5 Changes between 0.9.7c and 0.9.8 [xx XXX xxxx]
7 *) Update support for ECC-based TLS ciphersuites according to
8 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
9 SHA-1 now is only used for "small" curves (where the
10 representation of a field element takes up to 24 bytes); for
11 larger curves, the field element resulting from ECDH is directly
12 used as premaster secret.
13 [Douglas Stebila (Sun Microsystems Laboratories)]
15 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
16 curve secp160r1 to the tests.
17 [Douglas Stebila (Sun Microsystems Laboratories)]
19 *) Add the possibility to load symbols globally with DSO.
20 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
22 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
23 control of the error stack.
26 *) Add support for STORE in ENGINE.
29 *) Add the STORE type. The intention is to provide a common interface
30 to certificate and key stores, be they simple file-based stores, or
31 HSM-type store, or LDAP stores, or...
32 NOTE: The code is currently UNTESTED and isn't really used anywhere.
35 *) Add a generic structure called OPENSSL_ITEM. This can be used to
36 pass a list of arguments to any function as well as provide a way
37 for a function to pass data back to the caller.
40 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
41 works like BUF_strdup() but can be used to duplicate a portion of
42 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
46 *) Add the function sk_find_ex() which works like sk_find(), but will
47 return an index to an element even if an exact match couldn't be
48 found. The index is guaranteed to point at the element where the
49 searched-for key would be inserted to preserve sorting order.
52 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
53 takes an extra flags argument for optional functionality. Currently,
54 the following flags are defined:
56 OBJ_BSEARCH_VALUE_ON_NOMATCH
57 This one gets OBJ_bsearch_ex() to return a pointer to the first
58 element where the comparing function returns a negative or zero
61 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
62 This one gets OBJ_bsearch_ex() to return a pointer to the first
63 element where the comparing function returns zero. This is useful
64 if there are more than one element where the comparing function
68 *) Make it possible to create self-signed certificates with 'openssl ca'
69 in such a way that the self-signed certificate becomes part of the
70 CA database and uses the same mechanisms for serial number generation
71 as all other certificate signing. The new flag '-selfsign' enables
72 this functionality. Adapt CA.sh and CA.pl.in.
75 *) Add functionality to check the public key of a certificate request
76 against a given private. This is useful to check that a certificate
77 request can be signed by that key (self-signing).
80 *) Make it possible to have multiple active certificates with the same
81 subject in the CA index file. This is done only if the keyword
82 'unique_subject' is set to 'no' in the main CA section (default
83 if 'CA_default') of the configuration file. The value is saved
84 with the database itself in a separate index attribute file,
85 named like the index file with '.attr' appended to the name.
88 *) Generate muti valued AVAs using '+' notation in config files for
92 *) Support for nameConstraints certificate extension.
95 *) Support for policyConstraints certificate extension.
98 *) Support for policyMappings certificate extension.
101 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
102 ENGINE as defaults for all supported algorithms irrespective of
103 the 'flags' parameter. 'flags' is now honoured, so applications
104 should make sure they are passing it correctly.
107 *) Make sure the default DSA_METHOD implementation only uses its
108 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
109 and change its own handlers to be NULL so as to remove unnecessary
110 indirection. This lets alternative implementations fallback to the
111 default implementation more easily.
114 *) Support for directoryName in GeneralName related extensions
118 *) Make it possible to link applications using Makefile.shared.
119 Make that possible even when linking against static libraries!
122 *) Support for single pass processing for S/MIME signing. This now
123 means that S/MIME signing can be done from a pipe, in addition
124 cleartext signing (multipart/signed type) is effectively streaming
125 and the signed data does not need to be all held in memory.
127 This is done with a new flag PKCS7_STREAM. When this flag is set
128 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
129 is done after the data is output (and digests calculated) in
133 *) Add full support for -rpath/-R, both in shared libraries and
134 applications, at least on the platforms where it's known how
138 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
139 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
140 will now compute a table of multiples of the generator that
141 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
142 faster (notably in the case of a single point multiplication,
144 [Nils Larsch, Bodo Moeller]
146 *) IPv6 support for certificate extensions. The various extensions
147 which use the IP:a.b.c.d can now take IPv6 addresses using the
148 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
152 *) Added an ENGINE that implements RSA by performing private key
153 exponentiations with the GMP library. The conversions to and from
154 GMP's mpz_t format aren't optimised nor are any montgomery forms
155 cached, and on x86 it appears OpenSSL's own performance has caught up.
156 However there are likely to be other architectures where GMP could
157 provide a boost. This ENGINE is not built in by default, but it can be
158 specified at Configure time and should be accompanied by the necessary
159 linker additions, eg;
160 ./config -DOPENSSL_USE_GMP -lgmp
163 *) "openssl engine" will not display ENGINE/DSO load failure errors when
164 testing availability of engines with "-t" - the old behaviour is
165 produced by increasing the feature's verbosity with "-tt".
168 *) ECDSA routines: under certain error conditions uninitialized BN objects
169 could be freed. Solution: make sure initialization is performed early
170 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
174 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
175 and DH_METHOD (eg. by ENGINE implementations) to override the normal
176 software implementations. For DSA and DH, parameter generation can
177 also be overriden by providing the appropriate method callbacks.
180 *) Change the "progress" mechanism used in key-generation and
181 primality testing to functions that take a new BN_GENCB pointer in
182 place of callback/argument pairs. The new API functions have "_ex"
183 postfixes and the older functions are reimplemented as wrappers for
184 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
185 declarations of the old functions to help (graceful) attempts to
186 migrate to the new functions. Also, the new key-generation API
187 functions operate on a caller-supplied key-structure and return
188 success/failure rather than returning a key or NULL - this is to
189 help make "keygen" another member function of RSA_METHOD etc.
191 Example for using the new callback interface:
193 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
197 BN_GENCB_set(&my_cb, my_callback, my_arg);
199 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
200 /* For the meaning of a, b in calls to my_callback(), see the
201 * documentation of the function that calls the callback.
202 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
203 * my_callback should return 1 if it wants BN_is_prime_ex()
204 * to continue, or 0 to stop.
209 *) Change the ZLIB compression method to be stateful, and make it
210 available to TLS with the number defined in
211 draft-ietf-tls-compression-04.txt.
214 *) Add the ASN.1 structures and functions for CertificatePair, which
215 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
217 CertificatePair ::= SEQUENCE {
218 forward [0] Certificate OPTIONAL,
219 reverse [1] Certificate OPTIONAL,
220 -- at least one of the pair shall be present -- }
222 Also implement the PEM functions to read and write certificate
223 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
225 This needed to be defined, mostly for the sake of the LDAP
226 attribute crossCertificatePair, but may prove useful elsewhere as
230 *) Make it possible to inhibit symlinking of shared libraries in
231 Makefile.shared, for Cygwin's sake.
234 *) Extend the BIGNUM API by creating new macros that behave like
237 void BN_set_sign(BIGNUM *a, int neg);
238 int BN_get_sign(const BIGNUM *a);
240 and avoid the need to access 'a->neg' directly in applications.
241 [Nils Larsch <nla@trustcenter.de>]
243 *) Implement fast modular reduction for pseudo-Mersenne primes
244 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
245 EC_GROUP_new_curve_GFp() will now automatically use this
247 [Nils Larsch <nla@trustcenter.de>]
249 *) Add new lock type (CRYPTO_LOCK_BN).
252 *) Change the ENGINE framework to automatically load engines
253 dynamically from specific directories unless they could be
254 found to already be built in or loaded. Move all the
255 current engines except for the cryptodev one to a new
257 The engines in engines/ are built as shared libraries if
258 the "shared" options was given to ./Configure or ./config.
259 Otherwise, they are inserted in libcrypto.a.
260 /usr/local/ssl/engines is the default directory for dynamic
261 engines, but that can be overriden at configure time through
262 the usual use of --prefix and/or --openssldir, and at run
263 time with the environment variable OPENSSL_ENGINES.
264 [Geoff Thorpe and Richard Levitte]
266 *) Add Makefile.shared, a helper makefile to build shared
267 libraries. Addapt Makefile.org.
270 *) Add version info to Win32 DLLs.
271 [Peter 'Luna' Runestig" <peter@runestig.com>]
273 *) Add new 'medium level' PKCS#12 API. Certificates and keys
274 can be added using this API to created arbitrary PKCS#12
275 files while avoiding the low level API.
277 New options to PKCS12_create(), key or cert can be NULL and
278 will then be omitted from the output file. The encryption
279 algorithm NIDs can be set to -1 for no encryption, the mac
280 iteration count can be set to 0 to omit the mac.
282 Enhance pkcs12 utility by making the -nokeys and -nocerts
283 options work when creating a PKCS#12 file. New option -nomac
284 to omit the mac, NONE can be set for an encryption algorithm.
285 New code is modified to use the enhanced PKCS12_create()
286 instead of the low level API.
289 *) Extend ASN1 encoder to support indefinite length constructed
290 encoding. This can output sequences tags and octet strings in
291 this form. Modify pk7_asn1.c to support indefinite length
292 encoding. This is experimental and needs additional code to
293 be useful, such as an ASN1 bio and some enhanced streaming
296 Extend template encode functionality so that tagging is passed
297 down to the template encoder.
300 *) Let 'openssl req' fail if an argument to '-newkey' is not
301 recognized instead of using RSA as a default.
304 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
305 As these are not official, they are not included in "ALL";
306 the "ECCdraft" ciphersuite group alias can be used to select them.
307 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
309 *) Add ECDH engine support.
310 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
312 *) Add ECDH in new directory crypto/ecdh/.
313 [Douglas Stebila (Sun Microsystems Laboratories)]
315 *) Let BN_rand_range() abort with an error after 100 iterations
316 without success (which indicates a broken PRNG).
319 *) Change BN_mod_sqrt() so that it verifies that the input value
320 is really the square of the return value. (Previously,
321 BN_mod_sqrt would show GIGO behaviour.)
324 *) Add named elliptic curves over binary fields from X9.62, SECG,
325 and WAP/WTLS; add OIDs that were still missing.
327 [Sheueling Chang Shantz and Douglas Stebila
328 (Sun Microsystems Laboratories)]
330 *) Extend the EC library for elliptic curves over binary fields
331 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
334 EC_GF2m_simple_method
338 EC_GROUP_new_curve_GF2m
339 EC_GROUP_set_curve_GF2m
340 EC_GROUP_get_curve_GF2m
341 EC_POINT_set_affine_coordinates_GF2m
342 EC_POINT_get_affine_coordinates_GF2m
343 EC_POINT_set_compressed_coordinates_GF2m
345 Point compression for binary fields is disabled by default for
346 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
349 As binary polynomials are represented as BIGNUMs, various members
350 of the EC_GROUP and EC_POINT data structures can be shared
351 between the implementations for prime fields and binary fields;
352 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
353 are essentially identical to their ..._GFp counterparts.
354 (For simplicity, the '..._GFp' prefix has been dropped from
355 various internal method names.)
357 An internal 'field_div' method (similar to 'field_mul' and
358 'field_sqr') has been added; this is used only for binary fields.
360 [Sheueling Chang Shantz and Douglas Stebila
361 (Sun Microsystems Laboratories)]
363 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
364 through methods ('mul', 'precompute_mult').
366 The generic implementations (now internally called 'ec_wNAF_mul'
367 and 'ec_wNAF_precomputed_mult') remain the default if these
368 methods are undefined.
370 [Sheueling Chang Shantz and Douglas Stebila
371 (Sun Microsystems Laboratories)]
373 *) New function EC_GROUP_get_degree, which is defined through
374 EC_METHOD. For curves over prime fields, this returns the bit
375 length of the modulus.
377 [Sheueling Chang Shantz and Douglas Stebila
378 (Sun Microsystems Laboratories)]
380 *) New functions EC_GROUP_dup, EC_POINT_dup.
381 (These simply call ..._new and ..._copy).
383 [Sheueling Chang Shantz and Douglas Stebila
384 (Sun Microsystems Laboratories)]
386 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
387 Polynomials are represented as BIGNUMs (where the sign bit is not
388 used) in the following functions [macros]:
391 BN_GF2m_sub [= BN_GF2m_add]
392 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
393 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
394 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
396 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
397 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
398 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
399 BN_GF2m_cmp [= BN_ucmp]
401 (Note that only the 'mod' functions are actually for fields GF(2^m).
402 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
404 For some functions, an the irreducible polynomial defining a
405 field can be given as an 'unsigned int[]' with strictly
406 decreasing elements giving the indices of those bits that are set;
407 i.e., p[] represents the polynomial
408 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
410 p[0] > p[1] > ... > p[k] = 0.
411 This applies to the following functions:
416 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
417 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
420 BN_GF2m_mod_solve_quad_arr
424 Conversion can be performed by the following functions:
429 bntest.c has additional tests for binary polynomial arithmetic.
431 Two implementations for BN_GF2m_mod_div() are available.
432 The default algorithm simply uses BN_GF2m_mod_inv() and
433 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
434 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
435 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
437 [Sheueling Chang Shantz and Douglas Stebila
438 (Sun Microsystems Laboratories)]
440 *) Add new error code 'ERR_R_DISABLED' that can be used when some
441 functionality is disabled at compile-time.
442 [Douglas Stebila <douglas.stebila@sun.com>]
444 *) Change default behaviour of 'openssl asn1parse' so that more
445 information is visible when viewing, e.g., a certificate:
447 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
448 mode the content of non-printable OCTET STRINGs is output in a
449 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
450 avoid the appearance of a printable string.
451 [Nils Larsch <nla@trustcenter.de>]
453 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
455 EC_GROUP_set_asn1_flag()
456 EC_GROUP_get_asn1_flag()
457 EC_GROUP_set_point_conversion_form()
458 EC_GROUP_get_point_conversion_form()
459 These control ASN1 encoding details:
460 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
461 has been set to OPENSSL_EC_NAMED_CURVE.
462 - Points are encoded in uncompressed form by default; options for
463 asn1_for are as for point2oct, namely
464 POINT_CONVERSION_COMPRESSED
465 POINT_CONVERSION_UNCOMPRESSED
466 POINT_CONVERSION_HYBRID
468 Also add 'seed' and 'seed_len' members to EC_GROUP with access
472 EC_GROUP_get_seed_len()
473 This is used only for ASN1 purposes (so far).
474 [Nils Larsch <nla@trustcenter.de>]
476 *) Add 'field_type' member to EC_METHOD, which holds the NID
477 of the appropriate field type OID. The new function
478 EC_METHOD_get_field_type() returns this value.
479 [Nils Larsch <nla@trustcenter.de>]
486 providing useful interfaces to EC_POINT_point2oct() and
487 EC_POINT_oct2point().
488 [Nils Larsch <nla@trustcenter.de>]
490 *) Change internals of the EC library so that the functions
491 EC_GROUP_set_generator()
492 EC_GROUP_get_generator()
494 EC_GROUP_get_cofactor()
495 are implemented directly in crypto/ec/ec_lib.c and not dispatched
496 to methods, which would lead to unnecessary code duplication when
497 adding different types of curves.
498 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
500 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
501 arithmetic, and such that modified wNAFs are generated
502 (which avoid length expansion in many cases).
505 *) Add a function EC_GROUP_check_discriminant() (defined via
506 EC_METHOD) that verifies that the curve discriminant is non-zero.
508 Add a function EC_GROUP_check() that makes some sanity tests
509 on a EC_GROUP, its generator and order. This includes
510 EC_GROUP_check_discriminant().
511 [Nils Larsch <nla@trustcenter.de>]
513 *) Add ECDSA in new directory crypto/ecdsa/.
515 Add applications 'openssl ecparam' and 'openssl ecdsa'
516 (these are based on 'openssl dsaparam' and 'openssl dsa').
518 ECDSA support is also included in various other files across the
519 library. Most notably,
520 - 'openssl req' now has a '-newkey ecdsa:file' option;
521 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
522 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
523 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
524 them suitable for ECDSA where domain parameters must be
525 extracted before the specific public key;
526 - ECDSA engine support has been added.
527 [Nils Larsch <nla@trustcenter.de>]
529 *) Include some named elliptic curves, and add OIDs from X9.62,
530 SECG, and WAP/WTLS. Each curve can be obtained from the new
532 EC_GROUP_new_by_nid(),
533 and the list of available named curves can be obtained with
534 EC_get_builtin_curves().
535 Also add a 'curve_name' member to EC_GROUP objects, which can be
539 [Nils Larsch <nla@trustcenter.de, Bodo Moeller]
541 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
542 was actually never needed) and in BN_mul(). The removal in BN_mul()
543 required a small change in bn_mul_part_recursive() and the addition
544 of the functions bn_cmp_part_words(), bn_sub_part_words() and
545 bn_add_part_words(), which do the same thing as bn_cmp_words(),
546 bn_sub_words() and bn_add_words() except they take arrays with
550 Changes between 0.9.7b and 0.9.7c [xx XXX 2003]
552 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
553 extra data after the compression methods not only for TLS 1.0
554 but also for SSL 3.0 (as required by the specification).
555 [Bodo Moeller; problem pointed out by Matthias Loepfe]
557 *) Change X509_certificate_type() to mark the key as exported/exportable
558 when it's 512 *bits* long, not 512 bytes.
561 *) Change AES_cbc_encrypt() so it outputs exact multiple of
562 blocks during encryption.
565 *) Various fixes to base64 BIO and non blocking I/O. On write
566 flushes were not handled properly if the BIO retried. On read
567 data was not being buffered properly and had various logic bugs.
568 This also affects blocking I/O when the data being decoded is a
572 *) Various S/MIME bugfixes and compatibility changes:
573 output correct application/pkcs7 MIME type if
574 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
575 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
576 of files as .eml work). Correctly handle very long lines in MIME
580 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
582 *) Countermeasure against the Klima-Pokorny-Rosa extension of
583 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
584 a protocol version number mismatch like a decryption error
585 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
588 *) Turn on RSA blinding by default in the default implementation
589 to avoid a timing attack. Applications that don't want it can call
590 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
591 They would be ill-advised to do so in most cases.
592 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
594 *) Change RSA blinding code so that it works when the PRNG is not
595 seeded (in this case, the secret RSA exponent is abused as
596 an unpredictable seed -- if it is not unpredictable, there
597 is no point in blinding anyway). Make RSA blinding thread-safe
598 by remembering the creator's thread ID in rsa->blinding and
599 having all other threads use local one-time blinding factors
600 (this requires more computation than sharing rsa->blinding, but
601 avoids excessive locking; and if an RSA object is not shared
602 between threads, blinding will still be very fast).
605 yet to be integrated into this CVS branch:
606 - Geoff's ENGINE_set_default() fix
608 *) Target "mingw" now allows native Windows code to be generated in
609 the Cygwin environment as well as with the MinGW compiler.
612 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
614 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
615 via timing by performing a MAC computation even if incorrrect
616 block cipher padding has been found. This is a countermeasure
617 against active attacks where the attacker has to distinguish
618 between bad padding and a MAC verification error. (CAN-2003-0078)
620 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
621 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
622 Martin Vuagnoux (EPFL, Ilion)]
624 *) Make the no-err option work as intended. The intention with no-err
625 is not to have the whole error stack handling routines removed from
626 libcrypto, it's only intended to remove all the function name and
627 reason texts, thereby removing some of the footprint that may not
628 be interesting if those errors aren't displayed anyway.
630 NOTE: it's still possible for any application or module to have it's
631 own set of error texts inserted. The routines are there, just not
632 used by default when no-err is given.
635 *) Add support for FreeBSD on IA64.
636 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
638 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
639 Kerberos function mit_des_cbc_cksum(). Before this change,
640 the value returned by DES_cbc_cksum() was like the one from
641 mit_des_cbc_cksum(), except the bytes were swapped.
642 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
644 *) Allow an application to disable the automatic SSL chain building.
645 Before this a rather primitive chain build was always performed in
646 ssl3_output_cert_chain(): an application had no way to send the
647 correct chain if the automatic operation produced an incorrect result.
649 Now the chain builder is disabled if either:
651 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
653 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
655 The reasoning behind this is that an application would not want the
656 auto chain building to take place if extra chain certificates are
657 present and it might also want a means of sending no additional
658 certificates (for example the chain has two certificates and the
662 *) Add the possibility to build without the ENGINE framework.
663 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
665 *) Under Win32 gmtime() can return NULL: check return value in
666 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
669 *) DSA routines: under certain error conditions uninitialized BN objects
670 could be freed. Solution: make sure initialization is performed early
671 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
672 Nils Larsch <nla@trustcenter.de> via PR#459)
675 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
676 checked on reconnect on the client side, therefore session resumption
677 could still fail with a "ssl session id is different" error. This
678 behaviour is masked when SSL_OP_ALL is used due to
679 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
680 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
684 *) IA-32 assembler support enhancements: unified ELF targets, support
685 for SCO/Caldera platforms, fix for Cygwin shared build.
688 *) Add support for FreeBSD on sparc64. As a consequence, support for
689 FreeBSD on non-x86 processors is separate from x86 processors on
690 the config script, much like the NetBSD support.
691 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
693 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
695 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
698 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
699 code (06) was taken as the first octet of the session ID and the last
700 octet was ignored consequently. As a result SSLv2 client side session
701 caching could not have worked due to the session ID mismatch between
703 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
707 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
708 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
712 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
713 seems that in spite of existing for more than a year, many application
714 author have done nothing to provide the necessary callbacks, which
715 means that this particular engine will not work properly anywhere.
716 This is a very unfortunate situation which forces us, in the name
717 of usability, to give the hw_ncipher.c a static lock, which is part
719 NOTE: This is for the 0.9.7 series ONLY. This hack will never
720 appear in 0.9.8 or later. We EXPECT application authors to have
721 dealt properly with this when 0.9.8 is released (unless we actually
722 make such changes in the libcrypto locking code that changes will
723 have to be made anyway).
726 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
727 octets have been read, EOF or an error occurs. Without this change
728 some truncated ASN1 structures will not produce an error.
731 *) Disable Heimdal support, since it hasn't been fully implemented.
732 Still give the possibility to force the use of Heimdal, but with
733 warnings and a request that patches get sent to openssl-dev.
736 *) Add the VC-CE target, introduce the WINCE sysname, and add
737 INSTALL.WCE and appropriate conditionals to make it build.
738 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
740 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
741 cygssl-x.y.z.dll, where x, y and z are the major, minor and
742 edit numbers of the version.
743 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
745 *) Introduce safe string copy and catenation functions
746 (BUF_strlcpy() and BUF_strlcat()).
747 [Ben Laurie (CHATS) and Richard Levitte]
749 *) Avoid using fixed-size buffers for one-line DNs.
752 *) Add BUF_MEM_grow_clean() to avoid information leakage when
753 resizing buffers containing secrets, and use where appropriate.
756 *) Avoid using fixed size buffers for configuration file location.
759 *) Avoid filename truncation for various CA files.
762 *) Use sizeof in preference to magic numbers.
765 *) Avoid filename truncation in cert requests.
768 *) Add assertions to check for (supposedly impossible) buffer
772 *) Don't cache truncated DNS entries in the local cache (this could
773 potentially lead to a spoofing attack).
776 *) Fix various buffers to be large enough for hex/decimal
777 representations in a platform independent manner.
780 *) Add CRYPTO_realloc_clean() to avoid information leakage when
781 resizing buffers containing secrets, and use where appropriate.
784 *) Add BIO_indent() to avoid much slightly worrying code to do
788 *) Convert sprintf()/BIO_puts() to BIO_printf().
791 *) buffer_gets() could terminate with the buffer only half
795 *) Add assertions to prevent user-supplied crypto functions from
796 overflowing internal buffers by having large block sizes, etc.
799 *) New OPENSSL_assert() macro (similar to assert(), but enabled
803 *) Eliminate unused copy of key in RC4.
806 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
809 *) Fix off-by-one error in EGD path.
812 *) If RANDFILE path is too long, ignore instead of truncating.
815 *) Eliminate unused and incorrectly sized X.509 structure
819 *) Eliminate unused and dangerous function knumber().
822 *) Eliminate unused and dangerous structure, KSSL_ERR.
825 *) Protect against overlong session ID context length in an encoded
826 session object. Since these are local, this does not appear to be
830 *) Change from security patch (see 0.9.6e below) that did not affect
831 the 0.9.6 release series:
833 Remote buffer overflow in SSL3 protocol - an attacker could
834 supply an oversized master key in Kerberos-enabled versions.
838 *) Change the SSL kerb5 codes to match RFC 2712.
841 *) Make -nameopt work fully for req and add -reqopt switch.
842 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
844 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
845 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
847 *) Make sure tests can be performed even if the corresponding algorithms
848 have been removed entirely. This was also the last step to make
849 OpenSSL compilable with DJGPP under all reasonable conditions.
850 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
852 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
853 to allow version independent disabling of normally unselected ciphers,
854 which may be activated as a side-effect of selecting a single cipher.
856 (E.g., cipher list string "RSA" enables ciphersuites that are left
857 out of "ALL" because they do not provide symmetric encryption.
858 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
859 [Lutz Jaenicke, Bodo Moeller]
861 *) Add appropriate support for separate platform-dependent build
862 directories. The recommended way to make a platform-dependent
863 build directory is the following (tested on Linux), maybe with
866 # Place yourself outside of the OpenSSL source tree. In
867 # this example, the environment variable OPENSSL_SOURCE
868 # is assumed to contain the absolute OpenSSL source directory.
869 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
870 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
871 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
872 mkdir -p `dirname $F`
873 ln -s $OPENSSL_SOURCE/$F $F
876 To be absolutely sure not to disturb the source tree, a "make clean"
877 is a good thing. If it isn't successfull, don't worry about it,
878 it probably means the source directory is very clean.
881 *) Make sure any ENGINE control commands make local copies of string
882 pointers passed to them whenever necessary. Otherwise it is possible
883 the caller may have overwritten (or deallocated) the original string
884 data when a later ENGINE operation tries to use the stored values.
885 [Götz Babin-Ebell <babinebell@trustcenter.de>]
887 *) Improve diagnostics in file reading and command-line digests.
888 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
890 *) Add AES modes CFB and OFB to the object database. Correct an
891 error in AES-CFB decryption.
894 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
895 allows existing EVP_CIPHER_CTX structures to be reused after
896 calling EVP_*Final(). This behaviour is used by encryption
897 BIOs and some applications. This has the side effect that
898 applications must explicitly clean up cipher contexts with
899 EVP_CIPHER_CTX_cleanup() or they will leak memory.
902 *) Check the values of dna and dnb in bn_mul_recursive before calling
903 bn_mul_comba (a non zero value means the a or b arrays do not contain
904 n2 elements) and fallback to bn_mul_normal if either is not zero.
907 *) Fix escaping of non-ASCII characters when using the -subj option
908 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
911 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
912 form for "surname", serialNumber has no short form.
913 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
914 therefore remove "mail" short name for "internet 7".
915 The OID for unique identifiers in X509 certificates is
916 x500UniqueIdentifier, not uniqueIdentifier.
917 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
920 *) Add an "init" command to the ENGINE config module and auto initialize
921 ENGINEs. Without any "init" command the ENGINE will be initialized
922 after all ctrl commands have been executed on it. If init=1 the
923 ENGINE is initailized at that point (ctrls before that point are run
924 on the uninitialized ENGINE and after on the initialized one). If
925 init=0 then the ENGINE will not be iniatialized at all.
928 *) Fix the 'app_verify_callback' interface so that the user-defined
929 argument is actually passed to the callback: In the
930 SSL_CTX_set_cert_verify_callback() prototype, the callback
931 declaration has been changed from
934 int (*cb)(X509_STORE_CTX *,void *);
935 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
936 i=s->ctx->app_verify_callback(&ctx)
937 has been changed into
938 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
940 To update applications using SSL_CTX_set_cert_verify_callback(),
941 a dummy argument can be added to their callback functions.
942 [D. K. Smetters <smetters@parc.xerox.com>]
944 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
945 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
947 *) Add and OPENSSL_LOAD_CONF define which will cause
948 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
949 This allows older applications to transparently support certain
950 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
951 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
952 load the config file and OPENSSL_add_all_algorithms_conf() which will
953 always load it have also been added.
956 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
957 Adjust NIDs and EVP layer.
958 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
960 *) Config modules support in openssl utility.
962 Most commands now load modules from the config file,
963 though in a few (such as version) this isn't done
964 because it couldn't be used for anything.
966 In the case of ca and req the config file used is
967 the same as the utility itself: that is the -config
968 command line option can be used to specify an
972 *) Move default behaviour from OPENSSL_config(). If appname is NULL
973 use "openssl_conf" if filename is NULL use default openssl config file.
976 *) Add an argument to OPENSSL_config() to allow the use of an alternative
977 config section name. Add a new flag to tolerate a missing config file
978 and move code to CONF_modules_load_file().
981 *) Support for crypto accelerator cards from Accelerated Encryption
982 Processing, www.aep.ie. (Use engine 'aep')
983 The support was copied from 0.9.6c [engine] and adapted/corrected
984 to work with the new engine framework.
985 [AEP Inc. and Richard Levitte]
987 *) Support for SureWare crypto accelerator cards from Baltimore
988 Technologies. (Use engine 'sureware')
989 The support was copied from 0.9.6c [engine] and adapted
990 to work with the new engine framework.
993 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
994 make the newer ENGINE framework commands for the CHIL engine work.
995 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
997 *) Make it possible to produce shared libraries on ReliantUNIX.
998 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
1000 *) Add the configuration target debug-linux-ppro.
1001 Make 'openssl rsa' use the general key loading routines
1002 implemented in apps.c, and make those routines able to
1003 handle the key format FORMAT_NETSCAPE and the variant
1005 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1007 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
1008 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1010 *) Add -keyform to rsautl, and document -engine.
1011 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
1013 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
1014 BIO_R_NO_SUCH_FILE error code rather than the generic
1015 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
1018 *) Add new functions
1020 ERR_peek_last_error_line
1021 ERR_peek_last_error_line_data.
1022 These are similar to
1025 ERR_peek_error_line_data,
1026 but report on the latest error recorded rather than the first one
1027 still in the error queue.
1028 [Ben Laurie, Bodo Moeller]
1030 *) default_algorithms option in ENGINE config module. This allows things
1032 default_algorithms = ALL
1033 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
1036 *) Prelminary ENGINE config module.
1039 *) New experimental application configuration code.
1042 *) Change the AES code to follow the same name structure as all other
1043 symmetric ciphers, and behave the same way. Move everything to
1044 the directory crypto/aes, thereby obsoleting crypto/rijndael.
1045 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1047 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
1048 [Ben Laurie and Theo de Raadt]
1050 *) Add option to output public keys in req command.
1051 [Massimiliano Pala madwolf@openca.org]
1053 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
1054 (up to about 10% better than before for P-192 and P-224).
1057 *) New functions/macros
1059 SSL_CTX_set_msg_callback(ctx, cb)
1060 SSL_CTX_set_msg_callback_arg(ctx, arg)
1061 SSL_set_msg_callback(ssl, cb)
1062 SSL_set_msg_callback_arg(ssl, arg)
1064 to request calling a callback function
1066 void cb(int write_p, int version, int content_type,
1067 const void *buf, size_t len, SSL *ssl, void *arg)
1069 whenever a protocol message has been completely received
1070 (write_p == 0) or sent (write_p == 1). Here 'version' is the
1071 protocol version according to which the SSL library interprets
1072 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
1073 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
1074 the content type as defined in the SSL 3.0/TLS 1.0 protocol
1075 specification (change_cipher_spec(20), alert(21), handshake(22)).
1076 'buf' and 'len' point to the actual message, 'ssl' to the
1077 SSL object, and 'arg' is the application-defined value set by
1078 SSL[_CTX]_set_msg_callback_arg().
1080 'openssl s_client' and 'openssl s_server' have new '-msg' options
1081 to enable a callback that displays all protocol messages.
1084 *) Change the shared library support so shared libraries are built as
1085 soon as the corresponding static library is finished, and thereby get
1086 openssl and the test programs linked against the shared library.
1087 This still only happens when the keyword "shard" has been given to
1088 the configuration scripts.
1090 NOTE: shared library support is still an experimental thing, and
1091 backward binary compatibility is still not guaranteed.
1092 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
1094 *) Add support for Subject Information Access extension.
1095 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1097 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
1098 additional bytes when new memory had to be allocated, not just
1099 when reusing an existing buffer.
1102 *) New command line and configuration option 'utf8' for the req command.
1103 This allows field values to be specified as UTF8 strings.
1106 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
1107 runs for the former and machine-readable output for the latter.
1110 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
1111 of the e-mail address in the DN (i.e., it will go into a certificate
1112 extension only). The new configuration file option 'email_in_dn = no'
1113 has the same effect.
1114 [Massimiliano Pala madwolf@openca.org]
1116 *) Change all functions with names starting with des_ to be starting
1117 with DES_ instead. Add wrappers that are compatible with libdes,
1118 but are named _ossl_old_des_*. Finally, add macros that map the
1119 des_* symbols to the corresponding _ossl_old_des_* if libdes
1120 compatibility is desired. If OpenSSL 0.9.6c compatibility is
1121 desired, the des_* symbols will be mapped to DES_*, with one
1124 Since we provide two compatibility mappings, the user needs to
1125 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
1126 compatibility is desired. The default (i.e., when that macro
1127 isn't defined) is OpenSSL 0.9.6c compatibility.
1129 There are also macros that enable and disable the support of old
1130 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
1131 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
1132 are defined, the default will apply: to support the old des routines.
1134 In either case, one must include openssl/des.h to get the correct
1135 definitions. Do not try to just include openssl/des_old.h, that
1138 NOTE: This is a major break of an old API into a new one. Software
1139 authors are encouraged to switch to the DES_ style functions. Some
1140 time in the future, des_old.h and the libdes compatibility functions
1141 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
1142 default), and then completely removed.
1145 *) Test for certificates which contain unsupported critical extensions.
1146 If such a certificate is found during a verify operation it is
1147 rejected by default: this behaviour can be overridden by either
1148 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
1149 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
1150 X509_supported_extension() has also been added which returns 1 if a
1151 particular extension is supported.
1154 *) Modify the behaviour of EVP cipher functions in similar way to digests
1155 to retain compatibility with existing code.
1158 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
1159 compatibility with existing code. In particular the 'ctx' parameter does
1160 not have to be to be initialized before the call to EVP_DigestInit() and
1161 it is tidied up after a call to EVP_DigestFinal(). New function
1162 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
1163 EVP_MD_CTX_copy() changed to not require the destination to be
1164 initialized valid and new function EVP_MD_CTX_copy_ex() added which
1165 requires the destination to be valid.
1167 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
1168 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
1171 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
1172 so that complete 'Handshake' protocol structures are kept in memory
1173 instead of overwriting 'msg_type' and 'length' with 'body' data.
1176 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
1177 [Massimo Santin via Richard Levitte]
1179 *) Major restructuring to the underlying ENGINE code. This includes
1180 reduction of linker bloat, separation of pure "ENGINE" manipulation
1181 (initialisation, etc) from functionality dealing with implementations
1182 of specific crypto iterfaces. This change also introduces integrated
1183 support for symmetric ciphers and digest implementations - so ENGINEs
1184 can now accelerate these by providing EVP_CIPHER and EVP_MD
1185 implementations of their own. This is detailed in crypto/engine/README
1186 as it couldn't be adequately described here. However, there are a few
1187 API changes worth noting - some RSA, DSA, DH, and RAND functions that
1188 were changed in the original introduction of ENGINE code have now
1189 reverted back - the hooking from this code to ENGINE is now a good
1190 deal more passive and at run-time, operations deal directly with
1191 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
1192 dereferencing through an ENGINE pointer any more. Also, the ENGINE
1193 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
1194 they were not being used by the framework as there is no concept of a
1195 BIGNUM_METHOD and they could not be generalised to the new
1196 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
1197 ENGINE_cpy() has been removed as it cannot be consistently defined in
1201 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
1204 *) Change mkdef.pl to sort symbols that get the same entry number,
1205 and make sure the automatically generated functions ERR_load_*
1206 become part of libeay.num as well.
1209 *) New function SSL_renegotiate_pending(). This returns true once
1210 renegotiation has been requested (either SSL_renegotiate() call
1211 or HelloRequest/ClientHello receveived from the peer) and becomes
1212 false once a handshake has been completed.
1213 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
1214 sends a HelloRequest, but does not ensure that a handshake takes
1215 place. SSL_renegotiate_pending() is useful for checking if the
1216 client has followed the request.)
1219 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
1220 By default, clients may request session resumption even during
1221 renegotiation (if session ID contexts permit); with this option,
1222 session resumption is possible only in the first handshake.
1224 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
1225 more bits available for options that should not be part of
1226 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
1229 *) Add some demos for certificate and certificate request creation.
1232 *) Make maximum certificate chain size accepted from the peer application
1233 settable (SSL*_get/set_max_cert_list()), as proposed by
1234 "Douglas E. Engert" <deengert@anl.gov>.
1237 *) Add support for shared libraries for Unixware-7
1238 (Boyd Lynn Gerber <gerberb@zenez.com>).
1241 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
1242 be done prior to destruction. Use this to unload error strings from
1243 ENGINEs that load their own error strings. NB: This adds two new API
1244 functions to "get" and "set" this destroy handler in an ENGINE.
1247 *) Alter all existing ENGINE implementations (except "openssl" and
1248 "openbsd") to dynamically instantiate their own error strings. This
1249 makes them more flexible to be built both as statically-linked ENGINEs
1250 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
1251 Also, add stub code to each that makes building them as self-contained
1252 shared-libraries easier (see README.ENGINE).
1255 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
1256 implementations into applications that are completely implemented in
1257 self-contained shared-libraries. The "dynamic" ENGINE exposes control
1258 commands that can be used to configure what shared-library to load and
1259 to control aspects of the way it is handled. Also, made an update to
1260 the README.ENGINE file that brings its information up-to-date and
1261 provides some information and instructions on the "dynamic" ENGINE
1262 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
1265 *) Make it possible to unload ranges of ERR strings with a new
1266 "ERR_unload_strings" function.
1269 *) Add a copy() function to EVP_MD.
1272 *) Make EVP_MD routines take a context pointer instead of just the
1273 md_data void pointer.
1276 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
1277 that the digest can only process a single chunk of data
1278 (typically because it is provided by a piece of
1279 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
1280 is only going to provide a single chunk of data, and hence the
1281 framework needn't accumulate the data for oneshot drivers.
1284 *) As with "ERR", make it possible to replace the underlying "ex_data"
1285 functions. This change also alters the storage and management of global
1286 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
1287 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
1288 index counters. The API functions that use this state have been changed
1289 to take a "class_index" rather than pointers to the class's local STACK
1290 and counter, and there is now an API function to dynamically create new
1291 classes. This centralisation allows us to (a) plug a lot of the
1292 thread-safety problems that existed, and (b) makes it possible to clean
1293 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
1294 such data would previously have always leaked in application code and
1295 workarounds were in place to make the memory debugging turn a blind eye
1296 to it. Application code that doesn't use this new function will still
1297 leak as before, but their memory debugging output will announce it now
1298 rather than letting it slide.
1300 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
1301 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
1302 has a return value to indicate success or failure.
1305 *) Make it possible to replace the underlying "ERR" functions such that the
1306 global state (2 LHASH tables and 2 locks) is only used by the "default"
1307 implementation. This change also adds two functions to "get" and "set"
1308 the implementation prior to it being automatically set the first time
1309 any other ERR function takes place. Ie. an application can call "get",
1310 pass the return value to a module it has just loaded, and that module
1311 can call its own "set" function using that value. This means the
1312 module's "ERR" operations will use (and modify) the error state in the
1313 application and not in its own statically linked copy of OpenSSL code.
1316 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
1317 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
1318 the operation, and provides a more encapsulated way for external code
1319 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
1320 to use these functions rather than manually incrementing the counts.
1322 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
1325 *) Add EVP test program.
1328 *) Add symmetric cipher support to ENGINE. Expect the API to change!
1331 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
1332 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
1333 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
1334 These allow a CRL to be built without having to access X509_CRL fields
1335 directly. Modify 'ca' application to use new functions.
1338 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
1339 bug workarounds. Rollback attack detection is a security feature.
1340 The problem will only arise on OpenSSL servers when TLSv1 is not
1341 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
1342 Software authors not wanting to support TLSv1 will have special reasons
1343 for their choice and can explicitly enable this option.
1344 [Bodo Moeller, Lutz Jaenicke]
1346 *) Rationalise EVP so it can be extended: don't include a union of
1347 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
1348 (similar to those existing for EVP_CIPHER_CTX).
1353 EVP_MD_CTX_init(&md); /* new function call */
1354 EVP_DigestInit(&md, EVP_sha1());
1355 EVP_DigestUpdate(&md, in, len);
1356 EVP_DigestFinal(&md, out, NULL);
1357 EVP_MD_CTX_cleanup(&md); /* new function call */
1361 *) Make DES key schedule conform to the usual scheme, as well as
1362 correcting its structure. This means that calls to DES functions
1363 now have to pass a pointer to a des_key_schedule instead of a
1364 plain des_key_schedule (which was actually always a pointer
1367 des_key_schedule ks;
1369 des_set_key_checked(..., &ks);
1370 des_ncbc_encrypt(..., &ks, ...);
1372 (Note that a later change renames 'des_...' into 'DES_...'.)
1375 *) Initial reduction of linker bloat: the use of some functions, such as
1376 PEM causes large amounts of unused functions to be linked in due to
1377 poor organisation. For example pem_all.c contains every PEM function
1378 which has a knock on effect of linking in large amounts of (unused)
1379 ASN1 code. Grouping together similar functions and splitting unrelated
1380 functions prevents this.
1383 *) Cleanup of EVP macros.
1386 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
1387 correct _ecb suffix.
1390 *) Add initial OCSP responder support to ocsp application. The
1391 revocation information is handled using the text based index
1392 use by the ca application. The responder can either handle
1393 requests generated internally, supplied in files (for example
1394 via a CGI script) or using an internal minimal server.
1397 *) Add configuration choices to get zlib compression for TLS.
1400 *) Changes to Kerberos SSL for RFC 2712 compliance:
1401 1. Implemented real KerberosWrapper, instead of just using
1402 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
1403 2. Implemented optional authenticator field of KerberosWrapper.
1405 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
1406 and authenticator structs; see crypto/krb5/.
1408 Generalized Kerberos calls to support multiple Kerberos libraries.
1409 [Vern Staats <staatsvr@asc.hpc.mil>,
1410 Jeffrey Altman <jaltman@columbia.edu>
1411 via Richard Levitte]
1413 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
1414 already does with RSA. testdsa.h now has 'priv_key/pub_key'
1415 values for each of the key sizes rather than having just
1416 parameters (and 'speed' generating keys each time).
1419 *) Speed up EVP routines.
1422 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
1423 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
1424 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
1425 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
1427 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
1428 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
1429 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
1432 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
1434 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
1437 *) Added the OS2-EMX target.
1438 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
1440 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
1441 to support NCONF routines in extension code. New function CONF_set_nconf()
1442 to allow functions which take an NCONF to also handle the old LHASH
1443 structure: this means that the old CONF compatible routines can be
1444 retained (in particular wrt extensions) without having to duplicate the
1445 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
1448 *) Enhance the general user interface with mechanisms for inner control
1449 and with possibilities to have yes/no kind of prompts.
1452 *) Change all calls to low level digest routines in the library and
1453 applications to use EVP. Add missing calls to HMAC_cleanup() and
1454 don't assume HMAC_CTX can be copied using memcpy().
1455 [Verdon Walker <VWalker@novell.com>, Steve Henson]
1457 *) Add the possibility to control engines through control names but with
1458 arbitrary arguments instead of just a string.
1459 Change the key loaders to take a UI_METHOD instead of a callback
1460 function pointer. NOTE: this breaks binary compatibility with earlier
1461 versions of OpenSSL [engine].
1462 Adapt the nCipher code for these new conditions and add a card insertion
1466 *) Enhance the general user interface with mechanisms to better support
1467 dialog box interfaces, application-defined prompts, the possibility
1468 to use defaults (for example default passwords from somewhere else)
1469 and interrupts/cancellations.
1472 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
1473 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
1476 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
1477 tidy up some unnecessarily weird code in 'sk_new()').
1478 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
1480 *) Change the key loading routines for ENGINEs to use the same kind
1481 callback (pem_password_cb) as all other routines that need this
1485 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
1486 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
1487 than this minimum value is recommended.
1490 *) New random seeder for OpenVMS, using the system process statistics
1491 that are easily reachable.
1494 *) Windows apparently can't transparently handle global
1495 variables defined in DLLs. Initialisations such as:
1497 const ASN1_ITEM *it = &ASN1_INTEGER_it;
1499 wont compile. This is used by the any applications that need to
1500 declare their own ASN1 modules. This was fixed by adding the option
1501 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
1502 needed for static libraries under Win32.
1505 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
1506 setting of purpose and trust fields. New X509_STORE trust and
1507 purpose functions and tidy up setting in other SSL functions.
1510 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
1511 structure. These are inherited by X509_STORE_CTX when it is
1512 initialised. This allows various defaults to be set in the
1513 X509_STORE structure (such as flags for CRL checking and custom
1514 purpose or trust settings) for functions which only use X509_STORE_CTX
1515 internally such as S/MIME.
1517 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
1518 trust settings if they are not set in X509_STORE. This allows X509_STORE
1519 purposes and trust (in S/MIME for example) to override any set by default.
1521 Add command line options for CRL checking to smime, s_client and s_server
1525 *) Initial CRL based revocation checking. If the CRL checking flag(s)
1526 are set then the CRL is looked up in the X509_STORE structure and
1527 its validity and signature checked, then if the certificate is found
1528 in the CRL the verify fails with a revoked error.
1530 Various new CRL related callbacks added to X509_STORE_CTX structure.
1532 Command line options added to 'verify' application to support this.
1534 This needs some additional work, such as being able to handle multiple
1535 CRLs with different times, extension based lookup (rather than just
1536 by subject name) and ultimately more complete V2 CRL extension
1540 *) Add a general user interface API (crypto/ui/). This is designed
1541 to replace things like des_read_password and friends (backward
1542 compatibility functions using this new API are provided).
1543 The purpose is to remove prompting functions from the DES code
1544 section as well as provide for prompting through dialog boxes in
1545 a window system and the like.
1548 *) Add "ex_data" support to ENGINE so implementations can add state at a
1549 per-structure level rather than having to store it globally.
1552 *) Make it possible for ENGINE structures to be copied when retrieved by
1553 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
1554 This causes the "original" ENGINE structure to act like a template,
1555 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
1556 operational state can be localised to each ENGINE structure, despite the
1557 fact they all share the same "methods". New ENGINE structures returned in
1558 this case have no functional references and the return value is the single
1559 structural reference. This matches the single structural reference returned
1560 by ENGINE_by_id() normally, when it is incremented on the pre-existing
1564 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
1565 needs to match any other type at all we need to manually clear the
1569 *) Changes to the "openssl engine" utility to include;
1570 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
1571 about an ENGINE's available control commands.
1572 - executing control commands from command line arguments using the
1573 '-pre' and '-post' switches. '-post' is only used if '-t' is
1574 specified and the ENGINE is successfully initialised. The syntax for
1575 the individual commands are colon-separated, for example;
1576 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
1579 *) New dynamic control command support for ENGINEs. ENGINEs can now
1580 declare their own commands (numbers), names (strings), descriptions,
1581 and input types for run-time discovery by calling applications. A
1582 subset of these commands are implicitly classed as "executable"
1583 depending on their input type, and only these can be invoked through
1584 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
1585 can be based on user input, config files, etc). The distinction is
1586 that "executable" commands cannot return anything other than a boolean
1587 result and can only support numeric or string input, whereas some
1588 discoverable commands may only be for direct use through
1589 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
1590 pointers, or other custom uses. The "executable" commands are to
1591 support parameterisations of ENGINE behaviour that can be
1592 unambiguously defined by ENGINEs and used consistently across any
1593 OpenSSL-based application. Commands have been added to all the
1594 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
1595 control over shared-library paths without source code alterations.
1598 *) Changed all ENGINE implementations to dynamically allocate their
1599 ENGINEs rather than declaring them statically. Apart from this being
1600 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
1601 this also allows the implementations to compile without using the
1602 internal engine_int.h header.
1605 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
1606 'const' value. Any code that should be able to modify a RAND_METHOD
1607 should already have non-const pointers to it (ie. they should only
1608 modify their own ones).
1611 *) Made a variety of little tweaks to the ENGINE code.
1612 - "atalla" and "ubsec" string definitions were moved from header files
1613 to C code. "nuron" string definitions were placed in variables
1614 rather than hard-coded - allowing parameterisation of these values
1615 later on via ctrl() commands.
1616 - Removed unused "#if 0"'d code.
1617 - Fixed engine list iteration code so it uses ENGINE_free() to release
1618 structural references.
1619 - Constified the RAND_METHOD element of ENGINE structures.
1620 - Constified various get/set functions as appropriate and added
1621 missing functions (including a catch-all ENGINE_cpy that duplicates
1622 all ENGINE values onto a new ENGINE except reference counts/state).
1623 - Removed NULL parameter checks in get/set functions. Setting a method
1624 or function to NULL is a way of cancelling out a previously set
1625 value. Passing a NULL ENGINE parameter is just plain stupid anyway
1626 and doesn't justify the extra error symbols and code.
1627 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
1628 flags from engine_int.h to engine.h.
1629 - Changed prototypes for ENGINE handler functions (init(), finish(),
1630 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
1633 *) Implement binary inversion algorithm for BN_mod_inverse in addition
1634 to the algorithm using long division. The binary algorithm can be
1635 used only if the modulus is odd. On 32-bit systems, it is faster
1636 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
1637 roughly 5-15% for 256-bit moduli), so we use it only for moduli
1638 up to 450 bits. In 64-bit environments, the binary algorithm
1639 appears to be advantageous for much longer moduli; here we use it
1640 for moduli up to 2048 bits.
1643 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
1644 could not support the combine flag in choice fields.
1647 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
1648 extensions from a certificate request to the certificate.
1651 *) Allow multiple 'certopt' and 'nameopt' options to be separated
1652 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
1653 file: this allows the display of the certificate about to be
1654 signed to be customised, to allow certain fields to be included
1655 or excluded and extension details. The old system didn't display
1656 multicharacter strings properly, omitted fields not in the policy
1657 and couldn't display additional details such as extensions.
1660 *) Function EC_POINTs_mul for multiple scalar multiplication
1661 of an arbitrary number of elliptic curve points
1662 \sum scalars[i]*points[i],
1663 optionally including the generator defined for the EC_GROUP:
1664 scalar*generator + \sum scalars[i]*points[i].
1666 EC_POINT_mul is a simple wrapper function for the typical case
1667 that the point list has just one item (besides the optional
1671 *) First EC_METHODs for curves over GF(p):
1673 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
1674 operations and provides various method functions that can also
1675 operate with faster implementations of modular arithmetic.
1677 EC_GFp_mont_method() reuses most functions that are part of
1678 EC_GFp_simple_method, but uses Montgomery arithmetic.
1680 [Bodo Moeller; point addition and point doubling
1681 implementation directly derived from source code provided by
1682 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
1684 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
1685 crypto/ec/ec_lib.c):
1687 Curves are EC_GROUP objects (with an optional group generator)
1688 based on EC_METHODs that are built into the library.
1690 Points are EC_POINT objects based on EC_GROUP objects.
1692 Most of the framework would be able to handle curves over arbitrary
1693 finite fields, but as there are no obvious types for fields other
1694 than GF(p), some functions are limited to that for now.
1697 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
1698 that the file contains a complete HTTP response.
1701 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
1702 change the def and num file printf format specifier from "%-40sXXX"
1703 to "%-39s XXX". The latter will always guarantee a space after the
1704 field while the former will cause them to run together if the field
1705 is 40 of more characters long.
1708 *) Constify the cipher and digest 'method' functions and structures
1709 and modify related functions to take constant EVP_MD and EVP_CIPHER
1713 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
1714 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
1717 *) Modify EVP_Digest*() routines so they now return values. Although the
1718 internal software routines can never fail additional hardware versions
1722 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
1724 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
1725 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
1728 ERR_R_NESTED_ASN1_ERROR
1730 ERR_R_MISSING_ASN1_EOS
1731 were 4 .. 9, conflicting with
1732 ERR_LIB_RSA (= ERR_R_RSA_LIB)
1734 ERR_LIB_PEM (= ERR_R_PEM_LIB).
1735 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
1737 Add new error code 'ERR_R_INTERNAL_ERROR'.
1740 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
1744 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
1745 sets the subject name for a new request or supersedes the
1746 subject name in a given request. Formats that can be parsed are
1747 'CN=Some Name, OU=myOU, C=IT'
1749 'CN=Some Name/OU=myOU/C=IT'.
1751 Add options '-batch' and '-verbose' to 'openssl req'.
1752 [Massimiliano Pala <madwolf@hackmasters.net>]
1754 *) Introduce the possibility to access global variables through
1755 functions on platform were that's the best way to handle exporting
1756 global variables in shared libraries. To enable this functionality,
1757 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1758 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1759 is normally done by Configure or something similar).
1761 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1762 in the source file (foo.c) like this:
1764 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1765 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1767 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1768 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1770 OPENSSL_DECLARE_GLOBAL(int,foo);
1771 #define foo OPENSSL_GLOBAL_REF(foo)
1772 OPENSSL_DECLARE_GLOBAL(double,bar);
1773 #define bar OPENSSL_GLOBAL_REF(bar)
1775 The #defines are very important, and therefore so is including the
1776 header file everywhere where the defined globals are used.
1778 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1779 of ASN.1 items, but that structure is a bit different.
1781 The largest change is in util/mkdef.pl which has been enhanced with
1782 better and easier to understand logic to choose which symbols should
1783 go into the Windows .def files as well as a number of fixes and code
1784 cleanup (among others, algorithm keywords are now sorted
1785 lexicographically to avoid constant rewrites).
1788 *) In BN_div() keep a copy of the sign of 'num' before writing the
1789 result to 'rm' because if rm==num the value will be overwritten
1790 and produce the wrong result if 'num' is negative: this caused
1791 problems with BN_mod() and BN_nnmod().
1794 *) Function OCSP_request_verify(). This checks the signature on an
1795 OCSP request and verifies the signer certificate. The signer
1796 certificate is just checked for a generic purpose and OCSP request
1800 *) Add OCSP_check_validity() function to check the validity of OCSP
1801 responses. OCSP responses are prepared in real time and may only
1802 be a few seconds old. Simply checking that the current time lies
1803 between thisUpdate and nextUpdate max reject otherwise valid responses
1804 caused by either OCSP responder or client clock inaccuracy. Instead
1805 we allow thisUpdate and nextUpdate to fall within a certain period of
1806 the current time. The age of the response can also optionally be
1807 checked. Two new options -validity_period and -status_age added to
1811 *) If signature or public key algorithm is unrecognized print out its
1812 OID rather that just UNKNOWN.
1815 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1816 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1817 ID to be generated from the issuer certificate alone which can then be
1818 passed to OCSP_id_issuer_cmp().
1821 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1822 ASN1 modules to export functions returning ASN1_ITEM pointers
1823 instead of the ASN1_ITEM structures themselves. This adds several
1824 new macros which allow the underlying ASN1 function/structure to
1825 be accessed transparently. As a result code should not use ASN1_ITEM
1826 references directly (such as &X509_it) but instead use the relevant
1827 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1828 use of the new ASN1 code on platforms where exporting structures
1829 is problematical (for example in shared libraries) but exporting
1830 functions returning pointers to structures is not.
1833 *) Add support for overriding the generation of SSL/TLS session IDs.
1834 These callbacks can be registered either in an SSL_CTX or per SSL.
1835 The purpose of this is to allow applications to control, if they wish,
1836 the arbitrary values chosen for use as session IDs, particularly as it
1837 can be useful for session caching in multiple-server environments. A
1838 command-line switch for testing this (and any client code that wishes
1839 to use such a feature) has been added to "s_server".
1840 [Geoff Thorpe, Lutz Jaenicke]
1842 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
1843 of the form '#if defined(...) || defined(...) || ...' and
1844 '#if !defined(...) && !defined(...) && ...'. This also avoids
1845 the growing number of special cases it was previously handling.
1848 *) Make all configuration macros available for application by making
1849 sure they are available in opensslconf.h, by giving them names starting
1850 with "OPENSSL_" to avoid conflicts with other packages and by making
1851 sure e_os2.h will cover all platform-specific cases together with
1853 Additionally, it is now possible to define configuration/platform-
1854 specific names (called "system identities"). In the C code, these
1855 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1856 macro with the name beginning with "OPENSSL_SYS_", which is determined
1857 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1861 *) New option -set_serial to 'req' and 'x509' this allows the serial
1862 number to use to be specified on the command line. Previously self
1863 signed certificates were hard coded with serial number 0 and the
1864 CA options of 'x509' had to use a serial number in a file which was
1868 *) New options to 'ca' utility to support V2 CRL entry extensions.
1869 Currently CRL reason, invalidity date and hold instruction are
1870 supported. Add new CRL extensions to V3 code and some new objects.
1873 *) New function EVP_CIPHER_CTX_set_padding() this is used to
1874 disable standard block padding (aka PKCS#5 padding) in the EVP
1875 API, which was previously mandatory. This means that the data is
1876 not padded in any way and so the total length much be a multiple
1877 of the block size, otherwise an error occurs.
1880 *) Initial (incomplete) OCSP SSL support.
1883 *) New function OCSP_parse_url(). This splits up a URL into its host,
1884 port and path components: primarily to parse OCSP URLs. New -url
1885 option to ocsp utility.
1888 *) New nonce behavior. The return value of OCSP_check_nonce() now
1889 reflects the various checks performed. Applications can decide
1890 whether to tolerate certain situations such as an absent nonce
1891 in a response when one was present in a request: the ocsp application
1892 just prints out a warning. New function OCSP_add1_basic_nonce()
1893 this is to allow responders to include a nonce in a response even if
1894 the request is nonce-less.
1897 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1898 skipped when using openssl x509 multiple times on a single input file,
1899 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1902 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1903 set string type: to handle setting ASN1_TIME structures. Fix ca
1904 utility to correctly initialize revocation date of CRLs.
1907 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1908 the clients preferred ciphersuites and rather use its own preferences.
1909 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1910 Internet Explorer by ensuring unchanged hash method during stepup.
1911 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1914 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1915 to aes and add a new 'exist' option to print out symbols that don't
1919 *) Additional options to ocsp utility to allow flags to be set and
1920 additional certificates supplied.
1923 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
1924 OCSP client a number of certificate to only verify the response
1928 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1929 handle the new API. Currently only ECB, CBC modes supported. Add new
1932 Add TLS AES ciphersuites as described in RFC3268, "Advanced
1933 Encryption Standard (AES) Ciphersuites for Transport Layer
1934 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
1935 not enabled by default and were not part of the "ALL" ciphersuite
1936 alias because they were not yet official; they could be
1937 explicitly requested by specifying the "AESdraft" ciphersuite
1938 group alias. In the final release of OpenSSL 0.9.7, the group
1939 alias is called "AES" and is part of "ALL".)
1940 [Ben Laurie, Steve Henson, Bodo Moeller]
1942 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
1943 request to response.
1946 *) Functions for OCSP responders. OCSP_request_onereq_count(),
1947 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1948 extract information from a certificate request. OCSP_response_create()
1949 creates a response and optionally adds a basic response structure.
1950 OCSP_basic_add1_status() adds a complete single response to a basic
1951 response and returns the OCSP_SINGLERESP structure just added (to allow
1952 extensions to be included for example). OCSP_basic_add1_cert() adds a
1953 certificate to a basic response and OCSP_basic_sign() signs a basic
1954 response with various flags. New helper functions ASN1_TIME_check()
1955 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1956 (converts ASN1_TIME to GeneralizedTime).
1959 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1960 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1961 structure from a certificate. X509_pubkey_digest() digests the public_key
1962 contents: this is used in various key identifiers.
1965 *) Make sk_sort() tolerate a NULL argument.
1966 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1968 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1969 passed by the function are trusted implicitly. If any of them signed the
1970 response then it is assumed to be valid and is not verified.
1973 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1974 to data. This was previously part of the PKCS7 ASN1 code. This
1975 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1976 [Steve Henson, reported by Kenneth R. Robinette
1977 <support@securenetterm.com>]
1979 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1980 routines: without these tracing memory leaks is very painful.
1981 Fix leaks in PKCS12 and PKCS7 routines.
1984 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1985 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1986 effectively meant GeneralizedTime would never be used. Now it
1987 is initialised to -1 but X509_time_adj() now has to check the value
1988 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1989 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1990 [Steve Henson, reported by Kenneth R. Robinette
1991 <support@securenetterm.com>]
1993 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1994 result in a zero length in the ASN1_INTEGER structure which was
1995 not consistent with the structure when d2i_ASN1_INTEGER() was used
1996 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1997 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1998 where it did not print out a minus for negative ASN1_INTEGER.
2001 *) Add summary printout to ocsp utility. The various functions which
2002 convert status values to strings have been renamed to:
2003 OCSP_response_status_str(), OCSP_cert_status_str() and
2004 OCSP_crl_reason_str() and are no longer static. New options
2005 to verify nonce values and to disable verification. OCSP response
2006 printout format cleaned up.
2009 *) Add additional OCSP certificate checks. These are those specified
2010 in RFC2560. This consists of two separate checks: the CA of the
2011 certificate being checked must either be the OCSP signer certificate
2012 or the issuer of the OCSP signer certificate. In the latter case the
2013 OCSP signer certificate must contain the OCSP signing extended key
2014 usage. This check is performed by attempting to match the OCSP
2015 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
2016 in the OCSP_CERTID structures of the response.
2019 *) Initial OCSP certificate verification added to OCSP_basic_verify()
2020 and related routines. This uses the standard OpenSSL certificate
2021 verify routines to perform initial checks (just CA validity) and
2022 to obtain the certificate chain. Then additional checks will be
2023 performed on the chain. Currently the root CA is checked to see
2024 if it is explicitly trusted for OCSP signing. This is used to set
2025 a root CA as a global signing root: that is any certificate that
2026 chains to that CA is an acceptable OCSP signing certificate.
2029 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
2030 extensions from a separate configuration file.
2031 As when reading extensions from the main configuration file,
2032 the '-extensions ...' option may be used for specifying the
2034 [Massimiliano Pala <madwolf@comune.modena.it>]
2036 *) New OCSP utility. Allows OCSP requests to be generated or
2037 read. The request can be sent to a responder and the output
2038 parsed, outputed or printed in text form. Not complete yet:
2039 still needs to check the OCSP response validity.
2042 *) New subcommands for 'openssl ca':
2043 'openssl ca -status <serial>' prints the status of the cert with
2044 the given serial number (according to the index file).
2045 'openssl ca -updatedb' updates the expiry status of certificates
2047 [Massimiliano Pala <madwolf@comune.modena.it>]
2049 *) New '-newreq-nodes' command option to CA.pl. This is like
2050 '-newreq', but calls 'openssl req' with the '-nodes' option
2051 so that the resulting key is not encrypted.
2052 [Damien Miller <djm@mindrot.org>]
2054 *) New configuration for the GNU Hurd.
2055 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
2057 *) Initial code to implement OCSP basic response verify. This
2058 is currently incomplete. Currently just finds the signer's
2059 certificate and verifies the signature on the response.
2062 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
2063 value of OPENSSLDIR. This is available via the new '-d' option
2064 to 'openssl version', and is also included in 'openssl version -a'.
2067 *) Allowing defining memory allocation callbacks that will be given
2068 file name and line number information in additional arguments
2069 (a const char* and an int). The basic functionality remains, as
2070 well as the original possibility to just replace malloc(),
2071 realloc() and free() by functions that do not know about these
2072 additional arguments. To register and find out the current
2073 settings for extended allocation functions, the following
2074 functions are provided:
2076 CRYPTO_set_mem_ex_functions
2077 CRYPTO_set_locked_mem_ex_functions
2078 CRYPTO_get_mem_ex_functions
2079 CRYPTO_get_locked_mem_ex_functions
2081 These work the same way as CRYPTO_set_mem_functions and friends.
2082 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
2083 extended allocation function is enabled.
2084 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
2085 a conventional allocation function is enabled.
2086 [Richard Levitte, Bodo Moeller]
2088 *) Finish off removing the remaining LHASH function pointer casts.
2089 There should no longer be any prototype-casting required when using
2090 the LHASH abstraction, and any casts that remain are "bugs". See
2091 the callback types and macros at the head of lhash.h for details
2092 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
2095 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
2096 If /dev/[u]random devices are not available or do not return enough
2097 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
2099 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
2100 /etc/entropy will be queried once each in this sequence, quering stops
2101 when enough entropy was collected without querying more sockets.
2104 *) Change the Unix RAND_poll() variant to be able to poll several
2105 random devices, as specified by DEVRANDOM, until a sufficient amount
2106 of data has been collected. We spend at most 10 ms on each file
2107 (select timeout) and read in non-blocking mode. DEVRANDOM now
2108 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
2109 (previously it was just the string "/dev/urandom"), so on typical
2110 platforms the 10 ms delay will never occur.
2111 Also separate out the Unix variant to its own file, rand_unix.c.
2112 For VMS, there's a currently-empty rand_vms.c.
2115 *) Move OCSP client related routines to ocsp_cl.c. These
2116 provide utility functions which an application needing
2117 to issue a request to an OCSP responder and analyse the
2118 response will typically need: as opposed to those which an
2119 OCSP responder itself would need which will be added later.
2121 OCSP_request_sign() signs an OCSP request with an API similar
2122 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
2123 response. OCSP_response_get1_basic() extracts basic response
2124 from response. OCSP_resp_find_status(): finds and extracts status
2125 information from an OCSP_CERTID structure (which will be created
2126 when the request structure is built). These are built from lower
2127 level functions which work on OCSP_SINGLERESP structures but
2128 wont normally be used unless the application wishes to examine
2129 extensions in the OCSP response for example.
2131 Replace nonce routines with a pair of functions.
2132 OCSP_request_add1_nonce() adds a nonce value and optionally
2133 generates a random value. OCSP_check_nonce() checks the
2134 validity of the nonce in an OCSP response.
2137 *) Change function OCSP_request_add() to OCSP_request_add0_id().
2138 This doesn't copy the supplied OCSP_CERTID and avoids the
2139 need to free up the newly created id. Change return type
2140 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
2141 This can then be used to add extensions to the request.
2142 Deleted OCSP_request_new(), since most of its functionality
2143 is now in OCSP_REQUEST_new() (and the case insensitive name
2144 clash) apart from the ability to set the request name which
2145 will be added elsewhere.
2148 *) Update OCSP API. Remove obsolete extensions argument from
2149 various functions. Extensions are now handled using the new
2150 OCSP extension code. New simple OCSP HTTP function which
2151 can be used to send requests and parse the response.
2154 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
2155 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
2156 uses the special reorder version of SET OF to sort the attributes
2157 and reorder them to match the encoded order. This resolves a long
2158 standing problem: a verify on a PKCS7 structure just after signing
2159 it used to fail because the attribute order did not match the
2160 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
2161 it uses the received order. This is necessary to tolerate some broken
2162 software that does not order SET OF. This is handled by encoding
2163 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
2164 to produce the required SET OF.
2167 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
2168 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
2169 files to get correct declarations of the ASN.1 item variables.
2172 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
2173 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
2174 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
2175 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
2176 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
2177 ASN1_ITEM and no wrapper functions.
2180 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
2181 replace the old function pointer based I/O routines. Change most of
2182 the *_d2i_bio() and *_d2i_fp() functions to use these.
2185 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
2186 lines, recognice more "algorithms" that can be deselected, and make
2187 it complain about algorithm deselection that isn't recognised.
2190 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
2191 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
2192 to use new functions. Add NO_ASN1_OLD which can be set to remove
2193 some old style ASN1 functions: this can be used to determine if old
2194 code will still work when these eventually go away.
2197 *) New extension functions for OCSP structures, these follow the
2198 same conventions as certificates and CRLs.
2201 *) New function X509V3_add1_i2d(). This automatically encodes and
2202 adds an extension. Its behaviour can be customised with various
2203 flags to append, replace or delete. Various wrappers added for
2204 certifcates and CRLs.
2207 *) Fix to avoid calling the underlying ASN1 print routine when
2208 an extension cannot be parsed. Correct a typo in the
2209 OCSP_SERVICELOC extension. Tidy up print OCSP format.
2212 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
2213 entries for variables.
2216 *) Add functionality to apps/openssl.c for detecting locking
2217 problems: As the program is single-threaded, all we have
2218 to do is register a locking callback using an array for
2219 storing which locks are currently held by the program.
2222 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
2223 SSL_get_ex_data_X509_STORE_idx(), which is used in
2224 ssl_verify_cert_chain() and thus can be called at any time
2225 during TLS/SSL handshakes so that thread-safety is essential.
2226 Unfortunately, the ex_data design is not at all suited
2227 for multi-threaded use, so it probably should be abolished.
2230 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
2231 [Broadcom, tweaked and integrated by Geoff Thorpe]
2233 *) Move common extension printing code to new function
2234 X509V3_print_extensions(). Reorganise OCSP print routines and
2235 implement some needed OCSP ASN1 functions. Add OCSP extensions.
2238 *) New function X509_signature_print() to remove duplication in some
2242 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
2243 set (this was treated exactly the same as SET OF previously). This
2244 is used to reorder the STACK representing the structure to match the
2245 encoding. This will be used to get round a problem where a PKCS7
2246 structure which was signed could not be verified because the STACK
2247 order did not reflect the encoded order.
2250 *) Reimplement the OCSP ASN1 module using the new code.
2253 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
2254 for its ASN1 operations. The old style function pointers still exist
2255 for now but they will eventually go away.
2258 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
2259 completely replaces the old ASN1 functionality with a table driven
2260 encoder and decoder which interprets an ASN1_ITEM structure describing
2261 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
2262 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
2263 has also been converted to the new form.
2266 *) Change BN_mod_exp_recp so that negative moduli are tolerated
2267 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
2268 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
2269 for negative moduli.
2272 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
2273 of not touching the result's sign bit.
2276 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
2280 *) Changed the LHASH code to use prototypes for callbacks, and created
2281 macros to declare and implement thin (optionally static) functions
2282 that provide type-safety and avoid function pointer casting for the
2283 type-specific callbacks.
2286 *) Added Kerberos Cipher Suites to be used with TLS, as written in
2288 [Veers Staats <staatsvr@asc.hpc.mil>,
2289 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
2291 *) Reformat the FAQ so the different questions and answers can be divided
2292 in sections depending on the subject.
2295 *) Have the zlib compression code load ZLIB.DLL dynamically under
2299 *) New function BN_mod_sqrt for computing square roots modulo a prime
2300 (using the probabilistic Tonelli-Shanks algorithm unless
2301 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
2302 be handled deterministically).
2303 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2305 *) Make BN_mod_inverse faster by explicitly handling small quotients
2306 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
2307 512 bits], about 30% for larger ones [1024 or 2048 bits].)
2310 *) New function BN_kronecker.
2313 *) Fix BN_gcd so that it works on negative inputs; the result is
2314 positive unless both parameters are zero.
2315 Previously something reasonably close to an infinite loop was
2316 possible because numbers could be growing instead of shrinking
2317 in the implementation of Euclid's algorithm.
2320 *) Fix BN_is_word() and BN_is_one() macros to take into account the
2321 sign of the number in question.
2323 Fix BN_is_word(a,w) to work correctly for w == 0.
2325 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
2326 because its test if the absolute value of 'a' equals 'w'.
2327 Note that BN_abs_is_word does *not* handle w == 0 reliably;
2328 it exists mostly for use in the implementations of BN_is_zero(),
2329 BN_is_one(), and BN_is_word().
2332 *) New function BN_swap.
2335 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
2336 the exponentiation functions are more likely to produce reasonable
2337 results on negative inputs.
2340 *) Change BN_mod_mul so that the result is always non-negative.
2341 Previously, it could be negative if one of the factors was negative;
2342 I don't think anyone really wanted that behaviour.
2345 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
2346 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
2347 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
2348 and add new functions:
2357 BN_mod_lshift1_quick
2361 These functions always generate non-negative results.
2363 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
2364 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
2366 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2367 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2368 be reduced modulo m.
2369 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2372 The following entry accidentily appeared in the CHANGES file
2373 distributed with OpenSSL 0.9.7. The modifications described in
2374 it do *not* apply to OpenSSL 0.9.7.
2376 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2377 was actually never needed) and in BN_mul(). The removal in BN_mul()
2378 required a small change in bn_mul_part_recursive() and the addition
2379 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2380 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2381 bn_sub_words() and bn_add_words() except they take arrays with
2386 *) In 'openssl passwd', verify passwords read from the terminal
2387 unless the '-salt' option is used (which usually means that
2388 verification would just waste user's time since the resulting
2389 hash is going to be compared with some given password hash)
2390 or the new '-noverify' option is used.
2392 This is an incompatible change, but it does not affect
2393 non-interactive use of 'openssl passwd' (passwords on the command
2394 line, '-stdin' option, '-in ...' option) and thus should not
2398 *) Remove all references to RSAref, since there's no more need for it.
2401 *) Make DSO load along a path given through an environment variable
2402 (SHLIB_PATH) with shl_load().
2405 *) Constify the ENGINE code as a result of BIGNUM constification.
2406 Also constify the RSA code and most things related to it. In a
2407 few places, most notable in the depth of the ASN.1 code, ugly
2408 casts back to non-const were required (to be solved at a later
2412 *) Make it so the openssl application has all engines loaded by default.
2415 *) Constify the BIGNUM routines a little more.
2418 *) Add the following functions:
2420 ENGINE_load_cswift()
2422 ENGINE_load_atalla()
2424 ENGINE_load_builtin_engines()
2426 That way, an application can itself choose if external engines that
2427 are built-in in OpenSSL shall ever be used or not. The benefit is
2428 that applications won't have to be linked with libdl or other dso
2429 libraries unless it's really needed.
2431 Changed 'openssl engine' to load all engines on demand.
2432 Changed the engine header files to avoid the duplication of some
2433 declarations (they differed!).
2436 *) 'openssl engine' can now list capabilities.
2439 *) Better error reporting in 'openssl engine'.
2442 *) Never call load_dh_param(NULL) in s_server.
2445 *) Add engine application. It can currently list engines by name and
2446 identity, and test if they are actually available.
2449 *) Improve RPM specification file by forcing symbolic linking and making
2450 sure the installed documentation is also owned by root.root.
2451 [Damien Miller <djm@mindrot.org>]
2453 *) Give the OpenSSL applications more possibilities to make use of
2454 keys (public as well as private) handled by engines.
2457 *) Add OCSP code that comes from CertCo.
2460 *) Add VMS support for the Rijndael code.
2463 *) Added untested support for Nuron crypto accelerator.
2466 *) Add support for external cryptographic devices. This code was
2467 previously distributed separately as the "engine" branch.
2468 [Geoff Thorpe, Richard Levitte]
2470 *) Rework the filename-translation in the DSO code. It is now possible to
2471 have far greater control over how a "name" is turned into a filename
2472 depending on the operating environment and any oddities about the
2473 different shared library filenames on each system.
2476 *) Support threads on FreeBSD-elf in Configure.
2479 *) Fix for SHA1 assembly problem with MASM: it produces
2480 warnings about corrupt line number information when assembling
2481 with debugging information. This is caused by the overlapping
2483 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2486 NCONF_get_number() has no error checking at all. As a replacement,
2487 NCONF_get_number_e() is defined (_e for "error checking") and is
2488 promoted strongly. The old NCONF_get_number is kept around for
2489 binary backward compatibility.
2490 Make it possible for methods to load from something other than a BIO,
2491 by providing a function pointer that is given a name instead of a BIO.
2492 For example, this could be used to load configuration data from an
2496 *) Fix for non blocking accept BIOs. Added new I/O special reason
2497 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2498 with non blocking I/O was not possible because no retry code was
2499 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2503 *) Added the beginnings of Rijndael support.
2506 *) Fix for bug in DirectoryString mask setting. Add support for
2507 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2508 to allow certificate printing to more controllable, additional
2509 'certopt' option to 'x509' to allow new printing options to be
2513 *) Clean old EAY MD5 hack from e_os.h.
2516 Changes between 0.9.6j and 0.9.6k [xx XXX 2003]
2518 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2519 extra data after the compression methods not only for TLS 1.0
2520 but also for SSL 3.0 (as required by the specification).
2521 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2523 *) Change X509_certificate_type() to mark the key as exported/exportable
2524 when it's 512 *bits* long, not 512 bytes.
2527 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
2529 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2530 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2531 a protocol version number mismatch like a decryption error
2532 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2535 *) Turn on RSA blinding by default in the default implementation
2536 to avoid a timing attack. Applications that don't want it can call
2537 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2538 They would be ill-advised to do so in most cases.
2539 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2541 *) Change RSA blinding code so that it works when the PRNG is not
2542 seeded (in this case, the secret RSA exponent is abused as
2543 an unpredictable seed -- if it is not unpredictable, there
2544 is no point in blinding anyway). Make RSA blinding thread-safe
2545 by remembering the creator's thread ID in rsa->blinding and
2546 having all other threads use local one-time blinding factors
2547 (this requires more computation than sharing rsa->blinding, but
2548 avoids excessive locking; and if an RSA object is not shared
2549 between threads, blinding will still be very fast).
2552 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
2554 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2555 via timing by performing a MAC computation even if incorrrect
2556 block cipher padding has been found. This is a countermeasure
2557 against active attacks where the attacker has to distinguish
2558 between bad padding and a MAC verification error. (CAN-2003-0078)
2560 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2561 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2562 Martin Vuagnoux (EPFL, Ilion)]
2564 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
2566 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
2567 memory from it's contents. This is done with a counter that will
2568 place alternating values in each byte. This can be used to solve
2569 two issues: 1) the removal of calls to memset() by highly optimizing
2570 compilers, and 2) cleansing with other values than 0, since those can
2571 be read through on certain media, for example a swap space on disk.
2574 *) Bugfix: client side session caching did not work with external caching,
2575 because the session->cipher setting was not restored when reloading
2576 from the external cache. This problem was masked, when
2577 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
2578 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
2581 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
2582 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
2583 [Zeev Lieber <zeev-l@yahoo.com>]
2585 *) Undo an undocumented change introduced in 0.9.6e which caused
2586 repeated calls to OpenSSL_add_all_ciphers() and
2587 OpenSSL_add_all_digests() to be ignored, even after calling
2591 *) Change the default configuration reader to deal with last line not
2592 being properly terminated.
2595 *) Change X509_NAME_cmp() so it applies the special rules on handling
2596 DN values that are of type PrintableString, as well as RDNs of type
2597 emailAddress where the value has the type ia5String.
2598 [stefank@valicert.com via Richard Levitte]
2600 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
2601 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
2602 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
2603 the bitwise-OR of the two for use by the majority of applications
2604 wanting this behaviour, and update the docs. The documented
2605 behaviour and actual behaviour were inconsistent and had been
2606 changing anyway, so this is more a bug-fix than a behavioural
2608 [Geoff Thorpe, diagnosed by Nadav Har'El]
2610 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
2611 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
2614 *) Fix initialization code race conditions in
2615 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
2616 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
2617 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
2618 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
2619 ssl2_get_cipher_by_char(),
2620 ssl3_get_cipher_by_char().
2621 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
2623 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
2624 the cached sessions are flushed, as the remove_cb() might use ex_data
2625 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
2626 (see [openssl.org #212]).
2627 [Geoff Thorpe, Lutz Jaenicke]
2629 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
2630 length, instead of the encoding length to d2i_ASN1_OBJECT.
2633 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
2635 *) [In 0.9.6g-engine release:]
2636 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
2637 [Lynn Gazis <lgazis@rainbow.com>]
2639 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
2641 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
2642 and get fix the header length calculation.
2643 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
2644 Alon Kantor <alonk@checkpoint.com> (and others),
2647 *) Use proper error handling instead of 'assertions' in buffer
2648 overflow checks added in 0.9.6e. This prevents DoS (the
2649 assertions could call abort()).
2650 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
2652 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
2654 *) Add various sanity checks to asn1_get_length() to reject
2655 the ASN1 length bytes if they exceed sizeof(long), will appear
2656 negative or the content length exceeds the length of the
2658 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2660 *) Fix cipher selection routines: ciphers without encryption had no flags
2661 for the cipher strength set and where therefore not handled correctly
2662 by the selection routines (PR #130).
2665 *) Fix EVP_dsa_sha macro.
2669 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
2670 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
2671 that was added in OpenSSL 0.9.6d.
2673 As the countermeasure turned out to be incompatible with some
2674 broken SSL implementations, the new option is part of SSL_OP_ALL.
2675 SSL_OP_ALL is usually employed when compatibility with weird SSL
2676 implementations is desired (e.g. '-bugs' option to 's_client' and
2677 's_server'), so the new option is automatically set in many
2681 *) Changes in security patch:
2683 Changes marked "(CHATS)" were sponsored by the Defense Advanced
2684 Research Projects Agency (DARPA) and Air Force Research Laboratory,
2685 Air Force Materiel Command, USAF, under agreement number
2688 *) Add various sanity checks to asn1_get_length() to reject
2689 the ASN1 length bytes if they exceed sizeof(long), will appear
2690 negative or the content length exceeds the length of the
2691 supplied buffer. (CAN-2002-0659)
2692 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2694 *) Assertions for various potential buffer overflows, not known to
2696 [Ben Laurie (CHATS)]
2698 *) Various temporary buffers to hold ASCII versions of integers were
2699 too small for 64 bit platforms. (CAN-2002-0655)
2700 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
2702 *) Remote buffer overflow in SSL3 protocol - an attacker could
2703 supply an oversized session ID to a client. (CAN-2002-0656)
2704 [Ben Laurie (CHATS)]
2706 *) Remote buffer overflow in SSL2 protocol - an attacker could
2707 supply an oversized client master key. (CAN-2002-0656)
2708 [Ben Laurie (CHATS)]
2710 Changes between 0.9.6c and 0.9.6d [9 May 2002]
2712 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
2713 encoded as NULL) with id-dsa-with-sha1.
2714 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
2716 *) Check various X509_...() return values in apps/req.c.
2717 [Nils Larsch <nla@trustcenter.de>]
2719 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
2720 an end-of-file condition would erronously be flagged, when the CRLF
2721 was just at the end of a processed block. The bug was discovered when
2722 processing data through a buffering memory BIO handing the data to a
2723 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
2724 <ptsekov@syntrex.com> and Nedelcho Stanev.
2727 *) Implement a countermeasure against a vulnerability recently found
2728 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
2729 before application data chunks to avoid the use of known IVs
2730 with data potentially chosen by the attacker.
2733 *) Fix length checks in ssl3_get_client_hello().
2736 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
2737 to prevent ssl3_read_internal() from incorrectly assuming that
2738 ssl3_read_bytes() found application data while handshake
2739 processing was enabled when in fact s->s3->in_read_app_data was
2740 merely automatically cleared during the initial handshake.
2741 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
2743 *) Fix object definitions for Private and Enterprise: they were not
2744 recognized in their shortname (=lowercase) representation. Extend
2745 obj_dat.pl to issue an error when using undefined keywords instead
2746 of silently ignoring the problem (Svenning Sorensen
2747 <sss@sss.dnsalias.net>).
2750 *) Fix DH_generate_parameters() so that it works for 'non-standard'
2751 generators, i.e. generators other than 2 and 5. (Previously, the
2752 code did not properly initialise the 'add' and 'rem' values to
2753 BN_generate_prime().)
2755 In the new general case, we do not insist that 'generator' is
2756 actually a primitive root: This requirement is rather pointless;
2757 a generator of the order-q subgroup is just as good, if not
2761 *) Map new X509 verification errors to alerts. Discovered and submitted by
2762 Tom Wu <tom@arcot.com>.
2765 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
2766 returning non-zero before the data has been completely received
2767 when using non-blocking I/O.
2768 [Bodo Moeller; problem pointed out by John Hughes]
2770 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
2771 [Ben Laurie, Lutz Jaenicke]
2773 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
2774 Yoram Zahavi <YoramZ@gilian.com>).
2777 *) Add information about CygWin 1.3 and on, and preserve proper
2778 configuration for the versions before that.
2779 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2781 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
2782 check whether we deal with a copy of a session and do not delete from
2783 the cache in this case. Problem reported by "Izhar Shoshani Levi"
2784 <izhar@checkpoint.com>.
2787 *) Do not store session data into the internal session cache, if it
2788 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
2789 flag is set). Proposed by Aslam <aslam@funk.com>.
2792 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
2796 *) [In 0.9.6d-engine release:]
2797 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2798 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2800 *) Add the configuration target linux-s390x.
2801 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
2803 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
2804 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
2805 variable as an indication that a ClientHello message has been
2806 received. As the flag value will be lost between multiple
2807 invocations of ssl3_accept when using non-blocking I/O, the
2808 function may not be aware that a handshake has actually taken
2809 place, thus preventing a new session from being added to the
2812 To avoid this problem, we now set s->new_session to 2 instead of
2813 using a local variable.
2814 [Lutz Jaenicke, Bodo Moeller]
2816 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
2817 if the SSL_R_LENGTH_MISMATCH error is detected.
2818 [Geoff Thorpe, Bodo Moeller]
2820 *) New 'shared_ldflag' column in Configure platform table.
2823 *) Fix EVP_CIPHER_mode macro.
2824 ["Dan S. Camper" <dan@bti.net>]
2826 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
2827 type, we must throw them away by setting rr->length to 0.
2828 [D P Chang <dpc@qualys.com>]
2830 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
2832 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
2833 <Dominikus.Scherkl@biodata.com>. (The previous implementation
2834 worked incorrectly for those cases where range = 10..._2 and
2835 3*range is two bits longer than range.)
2838 *) Only add signing time to PKCS7 structures if it is not already
2842 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
2843 OBJ_ld_ce should be OBJ_id_ce.
2844 Also some ip-pda OIDs in crypto/objects/objects.txt were
2845 incorrect (cf. RFC 3039).
2846 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
2848 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
2849 returns early because it has nothing to do.
2850 [Andy Schneider <andy.schneider@bjss.co.uk>]
2852 *) [In 0.9.6c-engine release:]
2853 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
2854 [Andy Schneider <andy.schneider@bjss.co.uk>]
2856 *) [In 0.9.6c-engine release:]
2857 Add support for Cryptographic Appliance's keyserver technology.
2858 (Use engine 'keyclient')
2859 [Cryptographic Appliances and Geoff Thorpe]
2861 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
2862 is called via tools/c89.sh because arguments have to be
2863 rearranged (all '-L' options must appear before the first object
2865 [Richard Shapiro <rshapiro@abinitio.com>]
2867 *) [In 0.9.6c-engine release:]
2868 Add support for Broadcom crypto accelerator cards, backported
2870 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
2872 *) [In 0.9.6c-engine release:]
2873 Add support for SureWare crypto accelerator cards from
2874 Baltimore Technologies. (Use engine 'sureware')
2875 [Baltimore Technologies and Mark Cox]
2877 *) [In 0.9.6c-engine release:]
2878 Add support for crypto accelerator cards from Accelerated
2879 Encryption Processing, www.aep.ie. (Use engine 'aep')
2880 [AEP Inc. and Mark Cox]
2882 *) Add a configuration entry for gcc on UnixWare.
2883 [Gary Benson <gbenson@redhat.com>]
2885 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
2886 messages are stored in a single piece (fixed-length part and
2887 variable-length part combined) and fix various bugs found on the way.
2890 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
2891 instead. BIO_gethostbyname() does not know what timeouts are
2892 appropriate, so entries would stay in cache even when they have
2894 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
2896 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
2897 faced with a pathologically small ClientHello fragment that does
2898 not contain client_version: Instead of aborting with an error,
2899 simply choose the highest available protocol version (i.e.,
2900 TLS 1.0 unless it is disabled). In practice, ClientHello
2901 messages are never sent like this, but this change gives us
2902 strictly correct behaviour at least for TLS.
2905 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
2906 never resets s->method to s->ctx->method when called from within
2907 one of the SSL handshake functions.
2908 [Bodo Moeller; problem pointed out by Niko Baric]
2910 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
2911 (sent using the client's version number) if client_version is
2912 smaller than the protocol version in use. Also change
2913 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
2914 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
2915 the client will at least see that alert.
2918 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
2922 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
2923 client receives HelloRequest while in a handshake.
2924 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
2926 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
2927 should end in 'break', not 'goto end' which circuments various
2928 cleanups done in state SSL_ST_OK. But session related stuff
2929 must be disabled for SSL_ST_OK in the case that we just sent a
2932 Also avoid some overhead by not calling ssl_init_wbio_buffer()
2933 before just sending a HelloRequest.
2934 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
2936 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
2937 reveal whether illegal block cipher padding was found or a MAC
2938 verification error occured. (Neither SSLerr() codes nor alerts
2939 are directly visible to potential attackers, but the information
2940 may leak via logfiles.)
2942 Similar changes are not required for the SSL 2.0 implementation
2943 because the number of padding bytes is sent in clear for SSL 2.0,
2944 and the extra bytes are just ignored. However ssl/s2_pkt.c
2945 failed to verify that the purported number of padding bytes is in
2949 *) Add OpenUNIX-8 support including shared libraries
2950 (Boyd Lynn Gerber <gerberb@zenez.com>).
2953 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
2954 'wristwatch attack' using huge encoding parameters (cf.
2955 James H. Manger's CRYPTO 2001 paper). Note that the
2956 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
2957 encoding parameters and hence was not vulnerable.
2960 *) BN_sqr() bug fix.
2961 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
2963 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
2964 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
2965 followed by modular reduction.
2966 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
2968 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
2969 equivalent based on BN_pseudo_rand() instead of BN_rand().
2972 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
2973 This function was broken, as the check for a new client hello message
2974 to handle SGC did not allow these large messages.
2975 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
2978 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
2981 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
2982 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
2985 *) Rework the configuration and shared library support for Tru64 Unix.
2986 The configuration part makes use of modern compiler features and
2987 still retains old compiler behavior for those that run older versions
2988 of the OS. The shared library support part includes a variant that
2989 uses the RPATH feature, and is available through the special
2990 configuration target "alpha-cc-rpath", which will never be selected
2992 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
2994 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
2995 with the same message size as in ssl3_get_certificate_request().
2996 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
2997 messages might inadvertently be reject as too long.
2998 [Petr Lampa <lampa@fee.vutbr.cz>]
3000 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
3003 *) Modified SSL library such that the verify_callback that has been set
3004 specificly for an SSL object with SSL_set_verify() is actually being
3005 used. Before the change, a verify_callback set with this function was
3006 ignored and the verify_callback() set in the SSL_CTX at the time of
3007 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
3008 to allow the necessary settings.
3011 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
3012 explicitly to NULL, as at least on Solaris 8 this seems not always to be
3013 done automatically (in contradiction to the requirements of the C
3014 standard). This made problems when used from OpenSSH.
3017 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
3018 dh->length and always used
3020 BN_rand_range(priv_key, dh->p).
3022 BN_rand_range() is not necessary for Diffie-Hellman, and this
3023 specific range makes Diffie-Hellman unnecessarily inefficient if
3024 dh->length (recommended exponent length) is much smaller than the
3025 length of dh->p. We could use BN_rand_range() if the order of
3026 the subgroup was stored in the DH structure, but we only have
3031 BN_rand(priv_key, l, ...)
3033 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
3039 RSA_eay_public_encrypt
3040 RSA_eay_private_decrypt
3041 RSA_eay_private_encrypt (signing)
3042 RSA_eay_public_decrypt (signature verification)
3044 (default implementations for RSA_public_encrypt,
3045 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
3046 always reject numbers >= n.
3049 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
3050 to synchronize access to 'locking_thread'. This is necessary on
3051 systems where access to 'locking_thread' (an 'unsigned long'
3052 variable) is not atomic.
3055 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
3056 *before* setting the 'crypto_lock_rand' flag. The previous code had
3057 a race condition if 0 is a valid thread ID.
3058 [Travis Vitek <vitek@roguewave.com>]
3060 *) Add support for shared libraries under Irix.
3061 [Albert Chin-A-Young <china@thewrittenword.com>]
3063 *) Add configuration option to build on Linux on both big-endian and
3065 [Ralf Baechle <ralf@uni-koblenz.de>]
3067 *) Add the possibility to create shared libraries on HP-UX.
3070 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
3072 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
3073 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
3074 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
3075 PRNG state recovery was possible based on the output of
3076 one PRNG request appropriately sized to gain knowledge on
3077 'md' followed by enough consecutive 1-byte PRNG requests
3078 to traverse all of 'state'.
3080 1. When updating 'md_local' (the current thread's copy of 'md')
3081 during PRNG output generation, hash all of the previous
3082 'md_local' value, not just the half used for PRNG output.
3084 2. Make the number of bytes from 'state' included into the hash
3085 independent from the number of PRNG bytes requested.
3087 The first measure alone would be sufficient to avoid
3088 Markku-Juhani's attack. (Actually it had never occurred
3089 to me that the half of 'md_local' used for chaining was the
3090 half from which PRNG output bytes were taken -- I had always
3091 assumed that the secret half would be used.) The second
3092 measure makes sure that additional data from 'state' is never
3093 mixed into 'md_local' in small portions; this heuristically
3094 further strengthens the PRNG.
3097 *) Fix crypto/bn/asm/mips3.s.
3100 *) When only the key is given to "enc", the IV is undefined. Print out
3101 an error message in this case.
3104 *) Handle special case when X509_NAME is empty in X509 printing routines.
3107 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
3108 positive and less than q.
3111 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
3112 used: it isn't thread safe and the add_lock_callback should handle
3114 [Paul Rose <Paul.Rose@bridge.com>]
3116 *) Verify that incoming data obeys the block size in
3117 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
3121 [Ulf Möller, Bodo Möller]
3123 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
3124 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
3125 when fixing the server behaviour for backwards-compatible 'client
3126 hello' messages. (Note that the attack is impractical against
3127 SSL 3.0 and TLS 1.0 anyway because length and version checking
3128 means that the probability of guessing a valid ciphertext is
3129 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
3132 Before 0.9.5, the countermeasure (hide the error by generating a
3133 random 'decryption result') did not work properly because
3134 ERR_clear_error() was missing, meaning that SSL_get_error() would
3135 detect the supposedly ignored error.
3137 Both problems are now fixed.
3140 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
3141 (previously it was 1024).
3144 *) Fix for compatibility mode trust settings: ignore trust settings
3145 unless some valid trust or reject settings are present.
3148 *) Fix for blowfish EVP: its a variable length cipher.
3151 *) Fix various bugs related to DSA S/MIME verification. Handle missing
3152 parameters in DSA public key structures and return an error in the
3153 DSA routines if parameters are absent.
3156 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
3157 in the current directory if neither $RANDFILE nor $HOME was set.
3158 RAND_file_name() in 0.9.6a returned NULL in this case. This has
3159 caused some confusion to Windows users who haven't defined $HOME.
3160 Thus RAND_file_name() is changed again: e_os.h can define a
3161 DEFAULT_HOME, which will be used if $HOME is not set.
3162 For Windows, we use "C:"; on other platforms, we still require
3163 environment variables.
3165 *) Move 'if (!initialized) RAND_poll()' into regions protected by
3166 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
3167 having multiple threads call RAND_poll() concurrently.
3170 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
3171 combination of a flag and a thread ID variable.
3172 Otherwise while one thread is in ssleay_rand_bytes (which sets the
3173 flag), *other* threads can enter ssleay_add_bytes without obeying
3174 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
3175 that they do not hold after the first thread unsets add_do_not_lock).
3178 *) Change bctest again: '-x' expressions are not available in all
3182 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
3184 *) Fix a couple of memory leaks in PKCS7_dataDecode()
3185 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
3187 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
3188 the default extension for executables, if any. Also, make the perl
3189 scripts that use symlink() to test if it really exists and use "cp"
3190 if it doesn't. All this made OpenSSL compilable and installable in
3194 *) Fix for asn1_GetSequence() for indefinite length constructed data.
3195 If SEQUENCE is length is indefinite just set c->slen to the total
3196 amount of data available.
3197 [Steve Henson, reported by shige@FreeBSD.org]
3198 [This change does not apply to 0.9.7.]
3200 *) Change bctest to avoid here-documents inside command substitution
3201 (workaround for FreeBSD /bin/sh bug).
3202 For compatibility with Ultrix, avoid shell functions (introduced
3203 in the bctest version that searches along $PATH).
3206 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
3207 with des_encrypt() defined on some operating systems, like Solaris
3211 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
3212 On the Importance of Eliminating Errors in Cryptographic
3213 Computations, J. Cryptology 14 (2001) 2, 101-119,
3214 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
3217 *) MIPS assembler BIGNUM division bug fix.
3220 *) Disabled incorrect Alpha assembler code.
3223 *) Fix PKCS#7 decode routines so they correctly update the length
3224 after reading an EOC for the EXPLICIT tag.
3226 [This change does not apply to 0.9.7.]
3228 *) Fix bug in PKCS#12 key generation routines. This was triggered
3229 if a 3DES key was generated with a 0 initial byte. Include
3230 PKCS12_BROKEN_KEYGEN compilation option to retain the old
3231 (but broken) behaviour.
3234 *) Enhance bctest to search for a working bc along $PATH and print
3236 [Tim Rice <tim@multitalents.net> via Richard Levitte]
3238 *) Fix memory leaks in err.c: free err_data string if necessary;
3239 don't write to the wrong index in ERR_set_error_data.
3242 *) Implement ssl23_peek (analogous to ssl23_read), which previously
3246 *) Replace rdtsc with _emit statements for VC++ version 5.
3247 [Jeremy Cooper <jeremy@baymoo.org>]
3249 *) Make it possible to reuse SSLv2 sessions.
3252 *) In copy_email() check for >= 0 as a return value for
3253 X509_NAME_get_index_by_NID() since 0 is a valid index.
3254 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
3256 *) Avoid coredump with unsupported or invalid public keys by checking if
3257 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
3258 PKCS7_verify() fails with non detached data.
3261 *) Don't use getenv in library functions when run as setuid/setgid.
3262 New function OPENSSL_issetugid().
3265 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
3266 due to incorrect handling of multi-threading:
3268 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
3270 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
3272 3. Count how many times MemCheck_off() has been called so that
3273 nested use can be treated correctly. This also avoids
3274 inband-signalling in the previous code (which relied on the
3275 assumption that thread ID 0 is impossible).
3278 *) Add "-rand" option also to s_client and s_server.
3281 *) Fix CPU detection on Irix 6.x.
3282 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
3283 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3285 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
3288 [This change does not apply to 0.9.7.]
3290 *) Use the cached encoding of an X509_NAME structure rather than
3291 copying it. This is apparently the reason for the libsafe "errors"
3292 but the code is actually correct.
3295 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
3296 Bleichenbacher's DSA attack.
3297 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
3298 to be set and top=0 forces the highest bit to be set; top=-1 is new
3299 and leaves the highest bit random.
3300 [Ulf Moeller, Bodo Moeller]
3302 *) In the NCONF_...-based implementations for CONF_... queries
3303 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
3304 a temporary CONF structure with the data component set to NULL
3305 (which gives segmentation faults in lh_retrieve).
3306 Instead, use NULL for the CONF pointer in CONF_get_string and
3307 CONF_get_number (which may use environment variables) and directly
3308 return NULL from CONF_get_section.
3311 *) Fix potential buffer overrun for EBCDIC.
3314 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
3315 keyUsage if basicConstraints absent for a CA.
3318 *) Make SMIME_write_PKCS7() write mail header values with a format that
3319 is more generally accepted (no spaces before the semicolon), since
3320 some programs can't parse those values properly otherwise. Also make
3321 sure BIO's that break lines after each write do not create invalid
3325 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
3326 macros previously used would not encode an empty SEQUENCE OF
3327 and break the signature.
3329 [This change does not apply to 0.9.7.]
3331 *) Zero the premaster secret after deriving the master secret in
3335 *) Add some EVP_add_digest_alias registrations (as found in
3336 OpenSSL_add_all_digests()) to SSL_library_init()
3337 aka OpenSSL_add_ssl_algorithms(). This provides improved
3338 compatibility with peers using X.509 certificates
3339 with unconventional AlgorithmIdentifier OIDs.
3342 *) Fix for Irix with NO_ASM.
3343 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3345 *) ./config script fixes.
3346 [Ulf Moeller, Richard Levitte]
3348 *) Fix 'openssl passwd -1'.
3351 *) Change PKCS12_key_gen_asc() so it can cope with non null
3352 terminated strings whose length is passed in the passlen
3353 parameter, for example from PEM callbacks. This was done
3354 by adding an extra length parameter to asc2uni().
3355 [Steve Henson, reported by <oddissey@samsung.co.kr>]
3357 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
3358 call failed, free the DSA structure.
3361 *) Fix to uni2asc() to cope with zero length Unicode strings.
3362 These are present in some PKCS#12 files.
3365 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
3366 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
3367 when writing a 32767 byte record.
3368 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
3370 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
3371 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
3373 (RSA objects have a reference count access to which is protected
3374 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
3375 so they are meant to be shared between threads.)
3376 [Bodo Moeller, Geoff Thorpe; original patch submitted by
3377 "Reddie, Steven" <Steven.Reddie@ca.com>]
3379 *) Fix a deadlock in CRYPTO_mem_leaks().
3382 *) Use better test patterns in bntest.
3385 *) rand_win.c fix for Borland C.
3388 *) BN_rshift bugfix for n == 0.
3391 *) Add a 'bctest' script that checks for some known 'bc' bugs
3392 so that 'make test' does not abort just because 'bc' is broken.
3395 *) Store verify_result within SSL_SESSION also for client side to
3396 avoid potential security hole. (Re-used sessions on the client side
3397 always resulted in verify_result==X509_V_OK, not using the original
3398 result of the server certificate verification.)
3401 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
3402 SSL3_RT_APPLICATION_DATA, return 0.
3403 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
3407 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
3408 releases, have been re-implemented by renaming the previous
3409 implementations of ssl2_read and ssl3_read to ssl2_read_internal
3410 and ssl3_read_internal, respectively, and adding 'peek' parameters
3411 to them. The new ssl[23]_{read,peek} functions are calls to
3412 ssl[23]_read_internal with the 'peek' flag set appropriately.
3413 A 'peek' parameter has also been added to ssl3_read_bytes, which
3414 does the actual work for ssl3_read_internal.
3417 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
3418 the method-specific "init()" handler. Also clean up ex_data after
3419 calling the method-specific "finish()" handler. Previously, this was
3420 happening the other way round.
3423 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
3424 The previous value, 12, was not always sufficient for BN_mod_exp().
3427 *) Make sure that shared libraries get the internal name engine with
3428 the full version number and not just 0. This should mark the
3429 shared libraries as not backward compatible. Of course, this should
3430 be changed again when we can guarantee backward binary compatibility.
3433 *) Fix typo in get_cert_by_subject() in by_dir.c
3434 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
3436 *) Rework the system to generate shared libraries:
3438 - Make note of the expected extension for the shared libraries and
3439 if there is a need for symbolic links from for example libcrypto.so.0
3440 to libcrypto.so.0.9.7. There is extended info in Configure for
3443 - Make as few rebuilds of the shared libraries as possible.
3445 - Still avoid linking the OpenSSL programs with the shared libraries.
3447 - When installing, install the shared libraries separately from the
3451 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
3453 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
3454 and not in SSL_clear because the latter is also used by the
3455 accept/connect functions; previously, the settings made by
3456 SSL_set_read_ahead would be lost during the handshake.
3457 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
3459 *) Correct util/mkdef.pl to be selective about disabled algorithms.
3460 Previously, it would create entries for disableed algorithms no
3464 *) Added several new manual pages for SSL_* function.
3467 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
3469 *) In ssl23_get_client_hello, generate an error message when faced
3470 with an initial SSL 3.0/TLS record that is too small to contain the
3471 first two bytes of the ClientHello message, i.e. client_version.
3472 (Note that this is a pathologic case that probably has never happened
3473 in real life.) The previous approach was to use the version number
3474 from the record header as a substitute; but our protocol choice
3475 should not depend on that one because it is not authenticated
3476 by the Finished messages.
3479 *) More robust randomness gathering functions for Windows.
3480 [Jeffrey Altman <jaltman@columbia.edu>]
3482 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
3483 not set then we don't setup the error code for issuer check errors
3484 to avoid possibly overwriting other errors which the callback does
3485 handle. If an application does set the flag then we assume it knows
3486 what it is doing and can handle the new informational codes
3490 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
3491 a general "ANY" type, as such it should be able to decode anything
3492 including tagged types. However it didn't check the class so it would
3493 wrongly interpret tagged types in the same way as their universal
3494 counterpart and unknown types were just rejected. Changed so that the
3495 tagged and unknown types are handled in the same way as a SEQUENCE:
3496 that is the encoding is stored intact. There is also a new type
3497 "V_ASN1_OTHER" which is used when the class is not universal, in this
3498 case we have no idea what the actual type is so we just lump them all
3502 *) On VMS, stdout may very well lead to a file that is written to
3503 in a record-oriented fashion. That means that every write() will
3504 write a separate record, which will be read separately by the
3505 programs trying to read from it. This can be very confusing.
3507 The solution is to put a BIO filter in the way that will buffer
3508 text until a linefeed is reached, and then write everything a
3509 line at a time, so every record written will be an actual line,
3510 not chunks of lines and not (usually doesn't happen, but I've
3511 seen it once) several lines in one record. BIO_f_linebuffer() is
3514 Currently, it's a VMS-only method, because that's where it has
3515 been tested well enough.
3518 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
3519 it can return incorrect results.
3520 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
3521 but it was in 0.9.6-beta[12].)
3524 *) Disable the check for content being present when verifying detached
3525 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
3526 include zero length content when signing messages.
3529 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
3530 BIO_ctrl (for BIO pairs).
3533 *) Add DSO method for VMS.
3536 *) Bug fix: Montgomery multiplication could produce results with the
3540 *) Add RPM specification openssl.spec and modify it to build three
3541 packages. The default package contains applications, application
3542 documentation and run-time libraries. The devel package contains
3543 include files, static libraries and function documentation. The
3544 doc package contains the contents of the doc directory. The original
3545 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
3548 *) Add a large number of documentation files for many SSL routines.
3549 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3551 *) Add a configuration entry for Sony News 4.
3552 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
3554 *) Don't set the two most significant bits to one when generating a
3555 random number < q in the DSA library.
3558 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
3559 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
3560 the underlying transport is blocking) if a handshake took place.
3561 (The default behaviour is needed by applications such as s_client
3562 and s_server that use select() to determine when to use SSL_read;
3563 but for applications that know in advance when to expect data, it
3564 just makes things more complicated.)
3567 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
3571 *) Add a few more EBCDIC conditionals that make `req' and `x509'
3572 work better on such systems.
3573 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3575 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
3576 Update PKCS12_parse() so it copies the friendlyName and the
3577 keyid to the certificates aux info.
3580 *) Fix bug in PKCS7_verify() which caused an infinite loop
3581 if there was more than one signature.
3582 [Sven Uszpelkat <su@celocom.de>]
3584 *) Major change in util/mkdef.pl to include extra information
3585 about each symbol, as well as presentig variables as well
3586 as functions. This change means that there's n more need
3587 to rebuild the .num files when some algorithms are excluded.
3590 *) Allow the verify time to be set by an application,
3591 rather than always using the current time.
3594 *) Phase 2 verify code reorganisation. The certificate
3595 verify code now looks up an issuer certificate by a
3596 number of criteria: subject name, authority key id
3597 and key usage. It also verifies self signed certificates
3598 by the same criteria. The main comparison function is
3599 X509_check_issued() which performs these checks.
3601 Lot of changes were necessary in order to support this
3602 without completely rewriting the lookup code.
3604 Authority and subject key identifier are now cached.
3606 The LHASH 'certs' is X509_STORE has now been replaced
3607 by a STACK_OF(X509_OBJECT). This is mainly because an
3608 LHASH can't store or retrieve multiple objects with
3609 the same hash value.
3611 As a result various functions (which were all internal
3612 use only) have changed to handle the new X509_STORE
3613 structure. This will break anything that messed round
3614 with X509_STORE internally.
3616 The functions X509_STORE_add_cert() now checks for an
3617 exact match, rather than just subject name.
3619 The X509_STORE API doesn't directly support the retrieval
3620 of multiple certificates matching a given criteria, however
3621 this can be worked round by performing a lookup first
3622 (which will fill the cache with candidate certificates)
3623 and then examining the cache for matches. This is probably
3624 the best we can do without throwing out X509_LOOKUP
3625 entirely (maybe later...).
3627 The X509_VERIFY_CTX structure has been enhanced considerably.
3629 All certificate lookup operations now go via a get_issuer()
3630 callback. Although this currently uses an X509_STORE it
3631 can be replaced by custom lookups. This is a simple way
3632 to bypass the X509_STORE hackery necessary to make this
3633 work and makes it possible to use more efficient techniques
3634 in future. A very simple version which uses a simple
3635 STACK for its trusted certificate store is also provided
3636 using X509_STORE_CTX_trusted_stack().
3638 The verify_cb() and verify() callbacks now have equivalents
3639 in the X509_STORE_CTX structure.
3641 X509_STORE_CTX also has a 'flags' field which can be used
3642 to customise the verify behaviour.
3645 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
3646 excludes S/MIME capabilities.
3649 *) When a certificate request is read in keep a copy of the
3650 original encoding of the signed data and use it when outputing
3651 again. Signatures then use the original encoding rather than
3652 a decoded, encoded version which may cause problems if the
3653 request is improperly encoded.
3656 *) For consistency with other BIO_puts implementations, call
3657 buffer_write(b, ...) directly in buffer_puts instead of calling
3660 In BIO_puts, increment b->num_write as in BIO_write.
3661 [Peter.Sylvester@EdelWeb.fr]
3663 *) Fix BN_mul_word for the case where the word is 0. (We have to use
3664 BN_zero, we may not return a BIGNUM with an array consisting of
3668 *) Avoid calling abort() from within the library when problems are
3669 detected, except if preprocessor symbols have been defined
3670 (such as REF_CHECK, BN_DEBUG etc.).
3673 *) New openssl application 'rsautl'. This utility can be
3674 used for low level RSA operations. DER public key
3675 BIO/fp routines also added.
3678 *) New Configure entry and patches for compiling on QNX 4.
3679 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
3681 *) A demo state-machine implementation was sponsored by
3682 Nuron (http://www.nuron.com/) and is now available in
3683 demos/state_machine.
3686 *) New options added to the 'dgst' utility for signature
3687 generation and verification.
3690 *) Unrecognized PKCS#7 content types are now handled via a
3691 catch all ASN1_TYPE structure. This allows unsupported
3692 types to be stored as a "blob" and an application can
3693 encode and decode it manually.
3696 *) Fix various signed/unsigned issues to make a_strex.c
3698 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
3700 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
3701 length if passed a buffer. ASN1_INTEGER_to_BN failed
3702 if passed a NULL BN and its argument was negative.
3703 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
3705 *) Modification to PKCS#7 encoding routines to output definite
3706 length encoding. Since currently the whole structures are in
3707 memory there's not real point in using indefinite length
3708 constructed encoding. However if OpenSSL is compiled with
3709 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
3712 *) Added BIO_vprintf() and BIO_vsnprintf().
3715 *) Added more prefixes to parse for in the the strings written
3716 through a logging bio, to cover all the levels that are available
3717 through syslog. The prefixes are now:
3719 PANIC, EMERG, EMR => LOG_EMERG
3720 ALERT, ALR => LOG_ALERT
3721 CRIT, CRI => LOG_CRIT
3722 ERROR, ERR => LOG_ERR
3723 WARNING, WARN, WAR => LOG_WARNING
3724 NOTICE, NOTE, NOT => LOG_NOTICE
3725 INFO, INF => LOG_INFO
3726 DEBUG, DBG => LOG_DEBUG
3728 and as before, if none of those prefixes are present at the
3729 beginning of the string, LOG_ERR is chosen.
3731 On Win32, the LOG_* levels are mapped according to this:
3733 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
3734 LOG_WARNING => EVENTLOG_WARNING_TYPE
3735 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
3739 *) Made it possible to reconfigure with just the configuration
3740 argument "reconf" or "reconfigure". The command line arguments
3741 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
3742 and are retrieved from there when reconfiguring.
3746 [Assar Westerlund <assar@sics.se>, Richard Levitte]
3748 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
3751 *) The obj_dat.pl script was messing up the sorting of object
3752 names. The reason was that it compared the quoted version
3753 of strings as a result "OCSP" > "OCSP Signing" because
3754 " > SPACE. Changed script to store unquoted versions of
3755 names and add quotes on output. It was also omitting some
3756 names from the lookup table if they were given a default
3757 value (that is if SN is missing it is given the same
3758 value as LN and vice versa), these are now added on the
3759 grounds that if an object has a name we should be able to
3760 look it up. Finally added warning output when duplicate
3761 short or long names are found.
3764 *) Changes needed for Tandem NSK.
3765 [Scott Uroff <scott@xypro.com>]
3767 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
3768 RSA_padding_check_SSLv23(), special padding was never detected
3769 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
3770 version rollback attacks was not effective.
3772 In s23_clnt.c, don't use special rollback-attack detection padding
3773 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
3774 client; similarly, in s23_srvr.c, don't do the rollback check if
3775 SSL 2.0 is the only protocol enabled in the server.
3778 *) Make it possible to get hexdumps of unprintable data with 'openssl
3779 asn1parse'. By implication, the functions ASN1_parse_dump() and
3780 BIO_dump_indent() are added.
3783 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
3784 these print out strings and name structures based on various
3785 flags including RFC2253 support and proper handling of
3786 multibyte characters. Added options to the 'x509' utility
3787 to allow the various flags to be set.
3790 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
3791 Also change the functions X509_cmp_current_time() and
3792 X509_gmtime_adj() work with an ASN1_TIME structure,
3793 this will enable certificates using GeneralizedTime in validity
3794 dates to be checked.
3797 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
3798 negative public key encodings) on by default,
3799 NO_NEG_PUBKEY_BUG can be set to disable it.
3802 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
3803 content octets. An i2c_ASN1_OBJECT is unnecessary because
3804 the encoding can be trivially obtained from the structure.
3807 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
3808 not read locks (CRYPTO_r_[un]lock).
3811 *) A first attempt at creating official support for shared
3812 libraries through configuration. I've kept it so the
3813 default is static libraries only, and the OpenSSL programs
3814 are always statically linked for now, but there are
3815 preparations for dynamic linking in place.
3816 This has been tested on Linux and Tru64.
3819 *) Randomness polling function for Win9x, as described in:
3820 Peter Gutmann, Software Generation of Practically Strong
3824 *) Fix so PRNG is seeded in req if using an already existing
3828 *) New options to smime application. -inform and -outform
3829 allow alternative formats for the S/MIME message including
3830 PEM and DER. The -content option allows the content to be
3831 specified separately. This should allow things like Netscape
3832 form signing output easier to verify.
3835 *) Fix the ASN1 encoding of tags using the 'long form'.
3838 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
3839 STRING types. These convert content octets to and from the
3840 underlying type. The actual tag and length octets are
3841 already assumed to have been read in and checked. These
3842 are needed because all other string types have virtually
3843 identical handling apart from the tag. By having versions
3844 of the ASN1 functions that just operate on content octets
3845 IMPLICIT tagging can be handled properly. It also allows
3846 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
3847 and ASN1_INTEGER are identical apart from the tag.
3850 *) Change the handling of OID objects as follows:
3852 - New object identifiers are inserted in objects.txt, following
3853 the syntax given in objects.README.
3854 - objects.pl is used to process obj_mac.num and create a new
3856 - obj_dat.pl is used to create a new obj_dat.h, using the data in
3859 This is currently kind of a hack, and the perl code in objects.pl
3860 isn't very elegant, but it works as I intended. The simplest way
3861 to check that it worked correctly is to look in obj_dat.h and
3862 check the array nid_objs and make sure the objects haven't moved
3863 around (this is important!). Additions are OK, as well as
3864 consistent name changes.
3867 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
3870 *) Addition of the command line parameter '-rand file' to 'openssl req'.
3871 The given file adds to whatever has already been seeded into the
3872 random pool through the RANDFILE configuration file option or
3873 environment variable, or the default random state file.
3876 *) mkstack.pl now sorts each macro group into lexical order.
3877 Previously the output order depended on the order the files
3878 appeared in the directory, resulting in needless rewriting
3882 *) Patches to make OpenSSL compile under Win32 again. Mostly
3883 work arounds for the VC++ problem that it treats func() as
3884 func(void). Also stripped out the parts of mkdef.pl that
3885 added extra typesafe functions: these no longer exist.
3888 *) Reorganisation of the stack code. The macros are now all
3889 collected in safestack.h . Each macro is defined in terms of
3890 a "stack macro" of the form SKM_<name>(type, a, b). The
3891 DEBUG_SAFESTACK is now handled in terms of function casts,
3892 this has the advantage of retaining type safety without the
3893 use of additional functions. If DEBUG_SAFESTACK is not defined
3894 then the non typesafe macros are used instead. Also modified the
3895 mkstack.pl script to handle the new form. Needs testing to see
3896 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
3897 the default if no major problems. Similar behaviour for ASN1_SET_OF
3898 and PKCS12_STACK_OF.
3901 *) When some versions of IIS use the 'NET' form of private key the
3902 key derivation algorithm is different. Normally MD5(password) is
3903 used as a 128 bit RC4 key. In the modified case
3904 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
3905 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
3906 as the old Netscape_RSA functions except they have an additional
3907 'sgckey' parameter which uses the modified algorithm. Also added
3908 an -sgckey command line option to the rsa utility. Thanks to
3909 Adrian Peck <bertie@ncipher.com> for posting details of the modified
3910 algorithm to openssl-dev.
3913 *) The evp_local.h macros were using 'c.##kname' which resulted in
3914 invalid expansion on some systems (SCO 5.0.5 for example).
3915 Corrected to 'c.kname'.
3916 [Phillip Porch <root@theporch.com>]
3918 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
3919 a STACK of email addresses from a certificate or request, these look
3920 in the subject name and the subject alternative name extensions and
3921 omit any duplicate addresses.
3924 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
3925 This makes DSA verification about 2 % faster.
3928 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
3929 (meaning that now 2^5 values will be precomputed, which is only 4 KB
3930 plus overhead for 1024 bit moduli).
3931 This makes exponentiations about 0.5 % faster for 1024 bit
3932 exponents (as measured by "openssl speed rsa2048").
3935 *) Rename memory handling macros to avoid conflicts with other
3937 Malloc => OPENSSL_malloc
3938 Malloc_locked => OPENSSL_malloc_locked
3939 Realloc => OPENSSL_realloc
3940 Free => OPENSSL_free
3943 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
3944 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
3947 *) CygWin32 support.
3948 [John Jarvie <jjarvie@newsguy.com>]
3950 *) The type-safe stack code has been rejigged. It is now only compiled
3951 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
3952 by default all type-specific stack functions are "#define"d back to
3953 standard stack functions. This results in more streamlined output
3954 but retains the type-safety checking possibilities of the original
3958 *) The STACK code has been cleaned up, and certain type declarations
3959 that didn't make a lot of sense have been brought in line. This has
3960 also involved a cleanup of sorts in safestack.h to more correctly
3961 map type-safe stack functions onto their plain stack counterparts.
3962 This work has also resulted in a variety of "const"ifications of
3963 lots of the code, especially "_cmp" operations which should normally
3964 be prototyped with "const" parameters anyway.
3967 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
3968 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
3969 (The PRNG state consists of two parts, the large pool 'state' and 'md',
3970 where all of 'md' is used each time the PRNG is used, but 'state'
3971 is used only indexed by a cyclic counter. As entropy may not be
3972 well distributed from the beginning, 'md' is important as a
3973 chaining variable. However, the output function chains only half
3974 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
3975 all of 'md', and seeding with STATE_SIZE dummy bytes will result
3976 in all of 'state' being rewritten, with the new values depending
3977 on virtually all of 'md'. This overcomes the 80 bit limitation.)
3980 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
3981 the handshake is continued after ssl_verify_cert_chain();
3982 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
3983 can lead to 'unexplainable' connection aborts later.
3984 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
3986 *) Major EVP API cipher revision.
3987 Add hooks for extra EVP features. This allows various cipher
3988 parameters to be set in the EVP interface. Support added for variable
3989 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
3990 setting of RC2 and RC5 parameters.
3992 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
3995 Remove lots of duplicated code from the EVP library. For example *every*
3996 cipher init() function handles the 'iv' in the same way according to the
3997 cipher mode. They also all do nothing if the 'key' parameter is NULL and
3998 for CFB and OFB modes they zero ctx->num.
4000 New functionality allows removal of S/MIME code RC2 hack.
4002 Most of the routines have the same form and so can be declared in terms
4005 By shifting this to the top level EVP_CipherInit() it can be removed from
4006 all individual ciphers. If the cipher wants to handle IVs or keys
4007 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
4010 Change lots of functions like EVP_EncryptUpdate() to now return a
4011 value: although software versions of the algorithms cannot fail
4012 any installed hardware versions can.
4015 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
4016 this option is set, tolerate broken clients that send the negotiated
4017 protocol version number instead of the requested protocol version
4021 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
4022 i.e. non-zero for export ciphersuites, zero otherwise.
4023 Previous versions had this flag inverted, inconsistent with
4024 rsa_tmp_cb (..._TMP_RSA_CB).
4025 [Bodo Moeller; problem reported by Amit Chopra]
4027 *) Add missing DSA library text string. Work around for some IIS
4028 key files with invalid SEQUENCE encoding.
4031 *) Add a document (doc/standards.txt) that list all kinds of standards
4032 and so on that are implemented in OpenSSL.
4035 *) Enhance c_rehash script. Old version would mishandle certificates
4036 with the same subject name hash and wouldn't handle CRLs at all.
4037 Added -fingerprint option to crl utility, to support new c_rehash
4041 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
4044 *) Fix for SSL server purpose checking. Server checking was
4045 rejecting certificates which had extended key usage present
4046 but no ssl client purpose.
4047 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
4049 *) Make PKCS#12 code work with no password. The PKCS#12 spec
4050 is a little unclear about how a blank password is handled.
4051 Since the password in encoded as a BMPString with terminating
4052 double NULL a zero length password would end up as just the
4053 double NULL. However no password at all is different and is
4054 handled differently in the PKCS#12 key generation code. NS
4055 treats a blank password as zero length. MSIE treats it as no
4056 password on export: but it will try both on import. We now do
4057 the same: PKCS12_parse() tries zero length and no password if
4058 the password is set to "" or NULL (NULL is now a valid password:
4059 it wasn't before) as does the pkcs12 application.
4062 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
4063 perror when PEM_read_bio_X509_REQ fails, the error message must
4064 be obtained from the error queue.
4067 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
4068 it in ERR_remove_state if appropriate, and change ERR_get_state
4069 accordingly to avoid race conditions (this is necessary because
4070 thread_hash is no longer constant once set).
4073 *) Bugfix for linux-elf makefile.one.
4076 *) RSA_get_default_method() will now cause a default
4077 RSA_METHOD to be chosen if one doesn't exist already.
4078 Previously this was only set during a call to RSA_new()
4079 or RSA_new_method(NULL) meaning it was possible for
4080 RSA_get_default_method() to return NULL.
4083 *) Added native name translation to the existing DSO code
4084 that will convert (if the flag to do so is set) filenames
4085 that are sufficiently small and have no path information
4086 into a canonical native form. Eg. "blah" converted to
4087 "libblah.so" or "blah.dll" etc.
4090 *) New function ERR_error_string_n(e, buf, len) which is like
4091 ERR_error_string(e, buf), but writes at most 'len' bytes
4092 including the 0 terminator. For ERR_error_string_n, 'buf'
4094 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
4096 *) CONF library reworked to become more general. A new CONF
4097 configuration file reader "class" is implemented as well as a
4098 new functions (NCONF_*, for "New CONF") to handle it. The now
4099 old CONF_* functions are still there, but are reimplemented to
4100 work in terms of the new functions. Also, a set of functions
4101 to handle the internal storage of the configuration data is
4102 provided to make it easier to write new configuration file
4103 reader "classes" (I can definitely see something reading a
4104 configuration file in XML format, for example), called _CONF_*,
4105 or "the configuration storage API"...
4107 The new configuration file reading functions are:
4109 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
4110 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
4112 NCONF_default, NCONF_WIN32
4114 NCONF_dump_fp, NCONF_dump_bio
4116 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
4117 NCONF_new creates a new CONF object. This works in the same way
4118 as other interfaces in OpenSSL, like the BIO interface.
4119 NCONF_dump_* dump the internal storage of the configuration file,
4120 which is useful for debugging. All other functions take the same
4121 arguments as the old CONF_* functions wth the exception of the
4122 first that must be a `CONF *' instead of a `LHASH *'.
4124 To make it easer to use the new classes with the old CONF_* functions,
4125 the function CONF_set_default_method is provided.
4128 *) Add '-tls1' option to 'openssl ciphers', which was already
4129 mentioned in the documentation but had not been implemented.
4130 (This option is not yet really useful because even the additional
4131 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
4134 *) Initial DSO code added into libcrypto for letting OpenSSL (and
4135 OpenSSL-based applications) load shared libraries and bind to
4136 them in a portable way.
4137 [Geoff Thorpe, with contributions from Richard Levitte]
4139 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
4141 *) Make sure _lrotl and _lrotr are only used with MSVC.
4143 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
4144 (the default implementation of RAND_status).
4146 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
4147 to '-clrext' (= clear extensions), as intended and documented.
4148 [Bodo Moeller; inconsistency pointed out by Michael Attili
4151 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
4152 was larger than the MD block size.
4153 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
4155 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
4156 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
4157 using the passed key: if the passed key was a private key the result
4158 of X509_print(), for example, would be to print out all the private key
4162 *) des_quad_cksum() byte order bug fix.
4163 [Ulf Möller, using the problem description in krb4-0.9.7, where
4164 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
4166 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
4168 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
4170 *) For easily testing in shell scripts whether some command
4171 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
4172 returns with exit code 0 iff no command of the given name is available.
4173 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
4174 the output goes to stdout and nothing is printed to stderr.
4175 Additional arguments are always ignored.
4177 Since for each cipher there is a command of the same name,
4178 the 'no-cipher' compilation switches can be tested this way.
4180 ('openssl no-XXX' is not able to detect pseudo-commands such
4181 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
4184 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
4187 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
4188 is set; it will be thrown away anyway because each handshake creates
4190 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
4191 to parameters -- in previous versions (since OpenSSL 0.9.3) the
4192 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
4193 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
4196 *) New s_client option -ign_eof: EOF at stdin is ignored, and
4197 'Q' and 'R' lose their special meanings (quit/renegotiate).
4198 This is part of what -quiet does; unlike -quiet, -ign_eof
4199 does not suppress any output.
4202 *) Add compatibility options to the purpose and trust code. The
4203 purpose X509_PURPOSE_ANY is "any purpose" which automatically
4204 accepts a certificate or CA, this was the previous behaviour,
4205 with all the associated security issues.
4207 X509_TRUST_COMPAT is the old trust behaviour: only and
4208 automatically trust self signed roots in certificate store. A
4209 new trust setting X509_TRUST_DEFAULT is used to specify that
4210 a purpose has no associated trust setting and it should instead
4211 use the value in the default purpose.
4214 *) Fix the PKCS#8 DSA private key code so it decodes keys again
4215 and fix a memory leak.
4218 *) In util/mkerr.pl (which implements 'make errors'), preserve
4219 reason strings from the previous version of the .c file, as
4220 the default to have only downcase letters (and digits) in
4221 automatically generated reasons codes is not always appropriate.
4224 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
4225 using strerror. Previously, ERR_reason_error_string() returned
4226 library names as reason strings for SYSerr; but SYSerr is a special
4227 case where small numbers are errno values, not library numbers.
4230 *) Add '-dsaparam' option to 'openssl dhparam' application. This
4231 converts DSA parameters into DH parameters. (When creating parameters,
4232 DSA_generate_parameters is used.)
4235 *) Include 'length' (recommended exponent length) in C code generated
4236 by 'openssl dhparam -C'.
4239 *) The second argument to set_label in perlasm was already being used
4240 so couldn't be used as a "file scope" flag. Moved to third argument
4244 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
4245 instead of RAND_bytes for encryption IVs and salts.
4248 *) Include RAND_status() into RAND_METHOD instead of implementing
4249 it only for md_rand.c Otherwise replacing the PRNG by calling
4250 RAND_set_rand_method would be impossible.
4253 *) Don't let DSA_generate_key() enter an infinite loop if the random
4254 number generation fails.
4257 *) New 'rand' application for creating pseudo-random output.
4260 *) Added configuration support for Linux/IA64
4261 [Rolf Haberrecker <rolf@suse.de>]
4263 *) Assembler module support for Mingw32.
4266 *) Shared library support for HPUX (in shlib/).
4267 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
4269 *) Shared library support for Solaris gcc.
4270 [Lutz Behnke <behnke@trustcenter.de>]
4272 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
4274 *) PKCS7_encrypt() was adding text MIME headers twice because they
4275 were added manually and by SMIME_crlf_copy().
4278 *) In bntest.c don't call BN_rand with zero bits argument.
4279 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
4281 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
4282 case was implemented. This caused BN_div_recp() to fail occasionally.
4285 *) Add an optional second argument to the set_label() in the perl
4286 assembly language builder. If this argument exists and is set
4287 to 1 it signals that the assembler should use a symbol whose
4288 scope is the entire file, not just the current function. This
4289 is needed with MASM which uses the format label:: for this scope.
4290 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
4292 *) Change the ASN1 types so they are typedefs by default. Before
4293 almost all types were #define'd to ASN1_STRING which was causing
4294 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
4298 *) Change names of new functions to the new get1/get0 naming
4299 convention: After 'get1', the caller owns a reference count
4300 and has to call ..._free; 'get0' returns a pointer to some
4301 data structure without incrementing reference counters.
4302 (Some of the existing 'get' functions increment a reference
4303 counter, some don't.)
4304 Similarly, 'set1' and 'add1' functions increase reference
4305 counters or duplicate objects.
4308 *) Allow for the possibility of temp RSA key generation failure:
4309 the code used to assume it always worked and crashed on failure.
4312 *) Fix potential buffer overrun problem in BIO_printf().
4313 [Ulf Möller, using public domain code by Patrick Powell; problem
4314 pointed out by David Sacerdote <das33@cornell.edu>]
4316 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
4317 RAND_egd() and RAND_status(). In the command line application,
4318 the EGD socket can be specified like a seed file using RANDFILE
4322 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
4323 Some CAs (e.g. Verisign) distribute certificates in this form.
4326 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
4327 list to exclude them. This means that no special compilation option
4328 is needed to use anonymous DH: it just needs to be included in the
4332 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
4333 EVP_MD_type. The old functionality is available in a new macro called
4334 EVP_MD_md(). Change code that uses it and update docs.
4337 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
4338 where the 'void *' argument is replaced by a function pointer argument.
4339 Previously 'void *' was abused to point to functions, which works on
4340 many platforms, but is not correct. As these functions are usually
4341 called by macros defined in OpenSSL header files, most source code
4342 should work without changes.
4345 *) <openssl/opensslconf.h> (which is created by Configure) now contains
4346 sections with information on -D... compiler switches used for
4347 compiling the library so that applications can see them. To enable
4348 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
4349 must be defined. E.g.,
4350 #define OPENSSL_ALGORITHM_DEFINES
4351 #include <openssl/opensslconf.h>
4352 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
4353 [Richard Levitte, Ulf and Bodo Möller]
4355 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
4359 *) Change the 'other' type in certificate aux info to a STACK_OF
4360 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
4361 the required ASN1 format: arbitrary types determined by an OID.
4364 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
4365 argument to 'req'. This is not because the function is newer or
4366 better than others it just uses the work 'NEW' in the certificate
4367 request header lines. Some software needs this.
4370 *) Reorganise password command line arguments: now passwords can be
4371 obtained from various sources. Delete the PEM_cb function and make
4372 it the default behaviour: i.e. if the callback is NULL and the
4373 usrdata argument is not NULL interpret it as a null terminated pass
4374 phrase. If usrdata and the callback are NULL then the pass phrase
4375 is prompted for as usual.
4378 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
4379 the support is automatically enabled. The resulting binaries will
4380 autodetect the card and use it if present.
4381 [Ben Laurie and Compaq Inc.]
4383 *) Work around for Netscape hang bug. This sends certificate request
4384 and server done in one record. Since this is perfectly legal in the
4385 SSL/TLS protocol it isn't a "bug" option and is on by default. See
4386 the bugs/SSLv3 entry for more info.
4389 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
4392 *) Add -rand argument to smime and pkcs12 applications and read/write
4396 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
4399 *) Add command line password options to the remaining applications.
4402 *) Bug fix for BN_div_recp() for numerators with an even number of
4406 *) More tests in bntest.c, and changed test_bn output.
4409 *) ./config recognizes MacOS X now.
4412 *) Bug fix for BN_div() when the first words of num and divsor are
4413 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
4416 *) Add support for various broken PKCS#8 formats, and command line
4417 options to produce them.
4420 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
4421 get temporary BIGNUMs from a BN_CTX.
4424 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
4428 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
4429 include a #define from the old name to the new. The original intent
4430 was that statically linked binaries could for example just call
4431 SSLeay_add_all_ciphers() to just add ciphers to the table and not
4432 link with digests. This never worked becayse SSLeay_add_all_digests()
4433 and SSLeay_add_all_ciphers() were in the same source file so calling
4434 one would link with the other. They are now in separate source files.
4437 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
4440 *) Use a less unusual form of the Miller-Rabin primality test (it used
4441 a binary algorithm for exponentiation integrated into the Miller-Rabin
4442 loop, our standard modexp algorithms are faster).
4445 *) Support for the EBCDIC character set completed.
4446 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
4448 *) Source code cleanups: use const where appropriate, eliminate casts,
4449 use void * instead of char * in lhash.
4452 *) Bugfix: ssl3_send_server_key_exchange was not restartable
4453 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
4454 this the server could overwrite ephemeral keys that the client
4458 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
4459 using 50 iterations of the Rabin-Miller test.
4461 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
4462 iterations of the Rabin-Miller test as required by the appendix
4463 to FIPS PUB 186[-1]) instead of DSA_is_prime.
4464 As BN_is_prime_fasttest includes trial division, DSA parameter
4465 generation becomes much faster.
4467 This implies a change for the callback functions in DSA_is_prime
4468 and DSA_generate_parameters: The callback function is called once
4469 for each positive witness in the Rabin-Miller test, not just
4470 occasionally in the inner loop; and the parameters to the
4471 callback function now provide an iteration count for the outer
4472 loop rather than for the current invocation of the inner loop.
4473 DSA_generate_parameters additionally can call the callback
4474 function with an 'iteration count' of -1, meaning that a
4475 candidate has passed the trial division test (when q is generated
4476 from an application-provided seed, trial division is skipped).
4479 *) New function BN_is_prime_fasttest that optionally does trial
4480 division before starting the Rabin-Miller test and has
4481 an additional BN_CTX * argument (whereas BN_is_prime always
4482 has to allocate at least one BN_CTX).
4483 'callback(1, -1, cb_arg)' is called when a number has passed the
4484 trial division stage.
4487 *) Fix for bug in CRL encoding. The validity dates weren't being handled
4491 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
4494 *) New function BN_pseudo_rand().
4497 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
4498 bignum version of BN_from_montgomery() with the working code from
4499 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
4503 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
4504 made it impossible to use the same SSL_SESSION data structure in
4505 SSL2 clients in multiple threads.
4508 *) The return value of RAND_load_file() no longer counts bytes obtained
4509 by stat(). RAND_load_file(..., -1) is new and uses the complete file
4510 to seed the PRNG (previously an explicit byte count was required).
4511 [Ulf Möller, Bodo Möller]
4513 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
4514 used (char *) instead of (void *) and had casts all over the place.
4517 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
4520 *) Retain source code compatibility for BN_prime_checks macro:
4521 BN_is_prime(..., BN_prime_checks, ...) now uses
4522 BN_prime_checks_for_size to determine the appropriate number of
4523 Rabin-Miller iterations.
4526 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
4527 DH_CHECK_P_NOT_SAFE_PRIME.
4528 (Check if this is true? OpenPGP calls them "strong".)
4531 *) Merge the functionality of "dh" and "gendh" programs into a new program
4532 "dhparam". The old programs are retained for now but will handle DH keys
4533 (instead of parameters) in future.
4536 *) Make the ciphers, s_server and s_client programs check the return values
4537 when a new cipher list is set.
4540 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
4541 ciphers. Before when the 56bit ciphers were enabled the sorting was
4544 The syntax for the cipher sorting has been extended to support sorting by
4545 cipher-strength (using the strength_bits hard coded in the tables).
4546 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
4548 Fix a bug in the cipher-command parser: when supplying a cipher command
4549 string with an "undefined" symbol (neither command nor alphanumeric
4550 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
4551 an error is flagged.
4553 Due to the strength-sorting extension, the code of the
4554 ssl_create_cipher_list() function was completely rearranged. I hope that
4555 the readability was also increased :-)
4556 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4558 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
4559 for the first serial number and places 2 in the serial number file. This
4560 avoids problems when the root CA is created with serial number zero and
4561 the first user certificate has the same issuer name and serial number
4565 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
4566 the new code. Add documentation for this stuff.
4569 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
4570 X509_*() to X509at_*() on the grounds that they don't handle X509
4571 structures and behave in an analagous way to the X509v3 functions:
4572 they shouldn't be called directly but wrapper functions should be used
4575 So we also now have some wrapper functions that call the X509at functions
4576 when passed certificate requests. (TO DO: similar things can be done with
4577 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
4578 things. Some of these need some d2i or i2d and print functionality
4579 because they handle more complex structures.)
4582 *) Add missing #ifndefs that caused missing symbols when building libssl
4583 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
4584 NO_RSA in ssl/s2*.c.
4585 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
4587 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
4588 has a return value which indicates the quality of the random data
4589 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
4590 error queue. New function RAND_pseudo_bytes() generates output that is
4591 guaranteed to be unique but not unpredictable. RAND_add is like
4592 RAND_seed, but takes an extra argument for an entropy estimate
4593 (RAND_seed always assumes full entropy).
4596 *) Do more iterations of Rabin-Miller probable prime test (specifically,
4597 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
4598 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
4599 in crypto/bn/bn_prime.c for the complete table). This guarantees a
4600 false-positive rate of at most 2^-80 for random input.
4603 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
4606 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
4607 in the 0.9.5 release), this returns the chain
4608 from an X509_CTX structure with a dup of the stack and all
4609 the X509 reference counts upped: so the stack will exist
4610 after X509_CTX_cleanup() has been called. Modify pkcs12.c
4613 Also make SSL_SESSION_print() print out the verify return
4617 *) Add manpage for the pkcs12 command. Also change the default
4618 behaviour so MAC iteration counts are used unless the new
4619 -nomaciter option is used. This improves file security and
4620 only older versions of MSIE (4.0 for example) need it.
4623 *) Honor the no-xxx Configure options when creating .DEF files.
4626 *) Add PKCS#10 attributes to field table: challengePassword,
4627 unstructuredName and unstructuredAddress. These are taken from
4628 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
4629 international characters are used.
4631 More changes to X509_ATTRIBUTE code: allow the setting of types
4632 based on strings. Remove the 'loc' parameter when adding
4633 attributes because these will be a SET OF encoding which is sorted
4637 *) Initial changes to the 'req' utility to allow request generation
4638 automation. This will allow an application to just generate a template
4639 file containing all the field values and have req construct the
4642 Initial support for X509_ATTRIBUTE handling. Stacks of these are
4643 used all over the place including certificate requests and PKCS#7
4644 structures. They are currently handled manually where necessary with
4645 some primitive wrappers for PKCS#7. The new functions behave in a
4646 manner analogous to the X509 extension functions: they allow
4647 attributes to be looked up by NID and added.
4649 Later something similar to the X509V3 code would be desirable to
4650 automatically handle the encoding, decoding and printing of the
4651 more complex types. The string types like challengePassword can
4652 be handled by the string table functions.
4654 Also modified the multi byte string table handling. Now there is
4655 a 'global mask' which masks out certain types. The table itself
4656 can use the flag STABLE_NO_MASK to ignore the mask setting: this
4657 is useful when for example there is only one permissible type
4658 (as in countryName) and using the mask might result in no valid
4662 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
4663 SSL_get_peer_finished to allow applications to obtain the latest
4664 Finished messages sent to the peer or expected from the peer,
4665 respectively. (SSL_get_peer_finished is usually the Finished message
4666 actually received from the peer, otherwise the protocol will be aborted.)
4668 As the Finished message are message digests of the complete handshake
4669 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
4670 be used for external authentication procedures when the authentication
4671 provided by SSL/TLS is not desired or is not enough.
4674 *) Enhanced support for Alpha Linux is added. Now ./config checks if
4675 the host supports BWX extension and if Compaq C is present on the
4676 $PATH. Just exploiting of the BWX extension results in 20-30%
4677 performance kick for some algorithms, e.g. DES and RC4 to mention
4678 a couple. Compaq C in turn generates ~20% faster code for MD5 and
4682 *) Add support for MS "fast SGC". This is arguably a violation of the
4683 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
4684 weak crypto and after checking the certificate is SGC a second one
4685 with strong crypto. MS SGC stops the first handshake after receiving
4686 the server certificate message and sends a second client hello. Since
4687 a server will typically do all the time consuming operations before
4688 expecting any further messages from the client (server key exchange
4689 is the most expensive) there is little difference between the two.
4691 To get OpenSSL to support MS SGC we have to permit a second client
4692 hello message after we have sent server done. In addition we have to
4693 reset the MAC if we do get this second client hello.
4696 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
4697 if a DER encoded private key is RSA or DSA traditional format. Changed
4698 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
4699 format DER encoded private key. Newer code should use PKCS#8 format which
4700 has the key type encoded in the ASN1 structure. Added DER private key
4701 support to pkcs8 application.
4704 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
4705 ciphersuites has been selected (as required by the SSL 3/TLS 1
4706 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
4707 is set, we interpret this as a request to violate the specification
4708 (the worst that can happen is a handshake failure, and 'correct'
4709 behaviour would result in a handshake failure anyway).
4712 *) In SSL_CTX_add_session, take into account that there might be multiple
4713 SSL_SESSION structures with the same session ID (e.g. when two threads
4714 concurrently obtain them from an external cache).
4715 The internal cache can handle only one SSL_SESSION with a given ID,
4716 so if there's a conflict, we now throw out the old one to achieve
4720 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
4721 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
4722 some routines that use cipher OIDs: some ciphers do not have OIDs
4723 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
4727 *) Simplify the trust setting structure and code. Now we just have
4728 two sequences of OIDs for trusted and rejected settings. These will
4729 typically have values the same as the extended key usage extension
4730 and any application specific purposes.
4732 The trust checking code now has a default behaviour: it will just
4733 check for an object with the same NID as the passed id. Functions can
4734 be provided to override either the default behaviour or the behaviour
4735 for a given id. SSL client, server and email already have functions
4736 in place for compatibility: they check the NID and also return "trusted"
4737 if the certificate is self signed.
4740 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
4741 traditional format into an EVP_PKEY structure.
4744 *) Add a password callback function PEM_cb() which either prompts for
4745 a password if usr_data is NULL or otherwise assumes it is a null
4746 terminated password. Allow passwords to be passed on command line
4747 environment or config files in a few more utilities.
4750 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
4751 keys. Add some short names for PKCS#8 PBE algorithms and allow them
4752 to be specified on the command line for the pkcs8 and pkcs12 utilities.
4753 Update documentation.
4756 *) Support for ASN1 "NULL" type. This could be handled before by using
4757 ASN1_TYPE but there wasn't any function that would try to read a NULL
4758 and produce an error if it couldn't. For compatibility we also have
4759 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
4760 don't allocate anything because they don't need to.
4763 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
4765 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
4767 *) Rebuild of the memory allocation routines used by OpenSSL code and
4768 possibly others as well. The purpose is to make an interface that
4769 provide hooks so anyone can build a separate set of allocation and
4770 deallocation routines to be used by OpenSSL, for example memory
4771 pool implementations, or something else, which was previously hard
4772 since Malloc(), Realloc() and Free() were defined as macros having
4773 the values malloc, realloc and free, respectively (except for Win32
4774 compilations). The same is provided for memory debugging code.
4775 OpenSSL already comes with functionality to find memory leaks, but
4776 this gives people a chance to debug other memory problems.
4778 With these changes, a new set of functions and macros have appeared:
4780 CRYPTO_set_mem_debug_functions() [F]
4781 CRYPTO_get_mem_debug_functions() [F]
4782 CRYPTO_dbg_set_options() [F]
4783 CRYPTO_dbg_get_options() [F]
4784 CRYPTO_malloc_debug_init() [M]
4786 The memory debug functions are NULL by default, unless the library
4787 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
4788 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
4789 gives the standard debugging functions that come with OpenSSL) or
4790 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
4791 provided by the library user) must be used. When the standard
4792 debugging functions are used, CRYPTO_dbg_set_options can be used to
4793 request additional information:
4794 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
4795 the CRYPTO_MDEBUG_xxx macro when compiling the library.
4797 Also, things like CRYPTO_set_mem_functions will always give the
4798 expected result (the new set of functions is used for allocation
4799 and deallocation) at all times, regardless of platform and compiler
4802 To finish it up, some functions that were never use in any other
4803 way than through macros have a new API and new semantic:
4806 CRYPTO_dbg_realloc()
4809 All macros of value have retained their old syntax.
4810 [Richard Levitte and Bodo Moeller]
4812 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
4813 ordering of SMIMECapabilities wasn't in "strength order" and there
4814 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
4818 *) Some ASN1 types with illegal zero length encoding (INTEGER,
4819 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
4820 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
4822 *) Merge in my S/MIME library for OpenSSL. This provides a simple
4823 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
4824 functionality to handle multipart/signed properly) and a utility
4825 called 'smime' to call all this stuff. This is based on code I
4826 originally wrote for Celo who have kindly allowed it to be
4827 included in OpenSSL.
4830 *) Add variants des_set_key_checked and des_set_key_unchecked of
4831 des_set_key (aka des_key_sched). Global variable des_check_key
4832 decides which of these is called by des_set_key; this way
4833 des_check_key behaves as it always did, but applications and
4834 the library itself, which was buggy for des_check_key == 1,
4835 have a cleaner way to pick the version they need.
4838 *) New function PKCS12_newpass() which changes the password of a
4842 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
4843 dynamic mix. In both cases the ids can be used as an index into the
4844 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
4845 functions so they accept a list of the field values and the
4846 application doesn't need to directly manipulate the X509_TRUST
4850 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
4854 *) Modify the way the V3 extension code looks up extensions. This now
4855 works in a similar way to the object code: we have some "standard"
4856 extensions in a static table which is searched with OBJ_bsearch()
4857 and the application can add dynamic ones if needed. The file
4858 crypto/x509v3/ext_dat.h now has the info: this file needs to be
4859 updated whenever a new extension is added to the core code and kept
4860 in ext_nid order. There is a simple program 'tabtest.c' which checks
4861 this. New extensions are not added too often so this file can readily
4862 be maintained manually.
4864 There are two big advantages in doing things this way. The extensions
4865 can be looked up immediately and no longer need to be "added" using
4866 X509V3_add_standard_extensions(): this function now does nothing.
4867 [Side note: I get *lots* of email saying the extension code doesn't
4868 work because people forget to call this function]
4869 Also no dynamic allocation is done unless new extensions are added:
4870 so if we don't add custom extensions there is no need to call
4871 X509V3_EXT_cleanup().
4874 *) Modify enc utility's salting as follows: make salting the default. Add a
4875 magic header, so unsalted files fail gracefully instead of just decrypting
4876 to garbage. This is because not salting is a big security hole, so people
4877 should be discouraged from doing it.
4880 *) Fixes and enhancements to the 'x509' utility. It allowed a message
4881 digest to be passed on the command line but it only used this
4882 parameter when signing a certificate. Modified so all relevant
4883 operations are affected by the digest parameter including the
4884 -fingerprint and -x509toreq options. Also -x509toreq choked if a
4885 DSA key was used because it didn't fix the digest.
4888 *) Initial certificate chain verify code. Currently tests the untrusted
4889 certificates for consistency with the verify purpose (which is set
4890 when the X509_STORE_CTX structure is set up) and checks the pathlength.
4892 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
4893 this is because it will reject chains with invalid extensions whereas
4894 every previous version of OpenSSL and SSLeay made no checks at all.
4896 Trust code: checks the root CA for the relevant trust settings. Trust
4897 settings have an initial value consistent with the verify purpose: e.g.
4898 if the verify purpose is for SSL client use it expects the CA to be
4899 trusted for SSL client use. However the default value can be changed to
4900 permit custom trust settings: one example of this would be to only trust
4901 certificates from a specific "secure" set of CAs.
4903 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
4904 which should be used for version portability: especially since the
4905 verify structure is likely to change more often now.
4907 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
4908 to set them. If not set then assume SSL clients will verify SSL servers
4911 Two new options to the verify program: -untrusted allows a set of
4912 untrusted certificates to be passed in and -purpose which sets the
4913 intended purpose of the certificate. If a purpose is set then the
4914 new chain verify code is used to check extension consistency.
4917 *) Support for the authority information access extension.
4920 *) Modify RSA and DSA PEM read routines to transparently handle
4921 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
4922 public keys in a format compatible with certificate
4923 SubjectPublicKeyInfo structures. Unfortunately there were already
4924 functions called *_PublicKey_* which used various odd formats so
4925 these are retained for compatibility: however the DSA variants were
4926 never in a public release so they have been deleted. Changed dsa/rsa
4927 utilities to handle the new format: note no releases ever handled public
4928 keys so we should be OK.
4930 The primary motivation for this change is to avoid the same fiasco
4931 that dogs private keys: there are several incompatible private key
4932 formats some of which are standard and some OpenSSL specific and
4933 require various evil hacks to allow partial transparent handling and
4934 even then it doesn't work with DER formats. Given the option anything
4935 other than PKCS#8 should be dumped: but the other formats have to
4936 stay in the name of compatibility.
4938 With public keys and the benefit of hindsight one standard format
4939 is used which works with EVP_PKEY, RSA or DSA structures: though
4940 it clearly returns an error if you try to read the wrong kind of key.
4942 Added a -pubkey option to the 'x509' utility to output the public key.
4943 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
4944 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
4945 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
4946 that do the same as the EVP_PKEY_assign_*() except they up the
4947 reference count of the added key (they don't "swallow" the
4951 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
4952 CRLs would fail if the file contained no certificates or no CRLs:
4953 added a new function to read in both types and return the number
4954 read: this means that if none are read it will be an error. The
4955 DER versions of the certificate and CRL reader would always fail
4956 because it isn't possible to mix certificates and CRLs in DER format
4957 without choking one or the other routine. Changed this to just read
4958 a certificate: this is the best we can do. Also modified the code
4959 in apps/verify.c to take notice of return codes: it was previously
4960 attempting to read in certificates from NULL pointers and ignoring
4961 any errors: this is one reason why the cert and CRL reader seemed
4962 to work. It doesn't check return codes from the default certificate
4963 routines: these may well fail if the certificates aren't installed.
4966 *) Code to support otherName option in GeneralName.
4969 *) First update to verify code. Change the verify utility
4970 so it warns if it is passed a self signed certificate:
4971 for consistency with the normal behaviour. X509_verify
4972 has been modified to it will now verify a self signed
4973 certificate if *exactly* the same certificate appears
4974 in the store: it was previously impossible to trust a
4975 single self signed certificate. This means that:
4976 openssl verify ss.pem
4977 now gives a warning about a self signed certificate but
4978 openssl verify -CAfile ss.pem ss.pem
4982 *) For servers, store verify_result in SSL_SESSION data structure
4983 (and add it to external session representation).
4984 This is needed when client certificate verifications fails,
4985 but an application-provided verification callback (set by
4986 SSL_CTX_set_cert_verify_callback) allows accepting the session
4987 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
4988 but returns 1): When the session is reused, we have to set
4989 ssl->verify_result to the appropriate error code to avoid
4991 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
4993 *) Fix a bug in the new PKCS#7 code: it didn't consider the
4994 case in PKCS7_dataInit() where the signed PKCS7 structure
4995 didn't contain any existing data because it was being created.
4996 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
4998 *) Add a salt to the key derivation routines in enc.c. This
4999 forms the first 8 bytes of the encrypted file. Also add a
5000 -S option to allow a salt to be input on the command line.
5003 *) New function X509_cmp(). Oddly enough there wasn't a function
5004 to compare two certificates. We do this by working out the SHA1
5005 hash and comparing that. X509_cmp() will be needed by the trust
5009 *) SSL_get1_session() is like SSL_get_session(), but increments
5010 the reference count in the SSL_SESSION returned.
5011 [Geoff Thorpe <geoff@eu.c2.net>]
5013 *) Fix for 'req': it was adding a null to request attributes.
5014 Also change the X509_LOOKUP and X509_INFO code to handle
5015 certificate auxiliary information.
5018 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
5022 *) Add the possibility to add extra information to the memory leak
5023 detecting output, to form tracebacks, showing from where each
5024 allocation was originated: CRYPTO_push_info("constant string") adds
5025 the string plus current file name and line number to a per-thread
5026 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
5027 is like calling CYRPTO_pop_info() until the stack is empty.
5028 Also updated memory leak detection code to be multi-thread-safe.
5031 *) Add options -text and -noout to pkcs7 utility and delete the
5032 encryption options which never did anything. Update docs.
5035 *) Add options to some of the utilities to allow the pass phrase
5036 to be included on either the command line (not recommended on
5037 OSes like Unix) or read from the environment. Update the
5038 manpages and fix a few bugs.
5041 *) Add a few manpages for some of the openssl commands.
5044 *) Fix the -revoke option in ca. It was freeing up memory twice,
5045 leaking and not finding already revoked certificates.
5048 *) Extensive changes to support certificate auxiliary information.
5049 This involves the use of X509_CERT_AUX structure and X509_AUX
5050 functions. An X509_AUX function such as PEM_read_X509_AUX()
5051 can still read in a certificate file in the usual way but it
5052 will also read in any additional "auxiliary information". By
5053 doing things this way a fair degree of compatibility can be
5054 retained: existing certificates can have this information added
5055 using the new 'x509' options.
5057 Current auxiliary information includes an "alias" and some trust
5058 settings. The trust settings will ultimately be used in enhanced
5059 certificate chain verification routines: currently a certificate
5060 can only be trusted if it is self signed and then it is trusted
5064 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
5065 The problem was that one of the replacement routines had not been working
5066 since SSLeay releases. For now the offending routine has been replaced
5067 with non-optimised assembler. Even so, this now gives around 95%
5068 performance improvement for 1024 bit RSA signs.
5071 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
5072 handling. Most clients have the effective key size in bits equal to
5073 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
5074 A few however don't do this and instead use the size of the decrypted key
5075 to determine the RC2 key length and the AlgorithmIdentifier to determine
5076 the effective key length. In this case the effective key length can still
5077 be 40 bits but the key length can be 168 bits for example. This is fixed
5078 by manually forcing an RC2 key into the EVP_PKEY structure because the
5079 EVP code can't currently handle unusual RC2 key sizes: it always assumes
5080 the key length and effective key length are equal.
5083 *) Add a bunch of functions that should simplify the creation of
5084 X509_NAME structures. Now you should be able to do:
5085 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
5086 and have it automatically work out the correct field type and fill in
5087 the structures. The more adventurous can try:
5088 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
5089 and it will (hopefully) work out the correct multibyte encoding.
5092 *) Change the 'req' utility to use the new field handling and multibyte
5093 copy routines. Before the DN field creation was handled in an ad hoc
5094 way in req, ca, and x509 which was rather broken and didn't support
5095 BMPStrings or UTF8Strings. Since some software doesn't implement
5096 BMPStrings or UTF8Strings yet, they can be enabled using the config file
5097 using the dirstring_type option. See the new comment in the default
5098 openssl.cnf for more info.
5101 *) Make crypto/rand/md_rand.c more robust:
5102 - Assure unique random numbers after fork().
5103 - Make sure that concurrent threads access the global counter and
5104 md serializably so that we never lose entropy in them
5105 or use exactly the same state in multiple threads.
5106 Access to the large state is not always serializable because
5107 the additional locking could be a performance killer, and
5108 md should be large enough anyway.
5111 *) New file apps/app_rand.c with commonly needed functionality
5112 for handling the random seed file.
5114 Use the random seed file in some applications that previously did not:
5116 dsaparam -genkey (which also ignored its '-rand' option),
5119 x509 (when signing).
5120 Except on systems with /dev/urandom, it is crucial to have a random
5121 seed file at least for key creation, DSA signing, and for DH exchanges;
5122 for RSA signatures we could do without one.
5124 gendh and gendsa (unlike genrsa) used to read only the first byte
5125 of each file listed in the '-rand' option. The function as previously
5126 found in genrsa is now in app_rand.c and is used by all programs
5127 that support '-rand'.
5130 *) In RAND_write_file, use mode 0600 for creating files;
5131 don't just chmod when it may be too late.
5134 *) Report an error from X509_STORE_load_locations
5135 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
5138 *) New function ASN1_mbstring_copy() this copies a string in either
5139 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
5140 into an ASN1_STRING type. A mask of permissible types is passed
5141 and it chooses the "minimal" type to use or an error if not type
5145 *) Add function equivalents to the various macros in asn1.h. The old
5146 macros are retained with an M_ prefix. Code inside the library can
5147 use the M_ macros. External code (including the openssl utility)
5148 should *NOT* in order to be "shared library friendly".
5151 *) Add various functions that can check a certificate's extensions
5152 to see if it usable for various purposes such as SSL client,
5153 server or S/MIME and CAs of these types. This is currently
5154 VERY EXPERIMENTAL but will ultimately be used for certificate chain
5155 verification. Also added a -purpose flag to x509 utility to
5156 print out all the purposes.
5159 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
5163 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
5164 for, obtain and decode and extension and obtain its critical flag.
5165 This allows all the necessary extension code to be handled in a
5166 single function call.
5169 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
5170 platforms. See crypto/rc4/rc4_enc.c for further details.
5173 *) New -noout option to asn1parse. This causes no output to be produced
5174 its main use is when combined with -strparse and -out to extract data
5175 from a file (which may not be in ASN.1 format).
5178 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
5179 when producing the local key id.
5180 [Richard Levitte <levitte@stacken.kth.se>]
5182 *) New option -dhparam in s_server. This allows a DH parameter file to be
5183 stated explicitly. If it is not stated then it tries the first server
5184 certificate file. The previous behaviour hard coded the filename
5188 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
5189 a public key to be input or output. For example:
5190 openssl rsa -in key.pem -pubout -out pubkey.pem
5191 Also added necessary DSA public key functions to handle this.
5194 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
5195 in the message. This was handled by allowing
5196 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
5197 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
5199 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
5200 to the end of the strings whereas this didn't. This would cause problems
5201 if strings read with d2i_ASN1_bytes() were later modified.
5202 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
5204 *) Fix for base64 decode bug. When a base64 bio reads only one line of
5205 data and it contains EOF it will end up returning an error. This is
5206 caused by input 46 bytes long. The cause is due to the way base64
5207 BIOs find the start of base64 encoded data. They do this by trying a
5208 trial decode on each line until they find one that works. When they
5209 do a flag is set and it starts again knowing it can pass all the
5210 data directly through the decoder. Unfortunately it doesn't reset
5211 the context it uses. This means that if EOF is reached an attempt
5212 is made to pass two EOFs through the context and this causes the
5213 resulting error. This can also cause other problems as well. As is
5214 usual with these problems it takes *ages* to find and the fix is
5215 trivial: move one line.
5216 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
5218 *) Ugly workaround to get s_client and s_server working under Windows. The
5219 old code wouldn't work because it needed to select() on sockets and the
5220 tty (for keypresses and to see if data could be written). Win32 only
5221 supports select() on sockets so we select() with a 1s timeout on the
5222 sockets and then see if any characters are waiting to be read, if none
5223 are present then we retry, we also assume we can always write data to
5224 the tty. This isn't nice because the code then blocks until we've
5225 received a complete line of data and it is effectively polling the
5226 keyboard at 1s intervals: however it's quite a bit better than not
5227 working at all :-) A dedicated Windows application might handle this
5228 with an event loop for example.
5231 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
5232 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
5233 will be called when RSA_sign() and RSA_verify() are used. This is useful
5234 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
5235 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
5236 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
5237 This necessitated the support of an extra signature type NID_md5_sha1
5238 for SSL signatures and modifications to the SSL library to use it instead
5239 of calling RSA_public_decrypt() and RSA_private_encrypt().
5242 *) Add new -verify -CAfile and -CApath options to the crl program, these
5243 will lookup a CRL issuers certificate and verify the signature in a
5244 similar way to the verify program. Tidy up the crl program so it
5245 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
5246 less strict. It will now permit CRL extensions even if it is not
5247 a V2 CRL: this will allow it to tolerate some broken CRLs.
5250 *) Initialize all non-automatic variables each time one of the openssl
5251 sub-programs is started (this is necessary as they may be started
5252 multiple times from the "OpenSSL>" prompt).
5253 [Lennart Bang, Bodo Moeller]
5255 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
5256 removing all other RSA functionality (this is what NO_RSA does). This
5257 is so (for example) those in the US can disable those operations covered
5258 by the RSA patent while allowing storage and parsing of RSA keys and RSA
5262 *) Non-copying interface to BIO pairs.
5263 (still largely untested)
5266 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
5267 ASCII string. This was handled independently in various places before.
5270 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
5271 UTF8 strings a character at a time.
5274 *) Use client_version from client hello to select the protocol
5275 (s23_srvr.c) and for RSA client key exchange verification
5276 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
5279 *) Add various utility functions to handle SPKACs, these were previously
5280 handled by poking round in the structure internals. Added new function
5281 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
5282 print, verify and generate SPKACs. Based on an original idea from
5283 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
5286 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
5289 *) Allow the config file extension section to be overwritten on the
5290 command line. Based on an original idea from Massimiliano Pala
5291 <madwolf@comune.modena.it>. The new option is called -extensions
5292 and can be applied to ca, req and x509. Also -reqexts to override
5293 the request extensions in req and -crlexts to override the crl extensions
5297 *) Add new feature to the SPKAC handling in ca. Now you can include
5298 the same field multiple times by preceding it by "XXXX." for example:
5301 this is the same syntax as used in the req config file.
5304 *) Allow certificate extensions to be added to certificate requests. These
5305 are specified in a 'req_extensions' option of the req section of the
5306 config file. They can be printed out with the -text option to req but
5307 are otherwise ignored at present.
5310 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
5311 data read consists of only the final block it would not decrypted because
5312 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
5313 A misplaced 'break' also meant the decrypted final block might not be
5314 copied until the next read.
5317 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
5318 a few extra parameters to the DH structure: these will be useful if
5319 for example we want the value of 'q' or implement X9.42 DH.
5322 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
5323 provides hooks that allow the default DSA functions or functions on a
5324 "per key" basis to be replaced. This allows hardware acceleration and
5325 hardware key storage to be handled without major modification to the
5326 library. Also added low level modexp hooks and CRYPTO_EX structure and
5327 associated functions.
5330 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
5331 as "read only": it can't be written to and the buffer it points to will
5332 not be freed. Reading from a read only BIO is much more efficient than
5333 a normal memory BIO. This was added because there are several times when
5334 an area of memory needs to be read from a BIO. The previous method was
5335 to create a memory BIO and write the data to it, this results in two
5336 copies of the data and an O(n^2) reading algorithm. There is a new
5337 function BIO_new_mem_buf() which creates a read only memory BIO from
5338 an area of memory. Also modified the PKCS#7 routines to use read only
5342 *) Bugfix: ssl23_get_client_hello did not work properly when called in
5343 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
5344 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
5345 but a retry condition occured while trying to read the rest.
5348 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
5349 NID_pkcs7_encrypted by default: this was wrong since this should almost
5350 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
5351 the encrypted data type: this is a more sensible place to put it and it
5352 allows the PKCS#12 code to be tidied up that duplicated this
5356 *) Changed obj_dat.pl script so it takes its input and output files on
5357 the command line. This should avoid shell escape redirection problems
5361 *) Initial support for certificate extension requests, these are included
5362 in things like Xenroll certificate requests. Included functions to allow
5363 extensions to be obtained and added.
5366 *) -crlf option to s_client and s_server for sending newlines as
5367 CRLF (as required by many protocols).
5370 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
5372 *) Install libRSAglue.a when OpenSSL is built with RSAref.
5373 [Ralf S. Engelschall]
5375 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
5376 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
5378 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
5382 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
5383 DH parameters/keys (q is lost during that conversion, but the resulting
5384 DH parameters contain its length).
5386 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
5387 much faster than DH_generate_parameters (which creates parameters
5388 where p = 2*q + 1), and also the smaller q makes DH computations
5389 much more efficient (160-bit exponentiation instead of 1024-bit
5390 exponentiation); so this provides a convenient way to support DHE
5391 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
5392 utter importance to use
5393 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5395 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5396 when such DH parameters are used, because otherwise small subgroup
5397 attacks may become possible!
5400 *) Avoid memory leak in i2d_DHparams.
5403 *) Allow the -k option to be used more than once in the enc program:
5404 this allows the same encrypted message to be read by multiple recipients.
5407 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
5408 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
5409 it will always use the numerical form of the OID, even if it has a short
5413 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
5414 method only got called if p,q,dmp1,dmq1,iqmp components were present,
5415 otherwise bn_mod_exp was called. In the case of hardware keys for example
5416 no private key components need be present and it might store extra data
5417 in the RSA structure, which cannot be accessed from bn_mod_exp.
5418 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
5419 private key operations.
5422 *) Added support for SPARC Linux.
5425 *) pem_password_cb function type incompatibly changed from
5426 typedef int pem_password_cb(char *buf, int size, int rwflag);
5428 ....(char *buf, int size, int rwflag, void *userdata);
5429 so that applications can pass data to their callbacks:
5430 The PEM[_ASN1]_{read,write}... functions and macros now take an
5431 additional void * argument, which is just handed through whenever
5432 the password callback is called.
5433 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
5435 New function SSL_CTX_set_default_passwd_cb_userdata.
5437 Compatibility note: As many C implementations push function arguments
5438 onto the stack in reverse order, the new library version is likely to
5439 interoperate with programs that have been compiled with the old
5440 pem_password_cb definition (PEM_whatever takes some data that
5441 happens to be on the stack as its last argument, and the callback
5442 just ignores this garbage); but there is no guarantee whatsoever that
5445 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
5446 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
5447 problems not only on Windows, but also on some Unix platforms.
5448 To avoid problematic command lines, these definitions are now in an
5449 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
5450 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
5453 *) MIPS III/IV assembler module is reimplemented.
5456 *) More DES library cleanups: remove references to srand/rand and
5457 delete an unused file.
5460 *) Add support for the the free Netwide assembler (NASM) under Win32,
5461 since not many people have MASM (ml) and it can be hard to obtain.
5462 This is currently experimental but it seems to work OK and pass all
5463 the tests. Check out INSTALL.W32 for info.
5466 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
5467 without temporary keys kept an extra copy of the server key,
5468 and connections with temporary keys did not free everything in case
5472 *) New function RSA_check_key and new openssl rsa option -check
5473 for verifying the consistency of RSA keys.
5474 [Ulf Moeller, Bodo Moeller]
5476 *) Various changes to make Win32 compile work:
5477 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
5478 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
5479 comparison" warnings.
5480 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
5483 *) Add a debugging option to PKCS#5 v2 key generation function: when
5484 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
5485 derived keys are printed to stderr.
5488 *) Copy the flags in ASN1_STRING_dup().
5489 [Roman E. Pavlov <pre@mo.msk.ru>]
5491 *) The x509 application mishandled signing requests containing DSA
5492 keys when the signing key was also DSA and the parameters didn't match.
5494 It was supposed to omit the parameters when they matched the signing key:
5495 the verifying software was then supposed to automatically use the CA's
5496 parameters if they were absent from the end user certificate.
5498 Omitting parameters is no longer recommended. The test was also
5499 the wrong way round! This was probably due to unusual behaviour in
5500 EVP_cmp_parameters() which returns 1 if the parameters match.
5501 This meant that parameters were omitted when they *didn't* match and
5502 the certificate was useless. Certificates signed with 'ca' didn't have
5504 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
5506 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
5507 The interface is as follows:
5508 Applications can use
5509 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
5510 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
5511 "off" is now the default.
5512 The library internally uses
5513 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
5514 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
5515 to disable memory-checking temporarily.
5517 Some inconsistent states that previously were possible (and were
5518 even the default) are now avoided.
5520 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
5521 with each memory chunk allocated; this is occasionally more helpful
5522 than just having a counter.
5524 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
5526 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
5530 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
5531 which largely parallels "options", but is for changing API behaviour,
5532 whereas "options" are about protocol behaviour.
5533 Initial "mode" flags are:
5535 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
5536 a single record has been written.
5537 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
5538 retries use the same buffer location.
5539 (But all of the contents must be
5543 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
5546 *) Fix problems with no-hmac etc.
5547 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
5549 *) New functions RSA_get_default_method(), RSA_set_method() and
5550 RSA_get_method(). These allows replacement of RSA_METHODs without having
5551 to mess around with the internals of an RSA structure.
5554 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
5555 Also really enable memory leak checks in openssl.c and in some
5557 [Chad C. Mulligan, Bodo Moeller]
5559 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
5560 up the length of negative integers. This has now been simplified to just
5561 store the length when it is first determined and use it later, rather
5562 than trying to keep track of where data is copied and updating it to
5564 [Steve Henson, reported by Brien Wheeler
5565 <bwheeler@authentica-security.com>]
5567 *) Add a new function PKCS7_signatureVerify. This allows the verification
5568 of a PKCS#7 signature but with the signing certificate passed to the
5569 function itself. This contrasts with PKCS7_dataVerify which assumes the
5570 certificate is present in the PKCS#7 structure. This isn't always the
5571 case: certificates can be omitted from a PKCS#7 structure and be
5572 distributed by "out of band" means (such as a certificate database).
5575 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
5576 function prototypes in pem.h, also change util/mkdef.pl to add the
5577 necessary function names.
5580 *) mk1mf.pl (used by Windows builds) did not properly read the
5581 options set by Configure in the top level Makefile, and Configure
5582 was not even able to write more than one option correctly.
5583 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
5586 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
5587 file to be loaded from a BIO or FILE pointer. The BIO version will
5588 for example allow memory BIOs to contain config info.
5591 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
5592 Whoever hopes to achieve shared-library compatibility across versions
5593 must use this, not the compile-time macro.
5594 (Exercise 0.9.4: Which is the minimum library version required by
5596 Note: All this applies only to multi-threaded programs, others don't
5600 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
5601 through a BIO pair triggered the default case, i.e.
5602 SSLerr(...,SSL_R_UNKNOWN_STATE).
5605 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
5606 can use the SSL library even if none of the specific BIOs is
5610 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
5611 for the encoded length.
5612 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
5614 *) Add initial documentation of the X509V3 functions.
5617 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
5618 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
5619 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
5620 secure PKCS#8 private key format with a high iteration count.
5623 *) Fix determination of Perl interpreter: A perl or perl5
5624 _directory_ in $PATH was also accepted as the interpreter.
5625 [Ralf S. Engelschall]
5627 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
5628 wrong with it but it was very old and did things like calling
5629 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
5633 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
5634 to use the new extension code.
5637 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
5638 with macros. This should make it easier to change their form, add extra
5639 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
5643 *) Add to configuration table a new entry that can specify an alternative
5644 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
5645 according to Mark Crispin <MRC@Panda.COM>.
5649 *) DES CBC did not update the IV. Weird.
5652 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
5653 Changing the behaviour of the former might break existing programs --
5654 where IV updating is needed, des_ncbc_encrypt can be used.
5657 *) When bntest is run from "make test" it drives bc to check its
5658 calculations, as well as internally checking them. If an internal check
5659 fails, it needs to cause bc to give a non-zero result or make test carries
5660 on without noticing the failure. Fixed.
5663 *) DES library cleanups.
5666 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
5667 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
5668 ciphers. NOTE: although the key derivation function has been verified
5669 against some published test vectors it has not been extensively tested
5670 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
5674 *) Instead of "mkdir -p", which is not fully portable, use new
5675 Perl script "util/mkdir-p.pl".
5678 *) Rewrite the way password based encryption (PBE) is handled. It used to
5679 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
5680 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
5681 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
5682 the 'parameter' field of the AlgorithmIdentifier is passed to the
5683 underlying key generation function so it must do its own ASN1 parsing.
5684 This has also changed the EVP_PBE_CipherInit() function which now has a
5685 'parameter' argument instead of literal salt and iteration count values
5686 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
5689 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
5690 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
5691 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
5692 KEY" because this clashed with PKCS#8 unencrypted string. Since this
5693 value was just used as a "magic string" and not used directly its
5694 value doesn't matter.
5697 *) Introduce some semblance of const correctness to BN. Shame C doesn't
5701 *) "linux-sparc64" configuration (ultrapenguin).
5702 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
5703 "linux-sparc" configuration.
5704 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
5706 *) config now generates no-xxx options for missing ciphers.
5709 *) Support the EBCDIC character set (work in progress).
5710 File ebcdic.c not yet included because it has a different license.
5711 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5713 *) Support BS2000/OSD-POSIX.
5714 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5716 *) Make callbacks for key generation use void * instead of char *.
5719 *) Make S/MIME samples compile (not yet tested).
5722 *) Additional typesafe stacks.
5725 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
5729 Changes between 0.9.3 and 0.9.3a [29 May 1999]
5731 *) New configuration variant "sco5-gcc".
5733 *) Updated some demos.
5734 [Sean O Riordain, Wade Scholine]
5736 *) Add missing BIO_free at exit of pkcs12 application.
5739 *) Fix memory leak in conf.c.
5742 *) Updates for Win32 to assembler version of MD5.
5745 *) Set #! path to perl in apps/der_chop to where we found it
5746 instead of using a fixed path.
5749 *) SHA library changes for irix64-mips4-cc.
5752 *) Improvements for VMS support.
5756 Changes between 0.9.2b and 0.9.3 [24 May 1999]
5758 *) Bignum library bug fix. IRIX 6 passes "make test" now!
5759 This also avoids the problems with SC4.2 and unpatched SC5.
5760 [Andy Polyakov <appro@fy.chalmers.se>]
5762 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
5763 These are required because of the typesafe stack would otherwise break
5764 existing code. If old code used a structure member which used to be STACK
5765 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
5766 sk_num or sk_value it would produce an error because the num, data members
5767 are not present in STACK_OF. Now it just produces a warning. sk_set
5768 replaces the old method of assigning a value to sk_value
5769 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
5770 that does this will no longer work (and should use sk_set instead) but
5771 this could be regarded as a "questionable" behaviour anyway.
5774 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
5775 correctly handle encrypted S/MIME data.
5778 *) Change type of various DES function arguments from des_cblock
5779 (which means, in function argument declarations, pointer to char)
5780 to des_cblock * (meaning pointer to array with 8 char elements),
5781 which allows the compiler to do more typechecking; it was like
5782 that back in SSLeay, but with lots of ugly casts.
5784 Introduce new type const_des_cblock.
5787 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
5788 problems: find RecipientInfo structure that matches recipient certificate
5789 and initialise the ASN1 structures properly based on passed cipher.
5792 *) Belatedly make the BN tests actually check the results.
5795 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
5796 to and from BNs: it was completely broken. New compilation option
5797 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
5798 key elements as negative integers.
5801 *) Reorganize and speed up MD5.
5802 [Andy Polyakov <appro@fy.chalmers.se>]
5805 [Richard Levitte <richard@levitte.org>]
5807 *) New option -out to asn1parse to allow the parsed structure to be
5808 output to a file. This is most useful when combined with the -strparse
5809 option to examine the output of things like OCTET STRINGS.
5812 *) Make SSL library a little more fool-proof by not requiring any longer
5813 that SSL_set_{accept,connect}_state be called before
5814 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
5815 in many applications because usually everything *appeared* to work as
5816 intended anyway -- now it really works as intended).
5819 *) Move openssl.cnf out of lib/.
5822 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
5823 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
5824 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
5825 [Ralf S. Engelschall]
5827 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
5828 handle PKCS#7 enveloped data properly.
5829 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
5831 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
5832 copying pointers. The cert_st handling is changed by this in
5833 various ways (and thus what used to be known as ctx->default_cert
5834 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
5835 any longer when s->cert does not give us what we need).
5836 ssl_cert_instantiate becomes obsolete by this change.
5837 As soon as we've got the new code right (possibly it already is?),
5838 we have solved a couple of bugs of the earlier code where s->cert
5839 was used as if it could not have been shared with other SSL structures.
5841 Note that using the SSL API in certain dirty ways now will result
5842 in different behaviour than observed with earlier library versions:
5843 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
5844 does not influence s as it used to.
5846 In order to clean up things more thoroughly, inside SSL_SESSION
5847 we don't use CERT any longer, but a new structure SESS_CERT
5848 that holds per-session data (if available); currently, this is
5849 the peer's certificate chain and, for clients, the server's certificate
5850 and temporary key. CERT holds only those values that can have
5851 meaningful defaults in an SSL_CTX.
5854 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
5855 from the internal representation. Various PKCS#7 fixes: remove some
5856 evil casts and set the enc_dig_alg field properly based on the signing
5860 *) Allow PKCS#12 password to be set from the command line or the
5861 environment. Let 'ca' get its config file name from the environment
5862 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
5866 *) Allow certificate policies extension to use an IA5STRING for the
5867 organization field. This is contrary to the PKIX definition but
5868 VeriSign uses it and IE5 only recognises this form. Document 'x509'
5872 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
5873 without disallowing inline assembler and the like for non-pedantic builds.
5876 *) Support Borland C++ builder.
5877 [Janez Jere <jj@void.si>, modified by Ulf Möller]
5882 *) SHA-1 cleanups and performance enhancements.
5883 [Andy Polyakov <appro@fy.chalmers.se>]
5885 *) Sparc v8plus assembler for the bignum library.
5886 [Andy Polyakov <appro@fy.chalmers.se>]
5888 *) Accept any -xxx and +xxx compiler options in Configure.
5891 *) Update HPUX configuration.
5894 *) Add missing sk_<type>_unshift() function to safestack.h
5895 [Ralf S. Engelschall]
5897 *) New function SSL_CTX_use_certificate_chain_file that sets the
5898 "extra_cert"s in addition to the certificate. (This makes sense
5899 only for "PEM" format files, as chains as a whole are not
5903 *) Support verify_depth from the SSL API.
5904 x509_vfy.c had what can be considered an off-by-one-error:
5905 Its depth (which was not part of the external interface)
5906 was actually counting the number of certificates in a chain;
5907 now it really counts the depth.
5910 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
5911 instead of X509err, which often resulted in confusing error
5912 messages since the error codes are not globally unique
5913 (e.g. an alleged error in ssl3_accept when a certificate
5914 didn't match the private key).
5916 *) New function SSL_CTX_set_session_id_context that allows to set a default
5917 value (so that you don't need SSL_set_session_id_context for each
5918 connection using the SSL_CTX).
5921 *) OAEP decoding bug fix.
5924 *) Support INSTALL_PREFIX for package builders, as proposed by
5928 *) New Configure options "threads" and "no-threads". For systems
5929 where the proper compiler options are known (currently Solaris
5930 and Linux), "threads" is the default.
5933 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
5936 *) Install various scripts to $(OPENSSLDIR)/misc, not to
5937 $(INSTALLTOP)/bin -- they shouldn't clutter directories
5938 such as /usr/local/bin.
5941 *) "make linux-shared" to build shared libraries.
5942 [Niels Poppe <niels@netbox.org>]
5944 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
5947 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
5948 extension adding in x509 utility.
5951 *) Remove NOPROTO sections and error code comments.
5954 *) Partial rewrite of the DEF file generator to now parse the ANSI
5958 *) New Configure options --prefix=DIR and --openssldir=DIR.
5961 *) Complete rewrite of the error code script(s). It is all now handled
5962 by one script at the top level which handles error code gathering,
5963 header rewriting and C source file generation. It should be much better
5964 than the old method: it now uses a modified version of Ulf's parser to
5965 read the ANSI prototypes in all header files (thus the old K&R definitions
5966 aren't needed for error creation any more) and do a better job of
5967 translating function codes into names. The old 'ASN1 error code imbedded
5968 in a comment' is no longer necessary and it doesn't use .err files which
5969 have now been deleted. Also the error code call doesn't have to appear all
5970 on one line (which resulted in some large lines...).
5973 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
5976 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
5977 0 (which usually indicates a closed connection), but continue reading.
5980 *) Fix some race conditions.
5983 *) Add support for CRL distribution points extension. Add Certificate
5984 Policies and CRL distribution points documentation.
5987 *) Move the autogenerated header file parts to crypto/opensslconf.h.
5990 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
5991 8 of keying material. Merlin has also confirmed interop with this fix
5992 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
5993 [Merlin Hughes <merlin@baltimore.ie>]
5995 *) Fix lots of warnings.
5996 [Richard Levitte <levitte@stacken.kth.se>]
5998 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
5999 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
6000 [Richard Levitte <levitte@stacken.kth.se>]
6002 *) Fix problems with sizeof(long) == 8.
6003 [Andy Polyakov <appro@fy.chalmers.se>]
6005 *) Change functions to ANSI C.
6008 *) Fix typos in error codes.
6009 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
6011 *) Remove defunct assembler files from Configure.
6014 *) SPARC v8 assembler BIGNUM implementation.
6015 [Andy Polyakov <appro@fy.chalmers.se>]
6017 *) Support for Certificate Policies extension: both print and set.
6018 Various additions to support the r2i method this uses.
6021 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
6022 return a const string when you are expecting an allocated buffer.
6025 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
6026 types DirectoryString and DisplayText.
6029 *) Add code to allow r2i extensions to access the configuration database,
6030 add an LHASH database driver and add several ctx helper functions.
6033 *) Fix an evil bug in bn_expand2() which caused various BN functions to
6034 fail when they extended the size of a BIGNUM.
6037 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
6038 support typesafe stack.
6041 *) Fix typo in SSL_[gs]et_options().
6042 [Nils Frostberg <nils@medcom.se>]
6044 *) Delete various functions and files that belonged to the (now obsolete)
6045 old X509V3 handling code.
6048 *) New Configure option "rsaref".
6051 *) Don't auto-generate pem.h.
6054 *) Introduce type-safe ASN.1 SETs.
6057 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
6058 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
6060 *) Introduce type-safe STACKs. This will almost certainly break lots of code
6061 that links with OpenSSL (well at least cause lots of warnings), but fear
6062 not: the conversion is trivial, and it eliminates loads of evil casts. A
6063 few STACKed things have been converted already. Feel free to convert more.
6064 In the fullness of time, I'll do away with the STACK type altogether.
6067 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
6068 specified in <certfile> by updating the entry in the index.txt file.
6069 This way one no longer has to edit the index.txt file manually for
6070 revoking a certificate. The -revoke option does the gory details now.
6071 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
6073 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
6074 `-text' option at all and this way the `-noout -text' combination was
6075 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
6076 [Ralf S. Engelschall]
6078 *) Make sure a corresponding plain text error message exists for the
6079 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
6080 verify callback function determined that a certificate was revoked.
6081 [Ralf S. Engelschall]
6083 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
6084 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
6085 all available cipers including rc5, which was forgotten until now.
6086 In order to let the testing shell script know which algorithms
6087 are available, a new (up to now undocumented) command
6088 "openssl list-cipher-commands" is used.
6091 *) Bugfix: s_client occasionally would sleep in select() when
6092 it should have checked SSL_pending() first.
6095 *) New functions DSA_do_sign and DSA_do_verify to provide access to
6096 the raw DSA values prior to ASN.1 encoding.
6099 *) Tweaks to Configure
6100 [Niels Poppe <niels@netbox.org>]
6102 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
6106 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
6109 *) New config option to avoid instructions that are illegal on the 80386.
6110 The default code is faster, but requires at least a 486.
6113 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
6114 SSL2_SERVER_VERSION (not used at all) macros, which are now the
6115 same as SSL2_VERSION anyway.
6118 *) New "-showcerts" option for s_client.
6121 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
6122 application. Various cleanups and fixes.
6125 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
6126 modify error routines to work internally. Add error codes and PBE init
6127 to library startup routines.
6130 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
6131 packing functions to asn1 and evp. Changed function names and error
6132 codes along the way.
6135 *) PKCS12 integration: and so it begins... First of several patches to
6136 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
6137 objects to objects.h
6140 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
6141 and display support for Thawte strong extranet extension.
6144 *) Add LinuxPPC support.
6145 [Jeff Dubrule <igor@pobox.org>]
6147 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
6148 bn_div_words in alpha.s.
6149 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
6151 *) Make sure the RSA OAEP test is skipped under -DRSAref because
6152 OAEP isn't supported when OpenSSL is built with RSAref.
6153 [Ulf Moeller <ulf@fitug.de>]
6155 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
6156 so they no longer are missing under -DNOPROTO.
6157 [Soren S. Jorvang <soren@t.dk>]
6160 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
6162 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
6163 doesn't work when the session is reused. Coming soon!
6166 *) Fix a security hole, that allows sessions to be reused in the wrong
6167 context thus bypassing client cert protection! All software that uses
6168 client certs and session caches in multiple contexts NEEDS PATCHING to
6169 allow session reuse! A fuller solution is in the works.
6170 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
6172 *) Some more source tree cleanups (removed obsolete files
6173 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
6174 permission on "config" script to be executable) and a fix for the INSTALL
6176 [Ulf Moeller <ulf@fitug.de>]
6178 *) Remove some legacy and erroneous uses of malloc, free instead of
6180 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
6182 *) Make rsa_oaep_test return non-zero on error.
6183 [Ulf Moeller <ulf@fitug.de>]
6185 *) Add support for native Solaris shared libraries. Configure
6186 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
6187 if someone would make that last step automatic.
6188 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
6190 *) ctx_size was not built with the right compiler during "make links". Fixed.
6193 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
6194 except NULL ciphers". This means the default cipher list will no longer
6195 enable NULL ciphers. They need to be specifically enabled e.g. with
6196 the string "DEFAULT:eNULL".
6199 *) Fix to RSA private encryption routines: if p < q then it would
6200 occasionally produce an invalid result. This will only happen with
6201 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
6204 *) Be less restrictive and allow also `perl util/perlpath.pl
6205 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
6206 because this way one can also use an interpreter named `perl5' (which is
6207 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
6208 installed as `perl').
6209 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6211 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
6212 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6214 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
6215 advapi32.lib to Win32 build and change the pem test comparision
6216 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
6217 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
6218 and crypto/des/ede_cbcm_enc.c.
6221 *) DES quad checksum was broken on big-endian architectures. Fixed.
6224 *) Comment out two functions in bio.h that aren't implemented. Fix up the
6225 Win32 test batch file so it (might) work again. The Win32 test batch file
6226 is horrible: I feel ill....
6229 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
6230 in e_os.h. Audit of header files to check ANSI and non ANSI
6231 sections: 10 functions were absent from non ANSI section and not exported
6232 from Windows DLLs. Fixed up libeay.num for new functions.
6235 *) Make `openssl version' output lines consistent.
6236 [Ralf S. Engelschall]
6238 *) Fix Win32 symbol export lists for BIO functions: Added
6239 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
6240 to ms/libeay{16,32}.def.
6241 [Ralf S. Engelschall]
6243 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
6244 fine under Unix and passes some trivial tests I've now added. But the
6245 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
6246 added to make sure no one expects that this stuff really works in the
6247 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
6248 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
6250 [Ralf S. Engelschall]
6252 *) Fix the generation of two part addresses in perl.
6253 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
6255 *) Add config entry for Linux on MIPS.
6256 [John Tobey <jtobey@channel1.com>]
6258 *) Make links whenever Configure is run, unless we are on Windoze.
6261 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
6262 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
6266 *) Add a useful kludge to allow package maintainers to specify compiler and
6267 other platforms details on the command line without having to patch the
6268 Configure script everytime: One now can use ``perl Configure
6269 <id>:<details>'', i.e. platform ids are allowed to have details appended
6270 to them (seperated by colons). This is treated as there would be a static
6271 pre-configured entry in Configure's %table under key <id> with value
6272 <details> and ``perl Configure <id>'' is called. So, when you want to
6273 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
6274 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
6275 now, which overrides the FreeBSD-elf entry on-the-fly.
6276 [Ralf S. Engelschall]
6278 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
6281 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
6282 on the `perl Configure ...' command line. This way one can compile
6283 OpenSSL libraries with Position Independent Code (PIC) which is needed
6284 for linking it into DSOs.
6285 [Ralf S. Engelschall]
6287 *) Remarkably, export ciphers were totally broken and no-one had noticed!
6291 *) Cleaned up the LICENSE document: The official contact for any license
6292 questions now is the OpenSSL core team under openssl-core@openssl.org.
6293 And add a paragraph about the dual-license situation to make sure people
6294 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
6295 to the OpenSSL toolkit.
6296 [Ralf S. Engelschall]
6298 *) General source tree makefile cleanups: Made `making xxx in yyy...'
6299 display consistent in the source tree and replaced `/bin/rm' by `rm'.
6300 Additonally cleaned up the `make links' target: Remove unnecessary
6301 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
6302 to speed processing and no longer clutter the display with confusing
6303 stuff. Instead only the actually done links are displayed.
6304 [Ralf S. Engelschall]
6306 *) Permit null encryption ciphersuites, used for authentication only. It used
6307 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
6308 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
6312 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
6313 signed attributes when verifying signatures (this would break them),
6314 the detached data encoding was wrong and public keys obtained using
6315 X509_get_pubkey() weren't freed.
6318 *) Add text documentation for the BUFFER functions. Also added a work around
6319 to a Win95 console bug. This was triggered by the password read stuff: the
6320 last character typed gets carried over to the next fread(). If you were
6321 generating a new cert request using 'req' for example then the last
6322 character of the passphrase would be CR which would then enter the first
6326 *) Added the new `Includes OpenSSL Cryptography Software' button as
6327 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
6328 button and can be used by applications based on OpenSSL to show the
6329 relationship to the OpenSSL project.
6330 [Ralf S. Engelschall]
6332 *) Remove confusing variables in function signatures in files
6333 ssl/ssl_lib.c and ssl/ssl.h.
6334 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6336 *) Don't install bss_file.c under PREFIX/include/
6337 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6339 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
6340 functions that return function pointers and has support for NT specific
6341 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
6342 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
6343 unsigned to signed types: this was killing the Win32 compile.
6346 *) Add new certificate file to stack functions,
6347 SSL_add_dir_cert_subjects_to_stack() and
6348 SSL_add_file_cert_subjects_to_stack(). These largely supplant
6349 SSL_load_client_CA_file(), and can be used to add multiple certs easily
6350 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
6351 This means that Apache-SSL and similar packages don't have to mess around
6352 to add as many CAs as they want to the preferred list.
6355 *) Experiment with doxygen documentation. Currently only partially applied to
6357 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
6358 openssl.doxy as the configuration file.
6361 *) Get rid of remaining C++-style comments which strict C compilers hate.
6362 [Ralf S. Engelschall, pointed out by Carlos Amengual]
6364 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
6365 compiled in by default: it has problems with large keys.
6368 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
6369 DH private keys and/or callback functions which directly correspond to
6370 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
6371 is needed for applications which have to configure certificates on a
6372 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
6374 For the RSA certificate situation is makes no difference, but
6375 for the DSA certificate situation this fixes the "no shared cipher"
6376 problem where the OpenSSL cipher selection procedure failed because the
6377 temporary keys were not overtaken from the context and the API provided
6378 no way to reconfigure them.
6379 The new functions now let applications reconfigure the stuff and they
6380 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
6381 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
6382 non-public-API function ssl_cert_instantiate() is used as a helper
6383 function and also to reduce code redundancy inside ssl_rsa.c.
6384 [Ralf S. Engelschall]
6386 *) Move s_server -dcert and -dkey options out of the undocumented feature
6387 area because they are useful for the DSA situation and should be
6388 recognized by the users.
6389 [Ralf S. Engelschall]
6391 *) Fix the cipher decision scheme for export ciphers: the export bits are
6392 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
6393 SSL_EXP_MASK. So, the original variable has to be used instead of the
6394 already masked variable.
6395 [Richard Levitte <levitte@stacken.kth.se>]
6397 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
6398 [Richard Levitte <levitte@stacken.kth.se>]
6400 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
6401 from `int' to `unsigned int' because it's a length and initialized by
6402 EVP_DigestFinal() which expects an `unsigned int *'.
6403 [Richard Levitte <levitte@stacken.kth.se>]
6405 *) Don't hard-code path to Perl interpreter on shebang line of Configure
6406 script. Instead use the usual Shell->Perl transition trick.
6407 [Ralf S. Engelschall]
6409 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
6410 (in addition to RSA certificates) to match the behaviour of `openssl dsa
6411 -noout -modulus' as it's already the case for `openssl rsa -noout
6412 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
6413 currently the public key is printed (a decision which was already done by
6414 `openssl dsa -modulus' in the past) which serves a similar purpose.
6415 Additionally the NO_RSA no longer completely removes the whole -modulus
6416 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
6418 [Ralf S. Engelschall]
6420 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
6421 BIO. See the source (crypto/evp/bio_ok.c) for more info.
6422 [Arne Ansper <arne@ats.cyber.ee>]
6424 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
6425 to be added. Now both 'req' and 'ca' can use new objects defined in the
6429 *) Add cool BIO that does syslog (or event log on NT).
6430 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
6432 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
6433 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
6434 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
6435 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
6438 *) Add preliminary config info for new extension code.
6441 *) Make RSA_NO_PADDING really use no padding.
6442 [Ulf Moeller <ulf@fitug.de>]
6444 *) Generate errors when private/public key check is done.
6447 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
6448 for some CRL extensions and new objects added.
6451 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
6452 key usage extension and fuller support for authority key id.
6455 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
6456 padding method for RSA, which is recommended for new applications in PKCS
6457 #1 v2.0 (RFC 2437, October 1998).
6458 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
6459 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
6460 against Bleichbacher's attack on RSA.
6461 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
6464 *) Updates to the new SSL compression code
6465 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6467 *) Fix so that the version number in the master secret, when passed
6468 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
6469 (because the server will not accept higher), that the version number
6470 is 0x03,0x01, not 0x03,0x00
6471 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6473 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
6474 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
6475 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
6478 *) Support for RAW extensions where an arbitrary extension can be
6479 created by including its DER encoding. See apps/openssl.cnf for
6483 *) Make sure latest Perl versions don't interpret some generated C array
6484 code as Perl array code in the crypto/err/err_genc.pl script.
6485 [Lars Weber <3weber@informatik.uni-hamburg.de>]
6487 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
6488 not many people have the assembler. Various Win32 compilation fixes and
6489 update to the INSTALL.W32 file with (hopefully) more accurate Win32
6493 *) Modify configure script 'Configure' to automatically create crypto/date.h
6494 file under Win32 and also build pem.h from pem.org. New script
6495 util/mkfiles.pl to create the MINFO file on environments that can't do a
6496 'make files': perl util/mkfiles.pl >MINFO should work.
6499 *) Major rework of DES function declarations, in the pursuit of correctness
6500 and purity. As a result, many evil casts evaporated, and some weirdness,
6501 too. You may find this causes warnings in your code. Zapping your evil
6502 casts will probably fix them. Mostly.
6505 *) Fix for a typo in asn1.h. Bug fix to object creation script
6506 obj_dat.pl. It considered a zero in an object definition to mean
6507 "end of object": none of the objects in objects.h have any zeros
6508 so it wasn't spotted.
6509 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
6511 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
6512 Masking (CBCM). In the absence of test vectors, the best I have been able
6513 to do is check that the decrypt undoes the encrypt, so far. Send me test
6514 vectors if you have them.
6517 *) Correct calculation of key length for export ciphers (too much space was
6518 allocated for null ciphers). This has not been tested!
6521 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
6522 message is now correct (it understands "crypto" and "ssl" on its
6523 command line). There is also now an "update" option. This will update
6524 the util/ssleay.num and util/libeay.num files with any new functions.
6526 perl util/mkdef.pl crypto ssl update
6527 it will update them.
6530 *) Overhauled the Perl interface (perl/*):
6531 - ported BN stuff to OpenSSL's different BN library
6532 - made the perl/ source tree CVS-aware
6533 - renamed the package from SSLeay to OpenSSL (the files still contain
6534 their history because I've copied them in the repository)
6535 - removed obsolete files (the test scripts will be replaced
6536 by better Test::Harness variants in the future)
6537 [Ralf S. Engelschall]
6539 *) First cut for a very conservative source tree cleanup:
6540 1. merge various obsolete readme texts into doc/ssleay.txt
6541 where we collect the old documents and readme texts.
6542 2. remove the first part of files where I'm already sure that we no
6543 longer need them because of three reasons: either they are just temporary
6544 files which were left by Eric or they are preserved original files where
6545 I've verified that the diff is also available in the CVS via "cvs diff
6546 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
6547 the crypto/md/ stuff).
6548 [Ralf S. Engelschall]
6550 *) More extension code. Incomplete support for subject and issuer alt
6551 name, issuer and authority key id. Change the i2v function parameters
6552 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
6553 what that's for :-) Fix to ASN1 macro which messed up
6554 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
6557 *) Preliminary support for ENUMERATED type. This is largely copied from the
6561 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
6562 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6564 *) Make sure `make rehash' target really finds the `openssl' program.
6565 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6567 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
6568 like to hear about it if this slows down other processors.
6571 *) Add CygWin32 platform information to Configure script.
6572 [Alan Batie <batie@aahz.jf.intel.com>]
6574 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
6575 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
6577 *) New program nseq to manipulate netscape certificate sequences
6580 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
6584 *) Fixes to BN code. Previously the default was to define BN_RECURSION
6585 but the BN code had some problems that would cause failures when
6586 doing certificate verification and some other functions.
6587 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6589 *) Add ASN1 and PEM code to support netscape certificate sequences.
6592 *) Add ASN1 and PEM code to support netscape certificate sequences.
6595 *) Add several PKIX and private extended key usage OIDs.
6598 *) Modify the 'ca' program to handle the new extension code. Modify
6599 openssl.cnf for new extension format, add comments.
6602 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
6603 and add a sample to openssl.cnf so req -x509 now adds appropriate
6607 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
6608 error code, add initial support to X509_print() and x509 application.
6611 *) Takes a deep breath and start addding X509 V3 extension support code. Add
6612 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
6613 stuff is currently isolated and isn't even compiled yet.
6616 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
6617 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
6618 Removed the versions check from X509 routines when loading extensions:
6619 this allows certain broken certificates that don't set the version
6620 properly to be processed.
6623 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
6624 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
6625 can still be regenerated with "make depend".
6628 *) Spelling mistake in C version of CAST-128.
6629 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
6631 *) Changes to the error generation code. The perl script err-code.pl
6632 now reads in the old error codes and retains the old numbers, only
6633 adding new ones if necessary. It also only changes the .err files if new
6634 codes are added. The makefiles have been modified to only insert errors
6635 when needed (to avoid needlessly modifying header files). This is done
6636 by only inserting errors if the .err file is newer than the auto generated
6637 C file. To rebuild all the error codes from scratch (the old behaviour)
6638 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
6639 or delete all the .err files.
6642 *) CAST-128 was incorrectly implemented for short keys. The C version has
6643 been fixed, but is untested. The assembler versions are also fixed, but
6644 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
6645 to regenerate it if needed.
6646 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
6647 Hagino <itojun@kame.net>]
6649 *) File was opened incorrectly in randfile.c.
6650 [Ulf Möller <ulf@fitug.de>]
6652 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
6653 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
6654 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
6655 al: it's just almost always a UTCTime. Note this patch adds new error
6656 codes so do a "make errors" if there are problems.
6659 *) Correct Linux 1 recognition in config.
6660 [Ulf Möller <ulf@fitug.de>]
6662 *) Remove pointless MD5 hash when using DSA keys in ca.
6663 [Anonymous <nobody@replay.com>]
6665 *) Generate an error if given an empty string as a cert directory. Also
6666 generate an error if handed NULL (previously returned 0 to indicate an
6667 error, but didn't set one).
6668 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
6670 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
6673 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
6674 parameters. This was causing a warning which killed off the Win32 compile.
6677 *) Remove C++ style comments from crypto/bn/bn_local.h.
6678 [Neil Costigan <neil.costigan@celocom.com>]
6680 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
6681 based on a text string, looking up short and long names and finally
6682 "dot" format. The "dot" format stuff didn't work. Added new function
6683 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
6684 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
6685 OID is not part of the table.
6688 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
6689 X509_LOOKUP_by_alias().
6692 *) Sort openssl functions by name.
6695 *) Get the gendsa program working (hopefully) and add it to app list. Remove
6696 encryption from sample DSA keys (in case anyone is interested the password
6700 *) Make _all_ *_free functions accept a NULL pointer.
6701 [Frans Heymans <fheymans@isaserver.be>]
6703 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
6705 [Anonymous <nobody@replay.com>]
6707 *) s_server should send the CAfile as acceptable CAs, not its own cert.
6708 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6710 *) Don't blow it for numeric -newkey arguments to apps/req.
6711 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6713 *) Temp key "for export" tests were wrong in s3_srvr.c.
6714 [Anonymous <nobody@replay.com>]
6716 *) Add prototype for temp key callback functions
6717 SSL_CTX_set_tmp_{rsa,dh}_callback().
6720 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
6721 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
6724 *) X509_name_add_entry() freed the wrong thing after an error.
6725 [Arne Ansper <arne@ats.cyber.ee>]
6727 *) rsa_eay.c would attempt to free a NULL context.
6728 [Arne Ansper <arne@ats.cyber.ee>]
6730 *) BIO_s_socket() had a broken should_retry() on Windoze.
6731 [Arne Ansper <arne@ats.cyber.ee>]
6733 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
6734 [Arne Ansper <arne@ats.cyber.ee>]
6736 *) Make sure the already existing X509_STORE->depth variable is initialized
6737 in X509_STORE_new(), but document the fact that this variable is still
6738 unused in the certificate verification process.
6739 [Ralf S. Engelschall]
6741 *) Fix the various library and apps files to free up pkeys obtained from
6742 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
6745 *) Fix reference counting in X509_PUBKEY_get(). This makes
6746 demos/maurice/example2.c work, amongst others, probably.
6747 [Steve Henson and Ben Laurie]
6749 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
6750 `openssl' and second, the shortcut symlinks for the `openssl <command>'
6751 are no longer created. This way we have a single and consistent command
6752 line interface `openssl <command>', similar to `cvs <command>'.
6753 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
6755 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
6756 BIT STRING wrapper always have zero unused bits.
6759 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
6762 *) Make the top-level INSTALL documentation easier to understand.
6765 *) Makefiles updated to exit if an error occurs in a sub-directory
6766 make (including if user presses ^C) [Paul Sutton]
6768 *) Make Montgomery context stuff explicit in RSA data structure.
6771 *) Fix build order of pem and err to allow for generated pem.h.
6774 *) Fix renumbering bug in X509_NAME_delete_entry().
6777 *) Enhanced the err-ins.pl script so it makes the error library number
6778 global and can add a library name. This is needed for external ASN1 and
6779 other error libraries.
6782 *) Fixed sk_insert which never worked properly.
6785 *) Fix ASN1 macros so they can handle indefinite length construted
6786 EXPLICIT tags. Some non standard certificates use these: they can now
6790 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
6791 into a single doc/ssleay.txt bundle. This way the information is still
6792 preserved but no longer messes up this directory. Now it's new room for
6793 the new set of documenation files.
6794 [Ralf S. Engelschall]
6796 *) SETs were incorrectly DER encoded. This was a major pain, because they
6797 shared code with SEQUENCEs, which aren't coded the same. This means that
6798 almost everything to do with SETs or SEQUENCEs has either changed name or
6799 number of arguments.
6800 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
6802 *) Fix test data to work with the above.
6805 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
6806 was already fixed by Eric for 0.9.1 it seems.
6807 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
6809 *) Autodetect FreeBSD3.
6812 *) Fix various bugs in Configure. This affects the following platforms:
6816 unixware-2.0-pentium
6820 *) Eliminate generated files from CVS. Reorder tests to regenerate files
6821 before they are needed.
6824 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
6828 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
6830 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
6831 changed SSLeay to OpenSSL in version strings.
6832 [Ralf S. Engelschall]
6834 *) Some fixups to the top-level documents.
6837 *) Fixed the nasty bug where rsaref.h was not found under compile-time
6838 because the symlink to include/ was missing.
6839 [Ralf S. Engelschall]
6841 *) Incorporated the popular no-RSA/DSA-only patches
6842 which allow to compile a RSA-free SSLeay.
6843 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
6845 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
6846 when "ssleay" is still not found.
6847 [Ralf S. Engelschall]
6849 *) Added more platforms to Configure: Cray T3E, HPUX 11,
6850 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
6852 *) Updated the README file.
6853 [Ralf S. Engelschall]
6855 *) Added various .cvsignore files in the CVS repository subdirs
6856 to make a "cvs update" really silent.
6857 [Ralf S. Engelschall]
6859 *) Recompiled the error-definition header files and added
6860 missing symbols to the Win32 linker tables.
6861 [Ralf S. Engelschall]
6863 *) Cleaned up the top-level documents;
6864 o new files: CHANGES and LICENSE
6865 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
6866 o merged COPYRIGHT into LICENSE
6867 o removed obsolete TODO file
6868 o renamed MICROSOFT to INSTALL.W32
6869 [Ralf S. Engelschall]
6871 *) Removed dummy files from the 0.9.1b source tree:
6872 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
6873 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
6874 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
6875 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
6876 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
6877 [Ralf S. Engelschall]
6879 *) Added various platform portability fixes.
6882 *) The Genesis of the OpenSSL rpject:
6883 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
6884 Young and Tim J. Hudson created while they were working for C2Net until
6886 [The OpenSSL Project]
6889 Changes between 0.9.0b and 0.9.1b [not released]
6891 *) Updated a few CA certificates under certs/
6894 *) Changed some BIGNUM api stuff.
6897 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
6898 DGUX x86, Linux Alpha, etc.
6901 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
6902 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
6906 *) Add -strparse option to asn1pars program which parses nested
6908 [Dr Stephen Henson <shenson@bigfoot.com>]
6910 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
6913 *) DSA fix for "ca" program.
6916 *) Added "-genkey" option to "dsaparam" program.
6919 *) Added RIPE MD160 (rmd160) message digest.
6922 *) Added -a (all) option to "ssleay version" command.
6925 *) Added PLATFORM define which is the id given to Configure.
6928 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
6931 *) Extended the ASN.1 parser routines.
6934 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
6937 *) Added a BN_CTX to the BN library.
6940 *) Fixed the weak key values in DES library
6943 *) Changed API in EVP library for cipher aliases.
6946 *) Added support for RC2/64bit cipher.
6949 *) Converted the lhash library to the crypto/mem.c functions.
6952 *) Added more recognized ASN.1 object ids.
6955 *) Added more RSA padding checks for SSL/TLS.
6958 *) Added BIO proxy/filter functionality.
6961 *) Added extra_certs to SSL_CTX which can be used
6962 send extra CA certificates to the client in the CA cert chain sending
6963 process. It can be configured with SSL_CTX_add_extra_chain_cert().
6966 *) Now Fortezza is denied in the authentication phase because
6967 this is key exchange mechanism is not supported by SSLeay at all.
6970 *) Additional PKCS1 checks.
6973 *) Support the string "TLSv1" for all TLS v1 ciphers.
6976 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
6977 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
6980 *) Fixed a few memory leaks.
6983 *) Fixed various code and comment typos.
6986 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
6987 bytes sent in the client random.
6988 [Edward Bishop <ebishop@spyglass.com>]