kick for some algorithms, e.g. DES and RC4 to mention
7482 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7486 *) Add support for MS "fast SGC". This is arguably a violation of the
7487 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7488 weak crypto and after checking the certificate is SGC a second one
7489 with strong crypto. MS SGC stops the first handshake after receiving
7490 the server certificate message and sends a second client hello. Since
7491 a server will typically do all the time consuming operations before
7492 expecting any further messages from the client (server key exchange
7493 is the most expensive) there is little difference between the two.
7495 To get OpenSSL to support MS SGC we have to permit a second client
7496 hello message after we have sent server done. In addition we have to
7497 reset the MAC if we do get this second client hello.
7500 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7501 if a DER encoded private key is RSA or DSA traditional format. Changed
7502 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7503 format DER encoded private key. Newer code should use PKCS#8 format which
7504 has the key type encoded in the ASN1 structure. Added DER private key
7505 support to pkcs8 application.
7508 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7509 ciphersuites has been selected (as required by the SSL 3/TLS 1
7510 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7511 is set, we interpret this as a request to violate the specification
7512 (the worst that can happen is a handshake failure, and 'correct'
7513 behaviour would result in a handshake failure anyway).
7516 *) In SSL_CTX_add_session, take into account that there might be multiple
7517 SSL_SESSION structures with the same session ID (e.g. when two threads
7518 concurrently obtain them from an external cache).
7519 The internal cache can handle only one SSL_SESSION with a given ID,
7520 so if there's a conflict, we now throw out the old one to achieve
7524 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7525 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7526 some routines that use cipher OIDs: some ciphers do not have OIDs
7527 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7531 *) Simplify the trust setting structure and code. Now we just have
7532 two sequences of OIDs for trusted and rejected settings. These will
7533 typically have values the same as the extended key usage extension
7534 and any application specific purposes.
7536 The trust checking code now has a default behaviour: it will just
7537 check for an object with the same NID as the passed id. Functions can
7538 be provided to override either the default behaviour or the behaviour
7539 for a given id. SSL client, server and email already have functions
7540 in place for compatibility: they check the NID and also return "trusted"
7541 if the certificate is self signed.
7544 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7545 traditional format into an EVP_PKEY structure.
7548 *) Add a password callback function PEM_cb() which either prompts for
7549 a password if usr_data is NULL or otherwise assumes it is a null
7550 terminated password. Allow passwords to be passed on command line
7551 environment or config files in a few more utilities.
7554 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7555 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7556 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7557 Update documentation.
7560 *) Support for ASN1 "NULL" type. This could be handled before by using
7561 ASN1_TYPE but there wasn't any function that would try to read a NULL
7562 and produce an error if it couldn't. For compatibility we also have
7563 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7564 don't allocate anything because they don't need to.
7567 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7569 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7571 *) Rebuild of the memory allocation routines used by OpenSSL code and
7572 possibly others as well. The purpose is to make an interface that
7573 provide hooks so anyone can build a separate set of allocation and
7574 deallocation routines to be used by OpenSSL, for example memory
7575 pool implementations, or something else, which was previously hard
7576 since Malloc(), Realloc() and Free() were defined as macros having
7577 the values malloc, realloc and free, respectively (except for Win32
7578 compilations). The same is provided for memory debugging code.
7579 OpenSSL already comes with functionality to find memory leaks, but
7580 this gives people a chance to debug other memory problems.
7582 With these changes, a new set of functions and macros have appeared:
7584 CRYPTO_set_mem_debug_functions() [F]
7585 CRYPTO_get_mem_debug_functions() [F]
7586 CRYPTO_dbg_set_options() [F]
7587 CRYPTO_dbg_get_options() [F]
7588 CRYPTO_malloc_debug_init() [M]
7590 The memory debug functions are NULL by default, unless the library
7591 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7592 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7593 gives the standard debugging functions that come with OpenSSL) or
7594 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7595 provided by the library user) must be used. When the standard
7596 debugging functions are used, CRYPTO_dbg_set_options can be used to
7597 request additional information:
7598 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7599 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7601 Also, things like CRYPTO_set_mem_functions will always give the
7602 expected result (the new set of functions is used for allocation
7603 and deallocation) at all times, regardless of platform and compiler
7606 To finish it up, some functions that were never use in any other
7607 way than through macros have a new API and new semantic:
7610 CRYPTO_dbg_realloc()
7613 All macros of value have retained their old syntax.
7614 [Richard Levitte and Bodo Moeller]
7616 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7617 ordering of SMIMECapabilities wasn't in "strength order" and there
7618 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7622 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7623 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7624 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7626 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7627 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7628 functionality to handle multipart/signed properly) and a utility
7629 called 'smime' to call all this stuff. This is based on code I
7630 originally wrote for Celo who have kindly allowed it to be
7631 included in OpenSSL.
7634 *) Add variants des_set_key_checked and des_set_key_unchecked of
7635 des_set_key (aka des_key_sched). Global variable des_check_key
7636 decides which of these is called by des_set_key; this way
7637 des_check_key behaves as it always did, but applications and
7638 the library itself, which was buggy for des_check_key == 1,
7639 have a cleaner way to pick the version they need.
7642 *) New function PKCS12_newpass() which changes the password of a
7646 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7647 dynamic mix. In both cases the ids can be used as an index into the
7648 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7649 functions so they accept a list of the field values and the
7650 application doesn't need to directly manipulate the X509_TRUST
7654 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7658 *) Modify the way the V3 extension code looks up extensions. This now
7659 works in a similar way to the object code: we have some "standard"
7660 extensions in a static table which is searched with OBJ_bsearch()
7661 and the application can add dynamic ones if needed. The file
7662 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7663 updated whenever a new extension is added to the core code and kept
7664 in ext_nid order. There is a simple program 'tabtest.c' which checks
7665 this. New extensions are not added too often so this file can readily
7666 be maintained manually.
7668 There are two big advantages in doing things this way. The extensions
7669 can be looked up immediately and no longer need to be "added" using
7670 X509V3_add_standard_extensions(): this function now does nothing.
7671 [Side note: I get *lots* of email saying the extension code doesn't
7672 work because people forget to call this function]
7673 Also no dynamic allocation is done unless new extensions are added:
7674 so if we don't add custom extensions there is no need to call
7675 X509V3_EXT_cleanup().
7678 *) Modify enc utility's salting as follows: make salting the default. Add a
7679 magic header, so unsalted files fail gracefully instead of just decrypting
7680 to garbage. This is because not salting is a big security hole, so people
7681 should be discouraged from doing it.
7684 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7685 digest to be passed on the command line but it only used this
7686 parameter when signing a certificate. Modified so all relevant
7687 operations are affected by the digest parameter including the
7688 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7689 DSA key was used because it didn't fix the digest.
7692 *) Initial certificate chain verify code. Currently tests the untrusted
7693 certificates for consistency with the verify purpose (which is set
7694 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7696 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7697 this is because it will reject chains with invalid extensions whereas
7698 every previous version of OpenSSL and SSLeay made no checks at all.
7700 Trust code: checks the root CA for the relevant trust settings. Trust
7701 settings have an initial value consistent with the verify purpose: e.g.
7702 if the verify purpose is for SSL client use it expects the CA to be
7703 trusted for SSL client use. However the default value can be changed to
7704 permit custom trust settings: one example of this would be to only trust
7705 certificates from a specific "secure" set of CAs.
7707 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7708 which should be used for version portability: especially since the
7709 verify structure is likely to change more often now.
7711 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7712 to set them. If not set then assume SSL clients will verify SSL servers
7715 Two new options to the verify program: -untrusted allows a set of
7716 untrusted certificates to be passed in and -purpose which sets the
7717 intended purpose of the certificate. If a purpose is set then the
7718 new chain verify code is used to check extension consistency.
7721 *) Support for the authority information access extension.
7724 *) Modify RSA and DSA PEM read routines to transparently handle
7725 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7726 public keys in a format compatible with certificate
7727 SubjectPublicKeyInfo structures. Unfortunately there were already
7728 functions called *_PublicKey_* which used various odd formats so
7729 these are retained for compatibility: however the DSA variants were
7730 never in a public release so they have been deleted. Changed dsa/rsa
7731 utilities to handle the new format: note no releases ever handled public
7732 keys so we should be OK.
7734 The primary motivation for this change is to avoid the same fiasco
7735 that dogs private keys: there are several incompatible private key
7736 formats some of which are standard and some OpenSSL specific and
7737 require various evil hacks to allow partial transparent handling and
7738 even then it doesn't work with DER formats. Given the option anything
7739 other than PKCS#8 should be dumped: but the other formats have to
7740 stay in the name of compatibility.
7742 With public keys and the benefit of hindsight one standard format
7743 is used which works with EVP_PKEY, RSA or DSA structures: though
7744 it clearly returns an error if you try to read the wrong kind of key.
7746 Added a -pubkey option to the 'x509' utility to output the public key.
7747 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7748 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7749 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7750 that do the same as the EVP_PKEY_assign_*() except they up the
7751 reference count of the added key (they don't "swallow" the
7755 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7756 CRLs would fail if the file contained no certificates or no CRLs:
7757 added a new function to read in both types and return the number
7758 read: this means that if none are read it will be an error. The
7759 DER versions of the certificate and CRL reader would always fail
7760 because it isn't possible to mix certificates and CRLs in DER format
7761 without choking one or the other routine. Changed this to just read
7762 a certificate: this is the best we can do. Also modified the code
7763 in apps/verify.c to take notice of return codes: it was previously
7764 attempting to read in certificates from NULL pointers and ignoring
7765 any errors: this is one reason why the cert and CRL reader seemed
7766 to work. It doesn't check return codes from the default certificate
7767 routines: these may well fail if the certificates aren't installed.
7770 *) Code to support otherName option in GeneralName.
7773 *) First update to verify code. Change the verify utility
7774 so it warns if it is passed a self signed certificate:
7775 for consistency with the normal behaviour. X509_verify
7776 has been modified to it will now verify a self signed
7777 certificate if *exactly* the same certificate appears
7778 in the store: it was previously impossible to trust a
7779 single self signed certificate. This means that:
7780 openssl verify ss.pem
7781 now gives a warning about a self signed certificate but
7782 openssl verify -CAfile ss.pem ss.pem
7786 *) For servers, store verify_result in SSL_SESSION data structure
7787 (and add it to external session representation).
7788 This is needed when client certificate verifications fails,
7789 but an application-provided verification callback (set by
7790 SSL_CTX_set_cert_verify_callback) allows accepting the session
7791 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7792 but returns 1): When the session is reused, we have to set
7793 ssl->verify_result to the appropriate error code to avoid
7795 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7797 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7798 case in PKCS7_dataInit() where the signed PKCS7 structure
7799 didn't contain any existing data because it was being created.
7800 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7802 *) Add a salt to the key derivation routines in enc.c. This
7803 forms the first 8 bytes of the encrypted file. Also add a
7804 -S option to allow a salt to be input on the command line.
7807 *) New function X509_cmp(). Oddly enough there wasn't a function
7808 to compare two certificates. We do this by working out the SHA1
7809 hash and comparing that. X509_cmp() will be needed by the trust
7813 *) SSL_get1_session() is like SSL_get_session(), but increments
7814 the reference count in the SSL_SESSION returned.
7815 [Geoff Thorpe <geoff@eu.c2.net>]
7817 *) Fix for 'req': it was adding a null to request attributes.
7818 Also change the X509_LOOKUP and X509_INFO code to handle
7819 certificate auxiliary information.
7822 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7826 *) Add the possibility to add extra information to the memory leak
7827 detecting output, to form tracebacks, showing from where each
7828 allocation was originated: CRYPTO_push_info("constant string") adds
7829 the string plus current file name and line number to a per-thread
7830 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7831 is like calling CYRPTO_pop_info() until the stack is empty.
7832 Also updated memory leak detection code to be multi-thread-safe.
7835 *) Add options -text and -noout to pkcs7 utility and delete the
7836 encryption options which never did anything. Update docs.
7839 *) Add options to some of the utilities to allow the pass phrase
7840 to be included on either the command line (not recommended on
7841 OSes like Unix) or read from the environment. Update the
7842 manpages and fix a few bugs.
7845 *) Add a few manpages for some of the openssl commands.
7848 *) Fix the -revoke option in ca. It was freeing up memory twice,
7849 leaking and not finding already revoked certificates.
7852 *) Extensive changes to support certificate auxiliary information.
7853 This involves the use of X509_CERT_AUX structure and X509_AUX
7854 functions. An X509_AUX function such as PEM_read_X509_AUX()
7855 can still read in a certificate file in the usual way but it
7856 will also read in any additional "auxiliary information". By
7857 doing things this way a fair degree of compatibility can be
7858 retained: existing certificates can have this information added
7859 using the new 'x509' options.
7861 Current auxiliary information includes an "alias" and some trust
7862 settings. The trust settings will ultimately be used in enhanced
7863 certificate chain verification routines: currently a certificate
7864 can only be trusted if it is self signed and then it is trusted
7868 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7869 The problem was that one of the replacement routines had not been working
7870 since SSLeay releases. For now the offending routine has been replaced
7871 with non-optimised assembler. Even so, this now gives around 95%
7872 performance improvement for 1024 bit RSA signs.
7875 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7876 handling. Most clients have the effective key size in bits equal to
7877 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
7878 A few however don't do this and instead use the size of the decrypted key
7879 to determine the RC2 key length and the AlgorithmIdentifier to determine
7880 the effective key length. In this case the effective key length can still
7881 be 40 bits but the key length can be 168 bits for example. This is fixed
7882 by manually forcing an RC2 key into the EVP_PKEY structure because the
7883 EVP code can't currently handle unusual RC2 key sizes: it always assumes
7884 the key length and effective key length are equal.
7887 *) Add a bunch of functions that should simplify the creation of
7888 X509_NAME structures. Now you should be able to do:
7889 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
7890 and have it automatically work out the correct field type and fill in
7891 the structures. The more adventurous can try:
7892 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
7893 and it will (hopefully) work out the correct multibyte encoding.
7896 *) Change the 'req' utility to use the new field handling and multibyte
7897 copy routines. Before the DN field creation was handled in an ad hoc
7898 way in req, ca, and x509 which was rather broken and didn't support
7899 BMPStrings or UTF8Strings. Since some software doesn't implement
7900 BMPStrings or UTF8Strings yet, they can be enabled using the config file
7901 using the dirstring_type option. See the new comment in the default
7902 openssl.cnf for more info.
7905 *) Make crypto/rand/md_rand.c more robust:
7906 - Assure unique random numbers after fork().
7907 - Make sure that concurrent threads access the global counter and
7908 md serializably so that we never lose entropy in them
7909 or use exactly the same state in multiple threads.
7910 Access to the large state is not always serializable because
7911 the additional locking could be a performance killer, and
7912 md should be large enough anyway.
7915 *) New file apps/app_rand.c with commonly needed functionality
7916 for handling the random seed file.
7918 Use the random seed file in some applications that previously did not:
7920 dsaparam -genkey (which also ignored its '-rand' option),
7923 x509 (when signing).
7924 Except on systems with /dev/urandom, it is crucial to have a random
7925 seed file at least for key creation, DSA signing, and for DH exchanges;
7926 for RSA signatures we could do without one.
7928 gendh and gendsa (unlike genrsa) used to read only the first byte
7929 of each file listed in the '-rand' option. The function as previously
7930 found in genrsa is now in app_rand.c and is used by all programs
7931 that support '-rand'.
7934 *) In RAND_write_file, use mode 0600 for creating files;
7935 don't just chmod when it may be too late.
7938 *) Report an error from X509_STORE_load_locations
7939 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
7942 *) New function ASN1_mbstring_copy() this copies a string in either
7943 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
7944 into an ASN1_STRING type. A mask of permissible types is passed
7945 and it chooses the "minimal" type to use or an error if not type
7949 *) Add function equivalents to the various macros in asn1.h. The old
7950 macros are retained with an M_ prefix. Code inside the library can
7951 use the M_ macros. External code (including the openssl utility)
7952 should *NOT* in order to be "shared library friendly".
7955 *) Add various functions that can check a certificate's extensions
7956 to see if it usable for various purposes such as SSL client,
7957 server or S/MIME and CAs of these types. This is currently
7958 VERY EXPERIMENTAL but will ultimately be used for certificate chain
7959 verification. Also added a -purpose flag to x509 utility to
7960 print out all the purposes.
7963 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
7967 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
7968 for, obtain and decode and extension and obtain its critical flag.
7969 This allows all the necessary extension code to be handled in a
7970 single function call.
7973 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
7974 platforms. See crypto/rc4/rc4_enc.c for further details.
7977 *) New -noout option to asn1parse. This causes no output to be produced
7978 its main use is when combined with -strparse and -out to extract data
7979 from a file (which may not be in ASN.1 format).
7982 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
7983 when producing the local key id.
7984 [Richard Levitte <levitte@stacken.kth.se>]
7986 *) New option -dhparam in s_server. This allows a DH parameter file to be
7987 stated explicitly. If it is not stated then it tries the first server
7988 certificate file. The previous behaviour hard coded the filename
7992 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
7993 a public key to be input or output. For example:
7994 openssl rsa -in key.pem -pubout -out pubkey.pem
7995 Also added necessary DSA public key functions to handle this.
7998 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
7999 in the message. This was handled by allowing
8000 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8001 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8003 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8004 to the end of the strings whereas this didn't. This would cause problems
8005 if strings read with d2i_ASN1_bytes() were later modified.
8006 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8008 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8009 data and it contains EOF it will end up returning an error. This is
8010 caused by input 46 bytes long. The cause is due to the way base64
8011 BIOs find the start of base64 encoded data. They do this by trying a
8012 trial decode on each line until they find one that works. When they
8013 do a flag is set and it starts again knowing it can pass all the
8014 data directly through the decoder. Unfortunately it doesn't reset
8015 the context it uses. This means that if EOF is reached an attempt
8016 is made to pass two EOFs through the context and this causes the
8017 resulting error. This can also cause other problems as well. As is
8018 usual with these problems it takes *ages* to find and the fix is
8019 trivial: move one line.
8020 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8022 *) Ugly workaround to get s_client and s_server working under Windows. The
8023 old code wouldn't work because it needed to select() on sockets and the
8024 tty (for keypresses and to see if data could be written). Win32 only
8025 supports select() on sockets so we select() with a 1s timeout on the
8026 sockets and then see if any characters are waiting to be read, if none
8027 are present then we retry, we also assume we can always write data to
8028 the tty. This isn't nice because the code then blocks until we've
8029 received a complete line of data and it is effectively polling the
8030 keyboard at 1s intervals: however it's quite a bit better than not
8031 working at all :-) A dedicated Windows application might handle this
8032 with an event loop for example.
8035 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8036 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8037 will be called when RSA_sign() and RSA_verify() are used. This is useful
8038 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8039 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8040 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8041 This necessitated the support of an extra signature type NID_md5_sha1
8042 for SSL signatures and modifications to the SSL library to use it instead
8043 of calling RSA_public_decrypt() and RSA_private_encrypt().
8046 *) Add new -verify -CAfile and -CApath options to the crl program, these
8047 will lookup a CRL issuers certificate and verify the signature in a
8048 similar way to the verify program. Tidy up the crl program so it
8049 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8050 less strict. It will now permit CRL extensions even if it is not
8051 a V2 CRL: this will allow it to tolerate some broken CRLs.
8054 *) Initialize all non-automatic variables each time one of the openssl
8055 sub-programs is started (this is necessary as they may be started
8056 multiple times from the "OpenSSL>" prompt).
8057 [Lennart Bang, Bodo Moeller]
8059 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8060 removing all other RSA functionality (this is what NO_RSA does). This
8061 is so (for example) those in the US can disable those operations covered
8062 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8066 *) Non-copying interface to BIO pairs.
8067 (still largely untested)
8070 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8071 ASCII string. This was handled independently in various places before.
8074 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8075 UTF8 strings a character at a time.
8078 *) Use client_version from client hello to select the protocol
8079 (s23_srvr.c) and for RSA client key exchange verification
8080 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8083 *) Add various utility functions to handle SPKACs, these were previously
8084 handled by poking round in the structure internals. Added new function
8085 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8086 print, verify and generate SPKACs. Based on an original idea from
8087 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8090 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8093 *) Allow the config file extension section to be overwritten on the
8094 command line. Based on an original idea from Massimiliano Pala
8095 <madwolf@comune.modena.it>. The new option is called -extensions
8096 and can be applied to ca, req and x509. Also -reqexts to override
8097 the request extensions in req and -crlexts to override the crl extensions
8101 *) Add new feature to the SPKAC handling in ca. Now you can include
8102 the same field multiple times by preceding it by "XXXX." for example:
8105 this is the same syntax as used in the req config file.
8108 *) Allow certificate extensions to be added to certificate requests. These
8109 are specified in a 'req_extensions' option of the req section of the
8110 config file. They can be printed out with the -text option to req but
8111 are otherwise ignored at present.
8114 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8115 data read consists of only the final block it would not decrypted because
8116 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8117 A misplaced 'break' also meant the decrypted final block might not be
8118 copied until the next read.
8121 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8122 a few extra parameters to the DH structure: these will be useful if
8123 for example we want the value of 'q' or implement X9.42 DH.
8126 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8127 provides hooks that allow the default DSA functions or functions on a
8128 "per key" basis to be replaced. This allows hardware acceleration and
8129 hardware key storage to be handled without major modification to the
8130 library. Also added low level modexp hooks and CRYPTO_EX structure and
8131 associated functions.
8134 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8135 as "read only": it can't be written to and the buffer it points to will
8136 not be freed. Reading from a read only BIO is much more efficient than
8137 a normal memory BIO. This was added because there are several times when
8138 an area of memory needs to be read from a BIO. The previous method was
8139 to create a memory BIO and write the data to it, this results in two
8140 copies of the data and an O(n^2) reading algorithm. There is a new
8141 function BIO_new_mem_buf() which creates a read only memory BIO from
8142 an area of memory. Also modified the PKCS#7 routines to use read only
8146 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8147 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8148 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8149 but a retry condition occured while trying to read the rest.
8152 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8153 NID_pkcs7_encrypted by default: this was wrong since this should almost
8154 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8155 the encrypted data type: this is a more sensible place to put it and it
8156 allows the PKCS#12 code to be tidied up that duplicated this
8160 *) Changed obj_dat.pl script so it takes its input and output files on
8161 the command line. This should avoid shell escape redirection problems
8165 *) Initial support for certificate extension requests, these are included
8166 in things like Xenroll certificate requests. Included functions to allow
8167 extensions to be obtained and added.
8170 *) -crlf option to s_client and s_server for sending newlines as
8171 CRLF (as required by many protocols).
8174 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8176 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8177 [Ralf S. Engelschall]
8179 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8180 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8182 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8186 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8187 DH parameters/keys (q is lost during that conversion, but the resulting
8188 DH parameters contain its length).
8190 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8191 much faster than DH_generate_parameters (which creates parameters
8192 where p = 2*q + 1), and also the smaller q makes DH computations
8193 much more efficient (160-bit exponentiation instead of 1024-bit
8194 exponentiation); so this provides a convenient way to support DHE
8195 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8196 utter importance to use
8197 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8199 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8200 when such DH parameters are used, because otherwise small subgroup
8201 attacks may become possible!
8204 *) Avoid memory leak in i2d_DHparams.
8207 *) Allow the -k option to be used more than once in the enc program:
8208 this allows the same encrypted message to be read by multiple recipients.
8211 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8212 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8213 it will always use the numerical form of the OID, even if it has a short
8217 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8218 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8219 otherwise bn_mod_exp was called. In the case of hardware keys for example
8220 no private key components need be present and it might store extra data
8221 in the RSA structure, which cannot be accessed from bn_mod_exp.
8222 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8223 private key operations.
8226 *) Added support for SPARC Linux.
8229 *) pem_password_cb function type incompatibly changed from
8230 typedef int pem_password_cb(char *buf, int size, int rwflag);
8232 ....(char *buf, int size, int rwflag, void *userdata);
8233 so that applications can pass data to their callbacks:
8234 The PEM[_ASN1]_{read,write}... functions and macros now take an
8235 additional void * argument, which is just handed through whenever
8236 the password callback is called.
8237 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8239 New function SSL_CTX_set_default_passwd_cb_userdata.
8241 Compatibility note: As many C implementations push function arguments
8242 onto the stack in reverse order, the new library version is likely to
8243 interoperate with programs that have been compiled with the old
8244 pem_password_cb definition (PEM_whatever takes some data that
8245 happens to be on the stack as its last argument, and the callback
8246 just ignores this garbage); but there is no guarantee whatsoever that
8249 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8250 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8251 problems not only on Windows, but also on some Unix platforms.
8252 To avoid problematic command lines, these definitions are now in an
8253 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8254 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8257 *) MIPS III/IV assembler module is reimplemented.
8260 *) More DES library cleanups: remove references to srand/rand and
8261 delete an unused file.
8264 *) Add support for the the free Netwide assembler (NASM) under Win32,
8265 since not many people have MASM (ml) and it can be hard to obtain.
8266 This is currently experimental but it seems to work OK and pass all
8267 the tests. Check out INSTALL.W32 for info.
8270 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8271 without temporary keys kept an extra copy of the server key,
8272 and connections with temporary keys did not free everything in case
8276 *) New function RSA_check_key and new openssl rsa option -check
8277 for verifying the consistency of RSA keys.
8278 [Ulf Moeller, Bodo Moeller]
8280 *) Various changes to make Win32 compile work:
8281 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8282 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8283 comparison" warnings.
8284 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8287 *) Add a debugging option to PKCS#5 v2 key generation function: when
8288 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8289 derived keys are printed to stderr.
8292 *) Copy the flags in ASN1_STRING_dup().
8293 [Roman E. Pavlov <pre@mo.msk.ru>]
8295 *) The x509 application mishandled signing requests containing DSA
8296 keys when the signing key was also DSA and the parameters didn't match.
8298 It was supposed to omit the parameters when they matched the signing key:
8299 the verifying software was then supposed to automatically use the CA's
8300 parameters if they were absent from the end user certificate.
8302 Omitting parameters is no longer recommended. The test was also
8303 the wrong way round! This was probably due to unusual behaviour in
8304 EVP_cmp_parameters() which returns 1 if the parameters match.
8305 This meant that parameters were omitted when they *didn't* match and
8306 the certificate was useless. Certificates signed with 'ca' didn't have
8308 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8310 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8311 The interface is as follows:
8312 Applications can use
8313 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8314 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8315 "off" is now the default.
8316 The library internally uses
8317 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8318 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8319 to disable memory-checking temporarily.
8321 Some inconsistent states that previously were possible (and were
8322 even the default) are now avoided.
8324 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8325 with each memory chunk allocated; this is occasionally more helpful
8326 than just having a counter.
8328 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8330 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8334 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8335 which largely parallels "options", but is for changing API behaviour,
8336 whereas "options" are about protocol behaviour.
8337 Initial "mode" flags are:
8339 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8340 a single record has been written.
8341 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8342 retries use the same buffer location.
8343 (But all of the contents must be
8347 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8350 *) Fix problems with no-hmac etc.
8351 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8353 *) New functions RSA_get_default_method(), RSA_set_method() and
8354 RSA_get_method(). These allows replacement of RSA_METHODs without having
8355 to mess around with the internals of an RSA structure.
8358 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8359 Also really enable memory leak checks in openssl.c and in some
8361 [Chad C. Mulligan, Bodo Moeller]
8363 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8364 up the length of negative integers. This has now been simplified to just
8365 store the length when it is first determined and use it later, rather
8366 than trying to keep track of where data is copied and updating it to
8368 [Steve Henson, reported by Brien Wheeler
8369 <bwheeler@authentica-security.com>]
8371 *) Add a new function PKCS7_signatureVerify. This allows the verification
8372 of a PKCS#7 signature but with the signing certificate passed to the
8373 function itself. This contrasts with PKCS7_dataVerify which assumes the
8374 certificate is present in the PKCS#7 structure. This isn't always the
8375 case: certificates can be omitted from a PKCS#7 structure and be
8376 distributed by "out of band" means (such as a certificate database).
8379 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8380 function prototypes in pem.h, also change util/mkdef.pl to add the
8381 necessary function names.
8384 *) mk1mf.pl (used by Windows builds) did not properly read the
8385 options set by Configure in the top level Makefile, and Configure
8386 was not even able to write more than one option correctly.
8387 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8390 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8391 file to be loaded from a BIO or FILE pointer. The BIO version will
8392 for example allow memory BIOs to contain config info.
8395 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8396 Whoever hopes to achieve shared-library compatibility across versions
8397 must use this, not the compile-time macro.
8398 (Exercise 0.9.4: Which is the minimum library version required by
8400 Note: All this applies only to multi-threaded programs, others don't
8404 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8405 through a BIO pair triggered the default case, i.e.
8406 SSLerr(...,SSL_R_UNKNOWN_STATE).
8409 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8410 can use the SSL library even if none of the specific BIOs is
8414 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8415 for the encoded length.
8416 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8418 *) Add initial documentation of the X509V3 functions.
8421 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8422 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8423 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8424 secure PKCS#8 private key format with a high iteration count.
8427 *) Fix determination of Perl interpreter: A perl or perl5
8428 _directory_ in $PATH was also accepted as the interpreter.
8429 [Ralf S. Engelschall]
8431 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8432 wrong with it but it was very old and did things like calling
8433 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8437 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8438 to use the new extension code.
8441 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8442 with macros. This should make it easier to change their form, add extra
8443 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8447 *) Add to configuration table a new entry that can specify an alternative
8448 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8449 according to Mark Crispin <MRC@Panda.COM>.
8453 *) DES CBC did not update the IV. Weird.
8456 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8457 Changing the behaviour of the former might break existing programs --
8458 where IV updating is needed, des_ncbc_encrypt can be used.
8461 *) When bntest is run from "make test" it drives bc to check its
8462 calculations, as well as internally checking them. If an internal check
8463 fails, it needs to cause bc to give a non-zero result or make test carries
8464 on without noticing the failure. Fixed.
8467 *) DES library cleanups.
8470 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8471 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8472 ciphers. NOTE: although the key derivation function has been verified
8473 against some published test vectors it has not been extensively tested
8474 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8478 *) Instead of "mkdir -p", which is not fully portable, use new
8479 Perl script "util/mkdir-p.pl".
8482 *) Rewrite the way password based encryption (PBE) is handled. It used to
8483 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8484 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8485 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8486 the 'parameter' field of the AlgorithmIdentifier is passed to the
8487 underlying key generation function so it must do its own ASN1 parsing.
8488 This has also changed the EVP_PBE_CipherInit() function which now has a
8489 'parameter' argument instead of literal salt and iteration count values
8490 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8493 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8494 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8495 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8496 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8497 value was just used as a "magic string" and not used directly its
8498 value doesn't matter.
8501 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8505 *) "linux-sparc64" configuration (ultrapenguin).
8506 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8507 "linux-sparc" configuration.
8508 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8510 *) config now generates no-xxx options for missing ciphers.
8513 *) Support the EBCDIC character set (work in progress).
8514 File ebcdic.c not yet included because it has a different license.
8515 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8517 *) Support BS2000/OSD-POSIX.
8518 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8520 *) Make callbacks for key generation use void * instead of char *.
8523 *) Make S/MIME samples compile (not yet tested).
8526 *) Additional typesafe stacks.
8529 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8533 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8535 *) New configuration variant "sco5-gcc".
8537 *) Updated some demos.
8538 [Sean O Riordain, Wade Scholine]
8540 *) Add missing BIO_free at exit of pkcs12 application.
8543 *) Fix memory leak in conf.c.
8546 *) Updates for Win32 to assembler version of MD5.
8549 *) Set #! path to perl in apps/der_chop to where we found it
8550 instead of using a fixed path.
8553 *) SHA library changes for irix64-mips4-cc.
8556 *) Improvements for VMS support.
8560 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8562 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8563 This also avoids the problems with SC4.2 and unpatched SC5.
8564 [Andy Polyakov <appro@fy.chalmers.se>]
8566 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8567 These are required because of the typesafe stack would otherwise break
8568 existing code. If old code used a structure member which used to be STACK
8569 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8570 sk_num or sk_value it would produce an error because the num, data members
8571 are not present in STACK_OF. Now it just produces a warning. sk_set
8572 replaces the old method of assigning a value to sk_value
8573 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8574 that does this will no longer work (and should use sk_set instead) but
8575 this could be regarded as a "questionable" behaviour anyway.
8578 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8579 correctly handle encrypted S/MIME data.
8582 *) Change type of various DES function arguments from des_cblock
8583 (which means, in function argument declarations, pointer to char)
8584 to des_cblock * (meaning pointer to array with 8 char elements),
8585 which allows the compiler to do more typechecking; it was like
8586 that back in SSLeay, but with lots of ugly casts.
8588 Introduce new type const_des_cblock.
8591 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8592 problems: find RecipientInfo structure that matches recipient certificate
8593 and initialise the ASN1 structures properly based on passed cipher.
8596 *) Belatedly make the BN tests actually check the results.
8599 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8600 to and from BNs: it was completely broken. New compilation option
8601 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8602 key elements as negative integers.
8605 *) Reorganize and speed up MD5.
8606 [Andy Polyakov <appro@fy.chalmers.se>]
8609 [Richard Levitte <richard@levitte.org>]
8611 *) New option -out to asn1parse to allow the parsed structure to be
8612 output to a file. This is most useful when combined with the -strparse
8613 option to examine the output of things like OCTET STRINGS.
8616 *) Make SSL library a little more fool-proof by not requiring any longer
8617 that SSL_set_{accept,connect}_state be called before
8618 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8619 in many applications because usually everything *appeared* to work as
8620 intended anyway -- now it really works as intended).
8623 *) Move openssl.cnf out of lib/.
8626 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8627 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8628 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8629 [Ralf S. Engelschall]
8631 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8632 handle PKCS#7 enveloped data properly.
8633 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8635 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8636 copying pointers. The cert_st handling is changed by this in
8637 various ways (and thus what used to be known as ctx->default_cert
8638 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8639 any longer when s->cert does not give us what we need).
8640 ssl_cert_instantiate becomes obsolete by this change.
8641 As soon as we've got the new code right (possibly it already is?),
8642 we have solved a couple of bugs of the earlier code where s->cert
8643 was used as if it could not have been shared with other SSL structures.
8645 Note that using the SSL API in certain dirty ways now will result
8646 in different behaviour than observed with earlier library versions:
8647 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8648 does not influence s as it used to.
8650 In order to clean up things more thoroughly, inside SSL_SESSION
8651 we don't use CERT any longer, but a new structure SESS_CERT
8652 that holds per-session data (if available); currently, this is
8653 the peer's certificate chain and, for clients, the server's certificate
8654 and temporary key. CERT holds only those values that can have
8655 meaningful defaults in an SSL_CTX.
8658 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8659 from the internal representation. Various PKCS#7 fixes: remove some
8660 evil casts and set the enc_dig_alg field properly based on the signing
8664 *) Allow PKCS#12 password to be set from the command line or the
8665 environment. Let 'ca' get its config file name from the environment
8666 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8670 *) Allow certificate policies extension to use an IA5STRING for the
8671 organization field. This is contrary to the PKIX definition but
8672 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8676 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8677 without disallowing inline assembler and the like for non-pedantic builds.
8680 *) Support Borland C++ builder.
8681 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8686 *) SHA-1 cleanups and performance enhancements.
8687 [Andy Polyakov <appro@fy.chalmers.se>]
8689 *) Sparc v8plus assembler for the bignum library.
8690 [Andy Polyakov <appro@fy.chalmers.se>]
8692 *) Accept any -xxx and +xxx compiler options in Configure.
8695 *) Update HPUX configuration.
8698 *) Add missing sk_<type>_unshift() function to safestack.h
8699 [Ralf S. Engelschall]
8701 *) New function SSL_CTX_use_certificate_chain_file that sets the
8702 "extra_cert"s in addition to the certificate. (This makes sense
8703 only for "PEM" format files, as chains as a whole are not
8707 *) Support verify_depth from the SSL API.
8708 x509_vfy.c had what can be considered an off-by-one-error:
8709 Its depth (which was not part of the external interface)
8710 was actually counting the number of certificates in a chain;
8711 now it really counts the depth.
8714 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8715 instead of X509err, which often resulted in confusing error
8716 messages since the error codes are not globally unique
8717 (e.g. an alleged error in ssl3_accept when a certificate
8718 didn't match the private key).
8720 *) New function SSL_CTX_set_session_id_context that allows to set a default
8721 value (so that you don't need SSL_set_session_id_context for each
8722 connection using the SSL_CTX).
8725 *) OAEP decoding bug fix.
8728 *) Support INSTALL_PREFIX for package builders, as proposed by
8732 *) New Configure options "threads" and "no-threads". For systems
8733 where the proper compiler options are known (currently Solaris
8734 and Linux), "threads" is the default.
8737 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8740 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8741 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8742 such as /usr/local/bin.
8745 *) "make linux-shared" to build shared libraries.
8746 [Niels Poppe <niels@netbox.org>]
8748 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8751 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8752 extension adding in x509 utility.
8755 *) Remove NOPROTO sections and error code comments.
8758 *) Partial rewrite of the DEF file generator to now parse the ANSI
8762 *) New Configure options --prefix=DIR and --openssldir=DIR.
8765 *) Complete rewrite of the error code script(s). It is all now handled
8766 by one script at the top level which handles error code gathering,
8767 header rewriting and C source file generation. It should be much better
8768 than the old method: it now uses a modified version of Ulf's parser to
8769 read the ANSI prototypes in all header files (thus the old K&R definitions
8770 aren't needed for error creation any more) and do a better job of
8771 translating function codes into names. The old 'ASN1 error code imbedded
8772 in a comment' is no longer necessary and it doesn't use .err files which
8773 have now been deleted. Also the error code call doesn't have to appear all
8774 on one line (which resulted in some large lines...).
8777 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8780 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8781 0 (which usually indicates a closed connection), but continue reading.
8784 *) Fix some race conditions.
8787 *) Add support for CRL distribution points extension. Add Certificate
8788 Policies and CRL distribution points documentation.
8791 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8794 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8795 8 of keying material. Merlin has also confirmed interop with this fix
8796 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8797 [Merlin Hughes <merlin@baltimore.ie>]
8799 *) Fix lots of warnings.
8800 [Richard Levitte <levitte@stacken.kth.se>]
8802 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8803 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8804 [Richard Levitte <levitte@stacken.kth.se>]
8806 *) Fix problems with sizeof(long) == 8.
8807 [Andy Polyakov <appro@fy.chalmers.se>]
8809 *) Change functions to ANSI C.
8812 *) Fix typos in error codes.
8813 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8815 *) Remove defunct assembler files from Configure.
8818 *) SPARC v8 assembler BIGNUM implementation.
8819 [Andy Polyakov <appro@fy.chalmers.se>]
8821 *) Support for Certificate Policies extension: both print and set.
8822 Various additions to support the r2i method this uses.
8825 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8826 return a const string when you are expecting an allocated buffer.
8829 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8830 types DirectoryString and DisplayText.
8833 *) Add code to allow r2i extensions to access the configuration database,
8834 add an LHASH database driver and add several ctx helper functions.
8837 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8838 fail when they extended the size of a BIGNUM.
8841 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8842 support typesafe stack.
8845 *) Fix typo in SSL_[gs]et_options().
8846 [Nils Frostberg <nils@medcom.se>]
8848 *) Delete various functions and files that belonged to the (now obsolete)
8849 old X509V3 handling code.
8852 *) New Configure option "rsaref".
8855 *) Don't auto-generate pem.h.
8858 *) Introduce type-safe ASN.1 SETs.
8861 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8862 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8864 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8865 that links with OpenSSL (well at least cause lots of warnings), but fear
8866 not: the conversion is trivial, and it eliminates loads of evil casts. A
8867 few STACKed things have been converted already. Feel free to convert more.
8868 In the fullness of time, I'll do away with the STACK type altogether.
8871 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8872 specified in <certfile> by updating the entry in the index.txt file.
8873 This way one no longer has to edit the index.txt file manually for
8874 revoking a certificate. The -revoke option does the gory details now.
8875 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
8877 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
8878 `-text' option at all and this way the `-noout -text' combination was
8879 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
8880 [Ralf S. Engelschall]
8882 *) Make sure a corresponding plain text error message exists for the
8883 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
8884 verify callback function determined that a certificate was revoked.
8885 [Ralf S. Engelschall]
8887 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
8888 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
8889 all available cipers including rc5, which was forgotten until now.
8890 In order to let the testing shell script know which algorithms
8891 are available, a new (up to now undocumented) command
8892 "openssl list-cipher-commands" is used.
8895 *) Bugfix: s_client occasionally would sleep in select() when
8896 it should have checked SSL_pending() first.
8899 *) New functions DSA_do_sign and DSA_do_verify to provide access to
8900 the raw DSA values prior to ASN.1 encoding.
8903 *) Tweaks to Configure
8904 [Niels Poppe <niels@netbox.org>]
8906 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
8910 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
8913 *) New config option to avoid instructions that are illegal on the 80386.
8914 The default code is faster, but requires at least a 486.
8917 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
8918 SSL2_SERVER_VERSION (not used at all) macros, which are now the
8919 same as SSL2_VERSION anyway.
8922 *) New "-showcerts" option for s_client.
8925 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
8926 application. Various cleanups and fixes.
8929 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
8930 modify error routines to work internally. Add error codes and PBE init
8931 to library startup routines.
8934 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
8935 packing functions to asn1 and evp. Changed function names and error
8936 codes along the way.
8939 *) PKCS12 integration: and so it begins... First of several patches to
8940 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
8941 objects to objects.h
8944 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
8945 and display support for Thawte strong extranet extension.
8948 *) Add LinuxPPC support.
8949 [Jeff Dubrule <igor@pobox.org>]
8951 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
8952 bn_div_words in alpha.s.
8953 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
8955 *) Make sure the RSA OAEP test is skipped under -DRSAref because
8956 OAEP isn't supported when OpenSSL is built with RSAref.
8957 [Ulf Moeller <ulf@fitug.de>]
8959 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
8960 so they no longer are missing under -DNOPROTO.
8961 [Soren S. Jorvang <soren@t.dk>]
8964 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
8966 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
8967 doesn't work when the session is reused. Coming soon!
8970 *) Fix a security hole, that allows sessions to be reused in the wrong
8971 context thus bypassing client cert protection! All software that uses
8972 client certs and session caches in multiple contexts NEEDS PATCHING to
8973 allow session reuse! A fuller solution is in the works.
8974 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
8976 *) Some more source tree cleanups (removed obsolete files
8977 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
8978 permission on "config" script to be executable) and a fix for the INSTALL
8980 [Ulf Moeller <ulf@fitug.de>]
8982 *) Remove some legacy and erroneous uses of malloc, free instead of
8984 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
8986 *) Make rsa_oaep_test return non-zero on error.
8987 [Ulf Moeller <ulf@fitug.de>]
8989 *) Add support for native Solaris shared libraries. Configure
8990 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
8991 if someone would make that last step automatic.
8992 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
8994 *) ctx_size was not built with the right compiler during "make links". Fixed.
8997 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
8998 except NULL ciphers". This means the default cipher list will no longer
8999 enable NULL ciphers. They need to be specifically enabled e.g. with
9000 the string "DEFAULT:eNULL".
9003 *) Fix to RSA private encryption routines: if p < q then it would
9004 occasionally produce an invalid result. This will only happen with
9005 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9008 *) Be less restrictive and allow also `perl util/perlpath.pl
9009 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9010 because this way one can also use an interpreter named `perl5' (which is
9011 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9012 installed as `perl').
9013 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9015 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9016 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9018 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9019 advapi32.lib to Win32 build and change the pem test comparision
9020 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9021 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9022 and crypto/des/ede_cbcm_enc.c.
9025 *) DES quad checksum was broken on big-endian architectures. Fixed.
9028 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9029 Win32 test batch file so it (might) work again. The Win32 test batch file
9030 is horrible: I feel ill....
9033 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9034 in e_os.h. Audit of header files to check ANSI and non ANSI
9035 sections: 10 functions were absent from non ANSI section and not exported
9036 from Windows DLLs. Fixed up libeay.num for new functions.
9039 *) Make `openssl version' output lines consistent.
9040 [Ralf S. Engelschall]
9042 *) Fix Win32 symbol export lists for BIO functions: Added
9043 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9044 to ms/libeay{16,32}.def.
9045 [Ralf S. Engelschall]
9047 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9048 fine under Unix and passes some trivial tests I've now added. But the
9049 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9050 added to make sure no one expects that this stuff really works in the
9051 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9052 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9054 [Ralf S. Engelschall]
9056 *) Fix the generation of two part addresses in perl.
9057 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9059 *) Add config entry for Linux on MIPS.
9060 [John Tobey <jtobey@channel1.com>]
9062 *) Make links whenever Configure is run, unless we are on Windoze.
9065 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9066 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9070 *) Add a useful kludge to allow package maintainers to specify compiler and
9071 other platforms details on the command line without having to patch the
9072 Configure script everytime: One now can use ``perl Configure
9073 <id>:<details>'', i.e. platform ids are allowed to have details appended
9074 to them (seperated by colons). This is treated as there would be a static
9075 pre-configured entry in Configure's %table under key <id> with value
9076 <details> and ``perl Configure <id>'' is called. So, when you want to
9077 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9078 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9079 now, which overrides the FreeBSD-elf entry on-the-fly.
9080 [Ralf S. Engelschall]
9082 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9085 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9086 on the `perl Configure ...' command line. This way one can compile
9087 OpenSSL libraries with Position Independent Code (PIC) which is needed
9088 for linking it into DSOs.
9089 [Ralf S. Engelschall]
9091 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9095 *) Cleaned up the LICENSE document: The official contact for any license
9096 questions now is the OpenSSL core team under openssl-core@openssl.org.
9097 And add a paragraph about the dual-license situation to make sure people
9098 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9099 to the OpenSSL toolkit.
9100 [Ralf S. Engelschall]
9102 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9103 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9104 Additonally cleaned up the `make links' target: Remove unnecessary
9105 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9106 to speed processing and no longer clutter the display with confusing
9107 stuff. Instead only the actually done links are displayed.
9108 [Ralf S. Engelschall]
9110 *) Permit null encryption ciphersuites, used for authentication only. It used
9111 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9112 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9116 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9117 signed attributes when verifying signatures (this would break them),
9118 the detached data encoding was wrong and public keys obtained using
9119 X509_get_pubkey() weren't freed.
9122 *) Add text documentation for the BUFFER functions. Also added a work around
9123 to a Win95 console bug. This was triggered by the password read stuff: the
9124 last character typed gets carried over to the next fread(). If you were
9125 generating a new cert request using 'req' for example then the last
9126 character of the passphrase would be CR which would then enter the first
9130 *) Added the new `Includes OpenSSL Cryptography Software' button as
9131 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9132 button and can be used by applications based on OpenSSL to show the
9133 relationship to the OpenSSL project.
9134 [Ralf S. Engelschall]
9136 *) Remove confusing variables in function signatures in files
9137 ssl/ssl_lib.c and ssl/ssl.h.
9138 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9140 *) Don't install bss_file.c under PREFIX/include/
9141 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9143 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9144 functions that return function pointers and has support for NT specific
9145 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9146 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9147 unsigned to signed types: this was killing the Win32 compile.
9150 *) Add new certificate file to stack functions,
9151 SSL_add_dir_cert_subjects_to_stack() and
9152 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9153 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9154 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9155 This means that Apache-SSL and similar packages don't have to mess around
9156 to add as many CAs as they want to the preferred list.
9159 *) Experiment with doxygen documentation. Currently only partially applied to
9161 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9162 openssl.doxy as the configuration file.
9165 *) Get rid of remaining C++-style comments which strict C compilers hate.
9166 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9168 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9169 compiled in by default: it has problems with large keys.
9172 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9173 DH private keys and/or callback functions which directly correspond to
9174 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9175 is needed for applications which have to configure certificates on a
9176 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9178 For the RSA certificate situation is makes no difference, but
9179 for the DSA certificate situation this fixes the "no shared cipher"
9180 problem where the OpenSSL cipher selection procedure failed because the
9181 temporary keys were not overtaken from the context and the API provided
9182 no way to reconfigure them.
9183 The new functions now let applications reconfigure the stuff and they
9184 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9185 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9186 non-public-API function ssl_cert_instantiate() is used as a helper
9187 function and also to reduce code redundancy inside ssl_rsa.c.
9188 [Ralf S. Engelschall]
9190 *) Move s_server -dcert and -dkey options out of the undocumented feature
9191 area because they are useful for the DSA situation and should be
9192 recognized by the users.
9193 [Ralf S. Engelschall]
9195 *) Fix the cipher decision scheme for export ciphers: the export bits are
9196 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9197 SSL_EXP_MASK. So, the original variable has to be used instead of the
9198 already masked variable.
9199 [Richard Levitte <levitte@stacken.kth.se>]
9201 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9202 [Richard Levitte <levitte@stacken.kth.se>]
9204 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9205 from `int' to `unsigned int' because it's a length and initialized by
9206 EVP_DigestFinal() which expects an `unsigned int *'.
9207 [Richard Levitte <levitte@stacken.kth.se>]
9209 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9210 script. Instead use the usual Shell->Perl transition trick.
9211 [Ralf S. Engelschall]
9213 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9214 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9215 -noout -modulus' as it's already the case for `openssl rsa -noout
9216 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9217 currently the public key is printed (a decision which was already done by
9218 `openssl dsa -modulus' in the past) which serves a similar purpose.
9219 Additionally the NO_RSA no longer completely removes the whole -modulus
9220 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9222 [Ralf S. Engelschall]
9224 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9225 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9226 [Arne Ansper <arne@ats.cyber.ee>]
9228 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9229 to be added. Now both 'req' and 'ca' can use new objects defined in the
9233 *) Add cool BIO that does syslog (or event log on NT).
9234 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9236 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9237 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9238 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9239 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9242 *) Add preliminary config info for new extension code.
9245 *) Make RSA_NO_PADDING really use no padding.
9246 [Ulf Moeller <ulf@fitug.de>]
9248 *) Generate errors when private/public key check is done.
9251 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9252 for some CRL extensions and new objects added.
9255 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9256 key usage extension and fuller support for authority key id.
9259 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9260 padding method for RSA, which is recommended for new applications in PKCS
9261 #1 v2.0 (RFC 2437, October 1998).
9262 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9263 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9264 against Bleichbacher's attack on RSA.
9265 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9268 *) Updates to the new SSL compression code
9269 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9271 *) Fix so that the version number in the master secret, when passed
9272 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9273 (because the server will not accept higher), that the version number
9274 is 0x03,0x01, not 0x03,0x00
9275 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9277 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9278 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9279 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9282 *) Support for RAW extensions where an arbitrary extension can be
9283 created by including its DER encoding. See apps/openssl.cnf for
9287 *) Make sure latest Perl versions don't interpret some generated C array
9288 code as Perl array code in the crypto/err/err_genc.pl script.
9289 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9291 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9292 not many people have the assembler. Various Win32 compilation fixes and
9293 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9297 *) Modify configure script 'Configure' to automatically create crypto/date.h
9298 file under Win32 and also build pem.h from pem.org. New script
9299 util/mkfiles.pl to create the MINFO file on environments that can't do a
9300 'make files': perl util/mkfiles.pl >MINFO should work.
9303 *) Major rework of DES function declarations, in the pursuit of correctness
9304 and purity. As a result, many evil casts evaporated, and some weirdness,
9305 too. You may find this causes warnings in your code. Zapping your evil
9306 casts will probably fix them. Mostly.
9309 *) Fix for a typo in asn1.h. Bug fix to object creation script
9310 obj_dat.pl. It considered a zero in an object definition to mean
9311 "end of object": none of the objects in objects.h have any zeros
9312 so it wasn't spotted.
9313 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9315 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9316 Masking (CBCM). In the absence of test vectors, the best I have been able
9317 to do is check that the decrypt undoes the encrypt, so far. Send me test
9318 vectors if you have them.
9321 *) Correct calculation of key length for export ciphers (too much space was
9322 allocated for null ciphers). This has not been tested!
9325 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9326 message is now correct (it understands "crypto" and "ssl" on its
9327 command line). There is also now an "update" option. This will update
9328 the util/ssleay.num and util/libeay.num files with any new functions.
9330 perl util/mkdef.pl crypto ssl update
9331 it will update them.
9334 *) Overhauled the Perl interface (perl/*):
9335 - ported BN stuff to OpenSSL's different BN library
9336 - made the perl/ source tree CVS-aware
9337 - renamed the package from SSLeay to OpenSSL (the files still contain
9338 their history because I've copied them in the repository)
9339 - removed obsolete files (the test scripts will be replaced
9340 by better Test::Harness variants in the future)
9341 [Ralf S. Engelschall]
9343 *) First cut for a very conservative source tree cleanup:
9344 1. merge various obsolete readme texts into doc/ssleay.txt
9345 where we collect the old documents and readme texts.
9346 2. remove the first part of files where I'm already sure that we no
9347 longer need them because of three reasons: either they are just temporary
9348 files which were left by Eric or they are preserved original files where
9349 I've verified that the diff is also available in the CVS via "cvs diff
9350 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9351 the crypto/md/ stuff).
9352 [Ralf S. Engelschall]
9354 *) More extension code. Incomplete support for subject and issuer alt
9355 name, issuer and authority key id. Change the i2v function parameters
9356 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9357 what that's for :-) Fix to ASN1 macro which messed up
9358 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9361 *) Preliminary support for ENUMERATED type. This is largely copied from the
9365 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9366 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9368 *) Make sure `make rehash' target really finds the `openssl' program.
9369 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9371 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9372 like to hear about it if this slows down other processors.
9375 *) Add CygWin32 platform information to Configure script.
9376 [Alan Batie <batie@aahz.jf.intel.com>]
9378 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9379 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9381 *) New program nseq to manipulate netscape certificate sequences
9384 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9388 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9389 but the BN code had some problems that would cause failures when
9390 doing certificate verification and some other functions.
9391 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9393 *) Add ASN1 and PEM code to support netscape certificate sequences.
9396 *) Add ASN1 and PEM code to support netscape certificate sequences.
9399 *) Add several PKIX and private extended key usage OIDs.
9402 *) Modify the 'ca' program to handle the new extension code. Modify
9403 openssl.cnf for new extension format, add comments.
9406 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9407 and add a sample to openssl.cnf so req -x509 now adds appropriate
9411 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9412 error code, add initial support to X509_print() and x509 application.
9415 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9416 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9417 stuff is currently isolated and isn't even compiled yet.
9420 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9421 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9422 Removed the versions check from X509 routines when loading extensions:
9423 this allows certain broken certificates that don't set the version
9424 properly to be processed.
9427 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9428 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9429 can still be regenerated with "make depend".
9432 *) Spelling mistake in C version of CAST-128.
9433 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9435 *) Changes to the error generation code. The perl script err-code.pl
9436 now reads in the old error codes and retains the old numbers, only
9437 adding new ones if necessary. It also only changes the .err files if new
9438 codes are added. The makefiles have been modified to only insert errors
9439 when needed (to avoid needlessly modifying header files). This is done
9440 by only inserting errors if the .err file is newer than the auto generated
9441 C file. To rebuild all the error codes from scratch (the old behaviour)
9442 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9443 or delete all the .err files.
9446 *) CAST-128 was incorrectly implemented for short keys. The C version has
9447 been fixed, but is untested. The assembler versions are also fixed, but
9448 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9449 to regenerate it if needed.
9450 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9451 Hagino <itojun@kame.net>]
9453 *) File was opened incorrectly in randfile.c.
9454 [Ulf Möller <ulf@fitug.de>]
9456 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9457 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9458 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9459 al: it's just almost always a UTCTime. Note this patch adds new error
9460 codes so do a "make errors" if there are problems.
9463 *) Correct Linux 1 recognition in config.
9464 [Ulf Möller <ulf@fitug.de>]
9466 *) Remove pointless MD5 hash when using DSA keys in ca.
9467 [Anonymous <nobody@replay.com>]
9469 *) Generate an error if given an empty string as a cert directory. Also
9470 generate an error if handed NULL (previously returned 0 to indicate an
9471 error, but didn't set one).
9472 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9474 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9477 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9478 parameters. This was causing a warning which killed off the Win32 compile.
9481 *) Remove C++ style comments from crypto/bn/bn_local.h.
9482 [Neil Costigan <neil.costigan@celocom.com>]
9484 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9485 based on a text string, looking up short and long names and finally
9486 "dot" format. The "dot" format stuff didn't work. Added new function
9487 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9488 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9489 OID is not part of the table.
9492 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9493 X509_LOOKUP_by_alias().
9496 *) Sort openssl functions by name.
9499 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9500 encryption from sample DSA keys (in case anyone is interested the password
9504 *) Make _all_ *_free functions accept a NULL pointer.
9505 [Frans Heymans <fheymans@isaserver.be>]
9507 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9509 [Anonymous <nobody@replay.com>]
9511 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9512 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9514 *) Don't blow it for numeric -newkey arguments to apps/req.
9515 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9517 *) Temp key "for export" tests were wrong in s3_srvr.c.
9518 [Anonymous <nobody@replay.com>]
9520 *) Add prototype for temp key callback functions
9521 SSL_CTX_set_tmp_{rsa,dh}_callback().
9524 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9525 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9528 *) X509_name_add_entry() freed the wrong thing after an error.
9529 [Arne Ansper <arne@ats.cyber.ee>]
9531 *) rsa_eay.c would attempt to free a NULL context.
9532 [Arne Ansper <arne@ats.cyber.ee>]
9534 *) BIO_s_socket() had a broken should_retry() on Windoze.
9535 [Arne Ansper <arne@ats.cyber.ee>]
9537 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9538 [Arne Ansper <arne@ats.cyber.ee>]
9540 *) Make sure the already existing X509_STORE->depth variable is initialized
9541 in X509_STORE_new(), but document the fact that this variable is still
9542 unused in the certificate verification process.
9543 [Ralf S. Engelschall]
9545 *) Fix the various library and apps files to free up pkeys obtained from
9546 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9549 *) Fix reference counting in X509_PUBKEY_get(). This makes
9550 demos/maurice/example2.c work, amongst others, probably.
9551 [Steve Henson and Ben Laurie]
9553 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9554 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9555 are no longer created. This way we have a single and consistent command
9556 line interface `openssl <command>', similar to `cvs <command>'.
9557 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9559 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9560 BIT STRING wrapper always have zero unused bits.
9563 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9566 *) Make the top-level INSTALL documentation easier to understand.
9569 *) Makefiles updated to exit if an error occurs in a sub-directory
9570 make (including if user presses ^C) [Paul Sutton]
9572 *) Make Montgomery context stuff explicit in RSA data structure.
9575 *) Fix build order of pem and err to allow for generated pem.h.
9578 *) Fix renumbering bug in X509_NAME_delete_entry().
9581 *) Enhanced the err-ins.pl script so it makes the error library number
9582 global and can add a library name. This is needed for external ASN1 and
9583 other error libraries.
9586 *) Fixed sk_insert which never worked properly.
9589 *) Fix ASN1 macros so they can handle indefinite length construted
9590 EXPLICIT tags. Some non standard certificates use these: they can now
9594 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9595 into a single doc/ssleay.txt bundle. This way the information is still
9596 preserved but no longer messes up this directory. Now it's new room for
9597 the new set of documenation files.
9598 [Ralf S. Engelschall]
9600 *) SETs were incorrectly DER encoded. This was a major pain, because they
9601 shared code with SEQUENCEs, which aren't coded the same. This means that
9602 almost everything to do with SETs or SEQUENCEs has either changed name or
9603 number of arguments.
9604 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9606 *) Fix test data to work with the above.
9609 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9610 was already fixed by Eric for 0.9.1 it seems.
9611 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9613 *) Autodetect FreeBSD3.
9616 *) Fix various bugs in Configure. This affects the following platforms:
9620 unixware-2.0-pentium
9624 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9625 before they are needed.
9628 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9632 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9634 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9635 changed SSLeay to OpenSSL in version strings.
9636 [Ralf S. Engelschall]
9638 *) Some fixups to the top-level documents.
9641 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9642 because the symlink to include/ was missing.
9643 [Ralf S. Engelschall]
9645 *) Incorporated the popular no-RSA/DSA-only patches
9646 which allow to compile a RSA-free SSLeay.
9647 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9649 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9650 when "ssleay" is still not found.
9651 [Ralf S. Engelschall]
9653 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9654 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9656 *) Updated the README file.
9657 [Ralf S. Engelschall]
9659 *) Added various .cvsignore files in the CVS repository subdirs
9660 to make a "cvs update" really silent.
9661 [Ralf S. Engelschall]
9663 *) Recompiled the error-definition header files and added
9664 missing symbols to the Win32 linker tables.
9665 [Ralf S. Engelschall]
9667 *) Cleaned up the top-level documents;
9668 o new files: CHANGES and LICENSE
9669 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9670 o merged COPYRIGHT into LICENSE
9671 o removed obsolete TODO file
9672 o renamed MICROSOFT to INSTALL.W32
9673 [Ralf S. Engelschall]
9675 *) Removed dummy files from the 0.9.1b source tree:
9676 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9677 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9678 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9679 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9680 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9681 [Ralf S. Engelschall]
9683 *) Added various platform portability fixes.
9686 *) The Genesis of the OpenSSL rpject:
9687 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9688 Young and Tim J. Hudson created while they were working for C2Net until
9690 [The OpenSSL Project]
9693 Changes between 0.9.0b and 0.9.1b [not released]
9695 *) Updated a few CA certificates under certs/
9698 *) Changed some BIGNUM api stuff.
9701 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9702 DGUX x86, Linux Alpha, etc.
9705 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9706 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9710 *) Add -strparse option to asn1pars program which parses nested
9712 [Dr Stephen Henson <shenson@bigfoot.com>]
9714 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9717 *) DSA fix for "ca" program.
9720 *) Added "-genkey" option to "dsaparam" program.
9723 *) Added RIPE MD160 (rmd160) message digest.
9726 *) Added -a (all) option to "ssleay version" command.
9729 *) Added PLATFORM define which is the id given to Configure.
9732 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9735 *) Extended the ASN.1 parser routines.
9738 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9741 *) Added a BN_CTX to the BN library.
9744 *) Fixed the weak key values in DES library
9747 *) Changed API in EVP library for cipher aliases.
9750 *) Added support for RC2/64bit cipher.
9753 *) Converted the lhash library to the crypto/mem.c functions.
9756 *) Added more recognized ASN.1 object ids.
9759 *) Added more RSA padding checks for SSL/TLS.
9762 *) Added BIO proxy/filter functionality.
9765 *) Added extra_certs to SSL_CTX which can be used
9766 send extra CA certificates to the client in the CA cert chain sending
9767 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9770 *) Now Fortezza is denied in the authentication phase because
9771 this is key exchange mechanism is not supported by SSLeay at all.
9774 *) Additional PKCS1 checks.
9777 *) Support the string "TLSv1" for all TLS v1 ciphers.
9780 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9781 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9784 *) Fixed a few memory leaks.
9787 *) Fixed various code and comment typos.
9790 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9791 bytes sent in the client random.
9792 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob