5 Changes between 0.9.7i and 0.9.7j [XX xxx XXXX]
7 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
9 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10 The value now differs depending on if you build for FIPS or not.
11 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12 safely run with a non-FIPSed libcrypto, as it may crash because of
13 the difference induced by this change.
16 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
18 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
19 (part of SSL_OP_ALL). This option used to disable the
20 countermeasure against man-in-the-middle protocol-version
21 rollback in the SSL 2.0 server implementation, which is a bad
24 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
25 for Information Security, National Institute of Advanced Industrial
26 Science and Technology [AIST], Japan)]
28 *) Minimal support for X9.31 signatures and PSS padding modes. This is
29 mainly for FIPS compliance and not fully integrated at this stage.
32 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
33 the exponentiation using a fixed-length exponent. (Otherwise,
34 the information leaked through timing could expose the secret key
35 after many signatures; cf. Bleichenbacher's attack on DSA with
39 *) Make a new fixed-window mod_exp implementation the default for
40 RSA, DSA, and DH private-key operations so that the sequence of
41 squares and multiplies and the memory access pattern are
42 independent of the particular secret key. This will mitigate
43 cache-timing and potential related attacks.
45 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
46 and this is automatically used by BN_mod_exp_mont() if the new flag
47 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
48 will use this BN flag for private exponents unless the flag
49 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
50 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
52 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
54 *) Change the client implementation for SSLv23_method() and
55 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
56 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
57 (Previously, the SSL 2.0 backwards compatible Client Hello
58 message format would be used even with SSL_OP_NO_SSLv2.)
61 *) Add support for smime-type MIME parameter in S/MIME messages which some
65 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
66 a threadsafe manner. Modify rsa code to use new function and add calls
67 to dsa and dh code (which had race conditions before).
70 *) Include the fixed error library code in the C error file definitions
71 instead of fixing them up at runtime. This keeps the error code
75 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
77 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
80 *) Fixes for newer kerberos headers. NB: the casts are needed because
81 the 'length' field is signed on one version and unsigned on another
82 with no (?) obvious way to tell the difference, without these VC++
83 complains. Also the "definition" of FAR (blank) is no longer included
84 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
85 some needed definitions.
88 *) Undo Cygwin change.
91 *) Added support for proxy certificates according to RFC 3820.
92 Because they may be a security thread to unaware applications,
93 they must be explicitely allowed in run-time. See
94 docs/HOWTO/proxy_certificates.txt for further information.
97 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
99 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
100 server and client random values. Previously
101 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
102 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
104 This change has negligible security impact because:
106 1. Server and client random values still have 24 bytes of pseudo random
109 2. Server and client random values are sent in the clear in the initial
112 3. The master secret is derived using the premaster secret (48 bytes in
113 size for static RSA ciphersuites) as well as client server and random
116 The OpenSSL team would like to thank the UK NISCC for bringing this issue
119 [Stephen Henson, reported by UK NISCC]
121 *) Use Windows randomness collection on Cygwin.
124 *) Fix hang in EGD/PRNGD query when communication socket is closed
125 prematurely by EGD/PRNGD.
126 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
128 *) Prompt for pass phrases when appropriate for PKCS12 input format.
131 *) Back-port of selected performance improvements from development
132 branch, as well as improved support for PowerPC platforms.
135 *) Add lots of checks for memory allocation failure, error codes to indicate
136 failure and freeing up memory if a failure occurs.
137 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
139 *) Add new -passin argument to dgst.
142 *) Perform some character comparisons of different types in X509_NAME_cmp:
143 this is needed for some certificates that reencode DNs into UTF8Strings
144 (in violation of RFC3280) and can't or wont issue name rollover
148 *) Make an explicit check during certificate validation to see that
149 the CA setting in each certificate on the chain is correct. As a
150 side effect always do the following basic checks on extensions,
151 not just when there's an associated purpose to the check:
153 - if there is an unhandled critical extension (unless the user
154 has chosen to ignore this fault)
155 - if the path length has been exceeded (if one is set at all)
156 - that certain extensions fit the associated purpose (if one has
160 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
162 *) Avoid a race condition when CRLs are checked in a multi threaded
163 environment. This would happen due to the reordering of the revoked
164 entries during signature checking and serial number lookup. Now the
165 encoding is cached and the serial number sort performed under a lock.
166 Add new STACK function sk_is_sorted().
169 *) Add Delta CRL to the extension code.
172 *) Various fixes to s3_pkt.c so alerts are sent properly.
173 [David Holmes <d.holmes@f5.com>]
175 *) Reduce the chances of duplicate issuer name and serial numbers (in
176 violation of RFC3280) using the OpenSSL certificate creation utilities.
177 This is done by creating a random 64 bit value for the initial serial
178 number when a serial number file is created or when a self signed
179 certificate is created using 'openssl req -x509'. The initial serial
180 number file is created using 'openssl x509 -next_serial' in CA.pl
181 rather than being initialized to 1.
184 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
186 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
187 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
188 [Joe Orton, Steve Henson]
190 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
192 [Joe Orton, Steve Henson]
194 *) Make it possible to have multiple active certificates with the same
195 subject in the CA index file. This is done only if the keyword
196 'unique_subject' is set to 'no' in the main CA section (default
197 if 'CA_default') of the configuration file. The value is saved
198 with the database itself in a separate index attribute file,
199 named like the index file with '.attr' appended to the name.
202 *) X509 verify fixes. Disable broken certificate workarounds when
203 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
204 keyUsage extension present. Don't accept CRLs with unhandled critical
205 extensions: since verify currently doesn't process CRL extensions this
206 rejects a CRL with *any* critical extensions. Add new verify error codes
210 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
211 A clarification of RFC2560 will require the use of OCTET STRINGs and
212 some implementations cannot handle the current raw format. Since OpenSSL
213 copies and compares OCSP nonces as opaque blobs without any attempt at
214 parsing them this should not create any compatibility issues.
217 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
218 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
219 this HMAC (and other) operations are several times slower than OpenSSL
223 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
224 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
226 *) Use the correct content when signing type "other".
229 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
231 *) Fix various bugs revealed by running the NISCC test suite:
233 Stop out of bounds reads in the ASN1 code when presented with
234 invalid tags (CAN-2003-0543 and CAN-2003-0544).
236 Free up ASN1_TYPE correctly if ANY type is invalid (CAN-2003-0545).
238 If verify callback ignores invalid public key errors don't try to check
239 certificate signature with the NULL public key.
243 *) New -ignore_err option in ocsp application to stop the server
244 exiting on the first error in a request.
247 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
248 if the server requested one: as stated in TLS 1.0 and SSL 3.0
252 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
253 extra data after the compression methods not only for TLS 1.0
254 but also for SSL 3.0 (as required by the specification).
255 [Bodo Moeller; problem pointed out by Matthias Loepfe]
257 *) Change X509_certificate_type() to mark the key as exported/exportable
258 when it's 512 *bits* long, not 512 bytes.
261 *) Change AES_cbc_encrypt() so it outputs exact multiple of
262 blocks during encryption.
265 *) Various fixes to base64 BIO and non blocking I/O. On write
266 flushes were not handled properly if the BIO retried. On read
267 data was not being buffered properly and had various logic bugs.
268 This also affects blocking I/O when the data being decoded is a
272 *) Various S/MIME bugfixes and compatibility changes:
273 output correct application/pkcs7 MIME type if
274 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
275 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
276 of files as .eml work). Correctly handle very long lines in MIME
280 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
282 *) Countermeasure against the Klima-Pokorny-Rosa extension of
283 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
284 a protocol version number mismatch like a decryption error
285 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
288 *) Turn on RSA blinding by default in the default implementation
289 to avoid a timing attack. Applications that don't want it can call
290 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
291 They would be ill-advised to do so in most cases.
292 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
294 *) Change RSA blinding code so that it works when the PRNG is not
295 seeded (in this case, the secret RSA exponent is abused as
296 an unpredictable seed -- if it is not unpredictable, there
297 is no point in blinding anyway). Make RSA blinding thread-safe
298 by remembering the creator's thread ID in rsa->blinding and
299 having all other threads use local one-time blinding factors
300 (this requires more computation than sharing rsa->blinding, but
301 avoids excessive locking; and if an RSA object is not shared
302 between threads, blinding will still be very fast).
305 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
306 ENGINE as defaults for all supported algorithms irrespective of
307 the 'flags' parameter. 'flags' is now honoured, so applications
308 should make sure they are passing it correctly.
311 *) Target "mingw" now allows native Windows code to be generated in
312 the Cygwin environment as well as with the MinGW compiler.
315 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
317 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
318 via timing by performing a MAC computation even if incorrrect
319 block cipher padding has been found. This is a countermeasure
320 against active attacks where the attacker has to distinguish
321 between bad padding and a MAC verification error. (CAN-2003-0078)
323 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
324 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
325 Martin Vuagnoux (EPFL, Ilion)]
327 *) Make the no-err option work as intended. The intention with no-err
328 is not to have the whole error stack handling routines removed from
329 libcrypto, it's only intended to remove all the function name and
330 reason texts, thereby removing some of the footprint that may not
331 be interesting if those errors aren't displayed anyway.
333 NOTE: it's still possible for any application or module to have it's
334 own set of error texts inserted. The routines are there, just not
335 used by default when no-err is given.
338 *) Add support for FreeBSD on IA64.
339 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
341 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
342 Kerberos function mit_des_cbc_cksum(). Before this change,
343 the value returned by DES_cbc_cksum() was like the one from
344 mit_des_cbc_cksum(), except the bytes were swapped.
345 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
347 *) Allow an application to disable the automatic SSL chain building.
348 Before this a rather primitive chain build was always performed in
349 ssl3_output_cert_chain(): an application had no way to send the
350 correct chain if the automatic operation produced an incorrect result.
352 Now the chain builder is disabled if either:
354 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
356 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
358 The reasoning behind this is that an application would not want the
359 auto chain building to take place if extra chain certificates are
360 present and it might also want a means of sending no additional
361 certificates (for example the chain has two certificates and the
365 *) Add the possibility to build without the ENGINE framework.
366 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
368 *) Under Win32 gmtime() can return NULL: check return value in
369 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
372 *) DSA routines: under certain error conditions uninitialized BN objects
373 could be freed. Solution: make sure initialization is performed early
374 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
375 Nils Larsch <nla@trustcenter.de> via PR#459)
378 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
379 checked on reconnect on the client side, therefore session resumption
380 could still fail with a "ssl session id is different" error. This
381 behaviour is masked when SSL_OP_ALL is used due to
382 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
383 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
387 *) IA-32 assembler support enhancements: unified ELF targets, support
388 for SCO/Caldera platforms, fix for Cygwin shared build.
391 *) Add support for FreeBSD on sparc64. As a consequence, support for
392 FreeBSD on non-x86 processors is separate from x86 processors on
393 the config script, much like the NetBSD support.
394 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
396 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
398 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
401 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
402 code (06) was taken as the first octet of the session ID and the last
403 octet was ignored consequently. As a result SSLv2 client side session
404 caching could not have worked due to the session ID mismatch between
406 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
410 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
411 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
415 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
416 seems that in spite of existing for more than a year, many application
417 author have done nothing to provide the necessary callbacks, which
418 means that this particular engine will not work properly anywhere.
419 This is a very unfortunate situation which forces us, in the name
420 of usability, to give the hw_ncipher.c a static lock, which is part
422 NOTE: This is for the 0.9.7 series ONLY. This hack will never
423 appear in 0.9.8 or later. We EXPECT application authors to have
424 dealt properly with this when 0.9.8 is released (unless we actually
425 make such changes in the libcrypto locking code that changes will
426 have to be made anyway).
429 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
430 octets have been read, EOF or an error occurs. Without this change
431 some truncated ASN1 structures will not produce an error.
434 *) Disable Heimdal support, since it hasn't been fully implemented.
435 Still give the possibility to force the use of Heimdal, but with
436 warnings and a request that patches get sent to openssl-dev.
439 *) Add the VC-CE target, introduce the WINCE sysname, and add
440 INSTALL.WCE and appropriate conditionals to make it build.
441 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
443 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
444 cygssl-x.y.z.dll, where x, y and z are the major, minor and
445 edit numbers of the version.
446 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
448 *) Introduce safe string copy and catenation functions
449 (BUF_strlcpy() and BUF_strlcat()).
450 [Ben Laurie (CHATS) and Richard Levitte]
452 *) Avoid using fixed-size buffers for one-line DNs.
455 *) Add BUF_MEM_grow_clean() to avoid information leakage when
456 resizing buffers containing secrets, and use where appropriate.
459 *) Avoid using fixed size buffers for configuration file location.
462 *) Avoid filename truncation for various CA files.
465 *) Use sizeof in preference to magic numbers.
468 *) Avoid filename truncation in cert requests.
471 *) Add assertions to check for (supposedly impossible) buffer
475 *) Don't cache truncated DNS entries in the local cache (this could
476 potentially lead to a spoofing attack).
479 *) Fix various buffers to be large enough for hex/decimal
480 representations in a platform independent manner.
483 *) Add CRYPTO_realloc_clean() to avoid information leakage when
484 resizing buffers containing secrets, and use where appropriate.
487 *) Add BIO_indent() to avoid much slightly worrying code to do
491 *) Convert sprintf()/BIO_puts() to BIO_printf().
494 *) buffer_gets() could terminate with the buffer only half
498 *) Add assertions to prevent user-supplied crypto functions from
499 overflowing internal buffers by having large block sizes, etc.
502 *) New OPENSSL_assert() macro (similar to assert(), but enabled
506 *) Eliminate unused copy of key in RC4.
509 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
512 *) Fix off-by-one error in EGD path.
515 *) If RANDFILE path is too long, ignore instead of truncating.
518 *) Eliminate unused and incorrectly sized X.509 structure
522 *) Eliminate unused and dangerous function knumber().
525 *) Eliminate unused and dangerous structure, KSSL_ERR.
528 *) Protect against overlong session ID context length in an encoded
529 session object. Since these are local, this does not appear to be
533 *) Change from security patch (see 0.9.6e below) that did not affect
534 the 0.9.6 release series:
536 Remote buffer overflow in SSL3 protocol - an attacker could
537 supply an oversized master key in Kerberos-enabled versions.
541 *) Change the SSL kerb5 codes to match RFC 2712.
544 *) Make -nameopt work fully for req and add -reqopt switch.
545 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
547 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
548 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
550 *) Make sure tests can be performed even if the corresponding algorithms
551 have been removed entirely. This was also the last step to make
552 OpenSSL compilable with DJGPP under all reasonable conditions.
553 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
555 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
556 to allow version independent disabling of normally unselected ciphers,
557 which may be activated as a side-effect of selecting a single cipher.
559 (E.g., cipher list string "RSA" enables ciphersuites that are left
560 out of "ALL" because they do not provide symmetric encryption.
561 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
562 [Lutz Jaenicke, Bodo Moeller]
564 *) Add appropriate support for separate platform-dependent build
565 directories. The recommended way to make a platform-dependent
566 build directory is the following (tested on Linux), maybe with
569 # Place yourself outside of the OpenSSL source tree. In
570 # this example, the environment variable OPENSSL_SOURCE
571 # is assumed to contain the absolute OpenSSL source directory.
572 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
573 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
574 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
575 mkdir -p `dirname $F`
576 ln -s $OPENSSL_SOURCE/$F $F
579 To be absolutely sure not to disturb the source tree, a "make clean"
580 is a good thing. If it isn't successfull, don't worry about it,
581 it probably means the source directory is very clean.
584 *) Make sure any ENGINE control commands make local copies of string
585 pointers passed to them whenever necessary. Otherwise it is possible
586 the caller may have overwritten (or deallocated) the original string
587 data when a later ENGINE operation tries to use the stored values.
588 [Götz Babin-Ebell <babinebell@trustcenter.de>]
590 *) Improve diagnostics in file reading and command-line digests.
591 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
593 *) Add AES modes CFB and OFB to the object database. Correct an
594 error in AES-CFB decryption.
597 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
598 allows existing EVP_CIPHER_CTX structures to be reused after
599 calling EVP_*Final(). This behaviour is used by encryption
600 BIOs and some applications. This has the side effect that
601 applications must explicitly clean up cipher contexts with
602 EVP_CIPHER_CTX_cleanup() or they will leak memory.
605 *) Check the values of dna and dnb in bn_mul_recursive before calling
606 bn_mul_comba (a non zero value means the a or b arrays do not contain
607 n2 elements) and fallback to bn_mul_normal if either is not zero.
610 *) Fix escaping of non-ASCII characters when using the -subj option
611 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
614 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
615 form for "surname", serialNumber has no short form.
616 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
617 therefore remove "mail" short name for "internet 7".
618 The OID for unique identifiers in X509 certificates is
619 x500UniqueIdentifier, not uniqueIdentifier.
620 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
623 *) Add an "init" command to the ENGINE config module and auto initialize
624 ENGINEs. Without any "init" command the ENGINE will be initialized
625 after all ctrl commands have been executed on it. If init=1 the
626 ENGINE is initailized at that point (ctrls before that point are run
627 on the uninitialized ENGINE and after on the initialized one). If
628 init=0 then the ENGINE will not be iniatialized at all.
631 *) Fix the 'app_verify_callback' interface so that the user-defined
632 argument is actually passed to the callback: In the
633 SSL_CTX_set_cert_verify_callback() prototype, the callback
634 declaration has been changed from
637 int (*cb)(X509_STORE_CTX *,void *);
638 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
639 i=s->ctx->app_verify_callback(&ctx)
640 has been changed into
641 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
643 To update applications using SSL_CTX_set_cert_verify_callback(),
644 a dummy argument can be added to their callback functions.
645 [D. K. Smetters <smetters@parc.xerox.com>]
647 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
648 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
650 *) Add and OPENSSL_LOAD_CONF define which will cause
651 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
652 This allows older applications to transparently support certain
653 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
654 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
655 load the config file and OPENSSL_add_all_algorithms_conf() which will
656 always load it have also been added.
659 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
660 Adjust NIDs and EVP layer.
661 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
663 *) Config modules support in openssl utility.
665 Most commands now load modules from the config file,
666 though in a few (such as version) this isn't done
667 because it couldn't be used for anything.
669 In the case of ca and req the config file used is
670 the same as the utility itself: that is the -config
671 command line option can be used to specify an
675 *) Move default behaviour from OPENSSL_config(). If appname is NULL
676 use "openssl_conf" if filename is NULL use default openssl config file.
679 *) Add an argument to OPENSSL_config() to allow the use of an alternative
680 config section name. Add a new flag to tolerate a missing config file
681 and move code to CONF_modules_load_file().
684 *) Support for crypto accelerator cards from Accelerated Encryption
685 Processing, www.aep.ie. (Use engine 'aep')
686 The support was copied from 0.9.6c [engine] and adapted/corrected
687 to work with the new engine framework.
688 [AEP Inc. and Richard Levitte]
690 *) Support for SureWare crypto accelerator cards from Baltimore
691 Technologies. (Use engine 'sureware')
692 The support was copied from 0.9.6c [engine] and adapted
693 to work with the new engine framework.
696 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
697 make the newer ENGINE framework commands for the CHIL engine work.
698 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
700 *) Make it possible to produce shared libraries on ReliantUNIX.
701 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
703 *) Add the configuration target debug-linux-ppro.
704 Make 'openssl rsa' use the general key loading routines
705 implemented in apps.c, and make those routines able to
706 handle the key format FORMAT_NETSCAPE and the variant
708 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
710 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
711 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
713 *) Add -keyform to rsautl, and document -engine.
714 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
716 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
717 BIO_R_NO_SUCH_FILE error code rather than the generic
718 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
723 ERR_peek_last_error_line
724 ERR_peek_last_error_line_data.
728 ERR_peek_error_line_data,
729 but report on the latest error recorded rather than the first one
730 still in the error queue.
731 [Ben Laurie, Bodo Moeller]
733 *) default_algorithms option in ENGINE config module. This allows things
735 default_algorithms = ALL
736 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
739 *) Prelminary ENGINE config module.
742 *) New experimental application configuration code.
745 *) Change the AES code to follow the same name structure as all other
746 symmetric ciphers, and behave the same way. Move everything to
747 the directory crypto/aes, thereby obsoleting crypto/rijndael.
748 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
750 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
751 [Ben Laurie and Theo de Raadt]
753 *) Add option to output public keys in req command.
754 [Massimiliano Pala madwolf@openca.org]
756 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
757 (up to about 10% better than before for P-192 and P-224).
760 *) New functions/macros
762 SSL_CTX_set_msg_callback(ctx, cb)
763 SSL_CTX_set_msg_callback_arg(ctx, arg)
764 SSL_set_msg_callback(ssl, cb)
765 SSL_set_msg_callback_arg(ssl, arg)
767 to request calling a callback function
769 void cb(int write_p, int version, int content_type,
770 const void *buf, size_t len, SSL *ssl, void *arg)
772 whenever a protocol message has been completely received
773 (write_p == 0) or sent (write_p == 1). Here 'version' is the
774 protocol version according to which the SSL library interprets
775 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
776 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
777 the content type as defined in the SSL 3.0/TLS 1.0 protocol
778 specification (change_cipher_spec(20), alert(21), handshake(22)).
779 'buf' and 'len' point to the actual message, 'ssl' to the
780 SSL object, and 'arg' is the application-defined value set by
781 SSL[_CTX]_set_msg_callback_arg().
783 'openssl s_client' and 'openssl s_server' have new '-msg' options
784 to enable a callback that displays all protocol messages.
787 *) Change the shared library support so shared libraries are built as
788 soon as the corresponding static library is finished, and thereby get
789 openssl and the test programs linked against the shared library.
790 This still only happens when the keyword "shard" has been given to
791 the configuration scripts.
793 NOTE: shared library support is still an experimental thing, and
794 backward binary compatibility is still not guaranteed.
795 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
797 *) Add support for Subject Information Access extension.
798 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
800 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
801 additional bytes when new memory had to be allocated, not just
802 when reusing an existing buffer.
805 *) New command line and configuration option 'utf8' for the req command.
806 This allows field values to be specified as UTF8 strings.
809 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
810 runs for the former and machine-readable output for the latter.
813 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
814 of the e-mail address in the DN (i.e., it will go into a certificate
815 extension only). The new configuration file option 'email_in_dn = no'
817 [Massimiliano Pala madwolf@openca.org]
819 *) Change all functions with names starting with des_ to be starting
820 with DES_ instead. Add wrappers that are compatible with libdes,
821 but are named _ossl_old_des_*. Finally, add macros that map the
822 des_* symbols to the corresponding _ossl_old_des_* if libdes
823 compatibility is desired. If OpenSSL 0.9.6c compatibility is
824 desired, the des_* symbols will be mapped to DES_*, with one
827 Since we provide two compatibility mappings, the user needs to
828 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
829 compatibility is desired. The default (i.e., when that macro
830 isn't defined) is OpenSSL 0.9.6c compatibility.
832 There are also macros that enable and disable the support of old
833 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
834 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
835 are defined, the default will apply: to support the old des routines.
837 In either case, one must include openssl/des.h to get the correct
838 definitions. Do not try to just include openssl/des_old.h, that
841 NOTE: This is a major break of an old API into a new one. Software
842 authors are encouraged to switch to the DES_ style functions. Some
843 time in the future, des_old.h and the libdes compatibility functions
844 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
845 default), and then completely removed.
848 *) Test for certificates which contain unsupported critical extensions.
849 If such a certificate is found during a verify operation it is
850 rejected by default: this behaviour can be overridden by either
851 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
852 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
853 X509_supported_extension() has also been added which returns 1 if a
854 particular extension is supported.
857 *) Modify the behaviour of EVP cipher functions in similar way to digests
858 to retain compatibility with existing code.
861 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
862 compatibility with existing code. In particular the 'ctx' parameter does
863 not have to be to be initialized before the call to EVP_DigestInit() and
864 it is tidied up after a call to EVP_DigestFinal(). New function
865 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
866 EVP_MD_CTX_copy() changed to not require the destination to be
867 initialized valid and new function EVP_MD_CTX_copy_ex() added which
868 requires the destination to be valid.
870 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
871 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
874 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
875 so that complete 'Handshake' protocol structures are kept in memory
876 instead of overwriting 'msg_type' and 'length' with 'body' data.
879 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
880 [Massimo Santin via Richard Levitte]
882 *) Major restructuring to the underlying ENGINE code. This includes
883 reduction of linker bloat, separation of pure "ENGINE" manipulation
884 (initialisation, etc) from functionality dealing with implementations
885 of specific crypto iterfaces. This change also introduces integrated
886 support for symmetric ciphers and digest implementations - so ENGINEs
887 can now accelerate these by providing EVP_CIPHER and EVP_MD
888 implementations of their own. This is detailed in crypto/engine/README
889 as it couldn't be adequately described here. However, there are a few
890 API changes worth noting - some RSA, DSA, DH, and RAND functions that
891 were changed in the original introduction of ENGINE code have now
892 reverted back - the hooking from this code to ENGINE is now a good
893 deal more passive and at run-time, operations deal directly with
894 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
895 dereferencing through an ENGINE pointer any more. Also, the ENGINE
896 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
897 they were not being used by the framework as there is no concept of a
898 BIGNUM_METHOD and they could not be generalised to the new
899 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
900 ENGINE_cpy() has been removed as it cannot be consistently defined in
904 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
907 *) Change mkdef.pl to sort symbols that get the same entry number,
908 and make sure the automatically generated functions ERR_load_*
909 become part of libeay.num as well.
912 *) New function SSL_renegotiate_pending(). This returns true once
913 renegotiation has been requested (either SSL_renegotiate() call
914 or HelloRequest/ClientHello receveived from the peer) and becomes
915 false once a handshake has been completed.
916 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
917 sends a HelloRequest, but does not ensure that a handshake takes
918 place. SSL_renegotiate_pending() is useful for checking if the
919 client has followed the request.)
922 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
923 By default, clients may request session resumption even during
924 renegotiation (if session ID contexts permit); with this option,
925 session resumption is possible only in the first handshake.
927 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
928 more bits available for options that should not be part of
929 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
932 *) Add some demos for certificate and certificate request creation.
935 *) Make maximum certificate chain size accepted from the peer application
936 settable (SSL*_get/set_max_cert_list()), as proposed by
937 "Douglas E. Engert" <deengert@anl.gov>.
940 *) Add support for shared libraries for Unixware-7
941 (Boyd Lynn Gerber <gerberb@zenez.com>).
944 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
945 be done prior to destruction. Use this to unload error strings from
946 ENGINEs that load their own error strings. NB: This adds two new API
947 functions to "get" and "set" this destroy handler in an ENGINE.
950 *) Alter all existing ENGINE implementations (except "openssl" and
951 "openbsd") to dynamically instantiate their own error strings. This
952 makes them more flexible to be built both as statically-linked ENGINEs
953 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
954 Also, add stub code to each that makes building them as self-contained
955 shared-libraries easier (see README.ENGINE).
958 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
959 implementations into applications that are completely implemented in
960 self-contained shared-libraries. The "dynamic" ENGINE exposes control
961 commands that can be used to configure what shared-library to load and
962 to control aspects of the way it is handled. Also, made an update to
963 the README.ENGINE file that brings its information up-to-date and
964 provides some information and instructions on the "dynamic" ENGINE
965 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
968 *) Make it possible to unload ranges of ERR strings with a new
969 "ERR_unload_strings" function.
972 *) Add a copy() function to EVP_MD.
975 *) Make EVP_MD routines take a context pointer instead of just the
976 md_data void pointer.
979 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
980 that the digest can only process a single chunk of data
981 (typically because it is provided by a piece of
982 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
983 is only going to provide a single chunk of data, and hence the
984 framework needn't accumulate the data for oneshot drivers.
987 *) As with "ERR", make it possible to replace the underlying "ex_data"
988 functions. This change also alters the storage and management of global
989 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
990 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
991 index counters. The API functions that use this state have been changed
992 to take a "class_index" rather than pointers to the class's local STACK
993 and counter, and there is now an API function to dynamically create new
994 classes. This centralisation allows us to (a) plug a lot of the
995 thread-safety problems that existed, and (b) makes it possible to clean
996 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
997 such data would previously have always leaked in application code and
998 workarounds were in place to make the memory debugging turn a blind eye
999 to it. Application code that doesn't use this new function will still
1000 leak as before, but their memory debugging output will announce it now
1001 rather than letting it slide.
1003 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
1004 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
1005 has a return value to indicate success or failure.
1008 *) Make it possible to replace the underlying "ERR" functions such that the
1009 global state (2 LHASH tables and 2 locks) is only used by the "default"
1010 implementation. This change also adds two functions to "get" and "set"
1011 the implementation prior to it being automatically set the first time
1012 any other ERR function takes place. Ie. an application can call "get",
1013 pass the return value to a module it has just loaded, and that module
1014 can call its own "set" function using that value. This means the
1015 module's "ERR" operations will use (and modify) the error state in the
1016 application and not in its own statically linked copy of OpenSSL code.
1019 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
1020 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
1021 the operation, and provides a more encapsulated way for external code
1022 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
1023 to use these functions rather than manually incrementing the counts.
1025 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
1028 *) Add EVP test program.
1031 *) Add symmetric cipher support to ENGINE. Expect the API to change!
1034 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
1035 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
1036 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
1037 These allow a CRL to be built without having to access X509_CRL fields
1038 directly. Modify 'ca' application to use new functions.
1041 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
1042 bug workarounds. Rollback attack detection is a security feature.
1043 The problem will only arise on OpenSSL servers when TLSv1 is not
1044 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
1045 Software authors not wanting to support TLSv1 will have special reasons
1046 for their choice and can explicitly enable this option.
1047 [Bodo Moeller, Lutz Jaenicke]
1049 *) Rationalise EVP so it can be extended: don't include a union of
1050 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
1051 (similar to those existing for EVP_CIPHER_CTX).
1056 EVP_MD_CTX_init(&md); /* new function call */
1057 EVP_DigestInit(&md, EVP_sha1());
1058 EVP_DigestUpdate(&md, in, len);
1059 EVP_DigestFinal(&md, out, NULL);
1060 EVP_MD_CTX_cleanup(&md); /* new function call */
1064 *) Make DES key schedule conform to the usual scheme, as well as
1065 correcting its structure. This means that calls to DES functions
1066 now have to pass a pointer to a des_key_schedule instead of a
1067 plain des_key_schedule (which was actually always a pointer
1070 des_key_schedule ks;
1072 des_set_key_checked(..., &ks);
1073 des_ncbc_encrypt(..., &ks, ...);
1075 (Note that a later change renames 'des_...' into 'DES_...'.)
1078 *) Initial reduction of linker bloat: the use of some functions, such as
1079 PEM causes large amounts of unused functions to be linked in due to
1080 poor organisation. For example pem_all.c contains every PEM function
1081 which has a knock on effect of linking in large amounts of (unused)
1082 ASN1 code. Grouping together similar functions and splitting unrelated
1083 functions prevents this.
1086 *) Cleanup of EVP macros.
1089 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
1090 correct _ecb suffix.
1093 *) Add initial OCSP responder support to ocsp application. The
1094 revocation information is handled using the text based index
1095 use by the ca application. The responder can either handle
1096 requests generated internally, supplied in files (for example
1097 via a CGI script) or using an internal minimal server.
1100 *) Add configuration choices to get zlib compression for TLS.
1103 *) Changes to Kerberos SSL for RFC 2712 compliance:
1104 1. Implemented real KerberosWrapper, instead of just using
1105 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
1106 2. Implemented optional authenticator field of KerberosWrapper.
1108 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
1109 and authenticator structs; see crypto/krb5/.
1111 Generalized Kerberos calls to support multiple Kerberos libraries.
1112 [Vern Staats <staatsvr@asc.hpc.mil>,
1113 Jeffrey Altman <jaltman@columbia.edu>
1114 via Richard Levitte]
1116 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
1117 already does with RSA. testdsa.h now has 'priv_key/pub_key'
1118 values for each of the key sizes rather than having just
1119 parameters (and 'speed' generating keys each time).
1122 *) Speed up EVP routines.
1125 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
1126 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
1127 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
1128 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
1130 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
1131 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
1132 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
1135 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
1137 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
1140 *) Added the OS2-EMX target.
1141 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
1143 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
1144 to support NCONF routines in extension code. New function CONF_set_nconf()
1145 to allow functions which take an NCONF to also handle the old LHASH
1146 structure: this means that the old CONF compatible routines can be
1147 retained (in particular wrt extensions) without having to duplicate the
1148 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
1151 *) Enhance the general user interface with mechanisms for inner control
1152 and with possibilities to have yes/no kind of prompts.
1155 *) Change all calls to low level digest routines in the library and
1156 applications to use EVP. Add missing calls to HMAC_cleanup() and
1157 don't assume HMAC_CTX can be copied using memcpy().
1158 [Verdon Walker <VWalker@novell.com>, Steve Henson]
1160 *) Add the possibility to control engines through control names but with
1161 arbitrary arguments instead of just a string.
1162 Change the key loaders to take a UI_METHOD instead of a callback
1163 function pointer. NOTE: this breaks binary compatibility with earlier
1164 versions of OpenSSL [engine].
1165 Adapt the nCipher code for these new conditions and add a card insertion
1169 *) Enhance the general user interface with mechanisms to better support
1170 dialog box interfaces, application-defined prompts, the possibility
1171 to use defaults (for example default passwords from somewhere else)
1172 and interrupts/cancellations.
1175 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
1176 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
1179 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
1180 tidy up some unnecessarily weird code in 'sk_new()').
1181 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
1183 *) Change the key loading routines for ENGINEs to use the same kind
1184 callback (pem_password_cb) as all other routines that need this
1188 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
1189 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
1190 than this minimum value is recommended.
1193 *) New random seeder for OpenVMS, using the system process statistics
1194 that are easily reachable.
1197 *) Windows apparently can't transparently handle global
1198 variables defined in DLLs. Initialisations such as:
1200 const ASN1_ITEM *it = &ASN1_INTEGER_it;
1202 wont compile. This is used by the any applications that need to
1203 declare their own ASN1 modules. This was fixed by adding the option
1204 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
1205 needed for static libraries under Win32.
1208 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
1209 setting of purpose and trust fields. New X509_STORE trust and
1210 purpose functions and tidy up setting in other SSL functions.
1213 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
1214 structure. These are inherited by X509_STORE_CTX when it is
1215 initialised. This allows various defaults to be set in the
1216 X509_STORE structure (such as flags for CRL checking and custom
1217 purpose or trust settings) for functions which only use X509_STORE_CTX
1218 internally such as S/MIME.
1220 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
1221 trust settings if they are not set in X509_STORE. This allows X509_STORE
1222 purposes and trust (in S/MIME for example) to override any set by default.
1224 Add command line options for CRL checking to smime, s_client and s_server
1228 *) Initial CRL based revocation checking. If the CRL checking flag(s)
1229 are set then the CRL is looked up in the X509_STORE structure and
1230 its validity and signature checked, then if the certificate is found
1231 in the CRL the verify fails with a revoked error.
1233 Various new CRL related callbacks added to X509_STORE_CTX structure.
1235 Command line options added to 'verify' application to support this.
1237 This needs some additional work, such as being able to handle multiple
1238 CRLs with different times, extension based lookup (rather than just
1239 by subject name) and ultimately more complete V2 CRL extension
1243 *) Add a general user interface API (crypto/ui/). This is designed
1244 to replace things like des_read_password and friends (backward
1245 compatibility functions using this new API are provided).
1246 The purpose is to remove prompting functions from the DES code
1247 section as well as provide for prompting through dialog boxes in
1248 a window system and the like.
1251 *) Add "ex_data" support to ENGINE so implementations can add state at a
1252 per-structure level rather than having to store it globally.
1255 *) Make it possible for ENGINE structures to be copied when retrieved by
1256 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
1257 This causes the "original" ENGINE structure to act like a template,
1258 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
1259 operational state can be localised to each ENGINE structure, despite the
1260 fact they all share the same "methods". New ENGINE structures returned in
1261 this case have no functional references and the return value is the single
1262 structural reference. This matches the single structural reference returned
1263 by ENGINE_by_id() normally, when it is incremented on the pre-existing
1267 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
1268 needs to match any other type at all we need to manually clear the
1272 *) Changes to the "openssl engine" utility to include;
1273 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
1274 about an ENGINE's available control commands.
1275 - executing control commands from command line arguments using the
1276 '-pre' and '-post' switches. '-post' is only used if '-t' is
1277 specified and the ENGINE is successfully initialised. The syntax for
1278 the individual commands are colon-separated, for example;
1279 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
1282 *) New dynamic control command support for ENGINEs. ENGINEs can now
1283 declare their own commands (numbers), names (strings), descriptions,
1284 and input types for run-time discovery by calling applications. A
1285 subset of these commands are implicitly classed as "executable"
1286 depending on their input type, and only these can be invoked through
1287 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
1288 can be based on user input, config files, etc). The distinction is
1289 that "executable" commands cannot return anything other than a boolean
1290 result and can only support numeric or string input, whereas some
1291 discoverable commands may only be for direct use through
1292 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
1293 pointers, or other custom uses. The "executable" commands are to
1294 support parameterisations of ENGINE behaviour that can be
1295 unambiguously defined by ENGINEs and used consistently across any
1296 OpenSSL-based application. Commands have been added to all the
1297 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
1298 control over shared-library paths without source code alterations.
1301 *) Changed all ENGINE implementations to dynamically allocate their
1302 ENGINEs rather than declaring them statically. Apart from this being
1303 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
1304 this also allows the implementations to compile without using the
1305 internal engine_int.h header.
1308 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
1309 'const' value. Any code that should be able to modify a RAND_METHOD
1310 should already have non-const pointers to it (ie. they should only
1311 modify their own ones).
1314 *) Made a variety of little tweaks to the ENGINE code.
1315 - "atalla" and "ubsec" string definitions were moved from header files
1316 to C code. "nuron" string definitions were placed in variables
1317 rather than hard-coded - allowing parameterisation of these values
1318 later on via ctrl() commands.
1319 - Removed unused "#if 0"'d code.
1320 - Fixed engine list iteration code so it uses ENGINE_free() to release
1321 structural references.
1322 - Constified the RAND_METHOD element of ENGINE structures.
1323 - Constified various get/set functions as appropriate and added
1324 missing functions (including a catch-all ENGINE_cpy that duplicates
1325 all ENGINE values onto a new ENGINE except reference counts/state).
1326 - Removed NULL parameter checks in get/set functions. Setting a method
1327 or function to NULL is a way of cancelling out a previously set
1328 value. Passing a NULL ENGINE parameter is just plain stupid anyway
1329 and doesn't justify the extra error symbols and code.
1330 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
1331 flags from engine_int.h to engine.h.
1332 - Changed prototypes for ENGINE handler functions (init(), finish(),
1333 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
1336 *) Implement binary inversion algorithm for BN_mod_inverse in addition
1337 to the algorithm using long division. The binary algorithm can be
1338 used only if the modulus is odd. On 32-bit systems, it is faster
1339 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
1340 roughly 5-15% for 256-bit moduli), so we use it only for moduli
1341 up to 450 bits. In 64-bit environments, the binary algorithm
1342 appears to be advantageous for much longer moduli; here we use it
1343 for moduli up to 2048 bits.
1346 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
1347 could not support the combine flag in choice fields.
1350 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
1351 extensions from a certificate request to the certificate.
1354 *) Allow multiple 'certopt' and 'nameopt' options to be separated
1355 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
1356 file: this allows the display of the certificate about to be
1357 signed to be customised, to allow certain fields to be included
1358 or excluded and extension details. The old system didn't display
1359 multicharacter strings properly, omitted fields not in the policy
1360 and couldn't display additional details such as extensions.
1363 *) Function EC_POINTs_mul for multiple scalar multiplication
1364 of an arbitrary number of elliptic curve points
1365 \sum scalars[i]*points[i],
1366 optionally including the generator defined for the EC_GROUP:
1367 scalar*generator + \sum scalars[i]*points[i].
1369 EC_POINT_mul is a simple wrapper function for the typical case
1370 that the point list has just one item (besides the optional
1374 *) First EC_METHODs for curves over GF(p):
1376 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
1377 operations and provides various method functions that can also
1378 operate with faster implementations of modular arithmetic.
1380 EC_GFp_mont_method() reuses most functions that are part of
1381 EC_GFp_simple_method, but uses Montgomery arithmetic.
1383 [Bodo Moeller; point addition and point doubling
1384 implementation directly derived from source code provided by
1385 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
1387 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
1388 crypto/ec/ec_lib.c):
1390 Curves are EC_GROUP objects (with an optional group generator)
1391 based on EC_METHODs that are built into the library.
1393 Points are EC_POINT objects based on EC_GROUP objects.
1395 Most of the framework would be able to handle curves over arbitrary
1396 finite fields, but as there are no obvious types for fields other
1397 than GF(p), some functions are limited to that for now.
1400 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
1401 that the file contains a complete HTTP response.
1404 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
1405 change the def and num file printf format specifier from "%-40sXXX"
1406 to "%-39s XXX". The latter will always guarantee a space after the
1407 field while the former will cause them to run together if the field
1408 is 40 of more characters long.
1411 *) Constify the cipher and digest 'method' functions and structures
1412 and modify related functions to take constant EVP_MD and EVP_CIPHER
1416 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
1417 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
1420 *) Modify EVP_Digest*() routines so they now return values. Although the
1421 internal software routines can never fail additional hardware versions
1425 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
1427 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
1428 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
1431 ERR_R_NESTED_ASN1_ERROR
1433 ERR_R_MISSING_ASN1_EOS
1434 were 4 .. 9, conflicting with
1435 ERR_LIB_RSA (= ERR_R_RSA_LIB)
1437 ERR_LIB_PEM (= ERR_R_PEM_LIB).
1438 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
1440 Add new error code 'ERR_R_INTERNAL_ERROR'.
1443 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
1447 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
1448 sets the subject name for a new request or supersedes the
1449 subject name in a given request. Formats that can be parsed are
1450 'CN=Some Name, OU=myOU, C=IT'
1452 'CN=Some Name/OU=myOU/C=IT'.
1454 Add options '-batch' and '-verbose' to 'openssl req'.
1455 [Massimiliano Pala <madwolf@hackmasters.net>]
1457 *) Introduce the possibility to access global variables through
1458 functions on platform were that's the best way to handle exporting
1459 global variables in shared libraries. To enable this functionality,
1460 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1461 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1462 is normally done by Configure or something similar).
1464 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1465 in the source file (foo.c) like this:
1467 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1468 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1470 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1471 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1473 OPENSSL_DECLARE_GLOBAL(int,foo);
1474 #define foo OPENSSL_GLOBAL_REF(foo)
1475 OPENSSL_DECLARE_GLOBAL(double,bar);
1476 #define bar OPENSSL_GLOBAL_REF(bar)
1478 The #defines are very important, and therefore so is including the
1479 header file everywhere where the defined globals are used.
1481 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1482 of ASN.1 items, but that structure is a bit different.
1484 The largest change is in util/mkdef.pl which has been enhanced with
1485 better and easier to understand logic to choose which symbols should
1486 go into the Windows .def files as well as a number of fixes and code
1487 cleanup (among others, algorithm keywords are now sorted
1488 lexicographically to avoid constant rewrites).
1491 *) In BN_div() keep a copy of the sign of 'num' before writing the
1492 result to 'rm' because if rm==num the value will be overwritten
1493 and produce the wrong result if 'num' is negative: this caused
1494 problems with BN_mod() and BN_nnmod().
1497 *) Function OCSP_request_verify(). This checks the signature on an
1498 OCSP request and verifies the signer certificate. The signer
1499 certificate is just checked for a generic purpose and OCSP request
1503 *) Add OCSP_check_validity() function to check the validity of OCSP
1504 responses. OCSP responses are prepared in real time and may only
1505 be a few seconds old. Simply checking that the current time lies
1506 between thisUpdate and nextUpdate max reject otherwise valid responses
1507 caused by either OCSP responder or client clock inaccuracy. Instead
1508 we allow thisUpdate and nextUpdate to fall within a certain period of
1509 the current time. The age of the response can also optionally be
1510 checked. Two new options -validity_period and -status_age added to
1514 *) If signature or public key algorithm is unrecognized print out its
1515 OID rather that just UNKNOWN.
1518 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1519 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1520 ID to be generated from the issuer certificate alone which can then be
1521 passed to OCSP_id_issuer_cmp().
1524 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1525 ASN1 modules to export functions returning ASN1_ITEM pointers
1526 instead of the ASN1_ITEM structures themselves. This adds several
1527 new macros which allow the underlying ASN1 function/structure to
1528 be accessed transparently. As a result code should not use ASN1_ITEM
1529 references directly (such as &X509_it) but instead use the relevant
1530 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1531 use of the new ASN1 code on platforms where exporting structures
1532 is problematical (for example in shared libraries) but exporting
1533 functions returning pointers to structures is not.
1536 *) Add support for overriding the generation of SSL/TLS session IDs.
1537 These callbacks can be registered either in an SSL_CTX or per SSL.
1538 The purpose of this is to allow applications to control, if they wish,
1539 the arbitrary values chosen for use as session IDs, particularly as it
1540 can be useful for session caching in multiple-server environments. A
1541 command-line switch for testing this (and any client code that wishes
1542 to use such a feature) has been added to "s_server".
1543 [Geoff Thorpe, Lutz Jaenicke]
1545 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
1546 of the form '#if defined(...) || defined(...) || ...' and
1547 '#if !defined(...) && !defined(...) && ...'. This also avoids
1548 the growing number of special cases it was previously handling.
1551 *) Make all configuration macros available for application by making
1552 sure they are available in opensslconf.h, by giving them names starting
1553 with "OPENSSL_" to avoid conflicts with other packages and by making
1554 sure e_os2.h will cover all platform-specific cases together with
1556 Additionally, it is now possible to define configuration/platform-
1557 specific names (called "system identities"). In the C code, these
1558 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1559 macro with the name beginning with "OPENSSL_SYS_", which is determined
1560 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1564 *) New option -set_serial to 'req' and 'x509' this allows the serial
1565 number to use to be specified on the command line. Previously self
1566 signed certificates were hard coded with serial number 0 and the
1567 CA options of 'x509' had to use a serial number in a file which was
1571 *) New options to 'ca' utility to support V2 CRL entry extensions.
1572 Currently CRL reason, invalidity date and hold instruction are
1573 supported. Add new CRL extensions to V3 code and some new objects.
1576 *) New function EVP_CIPHER_CTX_set_padding() this is used to
1577 disable standard block padding (aka PKCS#5 padding) in the EVP
1578 API, which was previously mandatory. This means that the data is
1579 not padded in any way and so the total length much be a multiple
1580 of the block size, otherwise an error occurs.
1583 *) Initial (incomplete) OCSP SSL support.
1586 *) New function OCSP_parse_url(). This splits up a URL into its host,
1587 port and path components: primarily to parse OCSP URLs. New -url
1588 option to ocsp utility.
1591 *) New nonce behavior. The return value of OCSP_check_nonce() now
1592 reflects the various checks performed. Applications can decide
1593 whether to tolerate certain situations such as an absent nonce
1594 in a response when one was present in a request: the ocsp application
1595 just prints out a warning. New function OCSP_add1_basic_nonce()
1596 this is to allow responders to include a nonce in a response even if
1597 the request is nonce-less.
1600 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1601 skipped when using openssl x509 multiple times on a single input file,
1602 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1605 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1606 set string type: to handle setting ASN1_TIME structures. Fix ca
1607 utility to correctly initialize revocation date of CRLs.
1610 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1611 the clients preferred ciphersuites and rather use its own preferences.
1612 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1613 Internet Explorer by ensuring unchanged hash method during stepup.
1614 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1617 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1618 to aes and add a new 'exist' option to print out symbols that don't
1622 *) Additional options to ocsp utility to allow flags to be set and
1623 additional certificates supplied.
1626 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
1627 OCSP client a number of certificate to only verify the response
1631 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1632 handle the new API. Currently only ECB, CBC modes supported. Add new
1635 Add TLS AES ciphersuites as described in RFC3268, "Advanced
1636 Encryption Standard (AES) Ciphersuites for Transport Layer
1637 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
1638 not enabled by default and were not part of the "ALL" ciphersuite
1639 alias because they were not yet official; they could be
1640 explicitly requested by specifying the "AESdraft" ciphersuite
1641 group alias. In the final release of OpenSSL 0.9.7, the group
1642 alias is called "AES" and is part of "ALL".)
1643 [Ben Laurie, Steve Henson, Bodo Moeller]
1645 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
1646 request to response.
1649 *) Functions for OCSP responders. OCSP_request_onereq_count(),
1650 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1651 extract information from a certificate request. OCSP_response_create()
1652 creates a response and optionally adds a basic response structure.
1653 OCSP_basic_add1_status() adds a complete single response to a basic
1654 response and returns the OCSP_SINGLERESP structure just added (to allow
1655 extensions to be included for example). OCSP_basic_add1_cert() adds a
1656 certificate to a basic response and OCSP_basic_sign() signs a basic
1657 response with various flags. New helper functions ASN1_TIME_check()
1658 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1659 (converts ASN1_TIME to GeneralizedTime).
1662 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1663 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1664 structure from a certificate. X509_pubkey_digest() digests the public_key
1665 contents: this is used in various key identifiers.
1668 *) Make sk_sort() tolerate a NULL argument.
1669 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1671 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1672 passed by the function are trusted implicitly. If any of them signed the
1673 response then it is assumed to be valid and is not verified.
1676 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1677 to data. This was previously part of the PKCS7 ASN1 code. This
1678 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1679 [Steve Henson, reported by Kenneth R. Robinette
1680 <support@securenetterm.com>]
1682 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1683 routines: without these tracing memory leaks is very painful.
1684 Fix leaks in PKCS12 and PKCS7 routines.
1687 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1688 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1689 effectively meant GeneralizedTime would never be used. Now it
1690 is initialised to -1 but X509_time_adj() now has to check the value
1691 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1692 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1693 [Steve Henson, reported by Kenneth R. Robinette
1694 <support@securenetterm.com>]
1696 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1697 result in a zero length in the ASN1_INTEGER structure which was
1698 not consistent with the structure when d2i_ASN1_INTEGER() was used
1699 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1700 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1701 where it did not print out a minus for negative ASN1_INTEGER.
1704 *) Add summary printout to ocsp utility. The various functions which
1705 convert status values to strings have been renamed to:
1706 OCSP_response_status_str(), OCSP_cert_status_str() and
1707 OCSP_crl_reason_str() and are no longer static. New options
1708 to verify nonce values and to disable verification. OCSP response
1709 printout format cleaned up.
1712 *) Add additional OCSP certificate checks. These are those specified
1713 in RFC2560. This consists of two separate checks: the CA of the
1714 certificate being checked must either be the OCSP signer certificate
1715 or the issuer of the OCSP signer certificate. In the latter case the
1716 OCSP signer certificate must contain the OCSP signing extended key
1717 usage. This check is performed by attempting to match the OCSP
1718 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1719 in the OCSP_CERTID structures of the response.
1722 *) Initial OCSP certificate verification added to OCSP_basic_verify()
1723 and related routines. This uses the standard OpenSSL certificate
1724 verify routines to perform initial checks (just CA validity) and
1725 to obtain the certificate chain. Then additional checks will be
1726 performed on the chain. Currently the root CA is checked to see
1727 if it is explicitly trusted for OCSP signing. This is used to set
1728 a root CA as a global signing root: that is any certificate that
1729 chains to that CA is an acceptable OCSP signing certificate.
1732 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1733 extensions from a separate configuration file.
1734 As when reading extensions from the main configuration file,
1735 the '-extensions ...' option may be used for specifying the
1737 [Massimiliano Pala <madwolf@comune.modena.it>]
1739 *) New OCSP utility. Allows OCSP requests to be generated or
1740 read. The request can be sent to a responder and the output
1741 parsed, outputed or printed in text form. Not complete yet:
1742 still needs to check the OCSP response validity.
1745 *) New subcommands for 'openssl ca':
1746 'openssl ca -status <serial>' prints the status of the cert with
1747 the given serial number (according to the index file).
1748 'openssl ca -updatedb' updates the expiry status of certificates
1750 [Massimiliano Pala <madwolf@comune.modena.it>]
1752 *) New '-newreq-nodes' command option to CA.pl. This is like
1753 '-newreq', but calls 'openssl req' with the '-nodes' option
1754 so that the resulting key is not encrypted.
1755 [Damien Miller <djm@mindrot.org>]
1757 *) New configuration for the GNU Hurd.
1758 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1760 *) Initial code to implement OCSP basic response verify. This
1761 is currently incomplete. Currently just finds the signer's
1762 certificate and verifies the signature on the response.
1765 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1766 value of OPENSSLDIR. This is available via the new '-d' option
1767 to 'openssl version', and is also included in 'openssl version -a'.
1770 *) Allowing defining memory allocation callbacks that will be given
1771 file name and line number information in additional arguments
1772 (a const char* and an int). The basic functionality remains, as
1773 well as the original possibility to just replace malloc(),
1774 realloc() and free() by functions that do not know about these
1775 additional arguments. To register and find out the current
1776 settings for extended allocation functions, the following
1777 functions are provided:
1779 CRYPTO_set_mem_ex_functions
1780 CRYPTO_set_locked_mem_ex_functions
1781 CRYPTO_get_mem_ex_functions
1782 CRYPTO_get_locked_mem_ex_functions
1784 These work the same way as CRYPTO_set_mem_functions and friends.
1785 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1786 extended allocation function is enabled.
1787 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1788 a conventional allocation function is enabled.
1789 [Richard Levitte, Bodo Moeller]
1791 *) Finish off removing the remaining LHASH function pointer casts.
1792 There should no longer be any prototype-casting required when using
1793 the LHASH abstraction, and any casts that remain are "bugs". See
1794 the callback types and macros at the head of lhash.h for details
1795 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1798 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1799 If /dev/[u]random devices are not available or do not return enough
1800 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1802 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1803 /etc/entropy will be queried once each in this sequence, quering stops
1804 when enough entropy was collected without querying more sockets.
1807 *) Change the Unix RAND_poll() variant to be able to poll several
1808 random devices, as specified by DEVRANDOM, until a sufficient amount
1809 of data has been collected. We spend at most 10 ms on each file
1810 (select timeout) and read in non-blocking mode. DEVRANDOM now
1811 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1812 (previously it was just the string "/dev/urandom"), so on typical
1813 platforms the 10 ms delay will never occur.
1814 Also separate out the Unix variant to its own file, rand_unix.c.
1815 For VMS, there's a currently-empty rand_vms.c.
1818 *) Move OCSP client related routines to ocsp_cl.c. These
1819 provide utility functions which an application needing
1820 to issue a request to an OCSP responder and analyse the
1821 response will typically need: as opposed to those which an
1822 OCSP responder itself would need which will be added later.
1824 OCSP_request_sign() signs an OCSP request with an API similar
1825 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1826 response. OCSP_response_get1_basic() extracts basic response
1827 from response. OCSP_resp_find_status(): finds and extracts status
1828 information from an OCSP_CERTID structure (which will be created
1829 when the request structure is built). These are built from lower
1830 level functions which work on OCSP_SINGLERESP structures but
1831 wont normally be used unless the application wishes to examine
1832 extensions in the OCSP response for example.
1834 Replace nonce routines with a pair of functions.
1835 OCSP_request_add1_nonce() adds a nonce value and optionally
1836 generates a random value. OCSP_check_nonce() checks the
1837 validity of the nonce in an OCSP response.
1840 *) Change function OCSP_request_add() to OCSP_request_add0_id().
1841 This doesn't copy the supplied OCSP_CERTID and avoids the
1842 need to free up the newly created id. Change return type
1843 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1844 This can then be used to add extensions to the request.
1845 Deleted OCSP_request_new(), since most of its functionality
1846 is now in OCSP_REQUEST_new() (and the case insensitive name
1847 clash) apart from the ability to set the request name which
1848 will be added elsewhere.
1851 *) Update OCSP API. Remove obsolete extensions argument from
1852 various functions. Extensions are now handled using the new
1853 OCSP extension code. New simple OCSP HTTP function which
1854 can be used to send requests and parse the response.
1857 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1858 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1859 uses the special reorder version of SET OF to sort the attributes
1860 and reorder them to match the encoded order. This resolves a long
1861 standing problem: a verify on a PKCS7 structure just after signing
1862 it used to fail because the attribute order did not match the
1863 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1864 it uses the received order. This is necessary to tolerate some broken
1865 software that does not order SET OF. This is handled by encoding
1866 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1867 to produce the required SET OF.
1870 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1871 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1872 files to get correct declarations of the ASN.1 item variables.
1875 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1876 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1877 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1878 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1879 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1880 ASN1_ITEM and no wrapper functions.
1883 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1884 replace the old function pointer based I/O routines. Change most of
1885 the *_d2i_bio() and *_d2i_fp() functions to use these.
1888 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1889 lines, recognice more "algorithms" that can be deselected, and make
1890 it complain about algorithm deselection that isn't recognised.
1893 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
1894 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1895 to use new functions. Add NO_ASN1_OLD which can be set to remove
1896 some old style ASN1 functions: this can be used to determine if old
1897 code will still work when these eventually go away.
1900 *) New extension functions for OCSP structures, these follow the
1901 same conventions as certificates and CRLs.
1904 *) New function X509V3_add1_i2d(). This automatically encodes and
1905 adds an extension. Its behaviour can be customised with various
1906 flags to append, replace or delete. Various wrappers added for
1907 certifcates and CRLs.
1910 *) Fix to avoid calling the underlying ASN1 print routine when
1911 an extension cannot be parsed. Correct a typo in the
1912 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1915 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1916 entries for variables.
1919 *) Add functionality to apps/openssl.c for detecting locking
1920 problems: As the program is single-threaded, all we have
1921 to do is register a locking callback using an array for
1922 storing which locks are currently held by the program.
1925 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
1926 SSL_get_ex_data_X509_STORE_idx(), which is used in
1927 ssl_verify_cert_chain() and thus can be called at any time
1928 during TLS/SSL handshakes so that thread-safety is essential.
1929 Unfortunately, the ex_data design is not at all suited
1930 for multi-threaded use, so it probably should be abolished.
1933 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
1934 [Broadcom, tweaked and integrated by Geoff Thorpe]
1936 *) Move common extension printing code to new function
1937 X509V3_print_extensions(). Reorganise OCSP print routines and
1938 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1941 *) New function X509_signature_print() to remove duplication in some
1945 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
1946 set (this was treated exactly the same as SET OF previously). This
1947 is used to reorder the STACK representing the structure to match the
1948 encoding. This will be used to get round a problem where a PKCS7
1949 structure which was signed could not be verified because the STACK
1950 order did not reflect the encoded order.
1953 *) Reimplement the OCSP ASN1 module using the new code.
1956 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1957 for its ASN1 operations. The old style function pointers still exist
1958 for now but they will eventually go away.
1961 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
1962 completely replaces the old ASN1 functionality with a table driven
1963 encoder and decoder which interprets an ASN1_ITEM structure describing
1964 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1965 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1966 has also been converted to the new form.
1969 *) Change BN_mod_exp_recp so that negative moduli are tolerated
1970 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1971 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1972 for negative moduli.
1975 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
1976 of not touching the result's sign bit.
1979 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1983 *) Changed the LHASH code to use prototypes for callbacks, and created
1984 macros to declare and implement thin (optionally static) functions
1985 that provide type-safety and avoid function pointer casting for the
1986 type-specific callbacks.
1989 *) Added Kerberos Cipher Suites to be used with TLS, as written in
1991 [Veers Staats <staatsvr@asc.hpc.mil>,
1992 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1994 *) Reformat the FAQ so the different questions and answers can be divided
1995 in sections depending on the subject.
1998 *) Have the zlib compression code load ZLIB.DLL dynamically under
2002 *) New function BN_mod_sqrt for computing square roots modulo a prime
2003 (using the probabilistic Tonelli-Shanks algorithm unless
2004 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
2005 be handled deterministically).
2006 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2008 *) Make BN_mod_inverse faster by explicitly handling small quotients
2009 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
2010 512 bits], about 30% for larger ones [1024 or 2048 bits].)
2013 *) New function BN_kronecker.
2016 *) Fix BN_gcd so that it works on negative inputs; the result is
2017 positive unless both parameters are zero.
2018 Previously something reasonably close to an infinite loop was
2019 possible because numbers could be growing instead of shrinking
2020 in the implementation of Euclid's algorithm.
2023 *) Fix BN_is_word() and BN_is_one() macros to take into account the
2024 sign of the number in question.
2026 Fix BN_is_word(a,w) to work correctly for w == 0.
2028 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
2029 because its test if the absolute value of 'a' equals 'w'.
2030 Note that BN_abs_is_word does *not* handle w == 0 reliably;
2031 it exists mostly for use in the implementations of BN_is_zero(),
2032 BN_is_one(), and BN_is_word().
2035 *) New function BN_swap.
2038 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
2039 the exponentiation functions are more likely to produce reasonable
2040 results on negative inputs.
2043 *) Change BN_mod_mul so that the result is always non-negative.
2044 Previously, it could be negative if one of the factors was negative;
2045 I don't think anyone really wanted that behaviour.
2048 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
2049 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
2050 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
2051 and add new functions:
2060 BN_mod_lshift1_quick
2064 These functions always generate non-negative results.
2066 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
2067 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
2069 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2070 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2071 be reduced modulo m.
2072 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2075 The following entry accidentily appeared in the CHANGES file
2076 distributed with OpenSSL 0.9.7. The modifications described in
2077 it do *not* apply to OpenSSL 0.9.7.
2079 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2080 was actually never needed) and in BN_mul(). The removal in BN_mul()
2081 required a small change in bn_mul_part_recursive() and the addition
2082 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2083 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2084 bn_sub_words() and bn_add_words() except they take arrays with
2089 *) In 'openssl passwd', verify passwords read from the terminal
2090 unless the '-salt' option is used (which usually means that
2091 verification would just waste user's time since the resulting
2092 hash is going to be compared with some given password hash)
2093 or the new '-noverify' option is used.
2095 This is an incompatible change, but it does not affect
2096 non-interactive use of 'openssl passwd' (passwords on the command
2097 line, '-stdin' option, '-in ...' option) and thus should not
2101 *) Remove all references to RSAref, since there's no more need for it.
2104 *) Make DSO load along a path given through an environment variable
2105 (SHLIB_PATH) with shl_load().
2108 *) Constify the ENGINE code as a result of BIGNUM constification.
2109 Also constify the RSA code and most things related to it. In a
2110 few places, most notable in the depth of the ASN.1 code, ugly
2111 casts back to non-const were required (to be solved at a later
2115 *) Make it so the openssl application has all engines loaded by default.
2118 *) Constify the BIGNUM routines a little more.
2121 *) Add the following functions:
2123 ENGINE_load_cswift()
2125 ENGINE_load_atalla()
2127 ENGINE_load_builtin_engines()
2129 That way, an application can itself choose if external engines that
2130 are built-in in OpenSSL shall ever be used or not. The benefit is
2131 that applications won't have to be linked with libdl or other dso
2132 libraries unless it's really needed.
2134 Changed 'openssl engine' to load all engines on demand.
2135 Changed the engine header files to avoid the duplication of some
2136 declarations (they differed!).
2139 *) 'openssl engine' can now list capabilities.
2142 *) Better error reporting in 'openssl engine'.
2145 *) Never call load_dh_param(NULL) in s_server.
2148 *) Add engine application. It can currently list engines by name and
2149 identity, and test if they are actually available.
2152 *) Improve RPM specification file by forcing symbolic linking and making
2153 sure the installed documentation is also owned by root.root.
2154 [Damien Miller <djm@mindrot.org>]
2156 *) Give the OpenSSL applications more possibilities to make use of
2157 keys (public as well as private) handled by engines.
2160 *) Add OCSP code that comes from CertCo.
2163 *) Add VMS support for the Rijndael code.
2166 *) Added untested support for Nuron crypto accelerator.
2169 *) Add support for external cryptographic devices. This code was
2170 previously distributed separately as the "engine" branch.
2171 [Geoff Thorpe, Richard Levitte]
2173 *) Rework the filename-translation in the DSO code. It is now possible to
2174 have far greater control over how a "name" is turned into a filename
2175 depending on the operating environment and any oddities about the
2176 different shared library filenames on each system.
2179 *) Support threads on FreeBSD-elf in Configure.
2182 *) Fix for SHA1 assembly problem with MASM: it produces
2183 warnings about corrupt line number information when assembling
2184 with debugging information. This is caused by the overlapping
2186 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2189 NCONF_get_number() has no error checking at all. As a replacement,
2190 NCONF_get_number_e() is defined (_e for "error checking") and is
2191 promoted strongly. The old NCONF_get_number is kept around for
2192 binary backward compatibility.
2193 Make it possible for methods to load from something other than a BIO,
2194 by providing a function pointer that is given a name instead of a BIO.
2195 For example, this could be used to load configuration data from an
2199 *) Fix for non blocking accept BIOs. Added new I/O special reason
2200 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2201 with non blocking I/O was not possible because no retry code was
2202 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2206 *) Added the beginnings of Rijndael support.
2209 *) Fix for bug in DirectoryString mask setting. Add support for
2210 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2211 to allow certificate printing to more controllable, additional
2212 'certopt' option to 'x509' to allow new printing options to be
2216 *) Clean old EAY MD5 hack from e_os.h.
2219 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
2221 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2222 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
2223 [Joe Orton, Steve Henson]
2225 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
2227 *) Fix additional bug revealed by the NISCC test suite:
2229 Stop bug triggering large recursion when presented with
2230 certain ASN.1 tags (CAN-2003-0851)
2233 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
2235 *) Fix various bugs revealed by running the NISCC test suite:
2237 Stop out of bounds reads in the ASN1 code when presented with
2238 invalid tags (CAN-2003-0543 and CAN-2003-0544).
2240 If verify callback ignores invalid public key errors don't try to check
2241 certificate signature with the NULL public key.
2245 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2246 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2250 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2251 extra data after the compression methods not only for TLS 1.0
2252 but also for SSL 3.0 (as required by the specification).
2253 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2255 *) Change X509_certificate_type() to mark the key as exported/exportable
2256 when it's 512 *bits* long, not 512 bytes.
2259 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
2261 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2262 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2263 a protocol version number mismatch like a decryption error
2264 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2267 *) Turn on RSA blinding by default in the default implementation
2268 to avoid a timing attack. Applications that don't want it can call
2269 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2270 They would be ill-advised to do so in most cases.
2271 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2273 *) Change RSA blinding code so that it works when the PRNG is not
2274 seeded (in this case, the secret RSA exponent is abused as
2275 an unpredictable seed -- if it is not unpredictable, there
2276 is no point in blinding anyway). Make RSA blinding thread-safe
2277 by remembering the creator's thread ID in rsa->blinding and
2278 having all other threads use local one-time blinding factors
2279 (this requires more computation than sharing rsa->blinding, but
2280 avoids excessive locking; and if an RSA object is not shared
2281 between threads, blinding will still be very fast).
2284 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
2286 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2287 via timing by performing a MAC computation even if incorrrect
2288 block cipher padding has been found. This is a countermeasure
2289 against active attacks where the attacker has to distinguish
2290 between bad padding and a MAC verification error. (CAN-2003-0078)
2292 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2293 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2294 Martin Vuagnoux (EPFL, Ilion)]
2296 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
2298 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
2299 memory from it's contents. This is done with a counter that will
2300 place alternating values in each byte. This can be used to solve
2301 two issues: 1) the removal of calls to memset() by highly optimizing
2302 compilers, and 2) cleansing with other values than 0, since those can
2303 be read through on certain media, for example a swap space on disk.
2306 *) Bugfix: client side session caching did not work with external caching,
2307 because the session->cipher setting was not restored when reloading
2308 from the external cache. This problem was masked, when
2309 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
2310 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
2313 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
2314 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
2315 [Zeev Lieber <zeev-l@yahoo.com>]
2317 *) Undo an undocumented change introduced in 0.9.6e which caused
2318 repeated calls to OpenSSL_add_all_ciphers() and
2319 OpenSSL_add_all_digests() to be ignored, even after calling
2323 *) Change the default configuration reader to deal with last line not
2324 being properly terminated.
2327 *) Change X509_NAME_cmp() so it applies the special rules on handling
2328 DN values that are of type PrintableString, as well as RDNs of type
2329 emailAddress where the value has the type ia5String.
2330 [stefank@valicert.com via Richard Levitte]
2332 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
2333 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
2334 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
2335 the bitwise-OR of the two for use by the majority of applications
2336 wanting this behaviour, and update the docs. The documented
2337 behaviour and actual behaviour were inconsistent and had been
2338 changing anyway, so this is more a bug-fix than a behavioural
2340 [Geoff Thorpe, diagnosed by Nadav Har'El]
2342 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
2343 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
2346 *) Fix initialization code race conditions in
2347 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
2348 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
2349 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
2350 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
2351 ssl2_get_cipher_by_char(),
2352 ssl3_get_cipher_by_char().
2353 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
2355 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
2356 the cached sessions are flushed, as the remove_cb() might use ex_data
2357 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
2358 (see [openssl.org #212]).
2359 [Geoff Thorpe, Lutz Jaenicke]
2361 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
2362 length, instead of the encoding length to d2i_ASN1_OBJECT.
2365 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
2367 *) [In 0.9.6g-engine release:]
2368 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
2369 [Lynn Gazis <lgazis@rainbow.com>]
2371 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
2373 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
2374 and get fix the header length calculation.
2375 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
2376 Alon Kantor <alonk@checkpoint.com> (and others),
2379 *) Use proper error handling instead of 'assertions' in buffer
2380 overflow checks added in 0.9.6e. This prevents DoS (the
2381 assertions could call abort()).
2382 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
2384 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
2386 *) Add various sanity checks to asn1_get_length() to reject
2387 the ASN1 length bytes if they exceed sizeof(long), will appear
2388 negative or the content length exceeds the length of the
2390 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2392 *) Fix cipher selection routines: ciphers without encryption had no flags
2393 for the cipher strength set and where therefore not handled correctly
2394 by the selection routines (PR #130).
2397 *) Fix EVP_dsa_sha macro.
2401 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
2402 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
2403 that was added in OpenSSL 0.9.6d.
2405 As the countermeasure turned out to be incompatible with some
2406 broken SSL implementations, the new option is part of SSL_OP_ALL.
2407 SSL_OP_ALL is usually employed when compatibility with weird SSL
2408 implementations is desired (e.g. '-bugs' option to 's_client' and
2409 's_server'), so the new option is automatically set in many
2413 *) Changes in security patch:
2415 Changes marked "(CHATS)" were sponsored by the Defense Advanced
2416 Research Projects Agency (DARPA) and Air Force Research Laboratory,
2417 Air Force Materiel Command, USAF, under agreement number
2420 *) Add various sanity checks to asn1_get_length() to reject
2421 the ASN1 length bytes if they exceed sizeof(long), will appear
2422 negative or the content length exceeds the length of the
2423 supplied buffer. (CAN-2002-0659)
2424 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2426 *) Assertions for various potential buffer overflows, not known to
2428 [Ben Laurie (CHATS)]
2430 *) Various temporary buffers to hold ASCII versions of integers were
2431 too small for 64 bit platforms. (CAN-2002-0655)
2432 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
2434 *) Remote buffer overflow in SSL3 protocol - an attacker could
2435 supply an oversized session ID to a client. (CAN-2002-0656)
2436 [Ben Laurie (CHATS)]
2438 *) Remote buffer overflow in SSL2 protocol - an attacker could
2439 supply an oversized client master key. (CAN-2002-0656)
2440 [Ben Laurie (CHATS)]
2442 Changes between 0.9.6c and 0.9.6d [9 May 2002]
2444 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
2445 encoded as NULL) with id-dsa-with-sha1.
2446 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
2448 *) Check various X509_...() return values in apps/req.c.
2449 [Nils Larsch <nla@trustcenter.de>]
2451 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
2452 an end-of-file condition would erronously be flagged, when the CRLF
2453 was just at the end of a processed block. The bug was discovered when
2454 processing data through a buffering memory BIO handing the data to a
2455 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
2456 <ptsekov@syntrex.com> and Nedelcho Stanev.
2459 *) Implement a countermeasure against a vulnerability recently found
2460 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
2461 before application data chunks to avoid the use of known IVs
2462 with data potentially chosen by the attacker.
2465 *) Fix length checks in ssl3_get_client_hello().
2468 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
2469 to prevent ssl3_read_internal() from incorrectly assuming that
2470 ssl3_read_bytes() found application data while handshake
2471 processing was enabled when in fact s->s3->in_read_app_data was
2472 merely automatically cleared during the initial handshake.
2473 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
2475 *) Fix object definitions for Private and Enterprise: they were not
2476 recognized in their shortname (=lowercase) representation. Extend
2477 obj_dat.pl to issue an error when using undefined keywords instead
2478 of silently ignoring the problem (Svenning Sorensen
2479 <sss@sss.dnsalias.net>).
2482 *) Fix DH_generate_parameters() so that it works for 'non-standard'
2483 generators, i.e. generators other than 2 and 5. (Previously, the
2484 code did not properly initialise the 'add' and 'rem' values to
2485 BN_generate_prime().)
2487 In the new general case, we do not insist that 'generator' is
2488 actually a primitive root: This requirement is rather pointless;
2489 a generator of the order-q subgroup is just as good, if not
2493 *) Map new X509 verification errors to alerts. Discovered and submitted by
2494 Tom Wu <tom@arcot.com>.
2497 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
2498 returning non-zero before the data has been completely received
2499 when using non-blocking I/O.
2500 [Bodo Moeller; problem pointed out by John Hughes]
2502 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
2503 [Ben Laurie, Lutz Jaenicke]
2505 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
2506 Yoram Zahavi <YoramZ@gilian.com>).
2509 *) Add information about CygWin 1.3 and on, and preserve proper
2510 configuration for the versions before that.
2511 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2513 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
2514 check whether we deal with a copy of a session and do not delete from
2515 the cache in this case. Problem reported by "Izhar Shoshani Levi"
2516 <izhar@checkpoint.com>.
2519 *) Do not store session data into the internal session cache, if it
2520 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
2521 flag is set). Proposed by Aslam <aslam@funk.com>.
2524 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
2528 *) [In 0.9.6d-engine release:]
2529 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2530 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2532 *) Add the configuration target linux-s390x.
2533 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
2535 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
2536 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
2537 variable as an indication that a ClientHello message has been
2538 received. As the flag value will be lost between multiple
2539 invocations of ssl3_accept when using non-blocking I/O, the
2540 function may not be aware that a handshake has actually taken
2541 place, thus preventing a new session from being added to the
2544 To avoid this problem, we now set s->new_session to 2 instead of
2545 using a local variable.
2546 [Lutz Jaenicke, Bodo Moeller]
2548 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
2549 if the SSL_R_LENGTH_MISMATCH error is detected.
2550 [Geoff Thorpe, Bodo Moeller]
2552 *) New 'shared_ldflag' column in Configure platform table.
2555 *) Fix EVP_CIPHER_mode macro.
2556 ["Dan S. Camper" <dan@bti.net>]
2558 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
2559 type, we must throw them away by setting rr->length to 0.
2560 [D P Chang <dpc@qualys.com>]
2562 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
2564 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
2565 <Dominikus.Scherkl@biodata.com>. (The previous implementation
2566 worked incorrectly for those cases where range = 10..._2 and
2567 3*range is two bits longer than range.)
2570 *) Only add signing time to PKCS7 structures if it is not already
2574 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
2575 OBJ_ld_ce should be OBJ_id_ce.
2576 Also some ip-pda OIDs in crypto/objects/objects.txt were
2577 incorrect (cf. RFC 3039).
2578 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
2580 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
2581 returns early because it has nothing to do.
2582 [Andy Schneider <andy.schneider@bjss.co.uk>]
2584 *) [In 0.9.6c-engine release:]
2585 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
2586 [Andy Schneider <andy.schneider@bjss.co.uk>]
2588 *) [In 0.9.6c-engine release:]
2589 Add support for Cryptographic Appliance's keyserver technology.
2590 (Use engine 'keyclient')
2591 [Cryptographic Appliances and Geoff Thorpe]
2593 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
2594 is called via tools/c89.sh because arguments have to be
2595 rearranged (all '-L' options must appear before the first object
2597 [Richard Shapiro <rshapiro@abinitio.com>]
2599 *) [In 0.9.6c-engine release:]
2600 Add support for Broadcom crypto accelerator cards, backported
2602 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
2604 *) [In 0.9.6c-engine release:]
2605 Add support for SureWare crypto accelerator cards from
2606 Baltimore Technologies. (Use engine 'sureware')
2607 [Baltimore Technologies and Mark Cox]
2609 *) [In 0.9.6c-engine release:]
2610 Add support for crypto accelerator cards from Accelerated
2611 Encryption Processing, www.aep.ie. (Use engine 'aep')
2612 [AEP Inc. and Mark Cox]
2614 *) Add a configuration entry for gcc on UnixWare.
2615 [Gary Benson <gbenson@redhat.com>]
2617 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
2618 messages are stored in a single piece (fixed-length part and
2619 variable-length part combined) and fix various bugs found on the way.
2622 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
2623 instead. BIO_gethostbyname() does not know what timeouts are
2624 appropriate, so entries would stay in cache even when they have
2626 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
2628 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
2629 faced with a pathologically small ClientHello fragment that does
2630 not contain client_version: Instead of aborting with an error,
2631 simply choose the highest available protocol version (i.e.,
2632 TLS 1.0 unless it is disabled). In practice, ClientHello
2633 messages are never sent like this, but this change gives us
2634 strictly correct behaviour at least for TLS.
2637 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
2638 never resets s->method to s->ctx->method when called from within
2639 one of the SSL handshake functions.
2640 [Bodo Moeller; problem pointed out by Niko Baric]
2642 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
2643 (sent using the client's version number) if client_version is
2644 smaller than the protocol version in use. Also change
2645 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
2646 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
2647 the client will at least see that alert.
2650 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
2654 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
2655 client receives HelloRequest while in a handshake.
2656 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
2658 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
2659 should end in 'break', not 'goto end' which circuments various
2660 cleanups done in state SSL_ST_OK. But session related stuff
2661 must be disabled for SSL_ST_OK in the case that we just sent a
2664 Also avoid some overhead by not calling ssl_init_wbio_buffer()
2665 before just sending a HelloRequest.
2666 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
2668 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
2669 reveal whether illegal block cipher padding was found or a MAC
2670 verification error occured. (Neither SSLerr() codes nor alerts
2671 are directly visible to potential attackers, but the information
2672 may leak via logfiles.)
2674 Similar changes are not required for the SSL 2.0 implementation
2675 because the number of padding bytes is sent in clear for SSL 2.0,
2676 and the extra bytes are just ignored. However ssl/s2_pkt.c
2677 failed to verify that the purported number of padding bytes is in
2681 *) Add OpenUNIX-8 support including shared libraries
2682 (Boyd Lynn Gerber <gerberb@zenez.com>).
2685 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
2686 'wristwatch attack' using huge encoding parameters (cf.
2687 James H. Manger's CRYPTO 2001 paper). Note that the
2688 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
2689 encoding parameters and hence was not vulnerable.
2692 *) BN_sqr() bug fix.
2693 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
2695 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
2696 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
2697 followed by modular reduction.
2698 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
2700 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
2701 equivalent based on BN_pseudo_rand() instead of BN_rand().
2704 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
2705 This function was broken, as the check for a new client hello message
2706 to handle SGC did not allow these large messages.
2707 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
2710 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
2713 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
2714 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
2717 *) Rework the configuration and shared library support for Tru64 Unix.
2718 The configuration part makes use of modern compiler features and
2719 still retains old compiler behavior for those that run older versions
2720 of the OS. The shared library support part includes a variant that
2721 uses the RPATH feature, and is available through the special
2722 configuration target "alpha-cc-rpath", which will never be selected
2724 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
2726 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
2727 with the same message size as in ssl3_get_certificate_request().
2728 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
2729 messages might inadvertently be reject as too long.
2730 [Petr Lampa <lampa@fee.vutbr.cz>]
2732 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
2735 *) Modified SSL library such that the verify_callback that has been set
2736 specificly for an SSL object with SSL_set_verify() is actually being
2737 used. Before the change, a verify_callback set with this function was
2738 ignored and the verify_callback() set in the SSL_CTX at the time of
2739 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
2740 to allow the necessary settings.
2743 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
2744 explicitly to NULL, as at least on Solaris 8 this seems not always to be
2745 done automatically (in contradiction to the requirements of the C
2746 standard). This made problems when used from OpenSSH.
2749 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
2750 dh->length and always used
2752 BN_rand_range(priv_key, dh->p).
2754 BN_rand_range() is not necessary for Diffie-Hellman, and this
2755 specific range makes Diffie-Hellman unnecessarily inefficient if
2756 dh->length (recommended exponent length) is much smaller than the
2757 length of dh->p. We could use BN_rand_range() if the order of
2758 the subgroup was stored in the DH structure, but we only have
2763 BN_rand(priv_key, l, ...)
2765 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
2771 RSA_eay_public_encrypt
2772 RSA_eay_private_decrypt
2773 RSA_eay_private_encrypt (signing)
2774 RSA_eay_public_decrypt (signature verification)
2776 (default implementations for RSA_public_encrypt,
2777 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
2778 always reject numbers >= n.
2781 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
2782 to synchronize access to 'locking_thread'. This is necessary on
2783 systems where access to 'locking_thread' (an 'unsigned long'
2784 variable) is not atomic.
2787 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
2788 *before* setting the 'crypto_lock_rand' flag. The previous code had
2789 a race condition if 0 is a valid thread ID.
2790 [Travis Vitek <vitek@roguewave.com>]
2792 *) Add support for shared libraries under Irix.
2793 [Albert Chin-A-Young <china@thewrittenword.com>]
2795 *) Add configuration option to build on Linux on both big-endian and
2797 [Ralf Baechle <ralf@uni-koblenz.de>]
2799 *) Add the possibility to create shared libraries on HP-UX.
2802 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
2804 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
2805 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
2806 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
2807 PRNG state recovery was possible based on the output of
2808 one PRNG request appropriately sized to gain knowledge on
2809 'md' followed by enough consecutive 1-byte PRNG requests
2810 to traverse all of 'state'.
2812 1. When updating 'md_local' (the current thread's copy of 'md')
2813 during PRNG output generation, hash all of the previous
2814 'md_local' value, not just the half used for PRNG output.
2816 2. Make the number of bytes from 'state' included into the hash
2817 independent from the number of PRNG bytes requested.
2819 The first measure alone would be sufficient to avoid
2820 Markku-Juhani's attack. (Actually it had never occurred
2821 to me that the half of 'md_local' used for chaining was the
2822 half from which PRNG output bytes were taken -- I had always
2823 assumed that the secret half would be used.) The second
2824 measure makes sure that additional data from 'state' is never
2825 mixed into 'md_local' in small portions; this heuristically
2826 further strengthens the PRNG.
2829 *) Fix crypto/bn/asm/mips3.s.
2832 *) When only the key is given to "enc", the IV is undefined. Print out
2833 an error message in this case.
2836 *) Handle special case when X509_NAME is empty in X509 printing routines.
2839 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
2840 positive and less than q.
2843 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
2844 used: it isn't thread safe and the add_lock_callback should handle
2846 [Paul Rose <Paul.Rose@bridge.com>]
2848 *) Verify that incoming data obeys the block size in
2849 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
2853 [Ulf Möller, Bodo Möller]
2855 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
2856 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
2857 when fixing the server behaviour for backwards-compatible 'client
2858 hello' messages. (Note that the attack is impractical against
2859 SSL 3.0 and TLS 1.0 anyway because length and version checking
2860 means that the probability of guessing a valid ciphertext is
2861 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
2864 Before 0.9.5, the countermeasure (hide the error by generating a
2865 random 'decryption result') did not work properly because
2866 ERR_clear_error() was missing, meaning that SSL_get_error() would
2867 detect the supposedly ignored error.
2869 Both problems are now fixed.
2872 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
2873 (previously it was 1024).
2876 *) Fix for compatibility mode trust settings: ignore trust settings
2877 unless some valid trust or reject settings are present.
2880 *) Fix for blowfish EVP: its a variable length cipher.
2883 *) Fix various bugs related to DSA S/MIME verification. Handle missing
2884 parameters in DSA public key structures and return an error in the
2885 DSA routines if parameters are absent.
2888 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
2889 in the current directory if neither $RANDFILE nor $HOME was set.
2890 RAND_file_name() in 0.9.6a returned NULL in this case. This has
2891 caused some confusion to Windows users who haven't defined $HOME.
2892 Thus RAND_file_name() is changed again: e_os.h can define a
2893 DEFAULT_HOME, which will be used if $HOME is not set.
2894 For Windows, we use "C:"; on other platforms, we still require
2895 environment variables.
2897 *) Move 'if (!initialized) RAND_poll()' into regions protected by
2898 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
2899 having multiple threads call RAND_poll() concurrently.
2902 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
2903 combination of a flag and a thread ID variable.
2904 Otherwise while one thread is in ssleay_rand_bytes (which sets the
2905 flag), *other* threads can enter ssleay_add_bytes without obeying
2906 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
2907 that they do not hold after the first thread unsets add_do_not_lock).
2910 *) Change bctest again: '-x' expressions are not available in all
2914 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
2916 *) Fix a couple of memory leaks in PKCS7_dataDecode()
2917 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
2919 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
2920 the default extension for executables, if any. Also, make the perl
2921 scripts that use symlink() to test if it really exists and use "cp"
2922 if it doesn't. All this made OpenSSL compilable and installable in
2926 *) Fix for asn1_GetSequence() for indefinite length constructed data.
2927 If SEQUENCE is length is indefinite just set c->slen to the total
2928 amount of data available.
2929 [Steve Henson, reported by shige@FreeBSD.org]
2930 [This change does not apply to 0.9.7.]
2932 *) Change bctest to avoid here-documents inside command substitution
2933 (workaround for FreeBSD /bin/sh bug).
2934 For compatibility with Ultrix, avoid shell functions (introduced
2935 in the bctest version that searches along $PATH).
2938 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
2939 with des_encrypt() defined on some operating systems, like Solaris
2943 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
2944 On the Importance of Eliminating Errors in Cryptographic
2945 Computations, J. Cryptology 14 (2001) 2, 101-119,
2946 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
2949 *) MIPS assembler BIGNUM division bug fix.
2952 *) Disabled incorrect Alpha assembler code.
2955 *) Fix PKCS#7 decode routines so they correctly update the length
2956 after reading an EOC for the EXPLICIT tag.
2958 [This change does not apply to 0.9.7.]
2960 *) Fix bug in PKCS#12 key generation routines. This was triggered
2961 if a 3DES key was generated with a 0 initial byte. Include
2962 PKCS12_BROKEN_KEYGEN compilation option to retain the old
2963 (but broken) behaviour.
2966 *) Enhance bctest to search for a working bc along $PATH and print
2968 [Tim Rice <tim@multitalents.net> via Richard Levitte]
2970 *) Fix memory leaks in err.c: free err_data string if necessary;
2971 don't write to the wrong index in ERR_set_error_data.
2974 *) Implement ssl23_peek (analogous to ssl23_read), which previously
2978 *) Replace rdtsc with _emit statements for VC++ version 5.
2979 [Jeremy Cooper <jeremy@baymoo.org>]
2981 *) Make it possible to reuse SSLv2 sessions.
2984 *) In copy_email() check for >= 0 as a return value for
2985 X509_NAME_get_index_by_NID() since 0 is a valid index.
2986 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
2988 *) Avoid coredump with unsupported or invalid public keys by checking if
2989 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
2990 PKCS7_verify() fails with non detached data.
2993 *) Don't use getenv in library functions when run as setuid/setgid.
2994 New function OPENSSL_issetugid().
2997 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
2998 due to incorrect handling of multi-threading:
3000 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
3002 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
3004 3. Count how many times MemCheck_off() has been called so that
3005 nested use can be treated correctly. This also avoids
3006 inband-signalling in the previous code (which relied on the
3007 assumption that thread ID 0 is impossible).
3010 *) Add "-rand" option also to s_client and s_server.
3013 *) Fix CPU detection on Irix 6.x.
3014 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
3015 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3017 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
3020 [This change does not apply to 0.9.7.]
3022 *) Use the cached encoding of an X509_NAME structure rather than
3023 copying it. This is apparently the reason for the libsafe "errors"
3024 but the code is actually correct.
3027 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
3028 Bleichenbacher's DSA attack.
3029 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
3030 to be set and top=0 forces the highest bit to be set; top=-1 is new
3031 and leaves the highest bit random.
3032 [Ulf Moeller, Bodo Moeller]
3034 *) In the NCONF_...-based implementations for CONF_... queries
3035 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
3036 a temporary CONF structure with the data component set to NULL
3037 (which gives segmentation faults in lh_retrieve).
3038 Instead, use NULL for the CONF pointer in CONF_get_string and
3039 CONF_get_number (which may use environment variables) and directly
3040 return NULL from CONF_get_section.
3043 *) Fix potential buffer overrun for EBCDIC.
3046 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
3047 keyUsage if basicConstraints absent for a CA.
3050 *) Make SMIME_write_PKCS7() write mail header values with a format that
3051 is more generally accepted (no spaces before the semicolon), since
3052 some programs can't parse those values properly otherwise. Also make
3053 sure BIO's that break lines after each write do not create invalid
3057 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
3058 macros previously used would not encode an empty SEQUENCE OF
3059 and break the signature.
3061 [This change does not apply to 0.9.7.]
3063 *) Zero the premaster secret after deriving the master secret in
3067 *) Add some EVP_add_digest_alias registrations (as found in
3068 OpenSSL_add_all_digests()) to SSL_library_init()
3069 aka OpenSSL_add_ssl_algorithms(). This provides improved
3070 compatibility with peers using X.509 certificates
3071 with unconventional AlgorithmIdentifier OIDs.
3074 *) Fix for Irix with NO_ASM.
3075 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3077 *) ./config script fixes.
3078 [Ulf Moeller, Richard Levitte]
3080 *) Fix 'openssl passwd -1'.
3083 *) Change PKCS12_key_gen_asc() so it can cope with non null
3084 terminated strings whose length is passed in the passlen
3085 parameter, for example from PEM callbacks. This was done
3086 by adding an extra length parameter to asc2uni().
3087 [Steve Henson, reported by <oddissey@samsung.co.kr>]
3089 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
3090 call failed, free the DSA structure.
3093 *) Fix to uni2asc() to cope with zero length Unicode strings.
3094 These are present in some PKCS#12 files.
3097 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
3098 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
3099 when writing a 32767 byte record.
3100 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
3102 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
3103 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
3105 (RSA objects have a reference count access to which is protected
3106 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
3107 so they are meant to be shared between threads.)
3108 [Bodo Moeller, Geoff Thorpe; original patch submitted by
3109 "Reddie, Steven" <Steven.Reddie@ca.com>]
3111 *) Fix a deadlock in CRYPTO_mem_leaks().
3114 *) Use better test patterns in bntest.
3117 *) rand_win.c fix for Borland C.
3120 *) BN_rshift bugfix for n == 0.
3123 *) Add a 'bctest' script that checks for some known 'bc' bugs
3124 so that 'make test' does not abort just because 'bc' is broken.
3127 *) Store verify_result within SSL_SESSION also for client side to
3128 avoid potential security hole. (Re-used sessions on the client side
3129 always resulted in verify_result==X509_V_OK, not using the original
3130 result of the server certificate verification.)
3133 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
3134 SSL3_RT_APPLICATION_DATA, return 0.
3135 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
3139 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
3140 releases, have been re-implemented by renaming the previous
3141 implementations of ssl2_read and ssl3_read to ssl2_read_internal
3142 and ssl3_read_internal, respectively, and adding 'peek' parameters
3143 to them. The new ssl[23]_{read,peek} functions are calls to
3144 ssl[23]_read_internal with the 'peek' flag set appropriately.
3145 A 'peek' parameter has also been added to ssl3_read_bytes, which
3146 does the actual work for ssl3_read_internal.
3149 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
3150 the method-specific "init()" handler. Also clean up ex_data after
3151 calling the method-specific "finish()" handler. Previously, this was
3152 happening the other way round.
3155 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
3156 The previous value, 12, was not always sufficient for BN_mod_exp().
3159 *) Make sure that shared libraries get the internal name engine with
3160 the full version number and not just 0. This should mark the
3161 shared libraries as not backward compatible. Of course, this should
3162 be changed again when we can guarantee backward binary compatibility.
3165 *) Fix typo in get_cert_by_subject() in by_dir.c
3166 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
3168 *) Rework the system to generate shared libraries:
3170 - Make note of the expected extension for the shared libraries and
3171 if there is a need for symbolic links from for example libcrypto.so.0
3172 to libcrypto.so.0.9.7. There is extended info in Configure for
3175 - Make as few rebuilds of the shared libraries as possible.
3177 - Still avoid linking the OpenSSL programs with the shared libraries.
3179 - When installing, install the shared libraries separately from the
3183 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
3185 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
3186 and not in SSL_clear because the latter is also used by the
3187 accept/connect functions; previously, the settings made by
3188 SSL_set_read_ahead would be lost during the handshake.
3189 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
3191 *) Correct util/mkdef.pl to be selective about disabled algorithms.
3192 Previously, it would create entries for disableed algorithms no
3196 *) Added several new manual pages for SSL_* function.
3199 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
3201 *) In ssl23_get_client_hello, generate an error message when faced
3202 with an initial SSL 3.0/TLS record that is too small to contain the
3203 first two bytes of the ClientHello message, i.e. client_version.
3204 (Note that this is a pathologic case that probably has never happened
3205 in real life.) The previous approach was to use the version number
3206 from the record header as a substitute; but our protocol choice
3207 should not depend on that one because it is not authenticated
3208 by the Finished messages.
3211 *) More robust randomness gathering functions for Windows.
3212 [Jeffrey Altman <jaltman@columbia.edu>]
3214 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
3215 not set then we don't setup the error code for issuer check errors
3216 to avoid possibly overwriting other errors which the callback does
3217 handle. If an application does set the flag then we assume it knows
3218 what it is doing and can handle the new informational codes
3222 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
3223 a general "ANY" type, as such it should be able to decode anything
3224 including tagged types. However it didn't check the class so it would
3225 wrongly interpret tagged types in the same way as their universal
3226 counterpart and unknown types were just rejected. Changed so that the
3227 tagged and unknown types are handled in the same way as a SEQUENCE:
3228 that is the encoding is stored intact. There is also a new type
3229 "V_ASN1_OTHER" which is used when the class is not universal, in this
3230 case we have no idea what the actual type is so we just lump them all
3234 *) On VMS, stdout may very well lead to a file that is written to
3235 in a record-oriented fashion. That means that every write() will
3236 write a separate record, which will be read separately by the
3237 programs trying to read from it. This can be very confusing.
3239 The solution is to put a BIO filter in the way that will buffer
3240 text until a linefeed is reached, and then write everything a
3241 line at a time, so every record written will be an actual line,
3242 not chunks of lines and not (usually doesn't happen, but I've
3243 seen it once) several lines in one record. BIO_f_linebuffer() is
3246 Currently, it's a VMS-only method, because that's where it has
3247 been tested well enough.
3250 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
3251 it can return incorrect results.
3252 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
3253 but it was in 0.9.6-beta[12].)
3256 *) Disable the check for content being present when verifying detached
3257 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
3258 include zero length content when signing messages.
3261 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
3262 BIO_ctrl (for BIO pairs).
3265 *) Add DSO method for VMS.
3268 *) Bug fix: Montgomery multiplication could produce results with the
3272 *) Add RPM specification openssl.spec and modify it to build three
3273 packages. The default package contains applications, application
3274 documentation and run-time libraries. The devel package contains
3275 include files, static libraries and function documentation. The
3276 doc package contains the contents of the doc directory. The original
3277 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
3280 *) Add a large number of documentation files for many SSL routines.
3281 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3283 *) Add a configuration entry for Sony News 4.
3284 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
3286 *) Don't set the two most significant bits to one when generating a
3287 random number < q in the DSA library.
3290 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
3291 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
3292 the underlying transport is blocking) if a handshake took place.
3293 (The default behaviour is needed by applications such as s_client
3294 and s_server that use select() to determine when to use SSL_read;
3295 but for applications that know in advance when to expect data, it
3296 just makes things more complicated.)
3299 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
3303 *) Add a few more EBCDIC conditionals that make `req' and `x509'
3304 work better on such systems.
3305 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3307 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
3308 Update PKCS12_parse() so it copies the friendlyName and the
3309 keyid to the certificates aux info.
3312 *) Fix bug in PKCS7_verify() which caused an infinite loop
3313 if there was more than one signature.
3314 [Sven Uszpelkat <su@celocom.de>]
3316 *) Major change in util/mkdef.pl to include extra information
3317 about each symbol, as well as presentig variables as well
3318 as functions. This change means that there's n more need
3319 to rebuild the .num files when some algorithms are excluded.
3322 *) Allow the verify time to be set by an application,
3323 rather than always using the current time.
3326 *) Phase 2 verify code reorganisation. The certificate
3327 verify code now looks up an issuer certificate by a
3328 number of criteria: subject name, authority key id
3329 and key usage. It also verifies self signed certificates
3330 by the same criteria. The main comparison function is
3331 X509_check_issued() which performs these checks.
3333 Lot of changes were necessary in order to support this
3334 without completely rewriting the lookup code.
3336 Authority and subject key identifier are now cached.
3338 The LHASH 'certs' is X509_STORE has now been replaced
3339 by a STACK_OF(X509_OBJECT). This is mainly because an
3340 LHASH can't store or retrieve multiple objects with
3341 the same hash value.
3343 As a result various functions (which were all internal
3344 use only) have changed to handle the new X509_STORE
3345 structure. This will break anything that messed round
3346 with X509_STORE internally.
3348 The functions X509_STORE_add_cert() now checks for an
3349 exact match, rather than just subject name.
3351 The X509_STORE API doesn't directly support the retrieval
3352 of multiple certificates matching a given criteria, however
3353 this can be worked round by performing a lookup first
3354 (which will fill the cache with candidate certificates)
3355 and then examining the cache for matches. This is probably
3356 the best we can do without throwing out X509_LOOKUP
3357 entirely (maybe later...).
3359 The X509_VERIFY_CTX structure has been enhanced considerably.
3361 All certificate lookup operations now go via a get_issuer()
3362 callback. Although this currently uses an X509_STORE it
3363 can be replaced by custom lookups. This is a simple way
3364 to bypass the X509_STORE hackery necessary to make this
3365 work and makes it possible to use more efficient techniques
3366 in future. A very simple version which uses a simple
3367 STACK for its trusted certificate store is also provided
3368 using X509_STORE_CTX_trusted_stack().
3370 The verify_cb() and verify() callbacks now have equivalents
3371 in the X509_STORE_CTX structure.
3373 X509_STORE_CTX also has a 'flags' field which can be used
3374 to customise the verify behaviour.
3377 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
3378 excludes S/MIME capabilities.
3381 *) When a certificate request is read in keep a copy of the
3382 original encoding of the signed data and use it when outputing
3383 again. Signatures then use the original encoding rather than
3384 a decoded, encoded version which may cause problems if the
3385 request is improperly encoded.
3388 *) For consistency with other BIO_puts implementations, call
3389 buffer_write(b, ...) directly in buffer_puts instead of calling
3392 In BIO_puts, increment b->num_write as in BIO_write.
3393 [Peter.Sylvester@EdelWeb.fr]
3395 *) Fix BN_mul_word for the case where the word is 0. (We have to use
3396 BN_zero, we may not return a BIGNUM with an array consisting of
3400 *) Avoid calling abort() from within the library when problems are
3401 detected, except if preprocessor symbols have been defined
3402 (such as REF_CHECK, BN_DEBUG etc.).
3405 *) New openssl application 'rsautl'. This utility can be
3406 used for low level RSA operations. DER public key
3407 BIO/fp routines also added.
3410 *) New Configure entry and patches for compiling on QNX 4.
3411 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
3413 *) A demo state-machine implementation was sponsored by
3414 Nuron (http://www.nuron.com/) and is now available in
3415 demos/state_machine.
3418 *) New options added to the 'dgst' utility for signature
3419 generation and verification.
3422 *) Unrecognized PKCS#7 content types are now handled via a
3423 catch all ASN1_TYPE structure. This allows unsupported
3424 types to be stored as a "blob" and an application can
3425 encode and decode it manually.
3428 *) Fix various signed/unsigned issues to make a_strex.c
3430 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
3432 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
3433 length if passed a buffer. ASN1_INTEGER_to_BN failed
3434 if passed a NULL BN and its argument was negative.
3435 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
3437 *) Modification to PKCS#7 encoding routines to output definite
3438 length encoding. Since currently the whole structures are in
3439 memory there's not real point in using indefinite length
3440 constructed encoding. However if OpenSSL is compiled with
3441 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
3444 *) Added BIO_vprintf() and BIO_vsnprintf().
3447 *) Added more prefixes to parse for in the the strings written
3448 through a logging bio, to cover all the levels that are available
3449 through syslog. The prefixes are now:
3451 PANIC, EMERG, EMR => LOG_EMERG
3452 ALERT, ALR => LOG_ALERT
3453 CRIT, CRI => LOG_CRIT
3454 ERROR, ERR => LOG_ERR
3455 WARNING, WARN, WAR => LOG_WARNING
3456 NOTICE, NOTE, NOT => LOG_NOTICE
3457 INFO, INF => LOG_INFO
3458 DEBUG, DBG => LOG_DEBUG
3460 and as before, if none of those prefixes are present at the
3461 beginning of the string, LOG_ERR is chosen.
3463 On Win32, the LOG_* levels are mapped according to this:
3465 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
3466 LOG_WARNING => EVENTLOG_WARNING_TYPE
3467 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
3471 *) Made it possible to reconfigure with just the configuration
3472 argument "reconf" or "reconfigure". The command line arguments
3473 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
3474 and are retrieved from there when reconfiguring.
3478 [Assar Westerlund <assar@sics.se>, Richard Levitte]
3480 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
3483 *) The obj_dat.pl script was messing up the sorting of object
3484 names. The reason was that it compared the quoted version
3485 of strings as a result "OCSP" > "OCSP Signing" because
3486 " > SPACE. Changed script to store unquoted versions of
3487 names and add quotes on output. It was also omitting some
3488 names from the lookup table if they were given a default
3489 value (that is if SN is missing it is given the same
3490 value as LN and vice versa), these are now added on the
3491 grounds that if an object has a name we should be able to
3492 look it up. Finally added warning output when duplicate
3493 short or long names are found.
3496 *) Changes needed for Tandem NSK.
3497 [Scott Uroff <scott@xypro.com>]
3499 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
3500 RSA_padding_check_SSLv23(), special padding was never detected
3501 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
3502 version rollback attacks was not effective.
3504 In s23_clnt.c, don't use special rollback-attack detection padding
3505 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
3506 client; similarly, in s23_srvr.c, don't do the rollback check if
3507 SSL 2.0 is the only protocol enabled in the server.
3510 *) Make it possible to get hexdumps of unprintable data with 'openssl
3511 asn1parse'. By implication, the functions ASN1_parse_dump() and
3512 BIO_dump_indent() are added.
3515 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
3516 these print out strings and name structures based on various
3517 flags including RFC2253 support and proper handling of
3518 multibyte characters. Added options to the 'x509' utility
3519 to allow the various flags to be set.
3522 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
3523 Also change the functions X509_cmp_current_time() and
3524 X509_gmtime_adj() work with an ASN1_TIME structure,
3525 this will enable certificates using GeneralizedTime in validity
3526 dates to be checked.
3529 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
3530 negative public key encodings) on by default,
3531 NO_NEG_PUBKEY_BUG can be set to disable it.
3534 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
3535 content octets. An i2c_ASN1_OBJECT is unnecessary because
3536 the encoding can be trivially obtained from the structure.
3539 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
3540 not read locks (CRYPTO_r_[un]lock).
3543 *) A first attempt at creating official support for shared
3544 libraries through configuration. I've kept it so the
3545 default is static libraries only, and the OpenSSL programs
3546 are always statically linked for now, but there are
3547 preparations for dynamic linking in place.
3548 This has been tested on Linux and Tru64.
3551 *) Randomness polling function for Win9x, as described in:
3552 Peter Gutmann, Software Generation of Practically Strong
3556 *) Fix so PRNG is seeded in req if using an already existing
3560 *) New options to smime application. -inform and -outform
3561 allow alternative formats for the S/MIME message including
3562 PEM and DER. The -content option allows the content to be
3563 specified separately. This should allow things like Netscape
3564 form signing output easier to verify.
3567 *) Fix the ASN1 encoding of tags using the 'long form'.
3570 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
3571 STRING types. These convert content octets to and from the
3572 underlying type. The actual tag and length octets are
3573 already assumed to have been read in and checked. These
3574 are needed because all other string types have virtually
3575 identical handling apart from the tag. By having versions
3576 of the ASN1 functions that just operate on content octets
3577 IMPLICIT tagging can be handled properly. It also allows
3578 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
3579 and ASN1_INTEGER are identical apart from the tag.
3582 *) Change the handling of OID objects as follows:
3584 - New object identifiers are inserted in objects.txt, following
3585 the syntax given in objects.README.
3586 - objects.pl is used to process obj_mac.num and create a new
3588 - obj_dat.pl is used to create a new obj_dat.h, using the data in
3591 This is currently kind of a hack, and the perl code in objects.pl
3592 isn't very elegant, but it works as I intended. The simplest way
3593 to check that it worked correctly is to look in obj_dat.h and
3594 check the array nid_objs and make sure the objects haven't moved
3595 around (this is important!). Additions are OK, as well as
3596 consistent name changes.
3599 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
3602 *) Addition of the command line parameter '-rand file' to 'openssl req'.
3603 The given file adds to whatever has already been seeded into the
3604 random pool through the RANDFILE configuration file option or
3605 environment variable, or the default random state file.
3608 *) mkstack.pl now sorts each macro group into lexical order.
3609 Previously the output order depended on the order the files
3610 appeared in the directory, resulting in needless rewriting
3614 *) Patches to make OpenSSL compile under Win32 again. Mostly
3615 work arounds for the VC++ problem that it treats func() as
3616 func(void). Also stripped out the parts of mkdef.pl that
3617 added extra typesafe functions: these no longer exist.
3620 *) Reorganisation of the stack code. The macros are now all
3621 collected in safestack.h . Each macro is defined in terms of
3622 a "stack macro" of the form SKM_<name>(type, a, b). The
3623 DEBUG_SAFESTACK is now handled in terms of function casts,
3624 this has the advantage of retaining type safety without the
3625 use of additional functions. If DEBUG_SAFESTACK is not defined
3626 then the non typesafe macros are used instead. Also modified the
3627 mkstack.pl script to handle the new form. Needs testing to see
3628 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
3629 the default if no major problems. Similar behaviour for ASN1_SET_OF
3630 and PKCS12_STACK_OF.
3633 *) When some versions of IIS use the 'NET' form of private key the
3634 key derivation algorithm is different. Normally MD5(password) is
3635 used as a 128 bit RC4 key. In the modified case
3636 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
3637 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
3638 as the old Netscape_RSA functions except they have an additional
3639 'sgckey' parameter which uses the modified algorithm. Also added
3640 an -sgckey command line option to the rsa utility. Thanks to
3641 Adrian Peck <bertie@ncipher.com> for posting details of the modified
3642 algorithm to openssl-dev.
3645 *) The evp_local.h macros were using 'c.##kname' which resulted in
3646 invalid expansion on some systems (SCO 5.0.5 for example).
3647 Corrected to 'c.kname'.
3648 [Phillip Porch <root@theporch.com>]
3650 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
3651 a STACK of email addresses from a certificate or request, these look
3652 in the subject name and the subject alternative name extensions and
3653 omit any duplicate addresses.
3656 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
3657 This makes DSA verification about 2 % faster.
3660 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
3661 (meaning that now 2^5 values will be precomputed, which is only 4 KB
3662 plus overhead for 1024 bit moduli).
3663 This makes exponentiations about 0.5 % faster for 1024 bit
3664 exponents (as measured by "openssl speed rsa2048").
3667 *) Rename memory handling macros to avoid conflicts with other
3669 Malloc => OPENSSL_malloc
3670 Malloc_locked => OPENSSL_malloc_locked
3671 Realloc => OPENSSL_realloc
3672 Free => OPENSSL_free
3675 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
3676 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
3679 *) CygWin32 support.
3680 [John Jarvie <jjarvie@newsguy.com>]
3682 *) The type-safe stack code has been rejigged. It is now only compiled
3683 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
3684 by default all type-specific stack functions are "#define"d back to
3685 standard stack functions. This results in more streamlined output
3686 but retains the type-safety checking possibilities of the original
3690 *) The STACK code has been cleaned up, and certain type declarations
3691 that didn't make a lot of sense have been brought in line. This has
3692 also involved a cleanup of sorts in safestack.h to more correctly
3693 map type-safe stack functions onto their plain stack counterparts.
3694 This work has also resulted in a variety of "const"ifications of
3695 lots of the code, especially "_cmp" operations which should normally
3696 be prototyped with "const" parameters anyway.
3699 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
3700 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
3701 (The PRNG state consists of two parts, the large pool 'state' and 'md',
3702 where all of 'md' is used each time the PRNG is used, but 'state'
3703 is used only indexed by a cyclic counter. As entropy may not be
3704 well distributed from the beginning, 'md' is important as a
3705 chaining variable. However, the output function chains only half
3706 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
3707 all of 'md', and seeding with STATE_SIZE dummy bytes will result
3708 in all of 'state' being rewritten, with the new values depending
3709 on virtually all of 'md'. This overcomes the 80 bit limitation.)
3712 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
3713 the handshake is continued after ssl_verify_cert_chain();
3714 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
3715 can lead to 'unexplainable' connection aborts later.
3716 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
3718 *) Major EVP API cipher revision.
3719 Add hooks for extra EVP features. This allows various cipher
3720 parameters to be set in the EVP interface. Support added for variable
3721 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
3722 setting of RC2 and RC5 parameters.
3724 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
3727 Remove lots of duplicated code from the EVP library. For example *every*
3728 cipher init() function handles the 'iv' in the same way according to the
3729 cipher mode. They also all do nothing if the 'key' parameter is NULL and
3730 for CFB and OFB modes they zero ctx->num.
3732 New functionality allows removal of S/MIME code RC2 hack.
3734 Most of the routines have the same form and so can be declared in terms
3737 By shifting this to the top level EVP_CipherInit() it can be removed from
3738 all individual ciphers. If the cipher wants to handle IVs or keys
3739 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
3742 Change lots of functions like EVP_EncryptUpdate() to now return a
3743 value: although software versions of the algorithms cannot fail
3744 any installed hardware versions can.
3747 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
3748 this option is set, tolerate broken clients that send the negotiated
3749 protocol version number instead of the requested protocol version
3753 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
3754 i.e. non-zero for export ciphersuites, zero otherwise.
3755 Previous versions had this flag inverted, inconsistent with
3756 rsa_tmp_cb (..._TMP_RSA_CB).
3757 [Bodo Moeller; problem reported by Amit Chopra]
3759 *) Add missing DSA library text string. Work around for some IIS
3760 key files with invalid SEQUENCE encoding.
3763 *) Add a document (doc/standards.txt) that list all kinds of standards
3764 and so on that are implemented in OpenSSL.
3767 *) Enhance c_rehash script. Old version would mishandle certificates
3768 with the same subject name hash and wouldn't handle CRLs at all.
3769 Added -fingerprint option to crl utility, to support new c_rehash
3773 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
3776 *) Fix for SSL server purpose checking. Server checking was
3777 rejecting certificates which had extended key usage present
3778 but no ssl client purpose.
3779 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
3781 *) Make PKCS#12 code work with no password. The PKCS#12 spec
3782 is a little unclear about how a blank password is handled.
3783 Since the password in encoded as a BMPString with terminating
3784 double NULL a zero length password would end up as just the
3785 double NULL. However no password at all is different and is
3786 handled differently in the PKCS#12 key generation code. NS
3787 treats a blank password as zero length. MSIE treats it as no
3788 password on export: but it will try both on import. We now do
3789 the same: PKCS12_parse() tries zero length and no password if
3790 the password is set to "" or NULL (NULL is now a valid password:
3791 it wasn't before) as does the pkcs12 application.
3794 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
3795 perror when PEM_read_bio_X509_REQ fails, the error message must
3796 be obtained from the error queue.
3799 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
3800 it in ERR_remove_state if appropriate, and change ERR_get_state
3801 accordingly to avoid race conditions (this is necessary because
3802 thread_hash is no longer constant once set).
3805 *) Bugfix for linux-elf makefile.one.
3808 *) RSA_get_default_method() will now cause a default
3809 RSA_METHOD to be chosen if one doesn't exist already.
3810 Previously this was only set during a call to RSA_new()
3811 or RSA_new_method(NULL) meaning it was possible for
3812 RSA_get_default_method() to return NULL.
3815 *) Added native name translation to the existing DSO code
3816 that will convert (if the flag to do so is set) filenames
3817 that are sufficiently small and have no path information
3818 into a canonical native form. Eg. "blah" converted to
3819 "libblah.so" or "blah.dll" etc.
3822 *) New function ERR_error_string_n(e, buf, len) which is like
3823 ERR_error_string(e, buf), but writes at most 'len' bytes
3824 including the 0 terminator. For ERR_error_string_n, 'buf'
3826 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
3828 *) CONF library reworked to become more general. A new CONF
3829 configuration file reader "class" is implemented as well as a
3830 new functions (NCONF_*, for "New CONF") to handle it. The now
3831 old CONF_* functions are still there, but are reimplemented to
3832 work in terms of the new functions. Also, a set of functions
3833 to handle the internal storage of the configuration data is
3834 provided to make it easier to write new configuration file
3835 reader "classes" (I can definitely see something reading a
3836 configuration file in XML format, for example), called _CONF_*,
3837 or "the configuration storage API"...
3839 The new configuration file reading functions are:
3841 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
3842 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
3844 NCONF_default, NCONF_WIN32
3846 NCONF_dump_fp, NCONF_dump_bio
3848 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
3849 NCONF_new creates a new CONF object. This works in the same way
3850 as other interfaces in OpenSSL, like the BIO interface.
3851 NCONF_dump_* dump the internal storage of the configuration file,
3852 which is useful for debugging. All other functions take the same
3853 arguments as the old CONF_* functions wth the exception of the
3854 first that must be a `CONF *' instead of a `LHASH *'.
3856 To make it easer to use the new classes with the old CONF_* functions,
3857 the function CONF_set_default_method is provided.
3860 *) Add '-tls1' option to 'openssl ciphers', which was already
3861 mentioned in the documentation but had not been implemented.
3862 (This option is not yet really useful because even the additional
3863 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
3866 *) Initial DSO code added into libcrypto for letting OpenSSL (and
3867 OpenSSL-based applications) load shared libraries and bind to
3868 them in a portable way.
3869 [Geoff Thorpe, with contributions from Richard Levitte]
3871 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
3873 *) Make sure _lrotl and _lrotr are only used with MSVC.
3875 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
3876 (the default implementation of RAND_status).
3878 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
3879 to '-clrext' (= clear extensions), as intended and documented.
3880 [Bodo Moeller; inconsistency pointed out by Michael Attili
3883 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
3884 was larger than the MD block size.
3885 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
3887 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
3888 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
3889 using the passed key: if the passed key was a private key the result
3890 of X509_print(), for example, would be to print out all the private key
3894 *) des_quad_cksum() byte order bug fix.
3895 [Ulf Möller, using the problem description in krb4-0.9.7, where
3896 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
3898 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
3900 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
3902 *) For easily testing in shell scripts whether some command
3903 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
3904 returns with exit code 0 iff no command of the given name is available.
3905 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
3906 the output goes to stdout and nothing is printed to stderr.
3907 Additional arguments are always ignored.
3909 Since for each cipher there is a command of the same name,
3910 the 'no-cipher' compilation switches can be tested this way.
3912 ('openssl no-XXX' is not able to detect pseudo-commands such
3913 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
3916 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
3919 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
3920 is set; it will be thrown away anyway because each handshake creates
3922 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
3923 to parameters -- in previous versions (since OpenSSL 0.9.3) the
3924 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
3925 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
3928 *) New s_client option -ign_eof: EOF at stdin is ignored, and
3929 'Q' and 'R' lose their special meanings (quit/renegotiate).
3930 This is part of what -quiet does; unlike -quiet, -ign_eof
3931 does not suppress any output.
3934 *) Add compatibility options to the purpose and trust code. The
3935 purpose X509_PURPOSE_ANY is "any purpose" which automatically
3936 accepts a certificate or CA, this was the previous behaviour,
3937 with all the associated security issues.
3939 X509_TRUST_COMPAT is the old trust behaviour: only and
3940 automatically trust self signed roots in certificate store. A
3941 new trust setting X509_TRUST_DEFAULT is used to specify that
3942 a purpose has no associated trust setting and it should instead
3943 use the value in the default purpose.
3946 *) Fix the PKCS#8 DSA private key code so it decodes keys again
3947 and fix a memory leak.
3950 *) In util/mkerr.pl (which implements 'make errors'), preserve
3951 reason strings from the previous version of the .c file, as
3952 the default to have only downcase letters (and digits) in
3953 automatically generated reasons codes is not always appropriate.
3956 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
3957 using strerror. Previously, ERR_reason_error_string() returned
3958 library names as reason strings for SYSerr; but SYSerr is a special
3959 case where small numbers are errno values, not library numbers.
3962 *) Add '-dsaparam' option to 'openssl dhparam' application. This
3963 converts DSA parameters into DH parameters. (When creating parameters,
3964 DSA_generate_parameters is used.)
3967 *) Include 'length' (recommended exponent length) in C code generated
3968 by 'openssl dhparam -C'.
3971 *) The second argument to set_label in perlasm was already being used
3972 so couldn't be used as a "file scope" flag. Moved to third argument
3976 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
3977 instead of RAND_bytes for encryption IVs and salts.
3980 *) Include RAND_status() into RAND_METHOD instead of implementing
3981 it only for md_rand.c Otherwise replacing the PRNG by calling
3982 RAND_set_rand_method would be impossible.
3985 *) Don't let DSA_generate_key() enter an infinite loop if the random
3986 number generation fails.
3989 *) New 'rand' application for creating pseudo-random output.
3992 *) Added configuration support for Linux/IA64
3993 [Rolf Haberrecker <rolf@suse.de>]
3995 *) Assembler module support for Mingw32.
3998 *) Shared library support for HPUX (in shlib/).
3999 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
4001 *) Shared library support for Solaris gcc.
4002 [Lutz Behnke <behnke@trustcenter.de>]
4004 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
4006 *) PKCS7_encrypt() was adding text MIME headers twice because they
4007 were added manually and by SMIME_crlf_copy().
4010 *) In bntest.c don't call BN_rand with zero bits argument.
4011 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
4013 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
4014 case was implemented. This caused BN_div_recp() to fail occasionally.
4017 *) Add an optional second argument to the set_label() in the perl
4018 assembly language builder. If this argument exists and is set
4019 to 1 it signals that the assembler should use a symbol whose
4020 scope is the entire file, not just the current function. This
4021 is needed with MASM which uses the format label:: for this scope.
4022 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
4024 *) Change the ASN1 types so they are typedefs by default. Before
4025 almost all types were #define'd to ASN1_STRING which was causing
4026 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
4030 *) Change names of new functions to the new get1/get0 naming
4031 convention: After 'get1', the caller owns a reference count
4032 and has to call ..._free; 'get0' returns a pointer to some
4033 data structure without incrementing reference counters.
4034 (Some of the existing 'get' functions increment a reference
4035 counter, some don't.)
4036 Similarly, 'set1' and 'add1' functions increase reference
4037 counters or duplicate objects.
4040 *) Allow for the possibility of temp RSA key generation failure:
4041 the code used to assume it always worked and crashed on failure.
4044 *) Fix potential buffer overrun problem in BIO_printf().
4045 [Ulf Möller, using public domain code by Patrick Powell; problem
4046 pointed out by David Sacerdote <das33@cornell.edu>]
4048 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
4049 RAND_egd() and RAND_status(). In the command line application,
4050 the EGD socket can be specified like a seed file using RANDFILE
4054 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
4055 Some CAs (e.g. Verisign) distribute certificates in this form.
4058 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
4059 list to exclude them. This means that no special compilation option
4060 is needed to use anonymous DH: it just needs to be included in the
4064 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
4065 EVP_MD_type. The old functionality is available in a new macro called
4066 EVP_MD_md(). Change code that uses it and update docs.
4069 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
4070 where the 'void *' argument is replaced by a function pointer argument.
4071 Previously 'void *' was abused to point to functions, which works on
4072 many platforms, but is not correct. As these functions are usually
4073 called by macros defined in OpenSSL header files, most source code
4074 should work without changes.
4077 *) <openssl/opensslconf.h> (which is created by Configure) now contains
4078 sections with information on -D... compiler switches used for
4079 compiling the library so that applications can see them. To enable
4080 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
4081 must be defined. E.g.,
4082 #define OPENSSL_ALGORITHM_DEFINES
4083 #include <openssl/opensslconf.h>
4084 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
4085 [Richard Levitte, Ulf and Bodo Möller]
4087 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
4091 *) Change the 'other' type in certificate aux info to a STACK_OF
4092 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
4093 the required ASN1 format: arbitrary types determined by an OID.
4096 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
4097 argument to 'req'. This is not because the function is newer or
4098 better than others it just uses the work 'NEW' in the certificate
4099 request header lines. Some software needs this.
4102 *) Reorganise password command line arguments: now passwords can be
4103 obtained from various sources. Delete the PEM_cb function and make
4104 it the default behaviour: i.e. if the callback is NULL and the
4105 usrdata argument is not NULL interpret it as a null terminated pass
4106 phrase. If usrdata and the callback are NULL then the pass phrase
4107 is prompted for as usual.
4110 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
4111 the support is automatically enabled. The resulting binaries will
4112 autodetect the card and use it if present.
4113 [Ben Laurie and Compaq Inc.]
4115 *) Work around for Netscape hang bug. This sends certificate request
4116 and server done in one record. Since this is perfectly legal in the
4117 SSL/TLS protocol it isn't a "bug" option and is on by default. See
4118 the bugs/SSLv3 entry for more info.
4121 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
4124 *) Add -rand argument to smime and pkcs12 applications and read/write
4128 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
4131 *) Add command line password options to the remaining applications.
4134 *) Bug fix for BN_div_recp() for numerators with an even number of
4138 *) More tests in bntest.c, and changed test_bn output.
4141 *) ./config recognizes MacOS X now.
4144 *) Bug fix for BN_div() when the first words of num and divsor are
4145 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
4148 *) Add support for various broken PKCS#8 formats, and command line
4149 options to produce them.
4152 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
4153 get temporary BIGNUMs from a BN_CTX.
4156 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
4160 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
4161 include a #define from the old name to the new. The original intent
4162 was that statically linked binaries could for example just call
4163 SSLeay_add_all_ciphers() to just add ciphers to the table and not
4164 link with digests. This never worked becayse SSLeay_add_all_digests()
4165 and SSLeay_add_all_ciphers() were in the same source file so calling
4166 one would link with the other. They are now in separate source files.
4169 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
4172 *) Use a less unusual form of the Miller-Rabin primality test (it used
4173 a binary algorithm for exponentiation integrated into the Miller-Rabin
4174 loop, our standard modexp algorithms are faster).
4177 *) Support for the EBCDIC character set completed.
4178 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
4180 *) Source code cleanups: use const where appropriate, eliminate casts,
4181 use void * instead of char * in lhash.
4184 *) Bugfix: ssl3_send_server_key_exchange was not restartable
4185 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
4186 this the server could overwrite ephemeral keys that the client
4190 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
4191 using 50 iterations of the Rabin-Miller test.
4193 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
4194 iterations of the Rabin-Miller test as required by the appendix
4195 to FIPS PUB 186[-1]) instead of DSA_is_prime.
4196 As BN_is_prime_fasttest includes trial division, DSA parameter
4197 generation becomes much faster.
4199 This implies a change for the callback functions in DSA_is_prime
4200 and DSA_generate_parameters: The callback function is called once
4201 for each positive witness in the Rabin-Miller test, not just
4202 occasionally in the inner loop; and the parameters to the
4203 callback function now provide an iteration count for the outer
4204 loop rather than for the current invocation of the inner loop.
4205 DSA_generate_parameters additionally can call the callback
4206 function with an 'iteration count' of -1, meaning that a
4207 candidate has passed the trial division test (when q is generated
4208 from an application-provided seed, trial division is skipped).
4211 *) New function BN_is_prime_fasttest that optionally does trial
4212 division before starting the Rabin-Miller test and has
4213 an additional BN_CTX * argument (whereas BN_is_prime always
4214 has to allocate at least one BN_CTX).
4215 'callback(1, -1, cb_arg)' is called when a number has passed the
4216 trial division stage.
4219 *) Fix for bug in CRL encoding. The validity dates weren't being handled
4223 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
4226 *) New function BN_pseudo_rand().
4229 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
4230 bignum version of BN_from_montgomery() with the working code from
4231 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
4235 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
4236 made it impossible to use the same SSL_SESSION data structure in
4237 SSL2 clients in multiple threads.
4240 *) The return value of RAND_load_file() no longer counts bytes obtained
4241 by stat(). RAND_load_file(..., -1) is new and uses the complete file
4242 to seed the PRNG (previously an explicit byte count was required).
4243 [Ulf Möller, Bodo Möller]
4245 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
4246 used (char *) instead of (void *) and had casts all over the place.
4249 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
4252 *) Retain source code compatibility for BN_prime_checks macro:
4253 BN_is_prime(..., BN_prime_checks, ...) now uses
4254 BN_prime_checks_for_size to determine the appropriate number of
4255 Rabin-Miller iterations.
4258 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
4259 DH_CHECK_P_NOT_SAFE_PRIME.
4260 (Check if this is true? OpenPGP calls them "strong".)
4263 *) Merge the functionality of "dh" and "gendh" programs into a new program
4264 "dhparam". The old programs are retained for now but will handle DH keys
4265 (instead of parameters) in future.
4268 *) Make the ciphers, s_server and s_client programs check the return values
4269 when a new cipher list is set.
4272 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
4273 ciphers. Before when the 56bit ciphers were enabled the sorting was
4276 The syntax for the cipher sorting has been extended to support sorting by
4277 cipher-strength (using the strength_bits hard coded in the tables).
4278 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
4280 Fix a bug in the cipher-command parser: when supplying a cipher command
4281 string with an "undefined" symbol (neither command nor alphanumeric
4282 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
4283 an error is flagged.
4285 Due to the strength-sorting extension, the code of the
4286 ssl_create_cipher_list() function was completely rearranged. I hope that
4287 the readability was also increased :-)
4288 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4290 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
4291 for the first serial number and places 2 in the serial number file. This
4292 avoids problems when the root CA is created with serial number zero and
4293 the first user certificate has the same issuer name and serial number
4297 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
4298 the new code. Add documentation for this stuff.
4301 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
4302 X509_*() to X509at_*() on the grounds that they don't handle X509
4303 structures and behave in an analagous way to the X509v3 functions:
4304 they shouldn't be called directly but wrapper functions should be used
4307 So we also now have some wrapper functions that call the X509at functions
4308 when passed certificate requests. (TO DO: similar things can be done with
4309 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
4310 things. Some of these need some d2i or i2d and print functionality
4311 because they handle more complex structures.)
4314 *) Add missing #ifndefs that caused missing symbols when building libssl
4315 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
4316 NO_RSA in ssl/s2*.c.
4317 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
4319 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
4320 has a return value which indicates the quality of the random data
4321 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
4322 error queue. New function RAND_pseudo_bytes() generates output that is
4323 guaranteed to be unique but not unpredictable. RAND_add is like
4324 RAND_seed, but takes an extra argument for an entropy estimate
4325 (RAND_seed always assumes full entropy).
4328 *) Do more iterations of Rabin-Miller probable prime test (specifically,
4329 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
4330 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
4331 in crypto/bn/bn_prime.c for the complete table). This guarantees a
4332 false-positive rate of at most 2^-80 for random input.
4335 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
4338 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
4339 in the 0.9.5 release), this returns the chain
4340 from an X509_CTX structure with a dup of the stack and all
4341 the X509 reference counts upped: so the stack will exist
4342 after X509_CTX_cleanup() has been called. Modify pkcs12.c
4345 Also make SSL_SESSION_print() print out the verify return
4349 *) Add manpage for the pkcs12 command. Also change the default
4350 behaviour so MAC iteration counts are used unless the new
4351 -nomaciter option is used. This improves file security and
4352 only older versions of MSIE (4.0 for example) need it.
4355 *) Honor the no-xxx Configure options when creating .DEF files.
4358 *) Add PKCS#10 attributes to field table: challengePassword,
4359 unstructuredName and unstructuredAddress. These are taken from
4360 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
4361 international characters are used.
4363 More changes to X509_ATTRIBUTE code: allow the setting of types
4364 based on strings. Remove the 'loc' parameter when adding
4365 attributes because these will be a SET OF encoding which is sorted
4369 *) Initial changes to the 'req' utility to allow request generation
4370 automation. This will allow an application to just generate a template
4371 file containing all the field values and have req construct the
4374 Initial support for X509_ATTRIBUTE handling. Stacks of these are
4375 used all over the place including certificate requests and PKCS#7
4376 structures. They are currently handled manually where necessary with
4377 some primitive wrappers for PKCS#7. The new functions behave in a
4378 manner analogous to the X509 extension functions: they allow
4379 attributes to be looked up by NID and added.
4381 Later something similar to the X509V3 code would be desirable to
4382 automatically handle the encoding, decoding and printing of the
4383 more complex types. The string types like challengePassword can
4384 be handled by the string table functions.
4386 Also modified the multi byte string table handling. Now there is
4387 a 'global mask' which masks out certain types. The table itself
4388 can use the flag STABLE_NO_MASK to ignore the mask setting: this
4389 is useful when for example there is only one permissible type
4390 (as in countryName) and using the mask might result in no valid
4394 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
4395 SSL_get_peer_finished to allow applications to obtain the latest
4396 Finished messages sent to the peer or expected from the peer,
4397 respectively. (SSL_get_peer_finished is usually the Finished message
4398 actually received from the peer, otherwise the protocol will be aborted.)
4400 As the Finished message are message digests of the complete handshake
4401 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
4402 be used for external authentication procedures when the authentication
4403 provided by SSL/TLS is not desired or is not enough.
4406 *) Enhanced support for Alpha Linux is added. Now ./config checks if
4407 the host supports BWX extension and if Compaq C is present on the
4408 $PATH. Just exploiting of the BWX extension results in 20-30%
4409 performance kick for some algorithms, e.g. DES and RC4 to mention
4410 a couple. Compaq C in turn generates ~20% faster code for MD5 and
4414 *) Add support for MS "fast SGC". This is arguably a violation of the
4415 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
4416 weak crypto and after checking the certificate is SGC a second one
4417 with strong crypto. MS SGC stops the first handshake after receiving
4418 the server certificate message and sends a second client hello. Since
4419 a server will typically do all the time consuming operations before
4420 expecting any further messages from the client (server key exchange
4421 is the most expensive) there is little difference between the two.
4423 To get OpenSSL to support MS SGC we have to permit a second client
4424 hello message after we have sent server done. In addition we have to
4425 reset the MAC if we do get this second client hello.
4428 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
4429 if a DER encoded private key is RSA or DSA traditional format. Changed
4430 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
4431 format DER encoded private key. Newer code should use PKCS#8 format which
4432 has the key type encoded in the ASN1 structure. Added DER private key
4433 support to pkcs8 application.
4436 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
4437 ciphersuites has been selected (as required by the SSL 3/TLS 1
4438 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
4439 is set, we interpret this as a request to violate the specification
4440 (the worst that can happen is a handshake failure, and 'correct'
4441 behaviour would result in a handshake failure anyway).
4444 *) In SSL_CTX_add_session, take into account that there might be multiple
4445 SSL_SESSION structures with the same session ID (e.g. when two threads
4446 concurrently obtain them from an external cache).
4447 The internal cache can handle only one SSL_SESSION with a given ID,
4448 so if there's a conflict, we now throw out the old one to achieve
4452 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
4453 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
4454 some routines that use cipher OIDs: some ciphers do not have OIDs
4455 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
4459 *) Simplify the trust setting structure and code. Now we just have
4460 two sequences of OIDs for trusted and rejected settings. These will
4461 typically have values the same as the extended key usage extension
4462 and any application specific purposes.
4464 The trust checking code now has a default behaviour: it will just
4465 check for an object with the same NID as the passed id. Functions can
4466 be provided to override either the default behaviour or the behaviour
4467 for a given id. SSL client, server and email already have functions
4468 in place for compatibility: they check the NID and also return "trusted"
4469 if the certificate is self signed.
4472 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
4473 traditional format into an EVP_PKEY structure.
4476 *) Add a password callback function PEM_cb() which either prompts for
4477 a password if usr_data is NULL or otherwise assumes it is a null
4478 terminated password. Allow passwords to be passed on command line
4479 environment or config files in a few more utilities.
4482 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
4483 keys. Add some short names for PKCS#8 PBE algorithms and allow them
4484 to be specified on the command line for the pkcs8 and pkcs12 utilities.
4485 Update documentation.
4488 *) Support for ASN1 "NULL" type. This could be handled before by using
4489 ASN1_TYPE but there wasn't any function that would try to read a NULL
4490 and produce an error if it couldn't. For compatibility we also have
4491 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
4492 don't allocate anything because they don't need to.
4495 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
4497 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
4499 *) Rebuild of the memory allocation routines used by OpenSSL code and
4500 possibly others as well. The purpose is to make an interface that
4501 provide hooks so anyone can build a separate set of allocation and
4502 deallocation routines to be used by OpenSSL, for example memory
4503 pool implementations, or something else, which was previously hard
4504 since Malloc(), Realloc() and Free() were defined as macros having
4505 the values malloc, realloc and free, respectively (except for Win32
4506 compilations). The same is provided for memory debugging code.
4507 OpenSSL already comes with functionality to find memory leaks, but
4508 this gives people a chance to debug other memory problems.
4510 With these changes, a new set of functions and macros have appeared:
4512 CRYPTO_set_mem_debug_functions() [F]
4513 CRYPTO_get_mem_debug_functions() [F]
4514 CRYPTO_dbg_set_options() [F]
4515 CRYPTO_dbg_get_options() [F]
4516 CRYPTO_malloc_debug_init() [M]
4518 The memory debug functions are NULL by default, unless the library
4519 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
4520 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
4521 gives the standard debugging functions that come with OpenSSL) or
4522 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
4523 provided by the library user) must be used. When the standard
4524 debugging functions are used, CRYPTO_dbg_set_options can be used to
4525 request additional information:
4526 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
4527 the CRYPTO_MDEBUG_xxx macro when compiling the library.
4529 Also, things like CRYPTO_set_mem_functions will always give the
4530 expected result (the new set of functions is used for allocation
4531 and deallocation) at all times, regardless of platform and compiler
4534 To finish it up, some functions that were never use in any other
4535 way than through macros have a new API and new semantic:
4538 CRYPTO_dbg_realloc()
4541 All macros of value have retained their old syntax.
4542 [Richard Levitte and Bodo Moeller]
4544 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
4545 ordering of SMIMECapabilities wasn't in "strength order" and there
4546 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
4550 *) Some ASN1 types with illegal zero length encoding (INTEGER,
4551 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
4552 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
4554 *) Merge in my S/MIME library for OpenSSL. This provides a simple
4555 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
4556 functionality to handle multipart/signed properly) and a utility
4557 called 'smime' to call all this stuff. This is based on code I
4558 originally wrote for Celo who have kindly allowed it to be
4559 included in OpenSSL.
4562 *) Add variants des_set_key_checked and des_set_key_unchecked of
4563 des_set_key (aka des_key_sched). Global variable des_check_key
4564 decides which of these is called by des_set_key; this way
4565 des_check_key behaves as it always did, but applications and
4566 the library itself, which was buggy for des_check_key == 1,
4567 have a cleaner way to pick the version they need.
4570 *) New function PKCS12_newpass() which changes the password of a
4574 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
4575 dynamic mix. In both cases the ids can be used as an index into the
4576 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
4577 functions so they accept a list of the field values and the
4578 application doesn't need to directly manipulate the X509_TRUST
4582 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
4586 *) Modify the way the V3 extension code looks up extensions. This now
4587 works in a similar way to the object code: we have some "standard"
4588 extensions in a static table which is searched with OBJ_bsearch()
4589 and the application can add dynamic ones if needed. The file
4590 crypto/x509v3/ext_dat.h now has the info: this file needs to be
4591 updated whenever a new extension is added to the core code and kept
4592 in ext_nid order. There is a simple program 'tabtest.c' which checks
4593 this. New extensions are not added too often so this file can readily
4594 be maintained manually.
4596 There are two big advantages in doing things this way. The extensions
4597 can be looked up immediately and no longer need to be "added" using
4598 X509V3_add_standard_extensions(): this function now does nothing.
4599 [Side note: I get *lots* of email saying the extension code doesn't
4600 work because people forget to call this function]
4601 Also no dynamic allocation is done unless new extensions are added:
4602 so if we don't add custom extensions there is no need to call
4603 X509V3_EXT_cleanup().
4606 *) Modify enc utility's salting as follows: make salting the default. Add a
4607 magic header, so unsalted files fail gracefully instead of just decrypting
4608 to garbage. This is because not salting is a big security hole, so people
4609 should be discouraged from doing it.
4612 *) Fixes and enhancements to the 'x509' utility. It allowed a message
4613 digest to be passed on the command line but it only used this
4614 parameter when signing a certificate. Modified so all relevant
4615 operations are affected by the digest parameter including the
4616 -fingerprint and -x509toreq options. Also -x509toreq choked if a
4617 DSA key was used because it didn't fix the digest.
4620 *) Initial certificate chain verify code. Currently tests the untrusted
4621 certificates for consistency with the verify purpose (which is set
4622 when the X509_STORE_CTX structure is set up) and checks the pathlength.
4624 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
4625 this is because it will reject chains with invalid extensions whereas
4626 every previous version of OpenSSL and SSLeay made no checks at all.
4628 Trust code: checks the root CA for the relevant trust settings. Trust
4629 settings have an initial value consistent with the verify purpose: e.g.
4630 if the verify purpose is for SSL client use it expects the CA to be
4631 trusted for SSL client use. However the default value can be changed to
4632 permit custom trust settings: one example of this would be to only trust
4633 certificates from a specific "secure" set of CAs.
4635 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
4636 which should be used for version portability: especially since the
4637 verify structure is likely to change more often now.
4639 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
4640 to set them. If not set then assume SSL clients will verify SSL servers
4643 Two new options to the verify program: -untrusted allows a set of
4644 untrusted certificates to be passed in and -purpose which sets the
4645 intended purpose of the certificate. If a purpose is set then the
4646 new chain verify code is used to check extension consistency.
4649 *) Support for the authority information access extension.
4652 *) Modify RSA and DSA PEM read routines to transparently handle
4653 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
4654 public keys in a format compatible with certificate
4655 SubjectPublicKeyInfo structures. Unfortunately there were already
4656 functions called *_PublicKey_* which used various odd formats so
4657 these are retained for compatibility: however the DSA variants were
4658 never in a public release so they have been deleted. Changed dsa/rsa
4659 utilities to handle the new format: note no releases ever handled public
4660 keys so we should be OK.
4662 The primary motivation for this change is to avoid the same fiasco
4663 that dogs private keys: there are several incompatible private key
4664 formats some of which are standard and some OpenSSL specific and
4665 require various evil hacks to allow partial transparent handling and
4666 even then it doesn't work with DER formats. Given the option anything
4667 other than PKCS#8 should be dumped: but the other formats have to
4668 stay in the name of compatibility.
4670 With public keys and the benefit of hindsight one standard format
4671 is used which works with EVP_PKEY, RSA or DSA structures: though
4672 it clearly returns an error if you try to read the wrong kind of key.
4674 Added a -pubkey option to the 'x509' utility to output the public key.
4675 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
4676 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
4677 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
4678 that do the same as the EVP_PKEY_assign_*() except they up the
4679 reference count of the added key (they don't "swallow" the
4683 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
4684 CRLs would fail if the file contained no certificates or no CRLs:
4685 added a new function to read in both types and return the number
4686 read: this means that if none are read it will be an error. The
4687 DER versions of the certificate and CRL reader would always fail
4688 because it isn't possible to mix certificates and CRLs in DER format
4689 without choking one or the other routine. Changed this to just read
4690 a certificate: this is the best we can do. Also modified the code
4691 in apps/verify.c to take notice of return codes: it was previously
4692 attempting to read in certificates from NULL pointers and ignoring
4693 any errors: this is one reason why the cert and CRL reader seemed
4694 to work. It doesn't check return codes from the default certificate
4695 routines: these may well fail if the certificates aren't installed.
4698 *) Code to support otherName option in GeneralName.
4701 *) First update to verify code. Change the verify utility
4702 so it warns if it is passed a self signed certificate:
4703 for consistency with the normal behaviour. X509_verify
4704 has been modified to it will now verify a self signed
4705 certificate if *exactly* the same certificate appears
4706 in the store: it was previously impossible to trust a
4707 single self signed certificate. This means that:
4708 openssl verify ss.pem
4709 now gives a warning about a self signed certificate but
4710 openssl verify -CAfile ss.pem ss.pem
4714 *) For servers, store verify_result in SSL_SESSION data structure
4715 (and add it to external session representation).
4716 This is needed when client certificate verifications fails,
4717 but an application-provided verification callback (set by
4718 SSL_CTX_set_cert_verify_callback) allows accepting the session
4719 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
4720 but returns 1): When the session is reused, we have to set
4721 ssl->verify_result to the appropriate error code to avoid
4723 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
4725 *) Fix a bug in the new PKCS#7 code: it didn't consider the
4726 case in PKCS7_dataInit() where the signed PKCS7 structure
4727 didn't contain any existing data because it was being created.
4728 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
4730 *) Add a salt to the key derivation routines in enc.c. This
4731 forms the first 8 bytes of the encrypted file. Also add a
4732 -S option to allow a salt to be input on the command line.
4735 *) New function X509_cmp(). Oddly enough there wasn't a function
4736 to compare two certificates. We do this by working out the SHA1
4737 hash and comparing that. X509_cmp() will be needed by the trust
4741 *) SSL_get1_session() is like SSL_get_session(), but increments
4742 the reference count in the SSL_SESSION returned.
4743 [Geoff Thorpe <geoff@eu.c2.net>]
4745 *) Fix for 'req': it was adding a null to request attributes.
4746 Also change the X509_LOOKUP and X509_INFO code to handle
4747 certificate auxiliary information.
4750 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
4754 *) Add the possibility to add extra information to the memory leak
4755 detecting output, to form tracebacks, showing from where each
4756 allocation was originated: CRYPTO_push_info("constant string") adds
4757 the string plus current file name and line number to a per-thread
4758 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
4759 is like calling CYRPTO_pop_info() until the stack is empty.
4760 Also updated memory leak detection code to be multi-thread-safe.
4763 *) Add options -text and -noout to pkcs7 utility and delete the
4764 encryption options which never did anything. Update docs.
4767 *) Add options to some of the utilities to allow the pass phrase
4768 to be included on either the command line (not recommended on
4769 OSes like Unix) or read from the environment. Update the
4770 manpages and fix a few bugs.
4773 *) Add a few manpages for some of the openssl commands.
4776 *) Fix the -revoke option in ca. It was freeing up memory twice,
4777 leaking and not finding already revoked certificates.
4780 *) Extensive changes to support certificate auxiliary information.
4781 This involves the use of X509_CERT_AUX structure and X509_AUX
4782 functions. An X509_AUX function such as PEM_read_X509_AUX()
4783 can still read in a certificate file in the usual way but it
4784 will also read in any additional "auxiliary information". By
4785 doing things this way a fair degree of compatibility can be
4786 retained: existing certificates can have this information added
4787 using the new 'x509' options.
4789 Current auxiliary information includes an "alias" and some trust
4790 settings. The trust settings will ultimately be used in enhanced
4791 certificate chain verification routines: currently a certificate
4792 can only be trusted if it is self signed and then it is trusted
4796 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
4797 The problem was that one of the replacement routines had not been working
4798 since SSLeay releases. For now the offending routine has been replaced
4799 with non-optimised assembler. Even so, this now gives around 95%
4800 performance improvement for 1024 bit RSA signs.
4803 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
4804 handling. Most clients have the effective key size in bits equal to
4805 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
4806 A few however don't do this and instead use the size of the decrypted key
4807 to determine the RC2 key length and the AlgorithmIdentifier to determine
4808 the effective key length. In this case the effective key length can still
4809 be 40 bits but the key length can be 168 bits for example. This is fixed
4810 by manually forcing an RC2 key into the EVP_PKEY structure because the
4811 EVP code can't currently handle unusual RC2 key sizes: it always assumes
4812 the key length and effective key length are equal.
4815 *) Add a bunch of functions that should simplify the creation of
4816 X509_NAME structures. Now you should be able to do:
4817 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
4818 and have it automatically work out the correct field type and fill in
4819 the structures. The more adventurous can try:
4820 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
4821 and it will (hopefully) work out the correct multibyte encoding.
4824 *) Change the 'req' utility to use the new field handling and multibyte
4825 copy routines. Before the DN field creation was handled in an ad hoc
4826 way in req, ca, and x509 which was rather broken and didn't support
4827 BMPStrings or UTF8Strings. Since some software doesn't implement
4828 BMPStrings or UTF8Strings yet, they can be enabled using the config file
4829 using the dirstring_type option. See the new comment in the default
4830 openssl.cnf for more info.
4833 *) Make crypto/rand/md_rand.c more robust:
4834 - Assure unique random numbers after fork().
4835 - Make sure that concurrent threads access the global counter and
4836 md serializably so that we never lose entropy in them
4837 or use exactly the same state in multiple threads.
4838 Access to the large state is not always serializable because
4839 the additional locking could be a performance killer, and
4840 md should be large enough anyway.
4843 *) New file apps/app_rand.c with commonly needed functionality
4844 for handling the random seed file.
4846 Use the random seed file in some applications that previously did not:
4848 dsaparam -genkey (which also ignored its '-rand' option),
4851 x509 (when signing).
4852 Except on systems with /dev/urandom, it is crucial to have a random
4853 seed file at least for key creation, DSA signing, and for DH exchanges;
4854 for RSA signatures we could do without one.
4856 gendh and gendsa (unlike genrsa) used to read only the first byte
4857 of each file listed in the '-rand' option. The function as previously
4858 found in genrsa is now in app_rand.c and is used by all programs
4859 that support '-rand'.
4862 *) In RAND_write_file, use mode 0600 for creating files;
4863 don't just chmod when it may be too late.
4866 *) Report an error from X509_STORE_load_locations
4867 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
4870 *) New function ASN1_mbstring_copy() this copies a string in either
4871 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
4872 into an ASN1_STRING type. A mask of permissible types is passed
4873 and it chooses the "minimal" type to use or an error if not type
4877 *) Add function equivalents to the various macros in asn1.h. The old
4878 macros are retained with an M_ prefix. Code inside the library can
4879 use the M_ macros. External code (including the openssl utility)
4880 should *NOT* in order to be "shared library friendly".
4883 *) Add various functions that can check a certificate's extensions
4884 to see if it usable for various purposes such as SSL client,
4885 server or S/MIME and CAs of these types. This is currently
4886 VERY EXPERIMENTAL but will ultimately be used for certificate chain
4887 verification. Also added a -purpose flag to x509 utility to
4888 print out all the purposes.
4891 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
4895 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
4896 for, obtain and decode and extension and obtain its critical flag.
4897 This allows all the necessary extension code to be handled in a
4898 single function call.
4901 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
4902 platforms. See crypto/rc4/rc4_enc.c for further details.
4905 *) New -noout option to asn1parse. This causes no output to be produced
4906 its main use is when combined with -strparse and -out to extract data
4907 from a file (which may not be in ASN.1 format).
4910 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
4911 when producing the local key id.
4912 [Richard Levitte <levitte@stacken.kth.se>]
4914 *) New option -dhparam in s_server. This allows a DH parameter file to be
4915 stated explicitly. If it is not stated then it tries the first server
4916 certificate file. The previous behaviour hard coded the filename
4920 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
4921 a public key to be input or output. For example:
4922 openssl rsa -in key.pem -pubout -out pubkey.pem
4923 Also added necessary DSA public key functions to handle this.
4926 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
4927 in the message. This was handled by allowing
4928 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
4929 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
4931 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
4932 to the end of the strings whereas this didn't. This would cause problems
4933 if strings read with d2i_ASN1_bytes() were later modified.
4934 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
4936 *) Fix for base64 decode bug. When a base64 bio reads only one line of
4937 data and it contains EOF it will end up returning an error. This is
4938 caused by input 46 bytes long. The cause is due to the way base64
4939 BIOs find the start of base64 encoded data. They do this by trying a
4940 trial decode on each line until they find one that works. When they
4941 do a flag is set and it starts again knowing it can pass all the
4942 data directly through the decoder. Unfortunately it doesn't reset
4943 the context it uses. This means that if EOF is reached an attempt
4944 is made to pass two EOFs through the context and this causes the
4945 resulting error. This can also cause other problems as well. As is
4946 usual with these problems it takes *ages* to find and the fix is
4947 trivial: move one line.
4948 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
4950 *) Ugly workaround to get s_client and s_server working under Windows. The
4951 old code wouldn't work because it needed to select() on sockets and the
4952 tty (for keypresses and to see if data could be written). Win32 only
4953 supports select() on sockets so we select() with a 1s timeout on the
4954 sockets and then see if any characters are waiting to be read, if none
4955 are present then we retry, we also assume we can always write data to
4956 the tty. This isn't nice because the code then blocks until we've
4957 received a complete line of data and it is effectively polling the
4958 keyboard at 1s intervals: however it's quite a bit better than not
4959 working at all :-) A dedicated Windows application might handle this
4960 with an event loop for example.
4963 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
4964 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
4965 will be called when RSA_sign() and RSA_verify() are used. This is useful
4966 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
4967 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
4968 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
4969 This necessitated the support of an extra signature type NID_md5_sha1
4970 for SSL signatures and modifications to the SSL library to use it instead
4971 of calling RSA_public_decrypt() and RSA_private_encrypt().
4974 *) Add new -verify -CAfile and -CApath options to the crl program, these
4975 will lookup a CRL issuers certificate and verify the signature in a
4976 similar way to the verify program. Tidy up the crl program so it
4977 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
4978 less strict. It will now permit CRL extensions even if it is not
4979 a V2 CRL: this will allow it to tolerate some broken CRLs.
4982 *) Initialize all non-automatic variables each time one of the openssl
4983 sub-programs is started (this is necessary as they may be started
4984 multiple times from the "OpenSSL>" prompt).
4985 [Lennart Bang, Bodo Moeller]
4987 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
4988 removing all other RSA functionality (this is what NO_RSA does). This
4989 is so (for example) those in the US can disable those operations covered
4990 by the RSA patent while allowing storage and parsing of RSA keys and RSA
4994 *) Non-copying interface to BIO pairs.
4995 (still largely untested)
4998 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
4999 ASCII string. This was handled independently in various places before.
5002 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
5003 UTF8 strings a character at a time.
5006 *) Use client_version from client hello to select the protocol
5007 (s23_srvr.c) and for RSA client key exchange verification
5008 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
5011 *) Add various utility functions to handle SPKACs, these were previously
5012 handled by poking round in the structure internals. Added new function
5013 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
5014 print, verify and generate SPKACs. Based on an original idea from
5015 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
5018 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
5021 *) Allow the config file extension section to be overwritten on the
5022 command line. Based on an original idea from Massimiliano Pala
5023 <madwolf@comune.modena.it>. The new option is called -extensions
5024 and can be applied to ca, req and x509. Also -reqexts to override
5025 the request extensions in req and -crlexts to override the crl extensions
5029 *) Add new feature to the SPKAC handling in ca. Now you can include
5030 the same field multiple times by preceding it by "XXXX." for example:
5033 this is the same syntax as used in the req config file.
5036 *) Allow certificate extensions to be added to certificate requests. These
5037 are specified in a 'req_extensions' option of the req section of the
5038 config file. They can be printed out with the -text option to req but
5039 are otherwise ignored at present.
5042 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
5043 data read consists of only the final block it would not decrypted because
5044 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
5045 A misplaced 'break' also meant the decrypted final block might not be
5046 copied until the next read.
5049 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
5050 a few extra parameters to the DH structure: these will be useful if
5051 for example we want the value of 'q' or implement X9.42 DH.
5054 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
5055 provides hooks that allow the default DSA functions or functions on a
5056 "per key" basis to be replaced. This allows hardware acceleration and
5057 hardware key storage to be handled without major modification to the
5058 library. Also added low level modexp hooks and CRYPTO_EX structure and
5059 associated functions.
5062 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
5063 as "read only": it can't be written to and the buffer it points to will
5064 not be freed. Reading from a read only BIO is much more efficient than
5065 a normal memory BIO. This was added because there are several times when
5066 an area of memory needs to be read from a BIO. The previous method was
5067 to create a memory BIO and write the data to it, this results in two
5068 copies of the data and an O(n^2) reading algorithm. There is a new
5069 function BIO_new_mem_buf() which creates a read only memory BIO from
5070 an area of memory. Also modified the PKCS#7 routines to use read only
5074 *) Bugfix: ssl23_get_client_hello did not work properly when called in
5075 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
5076 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
5077 but a retry condition occured while trying to read the rest.
5080 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
5081 NID_pkcs7_encrypted by default: this was wrong since this should almost
5082 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
5083 the encrypted data type: this is a more sensible place to put it and it
5084 allows the PKCS#12 code to be tidied up that duplicated this
5088 *) Changed obj_dat.pl script so it takes its input and output files on
5089 the command line. This should avoid shell escape redirection problems
5093 *) Initial support for certificate extension requests, these are included
5094 in things like Xenroll certificate requests. Included functions to allow
5095 extensions to be obtained and added.
5098 *) -crlf option to s_client and s_server for sending newlines as
5099 CRLF (as required by many protocols).
5102 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
5104 *) Install libRSAglue.a when OpenSSL is built with RSAref.
5105 [Ralf S. Engelschall]
5107 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
5108 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
5110 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
5114 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
5115 DH parameters/keys (q is lost during that conversion, but the resulting
5116 DH parameters contain its length).
5118 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
5119 much faster than DH_generate_parameters (which creates parameters
5120 where p = 2*q + 1), and also the smaller q makes DH computations
5121 much more efficient (160-bit exponentiation instead of 1024-bit
5122 exponentiation); so this provides a convenient way to support DHE
5123 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
5124 utter importance to use
5125 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5127 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5128 when such DH parameters are used, because otherwise small subgroup
5129 attacks may become possible!
5132 *) Avoid memory leak in i2d_DHparams.
5135 *) Allow the -k option to be used more than once in the enc program:
5136 this allows the same encrypted message to be read by multiple recipients.
5139 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
5140 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
5141 it will always use the numerical form of the OID, even if it has a short
5145 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
5146 method only got called if p,q,dmp1,dmq1,iqmp components were present,
5147 otherwise bn_mod_exp was called. In the case of hardware keys for example
5148 no private key components need be present and it might store extra data
5149 in the RSA structure, which cannot be accessed from bn_mod_exp.
5150 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
5151 private key operations.
5154 *) Added support for SPARC Linux.
5157 *) pem_password_cb function type incompatibly changed from
5158 typedef int pem_password_cb(char *buf, int size, int rwflag);
5160 ....(char *buf, int size, int rwflag, void *userdata);
5161 so that applications can pass data to their callbacks:
5162 The PEM[_ASN1]_{read,write}... functions and macros now take an
5163 additional void * argument, which is just handed through whenever
5164 the password callback is called.
5165 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
5167 New function SSL_CTX_set_default_passwd_cb_userdata.
5169 Compatibility note: As many C implementations push function arguments
5170 onto the stack in reverse order, the new library version is likely to
5171 interoperate with programs that have been compiled with the old
5172 pem_password_cb definition (PEM_whatever takes some data that
5173 happens to be on the stack as its last argument, and the callback
5174 just ignores this garbage); but there is no guarantee whatsoever that
5177 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
5178 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
5179 problems not only on Windows, but also on some Unix platforms.
5180 To avoid problematic command lines, these definitions are now in an
5181 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
5182 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
5185 *) MIPS III/IV assembler module is reimplemented.
5188 *) More DES library cleanups: remove references to srand/rand and
5189 delete an unused file.
5192 *) Add support for the the free Netwide assembler (NASM) under Win32,
5193 since not many people have MASM (ml) and it can be hard to obtain.
5194 This is currently experimental but it seems to work OK and pass all
5195 the tests. Check out INSTALL.W32 for info.
5198 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
5199 without temporary keys kept an extra copy of the server key,
5200 and connections with temporary keys did not free everything in case
5204 *) New function RSA_check_key and new openssl rsa option -check
5205 for verifying the consistency of RSA keys.
5206 [Ulf Moeller, Bodo Moeller]
5208 *) Various changes to make Win32 compile work:
5209 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
5210 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
5211 comparison" warnings.
5212 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
5215 *) Add a debugging option to PKCS#5 v2 key generation function: when
5216 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
5217 derived keys are printed to stderr.
5220 *) Copy the flags in ASN1_STRING_dup().
5221 [Roman E. Pavlov <pre@mo.msk.ru>]
5223 *) The x509 application mishandled signing requests containing DSA
5224 keys when the signing key was also DSA and the parameters didn't match.
5226 It was supposed to omit the parameters when they matched the signing key:
5227 the verifying software was then supposed to automatically use the CA's
5228 parameters if they were absent from the end user certificate.
5230 Omitting parameters is no longer recommended. The test was also
5231 the wrong way round! This was probably due to unusual behaviour in
5232 EVP_cmp_parameters() which returns 1 if the parameters match.
5233 This meant that parameters were omitted when they *didn't* match and
5234 the certificate was useless. Certificates signed with 'ca' didn't have
5236 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
5238 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
5239 The interface is as follows:
5240 Applications can use
5241 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
5242 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
5243 "off" is now the default.
5244 The library internally uses
5245 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
5246 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
5247 to disable memory-checking temporarily.
5249 Some inconsistent states that previously were possible (and were
5250 even the default) are now avoided.
5252 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
5253 with each memory chunk allocated; this is occasionally more helpful
5254 than just having a counter.
5256 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
5258 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
5262 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
5263 which largely parallels "options", but is for changing API behaviour,
5264 whereas "options" are about protocol behaviour.
5265 Initial "mode" flags are:
5267 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
5268 a single record has been written.
5269 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
5270 retries use the same buffer location.
5271 (But all of the contents must be
5275 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
5278 *) Fix problems with no-hmac etc.
5279 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
5281 *) New functions RSA_get_default_method(), RSA_set_method() and
5282 RSA_get_method(). These allows replacement of RSA_METHODs without having
5283 to mess around with the internals of an RSA structure.
5286 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
5287 Also really enable memory leak checks in openssl.c and in some
5289 [Chad C. Mulligan, Bodo Moeller]
5291 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
5292 up the length of negative integers. This has now been simplified to just
5293 store the length when it is first determined and use it later, rather
5294 than trying to keep track of where data is copied and updating it to
5296 [Steve Henson, reported by Brien Wheeler
5297 <bwheeler@authentica-security.com>]
5299 *) Add a new function PKCS7_signatureVerify. This allows the verification
5300 of a PKCS#7 signature but with the signing certificate passed to the
5301 function itself. This contrasts with PKCS7_dataVerify which assumes the
5302 certificate is present in the PKCS#7 structure. This isn't always the
5303 case: certificates can be omitted from a PKCS#7 structure and be
5304 distributed by "out of band" means (such as a certificate database).
5307 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
5308 function prototypes in pem.h, also change util/mkdef.pl to add the
5309 necessary function names.
5312 *) mk1mf.pl (used by Windows builds) did not properly read the
5313 options set by Configure in the top level Makefile, and Configure
5314 was not even able to write more than one option correctly.
5315 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
5318 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
5319 file to be loaded from a BIO or FILE pointer. The BIO version will
5320 for example allow memory BIOs to contain config info.
5323 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
5324 Whoever hopes to achieve shared-library compatibility across versions
5325 must use this, not the compile-time macro.
5326 (Exercise 0.9.4: Which is the minimum library version required by
5328 Note: All this applies only to multi-threaded programs, others don't
5332 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
5333 through a BIO pair triggered the default case, i.e.
5334 SSLerr(...,SSL_R_UNKNOWN_STATE).
5337 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
5338 can use the SSL library even if none of the specific BIOs is
5342 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
5343 for the encoded length.
5344 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
5346 *) Add initial documentation of the X509V3 functions.
5349 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
5350 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
5351 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
5352 secure PKCS#8 private key format with a high iteration count.
5355 *) Fix determination of Perl interpreter: A perl or perl5
5356 _directory_ in $PATH was also accepted as the interpreter.
5357 [Ralf S. Engelschall]
5359 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
5360 wrong with it but it was very old and did things like calling
5361 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
5365 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
5366 to use the new extension code.
5369 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
5370 with macros. This should make it easier to change their form, add extra
5371 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
5375 *) Add to configuration table a new entry that can specify an alternative
5376 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
5377 according to Mark Crispin <MRC@Panda.COM>.
5381 *) DES CBC did not update the IV. Weird.
5384 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
5385 Changing the behaviour of the former might break existing programs --
5386 where IV updating is needed, des_ncbc_encrypt can be used.
5389 *) When bntest is run from "make test" it drives bc to check its
5390 calculations, as well as internally checking them. If an internal check
5391 fails, it needs to cause bc to give a non-zero result or make test carries
5392 on without noticing the failure. Fixed.
5395 *) DES library cleanups.
5398 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
5399 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
5400 ciphers. NOTE: although the key derivation function has been verified
5401 against some published test vectors it has not been extensively tested
5402 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
5406 *) Instead of "mkdir -p", which is not fully portable, use new
5407 Perl script "util/mkdir-p.pl".
5410 *) Rewrite the way password based encryption (PBE) is handled. It used to
5411 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
5412 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
5413 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
5414 the 'parameter' field of the AlgorithmIdentifier is passed to the
5415 underlying key generation function so it must do its own ASN1 parsing.
5416 This has also changed the EVP_PBE_CipherInit() function which now has a
5417 'parameter' argument instead of literal salt and iteration count values
5418 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
5421 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
5422 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
5423 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
5424 KEY" because this clashed with PKCS#8 unencrypted string. Since this
5425 value was just used as a "magic string" and not used directly its
5426 value doesn't matter.
5429 *) Introduce some semblance of const correctness to BN. Shame C doesn't
5433 *) "linux-sparc64" configuration (ultrapenguin).
5434 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
5435 "linux-sparc" configuration.
5436 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
5438 *) config now generates no-xxx options for missing ciphers.
5441 *) Support the EBCDIC character set (work in progress).
5442 File ebcdic.c not yet included because it has a different license.
5443 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5445 *) Support BS2000/OSD-POSIX.
5446 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5448 *) Make callbacks for key generation use void * instead of char *.
5451 *) Make S/MIME samples compile (not yet tested).
5454 *) Additional typesafe stacks.
5457 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
5461 Changes between 0.9.3 and 0.9.3a [29 May 1999]
5463 *) New configuration variant "sco5-gcc".
5465 *) Updated some demos.
5466 [Sean O Riordain, Wade Scholine]
5468 *) Add missing BIO_free at exit of pkcs12 application.
5471 *) Fix memory leak in conf.c.
5474 *) Updates for Win32 to assembler version of MD5.
5477 *) Set #! path to perl in apps/der_chop to where we found it
5478 instead of using a fixed path.
5481 *) SHA library changes for irix64-mips4-cc.
5484 *) Improvements for VMS support.
5488 Changes between 0.9.2b and 0.9.3 [24 May 1999]
5490 *) Bignum library bug fix. IRIX 6 passes "make test" now!
5491 This also avoids the problems with SC4.2 and unpatched SC5.
5492 [Andy Polyakov <appro@fy.chalmers.se>]
5494 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
5495 These are required because of the typesafe stack would otherwise break
5496 existing code. If old code used a structure member which used to be STACK
5497 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
5498 sk_num or sk_value it would produce an error because the num, data members
5499 are not present in STACK_OF. Now it just produces a warning. sk_set
5500 replaces the old method of assigning a value to sk_value
5501 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
5502 that does this will no longer work (and should use sk_set instead) but
5503 this could be regarded as a "questionable" behaviour anyway.
5506 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
5507 correctly handle encrypted S/MIME data.
5510 *) Change type of various DES function arguments from des_cblock
5511 (which means, in function argument declarations, pointer to char)
5512 to des_cblock * (meaning pointer to array with 8 char elements),
5513 which allows the compiler to do more typechecking; it was like
5514 that back in SSLeay, but with lots of ugly casts.
5516 Introduce new type const_des_cblock.
5519 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
5520 problems: find RecipientInfo structure that matches recipient certificate
5521 and initialise the ASN1 structures properly based on passed cipher.
5524 *) Belatedly make the BN tests actually check the results.
5527 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
5528 to and from BNs: it was completely broken. New compilation option
5529 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
5530 key elements as negative integers.
5533 *) Reorganize and speed up MD5.
5534 [Andy Polyakov <appro@fy.chalmers.se>]
5537 [Richard Levitte <richard@levitte.org>]
5539 *) New option -out to asn1parse to allow the parsed structure to be
5540 output to a file. This is most useful when combined with the -strparse
5541 option to examine the output of things like OCTET STRINGS.
5544 *) Make SSL library a little more fool-proof by not requiring any longer
5545 that SSL_set_{accept,connect}_state be called before
5546 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
5547 in many applications because usually everything *appeared* to work as
5548 intended anyway -- now it really works as intended).
5551 *) Move openssl.cnf out of lib/.
5554 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
5555 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
5556 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
5557 [Ralf S. Engelschall]
5559 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
5560 handle PKCS#7 enveloped data properly.
5561 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
5563 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
5564 copying pointers. The cert_st handling is changed by this in
5565 various ways (and thus what used to be known as ctx->default_cert
5566 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
5567 any longer when s->cert does not give us what we need).
5568 ssl_cert_instantiate becomes obsolete by this change.
5569 As soon as we've got the new code right (possibly it already is?),
5570 we have solved a couple of bugs of the earlier code where s->cert
5571 was used as if it could not have been shared with other SSL structures.
5573 Note that using the SSL API in certain dirty ways now will result
5574 in different behaviour than observed with earlier library versions:
5575 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
5576 does not influence s as it used to.
5578 In order to clean up things more thoroughly, inside SSL_SESSION
5579 we don't use CERT any longer, but a new structure SESS_CERT
5580 that holds per-session data (if available); currently, this is
5581 the peer's certificate chain and, for clients, the server's certificate
5582 and temporary key. CERT holds only those values that can have
5583 meaningful defaults in an SSL_CTX.
5586 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
5587 from the internal representation. Various PKCS#7 fixes: remove some
5588 evil casts and set the enc_dig_alg field properly based on the signing
5592 *) Allow PKCS#12 password to be set from the command line or the
5593 environment. Let 'ca' get its config file name from the environment
5594 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
5598 *) Allow certificate policies extension to use an IA5STRING for the
5599 organization field. This is contrary to the PKIX definition but
5600 VeriSign uses it and IE5 only recognises this form. Document 'x509'
5604 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
5605 without disallowing inline assembler and the like for non-pedantic builds.
5608 *) Support Borland C++ builder.
5609 [Janez Jere <jj@void.si>, modified by Ulf Möller]
5614 *) SHA-1 cleanups and performance enhancements.
5615 [Andy Polyakov <appro@fy.chalmers.se>]
5617 *) Sparc v8plus assembler for the bignum library.
5618 [Andy Polyakov <appro@fy.chalmers.se>]
5620 *) Accept any -xxx and +xxx compiler options in Configure.
5623 *) Update HPUX configuration.
5626 *) Add missing sk_<type>_unshift() function to safestack.h
5627 [Ralf S. Engelschall]
5629 *) New function SSL_CTX_use_certificate_chain_file that sets the
5630 "extra_cert"s in addition to the certificate. (This makes sense
5631 only for "PEM" format files, as chains as a whole are not
5635 *) Support verify_depth from the SSL API.
5636 x509_vfy.c had what can be considered an off-by-one-error:
5637 Its depth (which was not part of the external interface)
5638 was actually counting the number of certificates in a chain;
5639 now it really counts the depth.
5642 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
5643 instead of X509err, which often resulted in confusing error
5644 messages since the error codes are not globally unique
5645 (e.g. an alleged error in ssl3_accept when a certificate
5646 didn't match the private key).
5648 *) New function SSL_CTX_set_session_id_context that allows to set a default
5649 value (so that you don't need SSL_set_session_id_context for each
5650 connection using the SSL_CTX).
5653 *) OAEP decoding bug fix.
5656 *) Support INSTALL_PREFIX for package builders, as proposed by
5660 *) New Configure options "threads" and "no-threads". For systems
5661 where the proper compiler options are known (currently Solaris
5662 and Linux), "threads" is the default.
5665 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
5668 *) Install various scripts to $(OPENSSLDIR)/misc, not to
5669 $(INSTALLTOP)/bin -- they shouldn't clutter directories
5670 such as /usr/local/bin.
5673 *) "make linux-shared" to build shared libraries.
5674 [Niels Poppe <niels@netbox.org>]
5676 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
5679 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
5680 extension adding in x509 utility.
5683 *) Remove NOPROTO sections and error code comments.
5686 *) Partial rewrite of the DEF file generator to now parse the ANSI
5690 *) New Configure options --prefix=DIR and --openssldir=DIR.
5693 *) Complete rewrite of the error code script(s). It is all now handled
5694 by one script at the top level which handles error code gathering,
5695 header rewriting and C source file generation. It should be much better
5696 than the old method: it now uses a modified version of Ulf's parser to
5697 read the ANSI prototypes in all header files (thus the old K&R definitions
5698 aren't needed for error creation any more) and do a better job of
5699 translating function codes into names. The old 'ASN1 error code imbedded
5700 in a comment' is no longer necessary and it doesn't use .err files which
5701 have now been deleted. Also the error code call doesn't have to appear all
5702 on one line (which resulted in some large lines...).
5705 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
5708 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
5709 0 (which usually indicates a closed connection), but continue reading.
5712 *) Fix some race conditions.
5715 *) Add support for CRL distribution points extension. Add Certificate
5716 Policies and CRL distribution points documentation.
5719 *) Move the autogenerated header file parts to crypto/opensslconf.h.
5722 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
5723 8 of keying material. Merlin has also confirmed interop with this fix
5724 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
5725 [Merlin Hughes <merlin@baltimore.ie>]
5727 *) Fix lots of warnings.
5728 [Richard Levitte <levitte@stacken.kth.se>]
5730 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
5731 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
5732 [Richard Levitte <levitte@stacken.kth.se>]
5734 *) Fix problems with sizeof(long) == 8.
5735 [Andy Polyakov <appro@fy.chalmers.se>]
5737 *) Change functions to ANSI C.
5740 *) Fix typos in error codes.
5741 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
5743 *) Remove defunct assembler files from Configure.
5746 *) SPARC v8 assembler BIGNUM implementation.
5747 [Andy Polyakov <appro@fy.chalmers.se>]
5749 *) Support for Certificate Policies extension: both print and set.
5750 Various additions to support the r2i method this uses.
5753 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
5754 return a const string when you are expecting an allocated buffer.
5757 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
5758 types DirectoryString and DisplayText.
5761 *) Add code to allow r2i extensions to access the configuration database,
5762 add an LHASH database driver and add several ctx helper functions.
5765 *) Fix an evil bug in bn_expand2() which caused various BN functions to
5766 fail when they extended the size of a BIGNUM.
5769 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
5770 support typesafe stack.
5773 *) Fix typo in SSL_[gs]et_options().
5774 [Nils Frostberg <nils@medcom.se>]
5776 *) Delete various functions and files that belonged to the (now obsolete)
5777 old X509V3 handling code.
5780 *) New Configure option "rsaref".
5783 *) Don't auto-generate pem.h.
5786 *) Introduce type-safe ASN.1 SETs.
5789 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
5790 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
5792 *) Introduce type-safe STACKs. This will almost certainly break lots of code
5793 that links with OpenSSL (well at least cause lots of warnings), but fear
5794 not: the conversion is trivial, and it eliminates loads of evil casts. A
5795 few STACKed things have been converted already. Feel free to convert more.
5796 In the fullness of time, I'll do away with the STACK type altogether.
5799 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
5800 specified in <certfile> by updating the entry in the index.txt file.
5801 This way one no longer has to edit the index.txt file manually for
5802 revoking a certificate. The -revoke option does the gory details now.
5803 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
5805 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
5806 `-text' option at all and this way the `-noout -text' combination was
5807 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
5808 [Ralf S. Engelschall]
5810 *) Make sure a corresponding plain text error message exists for the
5811 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
5812 verify callback function determined that a certificate was revoked.
5813 [Ralf S. Engelschall]
5815 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
5816 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
5817 all available cipers including rc5, which was forgotten until now.
5818 In order to let the testing shell script know which algorithms
5819 are available, a new (up to now undocumented) command
5820 "openssl list-cipher-commands" is used.
5823 *) Bugfix: s_client occasionally would sleep in select() when
5824 it should have checked SSL_pending() first.
5827 *) New functions DSA_do_sign and DSA_do_verify to provide access to
5828 the raw DSA values prior to ASN.1 encoding.
5831 *) Tweaks to Configure
5832 [Niels Poppe <niels@netbox.org>]
5834 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
5838 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
5841 *) New config option to avoid instructions that are illegal on the 80386.
5842 The default code is faster, but requires at least a 486.
5845 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
5846 SSL2_SERVER_VERSION (not used at all) macros, which are now the
5847 same as SSL2_VERSION anyway.
5850 *) New "-showcerts" option for s_client.
5853 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
5854 application. Various cleanups and fixes.
5857 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
5858 modify error routines to work internally. Add error codes and PBE init
5859 to library startup routines.
5862 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
5863 packing functions to asn1 and evp. Changed function names and error
5864 codes along the way.
5867 *) PKCS12 integration: and so it begins... First of several patches to
5868 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
5869 objects to objects.h
5872 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
5873 and display support for Thawte strong extranet extension.
5876 *) Add LinuxPPC support.
5877 [Jeff Dubrule <igor@pobox.org>]
5879 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
5880 bn_div_words in alpha.s.
5881 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
5883 *) Make sure the RSA OAEP test is skipped under -DRSAref because
5884 OAEP isn't supported when OpenSSL is built with RSAref.
5885 [Ulf Moeller <ulf@fitug.de>]
5887 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
5888 so they no longer are missing under -DNOPROTO.
5889 [Soren S. Jorvang <soren@t.dk>]
5892 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
5894 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
5895 doesn't work when the session is reused. Coming soon!
5898 *) Fix a security hole, that allows sessions to be reused in the wrong
5899 context thus bypassing client cert protection! All software that uses
5900 client certs and session caches in multiple contexts NEEDS PATCHING to
5901 allow session reuse! A fuller solution is in the works.
5902 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
5904 *) Some more source tree cleanups (removed obsolete files
5905 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
5906 permission on "config" script to be executable) and a fix for the INSTALL
5908 [Ulf Moeller <ulf@fitug.de>]
5910 *) Remove some legacy and erroneous uses of malloc, free instead of
5912 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
5914 *) Make rsa_oaep_test return non-zero on error.
5915 [Ulf Moeller <ulf@fitug.de>]
5917 *) Add support for native Solaris shared libraries. Configure
5918 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
5919 if someone would make that last step automatic.
5920 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
5922 *) ctx_size was not built with the right compiler during "make links". Fixed.
5925 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
5926 except NULL ciphers". This means the default cipher list will no longer
5927 enable NULL ciphers. They need to be specifically enabled e.g. with
5928 the string "DEFAULT:eNULL".
5931 *) Fix to RSA private encryption routines: if p < q then it would
5932 occasionally produce an invalid result. This will only happen with
5933 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
5936 *) Be less restrictive and allow also `perl util/perlpath.pl
5937 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
5938 because this way one can also use an interpreter named `perl5' (which is
5939 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
5940 installed as `perl').
5941 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5943 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
5944 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5946 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
5947 advapi32.lib to Win32 build and change the pem test comparision
5948 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
5949 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
5950 and crypto/des/ede_cbcm_enc.c.
5953 *) DES quad checksum was broken on big-endian architectures. Fixed.
5956 *) Comment out two functions in bio.h that aren't implemented. Fix up the
5957 Win32 test batch file so it (might) work again. The Win32 test batch file
5958 is horrible: I feel ill....
5961 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
5962 in e_os.h. Audit of header files to check ANSI and non ANSI
5963 sections: 10 functions were absent from non ANSI section and not exported
5964 from Windows DLLs. Fixed up libeay.num for new functions.
5967 *) Make `openssl version' output lines consistent.
5968 [Ralf S. Engelschall]
5970 *) Fix Win32 symbol export lists for BIO functions: Added
5971 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
5972 to ms/libeay{16,32}.def.
5973 [Ralf S. Engelschall]
5975 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
5976 fine under Unix and passes some trivial tests I've now added. But the
5977 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
5978 added to make sure no one expects that this stuff really works in the
5979 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
5980 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
5982 [Ralf S. Engelschall]
5984 *) Fix the generation of two part addresses in perl.
5985 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
5987 *) Add config entry for Linux on MIPS.
5988 [John Tobey <jtobey@channel1.com>]
5990 *) Make links whenever Configure is run, unless we are on Windoze.
5993 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
5994 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
5998 *) Add a useful kludge to allow package maintainers to specify compiler and
5999 other platforms details on the command line without having to patch the
6000 Configure script everytime: One now can use ``perl Configure
6001 <id>:<details>'', i.e. platform ids are allowed to have details appended
6002 to them (seperated by colons). This is treated as there would be a static
6003 pre-configured entry in Configure's %table under key <id> with value
6004 <details> and ``perl Configure <id>'' is called. So, when you want to
6005 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
6006 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
6007 now, which overrides the FreeBSD-elf entry on-the-fly.
6008 [Ralf S. Engelschall]
6010 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
6013 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
6014 on the `perl Configure ...' command line. This way one can compile
6015 OpenSSL libraries with Position Independent Code (PIC) which is needed
6016 for linking it into DSOs.
6017 [Ralf S. Engelschall]
6019 *) Remarkably, export ciphers were totally broken and no-one had noticed!
6023 *) Cleaned up the LICENSE document: The official contact for any license
6024 questions now is the OpenSSL core team under openssl-core@openssl.org.
6025 And add a paragraph about the dual-license situation to make sure people
6026 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
6027 to the OpenSSL toolkit.
6028 [Ralf S. Engelschall]
6030 *) General source tree makefile cleanups: Made `making xxx in yyy...'
6031 display consistent in the source tree and replaced `/bin/rm' by `rm'.
6032 Additonally cleaned up the `make links' target: Remove unnecessary
6033 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
6034 to speed processing and no longer clutter the display with confusing
6035 stuff. Instead only the actually done links are displayed.
6036 [Ralf S. Engelschall]
6038 *) Permit null encryption ciphersuites, used for authentication only. It used
6039 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
6040 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
6044 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
6045 signed attributes when verifying signatures (this would break them),
6046 the detached data encoding was wrong and public keys obtained using
6047 X509_get_pubkey() weren't freed.
6050 *) Add text documentation for the BUFFER functions. Also added a work around
6051 to a Win95 console bug. This was triggered by the password read stuff: the
6052 last character typed gets carried over to the next fread(). If you were
6053 generating a new cert request using 'req' for example then the last
6054 character of the passphrase would be CR which would then enter the first
6058 *) Added the new `Includes OpenSSL Cryptography Software' button as
6059 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
6060 button and can be used by applications based on OpenSSL to show the
6061 relationship to the OpenSSL project.
6062 [Ralf S. Engelschall]
6064 *) Remove confusing variables in function signatures in files
6065 ssl/ssl_lib.c and ssl/ssl.h.
6066 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6068 *) Don't install bss_file.c under PREFIX/include/
6069 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6071 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
6072 functions that return function pointers and has support for NT specific
6073 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
6074 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
6075 unsigned to signed types: this was killing the Win32 compile.
6078 *) Add new certificate file to stack functions,
6079 SSL_add_dir_cert_subjects_to_stack() and
6080 SSL_add_file_cert_subjects_to_stack(). These largely supplant
6081 SSL_load_client_CA_file(), and can be used to add multiple certs easily
6082 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
6083 This means that Apache-SSL and similar packages don't have to mess around
6084 to add as many CAs as they want to the preferred list.
6087 *) Experiment with doxygen documentation. Currently only partially applied to
6089 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
6090 openssl.doxy as the configuration file.
6093 *) Get rid of remaining C++-style comments which strict C compilers hate.
6094 [Ralf S. Engelschall, pointed out by Carlos Amengual]
6096 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
6097 compiled in by default: it has problems with large keys.
6100 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
6101 DH private keys and/or callback functions which directly correspond to
6102 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
6103 is needed for applications which have to configure certificates on a
6104 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
6106 For the RSA certificate situation is makes no difference, but
6107 for the DSA certificate situation this fixes the "no shared cipher"
6108 problem where the OpenSSL cipher selection procedure failed because the
6109 temporary keys were not overtaken from the context and the API provided
6110 no way to reconfigure them.
6111 The new functions now let applications reconfigure the stuff and they
6112 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
6113 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
6114 non-public-API function ssl_cert_instantiate() is used as a helper
6115 function and also to reduce code redundancy inside ssl_rsa.c.
6116 [Ralf S. Engelschall]
6118 *) Move s_server -dcert and -dkey options out of the undocumented feature
6119 area because they are useful for the DSA situation and should be
6120 recognized by the users.
6121 [Ralf S. Engelschall]
6123 *) Fix the cipher decision scheme for export ciphers: the export bits are
6124 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
6125 SSL_EXP_MASK. So, the original variable has to be used instead of the
6126 already masked variable.
6127 [Richard Levitte <levitte@stacken.kth.se>]
6129 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
6130 [Richard Levitte <levitte@stacken.kth.se>]
6132 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
6133 from `int' to `unsigned int' because it's a length and initialized by
6134 EVP_DigestFinal() which expects an `unsigned int *'.
6135 [Richard Levitte <levitte@stacken.kth.se>]
6137 *) Don't hard-code path to Perl interpreter on shebang line of Configure
6138 script. Instead use the usual Shell->Perl transition trick.
6139 [Ralf S. Engelschall]
6141 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
6142 (in addition to RSA certificates) to match the behaviour of `openssl dsa
6143 -noout -modulus' as it's already the case for `openssl rsa -noout
6144 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
6145 currently the public key is printed (a decision which was already done by
6146 `openssl dsa -modulus' in the past) which serves a similar purpose.
6147 Additionally the NO_RSA no longer completely removes the whole -modulus
6148 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
6150 [Ralf S. Engelschall]
6152 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
6153 BIO. See the source (crypto/evp/bio_ok.c) for more info.
6154 [Arne Ansper <arne@ats.cyber.ee>]
6156 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
6157 to be added. Now both 'req' and 'ca' can use new objects defined in the
6161 *) Add cool BIO that does syslog (or event log on NT).
6162 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
6164 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
6165 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
6166 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
6167 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
6170 *) Add preliminary config info for new extension code.
6173 *) Make RSA_NO_PADDING really use no padding.
6174 [Ulf Moeller <ulf@fitug.de>]
6176 *) Generate errors when private/public key check is done.
6179 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
6180 for some CRL extensions and new objects added.
6183 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
6184 key usage extension and fuller support for authority key id.
6187 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
6188 padding method for RSA, which is recommended for new applications in PKCS
6189 #1 v2.0 (RFC 2437, October 1998).
6190 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
6191 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
6192 against Bleichbacher's attack on RSA.
6193 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
6196 *) Updates to the new SSL compression code
6197 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6199 *) Fix so that the version number in the master secret, when passed
6200 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
6201 (because the server will not accept higher), that the version number
6202 is 0x03,0x01, not 0x03,0x00
6203 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6205 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
6206 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
6207 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
6210 *) Support for RAW extensions where an arbitrary extension can be
6211 created by including its DER encoding. See apps/openssl.cnf for
6215 *) Make sure latest Perl versions don't interpret some generated C array
6216 code as Perl array code in the crypto/err/err_genc.pl script.
6217 [Lars Weber <3weber@informatik.uni-hamburg.de>]
6219 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
6220 not many people have the assembler. Various Win32 compilation fixes and
6221 update to the INSTALL.W32 file with (hopefully) more accurate Win32
6225 *) Modify configure script 'Configure' to automatically create crypto/date.h
6226 file under Win32 and also build pem.h from pem.org. New script
6227 util/mkfiles.pl to create the MINFO file on environments that can't do a
6228 'make files': perl util/mkfiles.pl >MINFO should work.
6231 *) Major rework of DES function declarations, in the pursuit of correctness
6232 and purity. As a result, many evil casts evaporated, and some weirdness,
6233 too. You may find this causes warnings in your code. Zapping your evil
6234 casts will probably fix them. Mostly.
6237 *) Fix for a typo in asn1.h. Bug fix to object creation script
6238 obj_dat.pl. It considered a zero in an object definition to mean
6239 "end of object": none of the objects in objects.h have any zeros
6240 so it wasn't spotted.
6241 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
6243 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
6244 Masking (CBCM). In the absence of test vectors, the best I have been able
6245 to do is check that the decrypt undoes the encrypt, so far. Send me test
6246 vectors if you have them.
6249 *) Correct calculation of key length for export ciphers (too much space was
6250 allocated for null ciphers). This has not been tested!
6253 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
6254 message is now correct (it understands "crypto" and "ssl" on its
6255 command line). There is also now an "update" option. This will update
6256 the util/ssleay.num and util/libeay.num files with any new functions.
6258 perl util/mkdef.pl crypto ssl update
6259 it will update them.
6262 *) Overhauled the Perl interface (perl/*):
6263 - ported BN stuff to OpenSSL's different BN library
6264 - made the perl/ source tree CVS-aware
6265 - renamed the package from SSLeay to OpenSSL (the files still contain
6266 their history because I've copied them in the repository)
6267 - removed obsolete files (the test scripts will be replaced
6268 by better Test::Harness variants in the future)
6269 [Ralf S. Engelschall]
6271 *) First cut for a very conservative source tree cleanup:
6272 1. merge various obsolete readme texts into doc/ssleay.txt
6273 where we collect the old documents and readme texts.
6274 2. remove the first part of files where I'm already sure that we no
6275 longer need them because of three reasons: either they are just temporary
6276 files which were left by Eric or they are preserved original files where
6277 I've verified that the diff is also available in the CVS via "cvs diff
6278 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
6279 the crypto/md/ stuff).
6280 [Ralf S. Engelschall]
6282 *) More extension code. Incomplete support for subject and issuer alt
6283 name, issuer and authority key id. Change the i2v function parameters
6284 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
6285 what that's for :-) Fix to ASN1 macro which messed up
6286 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
6289 *) Preliminary support for ENUMERATED type. This is largely copied from the
6293 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
6294 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6296 *) Make sure `make rehash' target really finds the `openssl' program.
6297 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6299 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
6300 like to hear about it if this slows down other processors.
6303 *) Add CygWin32 platform information to Configure script.
6304 [Alan Batie <batie@aahz.jf.intel.com>]
6306 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
6307 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
6309 *) New program nseq to manipulate netscape certificate sequences
6312 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
6316 *) Fixes to BN code. Previously the default was to define BN_RECURSION
6317 but the BN code had some problems that would cause failures when
6318 doing certificate verification and some other functions.
6319 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6321 *) Add ASN1 and PEM code to support netscape certificate sequences.
6324 *) Add ASN1 and PEM code to support netscape certificate sequences.
6327 *) Add several PKIX and private extended key usage OIDs.
6330 *) Modify the 'ca' program to handle the new extension code. Modify
6331 openssl.cnf for new extension format, add comments.
6334 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
6335 and add a sample to openssl.cnf so req -x509 now adds appropriate
6339 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
6340 error code, add initial support to X509_print() and x509 application.
6343 *) Takes a deep breath and start addding X509 V3 extension support code. Add
6344 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
6345 stuff is currently isolated and isn't even compiled yet.
6348 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
6349 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
6350 Removed the versions check from X509 routines when loading extensions:
6351 this allows certain broken certificates that don't set the version
6352 properly to be processed.
6355 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
6356 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
6357 can still be regenerated with "make depend".
6360 *) Spelling mistake in C version of CAST-128.
6361 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
6363 *) Changes to the error generation code. The perl script err-code.pl
6364 now reads in the old error codes and retains the old numbers, only
6365 adding new ones if necessary. It also only changes the .err files if new
6366 codes are added. The makefiles have been modified to only insert errors
6367 when needed (to avoid needlessly modifying header files). This is done
6368 by only inserting errors if the .err file is newer than the auto generated
6369 C file. To rebuild all the error codes from scratch (the old behaviour)
6370 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
6371 or delete all the .err files.
6374 *) CAST-128 was incorrectly implemented for short keys. The C version has
6375 been fixed, but is untested. The assembler versions are also fixed, but
6376 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
6377 to regenerate it if needed.
6378 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
6379 Hagino <itojun@kame.net>]
6381 *) File was opened incorrectly in randfile.c.
6382 [Ulf Möller <ulf@fitug.de>]
6384 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
6385 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
6386 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
6387 al: it's just almost always a UTCTime. Note this patch adds new error
6388 codes so do a "make errors" if there are problems.
6391 *) Correct Linux 1 recognition in config.
6392 [Ulf Möller <ulf@fitug.de>]
6394 *) Remove pointless MD5 hash when using DSA keys in ca.
6395 [Anonymous <nobody@replay.com>]
6397 *) Generate an error if given an empty string as a cert directory. Also
6398 generate an error if handed NULL (previously returned 0 to indicate an
6399 error, but didn't set one).
6400 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
6402 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
6405 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
6406 parameters. This was causing a warning which killed off the Win32 compile.
6409 *) Remove C++ style comments from crypto/bn/bn_local.h.
6410 [Neil Costigan <neil.costigan@celocom.com>]
6412 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
6413 based on a text string, looking up short and long names and finally
6414 "dot" format. The "dot" format stuff didn't work. Added new function
6415 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
6416 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
6417 OID is not part of the table.
6420 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
6421 X509_LOOKUP_by_alias().
6424 *) Sort openssl functions by name.
6427 *) Get the gendsa program working (hopefully) and add it to app list. Remove
6428 encryption from sample DSA keys (in case anyone is interested the password
6432 *) Make _all_ *_free functions accept a NULL pointer.
6433 [Frans Heymans <fheymans@isaserver.be>]
6435 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
6437 [Anonymous <nobody@replay.com>]
6439 *) s_server should send the CAfile as acceptable CAs, not its own cert.
6440 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6442 *) Don't blow it for numeric -newkey arguments to apps/req.
6443 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6445 *) Temp key "for export" tests were wrong in s3_srvr.c.
6446 [Anonymous <nobody@replay.com>]
6448 *) Add prototype for temp key callback functions
6449 SSL_CTX_set_tmp_{rsa,dh}_callback().
6452 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
6453 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
6456 *) X509_name_add_entry() freed the wrong thing after an error.
6457 [Arne Ansper <arne@ats.cyber.ee>]
6459 *) rsa_eay.c would attempt to free a NULL context.
6460 [Arne Ansper <arne@ats.cyber.ee>]
6462 *) BIO_s_socket() had a broken should_retry() on Windoze.
6463 [Arne Ansper <arne@ats.cyber.ee>]
6465 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
6466 [Arne Ansper <arne@ats.cyber.ee>]
6468 *) Make sure the already existing X509_STORE->depth variable is initialized
6469 in X509_STORE_new(), but document the fact that this variable is still
6470 unused in the certificate verification process.
6471 [Ralf S. Engelschall]
6473 *) Fix the various library and apps files to free up pkeys obtained from
6474 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
6477 *) Fix reference counting in X509_PUBKEY_get(). This makes
6478 demos/maurice/example2.c work, amongst others, probably.
6479 [Steve Henson and Ben Laurie]
6481 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
6482 `openssl' and second, the shortcut symlinks for the `openssl <command>'
6483 are no longer created. This way we have a single and consistent command
6484 line interface `openssl <command>', similar to `cvs <command>'.
6485 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
6487 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
6488 BIT STRING wrapper always have zero unused bits.
6491 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
6494 *) Make the top-level INSTALL documentation easier to understand.
6497 *) Makefiles updated to exit if an error occurs in a sub-directory
6498 make (including if user presses ^C) [Paul Sutton]
6500 *) Make Montgomery context stuff explicit in RSA data structure.
6503 *) Fix build order of pem and err to allow for generated pem.h.
6506 *) Fix renumbering bug in X509_NAME_delete_entry().
6509 *) Enhanced the err-ins.pl script so it makes the error library number
6510 global and can add a library name. This is needed for external ASN1 and
6511 other error libraries.
6514 *) Fixed sk_insert which never worked properly.
6517 *) Fix ASN1 macros so they can handle indefinite length construted
6518 EXPLICIT tags. Some non standard certificates use these: they can now
6522 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
6523 into a single doc/ssleay.txt bundle. This way the information is still
6524 preserved but no longer messes up this directory. Now it's new room for
6525 the new set of documenation files.
6526 [Ralf S. Engelschall]
6528 *) SETs were incorrectly DER encoded. This was a major pain, because they
6529 shared code with SEQUENCEs, which aren't coded the same. This means that
6530 almost everything to do with SETs or SEQUENCEs has either changed name or
6531 number of arguments.
6532 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
6534 *) Fix test data to work with the above.
6537 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
6538 was already fixed by Eric for 0.9.1 it seems.
6539 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
6541 *) Autodetect FreeBSD3.
6544 *) Fix various bugs in Configure. This affects the following platforms:
6548 unixware-2.0-pentium
6552 *) Eliminate generated files from CVS. Reorder tests to regenerate files
6553 before they are needed.
6556 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
6560 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
6562 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
6563 changed SSLeay to OpenSSL in version strings.
6564 [Ralf S. Engelschall]
6566 *) Some fixups to the top-level documents.
6569 *) Fixed the nasty bug where rsaref.h was not found under compile-time
6570 because the symlink to include/ was missing.
6571 [Ralf S. Engelschall]
6573 *) Incorporated the popular no-RSA/DSA-only patches
6574 which allow to compile a RSA-free SSLeay.
6575 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
6577 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
6578 when "ssleay" is still not found.
6579 [Ralf S. Engelschall]
6581 *) Added more platforms to Configure: Cray T3E, HPUX 11,
6582 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
6584 *) Updated the README file.
6585 [Ralf S. Engelschall]
6587 *) Added various .cvsignore files in the CVS repository subdirs
6588 to make a "cvs update" really silent.
6589 [Ralf S. Engelschall]
6591 *) Recompiled the error-definition header files and added
6592 missing symbols to the Win32 linker tables.
6593 [Ralf S. Engelschall]
6595 *) Cleaned up the top-level documents;
6596 o new files: CHANGES and LICENSE
6597 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
6598 o merged COPYRIGHT into LICENSE
6599 o removed obsolete TODO file
6600 o renamed MICROSOFT to INSTALL.W32
6601 [Ralf S. Engelschall]
6603 *) Removed dummy files from the 0.9.1b source tree:
6604 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
6605 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
6606 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
6607 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
6608 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
6609 [Ralf S. Engelschall]
6611 *) Added various platform portability fixes.
6614 *) The Genesis of the OpenSSL rpject:
6615 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
6616 Young and Tim J. Hudson created while they were working for C2Net until
6618 [The OpenSSL Project]
6621 Changes between 0.9.0b and 0.9.1b [not released]
6623 *) Updated a few CA certificates under certs/
6626 *) Changed some BIGNUM api stuff.
6629 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
6630 DGUX x86, Linux Alpha, etc.
6633 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
6634 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
6638 *) Add -strparse option to asn1pars program which parses nested
6640 [Dr Stephen Henson <shenson@bigfoot.com>]
6642 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
6645 *) DSA fix for "ca" program.
6648 *) Added "-genkey" option to "dsaparam" program.
6651 *) Added RIPE MD160 (rmd160) message digest.
6654 *) Added -a (all) option to "ssleay version" command.
6657 *) Added PLATFORM define which is the id given to Configure.
6660 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
6663 *) Extended the ASN.1 parser routines.
6666 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
6669 *) Added a BN_CTX to the BN library.
6672 *) Fixed the weak key values in DES library
6675 *) Changed API in EVP library for cipher aliases.
6678 *) Added support for RC2/64bit cipher.
6681 *) Converted the lhash library to the crypto/mem.c functions.
6684 *) Added more recognized ASN.1 object ids.
6687 *) Added more RSA padding checks for SSL/TLS.
6690 *) Added BIO proxy/filter functionality.
6693 *) Added extra_certs to SSL_CTX which can be used
6694 send extra CA certificates to the client in the CA cert chain sending
6695 process. It can be configured with SSL_CTX_add_extra_chain_cert().
6698 *) Now Fortezza is denied in the authentication phase because
6699 this is key exchange mechanism is not supported by SSLeay at all.
6702 *) Additional PKCS1 checks.
6705 *) Support the string "TLSv1" for all TLS v1 ciphers.
6708 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
6709 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
6712 *) Fixed a few memory leaks.
6715 *) Fixed various code and comment typos.
6718 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
6719 bytes sent in the client random.
6720 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob