5 Changes between 0.9.7g and 0.9.7h [XX xxx XXXX]
7 *) Change the client implementation for SSLv23_method() and
8 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
9 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10 (Previously, the SSL 2.0 backwards compatible Client Hello
11 message format would be used even with SSL_OP_NO_SSLv2.)
14 *) Add support for smime-type MIME parameter in S/MIME messages which some
18 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
19 a threadsafe manner. Modify rsa code to use new function and add calls
20 to dsa and dh code (which had race conditions before).
23 *) Include the fixed error library code in the C error file definitions
24 instead of fixing them up at runtime. This keeps the error code
28 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
30 *) Fixes for newer kerberos headers. NB: the casts are needed because
31 the 'length' field is signed on one version and unsigned on another
32 with no (?) obvious way to tell the difference, without these VC++
33 complains. Also the "definition" of FAR (blank) is no longer included
34 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
35 some needed definitions.
38 *) Undo Cygwin change.
41 *) Added support for proxy certificates according to RFC 3820.
42 Because they may be a security thread to unaware applications,
43 they must be explicitely allowed in run-time. See
44 docs/HOWTO/proxy_certificates.txt for further information.
47 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
49 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
50 server and client random values. Previously
51 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
52 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
54 This change has negligible security impact because:
56 1. Server and client random values still have 24 bytes of pseudo random
59 2. Server and client random values are sent in the clear in the initial
62 3. The master secret is derived using the premaster secret (48 bytes in
63 size for static RSA ciphersuites) as well as client server and random
66 The OpenSSL team would like to thank the UK NISCC for bringing this issue
69 [Stephen Henson, reported by UK NISCC]
71 *) Use Windows randomness collection on Cygwin.
74 *) Fix hang in EGD/PRNGD query when communication socket is closed
75 prematurely by EGD/PRNGD.
76 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
78 *) Prompt for pass phrases when appropriate for PKCS12 input format.
81 *) Back-port of selected performance improvements from development
82 branch, as well as improved support for PowerPC platforms.
85 *) Add lots of checks for memory allocation failure, error codes to indicate
86 failure and freeing up memory if a failure occurs.
87 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
89 *) Add new -passin argument to dgst.
92 *) Perform some character comparisons of different types in X509_NAME_cmp:
93 this is needed for some certificates that reencode DNs into UTF8Strings
94 (in violation of RFC3280) and can't or wont issue name rollover
98 *) Make an explicit check during certificate validation to see that
99 the CA setting in each certificate on the chain is correct. As a
100 side effect always do the following basic checks on extensions,
101 not just when there's an associated purpose to the check:
103 - if there is an unhandled critical extension (unless the user
104 has chosen to ignore this fault)
105 - if the path length has been exceeded (if one is set at all)
106 - that certain extensions fit the associated purpose (if one has
110 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
112 *) Avoid a race condition when CRLs are checked in a multi threaded
113 environment. This would happen due to the reordering of the revoked
114 entries during signature checking and serial number lookup. Now the
115 encoding is cached and the serial number sort performed under a lock.
116 Add new STACK function sk_is_sorted().
119 *) Add Delta CRL to the extension code.
122 *) Various fixes to s3_pkt.c so alerts are sent properly.
123 [David Holmes <d.holmes@f5.com>]
125 *) Reduce the chances of duplicate issuer name and serial numbers (in
126 violation of RFC3280) using the OpenSSL certificate creation utilities.
127 This is done by creating a random 64 bit value for the initial serial
128 number when a serial number file is created or when a self signed
129 certificate is created using 'openssl req -x509'. The initial serial
130 number file is created using 'openssl x509 -next_serial' in CA.pl
131 rather than being initialized to 1.
134 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
136 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
137 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
138 [Joe Orton, Steve Henson]
140 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
142 [Joe Orton, Steve Henson]
144 *) Make it possible to have multiple active certificates with the same
145 subject in the CA index file. This is done only if the keyword
146 'unique_subject' is set to 'no' in the main CA section (default
147 if 'CA_default') of the configuration file. The value is saved
148 with the database itself in a separate index attribute file,
149 named like the index file with '.attr' appended to the name.
152 *) X509 verify fixes. Disable broken certificate workarounds when
153 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
154 keyUsage extension present. Don't accept CRLs with unhandled critical
155 extensions: since verify currently doesn't process CRL extensions this
156 rejects a CRL with *any* critical extensions. Add new verify error codes
160 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
161 A clarification of RFC2560 will require the use of OCTET STRINGs and
162 some implementations cannot handle the current raw format. Since OpenSSL
163 copies and compares OCSP nonces as opaque blobs without any attempt at
164 parsing them this should not create any compatibility issues.
167 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
168 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
169 this HMAC (and other) operations are several times slower than OpenSSL
173 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
174 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
176 *) Use the correct content when signing type "other".
179 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
181 *) Fix various bugs revealed by running the NISCC test suite:
183 Stop out of bounds reads in the ASN1 code when presented with
184 invalid tags (CAN-2003-0543 and CAN-2003-0544).
186 Free up ASN1_TYPE correctly if ANY type is invalid (CAN-2003-0545).
188 If verify callback ignores invalid public key errors don't try to check
189 certificate signature with the NULL public key.
193 *) New -ignore_err option in ocsp application to stop the server
194 exiting on the first error in a request.
197 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
198 if the server requested one: as stated in TLS 1.0 and SSL 3.0
202 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
203 extra data after the compression methods not only for TLS 1.0
204 but also for SSL 3.0 (as required by the specification).
205 [Bodo Moeller; problem pointed out by Matthias Loepfe]
207 *) Change X509_certificate_type() to mark the key as exported/exportable
208 when it's 512 *bits* long, not 512 bytes.
211 *) Change AES_cbc_encrypt() so it outputs exact multiple of
212 blocks during encryption.
215 *) Various fixes to base64 BIO and non blocking I/O. On write
216 flushes were not handled properly if the BIO retried. On read
217 data was not being buffered properly and had various logic bugs.
218 This also affects blocking I/O when the data being decoded is a
222 *) Various S/MIME bugfixes and compatibility changes:
223 output correct application/pkcs7 MIME type if
224 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
225 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
226 of files as .eml work). Correctly handle very long lines in MIME
230 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
232 *) Countermeasure against the Klima-Pokorny-Rosa extension of
233 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
234 a protocol version number mismatch like a decryption error
235 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
238 *) Turn on RSA blinding by default in the default implementation
239 to avoid a timing attack. Applications that don't want it can call
240 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
241 They would be ill-advised to do so in most cases.
242 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
244 *) Change RSA blinding code so that it works when the PRNG is not
245 seeded (in this case, the secret RSA exponent is abused as
246 an unpredictable seed -- if it is not unpredictable, there
247 is no point in blinding anyway). Make RSA blinding thread-safe
248 by remembering the creator's thread ID in rsa->blinding and
249 having all other threads use local one-time blinding factors
250 (this requires more computation than sharing rsa->blinding, but
251 avoids excessive locking; and if an RSA object is not shared
252 between threads, blinding will still be very fast).
255 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
256 ENGINE as defaults for all supported algorithms irrespective of
257 the 'flags' parameter. 'flags' is now honoured, so applications
258 should make sure they are passing it correctly.
261 *) Target "mingw" now allows native Windows code to be generated in
262 the Cygwin environment as well as with the MinGW compiler.
265 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
267 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
268 via timing by performing a MAC computation even if incorrrect
269 block cipher padding has been found. This is a countermeasure
270 against active attacks where the attacker has to distinguish
271 between bad padding and a MAC verification error. (CAN-2003-0078)
273 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
274 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
275 Martin Vuagnoux (EPFL, Ilion)]
277 *) Make the no-err option work as intended. The intention with no-err
278 is not to have the whole error stack handling routines removed from
279 libcrypto, it's only intended to remove all the function name and
280 reason texts, thereby removing some of the footprint that may not
281 be interesting if those errors aren't displayed anyway.
283 NOTE: it's still possible for any application or module to have it's
284 own set of error texts inserted. The routines are there, just not
285 used by default when no-err is given.
288 *) Add support for FreeBSD on IA64.
289 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
291 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
292 Kerberos function mit_des_cbc_cksum(). Before this change,
293 the value returned by DES_cbc_cksum() was like the one from
294 mit_des_cbc_cksum(), except the bytes were swapped.
295 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
297 *) Allow an application to disable the automatic SSL chain building.
298 Before this a rather primitive chain build was always performed in
299 ssl3_output_cert_chain(): an application had no way to send the
300 correct chain if the automatic operation produced an incorrect result.
302 Now the chain builder is disabled if either:
304 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
306 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
308 The reasoning behind this is that an application would not want the
309 auto chain building to take place if extra chain certificates are
310 present and it might also want a means of sending no additional
311 certificates (for example the chain has two certificates and the
315 *) Add the possibility to build without the ENGINE framework.
316 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
318 *) Under Win32 gmtime() can return NULL: check return value in
319 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
322 *) DSA routines: under certain error conditions uninitialized BN objects
323 could be freed. Solution: make sure initialization is performed early
324 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
325 Nils Larsch <nla@trustcenter.de> via PR#459)
328 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
329 checked on reconnect on the client side, therefore session resumption
330 could still fail with a "ssl session id is different" error. This
331 behaviour is masked when SSL_OP_ALL is used due to
332 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
333 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
337 *) IA-32 assembler support enhancements: unified ELF targets, support
338 for SCO/Caldera platforms, fix for Cygwin shared build.
341 *) Add support for FreeBSD on sparc64. As a consequence, support for
342 FreeBSD on non-x86 processors is separate from x86 processors on
343 the config script, much like the NetBSD support.
344 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
346 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
348 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
351 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
352 code (06) was taken as the first octet of the session ID and the last
353 octet was ignored consequently. As a result SSLv2 client side session
354 caching could not have worked due to the session ID mismatch between
356 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
360 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
361 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
365 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
366 seems that in spite of existing for more than a year, many application
367 author have done nothing to provide the necessary callbacks, which
368 means that this particular engine will not work properly anywhere.
369 This is a very unfortunate situation which forces us, in the name
370 of usability, to give the hw_ncipher.c a static lock, which is part
372 NOTE: This is for the 0.9.7 series ONLY. This hack will never
373 appear in 0.9.8 or later. We EXPECT application authors to have
374 dealt properly with this when 0.9.8 is released (unless we actually
375 make such changes in the libcrypto locking code that changes will
376 have to be made anyway).
379 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
380 octets have been read, EOF or an error occurs. Without this change
381 some truncated ASN1 structures will not produce an error.
384 *) Disable Heimdal support, since it hasn't been fully implemented.
385 Still give the possibility to force the use of Heimdal, but with
386 warnings and a request that patches get sent to openssl-dev.
389 *) Add the VC-CE target, introduce the WINCE sysname, and add
390 INSTALL.WCE and appropriate conditionals to make it build.
391 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
393 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
394 cygssl-x.y.z.dll, where x, y and z are the major, minor and
395 edit numbers of the version.
396 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
398 *) Introduce safe string copy and catenation functions
399 (BUF_strlcpy() and BUF_strlcat()).
400 [Ben Laurie (CHATS) and Richard Levitte]
402 *) Avoid using fixed-size buffers for one-line DNs.
405 *) Add BUF_MEM_grow_clean() to avoid information leakage when
406 resizing buffers containing secrets, and use where appropriate.
409 *) Avoid using fixed size buffers for configuration file location.
412 *) Avoid filename truncation for various CA files.
415 *) Use sizeof in preference to magic numbers.
418 *) Avoid filename truncation in cert requests.
421 *) Add assertions to check for (supposedly impossible) buffer
425 *) Don't cache truncated DNS entries in the local cache (this could
426 potentially lead to a spoofing attack).
429 *) Fix various buffers to be large enough for hex/decimal
430 representations in a platform independent manner.
433 *) Add CRYPTO_realloc_clean() to avoid information leakage when
434 resizing buffers containing secrets, and use where appropriate.
437 *) Add BIO_indent() to avoid much slightly worrying code to do
441 *) Convert sprintf()/BIO_puts() to BIO_printf().
444 *) buffer_gets() could terminate with the buffer only half
448 *) Add assertions to prevent user-supplied crypto functions from
449 overflowing internal buffers by having large block sizes, etc.
452 *) New OPENSSL_assert() macro (similar to assert(), but enabled
456 *) Eliminate unused copy of key in RC4.
459 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
462 *) Fix off-by-one error in EGD path.
465 *) If RANDFILE path is too long, ignore instead of truncating.
468 *) Eliminate unused and incorrectly sized X.509 structure
472 *) Eliminate unused and dangerous function knumber().
475 *) Eliminate unused and dangerous structure, KSSL_ERR.
478 *) Protect against overlong session ID context length in an encoded
479 session object. Since these are local, this does not appear to be
483 *) Change from security patch (see 0.9.6e below) that did not affect
484 the 0.9.6 release series:
486 Remote buffer overflow in SSL3 protocol - an attacker could
487 supply an oversized master key in Kerberos-enabled versions.
491 *) Change the SSL kerb5 codes to match RFC 2712.
494 *) Make -nameopt work fully for req and add -reqopt switch.
495 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
497 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
498 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
500 *) Make sure tests can be performed even if the corresponding algorithms
501 have been removed entirely. This was also the last step to make
502 OpenSSL compilable with DJGPP under all reasonable conditions.
503 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
505 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
506 to allow version independent disabling of normally unselected ciphers,
507 which may be activated as a side-effect of selecting a single cipher.
509 (E.g., cipher list string "RSA" enables ciphersuites that are left
510 out of "ALL" because they do not provide symmetric encryption.
511 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
512 [Lutz Jaenicke, Bodo Moeller]
514 *) Add appropriate support for separate platform-dependent build
515 directories. The recommended way to make a platform-dependent
516 build directory is the following (tested on Linux), maybe with
519 # Place yourself outside of the OpenSSL source tree. In
520 # this example, the environment variable OPENSSL_SOURCE
521 # is assumed to contain the absolute OpenSSL source directory.
522 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
523 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
524 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
525 mkdir -p `dirname $F`
526 ln -s $OPENSSL_SOURCE/$F $F
529 To be absolutely sure not to disturb the source tree, a "make clean"
530 is a good thing. If it isn't successfull, don't worry about it,
531 it probably means the source directory is very clean.
534 *) Make sure any ENGINE control commands make local copies of string
535 pointers passed to them whenever necessary. Otherwise it is possible
536 the caller may have overwritten (or deallocated) the original string
537 data when a later ENGINE operation tries to use the stored values.
538 [Götz Babin-Ebell <babinebell@trustcenter.de>]
540 *) Improve diagnostics in file reading and command-line digests.
541 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
543 *) Add AES modes CFB and OFB to the object database. Correct an
544 error in AES-CFB decryption.
547 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
548 allows existing EVP_CIPHER_CTX structures to be reused after
549 calling EVP_*Final(). This behaviour is used by encryption
550 BIOs and some applications. This has the side effect that
551 applications must explicitly clean up cipher contexts with
552 EVP_CIPHER_CTX_cleanup() or they will leak memory.
555 *) Check the values of dna and dnb in bn_mul_recursive before calling
556 bn_mul_comba (a non zero value means the a or b arrays do not contain
557 n2 elements) and fallback to bn_mul_normal if either is not zero.
560 *) Fix escaping of non-ASCII characters when using the -subj option
561 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
564 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
565 form for "surname", serialNumber has no short form.
566 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
567 therefore remove "mail" short name for "internet 7".
568 The OID for unique identifiers in X509 certificates is
569 x500UniqueIdentifier, not uniqueIdentifier.
570 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
573 *) Add an "init" command to the ENGINE config module and auto initialize
574 ENGINEs. Without any "init" command the ENGINE will be initialized
575 after all ctrl commands have been executed on it. If init=1 the
576 ENGINE is initailized at that point (ctrls before that point are run
577 on the uninitialized ENGINE and after on the initialized one). If
578 init=0 then the ENGINE will not be iniatialized at all.
581 *) Fix the 'app_verify_callback' interface so that the user-defined
582 argument is actually passed to the callback: In the
583 SSL_CTX_set_cert_verify_callback() prototype, the callback
584 declaration has been changed from
587 int (*cb)(X509_STORE_CTX *,void *);
588 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
589 i=s->ctx->app_verify_callback(&ctx)
590 has been changed into
591 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
593 To update applications using SSL_CTX_set_cert_verify_callback(),
594 a dummy argument can be added to their callback functions.
595 [D. K. Smetters <smetters@parc.xerox.com>]
597 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
598 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
600 *) Add and OPENSSL_LOAD_CONF define which will cause
601 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
602 This allows older applications to transparently support certain
603 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
604 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
605 load the config file and OPENSSL_add_all_algorithms_conf() which will
606 always load it have also been added.
609 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
610 Adjust NIDs and EVP layer.
611 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
613 *) Config modules support in openssl utility.
615 Most commands now load modules from the config file,
616 though in a few (such as version) this isn't done
617 because it couldn't be used for anything.
619 In the case of ca and req the config file used is
620 the same as the utility itself: that is the -config
621 command line option can be used to specify an
625 *) Move default behaviour from OPENSSL_config(). If appname is NULL
626 use "openssl_conf" if filename is NULL use default openssl config file.
629 *) Add an argument to OPENSSL_config() to allow the use of an alternative
630 config section name. Add a new flag to tolerate a missing config file
631 and move code to CONF_modules_load_file().
634 *) Support for crypto accelerator cards from Accelerated Encryption
635 Processing, www.aep.ie. (Use engine 'aep')
636 The support was copied from 0.9.6c [engine] and adapted/corrected
637 to work with the new engine framework.
638 [AEP Inc. and Richard Levitte]
640 *) Support for SureWare crypto accelerator cards from Baltimore
641 Technologies. (Use engine 'sureware')
642 The support was copied from 0.9.6c [engine] and adapted
643 to work with the new engine framework.
646 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
647 make the newer ENGINE framework commands for the CHIL engine work.
648 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
650 *) Make it possible to produce shared libraries on ReliantUNIX.
651 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
653 *) Add the configuration target debug-linux-ppro.
654 Make 'openssl rsa' use the general key loading routines
655 implemented in apps.c, and make those routines able to
656 handle the key format FORMAT_NETSCAPE and the variant
658 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
660 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
661 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
663 *) Add -keyform to rsautl, and document -engine.
664 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
666 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
667 BIO_R_NO_SUCH_FILE error code rather than the generic
668 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
673 ERR_peek_last_error_line
674 ERR_peek_last_error_line_data.
678 ERR_peek_error_line_data,
679 but report on the latest error recorded rather than the first one
680 still in the error queue.
681 [Ben Laurie, Bodo Moeller]
683 *) default_algorithms option in ENGINE config module. This allows things
685 default_algorithms = ALL
686 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
689 *) Prelminary ENGINE config module.
692 *) New experimental application configuration code.
695 *) Change the AES code to follow the same name structure as all other
696 symmetric ciphers, and behave the same way. Move everything to
697 the directory crypto/aes, thereby obsoleting crypto/rijndael.
698 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
700 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
701 [Ben Laurie and Theo de Raadt]
703 *) Add option to output public keys in req command.
704 [Massimiliano Pala madwolf@openca.org]
706 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
707 (up to about 10% better than before for P-192 and P-224).
710 *) New functions/macros
712 SSL_CTX_set_msg_callback(ctx, cb)
713 SSL_CTX_set_msg_callback_arg(ctx, arg)
714 SSL_set_msg_callback(ssl, cb)
715 SSL_set_msg_callback_arg(ssl, arg)
717 to request calling a callback function
719 void cb(int write_p, int version, int content_type,
720 const void *buf, size_t len, SSL *ssl, void *arg)
722 whenever a protocol message has been completely received
723 (write_p == 0) or sent (write_p == 1). Here 'version' is the
724 protocol version according to which the SSL library interprets
725 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
726 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
727 the content type as defined in the SSL 3.0/TLS 1.0 protocol
728 specification (change_cipher_spec(20), alert(21), handshake(22)).
729 'buf' and 'len' point to the actual message, 'ssl' to the
730 SSL object, and 'arg' is the application-defined value set by
731 SSL[_CTX]_set_msg_callback_arg().
733 'openssl s_client' and 'openssl s_server' have new '-msg' options
734 to enable a callback that displays all protocol messages.
737 *) Change the shared library support so shared libraries are built as
738 soon as the corresponding static library is finished, and thereby get
739 openssl and the test programs linked against the shared library.
740 This still only happens when the keyword "shard" has been given to
741 the configuration scripts.
743 NOTE: shared library support is still an experimental thing, and
744 backward binary compatibility is still not guaranteed.
745 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
747 *) Add support for Subject Information Access extension.
748 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
750 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
751 additional bytes when new memory had to be allocated, not just
752 when reusing an existing buffer.
755 *) New command line and configuration option 'utf8' for the req command.
756 This allows field values to be specified as UTF8 strings.
759 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
760 runs for the former and machine-readable output for the latter.
763 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
764 of the e-mail address in the DN (i.e., it will go into a certificate
765 extension only). The new configuration file option 'email_in_dn = no'
767 [Massimiliano Pala madwolf@openca.org]
769 *) Change all functions with names starting with des_ to be starting
770 with DES_ instead. Add wrappers that are compatible with libdes,
771 but are named _ossl_old_des_*. Finally, add macros that map the
772 des_* symbols to the corresponding _ossl_old_des_* if libdes
773 compatibility is desired. If OpenSSL 0.9.6c compatibility is
774 desired, the des_* symbols will be mapped to DES_*, with one
777 Since we provide two compatibility mappings, the user needs to
778 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
779 compatibility is desired. The default (i.e., when that macro
780 isn't defined) is OpenSSL 0.9.6c compatibility.
782 There are also macros that enable and disable the support of old
783 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
784 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
785 are defined, the default will apply: to support the old des routines.
787 In either case, one must include openssl/des.h to get the correct
788 definitions. Do not try to just include openssl/des_old.h, that
791 NOTE: This is a major break of an old API into a new one. Software
792 authors are encouraged to switch to the DES_ style functions. Some
793 time in the future, des_old.h and the libdes compatibility functions
794 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
795 default), and then completely removed.
798 *) Test for certificates which contain unsupported critical extensions.
799 If such a certificate is found during a verify operation it is
800 rejected by default: this behaviour can be overridden by either
801 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
802 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
803 X509_supported_extension() has also been added which returns 1 if a
804 particular extension is supported.
807 *) Modify the behaviour of EVP cipher functions in similar way to digests
808 to retain compatibility with existing code.
811 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
812 compatibility with existing code. In particular the 'ctx' parameter does
813 not have to be to be initialized before the call to EVP_DigestInit() and
814 it is tidied up after a call to EVP_DigestFinal(). New function
815 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
816 EVP_MD_CTX_copy() changed to not require the destination to be
817 initialized valid and new function EVP_MD_CTX_copy_ex() added which
818 requires the destination to be valid.
820 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
821 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
824 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
825 so that complete 'Handshake' protocol structures are kept in memory
826 instead of overwriting 'msg_type' and 'length' with 'body' data.
829 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
830 [Massimo Santin via Richard Levitte]
832 *) Major restructuring to the underlying ENGINE code. This includes
833 reduction of linker bloat, separation of pure "ENGINE" manipulation
834 (initialisation, etc) from functionality dealing with implementations
835 of specific crypto iterfaces. This change also introduces integrated
836 support for symmetric ciphers and digest implementations - so ENGINEs
837 can now accelerate these by providing EVP_CIPHER and EVP_MD
838 implementations of their own. This is detailed in crypto/engine/README
839 as it couldn't be adequately described here. However, there are a few
840 API changes worth noting - some RSA, DSA, DH, and RAND functions that
841 were changed in the original introduction of ENGINE code have now
842 reverted back - the hooking from this code to ENGINE is now a good
843 deal more passive and at run-time, operations deal directly with
844 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
845 dereferencing through an ENGINE pointer any more. Also, the ENGINE
846 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
847 they were not being used by the framework as there is no concept of a
848 BIGNUM_METHOD and they could not be generalised to the new
849 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
850 ENGINE_cpy() has been removed as it cannot be consistently defined in
854 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
857 *) Change mkdef.pl to sort symbols that get the same entry number,
858 and make sure the automatically generated functions ERR_load_*
859 become part of libeay.num as well.
862 *) New function SSL_renegotiate_pending(). This returns true once
863 renegotiation has been requested (either SSL_renegotiate() call
864 or HelloRequest/ClientHello receveived from the peer) and becomes
865 false once a handshake has been completed.
866 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
867 sends a HelloRequest, but does not ensure that a handshake takes
868 place. SSL_renegotiate_pending() is useful for checking if the
869 client has followed the request.)
872 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
873 By default, clients may request session resumption even during
874 renegotiation (if session ID contexts permit); with this option,
875 session resumption is possible only in the first handshake.
877 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
878 more bits available for options that should not be part of
879 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
882 *) Add some demos for certificate and certificate request creation.
885 *) Make maximum certificate chain size accepted from the peer application
886 settable (SSL*_get/set_max_cert_list()), as proposed by
887 "Douglas E. Engert" <deengert@anl.gov>.
890 *) Add support for shared libraries for Unixware-7
891 (Boyd Lynn Gerber <gerberb@zenez.com>).
894 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
895 be done prior to destruction. Use this to unload error strings from
896 ENGINEs that load their own error strings. NB: This adds two new API
897 functions to "get" and "set" this destroy handler in an ENGINE.
900 *) Alter all existing ENGINE implementations (except "openssl" and
901 "openbsd") to dynamically instantiate their own error strings. This
902 makes them more flexible to be built both as statically-linked ENGINEs
903 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
904 Also, add stub code to each that makes building them as self-contained
905 shared-libraries easier (see README.ENGINE).
908 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
909 implementations into applications that are completely implemented in
910 self-contained shared-libraries. The "dynamic" ENGINE exposes control
911 commands that can be used to configure what shared-library to load and
912 to control aspects of the way it is handled. Also, made an update to
913 the README.ENGINE file that brings its information up-to-date and
914 provides some information and instructions on the "dynamic" ENGINE
915 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
918 *) Make it possible to unload ranges of ERR strings with a new
919 "ERR_unload_strings" function.
922 *) Add a copy() function to EVP_MD.
925 *) Make EVP_MD routines take a context pointer instead of just the
926 md_data void pointer.
929 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
930 that the digest can only process a single chunk of data
931 (typically because it is provided by a piece of
932 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
933 is only going to provide a single chunk of data, and hence the
934 framework needn't accumulate the data for oneshot drivers.
937 *) As with "ERR", make it possible to replace the underlying "ex_data"
938 functions. This change also alters the storage and management of global
939 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
940 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
941 index counters. The API functions that use this state have been changed
942 to take a "class_index" rather than pointers to the class's local STACK
943 and counter, and there is now an API function to dynamically create new
944 classes. This centralisation allows us to (a) plug a lot of the
945 thread-safety problems that existed, and (b) makes it possible to clean
946 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
947 such data would previously have always leaked in application code and
948 workarounds were in place to make the memory debugging turn a blind eye
949 to it. Application code that doesn't use this new function will still
950 leak as before, but their memory debugging output will announce it now
951 rather than letting it slide.
953 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
954 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
955 has a return value to indicate success or failure.
958 *) Make it possible to replace the underlying "ERR" functions such that the
959 global state (2 LHASH tables and 2 locks) is only used by the "default"
960 implementation. This change also adds two functions to "get" and "set"
961 the implementation prior to it being automatically set the first time
962 any other ERR function takes place. Ie. an application can call "get",
963 pass the return value to a module it has just loaded, and that module
964 can call its own "set" function using that value. This means the
965 module's "ERR" operations will use (and modify) the error state in the
966 application and not in its own statically linked copy of OpenSSL code.
969 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
970 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
971 the operation, and provides a more encapsulated way for external code
972 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
973 to use these functions rather than manually incrementing the counts.
975 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
978 *) Add EVP test program.
981 *) Add symmetric cipher support to ENGINE. Expect the API to change!
984 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
985 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
986 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
987 These allow a CRL to be built without having to access X509_CRL fields
988 directly. Modify 'ca' application to use new functions.
991 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
992 bug workarounds. Rollback attack detection is a security feature.
993 The problem will only arise on OpenSSL servers when TLSv1 is not
994 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
995 Software authors not wanting to support TLSv1 will have special reasons
996 for their choice and can explicitly enable this option.
997 [Bodo Moeller, Lutz Jaenicke]
999 *) Rationalise EVP so it can be extended: don't include a union of
1000 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
1001 (similar to those existing for EVP_CIPHER_CTX).
1006 EVP_MD_CTX_init(&md); /* new function call */
1007 EVP_DigestInit(&md, EVP_sha1());
1008 EVP_DigestUpdate(&md, in, len);
1009 EVP_DigestFinal(&md, out, NULL);
1010 EVP_MD_CTX_cleanup(&md); /* new function call */
1014 *) Make DES key schedule conform to the usual scheme, as well as
1015 correcting its structure. This means that calls to DES functions
1016 now have to pass a pointer to a des_key_schedule instead of a
1017 plain des_key_schedule (which was actually always a pointer
1020 des_key_schedule ks;
1022 des_set_key_checked(..., &ks);
1023 des_ncbc_encrypt(..., &ks, ...);
1025 (Note that a later change renames 'des_...' into 'DES_...'.)
1028 *) Initial reduction of linker bloat: the use of some functions, such as
1029 PEM causes large amounts of unused functions to be linked in due to
1030 poor organisation. For example pem_all.c contains every PEM function
1031 which has a knock on effect of linking in large amounts of (unused)
1032 ASN1 code. Grouping together similar functions and splitting unrelated
1033 functions prevents this.
1036 *) Cleanup of EVP macros.
1039 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
1040 correct _ecb suffix.
1043 *) Add initial OCSP responder support to ocsp application. The
1044 revocation information is handled using the text based index
1045 use by the ca application. The responder can either handle
1046 requests generated internally, supplied in files (for example
1047 via a CGI script) or using an internal minimal server.
1050 *) Add configuration choices to get zlib compression for TLS.
1053 *) Changes to Kerberos SSL for RFC 2712 compliance:
1054 1. Implemented real KerberosWrapper, instead of just using
1055 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
1056 2. Implemented optional authenticator field of KerberosWrapper.
1058 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
1059 and authenticator structs; see crypto/krb5/.
1061 Generalized Kerberos calls to support multiple Kerberos libraries.
1062 [Vern Staats <staatsvr@asc.hpc.mil>,
1063 Jeffrey Altman <jaltman@columbia.edu>
1064 via Richard Levitte]
1066 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
1067 already does with RSA. testdsa.h now has 'priv_key/pub_key'
1068 values for each of the key sizes rather than having just
1069 parameters (and 'speed' generating keys each time).
1072 *) Speed up EVP routines.
1075 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
1076 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
1077 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
1078 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
1080 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
1081 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
1082 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
1085 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
1087 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
1090 *) Added the OS2-EMX target.
1091 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
1093 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
1094 to support NCONF routines in extension code. New function CONF_set_nconf()
1095 to allow functions which take an NCONF to also handle the old LHASH
1096 structure: this means that the old CONF compatible routines can be
1097 retained (in particular wrt extensions) without having to duplicate the
1098 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
1101 *) Enhance the general user interface with mechanisms for inner control
1102 and with possibilities to have yes/no kind of prompts.
1105 *) Change all calls to low level digest routines in the library and
1106 applications to use EVP. Add missing calls to HMAC_cleanup() and
1107 don't assume HMAC_CTX can be copied using memcpy().
1108 [Verdon Walker <VWalker@novell.com>, Steve Henson]
1110 *) Add the possibility to control engines through control names but with
1111 arbitrary arguments instead of just a string.
1112 Change the key loaders to take a UI_METHOD instead of a callback
1113 function pointer. NOTE: this breaks binary compatibility with earlier
1114 versions of OpenSSL [engine].
1115 Adapt the nCipher code for these new conditions and add a card insertion
1119 *) Enhance the general user interface with mechanisms to better support
1120 dialog box interfaces, application-defined prompts, the possibility
1121 to use defaults (for example default passwords from somewhere else)
1122 and interrupts/cancellations.
1125 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
1126 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
1129 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
1130 tidy up some unnecessarily weird code in 'sk_new()').
1131 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
1133 *) Change the key loading routines for ENGINEs to use the same kind
1134 callback (pem_password_cb) as all other routines that need this
1138 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
1139 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
1140 than this minimum value is recommended.
1143 *) New random seeder for OpenVMS, using the system process statistics
1144 that are easily reachable.
1147 *) Windows apparently can't transparently handle global
1148 variables defined in DLLs. Initialisations such as:
1150 const ASN1_ITEM *it = &ASN1_INTEGER_it;
1152 wont compile. This is used by the any applications that need to
1153 declare their own ASN1 modules. This was fixed by adding the option
1154 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
1155 needed for static libraries under Win32.
1158 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
1159 setting of purpose and trust fields. New X509_STORE trust and
1160 purpose functions and tidy up setting in other SSL functions.
1163 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
1164 structure. These are inherited by X509_STORE_CTX when it is
1165 initialised. This allows various defaults to be set in the
1166 X509_STORE structure (such as flags for CRL checking and custom
1167 purpose or trust settings) for functions which only use X509_STORE_CTX
1168 internally such as S/MIME.
1170 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
1171 trust settings if they are not set in X509_STORE. This allows X509_STORE
1172 purposes and trust (in S/MIME for example) to override any set by default.
1174 Add command line options for CRL checking to smime, s_client and s_server
1178 *) Initial CRL based revocation checking. If the CRL checking flag(s)
1179 are set then the CRL is looked up in the X509_STORE structure and
1180 its validity and signature checked, then if the certificate is found
1181 in the CRL the verify fails with a revoked error.
1183 Various new CRL related callbacks added to X509_STORE_CTX structure.
1185 Command line options added to 'verify' application to support this.
1187 This needs some additional work, such as being able to handle multiple
1188 CRLs with different times, extension based lookup (rather than just
1189 by subject name) and ultimately more complete V2 CRL extension
1193 *) Add a general user interface API (crypto/ui/). This is designed
1194 to replace things like des_read_password and friends (backward
1195 compatibility functions using this new API are provided).
1196 The purpose is to remove prompting functions from the DES code
1197 section as well as provide for prompting through dialog boxes in
1198 a window system and the like.
1201 *) Add "ex_data" support to ENGINE so implementations can add state at a
1202 per-structure level rather than having to store it globally.
1205 *) Make it possible for ENGINE structures to be copied when retrieved by
1206 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
1207 This causes the "original" ENGINE structure to act like a template,
1208 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
1209 operational state can be localised to each ENGINE structure, despite the
1210 fact they all share the same "methods". New ENGINE structures returned in
1211 this case have no functional references and the return value is the single
1212 structural reference. This matches the single structural reference returned
1213 by ENGINE_by_id() normally, when it is incremented on the pre-existing
1217 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
1218 needs to match any other type at all we need to manually clear the
1222 *) Changes to the "openssl engine" utility to include;
1223 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
1224 about an ENGINE's available control commands.
1225 - executing control commands from command line arguments using the
1226 '-pre' and '-post' switches. '-post' is only used if '-t' is
1227 specified and the ENGINE is successfully initialised. The syntax for
1228 the individual commands are colon-separated, for example;
1229 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
1232 *) New dynamic control command support for ENGINEs. ENGINEs can now
1233 declare their own commands (numbers), names (strings), descriptions,
1234 and input types for run-time discovery by calling applications. A
1235 subset of these commands are implicitly classed as "executable"
1236 depending on their input type, and only these can be invoked through
1237 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
1238 can be based on user input, config files, etc). The distinction is
1239 that "executable" commands cannot return anything other than a boolean
1240 result and can only support numeric or string input, whereas some
1241 discoverable commands may only be for direct use through
1242 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
1243 pointers, or other custom uses. The "executable" commands are to
1244 support parameterisations of ENGINE behaviour that can be
1245 unambiguously defined by ENGINEs and used consistently across any
1246 OpenSSL-based application. Commands have been added to all the
1247 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
1248 control over shared-library paths without source code alterations.
1251 *) Changed all ENGINE implementations to dynamically allocate their
1252 ENGINEs rather than declaring them statically. Apart from this being
1253 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
1254 this also allows the implementations to compile without using the
1255 internal engine_int.h header.
1258 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
1259 'const' value. Any code that should be able to modify a RAND_METHOD
1260 should already have non-const pointers to it (ie. they should only
1261 modify their own ones).
1264 *) Made a variety of little tweaks to the ENGINE code.
1265 - "atalla" and "ubsec" string definitions were moved from header files
1266 to C code. "nuron" string definitions were placed in variables
1267 rather than hard-coded - allowing parameterisation of these values
1268 later on via ctrl() commands.
1269 - Removed unused "#if 0"'d code.
1270 - Fixed engine list iteration code so it uses ENGINE_free() to release
1271 structural references.
1272 - Constified the RAND_METHOD element of ENGINE structures.
1273 - Constified various get/set functions as appropriate and added
1274 missing functions (including a catch-all ENGINE_cpy that duplicates
1275 all ENGINE values onto a new ENGINE except reference counts/state).
1276 - Removed NULL parameter checks in get/set functions. Setting a method
1277 or function to NULL is a way of cancelling out a previously set
1278 value. Passing a NULL ENGINE parameter is just plain stupid anyway
1279 and doesn't justify the extra error symbols and code.
1280 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
1281 flags from engine_int.h to engine.h.
1282 - Changed prototypes for ENGINE handler functions (init(), finish(),
1283 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
1286 *) Implement binary inversion algorithm for BN_mod_inverse in addition
1287 to the algorithm using long division. The binary algorithm can be
1288 used only if the modulus is odd. On 32-bit systems, it is faster
1289 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
1290 roughly 5-15% for 256-bit moduli), so we use it only for moduli
1291 up to 450 bits. In 64-bit environments, the binary algorithm
1292 appears to be advantageous for much longer moduli; here we use it
1293 for moduli up to 2048 bits.
1296 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
1297 could not support the combine flag in choice fields.
1300 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
1301 extensions from a certificate request to the certificate.
1304 *) Allow multiple 'certopt' and 'nameopt' options to be separated
1305 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
1306 file: this allows the display of the certificate about to be
1307 signed to be customised, to allow certain fields to be included
1308 or excluded and extension details. The old system didn't display
1309 multicharacter strings properly, omitted fields not in the policy
1310 and couldn't display additional details such as extensions.
1313 *) Function EC_POINTs_mul for multiple scalar multiplication
1314 of an arbitrary number of elliptic curve points
1315 \sum scalars[i]*points[i],
1316 optionally including the generator defined for the EC_GROUP:
1317 scalar*generator + \sum scalars[i]*points[i].
1319 EC_POINT_mul is a simple wrapper function for the typical case
1320 that the point list has just one item (besides the optional
1324 *) First EC_METHODs for curves over GF(p):
1326 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
1327 operations and provides various method functions that can also
1328 operate with faster implementations of modular arithmetic.
1330 EC_GFp_mont_method() reuses most functions that are part of
1331 EC_GFp_simple_method, but uses Montgomery arithmetic.
1333 [Bodo Moeller; point addition and point doubling
1334 implementation directly derived from source code provided by
1335 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
1337 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
1338 crypto/ec/ec_lib.c):
1340 Curves are EC_GROUP objects (with an optional group generator)
1341 based on EC_METHODs that are built into the library.
1343 Points are EC_POINT objects based on EC_GROUP objects.
1345 Most of the framework would be able to handle curves over arbitrary
1346 finite fields, but as there are no obvious types for fields other
1347 than GF(p), some functions are limited to that for now.
1350 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
1351 that the file contains a complete HTTP response.
1354 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
1355 change the def and num file printf format specifier from "%-40sXXX"
1356 to "%-39s XXX". The latter will always guarantee a space after the
1357 field while the former will cause them to run together if the field
1358 is 40 of more characters long.
1361 *) Constify the cipher and digest 'method' functions and structures
1362 and modify related functions to take constant EVP_MD and EVP_CIPHER
1366 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
1367 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
1370 *) Modify EVP_Digest*() routines so they now return values. Although the
1371 internal software routines can never fail additional hardware versions
1375 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
1377 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
1378 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
1381 ERR_R_NESTED_ASN1_ERROR
1383 ERR_R_MISSING_ASN1_EOS
1384 were 4 .. 9, conflicting with
1385 ERR_LIB_RSA (= ERR_R_RSA_LIB)
1387 ERR_LIB_PEM (= ERR_R_PEM_LIB).
1388 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
1390 Add new error code 'ERR_R_INTERNAL_ERROR'.
1393 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
1397 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
1398 sets the subject name for a new request or supersedes the
1399 subject name in a given request. Formats that can be parsed are
1400 'CN=Some Name, OU=myOU, C=IT'
1402 'CN=Some Name/OU=myOU/C=IT'.
1404 Add options '-batch' and '-verbose' to 'openssl req'.
1405 [Massimiliano Pala <madwolf@hackmasters.net>]
1407 *) Introduce the possibility to access global variables through
1408 functions on platform were that's the best way to handle exporting
1409 global variables in shared libraries. To enable this functionality,
1410 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1411 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1412 is normally done by Configure or something similar).
1414 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1415 in the source file (foo.c) like this:
1417 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1418 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1420 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1421 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1423 OPENSSL_DECLARE_GLOBAL(int,foo);
1424 #define foo OPENSSL_GLOBAL_REF(foo)
1425 OPENSSL_DECLARE_GLOBAL(double,bar);
1426 #define bar OPENSSL_GLOBAL_REF(bar)
1428 The #defines are very important, and therefore so is including the
1429 header file everywhere where the defined globals are used.
1431 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1432 of ASN.1 items, but that structure is a bit different.
1434 The largest change is in util/mkdef.pl which has been enhanced with
1435 better and easier to understand logic to choose which symbols should
1436 go into the Windows .def files as well as a number of fixes and code
1437 cleanup (among others, algorithm keywords are now sorted
1438 lexicographically to avoid constant rewrites).
1441 *) In BN_div() keep a copy of the sign of 'num' before writing the
1442 result to 'rm' because if rm==num the value will be overwritten
1443 and produce the wrong result if 'num' is negative: this caused
1444 problems with BN_mod() and BN_nnmod().
1447 *) Function OCSP_request_verify(). This checks the signature on an
1448 OCSP request and verifies the signer certificate. The signer
1449 certificate is just checked for a generic purpose and OCSP request
1453 *) Add OCSP_check_validity() function to check the validity of OCSP
1454 responses. OCSP responses are prepared in real time and may only
1455 be a few seconds old. Simply checking that the current time lies
1456 between thisUpdate and nextUpdate max reject otherwise valid responses
1457 caused by either OCSP responder or client clock inaccuracy. Instead
1458 we allow thisUpdate and nextUpdate to fall within a certain period of
1459 the current time. The age of the response can also optionally be
1460 checked. Two new options -validity_period and -status_age added to
1464 *) If signature or public key algorithm is unrecognized print out its
1465 OID rather that just UNKNOWN.
1468 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1469 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1470 ID to be generated from the issuer certificate alone which can then be
1471 passed to OCSP_id_issuer_cmp().
1474 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1475 ASN1 modules to export functions returning ASN1_ITEM pointers
1476 instead of the ASN1_ITEM structures themselves. This adds several
1477 new macros which allow the underlying ASN1 function/structure to
1478 be accessed transparently. As a result code should not use ASN1_ITEM
1479 references directly (such as &X509_it) but instead use the relevant
1480 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1481 use of the new ASN1 code on platforms where exporting structures
1482 is problematical (for example in shared libraries) but exporting
1483 functions returning pointers to structures is not.
1486 *) Add support for overriding the generation of SSL/TLS session IDs.
1487 These callbacks can be registered either in an SSL_CTX or per SSL.
1488 The purpose of this is to allow applications to control, if they wish,
1489 the arbitrary values chosen for use as session IDs, particularly as it
1490 can be useful for session caching in multiple-server environments. A
1491 command-line switch for testing this (and any client code that wishes
1492 to use such a feature) has been added to "s_server".
1493 [Geoff Thorpe, Lutz Jaenicke]
1495 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
1496 of the form '#if defined(...) || defined(...) || ...' and
1497 '#if !defined(...) && !defined(...) && ...'. This also avoids
1498 the growing number of special cases it was previously handling.
1501 *) Make all configuration macros available for application by making
1502 sure they are available in opensslconf.h, by giving them names starting
1503 with "OPENSSL_" to avoid conflicts with other packages and by making
1504 sure e_os2.h will cover all platform-specific cases together with
1506 Additionally, it is now possible to define configuration/platform-
1507 specific names (called "system identities"). In the C code, these
1508 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1509 macro with the name beginning with "OPENSSL_SYS_", which is determined
1510 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1514 *) New option -set_serial to 'req' and 'x509' this allows the serial
1515 number to use to be specified on the command line. Previously self
1516 signed certificates were hard coded with serial number 0 and the
1517 CA options of 'x509' had to use a serial number in a file which was
1521 *) New options to 'ca' utility to support V2 CRL entry extensions.
1522 Currently CRL reason, invalidity date and hold instruction are
1523 supported. Add new CRL extensions to V3 code and some new objects.
1526 *) New function EVP_CIPHER_CTX_set_padding() this is used to
1527 disable standard block padding (aka PKCS#5 padding) in the EVP
1528 API, which was previously mandatory. This means that the data is
1529 not padded in any way and so the total length much be a multiple
1530 of the block size, otherwise an error occurs.
1533 *) Initial (incomplete) OCSP SSL support.
1536 *) New function OCSP_parse_url(). This splits up a URL into its host,
1537 port and path components: primarily to parse OCSP URLs. New -url
1538 option to ocsp utility.
1541 *) New nonce behavior. The return value of OCSP_check_nonce() now
1542 reflects the various checks performed. Applications can decide
1543 whether to tolerate certain situations such as an absent nonce
1544 in a response when one was present in a request: the ocsp application
1545 just prints out a warning. New function OCSP_add1_basic_nonce()
1546 this is to allow responders to include a nonce in a response even if
1547 the request is nonce-less.
1550 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1551 skipped when using openssl x509 multiple times on a single input file,
1552 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1555 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1556 set string type: to handle setting ASN1_TIME structures. Fix ca
1557 utility to correctly initialize revocation date of CRLs.
1560 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1561 the clients preferred ciphersuites and rather use its own preferences.
1562 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1563 Internet Explorer by ensuring unchanged hash method during stepup.
1564 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1567 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1568 to aes and add a new 'exist' option to print out symbols that don't
1572 *) Additional options to ocsp utility to allow flags to be set and
1573 additional certificates supplied.
1576 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
1577 OCSP client a number of certificate to only verify the response
1581 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1582 handle the new API. Currently only ECB, CBC modes supported. Add new
1585 Add TLS AES ciphersuites as described in RFC3268, "Advanced
1586 Encryption Standard (AES) Ciphersuites for Transport Layer
1587 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
1588 not enabled by default and were not part of the "ALL" ciphersuite
1589 alias because they were not yet official; they could be
1590 explicitly requested by specifying the "AESdraft" ciphersuite
1591 group alias. In the final release of OpenSSL 0.9.7, the group
1592 alias is called "AES" and is part of "ALL".)
1593 [Ben Laurie, Steve Henson, Bodo Moeller]
1595 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
1596 request to response.
1599 *) Functions for OCSP responders. OCSP_request_onereq_count(),
1600 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1601 extract information from a certificate request. OCSP_response_create()
1602 creates a response and optionally adds a basic response structure.
1603 OCSP_basic_add1_status() adds a complete single response to a basic
1604 response and returns the OCSP_SINGLERESP structure just added (to allow
1605 extensions to be included for example). OCSP_basic_add1_cert() adds a
1606 certificate to a basic response and OCSP_basic_sign() signs a basic
1607 response with various flags. New helper functions ASN1_TIME_check()
1608 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1609 (converts ASN1_TIME to GeneralizedTime).
1612 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1613 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1614 structure from a certificate. X509_pubkey_digest() digests the public_key
1615 contents: this is used in various key identifiers.
1618 *) Make sk_sort() tolerate a NULL argument.
1619 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1621 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1622 passed by the function are trusted implicitly. If any of them signed the
1623 response then it is assumed to be valid and is not verified.
1626 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1627 to data. This was previously part of the PKCS7 ASN1 code. This
1628 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1629 [Steve Henson, reported by Kenneth R. Robinette
1630 <support@securenetterm.com>]
1632 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1633 routines: without these tracing memory leaks is very painful.
1634 Fix leaks in PKCS12 and PKCS7 routines.
1637 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1638 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1639 effectively meant GeneralizedTime would never be used. Now it
1640 is initialised to -1 but X509_time_adj() now has to check the value
1641 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1642 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1643 [Steve Henson, reported by Kenneth R. Robinette
1644 <support@securenetterm.com>]
1646 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1647 result in a zero length in the ASN1_INTEGER structure which was
1648 not consistent with the structure when d2i_ASN1_INTEGER() was used
1649 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1650 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1651 where it did not print out a minus for negative ASN1_INTEGER.
1654 *) Add summary printout to ocsp utility. The various functions which
1655 convert status values to strings have been renamed to:
1656 OCSP_response_status_str(), OCSP_cert_status_str() and
1657 OCSP_crl_reason_str() and are no longer static. New options
1658 to verify nonce values and to disable verification. OCSP response
1659 printout format cleaned up.
1662 *) Add additional OCSP certificate checks. These are those specified
1663 in RFC2560. This consists of two separate checks: the CA of the
1664 certificate being checked must either be the OCSP signer certificate
1665 or the issuer of the OCSP signer certificate. In the latter case the
1666 OCSP signer certificate must contain the OCSP signing extended key
1667 usage. This check is performed by attempting to match the OCSP
1668 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1669 in the OCSP_CERTID structures of the response.
1672 *) Initial OCSP certificate verification added to OCSP_basic_verify()
1673 and related routines. This uses the standard OpenSSL certificate
1674 verify routines to perform initial checks (just CA validity) and
1675 to obtain the certificate chain. Then additional checks will be
1676 performed on the chain. Currently the root CA is checked to see
1677 if it is explicitly trusted for OCSP signing. This is used to set
1678 a root CA as a global signing root: that is any certificate that
1679 chains to that CA is an acceptable OCSP signing certificate.
1682 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1683 extensions from a separate configuration file.
1684 As when reading extensions from the main configuration file,
1685 the '-extensions ...' option may be used for specifying the
1687 [Massimiliano Pala <madwolf@comune.modena.it>]
1689 *) New OCSP utility. Allows OCSP requests to be generated or
1690 read. The request can be sent to a responder and the output
1691 parsed, outputed or printed in text form. Not complete yet:
1692 still needs to check the OCSP response validity.
1695 *) New subcommands for 'openssl ca':
1696 'openssl ca -status <serial>' prints the status of the cert with
1697 the given serial number (according to the index file).
1698 'openssl ca -updatedb' updates the expiry status of certificates
1700 [Massimiliano Pala <madwolf@comune.modena.it>]
1702 *) New '-newreq-nodes' command option to CA.pl. This is like
1703 '-newreq', but calls 'openssl req' with the '-nodes' option
1704 so that the resulting key is not encrypted.
1705 [Damien Miller <djm@mindrot.org>]
1707 *) New configuration for the GNU Hurd.
1708 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1710 *) Initial code to implement OCSP basic response verify. This
1711 is currently incomplete. Currently just finds the signer's
1712 certificate and verifies the signature on the response.
1715 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1716 value of OPENSSLDIR. This is available via the new '-d' option
1717 to 'openssl version', and is also included in 'openssl version -a'.
1720 *) Allowing defining memory allocation callbacks that will be given
1721 file name and line number information in additional arguments
1722 (a const char* and an int). The basic functionality remains, as
1723 well as the original possibility to just replace malloc(),
1724 realloc() and free() by functions that do not know about these
1725 additional arguments. To register and find out the current
1726 settings for extended allocation functions, the following
1727 functions are provided:
1729 CRYPTO_set_mem_ex_functions
1730 CRYPTO_set_locked_mem_ex_functions
1731 CRYPTO_get_mem_ex_functions
1732 CRYPTO_get_locked_mem_ex_functions
1734 These work the same way as CRYPTO_set_mem_functions and friends.
1735 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1736 extended allocation function is enabled.
1737 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1738 a conventional allocation function is enabled.
1739 [Richard Levitte, Bodo Moeller]
1741 *) Finish off removing the remaining LHASH function pointer casts.
1742 There should no longer be any prototype-casting required when using
1743 the LHASH abstraction, and any casts that remain are "bugs". See
1744 the callback types and macros at the head of lhash.h for details
1745 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1748 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1749 If /dev/[u]random devices are not available or do not return enough
1750 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1752 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1753 /etc/entropy will be queried once each in this sequence, quering stops
1754 when enough entropy was collected without querying more sockets.
1757 *) Change the Unix RAND_poll() variant to be able to poll several
1758 random devices, as specified by DEVRANDOM, until a sufficient amount
1759 of data has been collected. We spend at most 10 ms on each file
1760 (select timeout) and read in non-blocking mode. DEVRANDOM now
1761 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1762 (previously it was just the string "/dev/urandom"), so on typical
1763 platforms the 10 ms delay will never occur.
1764 Also separate out the Unix variant to its own file, rand_unix.c.
1765 For VMS, there's a currently-empty rand_vms.c.
1768 *) Move OCSP client related routines to ocsp_cl.c. These
1769 provide utility functions which an application needing
1770 to issue a request to an OCSP responder and analyse the
1771 response will typically need: as opposed to those which an
1772 OCSP responder itself would need which will be added later.
1774 OCSP_request_sign() signs an OCSP request with an API similar
1775 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1776 response. OCSP_response_get1_basic() extracts basic response
1777 from response. OCSP_resp_find_status(): finds and extracts status
1778 information from an OCSP_CERTID structure (which will be created
1779 when the request structure is built). These are built from lower
1780 level functions which work on OCSP_SINGLERESP structures but
1781 wont normally be used unless the application wishes to examine
1782 extensions in the OCSP response for example.
1784 Replace nonce routines with a pair of functions.
1785 OCSP_request_add1_nonce() adds a nonce value and optionally
1786 generates a random value. OCSP_check_nonce() checks the
1787 validity of the nonce in an OCSP response.
1790 *) Change function OCSP_request_add() to OCSP_request_add0_id().
1791 This doesn't copy the supplied OCSP_CERTID and avoids the
1792 need to free up the newly created id. Change return type
1793 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1794 This can then be used to add extensions to the request.
1795 Deleted OCSP_request_new(), since most of its functionality
1796 is now in OCSP_REQUEST_new() (and the case insensitive name
1797 clash) apart from the ability to set the request name which
1798 will be added elsewhere.
1801 *) Update OCSP API. Remove obsolete extensions argument from
1802 various functions. Extensions are now handled using the new
1803 OCSP extension code. New simple OCSP HTTP function which
1804 can be used to send requests and parse the response.
1807 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1808 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1809 uses the special reorder version of SET OF to sort the attributes
1810 and reorder them to match the encoded order. This resolves a long
1811 standing problem: a verify on a PKCS7 structure just after signing
1812 it used to fail because the attribute order did not match the
1813 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1814 it uses the received order. This is necessary to tolerate some broken
1815 software that does not order SET OF. This is handled by encoding
1816 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1817 to produce the required SET OF.
1820 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1821 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1822 files to get correct declarations of the ASN.1 item variables.
1825 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1826 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1827 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1828 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1829 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1830 ASN1_ITEM and no wrapper functions.
1833 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1834 replace the old function pointer based I/O routines. Change most of
1835 the *_d2i_bio() and *_d2i_fp() functions to use these.
1838 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1839 lines, recognice more "algorithms" that can be deselected, and make
1840 it complain about algorithm deselection that isn't recognised.
1843 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
1844 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1845 to use new functions. Add NO_ASN1_OLD which can be set to remove
1846 some old style ASN1 functions: this can be used to determine if old
1847 code will still work when these eventually go away.
1850 *) New extension functions for OCSP structures, these follow the
1851 same conventions as certificates and CRLs.
1854 *) New function X509V3_add1_i2d(). This automatically encodes and
1855 adds an extension. Its behaviour can be customised with various
1856 flags to append, replace or delete. Various wrappers added for
1857 certifcates and CRLs.
1860 *) Fix to avoid calling the underlying ASN1 print routine when
1861 an extension cannot be parsed. Correct a typo in the
1862 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1865 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1866 entries for variables.
1869 *) Add functionality to apps/openssl.c for detecting locking
1870 problems: As the program is single-threaded, all we have
1871 to do is register a locking callback using an array for
1872 storing which locks are currently held by the program.
1875 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
1876 SSL_get_ex_data_X509_STORE_idx(), which is used in
1877 ssl_verify_cert_chain() and thus can be called at any time
1878 during TLS/SSL handshakes so that thread-safety is essential.
1879 Unfortunately, the ex_data design is not at all suited
1880 for multi-threaded use, so it probably should be abolished.
1883 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
1884 [Broadcom, tweaked and integrated by Geoff Thorpe]
1886 *) Move common extension printing code to new function
1887 X509V3_print_extensions(). Reorganise OCSP print routines and
1888 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1891 *) New function X509_signature_print() to remove duplication in some
1895 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
1896 set (this was treated exactly the same as SET OF previously). This
1897 is used to reorder the STACK representing the structure to match the
1898 encoding. This will be used to get round a problem where a PKCS7
1899 structure which was signed could not be verified because the STACK
1900 order did not reflect the encoded order.
1903 *) Reimplement the OCSP ASN1 module using the new code.
1906 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1907 for its ASN1 operations. The old style function pointers still exist
1908 for now but they will eventually go away.
1911 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
1912 completely replaces the old ASN1 functionality with a table driven
1913 encoder and decoder which interprets an ASN1_ITEM structure describing
1914 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1915 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1916 has also been converted to the new form.
1919 *) Change BN_mod_exp_recp so that negative moduli are tolerated
1920 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1921 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1922 for negative moduli.
1925 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
1926 of not touching the result's sign bit.
1929 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1933 *) Changed the LHASH code to use prototypes for callbacks, and created
1934 macros to declare and implement thin (optionally static) functions
1935 that provide type-safety and avoid function pointer casting for the
1936 type-specific callbacks.
1939 *) Added Kerberos Cipher Suites to be used with TLS, as written in
1941 [Veers Staats <staatsvr@asc.hpc.mil>,
1942 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1944 *) Reformat the FAQ so the different questions and answers can be divided
1945 in sections depending on the subject.
1948 *) Have the zlib compression code load ZLIB.DLL dynamically under
1952 *) New function BN_mod_sqrt for computing square roots modulo a prime
1953 (using the probabilistic Tonelli-Shanks algorithm unless
1954 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1955 be handled deterministically).
1956 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1958 *) Make BN_mod_inverse faster by explicitly handling small quotients
1959 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1960 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1963 *) New function BN_kronecker.
1966 *) Fix BN_gcd so that it works on negative inputs; the result is
1967 positive unless both parameters are zero.
1968 Previously something reasonably close to an infinite loop was
1969 possible because numbers could be growing instead of shrinking
1970 in the implementation of Euclid's algorithm.
1973 *) Fix BN_is_word() and BN_is_one() macros to take into account the
1974 sign of the number in question.
1976 Fix BN_is_word(a,w) to work correctly for w == 0.
1978 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1979 because its test if the absolute value of 'a' equals 'w'.
1980 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1981 it exists mostly for use in the implementations of BN_is_zero(),
1982 BN_is_one(), and BN_is_word().
1985 *) New function BN_swap.
1988 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1989 the exponentiation functions are more likely to produce reasonable
1990 results on negative inputs.
1993 *) Change BN_mod_mul so that the result is always non-negative.
1994 Previously, it could be negative if one of the factors was negative;
1995 I don't think anyone really wanted that behaviour.
1998 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1999 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
2000 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
2001 and add new functions:
2010 BN_mod_lshift1_quick
2014 These functions always generate non-negative results.
2016 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
2017 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
2019 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2020 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2021 be reduced modulo m.
2022 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2025 The following entry accidentily appeared in the CHANGES file
2026 distributed with OpenSSL 0.9.7. The modifications described in
2027 it do *not* apply to OpenSSL 0.9.7.
2029 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2030 was actually never needed) and in BN_mul(). The removal in BN_mul()
2031 required a small change in bn_mul_part_recursive() and the addition
2032 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2033 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2034 bn_sub_words() and bn_add_words() except they take arrays with
2039 *) In 'openssl passwd', verify passwords read from the terminal
2040 unless the '-salt' option is used (which usually means that
2041 verification would just waste user's time since the resulting
2042 hash is going to be compared with some given password hash)
2043 or the new '-noverify' option is used.
2045 This is an incompatible change, but it does not affect
2046 non-interactive use of 'openssl passwd' (passwords on the command
2047 line, '-stdin' option, '-in ...' option) and thus should not
2051 *) Remove all references to RSAref, since there's no more need for it.
2054 *) Make DSO load along a path given through an environment variable
2055 (SHLIB_PATH) with shl_load().
2058 *) Constify the ENGINE code as a result of BIGNUM constification.
2059 Also constify the RSA code and most things related to it. In a
2060 few places, most notable in the depth of the ASN.1 code, ugly
2061 casts back to non-const were required (to be solved at a later
2065 *) Make it so the openssl application has all engines loaded by default.
2068 *) Constify the BIGNUM routines a little more.
2071 *) Add the following functions:
2073 ENGINE_load_cswift()
2075 ENGINE_load_atalla()
2077 ENGINE_load_builtin_engines()
2079 That way, an application can itself choose if external engines that
2080 are built-in in OpenSSL shall ever be used or not. The benefit is
2081 that applications won't have to be linked with libdl or other dso
2082 libraries unless it's really needed.
2084 Changed 'openssl engine' to load all engines on demand.
2085 Changed the engine header files to avoid the duplication of some
2086 declarations (they differed!).
2089 *) 'openssl engine' can now list capabilities.
2092 *) Better error reporting in 'openssl engine'.
2095 *) Never call load_dh_param(NULL) in s_server.
2098 *) Add engine application. It can currently list engines by name and
2099 identity, and test if they are actually available.
2102 *) Improve RPM specification file by forcing symbolic linking and making
2103 sure the installed documentation is also owned by root.root.
2104 [Damien Miller <djm@mindrot.org>]
2106 *) Give the OpenSSL applications more possibilities to make use of
2107 keys (public as well as private) handled by engines.
2110 *) Add OCSP code that comes from CertCo.
2113 *) Add VMS support for the Rijndael code.
2116 *) Added untested support for Nuron crypto accelerator.
2119 *) Add support for external cryptographic devices. This code was
2120 previously distributed separately as the "engine" branch.
2121 [Geoff Thorpe, Richard Levitte]
2123 *) Rework the filename-translation in the DSO code. It is now possible to
2124 have far greater control over how a "name" is turned into a filename
2125 depending on the operating environment and any oddities about the
2126 different shared library filenames on each system.
2129 *) Support threads on FreeBSD-elf in Configure.
2132 *) Fix for SHA1 assembly problem with MASM: it produces
2133 warnings about corrupt line number information when assembling
2134 with debugging information. This is caused by the overlapping
2136 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2139 NCONF_get_number() has no error checking at all. As a replacement,
2140 NCONF_get_number_e() is defined (_e for "error checking") and is
2141 promoted strongly. The old NCONF_get_number is kept around for
2142 binary backward compatibility.
2143 Make it possible for methods to load from something other than a BIO,
2144 by providing a function pointer that is given a name instead of a BIO.
2145 For example, this could be used to load configuration data from an
2149 *) Fix for non blocking accept BIOs. Added new I/O special reason
2150 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2151 with non blocking I/O was not possible because no retry code was
2152 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2156 *) Added the beginnings of Rijndael support.
2159 *) Fix for bug in DirectoryString mask setting. Add support for
2160 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2161 to allow certificate printing to more controllable, additional
2162 'certopt' option to 'x509' to allow new printing options to be
2166 *) Clean old EAY MD5 hack from e_os.h.
2169 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
2171 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2172 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
2173 [Joe Orton, Steve Henson]
2175 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
2177 *) Fix additional bug revealed by the NISCC test suite:
2179 Stop bug triggering large recursion when presented with
2180 certain ASN.1 tags (CAN-2003-0851)
2183 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
2185 *) Fix various bugs revealed by running the NISCC test suite:
2187 Stop out of bounds reads in the ASN1 code when presented with
2188 invalid tags (CAN-2003-0543 and CAN-2003-0544).
2190 If verify callback ignores invalid public key errors don't try to check
2191 certificate signature with the NULL public key.
2195 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2196 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2200 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2201 extra data after the compression methods not only for TLS 1.0
2202 but also for SSL 3.0 (as required by the specification).
2203 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2205 *) Change X509_certificate_type() to mark the key as exported/exportable
2206 when it's 512 *bits* long, not 512 bytes.
2209 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
2211 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2212 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2213 a protocol version number mismatch like a decryption error
2214 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2217 *) Turn on RSA blinding by default in the default implementation
2218 to avoid a timing attack. Applications that don't want it can call
2219 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2220 They would be ill-advised to do so in most cases.
2221 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2223 *) Change RSA blinding code so that it works when the PRNG is not
2224 seeded (in this case, the secret RSA exponent is abused as
2225 an unpredictable seed -- if it is not unpredictable, there
2226 is no point in blinding anyway). Make RSA blinding thread-safe
2227 by remembering the creator's thread ID in rsa->blinding and
2228 having all other threads use local one-time blinding factors
2229 (this requires more computation than sharing rsa->blinding, but
2230 avoids excessive locking; and if an RSA object is not shared
2231 between threads, blinding will still be very fast).
2234 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
2236 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2237 via timing by performing a MAC computation even if incorrrect
2238 block cipher padding has been found. This is a countermeasure
2239 against active attacks where the attacker has to distinguish
2240 between bad padding and a MAC verification error. (CAN-2003-0078)
2242 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2243 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2244 Martin Vuagnoux (EPFL, Ilion)]
2246 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
2248 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
2249 memory from it's contents. This is done with a counter that will
2250 place alternating values in each byte. This can be used to solve
2251 two issues: 1) the removal of calls to memset() by highly optimizing
2252 compilers, and 2) cleansing with other values than 0, since those can
2253 be read through on certain media, for example a swap space on disk.
2256 *) Bugfix: client side session caching did not work with external caching,
2257 because the session->cipher setting was not restored when reloading
2258 from the external cache. This problem was masked, when
2259 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
2260 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
2263 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
2264 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
2265 [Zeev Lieber <zeev-l@yahoo.com>]
2267 *) Undo an undocumented change introduced in 0.9.6e which caused
2268 repeated calls to OpenSSL_add_all_ciphers() and
2269 OpenSSL_add_all_digests() to be ignored, even after calling
2273 *) Change the default configuration reader to deal with last line not
2274 being properly terminated.
2277 *) Change X509_NAME_cmp() so it applies the special rules on handling
2278 DN values that are of type PrintableString, as well as RDNs of type
2279 emailAddress where the value has the type ia5String.
2280 [stefank@valicert.com via Richard Levitte]
2282 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
2283 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
2284 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
2285 the bitwise-OR of the two for use by the majority of applications
2286 wanting this behaviour, and update the docs. The documented
2287 behaviour and actual behaviour were inconsistent and had been
2288 changing anyway, so this is more a bug-fix than a behavioural
2290 [Geoff Thorpe, diagnosed by Nadav Har'El]
2292 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
2293 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
2296 *) Fix initialization code race conditions in
2297 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
2298 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
2299 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
2300 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
2301 ssl2_get_cipher_by_char(),
2302 ssl3_get_cipher_by_char().
2303 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
2305 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
2306 the cached sessions are flushed, as the remove_cb() might use ex_data
2307 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
2308 (see [openssl.org #212]).
2309 [Geoff Thorpe, Lutz Jaenicke]
2311 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
2312 length, instead of the encoding length to d2i_ASN1_OBJECT.
2315 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
2317 *) [In 0.9.6g-engine release:]
2318 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
2319 [Lynn Gazis <lgazis@rainbow.com>]
2321 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
2323 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
2324 and get fix the header length calculation.
2325 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
2326 Alon Kantor <alonk@checkpoint.com> (and others),
2329 *) Use proper error handling instead of 'assertions' in buffer
2330 overflow checks added in 0.9.6e. This prevents DoS (the
2331 assertions could call abort()).
2332 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
2334 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
2336 *) Add various sanity checks to asn1_get_length() to reject
2337 the ASN1 length bytes if they exceed sizeof(long), will appear
2338 negative or the content length exceeds the length of the
2340 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2342 *) Fix cipher selection routines: ciphers without encryption had no flags
2343 for the cipher strength set and where therefore not handled correctly
2344 by the selection routines (PR #130).
2347 *) Fix EVP_dsa_sha macro.
2351 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
2352 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
2353 that was added in OpenSSL 0.9.6d.
2355 As the countermeasure turned out to be incompatible with some
2356 broken SSL implementations, the new option is part of SSL_OP_ALL.
2357 SSL_OP_ALL is usually employed when compatibility with weird SSL
2358 implementations is desired (e.g. '-bugs' option to 's_client' and
2359 's_server'), so the new option is automatically set in many
2363 *) Changes in security patch:
2365 Changes marked "(CHATS)" were sponsored by the Defense Advanced
2366 Research Projects Agency (DARPA) and Air Force Research Laboratory,
2367 Air Force Materiel Command, USAF, under agreement number
2370 *) Add various sanity checks to asn1_get_length() to reject
2371 the ASN1 length bytes if they exceed sizeof(long), will appear
2372 negative or the content length exceeds the length of the
2373 supplied buffer. (CAN-2002-0659)
2374 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2376 *) Assertions for various potential buffer overflows, not known to
2378 [Ben Laurie (CHATS)]
2380 *) Various temporary buffers to hold ASCII versions of integers were
2381 too small for 64 bit platforms. (CAN-2002-0655)
2382 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
2384 *) Remote buffer overflow in SSL3 protocol - an attacker could
2385 supply an oversized session ID to a client. (CAN-2002-0656)
2386 [Ben Laurie (CHATS)]
2388 *) Remote buffer overflow in SSL2 protocol - an attacker could
2389 supply an oversized client master key. (CAN-2002-0656)
2390 [Ben Laurie (CHATS)]
2392 Changes between 0.9.6c and 0.9.6d [9 May 2002]
2394 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
2395 encoded as NULL) with id-dsa-with-sha1.
2396 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
2398 *) Check various X509_...() return values in apps/req.c.
2399 [Nils Larsch <nla@trustcenter.de>]
2401 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
2402 an end-of-file condition would erronously be flagged, when the CRLF
2403 was just at the end of a processed block. The bug was discovered when
2404 processing data through a buffering memory BIO handing the data to a
2405 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
2406 <ptsekov@syntrex.com> and Nedelcho Stanev.
2409 *) Implement a countermeasure against a vulnerability recently found
2410 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
2411 before application data chunks to avoid the use of known IVs
2412 with data potentially chosen by the attacker.
2415 *) Fix length checks in ssl3_get_client_hello().
2418 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
2419 to prevent ssl3_read_internal() from incorrectly assuming that
2420 ssl3_read_bytes() found application data while handshake
2421 processing was enabled when in fact s->s3->in_read_app_data was
2422 merely automatically cleared during the initial handshake.
2423 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
2425 *) Fix object definitions for Private and Enterprise: they were not
2426 recognized in their shortname (=lowercase) representation. Extend
2427 obj_dat.pl to issue an error when using undefined keywords instead
2428 of silently ignoring the problem (Svenning Sorensen
2429 <sss@sss.dnsalias.net>).
2432 *) Fix DH_generate_parameters() so that it works for 'non-standard'
2433 generators, i.e. generators other than 2 and 5. (Previously, the
2434 code did not properly initialise the 'add' and 'rem' values to
2435 BN_generate_prime().)
2437 In the new general case, we do not insist that 'generator' is
2438 actually a primitive root: This requirement is rather pointless;
2439 a generator of the order-q subgroup is just as good, if not
2443 *) Map new X509 verification errors to alerts. Discovered and submitted by
2444 Tom Wu <tom@arcot.com>.
2447 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
2448 returning non-zero before the data has been completely received
2449 when using non-blocking I/O.
2450 [Bodo Moeller; problem pointed out by John Hughes]
2452 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
2453 [Ben Laurie, Lutz Jaenicke]
2455 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
2456 Yoram Zahavi <YoramZ@gilian.com>).
2459 *) Add information about CygWin 1.3 and on, and preserve proper
2460 configuration for the versions before that.
2461 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2463 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
2464 check whether we deal with a copy of a session and do not delete from
2465 the cache in this case. Problem reported by "Izhar Shoshani Levi"
2466 <izhar@checkpoint.com>.
2469 *) Do not store session data into the internal session cache, if it
2470 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
2471 flag is set). Proposed by Aslam <aslam@funk.com>.
2474 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
2478 *) [In 0.9.6d-engine release:]
2479 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2480 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2482 *) Add the configuration target linux-s390x.
2483 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
2485 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
2486 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
2487 variable as an indication that a ClientHello message has been
2488 received. As the flag value will be lost between multiple
2489 invocations of ssl3_accept when using non-blocking I/O, the
2490 function may not be aware that a handshake has actually taken
2491 place, thus preventing a new session from being added to the
2494 To avoid this problem, we now set s->new_session to 2 instead of
2495 using a local variable.
2496 [Lutz Jaenicke, Bodo Moeller]
2498 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
2499 if the SSL_R_LENGTH_MISMATCH error is detected.
2500 [Geoff Thorpe, Bodo Moeller]
2502 *) New 'shared_ldflag' column in Configure platform table.
2505 *) Fix EVP_CIPHER_mode macro.
2506 ["Dan S. Camper" <dan@bti.net>]
2508 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
2509 type, we must throw them away by setting rr->length to 0.
2510 [D P Chang <dpc@qualys.com>]
2512 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
2514 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
2515 <Dominikus.Scherkl@biodata.com>. (The previous implementation
2516 worked incorrectly for those cases where range = 10..._2 and
2517 3*range is two bits longer than range.)
2520 *) Only add signing time to PKCS7 structures if it is not already
2524 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
2525 OBJ_ld_ce should be OBJ_id_ce.
2526 Also some ip-pda OIDs in crypto/objects/objects.txt were
2527 incorrect (cf. RFC 3039).
2528 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
2530 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
2531 returns early because it has nothing to do.
2532 [Andy Schneider <andy.schneider@bjss.co.uk>]
2534 *) [In 0.9.6c-engine release:]
2535 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
2536 [Andy Schneider <andy.schneider@bjss.co.uk>]
2538 *) [In 0.9.6c-engine release:]
2539 Add support for Cryptographic Appliance's keyserver technology.
2540 (Use engine 'keyclient')
2541 [Cryptographic Appliances and Geoff Thorpe]
2543 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
2544 is called via tools/c89.sh because arguments have to be
2545 rearranged (all '-L' options must appear before the first object
2547 [Richard Shapiro <rshapiro@abinitio.com>]
2549 *) [In 0.9.6c-engine release:]
2550 Add support for Broadcom crypto accelerator cards, backported
2552 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
2554 *) [In 0.9.6c-engine release:]
2555 Add support for SureWare crypto accelerator cards from
2556 Baltimore Technologies. (Use engine 'sureware')
2557 [Baltimore Technologies and Mark Cox]
2559 *) [In 0.9.6c-engine release:]
2560 Add support for crypto accelerator cards from Accelerated
2561 Encryption Processing, www.aep.ie. (Use engine 'aep')
2562 [AEP Inc. and Mark Cox]
2564 *) Add a configuration entry for gcc on UnixWare.
2565 [Gary Benson <gbenson@redhat.com>]
2567 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
2568 messages are stored in a single piece (fixed-length part and
2569 variable-length part combined) and fix various bugs found on the way.
2572 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
2573 instead. BIO_gethostbyname() does not know what timeouts are
2574 appropriate, so entries would stay in cache even when they have
2576 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
2578 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
2579 faced with a pathologically small ClientHello fragment that does
2580 not contain client_version: Instead of aborting with an error,
2581 simply choose the highest available protocol version (i.e.,
2582 TLS 1.0 unless it is disabled). In practice, ClientHello
2583 messages are never sent like this, but this change gives us
2584 strictly correct behaviour at least for TLS.
2587 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
2588 never resets s->method to s->ctx->method when called from within
2589 one of the SSL handshake functions.
2590 [Bodo Moeller; problem pointed out by Niko Baric]
2592 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
2593 (sent using the client's version number) if client_version is
2594 smaller than the protocol version in use. Also change
2595 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
2596 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
2597 the client will at least see that alert.
2600 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
2604 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
2605 client receives HelloRequest while in a handshake.
2606 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
2608 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
2609 should end in 'break', not 'goto end' which circuments various
2610 cleanups done in state SSL_ST_OK. But session related stuff
2611 must be disabled for SSL_ST_OK in the case that we just sent a
2614 Also avoid some overhead by not calling ssl_init_wbio_buffer()
2615 before just sending a HelloRequest.
2616 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
2618 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
2619 reveal whether illegal block cipher padding was found or a MAC
2620 verification error occured. (Neither SSLerr() codes nor alerts
2621 are directly visible to potential attackers, but the information
2622 may leak via logfiles.)
2624 Similar changes are not required for the SSL 2.0 implementation
2625 because the number of padding bytes is sent in clear for SSL 2.0,
2626 and the extra bytes are just ignored. However ssl/s2_pkt.c
2627 failed to verify that the purported number of padding bytes is in
2631 *) Add OpenUNIX-8 support including shared libraries
2632 (Boyd Lynn Gerber <gerberb@zenez.com>).
2635 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
2636 'wristwatch attack' using huge encoding parameters (cf.
2637 James H. Manger's CRYPTO 2001 paper). Note that the
2638 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
2639 encoding parameters and hence was not vulnerable.
2642 *) BN_sqr() bug fix.
2643 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
2645 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
2646 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
2647 followed by modular reduction.
2648 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
2650 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
2651 equivalent based on BN_pseudo_rand() instead of BN_rand().
2654 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
2655 This function was broken, as the check for a new client hello message
2656 to handle SGC did not allow these large messages.
2657 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
2660 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
2663 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
2664 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
2667 *) Rework the configuration and shared library support for Tru64 Unix.
2668 The configuration part makes use of modern compiler features and
2669 still retains old compiler behavior for those that run older versions
2670 of the OS. The shared library support part includes a variant that
2671 uses the RPATH feature, and is available through the special
2672 configuration target "alpha-cc-rpath", which will never be selected
2674 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
2676 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
2677 with the same message size as in ssl3_get_certificate_request().
2678 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
2679 messages might inadvertently be reject as too long.
2680 [Petr Lampa <lampa@fee.vutbr.cz>]
2682 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
2685 *) Modified SSL library such that the verify_callback that has been set
2686 specificly for an SSL object with SSL_set_verify() is actually being
2687 used. Before the change, a verify_callback set with this function was
2688 ignored and the verify_callback() set in the SSL_CTX at the time of
2689 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
2690 to allow the necessary settings.
2693 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
2694 explicitly to NULL, as at least on Solaris 8 this seems not always to be
2695 done automatically (in contradiction to the requirements of the C
2696 standard). This made problems when used from OpenSSH.
2699 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
2700 dh->length and always used
2702 BN_rand_range(priv_key, dh->p).
2704 BN_rand_range() is not necessary for Diffie-Hellman, and this
2705 specific range makes Diffie-Hellman unnecessarily inefficient if
2706 dh->length (recommended exponent length) is much smaller than the
2707 length of dh->p. We could use BN_rand_range() if the order of
2708 the subgroup was stored in the DH structure, but we only have
2713 BN_rand(priv_key, l, ...)
2715 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
2721 RSA_eay_public_encrypt
2722 RSA_eay_private_decrypt
2723 RSA_eay_private_encrypt (signing)
2724 RSA_eay_public_decrypt (signature verification)
2726 (default implementations for RSA_public_encrypt,
2727 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
2728 always reject numbers >= n.
2731 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
2732 to synchronize access to 'locking_thread'. This is necessary on
2733 systems where access to 'locking_thread' (an 'unsigned long'
2734 variable) is not atomic.
2737 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
2738 *before* setting the 'crypto_lock_rand' flag. The previous code had
2739 a race condition if 0 is a valid thread ID.
2740 [Travis Vitek <vitek@roguewave.com>]
2742 *) Add support for shared libraries under Irix.
2743 [Albert Chin-A-Young <china@thewrittenword.com>]
2745 *) Add configuration option to build on Linux on both big-endian and
2747 [Ralf Baechle <ralf@uni-koblenz.de>]
2749 *) Add the possibility to create shared libraries on HP-UX.
2752 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
2754 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
2755 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
2756 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
2757 PRNG state recovery was possible based on the output of
2758 one PRNG request appropriately sized to gain knowledge on
2759 'md' followed by enough consecutive 1-byte PRNG requests
2760 to traverse all of 'state'.
2762 1. When updating 'md_local' (the current thread's copy of 'md')
2763 during PRNG output generation, hash all of the previous
2764 'md_local' value, not just the half used for PRNG output.
2766 2. Make the number of bytes from 'state' included into the hash
2767 independent from the number of PRNG bytes requested.
2769 The first measure alone would be sufficient to avoid
2770 Markku-Juhani's attack. (Actually it had never occurred
2771 to me that the half of 'md_local' used for chaining was the
2772 half from which PRNG output bytes were taken -- I had always
2773 assumed that the secret half would be used.) The second
2774 measure makes sure that additional data from 'state' is never
2775 mixed into 'md_local' in small portions; this heuristically
2776 further strengthens the PRNG.
2779 *) Fix crypto/bn/asm/mips3.s.
2782 *) When only the key is given to "enc", the IV is undefined. Print out
2783 an error message in this case.
2786 *) Handle special case when X509_NAME is empty in X509 printing routines.
2789 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
2790 positive and less than q.
2793 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
2794 used: it isn't thread safe and the add_lock_callback should handle
2796 [Paul Rose <Paul.Rose@bridge.com>]
2798 *) Verify that incoming data obeys the block size in
2799 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
2803 [Ulf Möller, Bodo Möller]
2805 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
2806 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
2807 when fixing the server behaviour for backwards-compatible 'client
2808 hello' messages. (Note that the attack is impractical against
2809 SSL 3.0 and TLS 1.0 anyway because length and version checking
2810 means that the probability of guessing a valid ciphertext is
2811 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
2814 Before 0.9.5, the countermeasure (hide the error by generating a
2815 random 'decryption result') did not work properly because
2816 ERR_clear_error() was missing, meaning that SSL_get_error() would
2817 detect the supposedly ignored error.
2819 Both problems are now fixed.
2822 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
2823 (previously it was 1024).
2826 *) Fix for compatibility mode trust settings: ignore trust settings
2827 unless some valid trust or reject settings are present.
2830 *) Fix for blowfish EVP: its a variable length cipher.
2833 *) Fix various bugs related to DSA S/MIME verification. Handle missing
2834 parameters in DSA public key structures and return an error in the
2835 DSA routines if parameters are absent.
2838 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
2839 in the current directory if neither $RANDFILE nor $HOME was set.
2840 RAND_file_name() in 0.9.6a returned NULL in this case. This has
2841 caused some confusion to Windows users who haven't defined $HOME.
2842 Thus RAND_file_name() is changed again: e_os.h can define a
2843 DEFAULT_HOME, which will be used if $HOME is not set.
2844 For Windows, we use "C:"; on other platforms, we still require
2845 environment variables.
2847 *) Move 'if (!initialized) RAND_poll()' into regions protected by
2848 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
2849 having multiple threads call RAND_poll() concurrently.
2852 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
2853 combination of a flag and a thread ID variable.
2854 Otherwise while one thread is in ssleay_rand_bytes (which sets the
2855 flag), *other* threads can enter ssleay_add_bytes without obeying
2856 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
2857 that they do not hold after the first thread unsets add_do_not_lock).
2860 *) Change bctest again: '-x' expressions are not available in all
2864 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
2866 *) Fix a couple of memory leaks in PKCS7_dataDecode()
2867 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
2869 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
2870 the default extension for executables, if any. Also, make the perl
2871 scripts that use symlink() to test if it really exists and use "cp"
2872 if it doesn't. All this made OpenSSL compilable and installable in
2876 *) Fix for asn1_GetSequence() for indefinite length constructed data.
2877 If SEQUENCE is length is indefinite just set c->slen to the total
2878 amount of data available.
2879 [Steve Henson, reported by shige@FreeBSD.org]
2880 [This change does not apply to 0.9.7.]
2882 *) Change bctest to avoid here-documents inside command substitution
2883 (workaround for FreeBSD /bin/sh bug).
2884 For compatibility with Ultrix, avoid shell functions (introduced
2885 in the bctest version that searches along $PATH).
2888 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
2889 with des_encrypt() defined on some operating systems, like Solaris
2893 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
2894 On the Importance of Eliminating Errors in Cryptographic
2895 Computations, J. Cryptology 14 (2001) 2, 101-119,
2896 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
2899 *) MIPS assembler BIGNUM division bug fix.
2902 *) Disabled incorrect Alpha assembler code.
2905 *) Fix PKCS#7 decode routines so they correctly update the length
2906 after reading an EOC for the EXPLICIT tag.
2908 [This change does not apply to 0.9.7.]
2910 *) Fix bug in PKCS#12 key generation routines. This was triggered
2911 if a 3DES key was generated with a 0 initial byte. Include
2912 PKCS12_BROKEN_KEYGEN compilation option to retain the old
2913 (but broken) behaviour.
2916 *) Enhance bctest to search for a working bc along $PATH and print
2918 [Tim Rice <tim@multitalents.net> via Richard Levitte]
2920 *) Fix memory leaks in err.c: free err_data string if necessary;
2921 don't write to the wrong index in ERR_set_error_data.
2924 *) Implement ssl23_peek (analogous to ssl23_read), which previously
2928 *) Replace rdtsc with _emit statements for VC++ version 5.
2929 [Jeremy Cooper <jeremy@baymoo.org>]
2931 *) Make it possible to reuse SSLv2 sessions.
2934 *) In copy_email() check for >= 0 as a return value for
2935 X509_NAME_get_index_by_NID() since 0 is a valid index.
2936 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
2938 *) Avoid coredump with unsupported or invalid public keys by checking if
2939 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
2940 PKCS7_verify() fails with non detached data.
2943 *) Don't use getenv in library functions when run as setuid/setgid.
2944 New function OPENSSL_issetugid().
2947 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
2948 due to incorrect handling of multi-threading:
2950 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
2952 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
2954 3. Count how many times MemCheck_off() has been called so that
2955 nested use can be treated correctly. This also avoids
2956 inband-signalling in the previous code (which relied on the
2957 assumption that thread ID 0 is impossible).
2960 *) Add "-rand" option also to s_client and s_server.
2963 *) Fix CPU detection on Irix 6.x.
2964 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
2965 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
2967 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
2970 [This change does not apply to 0.9.7.]
2972 *) Use the cached encoding of an X509_NAME structure rather than
2973 copying it. This is apparently the reason for the libsafe "errors"
2974 but the code is actually correct.
2977 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
2978 Bleichenbacher's DSA attack.
2979 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
2980 to be set and top=0 forces the highest bit to be set; top=-1 is new
2981 and leaves the highest bit random.
2982 [Ulf Moeller, Bodo Moeller]
2984 *) In the NCONF_...-based implementations for CONF_... queries
2985 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
2986 a temporary CONF structure with the data component set to NULL
2987 (which gives segmentation faults in lh_retrieve).
2988 Instead, use NULL for the CONF pointer in CONF_get_string and
2989 CONF_get_number (which may use environment variables) and directly
2990 return NULL from CONF_get_section.
2993 *) Fix potential buffer overrun for EBCDIC.
2996 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
2997 keyUsage if basicConstraints absent for a CA.
3000 *) Make SMIME_write_PKCS7() write mail header values with a format that
3001 is more generally accepted (no spaces before the semicolon), since
3002 some programs can't parse those values properly otherwise. Also make
3003 sure BIO's that break lines after each write do not create invalid
3007 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
3008 macros previously used would not encode an empty SEQUENCE OF
3009 and break the signature.
3011 [This change does not apply to 0.9.7.]
3013 *) Zero the premaster secret after deriving the master secret in
3017 *) Add some EVP_add_digest_alias registrations (as found in
3018 OpenSSL_add_all_digests()) to SSL_library_init()
3019 aka OpenSSL_add_ssl_algorithms(). This provides improved
3020 compatibility with peers using X.509 certificates
3021 with unconventional AlgorithmIdentifier OIDs.
3024 *) Fix for Irix with NO_ASM.
3025 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3027 *) ./config script fixes.
3028 [Ulf Moeller, Richard Levitte]
3030 *) Fix 'openssl passwd -1'.
3033 *) Change PKCS12_key_gen_asc() so it can cope with non null
3034 terminated strings whose length is passed in the passlen
3035 parameter, for example from PEM callbacks. This was done
3036 by adding an extra length parameter to asc2uni().
3037 [Steve Henson, reported by <oddissey@samsung.co.kr>]
3039 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
3040 call failed, free the DSA structure.
3043 *) Fix to uni2asc() to cope with zero length Unicode strings.
3044 These are present in some PKCS#12 files.
3047 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
3048 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
3049 when writing a 32767 byte record.
3050 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
3052 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
3053 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
3055 (RSA objects have a reference count access to which is protected
3056 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
3057 so they are meant to be shared between threads.)
3058 [Bodo Moeller, Geoff Thorpe; original patch submitted by
3059 "Reddie, Steven" <Steven.Reddie@ca.com>]
3061 *) Fix a deadlock in CRYPTO_mem_leaks().
3064 *) Use better test patterns in bntest.
3067 *) rand_win.c fix for Borland C.
3070 *) BN_rshift bugfix for n == 0.
3073 *) Add a 'bctest' script that checks for some known 'bc' bugs
3074 so that 'make test' does not abort just because 'bc' is broken.
3077 *) Store verify_result within SSL_SESSION also for client side to
3078 avoid potential security hole. (Re-used sessions on the client side
3079 always resulted in verify_result==X509_V_OK, not using the original
3080 result of the server certificate verification.)
3083 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
3084 SSL3_RT_APPLICATION_DATA, return 0.
3085 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
3089 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
3090 releases, have been re-implemented by renaming the previous
3091 implementations of ssl2_read and ssl3_read to ssl2_read_internal
3092 and ssl3_read_internal, respectively, and adding 'peek' parameters
3093 to them. The new ssl[23]_{read,peek} functions are calls to
3094 ssl[23]_read_internal with the 'peek' flag set appropriately.
3095 A 'peek' parameter has also been added to ssl3_read_bytes, which
3096 does the actual work for ssl3_read_internal.
3099 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
3100 the method-specific "init()" handler. Also clean up ex_data after
3101 calling the method-specific "finish()" handler. Previously, this was
3102 happening the other way round.
3105 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
3106 The previous value, 12, was not always sufficient for BN_mod_exp().
3109 *) Make sure that shared libraries get the internal name engine with
3110 the full version number and not just 0. This should mark the
3111 shared libraries as not backward compatible. Of course, this should
3112 be changed again when we can guarantee backward binary compatibility.
3115 *) Fix typo in get_cert_by_subject() in by_dir.c
3116 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
3118 *) Rework the system to generate shared libraries:
3120 - Make note of the expected extension for the shared libraries and
3121 if there is a need for symbolic links from for example libcrypto.so.0
3122 to libcrypto.so.0.9.7. There is extended info in Configure for
3125 - Make as few rebuilds of the shared libraries as possible.
3127 - Still avoid linking the OpenSSL programs with the shared libraries.
3129 - When installing, install the shared libraries separately from the
3133 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
3135 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
3136 and not in SSL_clear because the latter is also used by the
3137 accept/connect functions; previously, the settings made by
3138 SSL_set_read_ahead would be lost during the handshake.
3139 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
3141 *) Correct util/mkdef.pl to be selective about disabled algorithms.
3142 Previously, it would create entries for disableed algorithms no
3146 *) Added several new manual pages for SSL_* function.
3149 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
3151 *) In ssl23_get_client_hello, generate an error message when faced
3152 with an initial SSL 3.0/TLS record that is too small to contain the
3153 first two bytes of the ClientHello message, i.e. client_version.
3154 (Note that this is a pathologic case that probably has never happened
3155 in real life.) The previous approach was to use the version number
3156 from the record header as a substitute; but our protocol choice
3157 should not depend on that one because it is not authenticated
3158 by the Finished messages.
3161 *) More robust randomness gathering functions for Windows.
3162 [Jeffrey Altman <jaltman@columbia.edu>]
3164 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
3165 not set then we don't setup the error code for issuer check errors
3166 to avoid possibly overwriting other errors which the callback does
3167 handle. If an application does set the flag then we assume it knows
3168 what it is doing and can handle the new informational codes
3172 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
3173 a general "ANY" type, as such it should be able to decode anything
3174 including tagged types. However it didn't check the class so it would
3175 wrongly interpret tagged types in the same way as their universal
3176 counterpart and unknown types were just rejected. Changed so that the
3177 tagged and unknown types are handled in the same way as a SEQUENCE:
3178 that is the encoding is stored intact. There is also a new type
3179 "V_ASN1_OTHER" which is used when the class is not universal, in this
3180 case we have no idea what the actual type is so we just lump them all
3184 *) On VMS, stdout may very well lead to a file that is written to
3185 in a record-oriented fashion. That means that every write() will
3186 write a separate record, which will be read separately by the
3187 programs trying to read from it. This can be very confusing.
3189 The solution is to put a BIO filter in the way that will buffer
3190 text until a linefeed is reached, and then write everything a
3191 line at a time, so every record written will be an actual line,
3192 not chunks of lines and not (usually doesn't happen, but I've
3193 seen it once) several lines in one record. BIO_f_linebuffer() is
3196 Currently, it's a VMS-only method, because that's where it has
3197 been tested well enough.
3200 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
3201 it can return incorrect results.
3202 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
3203 but it was in 0.9.6-beta[12].)
3206 *) Disable the check for content being present when verifying detached
3207 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
3208 include zero length content when signing messages.
3211 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
3212 BIO_ctrl (for BIO pairs).
3215 *) Add DSO method for VMS.
3218 *) Bug fix: Montgomery multiplication could produce results with the
3222 *) Add RPM specification openssl.spec and modify it to build three
3223 packages. The default package contains applications, application
3224 documentation and run-time libraries. The devel package contains
3225 include files, static libraries and function documentation. The
3226 doc package contains the contents of the doc directory. The original
3227 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
3230 *) Add a large number of documentation files for many SSL routines.
3231 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3233 *) Add a configuration entry for Sony News 4.
3234 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
3236 *) Don't set the two most significant bits to one when generating a
3237 random number < q in the DSA library.
3240 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
3241 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
3242 the underlying transport is blocking) if a handshake took place.
3243 (The default behaviour is needed by applications such as s_client
3244 and s_server that use select() to determine when to use SSL_read;
3245 but for applications that know in advance when to expect data, it
3246 just makes things more complicated.)
3249 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
3253 *) Add a few more EBCDIC conditionals that make `req' and `x509'
3254 work better on such systems.
3255 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3257 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
3258 Update PKCS12_parse() so it copies the friendlyName and the
3259 keyid to the certificates aux info.
3262 *) Fix bug in PKCS7_verify() which caused an infinite loop
3263 if there was more than one signature.
3264 [Sven Uszpelkat <su@celocom.de>]
3266 *) Major change in util/mkdef.pl to include extra information
3267 about each symbol, as well as presentig variables as well
3268 as functions. This change means that there's n more need
3269 to rebuild the .num files when some algorithms are excluded.
3272 *) Allow the verify time to be set by an application,
3273 rather than always using the current time.
3276 *) Phase 2 verify code reorganisation. The certificate
3277 verify code now looks up an issuer certificate by a
3278 number of criteria: subject name, authority key id
3279 and key usage. It also verifies self signed certificates
3280 by the same criteria. The main comparison function is
3281 X509_check_issued() which performs these checks.
3283 Lot of changes were necessary in order to support this
3284 without completely rewriting the lookup code.
3286 Authority and subject key identifier are now cached.
3288 The LHASH 'certs' is X509_STORE has now been replaced
3289 by a STACK_OF(X509_OBJECT). This is mainly because an
3290 LHASH can't store or retrieve multiple objects with
3291 the same hash value.
3293 As a result various functions (which were all internal
3294 use only) have changed to handle the new X509_STORE
3295 structure. This will break anything that messed round
3296 with X509_STORE internally.
3298 The functions X509_STORE_add_cert() now checks for an
3299 exact match, rather than just subject name.
3301 The X509_STORE API doesn't directly support the retrieval
3302 of multiple certificates matching a given criteria, however
3303 this can be worked round by performing a lookup first
3304 (which will fill the cache with candidate certificates)
3305 and then examining the cache for matches. This is probably
3306 the best we can do without throwing out X509_LOOKUP
3307 entirely (maybe later...).
3309 The X509_VERIFY_CTX structure has been enhanced considerably.
3311 All certificate lookup operations now go via a get_issuer()
3312 callback. Although this currently uses an X509_STORE it
3313 can be replaced by custom lookups. This is a simple way
3314 to bypass the X509_STORE hackery necessary to make this
3315 work and makes it possible to use more efficient techniques
3316 in future. A very simple version which uses a simple
3317 STACK for its trusted certificate store is also provided
3318 using X509_STORE_CTX_trusted_stack().
3320 The verify_cb() and verify() callbacks now have equivalents
3321 in the X509_STORE_CTX structure.
3323 X509_STORE_CTX also has a 'flags' field which can be used
3324 to customise the verify behaviour.
3327 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
3328 excludes S/MIME capabilities.
3331 *) When a certificate request is read in keep a copy of the
3332 original encoding of the signed data and use it when outputing
3333 again. Signatures then use the original encoding rather than
3334 a decoded, encoded version which may cause problems if the
3335 request is improperly encoded.
3338 *) For consistency with other BIO_puts implementations, call
3339 buffer_write(b, ...) directly in buffer_puts instead of calling
3342 In BIO_puts, increment b->num_write as in BIO_write.
3343 [Peter.Sylvester@EdelWeb.fr]
3345 *) Fix BN_mul_word for the case where the word is 0. (We have to use
3346 BN_zero, we may not return a BIGNUM with an array consisting of
3350 *) Avoid calling abort() from within the library when problems are
3351 detected, except if preprocessor symbols have been defined
3352 (such as REF_CHECK, BN_DEBUG etc.).
3355 *) New openssl application 'rsautl'. This utility can be
3356 used for low level RSA operations. DER public key
3357 BIO/fp routines also added.
3360 *) New Configure entry and patches for compiling on QNX 4.
3361 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
3363 *) A demo state-machine implementation was sponsored by
3364 Nuron (http://www.nuron.com/) and is now available in
3365 demos/state_machine.
3368 *) New options added to the 'dgst' utility for signature
3369 generation and verification.
3372 *) Unrecognized PKCS#7 content types are now handled via a
3373 catch all ASN1_TYPE structure. This allows unsupported
3374 types to be stored as a "blob" and an application can
3375 encode and decode it manually.
3378 *) Fix various signed/unsigned issues to make a_strex.c
3380 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
3382 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
3383 length if passed a buffer. ASN1_INTEGER_to_BN failed
3384 if passed a NULL BN and its argument was negative.
3385 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
3387 *) Modification to PKCS#7 encoding routines to output definite
3388 length encoding. Since currently the whole structures are in
3389 memory there's not real point in using indefinite length
3390 constructed encoding. However if OpenSSL is compiled with
3391 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
3394 *) Added BIO_vprintf() and BIO_vsnprintf().
3397 *) Added more prefixes to parse for in the the strings written
3398 through a logging bio, to cover all the levels that are available
3399 through syslog. The prefixes are now:
3401 PANIC, EMERG, EMR => LOG_EMERG
3402 ALERT, ALR => LOG_ALERT
3403 CRIT, CRI => LOG_CRIT
3404 ERROR, ERR => LOG_ERR
3405 WARNING, WARN, WAR => LOG_WARNING
3406 NOTICE, NOTE, NOT => LOG_NOTICE
3407 INFO, INF => LOG_INFO
3408 DEBUG, DBG => LOG_DEBUG
3410 and as before, if none of those prefixes are present at the
3411 beginning of the string, LOG_ERR is chosen.
3413 On Win32, the LOG_* levels are mapped according to this:
3415 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
3416 LOG_WARNING => EVENTLOG_WARNING_TYPE
3417 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
3421 *) Made it possible to reconfigure with just the configuration
3422 argument "reconf" or "reconfigure". The command line arguments
3423 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
3424 and are retrieved from there when reconfiguring.
3428 [Assar Westerlund <assar@sics.se>, Richard Levitte]
3430 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
3433 *) The obj_dat.pl script was messing up the sorting of object
3434 names. The reason was that it compared the quoted version
3435 of strings as a result "OCSP" > "OCSP Signing" because
3436 " > SPACE. Changed script to store unquoted versions of
3437 names and add quotes on output. It was also omitting some
3438 names from the lookup table if they were given a default
3439 value (that is if SN is missing it is given the same
3440 value as LN and vice versa), these are now added on the
3441 grounds that if an object has a name we should be able to
3442 look it up. Finally added warning output when duplicate
3443 short or long names are found.
3446 *) Changes needed for Tandem NSK.
3447 [Scott Uroff <scott@xypro.com>]
3449 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
3450 RSA_padding_check_SSLv23(), special padding was never detected
3451 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
3452 version rollback attacks was not effective.
3454 In s23_clnt.c, don't use special rollback-attack detection padding
3455 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
3456 client; similarly, in s23_srvr.c, don't do the rollback check if
3457 SSL 2.0 is the only protocol enabled in the server.
3460 *) Make it possible to get hexdumps of unprintable data with 'openssl
3461 asn1parse'. By implication, the functions ASN1_parse_dump() and
3462 BIO_dump_indent() are added.
3465 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
3466 these print out strings and name structures based on various
3467 flags including RFC2253 support and proper handling of
3468 multibyte characters. Added options to the 'x509' utility
3469 to allow the various flags to be set.
3472 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
3473 Also change the functions X509_cmp_current_time() and
3474 X509_gmtime_adj() work with an ASN1_TIME structure,
3475 this will enable certificates using GeneralizedTime in validity
3476 dates to be checked.
3479 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
3480 negative public key encodings) on by default,
3481 NO_NEG_PUBKEY_BUG can be set to disable it.
3484 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
3485 content octets. An i2c_ASN1_OBJECT is unnecessary because
3486 the encoding can be trivially obtained from the structure.
3489 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
3490 not read locks (CRYPTO_r_[un]lock).
3493 *) A first attempt at creating official support for shared
3494 libraries through configuration. I've kept it so the
3495 default is static libraries only, and the OpenSSL programs
3496 are always statically linked for now, but there are
3497 preparations for dynamic linking in place.
3498 This has been tested on Linux and Tru64.
3501 *) Randomness polling function for Win9x, as described in:
3502 Peter Gutmann, Software Generation of Practically Strong
3506 *) Fix so PRNG is seeded in req if using an already existing
3510 *) New options to smime application. -inform and -outform
3511 allow alternative formats for the S/MIME message including
3512 PEM and DER. The -content option allows the content to be
3513 specified separately. This should allow things like Netscape
3514 form signing output easier to verify.
3517 *) Fix the ASN1 encoding of tags using the 'long form'.
3520 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
3521 STRING types. These convert content octets to and from the
3522 underlying type. The actual tag and length octets are
3523 already assumed to have been read in and checked. These
3524 are needed because all other string types have virtually
3525 identical handling apart from the tag. By having versions
3526 of the ASN1 functions that just operate on content octets
3527 IMPLICIT tagging can be handled properly. It also allows
3528 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
3529 and ASN1_INTEGER are identical apart from the tag.
3532 *) Change the handling of OID objects as follows:
3534 - New object identifiers are inserted in objects.txt, following
3535 the syntax given in objects.README.
3536 - objects.pl is used to process obj_mac.num and create a new
3538 - obj_dat.pl is used to create a new obj_dat.h, using the data in
3541 This is currently kind of a hack, and the perl code in objects.pl
3542 isn't very elegant, but it works as I intended. The simplest way
3543 to check that it worked correctly is to look in obj_dat.h and
3544 check the array nid_objs and make sure the objects haven't moved
3545 around (this is important!). Additions are OK, as well as
3546 consistent name changes.
3549 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
3552 *) Addition of the command line parameter '-rand file' to 'openssl req'.
3553 The given file adds to whatever has already been seeded into the
3554 random pool through the RANDFILE configuration file option or
3555 environment variable, or the default random state file.
3558 *) mkstack.pl now sorts each macro group into lexical order.
3559 Previously the output order depended on the order the files
3560 appeared in the directory, resulting in needless rewriting
3564 *) Patches to make OpenSSL compile under Win32 again. Mostly
3565 work arounds for the VC++ problem that it treats func() as
3566 func(void). Also stripped out the parts of mkdef.pl that
3567 added extra typesafe functions: these no longer exist.
3570 *) Reorganisation of the stack code. The macros are now all
3571 collected in safestack.h . Each macro is defined in terms of
3572 a "stack macro" of the form SKM_<name>(type, a, b). The
3573 DEBUG_SAFESTACK is now handled in terms of function casts,
3574 this has the advantage of retaining type safety without the
3575 use of additional functions. If DEBUG_SAFESTACK is not defined
3576 then the non typesafe macros are used instead. Also modified the
3577 mkstack.pl script to handle the new form. Needs testing to see
3578 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
3579 the default if no major problems. Similar behaviour for ASN1_SET_OF
3580 and PKCS12_STACK_OF.
3583 *) When some versions of IIS use the 'NET' form of private key the
3584 key derivation algorithm is different. Normally MD5(password) is
3585 used as a 128 bit RC4 key. In the modified case
3586 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
3587 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
3588 as the old Netscape_RSA functions except they have an additional
3589 'sgckey' parameter which uses the modified algorithm. Also added
3590 an -sgckey command line option to the rsa utility. Thanks to
3591 Adrian Peck <bertie@ncipher.com> for posting details of the modified
3592 algorithm to openssl-dev.
3595 *) The evp_local.h macros were using 'c.##kname' which resulted in
3596 invalid expansion on some systems (SCO 5.0.5 for example).
3597 Corrected to 'c.kname'.
3598 [Phillip Porch <root@theporch.com>]
3600 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
3601 a STACK of email addresses from a certificate or request, these look
3602 in the subject name and the subject alternative name extensions and
3603 omit any duplicate addresses.
3606 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
3607 This makes DSA verification about 2 % faster.
3610 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
3611 (meaning that now 2^5 values will be precomputed, which is only 4 KB
3612 plus overhead for 1024 bit moduli).
3613 This makes exponentiations about 0.5 % faster for 1024 bit
3614 exponents (as measured by "openssl speed rsa2048").
3617 *) Rename memory handling macros to avoid conflicts with other
3619 Malloc => OPENSSL_malloc
3620 Malloc_locked => OPENSSL_malloc_locked
3621 Realloc => OPENSSL_realloc
3622 Free => OPENSSL_free
3625 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
3626 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
3629 *) CygWin32 support.
3630 [John Jarvie <jjarvie@newsguy.com>]
3632 *) The type-safe stack code has been rejigged. It is now only compiled
3633 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
3634 by default all type-specific stack functions are "#define"d back to
3635 standard stack functions. This results in more streamlined output
3636 but retains the type-safety checking possibilities of the original
3640 *) The STACK code has been cleaned up, and certain type declarations
3641 that didn't make a lot of sense have been brought in line. This has
3642 also involved a cleanup of sorts in safestack.h to more correctly
3643 map type-safe stack functions onto their plain stack counterparts.
3644 This work has also resulted in a variety of "const"ifications of
3645 lots of the code, especially "_cmp" operations which should normally
3646 be prototyped with "const" parameters anyway.
3649 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
3650 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
3651 (The PRNG state consists of two parts, the large pool 'state' and 'md',
3652 where all of 'md' is used each time the PRNG is used, but 'state'
3653 is used only indexed by a cyclic counter. As entropy may not be
3654 well distributed from the beginning, 'md' is important as a
3655 chaining variable. However, the output function chains only half
3656 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
3657 all of 'md', and seeding with STATE_SIZE dummy bytes will result
3658 in all of 'state' being rewritten, with the new values depending
3659 on virtually all of 'md'. This overcomes the 80 bit limitation.)
3662 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
3663 the handshake is continued after ssl_verify_cert_chain();
3664 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
3665 can lead to 'unexplainable' connection aborts later.
3666 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
3668 *) Major EVP API cipher revision.
3669 Add hooks for extra EVP features. This allows various cipher
3670 parameters to be set in the EVP interface. Support added for variable
3671 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
3672 setting of RC2 and RC5 parameters.
3674 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
3677 Remove lots of duplicated code from the EVP library. For example *every*
3678 cipher init() function handles the 'iv' in the same way according to the
3679 cipher mode. They also all do nothing if the 'key' parameter is NULL and
3680 for CFB and OFB modes they zero ctx->num.
3682 New functionality allows removal of S/MIME code RC2 hack.
3684 Most of the routines have the same form and so can be declared in terms
3687 By shifting this to the top level EVP_CipherInit() it can be removed from
3688 all individual ciphers. If the cipher wants to handle IVs or keys
3689 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
3692 Change lots of functions like EVP_EncryptUpdate() to now return a
3693 value: although software versions of the algorithms cannot fail
3694 any installed hardware versions can.
3697 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
3698 this option is set, tolerate broken clients that send the negotiated
3699 protocol version number instead of the requested protocol version
3703 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
3704 i.e. non-zero for export ciphersuites, zero otherwise.
3705 Previous versions had this flag inverted, inconsistent with
3706 rsa_tmp_cb (..._TMP_RSA_CB).
3707 [Bodo Moeller; problem reported by Amit Chopra]
3709 *) Add missing DSA library text string. Work around for some IIS
3710 key files with invalid SEQUENCE encoding.
3713 *) Add a document (doc/standards.txt) that list all kinds of standards
3714 and so on that are implemented in OpenSSL.
3717 *) Enhance c_rehash script. Old version would mishandle certificates
3718 with the same subject name hash and wouldn't handle CRLs at all.
3719 Added -fingerprint option to crl utility, to support new c_rehash
3723 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
3726 *) Fix for SSL server purpose checking. Server checking was
3727 rejecting certificates which had extended key usage present
3728 but no ssl client purpose.
3729 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
3731 *) Make PKCS#12 code work with no password. The PKCS#12 spec
3732 is a little unclear about how a blank password is handled.
3733 Since the password in encoded as a BMPString with terminating
3734 double NULL a zero length password would end up as just the
3735 double NULL. However no password at all is different and is
3736 handled differently in the PKCS#12 key generation code. NS
3737 treats a blank password as zero length. MSIE treats it as no
3738 password on export: but it will try both on import. We now do
3739 the same: PKCS12_parse() tries zero length and no password if
3740 the password is set to "" or NULL (NULL is now a valid password:
3741 it wasn't before) as does the pkcs12 application.
3744 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
3745 perror when PEM_read_bio_X509_REQ fails, the error message must
3746 be obtained from the error queue.
3749 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
3750 it in ERR_remove_state if appropriate, and change ERR_get_state
3751 accordingly to avoid race conditions (this is necessary because
3752 thread_hash is no longer constant once set).
3755 *) Bugfix for linux-elf makefile.one.
3758 *) RSA_get_default_method() will now cause a default
3759 RSA_METHOD to be chosen if one doesn't exist already.
3760 Previously this was only set during a call to RSA_new()
3761 or RSA_new_method(NULL) meaning it was possible for
3762 RSA_get_default_method() to return NULL.
3765 *) Added native name translation to the existing DSO code
3766 that will convert (if the flag to do so is set) filenames
3767 that are sufficiently small and have no path information
3768 into a canonical native form. Eg. "blah" converted to
3769 "libblah.so" or "blah.dll" etc.
3772 *) New function ERR_error_string_n(e, buf, len) which is like
3773 ERR_error_string(e, buf), but writes at most 'len' bytes
3774 including the 0 terminator. For ERR_error_string_n, 'buf'
3776 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
3778 *) CONF library reworked to become more general. A new CONF
3779 configuration file reader "class" is implemented as well as a
3780 new functions (NCONF_*, for "New CONF") to handle it. The now
3781 old CONF_* functions are still there, but are reimplemented to
3782 work in terms of the new functions. Also, a set of functions
3783 to handle the internal storage of the configuration data is
3784 provided to make it easier to write new configuration file
3785 reader "classes" (I can definitely see something reading a
3786 configuration file in XML format, for example), called _CONF_*,
3787 or "the configuration storage API"...
3789 The new configuration file reading functions are:
3791 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
3792 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
3794 NCONF_default, NCONF_WIN32
3796 NCONF_dump_fp, NCONF_dump_bio
3798 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
3799 NCONF_new creates a new CONF object. This works in the same way
3800 as other interfaces in OpenSSL, like the BIO interface.
3801 NCONF_dump_* dump the internal storage of the configuration file,
3802 which is useful for debugging. All other functions take the same
3803 arguments as the old CONF_* functions wth the exception of the
3804 first that must be a `CONF *' instead of a `LHASH *'.
3806 To make it easer to use the new classes with the old CONF_* functions,
3807 the function CONF_set_default_method is provided.
3810 *) Add '-tls1' option to 'openssl ciphers', which was already
3811 mentioned in the documentation but had not been implemented.
3812 (This option is not yet really useful because even the additional
3813 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
3816 *) Initial DSO code added into libcrypto for letting OpenSSL (and
3817 OpenSSL-based applications) load shared libraries and bind to
3818 them in a portable way.
3819 [Geoff Thorpe, with contributions from Richard Levitte]
3821 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
3823 *) Make sure _lrotl and _lrotr are only used with MSVC.
3825 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
3826 (the default implementation of RAND_status).
3828 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
3829 to '-clrext' (= clear extensions), as intended and documented.
3830 [Bodo Moeller; inconsistency pointed out by Michael Attili
3833 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
3834 was larger than the MD block size.
3835 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
3837 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
3838 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
3839 using the passed key: if the passed key was a private key the result
3840 of X509_print(), for example, would be to print out all the private key
3844 *) des_quad_cksum() byte order bug fix.
3845 [Ulf Möller, using the problem description in krb4-0.9.7, where
3846 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
3848 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
3850 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
3852 *) For easily testing in shell scripts whether some command
3853 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
3854 returns with exit code 0 iff no command of the given name is available.
3855 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
3856 the output goes to stdout and nothing is printed to stderr.
3857 Additional arguments are always ignored.
3859 Since for each cipher there is a command of the same name,
3860 the 'no-cipher' compilation switches can be tested this way.
3862 ('openssl no-XXX' is not able to detect pseudo-commands such
3863 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
3866 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
3869 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
3870 is set; it will be thrown away anyway because each handshake creates
3872 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
3873 to parameters -- in previous versions (since OpenSSL 0.9.3) the
3874 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
3875 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
3878 *) New s_client option -ign_eof: EOF at stdin is ignored, and
3879 'Q' and 'R' lose their special meanings (quit/renegotiate).
3880 This is part of what -quiet does; unlike -quiet, -ign_eof
3881 does not suppress any output.
3884 *) Add compatibility options to the purpose and trust code. The
3885 purpose X509_PURPOSE_ANY is "any purpose" which automatically
3886 accepts a certificate or CA, this was the previous behaviour,
3887 with all the associated security issues.
3889 X509_TRUST_COMPAT is the old trust behaviour: only and
3890 automatically trust self signed roots in certificate store. A
3891 new trust setting X509_TRUST_DEFAULT is used to specify that
3892 a purpose has no associated trust setting and it should instead
3893 use the value in the default purpose.
3896 *) Fix the PKCS#8 DSA private key code so it decodes keys again
3897 and fix a memory leak.
3900 *) In util/mkerr.pl (which implements 'make errors'), preserve
3901 reason strings from the previous version of the .c file, as
3902 the default to have only downcase letters (and digits) in
3903 automatically generated reasons codes is not always appropriate.
3906 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
3907 using strerror. Previously, ERR_reason_error_string() returned
3908 library names as reason strings for SYSerr; but SYSerr is a special
3909 case where small numbers are errno values, not library numbers.
3912 *) Add '-dsaparam' option to 'openssl dhparam' application. This
3913 converts DSA parameters into DH parameters. (When creating parameters,
3914 DSA_generate_parameters is used.)
3917 *) Include 'length' (recommended exponent length) in C code generated
3918 by 'openssl dhparam -C'.
3921 *) The second argument to set_label in perlasm was already being used
3922 so couldn't be used as a "file scope" flag. Moved to third argument
3926 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
3927 instead of RAND_bytes for encryption IVs and salts.
3930 *) Include RAND_status() into RAND_METHOD instead of implementing
3931 it only for md_rand.c Otherwise replacing the PRNG by calling
3932 RAND_set_rand_method would be impossible.
3935 *) Don't let DSA_generate_key() enter an infinite loop if the random
3936 number generation fails.
3939 *) New 'rand' application for creating pseudo-random output.
3942 *) Added configuration support for Linux/IA64
3943 [Rolf Haberrecker <rolf@suse.de>]
3945 *) Assembler module support for Mingw32.
3948 *) Shared library support for HPUX (in shlib/).
3949 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
3951 *) Shared library support for Solaris gcc.
3952 [Lutz Behnke <behnke@trustcenter.de>]
3954 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
3956 *) PKCS7_encrypt() was adding text MIME headers twice because they
3957 were added manually and by SMIME_crlf_copy().
3960 *) In bntest.c don't call BN_rand with zero bits argument.
3961 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
3963 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
3964 case was implemented. This caused BN_div_recp() to fail occasionally.
3967 *) Add an optional second argument to the set_label() in the perl
3968 assembly language builder. If this argument exists and is set
3969 to 1 it signals that the assembler should use a symbol whose
3970 scope is the entire file, not just the current function. This
3971 is needed with MASM which uses the format label:: for this scope.
3972 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
3974 *) Change the ASN1 types so they are typedefs by default. Before
3975 almost all types were #define'd to ASN1_STRING which was causing
3976 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
3980 *) Change names of new functions to the new get1/get0 naming
3981 convention: After 'get1', the caller owns a reference count
3982 and has to call ..._free; 'get0' returns a pointer to some
3983 data structure without incrementing reference counters.
3984 (Some of the existing 'get' functions increment a reference
3985 counter, some don't.)
3986 Similarly, 'set1' and 'add1' functions increase reference
3987 counters or duplicate objects.
3990 *) Allow for the possibility of temp RSA key generation failure:
3991 the code used to assume it always worked and crashed on failure.
3994 *) Fix potential buffer overrun problem in BIO_printf().
3995 [Ulf Möller, using public domain code by Patrick Powell; problem
3996 pointed out by David Sacerdote <das33@cornell.edu>]
3998 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
3999 RAND_egd() and RAND_status(). In the command line application,
4000 the EGD socket can be specified like a seed file using RANDFILE
4004 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
4005 Some CAs (e.g. Verisign) distribute certificates in this form.
4008 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
4009 list to exclude them. This means that no special compilation option
4010 is needed to use anonymous DH: it just needs to be included in the
4014 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
4015 EVP_MD_type. The old functionality is available in a new macro called
4016 EVP_MD_md(). Change code that uses it and update docs.
4019 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
4020 where the 'void *' argument is replaced by a function pointer argument.
4021 Previously 'void *' was abused to point to functions, which works on
4022 many platforms, but is not correct. As these functions are usually
4023 called by macros defined in OpenSSL header files, most source code
4024 should work without changes.
4027 *) <openssl/opensslconf.h> (which is created by Configure) now contains
4028 sections with information on -D... compiler switches used for
4029 compiling the library so that applications can see them. To enable
4030 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
4031 must be defined. E.g.,
4032 #define OPENSSL_ALGORITHM_DEFINES
4033 #include <openssl/opensslconf.h>
4034 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
4035 [Richard Levitte, Ulf and Bodo Möller]
4037 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
4041 *) Change the 'other' type in certificate aux info to a STACK_OF
4042 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
4043 the required ASN1 format: arbitrary types determined by an OID.
4046 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
4047 argument to 'req'. This is not because the function is newer or
4048 better than others it just uses the work 'NEW' in the certificate
4049 request header lines. Some software needs this.
4052 *) Reorganise password command line arguments: now passwords can be
4053 obtained from various sources. Delete the PEM_cb function and make
4054 it the default behaviour: i.e. if the callback is NULL and the
4055 usrdata argument is not NULL interpret it as a null terminated pass
4056 phrase. If usrdata and the callback are NULL then the pass phrase
4057 is prompted for as usual.
4060 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
4061 the support is automatically enabled. The resulting binaries will
4062 autodetect the card and use it if present.
4063 [Ben Laurie and Compaq Inc.]
4065 *) Work around for Netscape hang bug. This sends certificate request
4066 and server done in one record. Since this is perfectly legal in the
4067 SSL/TLS protocol it isn't a "bug" option and is on by default. See
4068 the bugs/SSLv3 entry for more info.
4071 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
4074 *) Add -rand argument to smime and pkcs12 applications and read/write
4078 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
4081 *) Add command line password options to the remaining applications.
4084 *) Bug fix for BN_div_recp() for numerators with an even number of
4088 *) More tests in bntest.c, and changed test_bn output.
4091 *) ./config recognizes MacOS X now.
4094 *) Bug fix for BN_div() when the first words of num and divsor are
4095 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
4098 *) Add support for various broken PKCS#8 formats, and command line
4099 options to produce them.
4102 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
4103 get temporary BIGNUMs from a BN_CTX.
4106 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
4110 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
4111 include a #define from the old name to the new. The original intent
4112 was that statically linked binaries could for example just call
4113 SSLeay_add_all_ciphers() to just add ciphers to the table and not
4114 link with digests. This never worked becayse SSLeay_add_all_digests()
4115 and SSLeay_add_all_ciphers() were in the same source file so calling
4116 one would link with the other. They are now in separate source files.
4119 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
4122 *) Use a less unusual form of the Miller-Rabin primality test (it used
4123 a binary algorithm for exponentiation integrated into the Miller-Rabin
4124 loop, our standard modexp algorithms are faster).
4127 *) Support for the EBCDIC character set completed.
4128 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
4130 *) Source code cleanups: use const where appropriate, eliminate casts,
4131 use void * instead of char * in lhash.
4134 *) Bugfix: ssl3_send_server_key_exchange was not restartable
4135 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
4136 this the server could overwrite ephemeral keys that the client
4140 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
4141 using 50 iterations of the Rabin-Miller test.
4143 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
4144 iterations of the Rabin-Miller test as required by the appendix
4145 to FIPS PUB 186[-1]) instead of DSA_is_prime.
4146 As BN_is_prime_fasttest includes trial division, DSA parameter
4147 generation becomes much faster.
4149 This implies a change for the callback functions in DSA_is_prime
4150 and DSA_generate_parameters: The callback function is called once
4151 for each positive witness in the Rabin-Miller test, not just
4152 occasionally in the inner loop; and the parameters to the
4153 callback function now provide an iteration count for the outer
4154 loop rather than for the current invocation of the inner loop.
4155 DSA_generate_parameters additionally can call the callback
4156 function with an 'iteration count' of -1, meaning that a
4157 candidate has passed the trial division test (when q is generated
4158 from an application-provided seed, trial division is skipped).
4161 *) New function BN_is_prime_fasttest that optionally does trial
4162 division before starting the Rabin-Miller test and has
4163 an additional BN_CTX * argument (whereas BN_is_prime always
4164 has to allocate at least one BN_CTX).
4165 'callback(1, -1, cb_arg)' is called when a number has passed the
4166 trial division stage.
4169 *) Fix for bug in CRL encoding. The validity dates weren't being handled
4173 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
4176 *) New function BN_pseudo_rand().
4179 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
4180 bignum version of BN_from_montgomery() with the working code from
4181 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
4185 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
4186 made it impossible to use the same SSL_SESSION data structure in
4187 SSL2 clients in multiple threads.
4190 *) The return value of RAND_load_file() no longer counts bytes obtained
4191 by stat(). RAND_load_file(..., -1) is new and uses the complete file
4192 to seed the PRNG (previously an explicit byte count was required).
4193 [Ulf Möller, Bodo Möller]
4195 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
4196 used (char *) instead of (void *) and had casts all over the place.
4199 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
4202 *) Retain source code compatibility for BN_prime_checks macro:
4203 BN_is_prime(..., BN_prime_checks, ...) now uses
4204 BN_prime_checks_for_size to determine the appropriate number of
4205 Rabin-Miller iterations.
4208 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
4209 DH_CHECK_P_NOT_SAFE_PRIME.
4210 (Check if this is true? OpenPGP calls them "strong".)
4213 *) Merge the functionality of "dh" and "gendh" programs into a new program
4214 "dhparam". The old programs are retained for now but will handle DH keys
4215 (instead of parameters) in future.
4218 *) Make the ciphers, s_server and s_client programs check the return values
4219 when a new cipher list is set.
4222 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
4223 ciphers. Before when the 56bit ciphers were enabled the sorting was
4226 The syntax for the cipher sorting has been extended to support sorting by
4227 cipher-strength (using the strength_bits hard coded in the tables).
4228 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
4230 Fix a bug in the cipher-command parser: when supplying a cipher command
4231 string with an "undefined" symbol (neither command nor alphanumeric
4232 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
4233 an error is flagged.
4235 Due to the strength-sorting extension, the code of the
4236 ssl_create_cipher_list() function was completely rearranged. I hope that
4237 the readability was also increased :-)
4238 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4240 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
4241 for the first serial number and places 2 in the serial number file. This
4242 avoids problems when the root CA is created with serial number zero and
4243 the first user certificate has the same issuer name and serial number
4247 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
4248 the new code. Add documentation for this stuff.
4251 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
4252 X509_*() to X509at_*() on the grounds that they don't handle X509
4253 structures and behave in an analagous way to the X509v3 functions:
4254 they shouldn't be called directly but wrapper functions should be used
4257 So we also now have some wrapper functions that call the X509at functions
4258 when passed certificate requests. (TO DO: similar things can be done with
4259 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
4260 things. Some of these need some d2i or i2d and print functionality
4261 because they handle more complex structures.)
4264 *) Add missing #ifndefs that caused missing symbols when building libssl
4265 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
4266 NO_RSA in ssl/s2*.c.
4267 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
4269 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
4270 has a return value which indicates the quality of the random data
4271 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
4272 error queue. New function RAND_pseudo_bytes() generates output that is
4273 guaranteed to be unique but not unpredictable. RAND_add is like
4274 RAND_seed, but takes an extra argument for an entropy estimate
4275 (RAND_seed always assumes full entropy).
4278 *) Do more iterations of Rabin-Miller probable prime test (specifically,
4279 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
4280 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
4281 in crypto/bn/bn_prime.c for the complete table). This guarantees a
4282 false-positive rate of at most 2^-80 for random input.
4285 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
4288 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
4289 in the 0.9.5 release), this returns the chain
4290 from an X509_CTX structure with a dup of the stack and all
4291 the X509 reference counts upped: so the stack will exist
4292 after X509_CTX_cleanup() has been called. Modify pkcs12.c
4295 Also make SSL_SESSION_print() print out the verify return
4299 *) Add manpage for the pkcs12 command. Also change the default
4300 behaviour so MAC iteration counts are used unless the new
4301 -nomaciter option is used. This improves file security and
4302 only older versions of MSIE (4.0 for example) need it.
4305 *) Honor the no-xxx Configure options when creating .DEF files.
4308 *) Add PKCS#10 attributes to field table: challengePassword,
4309 unstructuredName and unstructuredAddress. These are taken from
4310 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
4311 international characters are used.
4313 More changes to X509_ATTRIBUTE code: allow the setting of types
4314 based on strings. Remove the 'loc' parameter when adding
4315 attributes because these will be a SET OF encoding which is sorted
4319 *) Initial changes to the 'req' utility to allow request generation
4320 automation. This will allow an application to just generate a template
4321 file containing all the field values and have req construct the
4324 Initial support for X509_ATTRIBUTE handling. Stacks of these are
4325 used all over the place including certificate requests and PKCS#7
4326 structures. They are currently handled manually where necessary with
4327 some primitive wrappers for PKCS#7. The new functions behave in a
4328 manner analogous to the X509 extension functions: they allow
4329 attributes to be looked up by NID and added.
4331 Later something similar to the X509V3 code would be desirable to
4332 automatically handle the encoding, decoding and printing of the
4333 more complex types. The string types like challengePassword can
4334 be handled by the string table functions.
4336 Also modified the multi byte string table handling. Now there is
4337 a 'global mask' which masks out certain types. The table itself
4338 can use the flag STABLE_NO_MASK to ignore the mask setting: this
4339 is useful when for example there is only one permissible type
4340 (as in countryName) and using the mask might result in no valid
4344 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
4345 SSL_get_peer_finished to allow applications to obtain the latest
4346 Finished messages sent to the peer or expected from the peer,
4347 respectively. (SSL_get_peer_finished is usually the Finished message
4348 actually received from the peer, otherwise the protocol will be aborted.)
4350 As the Finished message are message digests of the complete handshake
4351 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
4352 be used for external authentication procedures when the authentication
4353 provided by SSL/TLS is not desired or is not enough.
4356 *) Enhanced support for Alpha Linux is added. Now ./config checks if
4357 the host supports BWX extension and if Compaq C is present on the
4358 $PATH. Just exploiting of the BWX extension results in 20-30%
4359 performance kick for some algorithms, e.g. DES and RC4 to mention
4360 a couple. Compaq C in turn generates ~20% faster code for MD5 and
4364 *) Add support for MS "fast SGC". This is arguably a violation of the
4365 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
4366 weak crypto and after checking the certificate is SGC a second one
4367 with strong crypto. MS SGC stops the first handshake after receiving
4368 the server certificate message and sends a second client hello. Since
4369 a server will typically do all the time consuming operations before
4370 expecting any further messages from the client (server key exchange
4371 is the most expensive) there is little difference between the two.
4373 To get OpenSSL to support MS SGC we have to permit a second client
4374 hello message after we have sent server done. In addition we have to
4375 reset the MAC if we do get this second client hello.
4378 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
4379 if a DER encoded private key is RSA or DSA traditional format. Changed
4380 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
4381 format DER encoded private key. Newer code should use PKCS#8 format which
4382 has the key type encoded in the ASN1 structure. Added DER private key
4383 support to pkcs8 application.
4386 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
4387 ciphersuites has been selected (as required by the SSL 3/TLS 1
4388 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
4389 is set, we interpret this as a request to violate the specification
4390 (the worst that can happen is a handshake failure, and 'correct'
4391 behaviour would result in a handshake failure anyway).
4394 *) In SSL_CTX_add_session, take into account that there might be multiple
4395 SSL_SESSION structures with the same session ID (e.g. when two threads
4396 concurrently obtain them from an external cache).
4397 The internal cache can handle only one SSL_SESSION with a given ID,
4398 so if there's a conflict, we now throw out the old one to achieve
4402 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
4403 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
4404 some routines that use cipher OIDs: some ciphers do not have OIDs
4405 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
4409 *) Simplify the trust setting structure and code. Now we just have
4410 two sequences of OIDs for trusted and rejected settings. These will
4411 typically have values the same as the extended key usage extension
4412 and any application specific purposes.
4414 The trust checking code now has a default behaviour: it will just
4415 check for an object with the same NID as the passed id. Functions can
4416 be provided to override either the default behaviour or the behaviour
4417 for a given id. SSL client, server and email already have functions
4418 in place for compatibility: they check the NID and also return "trusted"
4419 if the certificate is self signed.
4422 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
4423 traditional format into an EVP_PKEY structure.
4426 *) Add a password callback function PEM_cb() which either prompts for
4427 a password if usr_data is NULL or otherwise assumes it is a null
4428 terminated password. Allow passwords to be passed on command line
4429 environment or config files in a few more utilities.
4432 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
4433 keys. Add some short names for PKCS#8 PBE algorithms and allow them
4434 to be specified on the command line for the pkcs8 and pkcs12 utilities.
4435 Update documentation.
4438 *) Support for ASN1 "NULL" type. This could be handled before by using
4439 ASN1_TYPE but there wasn't any function that would try to read a NULL
4440 and produce an error if it couldn't. For compatibility we also have
4441 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
4442 don't allocate anything because they don't need to.
4445 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
4447 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
4449 *) Rebuild of the memory allocation routines used by OpenSSL code and
4450 possibly others as well. The purpose is to make an interface that
4451 provide hooks so anyone can build a separate set of allocation and
4452 deallocation routines to be used by OpenSSL, for example memory
4453 pool implementations, or something else, which was previously hard
4454 since Malloc(), Realloc() and Free() were defined as macros having
4455 the values malloc, realloc and free, respectively (except for Win32
4456 compilations). The same is provided for memory debugging code.
4457 OpenSSL already comes with functionality to find memory leaks, but
4458 this gives people a chance to debug other memory problems.
4460 With these changes, a new set of functions and macros have appeared:
4462 CRYPTO_set_mem_debug_functions() [F]
4463 CRYPTO_get_mem_debug_functions() [F]
4464 CRYPTO_dbg_set_options() [F]
4465 CRYPTO_dbg_get_options() [F]
4466 CRYPTO_malloc_debug_init() [M]
4468 The memory debug functions are NULL by default, unless the library
4469 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
4470 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
4471 gives the standard debugging functions that come with OpenSSL) or
4472 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
4473 provided by the library user) must be used. When the standard
4474 debugging functions are used, CRYPTO_dbg_set_options can be used to
4475 request additional information:
4476 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
4477 the CRYPTO_MDEBUG_xxx macro when compiling the library.
4479 Also, things like CRYPTO_set_mem_functions will always give the
4480 expected result (the new set of functions is used for allocation
4481 and deallocation) at all times, regardless of platform and compiler
4484 To finish it up, some functions that were never use in any other
4485 way than through macros have a new API and new semantic:
4488 CRYPTO_dbg_realloc()
4491 All macros of value have retained their old syntax.
4492 [Richard Levitte and Bodo Moeller]
4494 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
4495 ordering of SMIMECapabilities wasn't in "strength order" and there
4496 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
4500 *) Some ASN1 types with illegal zero length encoding (INTEGER,
4501 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
4502 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
4504 *) Merge in my S/MIME library for OpenSSL. This provides a simple
4505 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
4506 functionality to handle multipart/signed properly) and a utility
4507 called 'smime' to call all this stuff. This is based on code I
4508 originally wrote for Celo who have kindly allowed it to be
4509 included in OpenSSL.
4512 *) Add variants des_set_key_checked and des_set_key_unchecked of
4513 des_set_key (aka des_key_sched). Global variable des_check_key
4514 decides which of these is called by des_set_key; this way
4515 des_check_key behaves as it always did, but applications and
4516 the library itself, which was buggy for des_check_key == 1,
4517 have a cleaner way to pick the version they need.
4520 *) New function PKCS12_newpass() which changes the password of a
4524 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
4525 dynamic mix. In both cases the ids can be used as an index into the
4526 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
4527 functions so they accept a list of the field values and the
4528 application doesn't need to directly manipulate the X509_TRUST
4532 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
4536 *) Modify the way the V3 extension code looks up extensions. This now
4537 works in a similar way to the object code: we have some "standard"
4538 extensions in a static table which is searched with OBJ_bsearch()
4539 and the application can add dynamic ones if needed. The file
4540 crypto/x509v3/ext_dat.h now has the info: this file needs to be
4541 updated whenever a new extension is added to the core code and kept
4542 in ext_nid order. There is a simple program 'tabtest.c' which checks
4543 this. New extensions are not added too often so this file can readily
4544 be maintained manually.
4546 There are two big advantages in doing things this way. The extensions
4547 can be looked up immediately and no longer need to be "added" using
4548 X509V3_add_standard_extensions(): this function now does nothing.
4549 [Side note: I get *lots* of email saying the extension code doesn't
4550 work because people forget to call this function]
4551 Also no dynamic allocation is done unless new extensions are added:
4552 so if we don't add custom extensions there is no need to call
4553 X509V3_EXT_cleanup().
4556 *) Modify enc utility's salting as follows: make salting the default. Add a
4557 magic header, so unsalted files fail gracefully instead of just decrypting
4558 to garbage. This is because not salting is a big security hole, so people
4559 should be discouraged from doing it.
4562 *) Fixes and enhancements to the 'x509' utility. It allowed a message
4563 digest to be passed on the command line but it only used this
4564 parameter when signing a certificate. Modified so all relevant
4565 operations are affected by the digest parameter including the
4566 -fingerprint and -x509toreq options. Also -x509toreq choked if a
4567 DSA key was used because it didn't fix the digest.
4570 *) Initial certificate chain verify code. Currently tests the untrusted
4571 certificates for consistency with the verify purpose (which is set
4572 when the X509_STORE_CTX structure is set up) and checks the pathlength.
4574 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
4575 this is because it will reject chains with invalid extensions whereas
4576 every previous version of OpenSSL and SSLeay made no checks at all.
4578 Trust code: checks the root CA for the relevant trust settings. Trust
4579 settings have an initial value consistent with the verify purpose: e.g.
4580 if the verify purpose is for SSL client use it expects the CA to be
4581 trusted for SSL client use. However the default value can be changed to
4582 permit custom trust settings: one example of this would be to only trust
4583 certificates from a specific "secure" set of CAs.
4585 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
4586 which should be used for version portability: especially since the
4587 verify structure is likely to change more often now.
4589 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
4590 to set them. If not set then assume SSL clients will verify SSL servers
4593 Two new options to the verify program: -untrusted allows a set of
4594 untrusted certificates to be passed in and -purpose which sets the
4595 intended purpose of the certificate. If a purpose is set then the
4596 new chain verify code is used to check extension consistency.
4599 *) Support for the authority information access extension.
4602 *) Modify RSA and DSA PEM read routines to transparently handle
4603 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
4604 public keys in a format compatible with certificate
4605 SubjectPublicKeyInfo structures. Unfortunately there were already
4606 functions called *_PublicKey_* which used various odd formats so
4607 these are retained for compatibility: however the DSA variants were
4608 never in a public release so they have been deleted. Changed dsa/rsa
4609 utilities to handle the new format: note no releases ever handled public
4610 keys so we should be OK.
4612 The primary motivation for this change is to avoid the same fiasco
4613 that dogs private keys: there are several incompatible private key
4614 formats some of which are standard and some OpenSSL specific and
4615 require various evil hacks to allow partial transparent handling and
4616 even then it doesn't work with DER formats. Given the option anything
4617 other than PKCS#8 should be dumped: but the other formats have to
4618 stay in the name of compatibility.
4620 With public keys and the benefit of hindsight one standard format
4621 is used which works with EVP_PKEY, RSA or DSA structures: though
4622 it clearly returns an error if you try to read the wrong kind of key.
4624 Added a -pubkey option to the 'x509' utility to output the public key.
4625 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
4626 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
4627 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
4628 that do the same as the EVP_PKEY_assign_*() except they up the
4629 reference count of the added key (they don't "swallow" the
4633 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
4634 CRLs would fail if the file contained no certificates or no CRLs:
4635 added a new function to read in both types and return the number
4636 read: this means that if none are read it will be an error. The
4637 DER versions of the certificate and CRL reader would always fail
4638 because it isn't possible to mix certificates and CRLs in DER format
4639 without choking one or the other routine. Changed this to just read
4640 a certificate: this is the best we can do. Also modified the code
4641 in apps/verify.c to take notice of return codes: it was previously
4642 attempting to read in certificates from NULL pointers and ignoring
4643 any errors: this is one reason why the cert and CRL reader seemed
4644 to work. It doesn't check return codes from the default certificate
4645 routines: these may well fail if the certificates aren't installed.
4648 *) Code to support otherName option in GeneralName.
4651 *) First update to verify code. Change the verify utility
4652 so it warns if it is passed a self signed certificate:
4653 for consistency with the normal behaviour. X509_verify
4654 has been modified to it will now verify a self signed
4655 certificate if *exactly* the same certificate appears
4656 in the store: it was previously impossible to trust a
4657 single self signed certificate. This means that:
4658 openssl verify ss.pem
4659 now gives a warning about a self signed certificate but
4660 openssl verify -CAfile ss.pem ss.pem
4664 *) For servers, store verify_result in SSL_SESSION data structure
4665 (and add it to external session representation).
4666 This is needed when client certificate verifications fails,
4667 but an application-provided verification callback (set by
4668 SSL_CTX_set_cert_verify_callback) allows accepting the session
4669 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
4670 but returns 1): When the session is reused, we have to set
4671 ssl->verify_result to the appropriate error code to avoid
4673 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
4675 *) Fix a bug in the new PKCS#7 code: it didn't consider the
4676 case in PKCS7_dataInit() where the signed PKCS7 structure
4677 didn't contain any existing data because it was being created.
4678 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
4680 *) Add a salt to the key derivation routines in enc.c. This
4681 forms the first 8 bytes of the encrypted file. Also add a
4682 -S option to allow a salt to be input on the command line.
4685 *) New function X509_cmp(). Oddly enough there wasn't a function
4686 to compare two certificates. We do this by working out the SHA1
4687 hash and comparing that. X509_cmp() will be needed by the trust
4691 *) SSL_get1_session() is like SSL_get_session(), but increments
4692 the reference count in the SSL_SESSION returned.
4693 [Geoff Thorpe <geoff@eu.c2.net>]
4695 *) Fix for 'req': it was adding a null to request attributes.
4696 Also change the X509_LOOKUP and X509_INFO code to handle
4697 certificate auxiliary information.
4700 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
4704 *) Add the possibility to add extra information to the memory leak
4705 detecting output, to form tracebacks, showing from where each
4706 allocation was originated: CRYPTO_push_info("constant string") adds
4707 the string plus current file name and line number to a per-thread
4708 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
4709 is like calling CYRPTO_pop_info() until the stack is empty.
4710 Also updated memory leak detection code to be multi-thread-safe.
4713 *) Add options -text and -noout to pkcs7 utility and delete the
4714 encryption options which never did anything. Update docs.
4717 *) Add options to some of the utilities to allow the pass phrase
4718 to be included on either the command line (not recommended on
4719 OSes like Unix) or read from the environment. Update the
4720 manpages and fix a few bugs.
4723 *) Add a few manpages for some of the openssl commands.
4726 *) Fix the -revoke option in ca. It was freeing up memory twice,
4727 leaking and not finding already revoked certificates.
4730 *) Extensive changes to support certificate auxiliary information.
4731 This involves the use of X509_CERT_AUX structure and X509_AUX
4732 functions. An X509_AUX function such as PEM_read_X509_AUX()
4733 can still read in a certificate file in the usual way but it
4734 will also read in any additional "auxiliary information". By
4735 doing things this way a fair degree of compatibility can be
4736 retained: existing certificates can have this information added
4737 using the new 'x509' options.
4739 Current auxiliary information includes an "alias" and some trust
4740 settings. The trust settings will ultimately be used in enhanced
4741 certificate chain verification routines: currently a certificate
4742 can only be trusted if it is self signed and then it is trusted
4746 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
4747 The problem was that one of the replacement routines had not been working
4748 since SSLeay releases. For now the offending routine has been replaced
4749 with non-optimised assembler. Even so, this now gives around 95%
4750 performance improvement for 1024 bit RSA signs.
4753 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
4754 handling. Most clients have the effective key size in bits equal to
4755 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
4756 A few however don't do this and instead use the size of the decrypted key
4757 to determine the RC2 key length and the AlgorithmIdentifier to determine
4758 the effective key length. In this case the effective key length can still
4759 be 40 bits but the key length can be 168 bits for example. This is fixed
4760 by manually forcing an RC2 key into the EVP_PKEY structure because the
4761 EVP code can't currently handle unusual RC2 key sizes: it always assumes
4762 the key length and effective key length are equal.
4765 *) Add a bunch of functions that should simplify the creation of
4766 X509_NAME structures. Now you should be able to do:
4767 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
4768 and have it automatically work out the correct field type and fill in
4769 the structures. The more adventurous can try:
4770 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
4771 and it will (hopefully) work out the correct multibyte encoding.
4774 *) Change the 'req' utility to use the new field handling and multibyte
4775 copy routines. Before the DN field creation was handled in an ad hoc
4776 way in req, ca, and x509 which was rather broken and didn't support
4777 BMPStrings or UTF8Strings. Since some software doesn't implement
4778 BMPStrings or UTF8Strings yet, they can be enabled using the config file
4779 using the dirstring_type option. See the new comment in the default
4780 openssl.cnf for more info.
4783 *) Make crypto/rand/md_rand.c more robust:
4784 - Assure unique random numbers after fork().
4785 - Make sure that concurrent threads access the global counter and
4786 md serializably so that we never lose entropy in them
4787 or use exactly the same state in multiple threads.
4788 Access to the large state is not always serializable because
4789 the additional locking could be a performance killer, and
4790 md should be large enough anyway.
4793 *) New file apps/app_rand.c with commonly needed functionality
4794 for handling the random seed file.
4796 Use the random seed file in some applications that previously did not:
4798 dsaparam -genkey (which also ignored its '-rand' option),
4801 x509 (when signing).
4802 Except on systems with /dev/urandom, it is crucial to have a random
4803 seed file at least for key creation, DSA signing, and for DH exchanges;
4804 for RSA signatures we could do without one.
4806 gendh and gendsa (unlike genrsa) used to read only the first byte
4807 of each file listed in the '-rand' option. The function as previously
4808 found in genrsa is now in app_rand.c and is used by all programs
4809 that support '-rand'.
4812 *) In RAND_write_file, use mode 0600 for creating files;
4813 don't just chmod when it may be too late.
4816 *) Report an error from X509_STORE_load_locations
4817 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
4820 *) New function ASN1_mbstring_copy() this copies a string in either
4821 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
4822 into an ASN1_STRING type. A mask of permissible types is passed
4823 and it chooses the "minimal" type to use or an error if not type
4827 *) Add function equivalents to the various macros in asn1.h. The old
4828 macros are retained with an M_ prefix. Code inside the library can
4829 use the M_ macros. External code (including the openssl utility)
4830 should *NOT* in order to be "shared library friendly".
4833 *) Add various functions that can check a certificate's extensions
4834 to see if it usable for various purposes such as SSL client,
4835 server or S/MIME and CAs of these types. This is currently
4836 VERY EXPERIMENTAL but will ultimately be used for certificate chain
4837 verification. Also added a -purpose flag to x509 utility to
4838 print out all the purposes.
4841 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
4845 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
4846 for, obtain and decode and extension and obtain its critical flag.
4847 This allows all the necessary extension code to be handled in a
4848 single function call.
4851 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
4852 platforms. See crypto/rc4/rc4_enc.c for further details.
4855 *) New -noout option to asn1parse. This causes no output to be produced
4856 its main use is when combined with -strparse and -out to extract data
4857 from a file (which may not be in ASN.1 format).
4860 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
4861 when producing the local key id.
4862 [Richard Levitte <levitte@stacken.kth.se>]
4864 *) New option -dhparam in s_server. This allows a DH parameter file to be
4865 stated explicitly. If it is not stated then it tries the first server
4866 certificate file. The previous behaviour hard coded the filename
4870 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
4871 a public key to be input or output. For example:
4872 openssl rsa -in key.pem -pubout -out pubkey.pem
4873 Also added necessary DSA public key functions to handle this.
4876 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
4877 in the message. This was handled by allowing
4878 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
4879 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
4881 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
4882 to the end of the strings whereas this didn't. This would cause problems
4883 if strings read with d2i_ASN1_bytes() were later modified.
4884 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
4886 *) Fix for base64 decode bug. When a base64 bio reads only one line of
4887 data and it contains EOF it will end up returning an error. This is
4888 caused by input 46 bytes long. The cause is due to the way base64
4889 BIOs find the start of base64 encoded data. They do this by trying a
4890 trial decode on each line until they find one that works. When they
4891 do a flag is set and it starts again knowing it can pass all the
4892 data directly through the decoder. Unfortunately it doesn't reset
4893 the context it uses. This means that if EOF is reached an attempt
4894 is made to pass two EOFs through the context and this causes the
4895 resulting error. This can also cause other problems as well. As is
4896 usual with these problems it takes *ages* to find and the fix is
4897 trivial: move one line.
4898 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
4900 *) Ugly workaround to get s_client and s_server working under Windows. The
4901 old code wouldn't work because it needed to select() on sockets and the
4902 tty (for keypresses and to see if data could be written). Win32 only
4903 supports select() on sockets so we select() with a 1s timeout on the
4904 sockets and then see if any characters are waiting to be read, if none
4905 are present then we retry, we also assume we can always write data to
4906 the tty. This isn't nice because the code then blocks until we've
4907 received a complete line of data and it is effectively polling the
4908 keyboard at 1s intervals: however it's quite a bit better than not
4909 working at all :-) A dedicated Windows application might handle this
4910 with an event loop for example.
4913 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
4914 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
4915 will be called when RSA_sign() and RSA_verify() are used. This is useful
4916 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
4917 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
4918 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
4919 This necessitated the support of an extra signature type NID_md5_sha1
4920 for SSL signatures and modifications to the SSL library to use it instead
4921 of calling RSA_public_decrypt() and RSA_private_encrypt().
4924 *) Add new -verify -CAfile and -CApath options to the crl program, these
4925 will lookup a CRL issuers certificate and verify the signature in a
4926 similar way to the verify program. Tidy up the crl program so it
4927 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
4928 less strict. It will now permit CRL extensions even if it is not
4929 a V2 CRL: this will allow it to tolerate some broken CRLs.
4932 *) Initialize all non-automatic variables each time one of the openssl
4933 sub-programs is started (this is necessary as they may be started
4934 multiple times from the "OpenSSL>" prompt).
4935 [Lennart Bang, Bodo Moeller]
4937 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
4938 removing all other RSA functionality (this is what NO_RSA does). This
4939 is so (for example) those in the US can disable those operations covered
4940 by the RSA patent while allowing storage and parsing of RSA keys and RSA
4944 *) Non-copying interface to BIO pairs.
4945 (still largely untested)
4948 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
4949 ASCII string. This was handled independently in various places before.
4952 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
4953 UTF8 strings a character at a time.
4956 *) Use client_version from client hello to select the protocol
4957 (s23_srvr.c) and for RSA client key exchange verification
4958 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
4961 *) Add various utility functions to handle SPKACs, these were previously
4962 handled by poking round in the structure internals. Added new function
4963 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
4964 print, verify and generate SPKACs. Based on an original idea from
4965 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
4968 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
4971 *) Allow the config file extension section to be overwritten on the
4972 command line. Based on an original idea from Massimiliano Pala
4973 <madwolf@comune.modena.it>. The new option is called -extensions
4974 and can be applied to ca, req and x509. Also -reqexts to override
4975 the request extensions in req and -crlexts to override the crl extensions
4979 *) Add new feature to the SPKAC handling in ca. Now you can include
4980 the same field multiple times by preceding it by "XXXX." for example:
4983 this is the same syntax as used in the req config file.
4986 *) Allow certificate extensions to be added to certificate requests. These
4987 are specified in a 'req_extensions' option of the req section of the
4988 config file. They can be printed out with the -text option to req but
4989 are otherwise ignored at present.
4992 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
4993 data read consists of only the final block it would not decrypted because
4994 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
4995 A misplaced 'break' also meant the decrypted final block might not be
4996 copied until the next read.
4999 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
5000 a few extra parameters to the DH structure: these will be useful if
5001 for example we want the value of 'q' or implement X9.42 DH.
5004 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
5005 provides hooks that allow the default DSA functions or functions on a
5006 "per key" basis to be replaced. This allows hardware acceleration and
5007 hardware key storage to be handled without major modification to the
5008 library. Also added low level modexp hooks and CRYPTO_EX structure and
5009 associated functions.
5012 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
5013 as "read only": it can't be written to and the buffer it points to will
5014 not be freed. Reading from a read only BIO is much more efficient than
5015 a normal memory BIO. This was added because there are several times when
5016 an area of memory needs to be read from a BIO. The previous method was
5017 to create a memory BIO and write the data to it, this results in two
5018 copies of the data and an O(n^2) reading algorithm. There is a new
5019 function BIO_new_mem_buf() which creates a read only memory BIO from
5020 an area of memory. Also modified the PKCS#7 routines to use read only
5024 *) Bugfix: ssl23_get_client_hello did not work properly when called in
5025 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
5026 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
5027 but a retry condition occured while trying to read the rest.
5030 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
5031 NID_pkcs7_encrypted by default: this was wrong since this should almost
5032 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
5033 the encrypted data type: this is a more sensible place to put it and it
5034 allows the PKCS#12 code to be tidied up that duplicated this
5038 *) Changed obj_dat.pl script so it takes its input and output files on
5039 the command line. This should avoid shell escape redirection problems
5043 *) Initial support for certificate extension requests, these are included
5044 in things like Xenroll certificate requests. Included functions to allow
5045 extensions to be obtained and added.
5048 *) -crlf option to s_client and s_server for sending newlines as
5049 CRLF (as required by many protocols).
5052 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
5054 *) Install libRSAglue.a when OpenSSL is built with RSAref.
5055 [Ralf S. Engelschall]
5057 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
5058 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
5060 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
5064 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
5065 DH parameters/keys (q is lost during that conversion, but the resulting
5066 DH parameters contain its length).
5068 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
5069 much faster than DH_generate_parameters (which creates parameters
5070 where p = 2*q + 1), and also the smaller q makes DH computations
5071 much more efficient (160-bit exponentiation instead of 1024-bit
5072 exponentiation); so this provides a convenient way to support DHE
5073 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
5074 utter importance to use
5075 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5077 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5078 when such DH parameters are used, because otherwise small subgroup
5079 attacks may become possible!
5082 *) Avoid memory leak in i2d_DHparams.
5085 *) Allow the -k option to be used more than once in the enc program:
5086 this allows the same encrypted message to be read by multiple recipients.
5089 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
5090 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
5091 it will always use the numerical form of the OID, even if it has a short
5095 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
5096 method only got called if p,q,dmp1,dmq1,iqmp components were present,
5097 otherwise bn_mod_exp was called. In the case of hardware keys for example
5098 no private key components need be present and it might store extra data
5099 in the RSA structure, which cannot be accessed from bn_mod_exp.
5100 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
5101 private key operations.
5104 *) Added support for SPARC Linux.
5107 *) pem_password_cb function type incompatibly changed from
5108 typedef int pem_password_cb(char *buf, int size, int rwflag);
5110 ....(char *buf, int size, int rwflag, void *userdata);
5111 so that applications can pass data to their callbacks:
5112 The PEM[_ASN1]_{read,write}... functions and macros now take an
5113 additional void * argument, which is just handed through whenever
5114 the password callback is called.
5115 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
5117 New function SSL_CTX_set_default_passwd_cb_userdata.
5119 Compatibility note: As many C implementations push function arguments
5120 onto the stack in reverse order, the new library version is likely to
5121 interoperate with programs that have been compiled with the old
5122 pem_password_cb definition (PEM_whatever takes some data that
5123 happens to be on the stack as its last argument, and the callback
5124 just ignores this garbage); but there is no guarantee whatsoever that
5127 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
5128 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
5129 problems not only on Windows, but also on some Unix platforms.
5130 To avoid problematic command lines, these definitions are now in an
5131 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
5132 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
5135 *) MIPS III/IV assembler module is reimplemented.
5138 *) More DES library cleanups: remove references to srand/rand and
5139 delete an unused file.
5142 *) Add support for the the free Netwide assembler (NASM) under Win32,
5143 since not many people have MASM (ml) and it can be hard to obtain.
5144 This is currently experimental but it seems to work OK and pass all
5145 the tests. Check out INSTALL.W32 for info.
5148 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
5149 without temporary keys kept an extra copy of the server key,
5150 and connections with temporary keys did not free everything in case
5154 *) New function RSA_check_key and new openssl rsa option -check
5155 for verifying the consistency of RSA keys.
5156 [Ulf Moeller, Bodo Moeller]
5158 *) Various changes to make Win32 compile work:
5159 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
5160 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
5161 comparison" warnings.
5162 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
5165 *) Add a debugging option to PKCS#5 v2 key generation function: when
5166 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
5167 derived keys are printed to stderr.
5170 *) Copy the flags in ASN1_STRING_dup().
5171 [Roman E. Pavlov <pre@mo.msk.ru>]
5173 *) The x509 application mishandled signing requests containing DSA
5174 keys when the signing key was also DSA and the parameters didn't match.
5176 It was supposed to omit the parameters when they matched the signing key:
5177 the verifying software was then supposed to automatically use the CA's
5178 parameters if they were absent from the end user certificate.
5180 Omitting parameters is no longer recommended. The test was also
5181 the wrong way round! This was probably due to unusual behaviour in
5182 EVP_cmp_parameters() which returns 1 if the parameters match.
5183 This meant that parameters were omitted when they *didn't* match and
5184 the certificate was useless. Certificates signed with 'ca' didn't have
5186 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
5188 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
5189 The interface is as follows:
5190 Applications can use
5191 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
5192 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
5193 "off" is now the default.
5194 The library internally uses
5195 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
5196 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
5197 to disable memory-checking temporarily.
5199 Some inconsistent states that previously were possible (and were
5200 even the default) are now avoided.
5202 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
5203 with each memory chunk allocated; this is occasionally more helpful
5204 than just having a counter.
5206 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
5208 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
5212 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
5213 which largely parallels "options", but is for changing API behaviour,
5214 whereas "options" are about protocol behaviour.
5215 Initial "mode" flags are:
5217 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
5218 a single record has been written.
5219 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
5220 retries use the same buffer location.
5221 (But all of the contents must be
5225 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
5228 *) Fix problems with no-hmac etc.
5229 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
5231 *) New functions RSA_get_default_method(), RSA_set_method() and
5232 RSA_get_method(). These allows replacement of RSA_METHODs without having
5233 to mess around with the internals of an RSA structure.
5236 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
5237 Also really enable memory leak checks in openssl.c and in some
5239 [Chad C. Mulligan, Bodo Moeller]
5241 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
5242 up the length of negative integers. This has now been simplified to just
5243 store the length when it is first determined and use it later, rather
5244 than trying to keep track of where data is copied and updating it to
5246 [Steve Henson, reported by Brien Wheeler
5247 <bwheeler@authentica-security.com>]
5249 *) Add a new function PKCS7_signatureVerify. This allows the verification
5250 of a PKCS#7 signature but with the signing certificate passed to the
5251 function itself. This contrasts with PKCS7_dataVerify which assumes the
5252 certificate is present in the PKCS#7 structure. This isn't always the
5253 case: certificates can be omitted from a PKCS#7 structure and be
5254 distributed by "out of band" means (such as a certificate database).
5257 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
5258 function prototypes in pem.h, also change util/mkdef.pl to add the
5259 necessary function names.
5262 *) mk1mf.pl (used by Windows builds) did not properly read the
5263 options set by Configure in the top level Makefile, and Configure
5264 was not even able to write more than one option correctly.
5265 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
5268 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
5269 file to be loaded from a BIO or FILE pointer. The BIO version will
5270 for example allow memory BIOs to contain config info.
5273 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
5274 Whoever hopes to achieve shared-library compatibility across versions
5275 must use this, not the compile-time macro.
5276 (Exercise 0.9.4: Which is the minimum library version required by
5278 Note: All this applies only to multi-threaded programs, others don't
5282 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
5283 through a BIO pair triggered the default case, i.e.
5284 SSLerr(...,SSL_R_UNKNOWN_STATE).
5287 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
5288 can use the SSL library even if none of the specific BIOs is
5292 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
5293 for the encoded length.
5294 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
5296 *) Add initial documentation of the X509V3 functions.
5299 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
5300 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
5301 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
5302 secure PKCS#8 private key format with a high iteration count.
5305 *) Fix determination of Perl interpreter: A perl or perl5
5306 _directory_ in $PATH was also accepted as the interpreter.
5307 [Ralf S. Engelschall]
5309 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
5310 wrong with it but it was very old and did things like calling
5311 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
5315 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
5316 to use the new extension code.
5319 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
5320 with macros. This should make it easier to change their form, add extra
5321 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
5325 *) Add to configuration table a new entry that can specify an alternative
5326 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
5327 according to Mark Crispin <MRC@Panda.COM>.
5331 *) DES CBC did not update the IV. Weird.
5334 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
5335 Changing the behaviour of the former might break existing programs --
5336 where IV updating is needed, des_ncbc_encrypt can be used.
5339 *) When bntest is run from "make test" it drives bc to check its
5340 calculations, as well as internally checking them. If an internal check
5341 fails, it needs to cause bc to give a non-zero result or make test carries
5342 on without noticing the failure. Fixed.
5345 *) DES library cleanups.
5348 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
5349 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
5350 ciphers. NOTE: although the key derivation function has been verified
5351 against some published test vectors it has not been extensively tested
5352 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
5356 *) Instead of "mkdir -p", which is not fully portable, use new
5357 Perl script "util/mkdir-p.pl".
5360 *) Rewrite the way password based encryption (PBE) is handled. It used to
5361 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
5362 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
5363 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
5364 the 'parameter' field of the AlgorithmIdentifier is passed to the
5365 underlying key generation function so it must do its own ASN1 parsing.
5366 This has also changed the EVP_PBE_CipherInit() function which now has a
5367 'parameter' argument instead of literal salt and iteration count values
5368 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
5371 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
5372 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
5373 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
5374 KEY" because this clashed with PKCS#8 unencrypted string. Since this
5375 value was just used as a "magic string" and not used directly its
5376 value doesn't matter.
5379 *) Introduce some semblance of const correctness to BN. Shame C doesn't
5383 *) "linux-sparc64" configuration (ultrapenguin).
5384 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
5385 "linux-sparc" configuration.
5386 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
5388 *) config now generates no-xxx options for missing ciphers.
5391 *) Support the EBCDIC character set (work in progress).
5392 File ebcdic.c not yet included because it has a different license.
5393 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5395 *) Support BS2000/OSD-POSIX.
5396 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5398 *) Make callbacks for key generation use void * instead of char *.
5401 *) Make S/MIME samples compile (not yet tested).
5404 *) Additional typesafe stacks.
5407 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
5411 Changes between 0.9.3 and 0.9.3a [29 May 1999]
5413 *) New configuration variant "sco5-gcc".
5415 *) Updated some demos.
5416 [Sean O Riordain, Wade Scholine]
5418 *) Add missing BIO_free at exit of pkcs12 application.
5421 *) Fix memory leak in conf.c.
5424 *) Updates for Win32 to assembler version of MD5.
5427 *) Set #! path to perl in apps/der_chop to where we found it
5428 instead of using a fixed path.
5431 *) SHA library changes for irix64-mips4-cc.
5434 *) Improvements for VMS support.
5438 Changes between 0.9.2b and 0.9.3 [24 May 1999]
5440 *) Bignum library bug fix. IRIX 6 passes "make test" now!
5441 This also avoids the problems with SC4.2 and unpatched SC5.
5442 [Andy Polyakov <appro@fy.chalmers.se>]
5444 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
5445 These are required because of the typesafe stack would otherwise break
5446 existing code. If old code used a structure member which used to be STACK
5447 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
5448 sk_num or sk_value it would produce an error because the num, data members
5449 are not present in STACK_OF. Now it just produces a warning. sk_set
5450 replaces the old method of assigning a value to sk_value
5451 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
5452 that does this will no longer work (and should use sk_set instead) but
5453 this could be regarded as a "questionable" behaviour anyway.
5456 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
5457 correctly handle encrypted S/MIME data.
5460 *) Change type of various DES function arguments from des_cblock
5461 (which means, in function argument declarations, pointer to char)
5462 to des_cblock * (meaning pointer to array with 8 char elements),
5463 which allows the compiler to do more typechecking; it was like
5464 that back in SSLeay, but with lots of ugly casts.
5466 Introduce new type const_des_cblock.
5469 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
5470 problems: find RecipientInfo structure that matches recipient certificate
5471 and initialise the ASN1 structures properly based on passed cipher.
5474 *) Belatedly make the BN tests actually check the results.
5477 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
5478 to and from BNs: it was completely broken. New compilation option
5479 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
5480 key elements as negative integers.
5483 *) Reorganize and speed up MD5.
5484 [Andy Polyakov <appro@fy.chalmers.se>]
5487 [Richard Levitte <richard@levitte.org>]
5489 *) New option -out to asn1parse to allow the parsed structure to be
5490 output to a file. This is most useful when combined with the -strparse
5491 option to examine the output of things like OCTET STRINGS.
5494 *) Make SSL library a little more fool-proof by not requiring any longer
5495 that SSL_set_{accept,connect}_state be called before
5496 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
5497 in many applications because usually everything *appeared* to work as
5498 intended anyway -- now it really works as intended).
5501 *) Move openssl.cnf out of lib/.
5504 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
5505 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
5506 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
5507 [Ralf S. Engelschall]
5509 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
5510 handle PKCS#7 enveloped data properly.
5511 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
5513 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
5514 copying pointers. The cert_st handling is changed by this in
5515 various ways (and thus what used to be known as ctx->default_cert
5516 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
5517 any longer when s->cert does not give us what we need).
5518 ssl_cert_instantiate becomes obsolete by this change.
5519 As soon as we've got the new code right (possibly it already is?),
5520 we have solved a couple of bugs of the earlier code where s->cert
5521 was used as if it could not have been shared with other SSL structures.
5523 Note that using the SSL API in certain dirty ways now will result
5524 in different behaviour than observed with earlier library versions:
5525 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
5526 does not influence s as it used to.
5528 In order to clean up things more thoroughly, inside SSL_SESSION
5529 we don't use CERT any longer, but a new structure SESS_CERT
5530 that holds per-session data (if available); currently, this is
5531 the peer's certificate chain and, for clients, the server's certificate
5532 and temporary key. CERT holds only those values that can have
5533 meaningful defaults in an SSL_CTX.
5536 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
5537 from the internal representation. Various PKCS#7 fixes: remove some
5538 evil casts and set the enc_dig_alg field properly based on the signing
5542 *) Allow PKCS#12 password to be set from the command line or the
5543 environment. Let 'ca' get its config file name from the environment
5544 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
5548 *) Allow certificate policies extension to use an IA5STRING for the
5549 organization field. This is contrary to the PKIX definition but
5550 VeriSign uses it and IE5 only recognises this form. Document 'x509'
5554 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
5555 without disallowing inline assembler and the like for non-pedantic builds.
5558 *) Support Borland C++ builder.
5559 [Janez Jere <jj@void.si>, modified by Ulf Möller]
5564 *) SHA-1 cleanups and performance enhancements.
5565 [Andy Polyakov <appro@fy.chalmers.se>]
5567 *) Sparc v8plus assembler for the bignum library.
5568 [Andy Polyakov <appro@fy.chalmers.se>]
5570 *) Accept any -xxx and +xxx compiler options in Configure.
5573 *) Update HPUX configuration.
5576 *) Add missing sk_<type>_unshift() function to safestack.h
5577 [Ralf S. Engelschall]
5579 *) New function SSL_CTX_use_certificate_chain_file that sets the
5580 "extra_cert"s in addition to the certificate. (This makes sense
5581 only for "PEM" format files, as chains as a whole are not
5585 *) Support verify_depth from the SSL API.
5586 x509_vfy.c had what can be considered an off-by-one-error:
5587 Its depth (which was not part of the external interface)
5588 was actually counting the number of certificates in a chain;
5589 now it really counts the depth.
5592 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
5593 instead of X509err, which often resulted in confusing error
5594 messages since the error codes are not globally unique
5595 (e.g. an alleged error in ssl3_accept when a certificate
5596 didn't match the private key).
5598 *) New function SSL_CTX_set_session_id_context that allows to set a default
5599 value (so that you don't need SSL_set_session_id_context for each
5600 connection using the SSL_CTX).
5603 *) OAEP decoding bug fix.
5606 *) Support INSTALL_PREFIX for package builders, as proposed by
5610 *) New Configure options "threads" and "no-threads". For systems
5611 where the proper compiler options are known (currently Solaris
5612 and Linux), "threads" is the default.
5615 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
5618 *) Install various scripts to $(OPENSSLDIR)/misc, not to
5619 $(INSTALLTOP)/bin -- they shouldn't clutter directories
5620 such as /usr/local/bin.
5623 *) "make linux-shared" to build shared libraries.
5624 [Niels Poppe <niels@netbox.org>]
5626 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
5629 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
5630 extension adding in x509 utility.
5633 *) Remove NOPROTO sections and error code comments.
5636 *) Partial rewrite of the DEF file generator to now parse the ANSI
5640 *) New Configure options --prefix=DIR and --openssldir=DIR.
5643 *) Complete rewrite of the error code script(s). It is all now handled
5644 by one script at the top level which handles error code gathering,
5645 header rewriting and C source file generation. It should be much better
5646 than the old method: it now uses a modified version of Ulf's parser to
5647 read the ANSI prototypes in all header files (thus the old K&R definitions
5648 aren't needed for error creation any more) and do a better job of
5649 translating function codes into names. The old 'ASN1 error code imbedded
5650 in a comment' is no longer necessary and it doesn't use .err files which
5651 have now been deleted. Also the error code call doesn't have to appear all
5652 on one line (which resulted in some large lines...).
5655 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
5658 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
5659 0 (which usually indicates a closed connection), but continue reading.
5662 *) Fix some race conditions.
5665 *) Add support for CRL distribution points extension. Add Certificate
5666 Policies and CRL distribution points documentation.
5669 *) Move the autogenerated header file parts to crypto/opensslconf.h.
5672 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
5673 8 of keying material. Merlin has also confirmed interop with this fix
5674 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
5675 [Merlin Hughes <merlin@baltimore.ie>]
5677 *) Fix lots of warnings.
5678 [Richard Levitte <levitte@stacken.kth.se>]
5680 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
5681 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
5682 [Richard Levitte <levitte@stacken.kth.se>]
5684 *) Fix problems with sizeof(long) == 8.
5685 [Andy Polyakov <appro@fy.chalmers.se>]
5687 *) Change functions to ANSI C.
5690 *) Fix typos in error codes.
5691 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
5693 *) Remove defunct assembler files from Configure.
5696 *) SPARC v8 assembler BIGNUM implementation.
5697 [Andy Polyakov <appro@fy.chalmers.se>]
5699 *) Support for Certificate Policies extension: both print and set.
5700 Various additions to support the r2i method this uses.
5703 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
5704 return a const string when you are expecting an allocated buffer.
5707 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
5708 types DirectoryString and DisplayText.
5711 *) Add code to allow r2i extensions to access the configuration database,
5712 add an LHASH database driver and add several ctx helper functions.
5715 *) Fix an evil bug in bn_expand2() which caused various BN functions to
5716 fail when they extended the size of a BIGNUM.
5719 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
5720 support typesafe stack.
5723 *) Fix typo in SSL_[gs]et_options().
5724 [Nils Frostberg <nils@medcom.se>]
5726 *) Delete various functions and files that belonged to the (now obsolete)
5727 old X509V3 handling code.
5730 *) New Configure option "rsaref".
5733 *) Don't auto-generate pem.h.
5736 *) Introduce type-safe ASN.1 SETs.
5739 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
5740 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
5742 *) Introduce type-safe STACKs. This will almost certainly break lots of code
5743 that links with OpenSSL (well at least cause lots of warnings), but fear
5744 not: the conversion is trivial, and it eliminates loads of evil casts. A
5745 few STACKed things have been converted already. Feel free to convert more.
5746 In the fullness of time, I'll do away with the STACK type altogether.
5749 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
5750 specified in <certfile> by updating the entry in the index.txt file.
5751 This way one no longer has to edit the index.txt file manually for
5752 revoking a certificate. The -revoke option does the gory details now.
5753 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
5755 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
5756 `-text' option at all and this way the `-noout -text' combination was
5757 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
5758 [Ralf S. Engelschall]
5760 *) Make sure a corresponding plain text error message exists for the
5761 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
5762 verify callback function determined that a certificate was revoked.
5763 [Ralf S. Engelschall]
5765 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
5766 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
5767 all available cipers including rc5, which was forgotten until now.
5768 In order to let the testing shell script know which algorithms
5769 are available, a new (up to now undocumented) command
5770 "openssl list-cipher-commands" is used.
5773 *) Bugfix: s_client occasionally would sleep in select() when
5774 it should have checked SSL_pending() first.
5777 *) New functions DSA_do_sign and DSA_do_verify to provide access to
5778 the raw DSA values prior to ASN.1 encoding.
5781 *) Tweaks to Configure
5782 [Niels Poppe <niels@netbox.org>]
5784 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
5788 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
5791 *) New config option to avoid instructions that are illegal on the 80386.
5792 The default code is faster, but requires at least a 486.
5795 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
5796 SSL2_SERVER_VERSION (not used at all) macros, which are now the
5797 same as SSL2_VERSION anyway.
5800 *) New "-showcerts" option for s_client.
5803 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
5804 application. Various cleanups and fixes.
5807 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
5808 modify error routines to work internally. Add error codes and PBE init
5809 to library startup routines.
5812 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
5813 packing functions to asn1 and evp. Changed function names and error
5814 codes along the way.
5817 *) PKCS12 integration: and so it begins... First of several patches to
5818 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
5819 objects to objects.h
5822 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
5823 and display support for Thawte strong extranet extension.
5826 *) Add LinuxPPC support.
5827 [Jeff Dubrule <igor@pobox.org>]
5829 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
5830 bn_div_words in alpha.s.
5831 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
5833 *) Make sure the RSA OAEP test is skipped under -DRSAref because
5834 OAEP isn't supported when OpenSSL is built with RSAref.
5835 [Ulf Moeller <ulf@fitug.de>]
5837 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
5838 so they no longer are missing under -DNOPROTO.
5839 [Soren S. Jorvang <soren@t.dk>]
5842 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
5844 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
5845 doesn't work when the session is reused. Coming soon!
5848 *) Fix a security hole, that allows sessions to be reused in the wrong
5849 context thus bypassing client cert protection! All software that uses
5850 client certs and session caches in multiple contexts NEEDS PATCHING to
5851 allow session reuse! A fuller solution is in the works.
5852 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
5854 *) Some more source tree cleanups (removed obsolete files
5855 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
5856 permission on "config" script to be executable) and a fix for the INSTALL
5858 [Ulf Moeller <ulf@fitug.de>]
5860 *) Remove some legacy and erroneous uses of malloc, free instead of
5862 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
5864 *) Make rsa_oaep_test return non-zero on error.
5865 [Ulf Moeller <ulf@fitug.de>]
5867 *) Add support for native Solaris shared libraries. Configure
5868 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
5869 if someone would make that last step automatic.
5870 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
5872 *) ctx_size was not built with the right compiler during "make links". Fixed.
5875 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
5876 except NULL ciphers". This means the default cipher list will no longer
5877 enable NULL ciphers. They need to be specifically enabled e.g. with
5878 the string "DEFAULT:eNULL".
5881 *) Fix to RSA private encryption routines: if p < q then it would
5882 occasionally produce an invalid result. This will only happen with
5883 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
5886 *) Be less restrictive and allow also `perl util/perlpath.pl
5887 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
5888 because this way one can also use an interpreter named `perl5' (which is
5889 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
5890 installed as `perl').
5891 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5893 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
5894 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5896 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
5897 advapi32.lib to Win32 build and change the pem test comparision
5898 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
5899 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
5900 and crypto/des/ede_cbcm_enc.c.
5903 *) DES quad checksum was broken on big-endian architectures. Fixed.
5906 *) Comment out two functions in bio.h that aren't implemented. Fix up the
5907 Win32 test batch file so it (might) work again. The Win32 test batch file
5908 is horrible: I feel ill....
5911 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
5912 in e_os.h. Audit of header files to check ANSI and non ANSI
5913 sections: 10 functions were absent from non ANSI section and not exported
5914 from Windows DLLs. Fixed up libeay.num for new functions.
5917 *) Make `openssl version' output lines consistent.
5918 [Ralf S. Engelschall]
5920 *) Fix Win32 symbol export lists for BIO functions: Added
5921 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
5922 to ms/libeay{16,32}.def.
5923 [Ralf S. Engelschall]
5925 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
5926 fine under Unix and passes some trivial tests I've now added. But the
5927 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
5928 added to make sure no one expects that this stuff really works in the
5929 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
5930 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
5932 [Ralf S. Engelschall]
5934 *) Fix the generation of two part addresses in perl.
5935 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
5937 *) Add config entry for Linux on MIPS.
5938 [John Tobey <jtobey@channel1.com>]
5940 *) Make links whenever Configure is run, unless we are on Windoze.
5943 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
5944 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
5948 *) Add a useful kludge to allow package maintainers to specify compiler and
5949 other platforms details on the command line without having to patch the
5950 Configure script everytime: One now can use ``perl Configure
5951 <id>:<details>'', i.e. platform ids are allowed to have details appended
5952 to them (seperated by colons). This is treated as there would be a static
5953 pre-configured entry in Configure's %table under key <id> with value
5954 <details> and ``perl Configure <id>'' is called. So, when you want to
5955 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
5956 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
5957 now, which overrides the FreeBSD-elf entry on-the-fly.
5958 [Ralf S. Engelschall]
5960 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
5963 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
5964 on the `perl Configure ...' command line. This way one can compile
5965 OpenSSL libraries with Position Independent Code (PIC) which is needed
5966 for linking it into DSOs.
5967 [Ralf S. Engelschall]
5969 *) Remarkably, export ciphers were totally broken and no-one had noticed!
5973 *) Cleaned up the LICENSE document: The official contact for any license
5974 questions now is the OpenSSL core team under openssl-core@openssl.org.
5975 And add a paragraph about the dual-license situation to make sure people
5976 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
5977 to the OpenSSL toolkit.
5978 [Ralf S. Engelschall]
5980 *) General source tree makefile cleanups: Made `making xxx in yyy...'
5981 display consistent in the source tree and replaced `/bin/rm' by `rm'.
5982 Additonally cleaned up the `make links' target: Remove unnecessary
5983 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
5984 to speed processing and no longer clutter the display with confusing
5985 stuff. Instead only the actually done links are displayed.
5986 [Ralf S. Engelschall]
5988 *) Permit null encryption ciphersuites, used for authentication only. It used
5989 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
5990 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
5994 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
5995 signed attributes when verifying signatures (this would break them),
5996 the detached data encoding was wrong and public keys obtained using
5997 X509_get_pubkey() weren't freed.
6000 *) Add text documentation for the BUFFER functions. Also added a work around
6001 to a Win95 console bug. This was triggered by the password read stuff: the
6002 last character typed gets carried over to the next fread(). If you were
6003 generating a new cert request using 'req' for example then the last
6004 character of the passphrase would be CR which would then enter the first
6008 *) Added the new `Includes OpenSSL Cryptography Software' button as
6009 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
6010 button and can be used by applications based on OpenSSL to show the
6011 relationship to the OpenSSL project.
6012 [Ralf S. Engelschall]
6014 *) Remove confusing variables in function signatures in files
6015 ssl/ssl_lib.c and ssl/ssl.h.
6016 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6018 *) Don't install bss_file.c under PREFIX/include/
6019 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6021 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
6022 functions that return function pointers and has support for NT specific
6023 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
6024 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
6025 unsigned to signed types: this was killing the Win32 compile.
6028 *) Add new certificate file to stack functions,
6029 SSL_add_dir_cert_subjects_to_stack() and
6030 SSL_add_file_cert_subjects_to_stack(). These largely supplant
6031 SSL_load_client_CA_file(), and can be used to add multiple certs easily
6032 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
6033 This means that Apache-SSL and similar packages don't have to mess around
6034 to add as many CAs as they want to the preferred list.
6037 *) Experiment with doxygen documentation. Currently only partially applied to
6039 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
6040 openssl.doxy as the configuration file.
6043 *) Get rid of remaining C++-style comments which strict C compilers hate.
6044 [Ralf S. Engelschall, pointed out by Carlos Amengual]
6046 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
6047 compiled in by default: it has problems with large keys.
6050 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
6051 DH private keys and/or callback functions which directly correspond to
6052 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
6053 is needed for applications which have to configure certificates on a
6054 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
6056 For the RSA certificate situation is makes no difference, but
6057 for the DSA certificate situation this fixes the "no shared cipher"
6058 problem where the OpenSSL cipher selection procedure failed because the
6059 temporary keys were not overtaken from the context and the API provided
6060 no way to reconfigure them.
6061 The new functions now let applications reconfigure the stuff and they
6062 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
6063 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
6064 non-public-API function ssl_cert_instantiate() is used as a helper
6065 function and also to reduce code redundancy inside ssl_rsa.c.
6066 [Ralf S. Engelschall]
6068 *) Move s_server -dcert and -dkey options out of the undocumented feature
6069 area because they are useful for the DSA situation and should be
6070 recognized by the users.
6071 [Ralf S. Engelschall]
6073 *) Fix the cipher decision scheme for export ciphers: the export bits are
6074 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
6075 SSL_EXP_MASK. So, the original variable has to be used instead of the
6076 already masked variable.
6077 [Richard Levitte <levitte@stacken.kth.se>]
6079 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
6080 [Richard Levitte <levitte@stacken.kth.se>]
6082 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
6083 from `int' to `unsigned int' because it's a length and initialized by
6084 EVP_DigestFinal() which expects an `unsigned int *'.
6085 [Richard Levitte <levitte@stacken.kth.se>]
6087 *) Don't hard-code path to Perl interpreter on shebang line of Configure
6088 script. Instead use the usual Shell->Perl transition trick.
6089 [Ralf S. Engelschall]
6091 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
6092 (in addition to RSA certificates) to match the behaviour of `openssl dsa
6093 -noout -modulus' as it's already the case for `openssl rsa -noout
6094 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
6095 currently the public key is printed (a decision which was already done by
6096 `openssl dsa -modulus' in the past) which serves a similar purpose.
6097 Additionally the NO_RSA no longer completely removes the whole -modulus
6098 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
6100 [Ralf S. Engelschall]
6102 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
6103 BIO. See the source (crypto/evp/bio_ok.c) for more info.
6104 [Arne Ansper <arne@ats.cyber.ee>]
6106 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
6107 to be added. Now both 'req' and 'ca' can use new objects defined in the
6111 *) Add cool BIO that does syslog (or event log on NT).
6112 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
6114 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
6115 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
6116 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
6117 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
6120 *) Add preliminary config info for new extension code.
6123 *) Make RSA_NO_PADDING really use no padding.
6124 [Ulf Moeller <ulf@fitug.de>]
6126 *) Generate errors when private/public key check is done.
6129 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
6130 for some CRL extensions and new objects added.
6133 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
6134 key usage extension and fuller support for authority key id.
6137 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
6138 padding method for RSA, which is recommended for new applications in PKCS
6139 #1 v2.0 (RFC 2437, October 1998).
6140 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
6141 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
6142 against Bleichbacher's attack on RSA.
6143 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
6146 *) Updates to the new SSL compression code
6147 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6149 *) Fix so that the version number in the master secret, when passed
6150 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
6151 (because the server will not accept higher), that the version number
6152 is 0x03,0x01, not 0x03,0x00
6153 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6155 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
6156 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
6157 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
6160 *) Support for RAW extensions where an arbitrary extension can be
6161 created by including its DER encoding. See apps/openssl.cnf for
6165 *) Make sure latest Perl versions don't interpret some generated C array
6166 code as Perl array code in the crypto/err/err_genc.pl script.
6167 [Lars Weber <3weber@informatik.uni-hamburg.de>]
6169 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
6170 not many people have the assembler. Various Win32 compilation fixes and
6171 update to the INSTALL.W32 file with (hopefully) more accurate Win32
6175 *) Modify configure script 'Configure' to automatically create crypto/date.h
6176 file under Win32 and also build pem.h from pem.org. New script
6177 util/mkfiles.pl to create the MINFO file on environments that can't do a
6178 'make files': perl util/mkfiles.pl >MINFO should work.
6181 *) Major rework of DES function declarations, in the pursuit of correctness
6182 and purity. As a result, many evil casts evaporated, and some weirdness,
6183 too. You may find this causes warnings in your code. Zapping your evil
6184 casts will probably fix them. Mostly.
6187 *) Fix for a typo in asn1.h. Bug fix to object creation script
6188 obj_dat.pl. It considered a zero in an object definition to mean
6189 "end of object": none of the objects in objects.h have any zeros
6190 so it wasn't spotted.
6191 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
6193 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
6194 Masking (CBCM). In the absence of test vectors, the best I have been able
6195 to do is check that the decrypt undoes the encrypt, so far. Send me test
6196 vectors if you have them.
6199 *) Correct calculation of key length for export ciphers (too much space was
6200 allocated for null ciphers). This has not been tested!
6203 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
6204 message is now correct (it understands "crypto" and "ssl" on its
6205 command line). There is also now an "update" option. This will update
6206 the util/ssleay.num and util/libeay.num files with any new functions.
6208 perl util/mkdef.pl crypto ssl update
6209 it will update them.
6212 *) Overhauled the Perl interface (perl/*):
6213 - ported BN stuff to OpenSSL's different BN library
6214 - made the perl/ source tree CVS-aware
6215 - renamed the package from SSLeay to OpenSSL (the files still contain
6216 their history because I've copied them in the repository)
6217 - removed obsolete files (the test scripts will be replaced
6218 by better Test::Harness variants in the future)
6219 [Ralf S. Engelschall]
6221 *) First cut for a very conservative source tree cleanup:
6222 1. merge various obsolete readme texts into doc/ssleay.txt
6223 where we collect the old documents and readme texts.
6224 2. remove the first part of files where I'm already sure that we no
6225 longer need them because of three reasons: either they are just temporary
6226 files which were left by Eric or they are preserved original files where
6227 I've verified that the diff is also available in the CVS via "cvs diff
6228 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
6229 the crypto/md/ stuff).
6230 [Ralf S. Engelschall]
6232 *) More extension code. Incomplete support for subject and issuer alt
6233 name, issuer and authority key id. Change the i2v function parameters
6234 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
6235 what that's for :-) Fix to ASN1 macro which messed up
6236 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
6239 *) Preliminary support for ENUMERATED type. This is largely copied from the
6243 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
6244 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6246 *) Make sure `make rehash' target really finds the `openssl' program.
6247 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6249 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
6250 like to hear about it if this slows down other processors.
6253 *) Add CygWin32 platform information to Configure script.
6254 [Alan Batie <batie@aahz.jf.intel.com>]
6256 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
6257 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
6259 *) New program nseq to manipulate netscape certificate sequences
6262 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
6266 *) Fixes to BN code. Previously the default was to define BN_RECURSION
6267 but the BN code had some problems that would cause failures when
6268 doing certificate verification and some other functions.
6269 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6271 *) Add ASN1 and PEM code to support netscape certificate sequences.
6274 *) Add ASN1 and PEM code to support netscape certificate sequences.
6277 *) Add several PKIX and private extended key usage OIDs.
6280 *) Modify the 'ca' program to handle the new extension code. Modify
6281 openssl.cnf for new extension format, add comments.
6284 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
6285 and add a sample to openssl.cnf so req -x509 now adds appropriate
6289 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
6290 error code, add initial support to X509_print() and x509 application.
6293 *) Takes a deep breath and start addding X509 V3 extension support code. Add
6294 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
6295 stuff is currently isolated and isn't even compiled yet.
6298 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
6299 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
6300 Removed the versions check from X509 routines when loading extensions:
6301 this allows certain broken certificates that don't set the version
6302 properly to be processed.
6305 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
6306 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
6307 can still be regenerated with "make depend".
6310 *) Spelling mistake in C version of CAST-128.
6311 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
6313 *) Changes to the error generation code. The perl script err-code.pl
6314 now reads in the old error codes and retains the old numbers, only
6315 adding new ones if necessary. It also only changes the .err files if new
6316 codes are added. The makefiles have been modified to only insert errors
6317 when needed (to avoid needlessly modifying header files). This is done
6318 by only inserting errors if the .err file is newer than the auto generated
6319 C file. To rebuild all the error codes from scratch (the old behaviour)
6320 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
6321 or delete all the .err files.
6324 *) CAST-128 was incorrectly implemented for short keys. The C version has
6325 been fixed, but is untested. The assembler versions are also fixed, but
6326 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
6327 to regenerate it if needed.
6328 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
6329 Hagino <itojun@kame.net>]
6331 *) File was opened incorrectly in randfile.c.
6332 [Ulf Möller <ulf@fitug.de>]
6334 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
6335 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
6336 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
6337 al: it's just almost always a UTCTime. Note this patch adds new error
6338 codes so do a "make errors" if there are problems.
6341 *) Correct Linux 1 recognition in config.
6342 [Ulf Möller <ulf@fitug.de>]
6344 *) Remove pointless MD5 hash when using DSA keys in ca.
6345 [Anonymous <nobody@replay.com>]
6347 *) Generate an error if given an empty string as a cert directory. Also
6348 generate an error if handed NULL (previously returned 0 to indicate an
6349 error, but didn't set one).
6350 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
6352 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
6355 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
6356 parameters. This was causing a warning which killed off the Win32 compile.
6359 *) Remove C++ style comments from crypto/bn/bn_local.h.
6360 [Neil Costigan <neil.costigan@celocom.com>]
6362 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
6363 based on a text string, looking up short and long names and finally
6364 "dot" format. The "dot" format stuff didn't work. Added new function
6365 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
6366 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
6367 OID is not part of the table.
6370 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
6371 X509_LOOKUP_by_alias().
6374 *) Sort openssl functions by name.
6377 *) Get the gendsa program working (hopefully) and add it to app list. Remove
6378 encryption from sample DSA keys (in case anyone is interested the password
6382 *) Make _all_ *_free functions accept a NULL pointer.
6383 [Frans Heymans <fheymans@isaserver.be>]
6385 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
6387 [Anonymous <nobody@replay.com>]
6389 *) s_server should send the CAfile as acceptable CAs, not its own cert.
6390 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6392 *) Don't blow it for numeric -newkey arguments to apps/req.
6393 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6395 *) Temp key "for export" tests were wrong in s3_srvr.c.
6396 [Anonymous <nobody@replay.com>]
6398 *) Add prototype for temp key callback functions
6399 SSL_CTX_set_tmp_{rsa,dh}_callback().
6402 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
6403 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
6406 *) X509_name_add_entry() freed the wrong thing after an error.
6407 [Arne Ansper <arne@ats.cyber.ee>]
6409 *) rsa_eay.c would attempt to free a NULL context.
6410 [Arne Ansper <arne@ats.cyber.ee>]
6412 *) BIO_s_socket() had a broken should_retry() on Windoze.
6413 [Arne Ansper <arne@ats.cyber.ee>]
6415 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
6416 [Arne Ansper <arne@ats.cyber.ee>]
6418 *) Make sure the already existing X509_STORE->depth variable is initialized
6419 in X509_STORE_new(), but document the fact that this variable is still
6420 unused in the certificate verification process.
6421 [Ralf S. Engelschall]
6423 *) Fix the various library and apps files to free up pkeys obtained from
6424 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
6427 *) Fix reference counting in X509_PUBKEY_get(). This makes
6428 demos/maurice/example2.c work, amongst others, probably.
6429 [Steve Henson and Ben Laurie]
6431 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
6432 `openssl' and second, the shortcut symlinks for the `openssl <command>'
6433 are no longer created. This way we have a single and consistent command
6434 line interface `openssl <command>', similar to `cvs <command>'.
6435 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
6437 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
6438 BIT STRING wrapper always have zero unused bits.
6441 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
6444 *) Make the top-level INSTALL documentation easier to understand.
6447 *) Makefiles updated to exit if an error occurs in a sub-directory
6448 make (including if user presses ^C) [Paul Sutton]
6450 *) Make Montgomery context stuff explicit in RSA data structure.
6453 *) Fix build order of pem and err to allow for generated pem.h.
6456 *) Fix renumbering bug in X509_NAME_delete_entry().
6459 *) Enhanced the err-ins.pl script so it makes the error library number
6460 global and can add a library name. This is needed for external ASN1 and
6461 other error libraries.
6464 *) Fixed sk_insert which never worked properly.
6467 *) Fix ASN1 macros so they can handle indefinite length construted
6468 EXPLICIT tags. Some non standard certificates use these: they can now
6472 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
6473 into a single doc/ssleay.txt bundle. This way the information is still
6474 preserved but no longer messes up this directory. Now it's new room for
6475 the new set of documenation files.
6476 [Ralf S. Engelschall]
6478 *) SETs were incorrectly DER encoded. This was a major pain, because they
6479 shared code with SEQUENCEs, which aren't coded the same. This means that
6480 almost everything to do with SETs or SEQUENCEs has either changed name or
6481 number of arguments.
6482 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
6484 *) Fix test data to work with the above.
6487 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
6488 was already fixed by Eric for 0.9.1 it seems.
6489 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
6491 *) Autodetect FreeBSD3.
6494 *) Fix various bugs in Configure. This affects the following platforms:
6498 unixware-2.0-pentium
6502 *) Eliminate generated files from CVS. Reorder tests to regenerate files
6503 before they are needed.
6506 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
6510 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
6512 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
6513 changed SSLeay to OpenSSL in version strings.
6514 [Ralf S. Engelschall]
6516 *) Some fixups to the top-level documents.
6519 *) Fixed the nasty bug where rsaref.h was not found under compile-time
6520 because the symlink to include/ was missing.
6521 [Ralf S. Engelschall]
6523 *) Incorporated the popular no-RSA/DSA-only patches
6524 which allow to compile a RSA-free SSLeay.
6525 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
6527 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
6528 when "ssleay" is still not found.
6529 [Ralf S. Engelschall]
6531 *) Added more platforms to Configure: Cray T3E, HPUX 11,
6532 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
6534 *) Updated the README file.
6535 [Ralf S. Engelschall]
6537 *) Added various .cvsignore files in the CVS repository subdirs
6538 to make a "cvs update" really silent.
6539 [Ralf S. Engelschall]
6541 *) Recompiled the error-definition header files and added
6542 missing symbols to the Win32 linker tables.
6543 [Ralf S. Engelschall]
6545 *) Cleaned up the top-level documents;
6546 o new files: CHANGES and LICENSE
6547 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
6548 o merged COPYRIGHT into LICENSE
6549 o removed obsolete TODO file
6550 o renamed MICROSOFT to INSTALL.W32
6551 [Ralf S. Engelschall]
6553 *) Removed dummy files from the 0.9.1b source tree:
6554 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
6555 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
6556 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
6557 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
6558 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
6559 [Ralf S. Engelschall]
6561 *) Added various platform portability fixes.
6564 *) The Genesis of the OpenSSL rpject:
6565 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
6566 Young and Tim J. Hudson created while they were working for C2Net until
6568 [The OpenSSL Project]
6571 Changes between 0.9.0b and 0.9.1b [not released]
6573 *) Updated a few CA certificates under certs/
6576 *) Changed some BIGNUM api stuff.
6579 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
6580 DGUX x86, Linux Alpha, etc.
6583 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
6584 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
6588 *) Add -strparse option to asn1pars program which parses nested
6590 [Dr Stephen Henson <shenson@bigfoot.com>]
6592 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
6595 *) DSA fix for "ca" program.
6598 *) Added "-genkey" option to "dsaparam" program.
6601 *) Added RIPE MD160 (rmd160) message digest.
6604 *) Added -a (all) option to "ssleay version" command.
6607 *) Added PLATFORM define which is the id given to Configure.
6610 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
6613 *) Extended the ASN.1 parser routines.
6616 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
6619 *) Added a BN_CTX to the BN library.
6622 *) Fixed the weak key values in DES library
6625 *) Changed API in EVP library for cipher aliases.
6628 *) Added support for RC2/64bit cipher.
6631 *) Converted the lhash library to the crypto/mem.c functions.
6634 *) Added more recognized ASN.1 object ids.
6637 *) Added more RSA padding checks for SSL/TLS.
6640 *) Added BIO proxy/filter functionality.
6643 *) Added extra_certs to SSL_CTX which can be used
6644 send extra CA certificates to the client in the CA cert chain sending
6645 process. It can be configured with SSL_CTX_add_extra_chain_cert().
6648 *) Now Fortezza is denied in the authentication phase because
6649 this is key exchange mechanism is not supported by SSLeay at all.
6652 *) Additional PKCS1 checks.
6655 *) Support the string "TLSv1" for all TLS v1 ciphers.
6658 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
6659 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
6662 *) Fixed a few memory leaks.
6665 *) Fixed various code and comment typos.
6668 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
6669 bytes sent in the client random.
6670 [Edward Bishop <ebishop@spyglass.com>]