5 Changes between 0.9.7c and 0.9.8 [xx XXX xxxx]
7 *) An audit of the BIGNUM code is underway, for which debugging code is
8 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
9 is considered valid when processing BIGNUMs, and causes execution to
10 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11 further steps are taken to deliberately pollute unused data in BIGNUM
12 structures to try and expose faulty code further on. For now, openssl will
13 (in its default mode of operation) continue to tolerate the inconsistent
14 forms that it has tolerated in the past, but authors and packagers should
15 consider trying openssl and their own applications when compiled with
16 these debugging symbols defined. It will help highlight potential bugs in
17 their own code, and will improve the test coverage for OpenSSL itself. At
18 some point, these tighter rules will become openssl's default to improve
19 maintainability, though the assert()s and other overheads will remain only
20 in debugging configurations. See bn.h for more details.
23 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
24 that can only be obtained through BN_CTX_new() (which implicitly
25 initialises it). The presence of this function only made it possible
26 to overwrite an existing structure (and cause memory leaks).
29 *) Because of the callback-based approach for implementing LHASH as a
30 template type, lh_insert() adds opaque objects to hash-tables and
31 lh_doall() or lh_doall_arg() are typically used with a destructor callback
32 to clean up those corresponding objects before destroying the hash table
33 (and losing the object pointers). So some over-zealous constifications in
34 LHASH have been relaxed so that lh_insert() does not take (nor store) the
35 objects as "const" and the lh_doall[_arg] callback wrappers are not
36 prototyped to have "const" restrictions on the object pointers they are
37 given (and so aren't required to cast them away any more).
40 *) The tmdiff.h API was so ugly and minimal that our own timing utility
41 (speed) prefers to use its own implementation. The two implementations
42 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
43 its object type properly exposed (MS_TM) instead of casting to/from "char
44 *". This may still change yet if someone realises MS_TM and "ms_time_***"
45 aren't necessarily the greatest nomenclatures - but this is what was used
46 internally to the implementation so I've used that for now.
49 *) Ensure that deprecated functions do not get compiled when
50 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
51 the self-tests were still using deprecated key-generation functions so
52 these have been updated also.
55 *) Reorganise PKCS#7 code to separate the digest location functionality
56 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
57 New function PKCS7_set_digest() to set the digest type for PKCS#7
58 digestedData type. Add additional code to correctly generate the
59 digestedData type and add support for this type in PKCS7 initialization
63 *) New function PKCS7_set0_type_other() this initializes a PKCS7
64 structure of type "other".
67 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
68 sure the loop does correctly stop and breaking ("division by zero")
69 modulus operations are not performed. The (pre-generated) prime
70 table crypto/bn/bn_prime.h was already correct, but it could not be
71 re-generated on some platforms because of the "division by zero"
72 situation in the script.
75 *) Update support for ECC-based TLS ciphersuites according to
76 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
77 SHA-1 now is only used for "small" curves (where the
78 representation of a field element takes up to 24 bytes); for
79 larger curves, the field element resulting from ECDH is directly
80 used as premaster secret.
81 [Douglas Stebila (Sun Microsystems Laboratories)]
83 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
84 curve secp160r1 to the tests.
85 [Douglas Stebila (Sun Microsystems Laboratories)]
87 *) Add the possibility to load symbols globally with DSO.
88 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
90 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
91 control of the error stack.
94 *) Add support for STORE in ENGINE.
97 *) Add the STORE type. The intention is to provide a common interface
98 to certificate and key stores, be they simple file-based stores, or
99 HSM-type store, or LDAP stores, or...
100 NOTE: The code is currently UNTESTED and isn't really used anywhere.
103 *) Add a generic structure called OPENSSL_ITEM. This can be used to
104 pass a list of arguments to any function as well as provide a way
105 for a function to pass data back to the caller.
108 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
109 works like BUF_strdup() but can be used to duplicate a portion of
110 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
114 *) Add the function sk_find_ex() which works like sk_find(), but will
115 return an index to an element even if an exact match couldn't be
116 found. The index is guaranteed to point at the element where the
117 searched-for key would be inserted to preserve sorting order.
120 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
121 takes an extra flags argument for optional functionality. Currently,
122 the following flags are defined:
124 OBJ_BSEARCH_VALUE_ON_NOMATCH
125 This one gets OBJ_bsearch_ex() to return a pointer to the first
126 element where the comparing function returns a negative or zero
129 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
130 This one gets OBJ_bsearch_ex() to return a pointer to the first
131 element where the comparing function returns zero. This is useful
132 if there are more than one element where the comparing function
136 *) Make it possible to create self-signed certificates with 'openssl ca'
137 in such a way that the self-signed certificate becomes part of the
138 CA database and uses the same mechanisms for serial number generation
139 as all other certificate signing. The new flag '-selfsign' enables
140 this functionality. Adapt CA.sh and CA.pl.in.
143 *) Add functionality to check the public key of a certificate request
144 against a given private. This is useful to check that a certificate
145 request can be signed by that key (self-signing).
148 *) Make it possible to have multiple active certificates with the same
149 subject in the CA index file. This is done only if the keyword
150 'unique_subject' is set to 'no' in the main CA section (default
151 if 'CA_default') of the configuration file. The value is saved
152 with the database itself in a separate index attribute file,
153 named like the index file with '.attr' appended to the name.
156 *) Generate muti valued AVAs using '+' notation in config files for
160 *) Support for nameConstraints certificate extension.
163 *) Support for policyConstraints certificate extension.
166 *) Support for policyMappings certificate extension.
169 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
170 ENGINE as defaults for all supported algorithms irrespective of
171 the 'flags' parameter. 'flags' is now honoured, so applications
172 should make sure they are passing it correctly.
175 *) Make sure the default DSA_METHOD implementation only uses its
176 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
177 and change its own handlers to be NULL so as to remove unnecessary
178 indirection. This lets alternative implementations fallback to the
179 default implementation more easily.
182 *) Support for directoryName in GeneralName related extensions
186 *) Make it possible to link applications using Makefile.shared.
187 Make that possible even when linking against static libraries!
190 *) Support for single pass processing for S/MIME signing. This now
191 means that S/MIME signing can be done from a pipe, in addition
192 cleartext signing (multipart/signed type) is effectively streaming
193 and the signed data does not need to be all held in memory.
195 This is done with a new flag PKCS7_STREAM. When this flag is set
196 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
197 is done after the data is output (and digests calculated) in
201 *) Add full support for -rpath/-R, both in shared libraries and
202 applications, at least on the platforms where it's known how
206 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
207 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
208 will now compute a table of multiples of the generator that
209 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
210 faster (notably in the case of a single point multiplication,
212 [Nils Larsch, Bodo Moeller]
214 *) IPv6 support for certificate extensions. The various extensions
215 which use the IP:a.b.c.d can now take IPv6 addresses using the
216 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
220 *) Added an ENGINE that implements RSA by performing private key
221 exponentiations with the GMP library. The conversions to and from
222 GMP's mpz_t format aren't optimised nor are any montgomery forms
223 cached, and on x86 it appears OpenSSL's own performance has caught up.
224 However there are likely to be other architectures where GMP could
225 provide a boost. This ENGINE is not built in by default, but it can be
226 specified at Configure time and should be accompanied by the necessary
227 linker additions, eg;
228 ./config -DOPENSSL_USE_GMP -lgmp
231 *) "openssl engine" will not display ENGINE/DSO load failure errors when
232 testing availability of engines with "-t" - the old behaviour is
233 produced by increasing the feature's verbosity with "-tt".
236 *) ECDSA routines: under certain error conditions uninitialized BN objects
237 could be freed. Solution: make sure initialization is performed early
238 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
242 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
243 and DH_METHOD (eg. by ENGINE implementations) to override the normal
244 software implementations. For DSA and DH, parameter generation can
245 also be overriden by providing the appropriate method callbacks.
248 *) Change the "progress" mechanism used in key-generation and
249 primality testing to functions that take a new BN_GENCB pointer in
250 place of callback/argument pairs. The new API functions have "_ex"
251 postfixes and the older functions are reimplemented as wrappers for
252 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
253 declarations of the old functions to help (graceful) attempts to
254 migrate to the new functions. Also, the new key-generation API
255 functions operate on a caller-supplied key-structure and return
256 success/failure rather than returning a key or NULL - this is to
257 help make "keygen" another member function of RSA_METHOD etc.
259 Example for using the new callback interface:
261 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
265 BN_GENCB_set(&my_cb, my_callback, my_arg);
267 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
268 /* For the meaning of a, b in calls to my_callback(), see the
269 * documentation of the function that calls the callback.
270 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
271 * my_callback should return 1 if it wants BN_is_prime_ex()
272 * to continue, or 0 to stop.
277 *) Change the ZLIB compression method to be stateful, and make it
278 available to TLS with the number defined in
279 draft-ietf-tls-compression-04.txt.
282 *) Add the ASN.1 structures and functions for CertificatePair, which
283 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
285 CertificatePair ::= SEQUENCE {
286 forward [0] Certificate OPTIONAL,
287 reverse [1] Certificate OPTIONAL,
288 -- at least one of the pair shall be present -- }
290 Also implement the PEM functions to read and write certificate
291 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
293 This needed to be defined, mostly for the sake of the LDAP
294 attribute crossCertificatePair, but may prove useful elsewhere as
298 *) Make it possible to inhibit symlinking of shared libraries in
299 Makefile.shared, for Cygwin's sake.
302 *) Extend the BIGNUM API by creating new macros that behave like
305 void BN_set_sign(BIGNUM *a, int neg);
306 int BN_get_sign(const BIGNUM *a);
308 and avoid the need to access 'a->neg' directly in applications.
309 [Nils Larsch <nla@trustcenter.de>]
311 *) Implement fast modular reduction for pseudo-Mersenne primes
312 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
313 EC_GROUP_new_curve_GFp() will now automatically use this
315 [Nils Larsch <nla@trustcenter.de>]
317 *) Add new lock type (CRYPTO_LOCK_BN).
320 *) Change the ENGINE framework to automatically load engines
321 dynamically from specific directories unless they could be
322 found to already be built in or loaded. Move all the
323 current engines except for the cryptodev one to a new
325 The engines in engines/ are built as shared libraries if
326 the "shared" options was given to ./Configure or ./config.
327 Otherwise, they are inserted in libcrypto.a.
328 /usr/local/ssl/engines is the default directory for dynamic
329 engines, but that can be overriden at configure time through
330 the usual use of --prefix and/or --openssldir, and at run
331 time with the environment variable OPENSSL_ENGINES.
332 [Geoff Thorpe and Richard Levitte]
334 *) Add Makefile.shared, a helper makefile to build shared
335 libraries. Addapt Makefile.org.
338 *) Add version info to Win32 DLLs.
339 [Peter 'Luna' Runestig" <peter@runestig.com>]
341 *) Add new 'medium level' PKCS#12 API. Certificates and keys
342 can be added using this API to created arbitrary PKCS#12
343 files while avoiding the low level API.
345 New options to PKCS12_create(), key or cert can be NULL and
346 will then be omitted from the output file. The encryption
347 algorithm NIDs can be set to -1 for no encryption, the mac
348 iteration count can be set to 0 to omit the mac.
350 Enhance pkcs12 utility by making the -nokeys and -nocerts
351 options work when creating a PKCS#12 file. New option -nomac
352 to omit the mac, NONE can be set for an encryption algorithm.
353 New code is modified to use the enhanced PKCS12_create()
354 instead of the low level API.
357 *) Extend ASN1 encoder to support indefinite length constructed
358 encoding. This can output sequences tags and octet strings in
359 this form. Modify pk7_asn1.c to support indefinite length
360 encoding. This is experimental and needs additional code to
361 be useful, such as an ASN1 bio and some enhanced streaming
364 Extend template encode functionality so that tagging is passed
365 down to the template encoder.
368 *) Let 'openssl req' fail if an argument to '-newkey' is not
369 recognized instead of using RSA as a default.
372 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
373 As these are not official, they are not included in "ALL";
374 the "ECCdraft" ciphersuite group alias can be used to select them.
375 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
377 *) Add ECDH engine support.
378 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
380 *) Add ECDH in new directory crypto/ecdh/.
381 [Douglas Stebila (Sun Microsystems Laboratories)]
383 *) Let BN_rand_range() abort with an error after 100 iterations
384 without success (which indicates a broken PRNG).
387 *) Change BN_mod_sqrt() so that it verifies that the input value
388 is really the square of the return value. (Previously,
389 BN_mod_sqrt would show GIGO behaviour.)
392 *) Add named elliptic curves over binary fields from X9.62, SECG,
393 and WAP/WTLS; add OIDs that were still missing.
395 [Sheueling Chang Shantz and Douglas Stebila
396 (Sun Microsystems Laboratories)]
398 *) Extend the EC library for elliptic curves over binary fields
399 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
402 EC_GF2m_simple_method
406 EC_GROUP_new_curve_GF2m
407 EC_GROUP_set_curve_GF2m
408 EC_GROUP_get_curve_GF2m
409 EC_POINT_set_affine_coordinates_GF2m
410 EC_POINT_get_affine_coordinates_GF2m
411 EC_POINT_set_compressed_coordinates_GF2m
413 Point compression for binary fields is disabled by default for
414 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
417 As binary polynomials are represented as BIGNUMs, various members
418 of the EC_GROUP and EC_POINT data structures can be shared
419 between the implementations for prime fields and binary fields;
420 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
421 are essentially identical to their ..._GFp counterparts.
422 (For simplicity, the '..._GFp' prefix has been dropped from
423 various internal method names.)
425 An internal 'field_div' method (similar to 'field_mul' and
426 'field_sqr') has been added; this is used only for binary fields.
428 [Sheueling Chang Shantz and Douglas Stebila
429 (Sun Microsystems Laboratories)]
431 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
432 through methods ('mul', 'precompute_mult').
434 The generic implementations (now internally called 'ec_wNAF_mul'
435 and 'ec_wNAF_precomputed_mult') remain the default if these
436 methods are undefined.
438 [Sheueling Chang Shantz and Douglas Stebila
439 (Sun Microsystems Laboratories)]
441 *) New function EC_GROUP_get_degree, which is defined through
442 EC_METHOD. For curves over prime fields, this returns the bit
443 length of the modulus.
445 [Sheueling Chang Shantz and Douglas Stebila
446 (Sun Microsystems Laboratories)]
448 *) New functions EC_GROUP_dup, EC_POINT_dup.
449 (These simply call ..._new and ..._copy).
451 [Sheueling Chang Shantz and Douglas Stebila
452 (Sun Microsystems Laboratories)]
454 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
455 Polynomials are represented as BIGNUMs (where the sign bit is not
456 used) in the following functions [macros]:
459 BN_GF2m_sub [= BN_GF2m_add]
460 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
461 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
462 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
464 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
465 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
466 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
467 BN_GF2m_cmp [= BN_ucmp]
469 (Note that only the 'mod' functions are actually for fields GF(2^m).
470 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
472 For some functions, an the irreducible polynomial defining a
473 field can be given as an 'unsigned int[]' with strictly
474 decreasing elements giving the indices of those bits that are set;
475 i.e., p[] represents the polynomial
476 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
478 p[0] > p[1] > ... > p[k] = 0.
479 This applies to the following functions:
484 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
485 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
488 BN_GF2m_mod_solve_quad_arr
492 Conversion can be performed by the following functions:
497 bntest.c has additional tests for binary polynomial arithmetic.
499 Two implementations for BN_GF2m_mod_div() are available.
500 The default algorithm simply uses BN_GF2m_mod_inv() and
501 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
502 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
503 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
505 [Sheueling Chang Shantz and Douglas Stebila
506 (Sun Microsystems Laboratories)]
508 *) Add new error code 'ERR_R_DISABLED' that can be used when some
509 functionality is disabled at compile-time.
510 [Douglas Stebila <douglas.stebila@sun.com>]
512 *) Change default behaviour of 'openssl asn1parse' so that more
513 information is visible when viewing, e.g., a certificate:
515 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
516 mode the content of non-printable OCTET STRINGs is output in a
517 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
518 avoid the appearance of a printable string.
519 [Nils Larsch <nla@trustcenter.de>]
521 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
523 EC_GROUP_set_asn1_flag()
524 EC_GROUP_get_asn1_flag()
525 EC_GROUP_set_point_conversion_form()
526 EC_GROUP_get_point_conversion_form()
527 These control ASN1 encoding details:
528 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
529 has been set to OPENSSL_EC_NAMED_CURVE.
530 - Points are encoded in uncompressed form by default; options for
531 asn1_for are as for point2oct, namely
532 POINT_CONVERSION_COMPRESSED
533 POINT_CONVERSION_UNCOMPRESSED
534 POINT_CONVERSION_HYBRID
536 Also add 'seed' and 'seed_len' members to EC_GROUP with access
540 EC_GROUP_get_seed_len()
541 This is used only for ASN1 purposes (so far).
542 [Nils Larsch <nla@trustcenter.de>]
544 *) Add 'field_type' member to EC_METHOD, which holds the NID
545 of the appropriate field type OID. The new function
546 EC_METHOD_get_field_type() returns this value.
547 [Nils Larsch <nla@trustcenter.de>]
554 providing useful interfaces to EC_POINT_point2oct() and
555 EC_POINT_oct2point().
556 [Nils Larsch <nla@trustcenter.de>]
558 *) Change internals of the EC library so that the functions
559 EC_GROUP_set_generator()
560 EC_GROUP_get_generator()
562 EC_GROUP_get_cofactor()
563 are implemented directly in crypto/ec/ec_lib.c and not dispatched
564 to methods, which would lead to unnecessary code duplication when
565 adding different types of curves.
566 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
568 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
569 arithmetic, and such that modified wNAFs are generated
570 (which avoid length expansion in many cases).
573 *) Add a function EC_GROUP_check_discriminant() (defined via
574 EC_METHOD) that verifies that the curve discriminant is non-zero.
576 Add a function EC_GROUP_check() that makes some sanity tests
577 on a EC_GROUP, its generator and order. This includes
578 EC_GROUP_check_discriminant().
579 [Nils Larsch <nla@trustcenter.de>]
581 *) Add ECDSA in new directory crypto/ecdsa/.
583 Add applications 'openssl ecparam' and 'openssl ecdsa'
584 (these are based on 'openssl dsaparam' and 'openssl dsa').
586 ECDSA support is also included in various other files across the
587 library. Most notably,
588 - 'openssl req' now has a '-newkey ecdsa:file' option;
589 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
590 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
591 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
592 them suitable for ECDSA where domain parameters must be
593 extracted before the specific public key;
594 - ECDSA engine support has been added.
595 [Nils Larsch <nla@trustcenter.de>]
597 *) Include some named elliptic curves, and add OIDs from X9.62,
598 SECG, and WAP/WTLS. Each curve can be obtained from the new
600 EC_GROUP_new_by_nid(),
601 and the list of available named curves can be obtained with
602 EC_get_builtin_curves().
603 Also add a 'curve_name' member to EC_GROUP objects, which can be
607 [Nils Larsch <nla@trustcenter.de, Bodo Moeller]
609 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
610 was actually never needed) and in BN_mul(). The removal in BN_mul()
611 required a small change in bn_mul_part_recursive() and the addition
612 of the functions bn_cmp_part_words(), bn_sub_part_words() and
613 bn_add_part_words(), which do the same thing as bn_cmp_words(),
614 bn_sub_words() and bn_add_words() except they take arrays with
618 Changes between 0.9.7c and 0.9.7d [xx XXX XXXX]
620 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
621 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
623 *) Use the correct content when signing type "other".
626 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
628 *) Fix various bugs revealed by running the NISCC test suite:
630 Stop out of bounds reads in the ASN1 code when presented with
631 invalid tags (CAN-2003-0543 and CAN-2003-0544).
633 Free up ASN1_TYPE correctly if ANY type is invalid (CAN-2003-0545).
635 If verify callback ignores invalid public key errors don't try to check
636 certificate signature with the NULL public key.
640 *) New -ignore_err option in ocsp application to stop the server
641 exiting on the first error in a request.
644 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
645 if the server requested one: as stated in TLS 1.0 and SSL 3.0
649 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
650 extra data after the compression methods not only for TLS 1.0
651 but also for SSL 3.0 (as required by the specification).
652 [Bodo Moeller; problem pointed out by Matthias Loepfe]
654 *) Change X509_certificate_type() to mark the key as exported/exportable
655 when it's 512 *bits* long, not 512 bytes.
658 *) Change AES_cbc_encrypt() so it outputs exact multiple of
659 blocks during encryption.
662 *) Various fixes to base64 BIO and non blocking I/O. On write
663 flushes were not handled properly if the BIO retried. On read
664 data was not being buffered properly and had various logic bugs.
665 This also affects blocking I/O when the data being decoded is a
669 *) Various S/MIME bugfixes and compatibility changes:
670 output correct application/pkcs7 MIME type if
671 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
672 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
673 of files as .eml work). Correctly handle very long lines in MIME
677 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
679 *) Countermeasure against the Klima-Pokorny-Rosa extension of
680 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
681 a protocol version number mismatch like a decryption error
682 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
685 *) Turn on RSA blinding by default in the default implementation
686 to avoid a timing attack. Applications that don't want it can call
687 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
688 They would be ill-advised to do so in most cases.
689 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
691 *) Change RSA blinding code so that it works when the PRNG is not
692 seeded (in this case, the secret RSA exponent is abused as
693 an unpredictable seed -- if it is not unpredictable, there
694 is no point in blinding anyway). Make RSA blinding thread-safe
695 by remembering the creator's thread ID in rsa->blinding and
696 having all other threads use local one-time blinding factors
697 (this requires more computation than sharing rsa->blinding, but
698 avoids excessive locking; and if an RSA object is not shared
699 between threads, blinding will still be very fast).
702 yet to be integrated into this CVS branch:
703 - Geoff's ENGINE_set_default() fix
705 *) Target "mingw" now allows native Windows code to be generated in
706 the Cygwin environment as well as with the MinGW compiler.
709 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
711 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
712 via timing by performing a MAC computation even if incorrrect
713 block cipher padding has been found. This is a countermeasure
714 against active attacks where the attacker has to distinguish
715 between bad padding and a MAC verification error. (CAN-2003-0078)
717 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
718 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
719 Martin Vuagnoux (EPFL, Ilion)]
721 *) Make the no-err option work as intended. The intention with no-err
722 is not to have the whole error stack handling routines removed from
723 libcrypto, it's only intended to remove all the function name and
724 reason texts, thereby removing some of the footprint that may not
725 be interesting if those errors aren't displayed anyway.
727 NOTE: it's still possible for any application or module to have it's
728 own set of error texts inserted. The routines are there, just not
729 used by default when no-err is given.
732 *) Add support for FreeBSD on IA64.
733 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
735 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
736 Kerberos function mit_des_cbc_cksum(). Before this change,
737 the value returned by DES_cbc_cksum() was like the one from
738 mit_des_cbc_cksum(), except the bytes were swapped.
739 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
741 *) Allow an application to disable the automatic SSL chain building.
742 Before this a rather primitive chain build was always performed in
743 ssl3_output_cert_chain(): an application had no way to send the
744 correct chain if the automatic operation produced an incorrect result.
746 Now the chain builder is disabled if either:
748 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
750 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
752 The reasoning behind this is that an application would not want the
753 auto chain building to take place if extra chain certificates are
754 present and it might also want a means of sending no additional
755 certificates (for example the chain has two certificates and the
759 *) Add the possibility to build without the ENGINE framework.
760 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
762 *) Under Win32 gmtime() can return NULL: check return value in
763 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
766 *) DSA routines: under certain error conditions uninitialized BN objects
767 could be freed. Solution: make sure initialization is performed early
768 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
769 Nils Larsch <nla@trustcenter.de> via PR#459)
772 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
773 checked on reconnect on the client side, therefore session resumption
774 could still fail with a "ssl session id is different" error. This
775 behaviour is masked when SSL_OP_ALL is used due to
776 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
777 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
781 *) IA-32 assembler support enhancements: unified ELF targets, support
782 for SCO/Caldera platforms, fix for Cygwin shared build.
785 *) Add support for FreeBSD on sparc64. As a consequence, support for
786 FreeBSD on non-x86 processors is separate from x86 processors on
787 the config script, much like the NetBSD support.
788 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
790 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
792 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
795 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
796 code (06) was taken as the first octet of the session ID and the last
797 octet was ignored consequently. As a result SSLv2 client side session
798 caching could not have worked due to the session ID mismatch between
800 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
804 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
805 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
809 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
810 seems that in spite of existing for more than a year, many application
811 author have done nothing to provide the necessary callbacks, which
812 means that this particular engine will not work properly anywhere.
813 This is a very unfortunate situation which forces us, in the name
814 of usability, to give the hw_ncipher.c a static lock, which is part
816 NOTE: This is for the 0.9.7 series ONLY. This hack will never
817 appear in 0.9.8 or later. We EXPECT application authors to have
818 dealt properly with this when 0.9.8 is released (unless we actually
819 make such changes in the libcrypto locking code that changes will
820 have to be made anyway).
823 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
824 octets have been read, EOF or an error occurs. Without this change
825 some truncated ASN1 structures will not produce an error.
828 *) Disable Heimdal support, since it hasn't been fully implemented.
829 Still give the possibility to force the use of Heimdal, but with
830 warnings and a request that patches get sent to openssl-dev.
833 *) Add the VC-CE target, introduce the WINCE sysname, and add
834 INSTALL.WCE and appropriate conditionals to make it build.
835 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
837 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
838 cygssl-x.y.z.dll, where x, y and z are the major, minor and
839 edit numbers of the version.
840 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
842 *) Introduce safe string copy and catenation functions
843 (BUF_strlcpy() and BUF_strlcat()).
844 [Ben Laurie (CHATS) and Richard Levitte]
846 *) Avoid using fixed-size buffers for one-line DNs.
849 *) Add BUF_MEM_grow_clean() to avoid information leakage when
850 resizing buffers containing secrets, and use where appropriate.
853 *) Avoid using fixed size buffers for configuration file location.
856 *) Avoid filename truncation for various CA files.
859 *) Use sizeof in preference to magic numbers.
862 *) Avoid filename truncation in cert requests.
865 *) Add assertions to check for (supposedly impossible) buffer
869 *) Don't cache truncated DNS entries in the local cache (this could
870 potentially lead to a spoofing attack).
873 *) Fix various buffers to be large enough for hex/decimal
874 representations in a platform independent manner.
877 *) Add CRYPTO_realloc_clean() to avoid information leakage when
878 resizing buffers containing secrets, and use where appropriate.
881 *) Add BIO_indent() to avoid much slightly worrying code to do
885 *) Convert sprintf()/BIO_puts() to BIO_printf().
888 *) buffer_gets() could terminate with the buffer only half
892 *) Add assertions to prevent user-supplied crypto functions from
893 overflowing internal buffers by having large block sizes, etc.
896 *) New OPENSSL_assert() macro (similar to assert(), but enabled
900 *) Eliminate unused copy of key in RC4.
903 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
906 *) Fix off-by-one error in EGD path.
909 *) If RANDFILE path is too long, ignore instead of truncating.
912 *) Eliminate unused and incorrectly sized X.509 structure
916 *) Eliminate unused and dangerous function knumber().
919 *) Eliminate unused and dangerous structure, KSSL_ERR.
922 *) Protect against overlong session ID context length in an encoded
923 session object. Since these are local, this does not appear to be
927 *) Change from security patch (see 0.9.6e below) that did not affect
928 the 0.9.6 release series:
930 Remote buffer overflow in SSL3 protocol - an attacker could
931 supply an oversized master key in Kerberos-enabled versions.
935 *) Change the SSL kerb5 codes to match RFC 2712.
938 *) Make -nameopt work fully for req and add -reqopt switch.
939 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
941 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
942 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
944 *) Make sure tests can be performed even if the corresponding algorithms
945 have been removed entirely. This was also the last step to make
946 OpenSSL compilable with DJGPP under all reasonable conditions.
947 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
949 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
950 to allow version independent disabling of normally unselected ciphers,
951 which may be activated as a side-effect of selecting a single cipher.
953 (E.g., cipher list string "RSA" enables ciphersuites that are left
954 out of "ALL" because they do not provide symmetric encryption.
955 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
956 [Lutz Jaenicke, Bodo Moeller]
958 *) Add appropriate support for separate platform-dependent build
959 directories. The recommended way to make a platform-dependent
960 build directory is the following (tested on Linux), maybe with
963 # Place yourself outside of the OpenSSL source tree. In
964 # this example, the environment variable OPENSSL_SOURCE
965 # is assumed to contain the absolute OpenSSL source directory.
966 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
967 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
968 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
969 mkdir -p `dirname $F`
970 ln -s $OPENSSL_SOURCE/$F $F
973 To be absolutely sure not to disturb the source tree, a "make clean"
974 is a good thing. If it isn't successfull, don't worry about it,
975 it probably means the source directory is very clean.
978 *) Make sure any ENGINE control commands make local copies of string
979 pointers passed to them whenever necessary. Otherwise it is possible
980 the caller may have overwritten (or deallocated) the original string
981 data when a later ENGINE operation tries to use the stored values.
982 [Götz Babin-Ebell <babinebell@trustcenter.de>]
984 *) Improve diagnostics in file reading and command-line digests.
985 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
987 *) Add AES modes CFB and OFB to the object database. Correct an
988 error in AES-CFB decryption.
991 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
992 allows existing EVP_CIPHER_CTX structures to be reused after
993 calling EVP_*Final(). This behaviour is used by encryption
994 BIOs and some applications. This has the side effect that
995 applications must explicitly clean up cipher contexts with
996 EVP_CIPHER_CTX_cleanup() or they will leak memory.
999 *) Check the values of dna and dnb in bn_mul_recursive before calling
1000 bn_mul_comba (a non zero value means the a or b arrays do not contain
1001 n2 elements) and fallback to bn_mul_normal if either is not zero.
1004 *) Fix escaping of non-ASCII characters when using the -subj option
1005 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
1008 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
1009 form for "surname", serialNumber has no short form.
1010 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
1011 therefore remove "mail" short name for "internet 7".
1012 The OID for unique identifiers in X509 certificates is
1013 x500UniqueIdentifier, not uniqueIdentifier.
1014 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
1017 *) Add an "init" command to the ENGINE config module and auto initialize
1018 ENGINEs. Without any "init" command the ENGINE will be initialized
1019 after all ctrl commands have been executed on it. If init=1 the
1020 ENGINE is initailized at that point (ctrls before that point are run
1021 on the uninitialized ENGINE and after on the initialized one). If
1022 init=0 then the ENGINE will not be iniatialized at all.
1025 *) Fix the 'app_verify_callback' interface so that the user-defined
1026 argument is actually passed to the callback: In the
1027 SSL_CTX_set_cert_verify_callback() prototype, the callback
1028 declaration has been changed from
1031 int (*cb)(X509_STORE_CTX *,void *);
1032 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
1033 i=s->ctx->app_verify_callback(&ctx)
1034 has been changed into
1035 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
1037 To update applications using SSL_CTX_set_cert_verify_callback(),
1038 a dummy argument can be added to their callback functions.
1039 [D. K. Smetters <smetters@parc.xerox.com>]
1041 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
1042 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
1044 *) Add and OPENSSL_LOAD_CONF define which will cause
1045 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
1046 This allows older applications to transparently support certain
1047 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
1048 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
1049 load the config file and OPENSSL_add_all_algorithms_conf() which will
1050 always load it have also been added.
1053 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
1054 Adjust NIDs and EVP layer.
1055 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1057 *) Config modules support in openssl utility.
1059 Most commands now load modules from the config file,
1060 though in a few (such as version) this isn't done
1061 because it couldn't be used for anything.
1063 In the case of ca and req the config file used is
1064 the same as the utility itself: that is the -config
1065 command line option can be used to specify an
1069 *) Move default behaviour from OPENSSL_config(). If appname is NULL
1070 use "openssl_conf" if filename is NULL use default openssl config file.
1073 *) Add an argument to OPENSSL_config() to allow the use of an alternative
1074 config section name. Add a new flag to tolerate a missing config file
1075 and move code to CONF_modules_load_file().
1078 *) Support for crypto accelerator cards from Accelerated Encryption
1079 Processing, www.aep.ie. (Use engine 'aep')
1080 The support was copied from 0.9.6c [engine] and adapted/corrected
1081 to work with the new engine framework.
1082 [AEP Inc. and Richard Levitte]
1084 *) Support for SureWare crypto accelerator cards from Baltimore
1085 Technologies. (Use engine 'sureware')
1086 The support was copied from 0.9.6c [engine] and adapted
1087 to work with the new engine framework.
1090 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
1091 make the newer ENGINE framework commands for the CHIL engine work.
1092 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
1094 *) Make it possible to produce shared libraries on ReliantUNIX.
1095 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
1097 *) Add the configuration target debug-linux-ppro.
1098 Make 'openssl rsa' use the general key loading routines
1099 implemented in apps.c, and make those routines able to
1100 handle the key format FORMAT_NETSCAPE and the variant
1102 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1104 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
1105 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1107 *) Add -keyform to rsautl, and document -engine.
1108 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
1110 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
1111 BIO_R_NO_SUCH_FILE error code rather than the generic
1112 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
1115 *) Add new functions
1117 ERR_peek_last_error_line
1118 ERR_peek_last_error_line_data.
1119 These are similar to
1122 ERR_peek_error_line_data,
1123 but report on the latest error recorded rather than the first one
1124 still in the error queue.
1125 [Ben Laurie, Bodo Moeller]
1127 *) default_algorithms option in ENGINE config module. This allows things
1129 default_algorithms = ALL
1130 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
1133 *) Prelminary ENGINE config module.
1136 *) New experimental application configuration code.
1139 *) Change the AES code to follow the same name structure as all other
1140 symmetric ciphers, and behave the same way. Move everything to
1141 the directory crypto/aes, thereby obsoleting crypto/rijndael.
1142 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1144 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
1145 [Ben Laurie and Theo de Raadt]
1147 *) Add option to output public keys in req command.
1148 [Massimiliano Pala madwolf@openca.org]
1150 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
1151 (up to about 10% better than before for P-192 and P-224).
1154 *) New functions/macros
1156 SSL_CTX_set_msg_callback(ctx, cb)
1157 SSL_CTX_set_msg_callback_arg(ctx, arg)
1158 SSL_set_msg_callback(ssl, cb)
1159 SSL_set_msg_callback_arg(ssl, arg)
1161 to request calling a callback function
1163 void cb(int write_p, int version, int content_type,
1164 const void *buf, size_t len, SSL *ssl, void *arg)
1166 whenever a protocol message has been completely received
1167 (write_p == 0) or sent (write_p == 1). Here 'version' is the
1168 protocol version according to which the SSL library interprets
1169 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
1170 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
1171 the content type as defined in the SSL 3.0/TLS 1.0 protocol
1172 specification (change_cipher_spec(20), alert(21), handshake(22)).
1173 'buf' and 'len' point to the actual message, 'ssl' to the
1174 SSL object, and 'arg' is the application-defined value set by
1175 SSL[_CTX]_set_msg_callback_arg().
1177 'openssl s_client' and 'openssl s_server' have new '-msg' options
1178 to enable a callback that displays all protocol messages.
1181 *) Change the shared library support so shared libraries are built as
1182 soon as the corresponding static library is finished, and thereby get
1183 openssl and the test programs linked against the shared library.
1184 This still only happens when the keyword "shard" has been given to
1185 the configuration scripts.
1187 NOTE: shared library support is still an experimental thing, and
1188 backward binary compatibility is still not guaranteed.
1189 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
1191 *) Add support for Subject Information Access extension.
1192 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1194 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
1195 additional bytes when new memory had to be allocated, not just
1196 when reusing an existing buffer.
1199 *) New command line and configuration option 'utf8' for the req command.
1200 This allows field values to be specified as UTF8 strings.
1203 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
1204 runs for the former and machine-readable output for the latter.
1207 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
1208 of the e-mail address in the DN (i.e., it will go into a certificate
1209 extension only). The new configuration file option 'email_in_dn = no'
1210 has the same effect.
1211 [Massimiliano Pala madwolf@openca.org]
1213 *) Change all functions with names starting with des_ to be starting
1214 with DES_ instead. Add wrappers that are compatible with libdes,
1215 but are named _ossl_old_des_*. Finally, add macros that map the
1216 des_* symbols to the corresponding _ossl_old_des_* if libdes
1217 compatibility is desired. If OpenSSL 0.9.6c compatibility is
1218 desired, the des_* symbols will be mapped to DES_*, with one
1221 Since we provide two compatibility mappings, the user needs to
1222 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
1223 compatibility is desired. The default (i.e., when that macro
1224 isn't defined) is OpenSSL 0.9.6c compatibility.
1226 There are also macros that enable and disable the support of old
1227 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
1228 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
1229 are defined, the default will apply: to support the old des routines.
1231 In either case, one must include openssl/des.h to get the correct
1232 definitions. Do not try to just include openssl/des_old.h, that
1235 NOTE: This is a major break of an old API into a new one. Software
1236 authors are encouraged to switch to the DES_ style functions. Some
1237 time in the future, des_old.h and the libdes compatibility functions
1238 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
1239 default), and then completely removed.
1242 *) Test for certificates which contain unsupported critical extensions.
1243 If such a certificate is found during a verify operation it is
1244 rejected by default: this behaviour can be overridden by either
1245 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
1246 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
1247 X509_supported_extension() has also been added which returns 1 if a
1248 particular extension is supported.
1251 *) Modify the behaviour of EVP cipher functions in similar way to digests
1252 to retain compatibility with existing code.
1255 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
1256 compatibility with existing code. In particular the 'ctx' parameter does
1257 not have to be to be initialized before the call to EVP_DigestInit() and
1258 it is tidied up after a call to EVP_DigestFinal(). New function
1259 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
1260 EVP_MD_CTX_copy() changed to not require the destination to be
1261 initialized valid and new function EVP_MD_CTX_copy_ex() added which
1262 requires the destination to be valid.
1264 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
1265 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
1268 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
1269 so that complete 'Handshake' protocol structures are kept in memory
1270 instead of overwriting 'msg_type' and 'length' with 'body' data.
1273 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
1274 [Massimo Santin via Richard Levitte]
1276 *) Major restructuring to the underlying ENGINE code. This includes
1277 reduction of linker bloat, separation of pure "ENGINE" manipulation
1278 (initialisation, etc) from functionality dealing with implementations
1279 of specific crypto iterfaces. This change also introduces integrated
1280 support for symmetric ciphers and digest implementations - so ENGINEs
1281 can now accelerate these by providing EVP_CIPHER and EVP_MD
1282 implementations of their own. This is detailed in crypto/engine/README
1283 as it couldn't be adequately described here. However, there are a few
1284 API changes worth noting - some RSA, DSA, DH, and RAND functions that
1285 were changed in the original introduction of ENGINE code have now
1286 reverted back - the hooking from this code to ENGINE is now a good
1287 deal more passive and at run-time, operations deal directly with
1288 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
1289 dereferencing through an ENGINE pointer any more. Also, the ENGINE
1290 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
1291 they were not being used by the framework as there is no concept of a
1292 BIGNUM_METHOD and they could not be generalised to the new
1293 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
1294 ENGINE_cpy() has been removed as it cannot be consistently defined in
1298 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
1301 *) Change mkdef.pl to sort symbols that get the same entry number,
1302 and make sure the automatically generated functions ERR_load_*
1303 become part of libeay.num as well.
1306 *) New function SSL_renegotiate_pending(). This returns true once
1307 renegotiation has been requested (either SSL_renegotiate() call
1308 or HelloRequest/ClientHello receveived from the peer) and becomes
1309 false once a handshake has been completed.
1310 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
1311 sends a HelloRequest, but does not ensure that a handshake takes
1312 place. SSL_renegotiate_pending() is useful for checking if the
1313 client has followed the request.)
1316 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
1317 By default, clients may request session resumption even during
1318 renegotiation (if session ID contexts permit); with this option,
1319 session resumption is possible only in the first handshake.
1321 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
1322 more bits available for options that should not be part of
1323 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
1326 *) Add some demos for certificate and certificate request creation.
1329 *) Make maximum certificate chain size accepted from the peer application
1330 settable (SSL*_get/set_max_cert_list()), as proposed by
1331 "Douglas E. Engert" <deengert@anl.gov>.
1334 *) Add support for shared libraries for Unixware-7
1335 (Boyd Lynn Gerber <gerberb@zenez.com>).
1338 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
1339 be done prior to destruction. Use this to unload error strings from
1340 ENGINEs that load their own error strings. NB: This adds two new API
1341 functions to "get" and "set" this destroy handler in an ENGINE.
1344 *) Alter all existing ENGINE implementations (except "openssl" and
1345 "openbsd") to dynamically instantiate their own error strings. This
1346 makes them more flexible to be built both as statically-linked ENGINEs
1347 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
1348 Also, add stub code to each that makes building them as self-contained
1349 shared-libraries easier (see README.ENGINE).
1352 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
1353 implementations into applications that are completely implemented in
1354 self-contained shared-libraries. The "dynamic" ENGINE exposes control
1355 commands that can be used to configure what shared-library to load and
1356 to control aspects of the way it is handled. Also, made an update to
1357 the README.ENGINE file that brings its information up-to-date and
1358 provides some information and instructions on the "dynamic" ENGINE
1359 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
1362 *) Make it possible to unload ranges of ERR strings with a new
1363 "ERR_unload_strings" function.
1366 *) Add a copy() function to EVP_MD.
1369 *) Make EVP_MD routines take a context pointer instead of just the
1370 md_data void pointer.
1373 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
1374 that the digest can only process a single chunk of data
1375 (typically because it is provided by a piece of
1376 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
1377 is only going to provide a single chunk of data, and hence the
1378 framework needn't accumulate the data for oneshot drivers.
1381 *) As with "ERR", make it possible to replace the underlying "ex_data"
1382 functions. This change also alters the storage and management of global
1383 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
1384 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
1385 index counters. The API functions that use this state have been changed
1386 to take a "class_index" rather than pointers to the class's local STACK
1387 and counter, and there is now an API function to dynamically create new
1388 classes. This centralisation allows us to (a) plug a lot of the
1389 thread-safety problems that existed, and (b) makes it possible to clean
1390 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
1391 such data would previously have always leaked in application code and
1392 workarounds were in place to make the memory debugging turn a blind eye
1393 to it. Application code that doesn't use this new function will still
1394 leak as before, but their memory debugging output will announce it now
1395 rather than letting it slide.
1397 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
1398 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
1399 has a return value to indicate success or failure.
1402 *) Make it possible to replace the underlying "ERR" functions such that the
1403 global state (2 LHASH tables and 2 locks) is only used by the "default"
1404 implementation. This change also adds two functions to "get" and "set"
1405 the implementation prior to it being automatically set the first time
1406 any other ERR function takes place. Ie. an application can call "get",
1407 pass the return value to a module it has just loaded, and that module
1408 can call its own "set" function using that value. This means the
1409 module's "ERR" operations will use (and modify) the error state in the
1410 application and not in its own statically linked copy of OpenSSL code.
1413 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
1414 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
1415 the operation, and provides a more encapsulated way for external code
1416 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
1417 to use these functions rather than manually incrementing the counts.
1419 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
1422 *) Add EVP test program.
1425 *) Add symmetric cipher support to ENGINE. Expect the API to change!
1428 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
1429 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
1430 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
1431 These allow a CRL to be built without having to access X509_CRL fields
1432 directly. Modify 'ca' application to use new functions.
1435 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
1436 bug workarounds. Rollback attack detection is a security feature.
1437 The problem will only arise on OpenSSL servers when TLSv1 is not
1438 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
1439 Software authors not wanting to support TLSv1 will have special reasons
1440 for their choice and can explicitly enable this option.
1441 [Bodo Moeller, Lutz Jaenicke]
1443 *) Rationalise EVP so it can be extended: don't include a union of
1444 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
1445 (similar to those existing for EVP_CIPHER_CTX).
1450 EVP_MD_CTX_init(&md); /* new function call */
1451 EVP_DigestInit(&md, EVP_sha1());
1452 EVP_DigestUpdate(&md, in, len);
1453 EVP_DigestFinal(&md, out, NULL);
1454 EVP_MD_CTX_cleanup(&md); /* new function call */
1458 *) Make DES key schedule conform to the usual scheme, as well as
1459 correcting its structure. This means that calls to DES functions
1460 now have to pass a pointer to a des_key_schedule instead of a
1461 plain des_key_schedule (which was actually always a pointer
1464 des_key_schedule ks;
1466 des_set_key_checked(..., &ks);
1467 des_ncbc_encrypt(..., &ks, ...);
1469 (Note that a later change renames 'des_...' into 'DES_...'.)
1472 *) Initial reduction of linker bloat: the use of some functions, such as
1473 PEM causes large amounts of unused functions to be linked in due to
1474 poor organisation. For example pem_all.c contains every PEM function
1475 which has a knock on effect of linking in large amounts of (unused)
1476 ASN1 code. Grouping together similar functions and splitting unrelated
1477 functions prevents this.
1480 *) Cleanup of EVP macros.
1483 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
1484 correct _ecb suffix.
1487 *) Add initial OCSP responder support to ocsp application. The
1488 revocation information is handled using the text based index
1489 use by the ca application. The responder can either handle
1490 requests generated internally, supplied in files (for example
1491 via a CGI script) or using an internal minimal server.
1494 *) Add configuration choices to get zlib compression for TLS.
1497 *) Changes to Kerberos SSL for RFC 2712 compliance:
1498 1. Implemented real KerberosWrapper, instead of just using
1499 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
1500 2. Implemented optional authenticator field of KerberosWrapper.
1502 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
1503 and authenticator structs; see crypto/krb5/.
1505 Generalized Kerberos calls to support multiple Kerberos libraries.
1506 [Vern Staats <staatsvr@asc.hpc.mil>,
1507 Jeffrey Altman <jaltman@columbia.edu>
1508 via Richard Levitte]
1510 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
1511 already does with RSA. testdsa.h now has 'priv_key/pub_key'
1512 values for each of the key sizes rather than having just
1513 parameters (and 'speed' generating keys each time).
1516 *) Speed up EVP routines.
1519 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
1520 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
1521 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
1522 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
1524 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
1525 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
1526 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
1529 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
1531 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
1534 *) Added the OS2-EMX target.
1535 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
1537 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
1538 to support NCONF routines in extension code. New function CONF_set_nconf()
1539 to allow functions which take an NCONF to also handle the old LHASH
1540 structure: this means that the old CONF compatible routines can be
1541 retained (in particular wrt extensions) without having to duplicate the
1542 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
1545 *) Enhance the general user interface with mechanisms for inner control
1546 and with possibilities to have yes/no kind of prompts.
1549 *) Change all calls to low level digest routines in the library and
1550 applications to use EVP. Add missing calls to HMAC_cleanup() and
1551 don't assume HMAC_CTX can be copied using memcpy().
1552 [Verdon Walker <VWalker@novell.com>, Steve Henson]
1554 *) Add the possibility to control engines through control names but with
1555 arbitrary arguments instead of just a string.
1556 Change the key loaders to take a UI_METHOD instead of a callback
1557 function pointer. NOTE: this breaks binary compatibility with earlier
1558 versions of OpenSSL [engine].
1559 Adapt the nCipher code for these new conditions and add a card insertion
1563 *) Enhance the general user interface with mechanisms to better support
1564 dialog box interfaces, application-defined prompts, the possibility
1565 to use defaults (for example default passwords from somewhere else)
1566 and interrupts/cancellations.
1569 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
1570 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
1573 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
1574 tidy up some unnecessarily weird code in 'sk_new()').
1575 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
1577 *) Change the key loading routines for ENGINEs to use the same kind
1578 callback (pem_password_cb) as all other routines that need this
1582 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
1583 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
1584 than this minimum value is recommended.
1587 *) New random seeder for OpenVMS, using the system process statistics
1588 that are easily reachable.
1591 *) Windows apparently can't transparently handle global
1592 variables defined in DLLs. Initialisations such as:
1594 const ASN1_ITEM *it = &ASN1_INTEGER_it;
1596 wont compile. This is used by the any applications that need to
1597 declare their own ASN1 modules. This was fixed by adding the option
1598 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
1599 needed for static libraries under Win32.
1602 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
1603 setting of purpose and trust fields. New X509_STORE trust and
1604 purpose functions and tidy up setting in other SSL functions.
1607 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
1608 structure. These are inherited by X509_STORE_CTX when it is
1609 initialised. This allows various defaults to be set in the
1610 X509_STORE structure (such as flags for CRL checking and custom
1611 purpose or trust settings) for functions which only use X509_STORE_CTX
1612 internally such as S/MIME.
1614 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
1615 trust settings if they are not set in X509_STORE. This allows X509_STORE
1616 purposes and trust (in S/MIME for example) to override any set by default.
1618 Add command line options for CRL checking to smime, s_client and s_server
1622 *) Initial CRL based revocation checking. If the CRL checking flag(s)
1623 are set then the CRL is looked up in the X509_STORE structure and
1624 its validity and signature checked, then if the certificate is found
1625 in the CRL the verify fails with a revoked error.
1627 Various new CRL related callbacks added to X509_STORE_CTX structure.
1629 Command line options added to 'verify' application to support this.
1631 This needs some additional work, such as being able to handle multiple
1632 CRLs with different times, extension based lookup (rather than just
1633 by subject name) and ultimately more complete V2 CRL extension
1637 *) Add a general user interface API (crypto/ui/). This is designed
1638 to replace things like des_read_password and friends (backward
1639 compatibility functions using this new API are provided).
1640 The purpose is to remove prompting functions from the DES code
1641 section as well as provide for prompting through dialog boxes in
1642 a window system and the like.
1645 *) Add "ex_data" support to ENGINE so implementations can add state at a
1646 per-structure level rather than having to store it globally.
1649 *) Make it possible for ENGINE structures to be copied when retrieved by
1650 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
1651 This causes the "original" ENGINE structure to act like a template,
1652 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
1653 operational state can be localised to each ENGINE structure, despite the
1654 fact they all share the same "methods". New ENGINE structures returned in
1655 this case have no functional references and the return value is the single
1656 structural reference. This matches the single structural reference returned
1657 by ENGINE_by_id() normally, when it is incremented on the pre-existing
1661 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
1662 needs to match any other type at all we need to manually clear the
1666 *) Changes to the "openssl engine" utility to include;
1667 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
1668 about an ENGINE's available control commands.
1669 - executing control commands from command line arguments using the
1670 '-pre' and '-post' switches. '-post' is only used if '-t' is
1671 specified and the ENGINE is successfully initialised. The syntax for
1672 the individual commands are colon-separated, for example;
1673 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
1676 *) New dynamic control command support for ENGINEs. ENGINEs can now
1677 declare their own commands (numbers), names (strings), descriptions,
1678 and input types for run-time discovery by calling applications. A
1679 subset of these commands are implicitly classed as "executable"
1680 depending on their input type, and only these can be invoked through
1681 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
1682 can be based on user input, config files, etc). The distinction is
1683 that "executable" commands cannot return anything other than a boolean
1684 result and can only support numeric or string input, whereas some
1685 discoverable commands may only be for direct use through
1686 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
1687 pointers, or other custom uses. The "executable" commands are to
1688 support parameterisations of ENGINE behaviour that can be
1689 unambiguously defined by ENGINEs and used consistently across any
1690 OpenSSL-based application. Commands have been added to all the
1691 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
1692 control over shared-library paths without source code alterations.
1695 *) Changed all ENGINE implementations to dynamically allocate their
1696 ENGINEs rather than declaring them statically. Apart from this being
1697 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
1698 this also allows the implementations to compile without using the
1699 internal engine_int.h header.
1702 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
1703 'const' value. Any code that should be able to modify a RAND_METHOD
1704 should already have non-const pointers to it (ie. they should only
1705 modify their own ones).
1708 *) Made a variety of little tweaks to the ENGINE code.
1709 - "atalla" and "ubsec" string definitions were moved from header files
1710 to C code. "nuron" string definitions were placed in variables
1711 rather than hard-coded - allowing parameterisation of these values
1712 later on via ctrl() commands.
1713 - Removed unused "#if 0"'d code.
1714 - Fixed engine list iteration code so it uses ENGINE_free() to release
1715 structural references.
1716 - Constified the RAND_METHOD element of ENGINE structures.
1717 - Constified various get/set functions as appropriate and added
1718 missing functions (including a catch-all ENGINE_cpy that duplicates
1719 all ENGINE values onto a new ENGINE except reference counts/state).
1720 - Removed NULL parameter checks in get/set functions. Setting a method
1721 or function to NULL is a way of cancelling out a previously set
1722 value. Passing a NULL ENGINE parameter is just plain stupid anyway
1723 and doesn't justify the extra error symbols and code.
1724 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
1725 flags from engine_int.h to engine.h.
1726 - Changed prototypes for ENGINE handler functions (init(), finish(),
1727 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
1730 *) Implement binary inversion algorithm for BN_mod_inverse in addition
1731 to the algorithm using long division. The binary algorithm can be
1732 used only if the modulus is odd. On 32-bit systems, it is faster
1733 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
1734 roughly 5-15% for 256-bit moduli), so we use it only for moduli
1735 up to 450 bits. In 64-bit environments, the binary algorithm
1736 appears to be advantageous for much longer moduli; here we use it
1737 for moduli up to 2048 bits.
1740 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
1741 could not support the combine flag in choice fields.
1744 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
1745 extensions from a certificate request to the certificate.
1748 *) Allow multiple 'certopt' and 'nameopt' options to be separated
1749 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
1750 file: this allows the display of the certificate about to be
1751 signed to be customised, to allow certain fields to be included
1752 or excluded and extension details. The old system didn't display
1753 multicharacter strings properly, omitted fields not in the policy
1754 and couldn't display additional details such as extensions.
1757 *) Function EC_POINTs_mul for multiple scalar multiplication
1758 of an arbitrary number of elliptic curve points
1759 \sum scalars[i]*points[i],
1760 optionally including the generator defined for the EC_GROUP:
1761 scalar*generator + \sum scalars[i]*points[i].
1763 EC_POINT_mul is a simple wrapper function for the typical case
1764 that the point list has just one item (besides the optional
1768 *) First EC_METHODs for curves over GF(p):
1770 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
1771 operations and provides various method functions that can also
1772 operate with faster implementations of modular arithmetic.
1774 EC_GFp_mont_method() reuses most functions that are part of
1775 EC_GFp_simple_method, but uses Montgomery arithmetic.
1777 [Bodo Moeller; point addition and point doubling
1778 implementation directly derived from source code provided by
1779 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
1781 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
1782 crypto/ec/ec_lib.c):
1784 Curves are EC_GROUP objects (with an optional group generator)
1785 based on EC_METHODs that are built into the library.
1787 Points are EC_POINT objects based on EC_GROUP objects.
1789 Most of the framework would be able to handle curves over arbitrary
1790 finite fields, but as there are no obvious types for fields other
1791 than GF(p), some functions are limited to that for now.
1794 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
1795 that the file contains a complete HTTP response.
1798 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
1799 change the def and num file printf format specifier from "%-40sXXX"
1800 to "%-39s XXX". The latter will always guarantee a space after the
1801 field while the former will cause them to run together if the field
1802 is 40 of more characters long.
1805 *) Constify the cipher and digest 'method' functions and structures
1806 and modify related functions to take constant EVP_MD and EVP_CIPHER
1810 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
1811 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
1814 *) Modify EVP_Digest*() routines so they now return values. Although the
1815 internal software routines can never fail additional hardware versions
1819 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
1821 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
1822 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
1825 ERR_R_NESTED_ASN1_ERROR
1827 ERR_R_MISSING_ASN1_EOS
1828 were 4 .. 9, conflicting with
1829 ERR_LIB_RSA (= ERR_R_RSA_LIB)
1831 ERR_LIB_PEM (= ERR_R_PEM_LIB).
1832 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
1834 Add new error code 'ERR_R_INTERNAL_ERROR'.
1837 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
1841 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
1842 sets the subject name for a new request or supersedes the
1843 subject name in a given request. Formats that can be parsed are
1844 'CN=Some Name, OU=myOU, C=IT'
1846 'CN=Some Name/OU=myOU/C=IT'.
1848 Add options '-batch' and '-verbose' to 'openssl req'.
1849 [Massimiliano Pala <madwolf@hackmasters.net>]
1851 *) Introduce the possibility to access global variables through
1852 functions on platform were that's the best way to handle exporting
1853 global variables in shared libraries. To enable this functionality,
1854 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1855 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1856 is normally done by Configure or something similar).
1858 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1859 in the source file (foo.c) like this:
1861 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1862 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1864 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1865 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1867 OPENSSL_DECLARE_GLOBAL(int,foo);
1868 #define foo OPENSSL_GLOBAL_REF(foo)
1869 OPENSSL_DECLARE_GLOBAL(double,bar);
1870 #define bar OPENSSL_GLOBAL_REF(bar)
1872 The #defines are very important, and therefore so is including the
1873 header file everywhere where the defined globals are used.
1875 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1876 of ASN.1 items, but that structure is a bit different.
1878 The largest change is in util/mkdef.pl which has been enhanced with
1879 better and easier to understand logic to choose which symbols should
1880 go into the Windows .def files as well as a number of fixes and code
1881 cleanup (among others, algorithm keywords are now sorted
1882 lexicographically to avoid constant rewrites).
1885 *) In BN_div() keep a copy of the sign of 'num' before writing the
1886 result to 'rm' because if rm==num the value will be overwritten
1887 and produce the wrong result if 'num' is negative: this caused
1888 problems with BN_mod() and BN_nnmod().
1891 *) Function OCSP_request_verify(). This checks the signature on an
1892 OCSP request and verifies the signer certificate. The signer
1893 certificate is just checked for a generic purpose and OCSP request
1897 *) Add OCSP_check_validity() function to check the validity of OCSP
1898 responses. OCSP responses are prepared in real time and may only
1899 be a few seconds old. Simply checking that the current time lies
1900 between thisUpdate and nextUpdate max reject otherwise valid responses
1901 caused by either OCSP responder or client clock inaccuracy. Instead
1902 we allow thisUpdate and nextUpdate to fall within a certain period of
1903 the current time. The age of the response can also optionally be
1904 checked. Two new options -validity_period and -status_age added to
1908 *) If signature or public key algorithm is unrecognized print out its
1909 OID rather that just UNKNOWN.
1912 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1913 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1914 ID to be generated from the issuer certificate alone which can then be
1915 passed to OCSP_id_issuer_cmp().
1918 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1919 ASN1 modules to export functions returning ASN1_ITEM pointers
1920 instead of the ASN1_ITEM structures themselves. This adds several
1921 new macros which allow the underlying ASN1 function/structure to
1922 be accessed transparently. As a result code should not use ASN1_ITEM
1923 references directly (such as &X509_it) but instead use the relevant
1924 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1925 use of the new ASN1 code on platforms where exporting structures
1926 is problematical (for example in shared libraries) but exporting
1927 functions returning pointers to structures is not.
1930 *) Add support for overriding the generation of SSL/TLS session IDs.
1931 These callbacks can be registered either in an SSL_CTX or per SSL.
1932 The purpose of this is to allow applications to control, if they wish,
1933 the arbitrary values chosen for use as session IDs, particularly as it
1934 can be useful for session caching in multiple-server environments. A
1935 command-line switch for testing this (and any client code that wishes
1936 to use such a feature) has been added to "s_server".
1937 [Geoff Thorpe, Lutz Jaenicke]
1939 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
1940 of the form '#if defined(...) || defined(...) || ...' and
1941 '#if !defined(...) && !defined(...) && ...'. This also avoids
1942 the growing number of special cases it was previously handling.
1945 *) Make all configuration macros available for application by making
1946 sure they are available in opensslconf.h, by giving them names starting
1947 with "OPENSSL_" to avoid conflicts with other packages and by making
1948 sure e_os2.h will cover all platform-specific cases together with
1950 Additionally, it is now possible to define configuration/platform-
1951 specific names (called "system identities"). In the C code, these
1952 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1953 macro with the name beginning with "OPENSSL_SYS_", which is determined
1954 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1958 *) New option -set_serial to 'req' and 'x509' this allows the serial
1959 number to use to be specified on the command line. Previously self
1960 signed certificates were hard coded with serial number 0 and the
1961 CA options of 'x509' had to use a serial number in a file which was
1965 *) New options to 'ca' utility to support V2 CRL entry extensions.
1966 Currently CRL reason, invalidity date and hold instruction are
1967 supported. Add new CRL extensions to V3 code and some new objects.
1970 *) New function EVP_CIPHER_CTX_set_padding() this is used to
1971 disable standard block padding (aka PKCS#5 padding) in the EVP
1972 API, which was previously mandatory. This means that the data is
1973 not padded in any way and so the total length much be a multiple
1974 of the block size, otherwise an error occurs.
1977 *) Initial (incomplete) OCSP SSL support.
1980 *) New function OCSP_parse_url(). This splits up a URL into its host,
1981 port and path components: primarily to parse OCSP URLs. New -url
1982 option to ocsp utility.
1985 *) New nonce behavior. The return value of OCSP_check_nonce() now
1986 reflects the various checks performed. Applications can decide
1987 whether to tolerate certain situations such as an absent nonce
1988 in a response when one was present in a request: the ocsp application
1989 just prints out a warning. New function OCSP_add1_basic_nonce()
1990 this is to allow responders to include a nonce in a response even if
1991 the request is nonce-less.
1994 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1995 skipped when using openssl x509 multiple times on a single input file,
1996 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1999 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
2000 set string type: to handle setting ASN1_TIME structures. Fix ca
2001 utility to correctly initialize revocation date of CRLs.
2004 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
2005 the clients preferred ciphersuites and rather use its own preferences.
2006 Should help to work around M$ SGC (Server Gated Cryptography) bug in
2007 Internet Explorer by ensuring unchanged hash method during stepup.
2008 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
2011 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
2012 to aes and add a new 'exist' option to print out symbols that don't
2016 *) Additional options to ocsp utility to allow flags to be set and
2017 additional certificates supplied.
2020 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
2021 OCSP client a number of certificate to only verify the response
2025 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
2026 handle the new API. Currently only ECB, CBC modes supported. Add new
2029 Add TLS AES ciphersuites as described in RFC3268, "Advanced
2030 Encryption Standard (AES) Ciphersuites for Transport Layer
2031 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
2032 not enabled by default and were not part of the "ALL" ciphersuite
2033 alias because they were not yet official; they could be
2034 explicitly requested by specifying the "AESdraft" ciphersuite
2035 group alias. In the final release of OpenSSL 0.9.7, the group
2036 alias is called "AES" and is part of "ALL".)
2037 [Ben Laurie, Steve Henson, Bodo Moeller]
2039 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
2040 request to response.
2043 *) Functions for OCSP responders. OCSP_request_onereq_count(),
2044 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
2045 extract information from a certificate request. OCSP_response_create()
2046 creates a response and optionally adds a basic response structure.
2047 OCSP_basic_add1_status() adds a complete single response to a basic
2048 response and returns the OCSP_SINGLERESP structure just added (to allow
2049 extensions to be included for example). OCSP_basic_add1_cert() adds a
2050 certificate to a basic response and OCSP_basic_sign() signs a basic
2051 response with various flags. New helper functions ASN1_TIME_check()
2052 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
2053 (converts ASN1_TIME to GeneralizedTime).
2056 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
2057 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
2058 structure from a certificate. X509_pubkey_digest() digests the public_key
2059 contents: this is used in various key identifiers.
2062 *) Make sk_sort() tolerate a NULL argument.
2063 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
2065 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
2066 passed by the function are trusted implicitly. If any of them signed the
2067 response then it is assumed to be valid and is not verified.
2070 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
2071 to data. This was previously part of the PKCS7 ASN1 code. This
2072 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
2073 [Steve Henson, reported by Kenneth R. Robinette
2074 <support@securenetterm.com>]
2076 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
2077 routines: without these tracing memory leaks is very painful.
2078 Fix leaks in PKCS12 and PKCS7 routines.
2081 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
2082 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
2083 effectively meant GeneralizedTime would never be used. Now it
2084 is initialised to -1 but X509_time_adj() now has to check the value
2085 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
2086 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
2087 [Steve Henson, reported by Kenneth R. Robinette
2088 <support@securenetterm.com>]
2090 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
2091 result in a zero length in the ASN1_INTEGER structure which was
2092 not consistent with the structure when d2i_ASN1_INTEGER() was used
2093 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
2094 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
2095 where it did not print out a minus for negative ASN1_INTEGER.
2098 *) Add summary printout to ocsp utility. The various functions which
2099 convert status values to strings have been renamed to:
2100 OCSP_response_status_str(), OCSP_cert_status_str() and
2101 OCSP_crl_reason_str() and are no longer static. New options
2102 to verify nonce values and to disable verification. OCSP response
2103 printout format cleaned up.
2106 *) Add additional OCSP certificate checks. These are those specified
2107 in RFC2560. This consists of two separate checks: the CA of the
2108 certificate being checked must either be the OCSP signer certificate
2109 or the issuer of the OCSP signer certificate. In the latter case the
2110 OCSP signer certificate must contain the OCSP signing extended key
2111 usage. This check is performed by attempting to match the OCSP
2112 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
2113 in the OCSP_CERTID structures of the response.
2116 *) Initial OCSP certificate verification added to OCSP_basic_verify()
2117 and related routines. This uses the standard OpenSSL certificate
2118 verify routines to perform initial checks (just CA validity) and
2119 to obtain the certificate chain. Then additional checks will be
2120 performed on the chain. Currently the root CA is checked to see
2121 if it is explicitly trusted for OCSP signing. This is used to set
2122 a root CA as a global signing root: that is any certificate that
2123 chains to that CA is an acceptable OCSP signing certificate.
2126 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
2127 extensions from a separate configuration file.
2128 As when reading extensions from the main configuration file,
2129 the '-extensions ...' option may be used for specifying the
2131 [Massimiliano Pala <madwolf@comune.modena.it>]
2133 *) New OCSP utility. Allows OCSP requests to be generated or
2134 read. The request can be sent to a responder and the output
2135 parsed, outputed or printed in text form. Not complete yet:
2136 still needs to check the OCSP response validity.
2139 *) New subcommands for 'openssl ca':
2140 'openssl ca -status <serial>' prints the status of the cert with
2141 the given serial number (according to the index file).
2142 'openssl ca -updatedb' updates the expiry status of certificates
2144 [Massimiliano Pala <madwolf@comune.modena.it>]
2146 *) New '-newreq-nodes' command option to CA.pl. This is like
2147 '-newreq', but calls 'openssl req' with the '-nodes' option
2148 so that the resulting key is not encrypted.
2149 [Damien Miller <djm@mindrot.org>]
2151 *) New configuration for the GNU Hurd.
2152 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
2154 *) Initial code to implement OCSP basic response verify. This
2155 is currently incomplete. Currently just finds the signer's
2156 certificate and verifies the signature on the response.
2159 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
2160 value of OPENSSLDIR. This is available via the new '-d' option
2161 to 'openssl version', and is also included in 'openssl version -a'.
2164 *) Allowing defining memory allocation callbacks that will be given
2165 file name and line number information in additional arguments
2166 (a const char* and an int). The basic functionality remains, as
2167 well as the original possibility to just replace malloc(),
2168 realloc() and free() by functions that do not know about these
2169 additional arguments. To register and find out the current
2170 settings for extended allocation functions, the following
2171 functions are provided:
2173 CRYPTO_set_mem_ex_functions
2174 CRYPTO_set_locked_mem_ex_functions
2175 CRYPTO_get_mem_ex_functions
2176 CRYPTO_get_locked_mem_ex_functions
2178 These work the same way as CRYPTO_set_mem_functions and friends.
2179 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
2180 extended allocation function is enabled.
2181 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
2182 a conventional allocation function is enabled.
2183 [Richard Levitte, Bodo Moeller]
2185 *) Finish off removing the remaining LHASH function pointer casts.
2186 There should no longer be any prototype-casting required when using
2187 the LHASH abstraction, and any casts that remain are "bugs". See
2188 the callback types and macros at the head of lhash.h for details
2189 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
2192 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
2193 If /dev/[u]random devices are not available or do not return enough
2194 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
2196 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
2197 /etc/entropy will be queried once each in this sequence, quering stops
2198 when enough entropy was collected without querying more sockets.
2201 *) Change the Unix RAND_poll() variant to be able to poll several
2202 random devices, as specified by DEVRANDOM, until a sufficient amount
2203 of data has been collected. We spend at most 10 ms on each file
2204 (select timeout) and read in non-blocking mode. DEVRANDOM now
2205 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
2206 (previously it was just the string "/dev/urandom"), so on typical
2207 platforms the 10 ms delay will never occur.
2208 Also separate out the Unix variant to its own file, rand_unix.c.
2209 For VMS, there's a currently-empty rand_vms.c.
2212 *) Move OCSP client related routines to ocsp_cl.c. These
2213 provide utility functions which an application needing
2214 to issue a request to an OCSP responder and analyse the
2215 response will typically need: as opposed to those which an
2216 OCSP responder itself would need which will be added later.
2218 OCSP_request_sign() signs an OCSP request with an API similar
2219 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
2220 response. OCSP_response_get1_basic() extracts basic response
2221 from response. OCSP_resp_find_status(): finds and extracts status
2222 information from an OCSP_CERTID structure (which will be created
2223 when the request structure is built). These are built from lower
2224 level functions which work on OCSP_SINGLERESP structures but
2225 wont normally be used unless the application wishes to examine
2226 extensions in the OCSP response for example.
2228 Replace nonce routines with a pair of functions.
2229 OCSP_request_add1_nonce() adds a nonce value and optionally
2230 generates a random value. OCSP_check_nonce() checks the
2231 validity of the nonce in an OCSP response.
2234 *) Change function OCSP_request_add() to OCSP_request_add0_id().
2235 This doesn't copy the supplied OCSP_CERTID and avoids the
2236 need to free up the newly created id. Change return type
2237 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
2238 This can then be used to add extensions to the request.
2239 Deleted OCSP_request_new(), since most of its functionality
2240 is now in OCSP_REQUEST_new() (and the case insensitive name
2241 clash) apart from the ability to set the request name which
2242 will be added elsewhere.
2245 *) Update OCSP API. Remove obsolete extensions argument from
2246 various functions. Extensions are now handled using the new
2247 OCSP extension code. New simple OCSP HTTP function which
2248 can be used to send requests and parse the response.
2251 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
2252 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
2253 uses the special reorder version of SET OF to sort the attributes
2254 and reorder them to match the encoded order. This resolves a long
2255 standing problem: a verify on a PKCS7 structure just after signing
2256 it used to fail because the attribute order did not match the
2257 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
2258 it uses the received order. This is necessary to tolerate some broken
2259 software that does not order SET OF. This is handled by encoding
2260 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
2261 to produce the required SET OF.
2264 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
2265 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
2266 files to get correct declarations of the ASN.1 item variables.
2269 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
2270 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
2271 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
2272 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
2273 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
2274 ASN1_ITEM and no wrapper functions.
2277 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
2278 replace the old function pointer based I/O routines. Change most of
2279 the *_d2i_bio() and *_d2i_fp() functions to use these.
2282 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
2283 lines, recognice more "algorithms" that can be deselected, and make
2284 it complain about algorithm deselection that isn't recognised.
2287 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
2288 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
2289 to use new functions. Add NO_ASN1_OLD which can be set to remove
2290 some old style ASN1 functions: this can be used to determine if old
2291 code will still work when these eventually go away.
2294 *) New extension functions for OCSP structures, these follow the
2295 same conventions as certificates and CRLs.
2298 *) New function X509V3_add1_i2d(). This automatically encodes and
2299 adds an extension. Its behaviour can be customised with various
2300 flags to append, replace or delete. Various wrappers added for
2301 certifcates and CRLs.
2304 *) Fix to avoid calling the underlying ASN1 print routine when
2305 an extension cannot be parsed. Correct a typo in the
2306 OCSP_SERVICELOC extension. Tidy up print OCSP format.
2309 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
2310 entries for variables.
2313 *) Add functionality to apps/openssl.c for detecting locking
2314 problems: As the program is single-threaded, all we have
2315 to do is register a locking callback using an array for
2316 storing which locks are currently held by the program.
2319 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
2320 SSL_get_ex_data_X509_STORE_idx(), which is used in
2321 ssl_verify_cert_chain() and thus can be called at any time
2322 during TLS/SSL handshakes so that thread-safety is essential.
2323 Unfortunately, the ex_data design is not at all suited
2324 for multi-threaded use, so it probably should be abolished.
2327 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
2328 [Broadcom, tweaked and integrated by Geoff Thorpe]
2330 *) Move common extension printing code to new function
2331 X509V3_print_extensions(). Reorganise OCSP print routines and
2332 implement some needed OCSP ASN1 functions. Add OCSP extensions.
2335 *) New function X509_signature_print() to remove duplication in some
2339 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
2340 set (this was treated exactly the same as SET OF previously). This
2341 is used to reorder the STACK representing the structure to match the
2342 encoding. This will be used to get round a problem where a PKCS7
2343 structure which was signed could not be verified because the STACK
2344 order did not reflect the encoded order.
2347 *) Reimplement the OCSP ASN1 module using the new code.
2350 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
2351 for its ASN1 operations. The old style function pointers still exist
2352 for now but they will eventually go away.
2355 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
2356 completely replaces the old ASN1 functionality with a table driven
2357 encoder and decoder which interprets an ASN1_ITEM structure describing
2358 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
2359 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
2360 has also been converted to the new form.
2363 *) Change BN_mod_exp_recp so that negative moduli are tolerated
2364 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
2365 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
2366 for negative moduli.
2369 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
2370 of not touching the result's sign bit.
2373 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
2377 *) Changed the LHASH code to use prototypes for callbacks, and created
2378 macros to declare and implement thin (optionally static) functions
2379 that provide type-safety and avoid function pointer casting for the
2380 type-specific callbacks.
2383 *) Added Kerberos Cipher Suites to be used with TLS, as written in
2385 [Veers Staats <staatsvr@asc.hpc.mil>,
2386 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
2388 *) Reformat the FAQ so the different questions and answers can be divided
2389 in sections depending on the subject.
2392 *) Have the zlib compression code load ZLIB.DLL dynamically under
2396 *) New function BN_mod_sqrt for computing square roots modulo a prime
2397 (using the probabilistic Tonelli-Shanks algorithm unless
2398 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
2399 be handled deterministically).
2400 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2402 *) Make BN_mod_inverse faster by explicitly handling small quotients
2403 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
2404 512 bits], about 30% for larger ones [1024 or 2048 bits].)
2407 *) New function BN_kronecker.
2410 *) Fix BN_gcd so that it works on negative inputs; the result is
2411 positive unless both parameters are zero.
2412 Previously something reasonably close to an infinite loop was
2413 possible because numbers could be growing instead of shrinking
2414 in the implementation of Euclid's algorithm.
2417 *) Fix BN_is_word() and BN_is_one() macros to take into account the
2418 sign of the number in question.
2420 Fix BN_is_word(a,w) to work correctly for w == 0.
2422 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
2423 because its test if the absolute value of 'a' equals 'w'.
2424 Note that BN_abs_is_word does *not* handle w == 0 reliably;
2425 it exists mostly for use in the implementations of BN_is_zero(),
2426 BN_is_one(), and BN_is_word().
2429 *) New function BN_swap.
2432 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
2433 the exponentiation functions are more likely to produce reasonable
2434 results on negative inputs.
2437 *) Change BN_mod_mul so that the result is always non-negative.
2438 Previously, it could be negative if one of the factors was negative;
2439 I don't think anyone really wanted that behaviour.
2442 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
2443 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
2444 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
2445 and add new functions:
2454 BN_mod_lshift1_quick
2458 These functions always generate non-negative results.
2460 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
2461 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
2463 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2464 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2465 be reduced modulo m.
2466 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2469 The following entry accidentily appeared in the CHANGES file
2470 distributed with OpenSSL 0.9.7. The modifications described in
2471 it do *not* apply to OpenSSL 0.9.7.
2473 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2474 was actually never needed) and in BN_mul(). The removal in BN_mul()
2475 required a small change in bn_mul_part_recursive() and the addition
2476 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2477 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2478 bn_sub_words() and bn_add_words() except they take arrays with
2483 *) In 'openssl passwd', verify passwords read from the terminal
2484 unless the '-salt' option is used (which usually means that
2485 verification would just waste user's time since the resulting
2486 hash is going to be compared with some given password hash)
2487 or the new '-noverify' option is used.
2489 This is an incompatible change, but it does not affect
2490 non-interactive use of 'openssl passwd' (passwords on the command
2491 line, '-stdin' option, '-in ...' option) and thus should not
2495 *) Remove all references to RSAref, since there's no more need for it.
2498 *) Make DSO load along a path given through an environment variable
2499 (SHLIB_PATH) with shl_load().
2502 *) Constify the ENGINE code as a result of BIGNUM constification.
2503 Also constify the RSA code and most things related to it. In a
2504 few places, most notable in the depth of the ASN.1 code, ugly
2505 casts back to non-const were required (to be solved at a later
2509 *) Make it so the openssl application has all engines loaded by default.
2512 *) Constify the BIGNUM routines a little more.
2515 *) Add the following functions:
2517 ENGINE_load_cswift()
2519 ENGINE_load_atalla()
2521 ENGINE_load_builtin_engines()
2523 That way, an application can itself choose if external engines that
2524 are built-in in OpenSSL shall ever be used or not. The benefit is
2525 that applications won't have to be linked with libdl or other dso
2526 libraries unless it's really needed.
2528 Changed 'openssl engine' to load all engines on demand.
2529 Changed the engine header files to avoid the duplication of some
2530 declarations (they differed!).
2533 *) 'openssl engine' can now list capabilities.
2536 *) Better error reporting in 'openssl engine'.
2539 *) Never call load_dh_param(NULL) in s_server.
2542 *) Add engine application. It can currently list engines by name and
2543 identity, and test if they are actually available.
2546 *) Improve RPM specification file by forcing symbolic linking and making
2547 sure the installed documentation is also owned by root.root.
2548 [Damien Miller <djm@mindrot.org>]
2550 *) Give the OpenSSL applications more possibilities to make use of
2551 keys (public as well as private) handled by engines.
2554 *) Add OCSP code that comes from CertCo.
2557 *) Add VMS support for the Rijndael code.
2560 *) Added untested support for Nuron crypto accelerator.
2563 *) Add support for external cryptographic devices. This code was
2564 previously distributed separately as the "engine" branch.
2565 [Geoff Thorpe, Richard Levitte]
2567 *) Rework the filename-translation in the DSO code. It is now possible to
2568 have far greater control over how a "name" is turned into a filename
2569 depending on the operating environment and any oddities about the
2570 different shared library filenames on each system.
2573 *) Support threads on FreeBSD-elf in Configure.
2576 *) Fix for SHA1 assembly problem with MASM: it produces
2577 warnings about corrupt line number information when assembling
2578 with debugging information. This is caused by the overlapping
2580 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2583 NCONF_get_number() has no error checking at all. As a replacement,
2584 NCONF_get_number_e() is defined (_e for "error checking") and is
2585 promoted strongly. The old NCONF_get_number is kept around for
2586 binary backward compatibility.
2587 Make it possible for methods to load from something other than a BIO,
2588 by providing a function pointer that is given a name instead of a BIO.
2589 For example, this could be used to load configuration data from an
2593 *) Fix for non blocking accept BIOs. Added new I/O special reason
2594 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2595 with non blocking I/O was not possible because no retry code was
2596 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2600 *) Added the beginnings of Rijndael support.
2603 *) Fix for bug in DirectoryString mask setting. Add support for
2604 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2605 to allow certificate printing to more controllable, additional
2606 'certopt' option to 'x509' to allow new printing options to be
2610 *) Clean old EAY MD5 hack from e_os.h.
2613 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
2615 *) Fix various bugs revealed by running the NISCC test suite:
2617 Stop out of bounds reads in the ASN1 code when presented with
2618 invalid tags (CAN-2003-0543 and CAN-2003-0544).
2620 If verify callback ignores invalid public key errors don't try to check
2621 certificate signature with the NULL public key.
2625 *) Fix various bugs revealed by running the NISCC test suite:
2627 Stop out of bounds reads in the ASN1 code when presented with
2628 invalid tags (CAN-2003-0543 and CAN-2003-0544).
2630 If verify callback ignores invalid public key errors don't try to check
2631 certificate signature with the NULL public key.
2635 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2636 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2640 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2641 extra data after the compression methods not only for TLS 1.0
2642 but also for SSL 3.0 (as required by the specification).
2643 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2645 *) Change X509_certificate_type() to mark the key as exported/exportable
2646 when it's 512 *bits* long, not 512 bytes.
2649 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
2651 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2652 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2653 a protocol version number mismatch like a decryption error
2654 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2657 *) Turn on RSA blinding by default in the default implementation
2658 to avoid a timing attack. Applications that don't want it can call
2659 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2660 They would be ill-advised to do so in most cases.
2661 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2663 *) Change RSA blinding code so that it works when the PRNG is not
2664 seeded (in this case, the secret RSA exponent is abused as
2665 an unpredictable seed -- if it is not unpredictable, there
2666 is no point in blinding anyway). Make RSA blinding thread-safe
2667 by remembering the creator's thread ID in rsa->blinding and
2668 having all other threads use local one-time blinding factors
2669 (this requires more computation than sharing rsa->blinding, but
2670 avoids excessive locking; and if an RSA object is not shared
2671 between threads, blinding will still be very fast).
2674 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
2676 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2677 via timing by performing a MAC computation even if incorrrect
2678 block cipher padding has been found. This is a countermeasure
2679 against active attacks where the attacker has to distinguish
2680 between bad padding and a MAC verification error. (CAN-2003-0078)
2682 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2683 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2684 Martin Vuagnoux (EPFL, Ilion)]
2686 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
2688 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
2689 memory from it's contents. This is done with a counter that will
2690 place alternating values in each byte. This can be used to solve
2691 two issues: 1) the removal of calls to memset() by highly optimizing
2692 compilers, and 2) cleansing with other values than 0, since those can
2693 be read through on certain media, for example a swap space on disk.
2696 *) Bugfix: client side session caching did not work with external caching,
2697 because the session->cipher setting was not restored when reloading
2698 from the external cache. This problem was masked, when
2699 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
2700 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
2703 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
2704 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
2705 [Zeev Lieber <zeev-l@yahoo.com>]
2707 *) Undo an undocumented change introduced in 0.9.6e which caused
2708 repeated calls to OpenSSL_add_all_ciphers() and
2709 OpenSSL_add_all_digests() to be ignored, even after calling
2713 *) Change the default configuration reader to deal with last line not
2714 being properly terminated.
2717 *) Change X509_NAME_cmp() so it applies the special rules on handling
2718 DN values that are of type PrintableString, as well as RDNs of type
2719 emailAddress where the value has the type ia5String.
2720 [stefank@valicert.com via Richard Levitte]
2722 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
2723 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
2724 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
2725 the bitwise-OR of the two for use by the majority of applications
2726 wanting this behaviour, and update the docs. The documented
2727 behaviour and actual behaviour were inconsistent and had been
2728 changing anyway, so this is more a bug-fix than a behavioural
2730 [Geoff Thorpe, diagnosed by Nadav Har'El]
2732 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
2733 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
2736 *) Fix initialization code race conditions in
2737 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
2738 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
2739 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
2740 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
2741 ssl2_get_cipher_by_char(),
2742 ssl3_get_cipher_by_char().
2743 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
2745 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
2746 the cached sessions are flushed, as the remove_cb() might use ex_data
2747 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
2748 (see [openssl.org #212]).
2749 [Geoff Thorpe, Lutz Jaenicke]
2751 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
2752 length, instead of the encoding length to d2i_ASN1_OBJECT.
2755 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
2757 *) [In 0.9.6g-engine release:]
2758 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
2759 [Lynn Gazis <lgazis@rainbow.com>]
2761 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
2763 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
2764 and get fix the header length calculation.
2765 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
2766 Alon Kantor <alonk@checkpoint.com> (and others),
2769 *) Use proper error handling instead of 'assertions' in buffer
2770 overflow checks added in 0.9.6e. This prevents DoS (the
2771 assertions could call abort()).
2772 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
2774 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
2776 *) Add various sanity checks to asn1_get_length() to reject
2777 the ASN1 length bytes if they exceed sizeof(long), will appear
2778 negative or the content length exceeds the length of the
2780 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2782 *) Fix cipher selection routines: ciphers without encryption had no flags
2783 for the cipher strength set and where therefore not handled correctly
2784 by the selection routines (PR #130).
2787 *) Fix EVP_dsa_sha macro.
2791 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
2792 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
2793 that was added in OpenSSL 0.9.6d.
2795 As the countermeasure turned out to be incompatible with some
2796 broken SSL implementations, the new option is part of SSL_OP_ALL.
2797 SSL_OP_ALL is usually employed when compatibility with weird SSL
2798 implementations is desired (e.g. '-bugs' option to 's_client' and
2799 's_server'), so the new option is automatically set in many
2803 *) Changes in security patch:
2805 Changes marked "(CHATS)" were sponsored by the Defense Advanced
2806 Research Projects Agency (DARPA) and Air Force Research Laboratory,
2807 Air Force Materiel Command, USAF, under agreement number
2810 *) Add various sanity checks to asn1_get_length() to reject
2811 the ASN1 length bytes if they exceed sizeof(long), will appear
2812 negative or the content length exceeds the length of the
2813 supplied buffer. (CAN-2002-0659)
2814 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2816 *) Assertions for various potential buffer overflows, not known to
2818 [Ben Laurie (CHATS)]
2820 *) Various temporary buffers to hold ASCII versions of integers were
2821 too small for 64 bit platforms. (CAN-2002-0655)
2822 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
2824 *) Remote buffer overflow in SSL3 protocol - an attacker could
2825 supply an oversized session ID to a client. (CAN-2002-0656)
2826 [Ben Laurie (CHATS)]
2828 *) Remote buffer overflow in SSL2 protocol - an attacker could
2829 supply an oversized client master key. (CAN-2002-0656)
2830 [Ben Laurie (CHATS)]
2832 Changes between 0.9.6c and 0.9.6d [9 May 2002]
2834 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
2835 encoded as NULL) with id-dsa-with-sha1.
2836 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
2838 *) Check various X509_...() return values in apps/req.c.
2839 [Nils Larsch <nla@trustcenter.de>]
2841 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
2842 an end-of-file condition would erronously be flagged, when the CRLF
2843 was just at the end of a processed block. The bug was discovered when
2844 processing data through a buffering memory BIO handing the data to a
2845 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
2846 <ptsekov@syntrex.com> and Nedelcho Stanev.
2849 *) Implement a countermeasure against a vulnerability recently found
2850 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
2851 before application data chunks to avoid the use of known IVs
2852 with data potentially chosen by the attacker.
2855 *) Fix length checks in ssl3_get_client_hello().
2858 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
2859 to prevent ssl3_read_internal() from incorrectly assuming that
2860 ssl3_read_bytes() found application data while handshake
2861 processing was enabled when in fact s->s3->in_read_app_data was
2862 merely automatically cleared during the initial handshake.
2863 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
2865 *) Fix object definitions for Private and Enterprise: they were not
2866 recognized in their shortname (=lowercase) representation. Extend
2867 obj_dat.pl to issue an error when using undefined keywords instead
2868 of silently ignoring the problem (Svenning Sorensen
2869 <sss@sss.dnsalias.net>).
2872 *) Fix DH_generate_parameters() so that it works for 'non-standard'
2873 generators, i.e. generators other than 2 and 5. (Previously, the
2874 code did not properly initialise the 'add' and 'rem' values to
2875 BN_generate_prime().)
2877 In the new general case, we do not insist that 'generator' is
2878 actually a primitive root: This requirement is rather pointless;
2879 a generator of the order-q subgroup is just as good, if not
2883 *) Map new X509 verification errors to alerts. Discovered and submitted by
2884 Tom Wu <tom@arcot.com>.
2887 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
2888 returning non-zero before the data has been completely received
2889 when using non-blocking I/O.
2890 [Bodo Moeller; problem pointed out by John Hughes]
2892 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
2893 [Ben Laurie, Lutz Jaenicke]
2895 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
2896 Yoram Zahavi <YoramZ@gilian.com>).
2899 *) Add information about CygWin 1.3 and on, and preserve proper
2900 configuration for the versions before that.
2901 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2903 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
2904 check whether we deal with a copy of a session and do not delete from
2905 the cache in this case. Problem reported by "Izhar Shoshani Levi"
2906 <izhar@checkpoint.com>.
2909 *) Do not store session data into the internal session cache, if it
2910 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
2911 flag is set). Proposed by Aslam <aslam@funk.com>.
2914 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
2918 *) [In 0.9.6d-engine release:]
2919 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2920 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2922 *) Add the configuration target linux-s390x.
2923 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
2925 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
2926 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
2927 variable as an indication that a ClientHello message has been
2928 received. As the flag value will be lost between multiple
2929 invocations of ssl3_accept when using non-blocking I/O, the
2930 function may not be aware that a handshake has actually taken
2931 place, thus preventing a new session from being added to the
2934 To avoid this problem, we now set s->new_session to 2 instead of
2935 using a local variable.
2936 [Lutz Jaenicke, Bodo Moeller]
2938 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
2939 if the SSL_R_LENGTH_MISMATCH error is detected.
2940 [Geoff Thorpe, Bodo Moeller]
2942 *) New 'shared_ldflag' column in Configure platform table.
2945 *) Fix EVP_CIPHER_mode macro.
2946 ["Dan S. Camper" <dan@bti.net>]
2948 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
2949 type, we must throw them away by setting rr->length to 0.
2950 [D P Chang <dpc@qualys.com>]
2952 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
2954 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
2955 <Dominikus.Scherkl@biodata.com>. (The previous implementation
2956 worked incorrectly for those cases where range = 10..._2 and
2957 3*range is two bits longer than range.)
2960 *) Only add signing time to PKCS7 structures if it is not already
2964 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
2965 OBJ_ld_ce should be OBJ_id_ce.
2966 Also some ip-pda OIDs in crypto/objects/objects.txt were
2967 incorrect (cf. RFC 3039).
2968 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
2970 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
2971 returns early because it has nothing to do.
2972 [Andy Schneider <andy.schneider@bjss.co.uk>]
2974 *) [In 0.9.6c-engine release:]
2975 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
2976 [Andy Schneider <andy.schneider@bjss.co.uk>]
2978 *) [In 0.9.6c-engine release:]
2979 Add support for Cryptographic Appliance's keyserver technology.
2980 (Use engine 'keyclient')
2981 [Cryptographic Appliances and Geoff Thorpe]
2983 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
2984 is called via tools/c89.sh because arguments have to be
2985 rearranged (all '-L' options must appear before the first object
2987 [Richard Shapiro <rshapiro@abinitio.com>]
2989 *) [In 0.9.6c-engine release:]
2990 Add support for Broadcom crypto accelerator cards, backported
2992 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
2994 *) [In 0.9.6c-engine release:]
2995 Add support for SureWare crypto accelerator cards from
2996 Baltimore Technologies. (Use engine 'sureware')
2997 [Baltimore Technologies and Mark Cox]
2999 *) [In 0.9.6c-engine release:]
3000 Add support for crypto accelerator cards from Accelerated
3001 Encryption Processing, www.aep.ie. (Use engine 'aep')
3002 [AEP Inc. and Mark Cox]
3004 *) Add a configuration entry for gcc on UnixWare.
3005 [Gary Benson <gbenson@redhat.com>]
3007 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
3008 messages are stored in a single piece (fixed-length part and
3009 variable-length part combined) and fix various bugs found on the way.
3012 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
3013 instead. BIO_gethostbyname() does not know what timeouts are
3014 appropriate, so entries would stay in cache even when they have
3016 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
3018 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
3019 faced with a pathologically small ClientHello fragment that does
3020 not contain client_version: Instead of aborting with an error,
3021 simply choose the highest available protocol version (i.e.,
3022 TLS 1.0 unless it is disabled). In practice, ClientHello
3023 messages are never sent like this, but this change gives us
3024 strictly correct behaviour at least for TLS.
3027 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
3028 never resets s->method to s->ctx->method when called from within
3029 one of the SSL handshake functions.
3030 [Bodo Moeller; problem pointed out by Niko Baric]
3032 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
3033 (sent using the client's version number) if client_version is
3034 smaller than the protocol version in use. Also change
3035 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
3036 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
3037 the client will at least see that alert.
3040 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
3044 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
3045 client receives HelloRequest while in a handshake.
3046 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
3048 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
3049 should end in 'break', not 'goto end' which circuments various
3050 cleanups done in state SSL_ST_OK. But session related stuff
3051 must be disabled for SSL_ST_OK in the case that we just sent a
3054 Also avoid some overhead by not calling ssl_init_wbio_buffer()
3055 before just sending a HelloRequest.
3056 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
3058 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
3059 reveal whether illegal block cipher padding was found or a MAC
3060 verification error occured. (Neither SSLerr() codes nor alerts
3061 are directly visible to potential attackers, but the information
3062 may leak via logfiles.)
3064 Similar changes are not required for the SSL 2.0 implementation
3065 because the number of padding bytes is sent in clear for SSL 2.0,
3066 and the extra bytes are just ignored. However ssl/s2_pkt.c
3067 failed to verify that the purported number of padding bytes is in
3071 *) Add OpenUNIX-8 support including shared libraries
3072 (Boyd Lynn Gerber <gerberb@zenez.com>).
3075 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
3076 'wristwatch attack' using huge encoding parameters (cf.
3077 James H. Manger's CRYPTO 2001 paper). Note that the
3078 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
3079 encoding parameters and hence was not vulnerable.
3082 *) BN_sqr() bug fix.
3083 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
3085 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
3086 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
3087 followed by modular reduction.
3088 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
3090 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
3091 equivalent based on BN_pseudo_rand() instead of BN_rand().
3094 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
3095 This function was broken, as the check for a new client hello message
3096 to handle SGC did not allow these large messages.
3097 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
3100 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
3103 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
3104 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
3107 *) Rework the configuration and shared library support for Tru64 Unix.
3108 The configuration part makes use of modern compiler features and
3109 still retains old compiler behavior for those that run older versions
3110 of the OS. The shared library support part includes a variant that
3111 uses the RPATH feature, and is available through the special
3112 configuration target "alpha-cc-rpath", which will never be selected
3114 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
3116 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
3117 with the same message size as in ssl3_get_certificate_request().
3118 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
3119 messages might inadvertently be reject as too long.
3120 [Petr Lampa <lampa@fee.vutbr.cz>]
3122 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
3125 *) Modified SSL library such that the verify_callback that has been set
3126 specificly for an SSL object with SSL_set_verify() is actually being
3127 used. Before the change, a verify_callback set with this function was
3128 ignored and the verify_callback() set in the SSL_CTX at the time of
3129 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
3130 to allow the necessary settings.
3133 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
3134 explicitly to NULL, as at least on Solaris 8 this seems not always to be
3135 done automatically (in contradiction to the requirements of the C
3136 standard). This made problems when used from OpenSSH.
3139 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
3140 dh->length and always used
3142 BN_rand_range(priv_key, dh->p).
3144 BN_rand_range() is not necessary for Diffie-Hellman, and this
3145 specific range makes Diffie-Hellman unnecessarily inefficient if
3146 dh->length (recommended exponent length) is much smaller than the
3147 length of dh->p. We could use BN_rand_range() if the order of
3148 the subgroup was stored in the DH structure, but we only have
3153 BN_rand(priv_key, l, ...)
3155 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
3161 RSA_eay_public_encrypt
3162 RSA_eay_private_decrypt
3163 RSA_eay_private_encrypt (signing)
3164 RSA_eay_public_decrypt (signature verification)
3166 (default implementations for RSA_public_encrypt,
3167 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
3168 always reject numbers >= n.
3171 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
3172 to synchronize access to 'locking_thread'. This is necessary on
3173 systems where access to 'locking_thread' (an 'unsigned long'
3174 variable) is not atomic.
3177 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
3178 *before* setting the 'crypto_lock_rand' flag. The previous code had
3179 a race condition if 0 is a valid thread ID.
3180 [Travis Vitek <vitek@roguewave.com>]
3182 *) Add support for shared libraries under Irix.
3183 [Albert Chin-A-Young <china@thewrittenword.com>]
3185 *) Add configuration option to build on Linux on both big-endian and
3187 [Ralf Baechle <ralf@uni-koblenz.de>]
3189 *) Add the possibility to create shared libraries on HP-UX.
3192 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
3194 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
3195 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
3196 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
3197 PRNG state recovery was possible based on the output of
3198 one PRNG request appropriately sized to gain knowledge on
3199 'md' followed by enough consecutive 1-byte PRNG requests
3200 to traverse all of 'state'.
3202 1. When updating 'md_local' (the current thread's copy of 'md')
3203 during PRNG output generation, hash all of the previous
3204 'md_local' value, not just the half used for PRNG output.
3206 2. Make the number of bytes from 'state' included into the hash
3207 independent from the number of PRNG bytes requested.
3209 The first measure alone would be sufficient to avoid
3210 Markku-Juhani's attack. (Actually it had never occurred
3211 to me that the half of 'md_local' used for chaining was the
3212 half from which PRNG output bytes were taken -- I had always
3213 assumed that the secret half would be used.) The second
3214 measure makes sure that additional data from 'state' is never
3215 mixed into 'md_local' in small portions; this heuristically
3216 further strengthens the PRNG.
3219 *) Fix crypto/bn/asm/mips3.s.
3222 *) When only the key is given to "enc", the IV is undefined. Print out
3223 an error message in this case.
3226 *) Handle special case when X509_NAME is empty in X509 printing routines.
3229 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
3230 positive and less than q.
3233 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
3234 used: it isn't thread safe and the add_lock_callback should handle
3236 [Paul Rose <Paul.Rose@bridge.com>]
3238 *) Verify that incoming data obeys the block size in
3239 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
3243 [Ulf Möller, Bodo Möller]
3245 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
3246 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
3247 when fixing the server behaviour for backwards-compatible 'client
3248 hello' messages. (Note that the attack is impractical against
3249 SSL 3.0 and TLS 1.0 anyway because length and version checking
3250 means that the probability of guessing a valid ciphertext is
3251 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
3254 Before 0.9.5, the countermeasure (hide the error by generating a
3255 random 'decryption result') did not work properly because
3256 ERR_clear_error() was missing, meaning that SSL_get_error() would
3257 detect the supposedly ignored error.
3259 Both problems are now fixed.
3262 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
3263 (previously it was 1024).
3266 *) Fix for compatibility mode trust settings: ignore trust settings
3267 unless some valid trust or reject settings are present.
3270 *) Fix for blowfish EVP: its a variable length cipher.
3273 *) Fix various bugs related to DSA S/MIME verification. Handle missing
3274 parameters in DSA public key structures and return an error in the
3275 DSA routines if parameters are absent.
3278 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
3279 in the current directory if neither $RANDFILE nor $HOME was set.
3280 RAND_file_name() in 0.9.6a returned NULL in this case. This has
3281 caused some confusion to Windows users who haven't defined $HOME.
3282 Thus RAND_file_name() is changed again: e_os.h can define a
3283 DEFAULT_HOME, which will be used if $HOME is not set.
3284 For Windows, we use "C:"; on other platforms, we still require
3285 environment variables.
3287 *) Move 'if (!initialized) RAND_poll()' into regions protected by
3288 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
3289 having multiple threads call RAND_poll() concurrently.
3292 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
3293 combination of a flag and a thread ID variable.
3294 Otherwise while one thread is in ssleay_rand_bytes (which sets the
3295 flag), *other* threads can enter ssleay_add_bytes without obeying
3296 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
3297 that they do not hold after the first thread unsets add_do_not_lock).
3300 *) Change bctest again: '-x' expressions are not available in all
3304 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
3306 *) Fix a couple of memory leaks in PKCS7_dataDecode()
3307 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
3309 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
3310 the default extension for executables, if any. Also, make the perl
3311 scripts that use symlink() to test if it really exists and use "cp"
3312 if it doesn't. All this made OpenSSL compilable and installable in
3316 *) Fix for asn1_GetSequence() for indefinite length constructed data.
3317 If SEQUENCE is length is indefinite just set c->slen to the total
3318 amount of data available.
3319 [Steve Henson, reported by shige@FreeBSD.org]
3320 [This change does not apply to 0.9.7.]
3322 *) Change bctest to avoid here-documents inside command substitution
3323 (workaround for FreeBSD /bin/sh bug).
3324 For compatibility with Ultrix, avoid shell functions (introduced
3325 in the bctest version that searches along $PATH).
3328 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
3329 with des_encrypt() defined on some operating systems, like Solaris
3333 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
3334 On the Importance of Eliminating Errors in Cryptographic
3335 Computations, J. Cryptology 14 (2001) 2, 101-119,
3336 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
3339 *) MIPS assembler BIGNUM division bug fix.
3342 *) Disabled incorrect Alpha assembler code.
3345 *) Fix PKCS#7 decode routines so they correctly update the length
3346 after reading an EOC for the EXPLICIT tag.
3348 [This change does not apply to 0.9.7.]
3350 *) Fix bug in PKCS#12 key generation routines. This was triggered
3351 if a 3DES key was generated with a 0 initial byte. Include
3352 PKCS12_BROKEN_KEYGEN compilation option to retain the old
3353 (but broken) behaviour.
3356 *) Enhance bctest to search for a working bc along $PATH and print
3358 [Tim Rice <tim@multitalents.net> via Richard Levitte]
3360 *) Fix memory leaks in err.c: free err_data string if necessary;
3361 don't write to the wrong index in ERR_set_error_data.
3364 *) Implement ssl23_peek (analogous to ssl23_read), which previously
3368 *) Replace rdtsc with _emit statements for VC++ version 5.
3369 [Jeremy Cooper <jeremy@baymoo.org>]
3371 *) Make it possible to reuse SSLv2 sessions.
3374 *) In copy_email() check for >= 0 as a return value for
3375 X509_NAME_get_index_by_NID() since 0 is a valid index.
3376 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
3378 *) Avoid coredump with unsupported or invalid public keys by checking if
3379 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
3380 PKCS7_verify() fails with non detached data.
3383 *) Don't use getenv in library functions when run as setuid/setgid.
3384 New function OPENSSL_issetugid().
3387 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
3388 due to incorrect handling of multi-threading:
3390 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
3392 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
3394 3. Count how many times MemCheck_off() has been called so that
3395 nested use can be treated correctly. This also avoids
3396 inband-signalling in the previous code (which relied on the
3397 assumption that thread ID 0 is impossible).
3400 *) Add "-rand" option also to s_client and s_server.
3403 *) Fix CPU detection on Irix 6.x.
3404 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
3405 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3407 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
3410 [This change does not apply to 0.9.7.]
3412 *) Use the cached encoding of an X509_NAME structure rather than
3413 copying it. This is apparently the reason for the libsafe "errors"
3414 but the code is actually correct.
3417 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
3418 Bleichenbacher's DSA attack.
3419 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
3420 to be set and top=0 forces the highest bit to be set; top=-1 is new
3421 and leaves the highest bit random.
3422 [Ulf Moeller, Bodo Moeller]
3424 *) In the NCONF_...-based implementations for CONF_... queries
3425 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
3426 a temporary CONF structure with the data component set to NULL
3427 (which gives segmentation faults in lh_retrieve).
3428 Instead, use NULL for the CONF pointer in CONF_get_string and
3429 CONF_get_number (which may use environment variables) and directly
3430 return NULL from CONF_get_section.
3433 *) Fix potential buffer overrun for EBCDIC.
3436 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
3437 keyUsage if basicConstraints absent for a CA.
3440 *) Make SMIME_write_PKCS7() write mail header values with a format that
3441 is more generally accepted (no spaces before the semicolon), since
3442 some programs can't parse those values properly otherwise. Also make
3443 sure BIO's that break lines after each write do not create invalid
3447 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
3448 macros previously used would not encode an empty SEQUENCE OF
3449 and break the signature.
3451 [This change does not apply to 0.9.7.]
3453 *) Zero the premaster secret after deriving the master secret in
3457 *) Add some EVP_add_digest_alias registrations (as found in
3458 OpenSSL_add_all_digests()) to SSL_library_init()
3459 aka OpenSSL_add_ssl_algorithms(). This provides improved
3460 compatibility with peers using X.509 certificates
3461 with unconventional AlgorithmIdentifier OIDs.
3464 *) Fix for Irix with NO_ASM.
3465 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3467 *) ./config script fixes.
3468 [Ulf Moeller, Richard Levitte]
3470 *) Fix 'openssl passwd -1'.
3473 *) Change PKCS12_key_gen_asc() so it can cope with non null
3474 terminated strings whose length is passed in the passlen
3475 parameter, for example from PEM callbacks. This was done
3476 by adding an extra length parameter to asc2uni().
3477 [Steve Henson, reported by <oddissey@samsung.co.kr>]
3479 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
3480 call failed, free the DSA structure.
3483 *) Fix to uni2asc() to cope with zero length Unicode strings.
3484 These are present in some PKCS#12 files.
3487 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
3488 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
3489 when writing a 32767 byte record.
3490 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
3492 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
3493 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
3495 (RSA objects have a reference count access to which is protected
3496 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
3497 so they are meant to be shared between threads.)
3498 [Bodo Moeller, Geoff Thorpe; original patch submitted by
3499 "Reddie, Steven" <Steven.Reddie@ca.com>]
3501 *) Fix a deadlock in CRYPTO_mem_leaks().
3504 *) Use better test patterns in bntest.
3507 *) rand_win.c fix for Borland C.
3510 *) BN_rshift bugfix for n == 0.
3513 *) Add a 'bctest' script that checks for some known 'bc' bugs
3514 so that 'make test' does not abort just because 'bc' is broken.
3517 *) Store verify_result within SSL_SESSION also for client side to
3518 avoid potential security hole. (Re-used sessions on the client side
3519 always resulted in verify_result==X509_V_OK, not using the original
3520 result of the server certificate verification.)
3523 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
3524 SSL3_RT_APPLICATION_DATA, return 0.
3525 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
3529 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
3530 releases, have been re-implemented by renaming the previous
3531 implementations of ssl2_read and ssl3_read to ssl2_read_internal
3532 and ssl3_read_internal, respectively, and adding 'peek' parameters
3533 to them. The new ssl[23]_{read,peek} functions are calls to
3534 ssl[23]_read_internal with the 'peek' flag set appropriately.
3535 A 'peek' parameter has also been added to ssl3_read_bytes, which
3536 does the actual work for ssl3_read_internal.
3539 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
3540 the method-specific "init()" handler. Also clean up ex_data after
3541 calling the method-specific "finish()" handler. Previously, this was
3542 happening the other way round.
3545 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
3546 The previous value, 12, was not always sufficient for BN_mod_exp().
3549 *) Make sure that shared libraries get the internal name engine with
3550 the full version number and not just 0. This should mark the
3551 shared libraries as not backward compatible. Of course, this should
3552 be changed again when we can guarantee backward binary compatibility.
3555 *) Fix typo in get_cert_by_subject() in by_dir.c
3556 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
3558 *) Rework the system to generate shared libraries:
3560 - Make note of the expected extension for the shared libraries and
3561 if there is a need for symbolic links from for example libcrypto.so.0
3562 to libcrypto.so.0.9.7. There is extended info in Configure for
3565 - Make as few rebuilds of the shared libraries as possible.
3567 - Still avoid linking the OpenSSL programs with the shared libraries.
3569 - When installing, install the shared libraries separately from the
3573 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
3575 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
3576 and not in SSL_clear because the latter is also used by the
3577 accept/connect functions; previously, the settings made by
3578 SSL_set_read_ahead would be lost during the handshake.
3579 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
3581 *) Correct util/mkdef.pl to be selective about disabled algorithms.
3582 Previously, it would create entries for disableed algorithms no
3586 *) Added several new manual pages for SSL_* function.
3589 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
3591 *) In ssl23_get_client_hello, generate an error message when faced
3592 with an initial SSL 3.0/TLS record that is too small to contain the
3593 first two bytes of the ClientHello message, i.e. client_version.
3594 (Note that this is a pathologic case that probably has never happened
3595 in real life.) The previous approach was to use the version number
3596 from the record header as a substitute; but our protocol choice
3597 should not depend on that one because it is not authenticated
3598 by the Finished messages.
3601 *) More robust randomness gathering functions for Windows.
3602 [Jeffrey Altman <jaltman@columbia.edu>]
3604 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
3605 not set then we don't setup the error code for issuer check errors
3606 to avoid possibly overwriting other errors which the callback does
3607 handle. If an application does set the flag then we assume it knows
3608 what it is doing and can handle the new informational codes
3612 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
3613 a general "ANY" type, as such it should be able to decode anything
3614 including tagged types. However it didn't check the class so it would
3615 wrongly interpret tagged types in the same way as their universal
3616 counterpart and unknown types were just rejected. Changed so that the
3617 tagged and unknown types are handled in the same way as a SEQUENCE:
3618 that is the encoding is stored intact. There is also a new type
3619 "V_ASN1_OTHER" which is used when the class is not universal, in this
3620 case we have no idea what the actual type is so we just lump them all
3624 *) On VMS, stdout may very well lead to a file that is written to
3625 in a record-oriented fashion. That means that every write() will
3626 write a separate record, which will be read separately by the
3627 programs trying to read from it. This can be very confusing.
3629 The solution is to put a BIO filter in the way that will buffer
3630 text until a linefeed is reached, and then write everything a
3631 line at a time, so every record written will be an actual line,
3632 not chunks of lines and not (usually doesn't happen, but I've
3633 seen it once) several lines in one record. BIO_f_linebuffer() is
3636 Currently, it's a VMS-only method, because that's where it has
3637 been tested well enough.
3640 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
3641 it can return incorrect results.
3642 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
3643 but it was in 0.9.6-beta[12].)
3646 *) Disable the check for content being present when verifying detached
3647 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
3648 include zero length content when signing messages.
3651 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
3652 BIO_ctrl (for BIO pairs).
3655 *) Add DSO method for VMS.
3658 *) Bug fix: Montgomery multiplication could produce results with the
3662 *) Add RPM specification openssl.spec and modify it to build three
3663 packages. The default package contains applications, application
3664 documentation and run-time libraries. The devel package contains
3665 include files, static libraries and function documentation. The
3666 doc package contains the contents of the doc directory. The original
3667 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
3670 *) Add a large number of documentation files for many SSL routines.
3671 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3673 *) Add a configuration entry for Sony News 4.
3674 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
3676 *) Don't set the two most significant bits to one when generating a
3677 random number < q in the DSA library.
3680 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
3681 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
3682 the underlying transport is blocking) if a handshake took place.
3683 (The default behaviour is needed by applications such as s_client
3684 and s_server that use select() to determine when to use SSL_read;
3685 but for applications that know in advance when to expect data, it
3686 just makes things more complicated.)
3689 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
3693 *) Add a few more EBCDIC conditionals that make `req' and `x509'
3694 work better on such systems.
3695 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3697 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
3698 Update PKCS12_parse() so it copies the friendlyName and the
3699 keyid to the certificates aux info.
3702 *) Fix bug in PKCS7_verify() which caused an infinite loop
3703 if there was more than one signature.
3704 [Sven Uszpelkat <su@celocom.de>]
3706 *) Major change in util/mkdef.pl to include extra information
3707 about each symbol, as well as presentig variables as well
3708 as functions. This change means that there's n more need
3709 to rebuild the .num files when some algorithms are excluded.
3712 *) Allow the verify time to be set by an application,
3713 rather than always using the current time.
3716 *) Phase 2 verify code reorganisation. The certificate
3717 verify code now looks up an issuer certificate by a
3718 number of criteria: subject name, authority key id
3719 and key usage. It also verifies self signed certificates
3720 by the same criteria. The main comparison function is
3721 X509_check_issued() which performs these checks.
3723 Lot of changes were necessary in order to support this
3724 without completely rewriting the lookup code.
3726 Authority and subject key identifier are now cached.
3728 The LHASH 'certs' is X509_STORE has now been replaced
3729 by a STACK_OF(X509_OBJECT). This is mainly because an
3730 LHASH can't store or retrieve multiple objects with
3731 the same hash value.
3733 As a result various functions (which were all internal
3734 use only) have changed to handle the new X509_STORE
3735 structure. This will break anything that messed round
3736 with X509_STORE internally.
3738 The functions X509_STORE_add_cert() now checks for an
3739 exact match, rather than just subject name.
3741 The X509_STORE API doesn't directly support the retrieval
3742 of multiple certificates matching a given criteria, however
3743 this can be worked round by performing a lookup first
3744 (which will fill the cache with candidate certificates)
3745 and then examining the cache for matches. This is probably
3746 the best we can do without throwing out X509_LOOKUP
3747 entirely (maybe later...).
3749 The X509_VERIFY_CTX structure has been enhanced considerably.
3751 All certificate lookup operations now go via a get_issuer()
3752 callback. Although this currently uses an X509_STORE it
3753 can be replaced by custom lookups. This is a simple way
3754 to bypass the X509_STORE hackery necessary to make this
3755 work and makes it possible to use more efficient techniques
3756 in future. A very simple version which uses a simple
3757 STACK for its trusted certificate store is also provided
3758 using X509_STORE_CTX_trusted_stack().
3760 The verify_cb() and verify() callbacks now have equivalents
3761 in the X509_STORE_CTX structure.
3763 X509_STORE_CTX also has a 'flags' field which can be used
3764 to customise the verify behaviour.
3767 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
3768 excludes S/MIME capabilities.
3771 *) When a certificate request is read in keep a copy of the
3772 original encoding of the signed data and use it when outputing
3773 again. Signatures then use the original encoding rather than
3774 a decoded, encoded version which may cause problems if the
3775 request is improperly encoded.
3778 *) For consistency with other BIO_puts implementations, call
3779 buffer_write(b, ...) directly in buffer_puts instead of calling
3782 In BIO_puts, increment b->num_write as in BIO_write.
3783 [Peter.Sylvester@EdelWeb.fr]
3785 *) Fix BN_mul_word for the case where the word is 0. (We have to use
3786 BN_zero, we may not return a BIGNUM with an array consisting of
3790 *) Avoid calling abort() from within the library when problems are
3791 detected, except if preprocessor symbols have been defined
3792 (such as REF_CHECK, BN_DEBUG etc.).
3795 *) New openssl application 'rsautl'. This utility can be
3796 used for low level RSA operations. DER public key
3797 BIO/fp routines also added.
3800 *) New Configure entry and patches for compiling on QNX 4.
3801 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
3803 *) A demo state-machine implementation was sponsored by
3804 Nuron (http://www.nuron.com/) and is now available in
3805 demos/state_machine.
3808 *) New options added to the 'dgst' utility for signature
3809 generation and verification.
3812 *) Unrecognized PKCS#7 content types are now handled via a
3813 catch all ASN1_TYPE structure. This allows unsupported
3814 types to be stored as a "blob" and an application can
3815 encode and decode it manually.
3818 *) Fix various signed/unsigned issues to make a_strex.c
3820 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
3822 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
3823 length if passed a buffer. ASN1_INTEGER_to_BN failed
3824 if passed a NULL BN and its argument was negative.
3825 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
3827 *) Modification to PKCS#7 encoding routines to output definite
3828 length encoding. Since currently the whole structures are in
3829 memory there's not real point in using indefinite length
3830 constructed encoding. However if OpenSSL is compiled with
3831 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
3834 *) Added BIO_vprintf() and BIO_vsnprintf().
3837 *) Added more prefixes to parse for in the the strings written
3838 through a logging bio, to cover all the levels that are available
3839 through syslog. The prefixes are now:
3841 PANIC, EMERG, EMR => LOG_EMERG
3842 ALERT, ALR => LOG_ALERT
3843 CRIT, CRI => LOG_CRIT
3844 ERROR, ERR => LOG_ERR
3845 WARNING, WARN, WAR => LOG_WARNING
3846 NOTICE, NOTE, NOT => LOG_NOTICE
3847 INFO, INF => LOG_INFO
3848 DEBUG, DBG => LOG_DEBUG
3850 and as before, if none of those prefixes are present at the
3851 beginning of the string, LOG_ERR is chosen.
3853 On Win32, the LOG_* levels are mapped according to this:
3855 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
3856 LOG_WARNING => EVENTLOG_WARNING_TYPE
3857 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
3861 *) Made it possible to reconfigure with just the configuration
3862 argument "reconf" or "reconfigure". The command line arguments
3863 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
3864 and are retrieved from there when reconfiguring.
3868 [Assar Westerlund <assar@sics.se>, Richard Levitte]
3870 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
3873 *) The obj_dat.pl script was messing up the sorting of object
3874 names. The reason was that it compared the quoted version
3875 of strings as a result "OCSP" > "OCSP Signing" because
3876 " > SPACE. Changed script to store unquoted versions of
3877 names and add quotes on output. It was also omitting some
3878 names from the lookup table if they were given a default
3879 value (that is if SN is missing it is given the same
3880 value as LN and vice versa), these are now added on the
3881 grounds that if an object has a name we should be able to
3882 look it up. Finally added warning output when duplicate
3883 short or long names are found.
3886 *) Changes needed for Tandem NSK.
3887 [Scott Uroff <scott@xypro.com>]
3889 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
3890 RSA_padding_check_SSLv23(), special padding was never detected
3891 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
3892 version rollback attacks was not effective.
3894 In s23_clnt.c, don't use special rollback-attack detection padding
3895 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
3896 client; similarly, in s23_srvr.c, don't do the rollback check if
3897 SSL 2.0 is the only protocol enabled in the server.
3900 *) Make it possible to get hexdumps of unprintable data with 'openssl
3901 asn1parse'. By implication, the functions ASN1_parse_dump() and
3902 BIO_dump_indent() are added.
3905 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
3906 these print out strings and name structures based on various
3907 flags including RFC2253 support and proper handling of
3908 multibyte characters. Added options to the 'x509' utility
3909 to allow the various flags to be set.
3912 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
3913 Also change the functions X509_cmp_current_time() and
3914 X509_gmtime_adj() work with an ASN1_TIME structure,
3915 this will enable certificates using GeneralizedTime in validity
3916 dates to be checked.
3919 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
3920 negative public key encodings) on by default,
3921 NO_NEG_PUBKEY_BUG can be set to disable it.
3924 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
3925 content octets. An i2c_ASN1_OBJECT is unnecessary because
3926 the encoding can be trivially obtained from the structure.
3929 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
3930 not read locks (CRYPTO_r_[un]lock).
3933 *) A first attempt at creating official support for shared
3934 libraries through configuration. I've kept it so the
3935 default is static libraries only, and the OpenSSL programs
3936 are always statically linked for now, but there are
3937 preparations for dynamic linking in place.
3938 This has been tested on Linux and Tru64.
3941 *) Randomness polling function for Win9x, as described in:
3942 Peter Gutmann, Software Generation of Practically Strong
3946 *) Fix so PRNG is seeded in req if using an already existing
3950 *) New options to smime application. -inform and -outform
3951 allow alternative formats for the S/MIME message including
3952 PEM and DER. The -content option allows the content to be
3953 specified separately. This should allow things like Netscape
3954 form signing output easier to verify.
3957 *) Fix the ASN1 encoding of tags using the 'long form'.
3960 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
3961 STRING types. These convert content octets to and from the
3962 underlying type. The actual tag and length octets are
3963 already assumed to have been read in and checked. These
3964 are needed because all other string types have virtually
3965 identical handling apart from the tag. By having versions
3966 of the ASN1 functions that just operate on content octets
3967 IMPLICIT tagging can be handled properly. It also allows
3968 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
3969 and ASN1_INTEGER are identical apart from the tag.
3972 *) Change the handling of OID objects as follows:
3974 - New object identifiers are inserted in objects.txt, following
3975 the syntax given in objects.README.
3976 - objects.pl is used to process obj_mac.num and create a new
3978 - obj_dat.pl is used to create a new obj_dat.h, using the data in
3981 This is currently kind of a hack, and the perl code in objects.pl
3982 isn't very elegant, but it works as I intended. The simplest way
3983 to check that it worked correctly is to look in obj_dat.h and
3984 check the array nid_objs and make sure the objects haven't moved
3985 around (this is important!). Additions are OK, as well as
3986 consistent name changes.
3989 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
3992 *) Addition of the command line parameter '-rand file' to 'openssl req'.
3993 The given file adds to whatever has already been seeded into the
3994 random pool through the RANDFILE configuration file option or
3995 environment variable, or the default random state file.
3998 *) mkstack.pl now sorts each macro group into lexical order.
3999 Previously the output order depended on the order the files
4000 appeared in the directory, resulting in needless rewriting
4004 *) Patches to make OpenSSL compile under Win32 again. Mostly
4005 work arounds for the VC++ problem that it treats func() as
4006 func(void). Also stripped out the parts of mkdef.pl that
4007 added extra typesafe functions: these no longer exist.
4010 *) Reorganisation of the stack code. The macros are now all
4011 collected in safestack.h . Each macro is defined in terms of
4012 a "stack macro" of the form SKM_<name>(type, a, b). The
4013 DEBUG_SAFESTACK is now handled in terms of function casts,
4014 this has the advantage of retaining type safety without the
4015 use of additional functions. If DEBUG_SAFESTACK is not defined
4016 then the non typesafe macros are used instead. Also modified the
4017 mkstack.pl script to handle the new form. Needs testing to see
4018 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
4019 the default if no major problems. Similar behaviour for ASN1_SET_OF
4020 and PKCS12_STACK_OF.
4023 *) When some versions of IIS use the 'NET' form of private key the
4024 key derivation algorithm is different. Normally MD5(password) is
4025 used as a 128 bit RC4 key. In the modified case
4026 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
4027 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
4028 as the old Netscape_RSA functions except they have an additional
4029 'sgckey' parameter which uses the modified algorithm. Also added
4030 an -sgckey command line option to the rsa utility. Thanks to
4031 Adrian Peck <bertie@ncipher.com> for posting details of the modified
4032 algorithm to openssl-dev.
4035 *) The evp_local.h macros were using 'c.##kname' which resulted in
4036 invalid expansion on some systems (SCO 5.0.5 for example).
4037 Corrected to 'c.kname'.
4038 [Phillip Porch <root@theporch.com>]
4040 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
4041 a STACK of email addresses from a certificate or request, these look
4042 in the subject name and the subject alternative name extensions and
4043 omit any duplicate addresses.
4046 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
4047 This makes DSA verification about 2 % faster.
4050 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
4051 (meaning that now 2^5 values will be precomputed, which is only 4 KB
4052 plus overhead for 1024 bit moduli).
4053 This makes exponentiations about 0.5 % faster for 1024 bit
4054 exponents (as measured by "openssl speed rsa2048").
4057 *) Rename memory handling macros to avoid conflicts with other
4059 Malloc => OPENSSL_malloc
4060 Malloc_locked => OPENSSL_malloc_locked
4061 Realloc => OPENSSL_realloc
4062 Free => OPENSSL_free
4065 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
4066 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
4069 *) CygWin32 support.
4070 [John Jarvie <jjarvie@newsguy.com>]
4072 *) The type-safe stack code has been rejigged. It is now only compiled
4073 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
4074 by default all type-specific stack functions are "#define"d back to
4075 standard stack functions. This results in more streamlined output
4076 but retains the type-safety checking possibilities of the original
4080 *) The STACK code has been cleaned up, and certain type declarations
4081 that didn't make a lot of sense have been brought in line. This has
4082 also involved a cleanup of sorts in safestack.h to more correctly
4083 map type-safe stack functions onto their plain stack counterparts.
4084 This work has also resulted in a variety of "const"ifications of
4085 lots of the code, especially "_cmp" operations which should normally
4086 be prototyped with "const" parameters anyway.
4089 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
4090 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
4091 (The PRNG state consists of two parts, the large pool 'state' and 'md',
4092 where all of 'md' is used each time the PRNG is used, but 'state'
4093 is used only indexed by a cyclic counter. As entropy may not be
4094 well distributed from the beginning, 'md' is important as a
4095 chaining variable. However, the output function chains only half
4096 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
4097 all of 'md', and seeding with STATE_SIZE dummy bytes will result
4098 in all of 'state' being rewritten, with the new values depending
4099 on virtually all of 'md'. This overcomes the 80 bit limitation.)
4102 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
4103 the handshake is continued after ssl_verify_cert_chain();
4104 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
4105 can lead to 'unexplainable' connection aborts later.
4106 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
4108 *) Major EVP API cipher revision.
4109 Add hooks for extra EVP features. This allows various cipher
4110 parameters to be set in the EVP interface. Support added for variable
4111 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
4112 setting of RC2 and RC5 parameters.
4114 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
4117 Remove lots of duplicated code from the EVP library. For example *every*
4118 cipher init() function handles the 'iv' in the same way according to the
4119 cipher mode. They also all do nothing if the 'key' parameter is NULL and
4120 for CFB and OFB modes they zero ctx->num.
4122 New functionality allows removal of S/MIME code RC2 hack.
4124 Most of the routines have the same form and so can be declared in terms
4127 By shifting this to the top level EVP_CipherInit() it can be removed from
4128 all individual ciphers. If the cipher wants to handle IVs or keys
4129 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
4132 Change lots of functions like EVP_EncryptUpdate() to now return a
4133 value: although software versions of the algorithms cannot fail
4134 any installed hardware versions can.
4137 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
4138 this option is set, tolerate broken clients that send the negotiated
4139 protocol version number instead of the requested protocol version
4143 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
4144 i.e. non-zero for export ciphersuites, zero otherwise.
4145 Previous versions had this flag inverted, inconsistent with
4146 rsa_tmp_cb (..._TMP_RSA_CB).
4147 [Bodo Moeller; problem reported by Amit Chopra]
4149 *) Add missing DSA library text string. Work around for some IIS
4150 key files with invalid SEQUENCE encoding.
4153 *) Add a document (doc/standards.txt) that list all kinds of standards
4154 and so on that are implemented in OpenSSL.
4157 *) Enhance c_rehash script. Old version would mishandle certificates
4158 with the same subject name hash and wouldn't handle CRLs at all.
4159 Added -fingerprint option to crl utility, to support new c_rehash
4163 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
4166 *) Fix for SSL server purpose checking. Server checking was
4167 rejecting certificates which had extended key usage present
4168 but no ssl client purpose.
4169 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
4171 *) Make PKCS#12 code work with no password. The PKCS#12 spec
4172 is a little unclear about how a blank password is handled.
4173 Since the password in encoded as a BMPString with terminating
4174 double NULL a zero length password would end up as just the
4175 double NULL. However no password at all is different and is
4176 handled differently in the PKCS#12 key generation code. NS
4177 treats a blank password as zero length. MSIE treats it as no
4178 password on export: but it will try both on import. We now do
4179 the same: PKCS12_parse() tries zero length and no password if
4180 the password is set to "" or NULL (NULL is now a valid password:
4181 it wasn't before) as does the pkcs12 application.
4184 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
4185 perror when PEM_read_bio_X509_REQ fails, the error message must
4186 be obtained from the error queue.
4189 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
4190 it in ERR_remove_state if appropriate, and change ERR_get_state
4191 accordingly to avoid race conditions (this is necessary because
4192 thread_hash is no longer constant once set).
4195 *) Bugfix for linux-elf makefile.one.
4198 *) RSA_get_default_method() will now cause a default
4199 RSA_METHOD to be chosen if one doesn't exist already.
4200 Previously this was only set during a call to RSA_new()
4201 or RSA_new_method(NULL) meaning it was possible for
4202 RSA_get_default_method() to return NULL.
4205 *) Added native name translation to the existing DSO code
4206 that will convert (if the flag to do so is set) filenames
4207 that are sufficiently small and have no path information
4208 into a canonical native form. Eg. "blah" converted to
4209 "libblah.so" or "blah.dll" etc.
4212 *) New function ERR_error_string_n(e, buf, len) which is like
4213 ERR_error_string(e, buf), but writes at most 'len' bytes
4214 including the 0 terminator. For ERR_error_string_n, 'buf'
4216 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
4218 *) CONF library reworked to become more general. A new CONF
4219 configuration file reader "class" is implemented as well as a
4220 new functions (NCONF_*, for "New CONF") to handle it. The now
4221 old CONF_* functions are still there, but are reimplemented to
4222 work in terms of the new functions. Also, a set of functions
4223 to handle the internal storage of the configuration data is
4224 provided to make it easier to write new configuration file
4225 reader "classes" (I can definitely see something reading a
4226 configuration file in XML format, for example), called _CONF_*,
4227 or "the configuration storage API"...
4229 The new configuration file reading functions are:
4231 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
4232 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
4234 NCONF_default, NCONF_WIN32
4236 NCONF_dump_fp, NCONF_dump_bio
4238 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
4239 NCONF_new creates a new CONF object. This works in the same way
4240 as other interfaces in OpenSSL, like the BIO interface.
4241 NCONF_dump_* dump the internal storage of the configuration file,
4242 which is useful for debugging. All other functions take the same
4243 arguments as the old CONF_* functions wth the exception of the
4244 first that must be a `CONF *' instead of a `LHASH *'.
4246 To make it easer to use the new classes with the old CONF_* functions,
4247 the function CONF_set_default_method is provided.
4250 *) Add '-tls1' option to 'openssl ciphers', which was already
4251 mentioned in the documentation but had not been implemented.
4252 (This option is not yet really useful because even the additional
4253 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
4256 *) Initial DSO code added into libcrypto for letting OpenSSL (and
4257 OpenSSL-based applications) load shared libraries and bind to
4258 them in a portable way.
4259 [Geoff Thorpe, with contributions from Richard Levitte]
4261 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
4263 *) Make sure _lrotl and _lrotr are only used with MSVC.
4265 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
4266 (the default implementation of RAND_status).
4268 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
4269 to '-clrext' (= clear extensions), as intended and documented.
4270 [Bodo Moeller; inconsistency pointed out by Michael Attili
4273 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
4274 was larger than the MD block size.
4275 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
4277 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
4278 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
4279 using the passed key: if the passed key was a private key the result
4280 of X509_print(), for example, would be to print out all the private key
4284 *) des_quad_cksum() byte order bug fix.
4285 [Ulf Möller, using the problem description in krb4-0.9.7, where
4286 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
4288 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
4290 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
4292 *) For easily testing in shell scripts whether some command
4293 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
4294 returns with exit code 0 iff no command of the given name is available.
4295 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
4296 the output goes to stdout and nothing is printed to stderr.
4297 Additional arguments are always ignored.
4299 Since for each cipher there is a command of the same name,
4300 the 'no-cipher' compilation switches can be tested this way.
4302 ('openssl no-XXX' is not able to detect pseudo-commands such
4303 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
4306 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
4309 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
4310 is set; it will be thrown away anyway because each handshake creates
4312 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
4313 to parameters -- in previous versions (since OpenSSL 0.9.3) the
4314 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
4315 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
4318 *) New s_client option -ign_eof: EOF at stdin is ignored, and
4319 'Q' and 'R' lose their special meanings (quit/renegotiate).
4320 This is part of what -quiet does; unlike -quiet, -ign_eof
4321 does not suppress any output.
4324 *) Add compatibility options to the purpose and trust code. The
4325 purpose X509_PURPOSE_ANY is "any purpose" which automatically
4326 accepts a certificate or CA, this was the previous behaviour,
4327 with all the associated security issues.
4329 X509_TRUST_COMPAT is the old trust behaviour: only and
4330 automatically trust self signed roots in certificate store. A
4331 new trust setting X509_TRUST_DEFAULT is used to specify that
4332 a purpose has no associated trust setting and it should instead
4333 use the value in the default purpose.
4336 *) Fix the PKCS#8 DSA private key code so it decodes keys again
4337 and fix a memory leak.
4340 *) In util/mkerr.pl (which implements 'make errors'), preserve
4341 reason strings from the previous version of the .c file, as
4342 the default to have only downcase letters (and digits) in
4343 automatically generated reasons codes is not always appropriate.
4346 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
4347 using strerror. Previously, ERR_reason_error_string() returned
4348 library names as reason strings for SYSerr; but SYSerr is a special
4349 case where small numbers are errno values, not library numbers.
4352 *) Add '-dsaparam' option to 'openssl dhparam' application. This
4353 converts DSA parameters into DH parameters. (When creating parameters,
4354 DSA_generate_parameters is used.)
4357 *) Include 'length' (recommended exponent length) in C code generated
4358 by 'openssl dhparam -C'.
4361 *) The second argument to set_label in perlasm was already being used
4362 so couldn't be used as a "file scope" flag. Moved to third argument
4366 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
4367 instead of RAND_bytes for encryption IVs and salts.
4370 *) Include RAND_status() into RAND_METHOD instead of implementing
4371 it only for md_rand.c Otherwise replacing the PRNG by calling
4372 RAND_set_rand_method would be impossible.
4375 *) Don't let DSA_generate_key() enter an infinite loop if the random
4376 number generation fails.
4379 *) New 'rand' application for creating pseudo-random output.
4382 *) Added configuration support for Linux/IA64
4383 [Rolf Haberrecker <rolf@suse.de>]
4385 *) Assembler module support for Mingw32.
4388 *) Shared library support for HPUX (in shlib/).
4389 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
4391 *) Shared library support for Solaris gcc.
4392 [Lutz Behnke <behnke@trustcenter.de>]
4394 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
4396 *) PKCS7_encrypt() was adding text MIME headers twice because they
4397 were added manually and by SMIME_crlf_copy().
4400 *) In bntest.c don't call BN_rand with zero bits argument.
4401 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
4403 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
4404 case was implemented. This caused BN_div_recp() to fail occasionally.
4407 *) Add an optional second argument to the set_label() in the perl
4408 assembly language builder. If this argument exists and is set
4409 to 1 it signals that the assembler should use a symbol whose
4410 scope is the entire file, not just the current function. This
4411 is needed with MASM which uses the format label:: for this scope.
4412 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
4414 *) Change the ASN1 types so they are typedefs by default. Before
4415 almost all types were #define'd to ASN1_STRING which was causing
4416 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
4420 *) Change names of new functions to the new get1/get0 naming
4421 convention: After 'get1', the caller owns a reference count
4422 and has to call ..._free; 'get0' returns a pointer to some
4423 data structure without incrementing reference counters.
4424 (Some of the existing 'get' functions increment a reference
4425 counter, some don't.)
4426 Similarly, 'set1' and 'add1' functions increase reference
4427 counters or duplicate objects.
4430 *) Allow for the possibility of temp RSA key generation failure:
4431 the code used to assume it always worked and crashed on failure.
4434 *) Fix potential buffer overrun problem in BIO_printf().
4435 [Ulf Möller, using public domain code by Patrick Powell; problem
4436 pointed out by David Sacerdote <das33@cornell.edu>]
4438 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
4439 RAND_egd() and RAND_status(). In the command line application,
4440 the EGD socket can be specified like a seed file using RANDFILE
4444 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
4445 Some CAs (e.g. Verisign) distribute certificates in this form.
4448 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
4449 list to exclude them. This means that no special compilation option
4450 is needed to use anonymous DH: it just needs to be included in the
4454 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
4455 EVP_MD_type. The old functionality is available in a new macro called
4456 EVP_MD_md(). Change code that uses it and update docs.
4459 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
4460 where the 'void *' argument is replaced by a function pointer argument.
4461 Previously 'void *' was abused to point to functions, which works on
4462 many platforms, but is not correct. As these functions are usually
4463 called by macros defined in OpenSSL header files, most source code
4464 should work without changes.
4467 *) <openssl/opensslconf.h> (which is created by Configure) now contains
4468 sections with information on -D... compiler switches used for
4469 compiling the library so that applications can see them. To enable
4470 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
4471 must be defined. E.g.,
4472 #define OPENSSL_ALGORITHM_DEFINES
4473 #include <openssl/opensslconf.h>
4474 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
4475 [Richard Levitte, Ulf and Bodo Möller]
4477 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
4481 *) Change the 'other' type in certificate aux info to a STACK_OF
4482 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
4483 the required ASN1 format: arbitrary types determined by an OID.
4486 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
4487 argument to 'req'. This is not because the function is newer or
4488 better than others it just uses the work 'NEW' in the certificate
4489 request header lines. Some software needs this.
4492 *) Reorganise password command line arguments: now passwords can be
4493 obtained from various sources. Delete the PEM_cb function and make
4494 it the default behaviour: i.e. if the callback is NULL and the
4495 usrdata argument is not NULL interpret it as a null terminated pass
4496 phrase. If usrdata and the callback are NULL then the pass phrase
4497 is prompted for as usual.
4500 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
4501 the support is automatically enabled. The resulting binaries will
4502 autodetect the card and use it if present.
4503 [Ben Laurie and Compaq Inc.]
4505 *) Work around for Netscape hang bug. This sends certificate request
4506 and server done in one record. Since this is perfectly legal in the
4507 SSL/TLS protocol it isn't a "bug" option and is on by default. See
4508 the bugs/SSLv3 entry for more info.
4511 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
4514 *) Add -rand argument to smime and pkcs12 applications and read/write
4518 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
4521 *) Add command line password options to the remaining applications.
4524 *) Bug fix for BN_div_recp() for numerators with an even number of
4528 *) More tests in bntest.c, and changed test_bn output.
4531 *) ./config recognizes MacOS X now.
4534 *) Bug fix for BN_div() when the first words of num and divsor are
4535 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
4538 *) Add support for various broken PKCS#8 formats, and command line
4539 options to produce them.
4542 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
4543 get temporary BIGNUMs from a BN_CTX.
4546 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
4550 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
4551 include a #define from the old name to the new. The original intent
4552 was that statically linked binaries could for example just call
4553 SSLeay_add_all_ciphers() to just add ciphers to the table and not
4554 link with digests. This never worked becayse SSLeay_add_all_digests()
4555 and SSLeay_add_all_ciphers() were in the same source file so calling
4556 one would link with the other. They are now in separate source files.
4559 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
4562 *) Use a less unusual form of the Miller-Rabin primality test (it used
4563 a binary algorithm for exponentiation integrated into the Miller-Rabin
4564 loop, our standard modexp algorithms are faster).
4567 *) Support for the EBCDIC character set completed.
4568 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
4570 *) Source code cleanups: use const where appropriate, eliminate casts,
4571 use void * instead of char * in lhash.
4574 *) Bugfix: ssl3_send_server_key_exchange was not restartable
4575 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
4576 this the server could overwrite ephemeral keys that the client
4580 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
4581 using 50 iterations of the Rabin-Miller test.
4583 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
4584 iterations of the Rabin-Miller test as required by the appendix
4585 to FIPS PUB 186[-1]) instead of DSA_is_prime.
4586 As BN_is_prime_fasttest includes trial division, DSA parameter
4587 generation becomes much faster.
4589 This implies a change for the callback functions in DSA_is_prime
4590 and DSA_generate_parameters: The callback function is called once
4591 for each positive witness in the Rabin-Miller test, not just
4592 occasionally in the inner loop; and the parameters to the
4593 callback function now provide an iteration count for the outer
4594 loop rather than for the current invocation of the inner loop.
4595 DSA_generate_parameters additionally can call the callback
4596 function with an 'iteration count' of -1, meaning that a
4597 candidate has passed the trial division test (when q is generated
4598 from an application-provided seed, trial division is skipped).
4601 *) New function BN_is_prime_fasttest that optionally does trial
4602 division before starting the Rabin-Miller test and has
4603 an additional BN_CTX * argument (whereas BN_is_prime always
4604 has to allocate at least one BN_CTX).
4605 'callback(1, -1, cb_arg)' is called when a number has passed the
4606 trial division stage.
4609 *) Fix for bug in CRL encoding. The validity dates weren't being handled
4613 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
4616 *) New function BN_pseudo_rand().
4619 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
4620 bignum version of BN_from_montgomery() with the working code from
4621 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
4625 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
4626 made it impossible to use the same SSL_SESSION data structure in
4627 SSL2 clients in multiple threads.
4630 *) The return value of RAND_load_file() no longer counts bytes obtained
4631 by stat(). RAND_load_file(..., -1) is new and uses the complete file
4632 to seed the PRNG (previously an explicit byte count was required).
4633 [Ulf Möller, Bodo Möller]
4635 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
4636 used (char *) instead of (void *) and had casts all over the place.
4639 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
4642 *) Retain source code compatibility for BN_prime_checks macro:
4643 BN_is_prime(..., BN_prime_checks, ...) now uses
4644 BN_prime_checks_for_size to determine the appropriate number of
4645 Rabin-Miller iterations.
4648 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
4649 DH_CHECK_P_NOT_SAFE_PRIME.
4650 (Check if this is true? OpenPGP calls them "strong".)
4653 *) Merge the functionality of "dh" and "gendh" programs into a new program
4654 "dhparam". The old programs are retained for now but will handle DH keys
4655 (instead of parameters) in future.
4658 *) Make the ciphers, s_server and s_client programs check the return values
4659 when a new cipher list is set.
4662 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
4663 ciphers. Before when the 56bit ciphers were enabled the sorting was
4666 The syntax for the cipher sorting has been extended to support sorting by
4667 cipher-strength (using the strength_bits hard coded in the tables).
4668 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
4670 Fix a bug in the cipher-command parser: when supplying a cipher command
4671 string with an "undefined" symbol (neither command nor alphanumeric
4672 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
4673 an error is flagged.
4675 Due to the strength-sorting extension, the code of the
4676 ssl_create_cipher_list() function was completely rearranged. I hope that
4677 the readability was also increased :-)
4678 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4680 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
4681 for the first serial number and places 2 in the serial number file. This
4682 avoids problems when the root CA is created with serial number zero and
4683 the first user certificate has the same issuer name and serial number
4687 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
4688 the new code. Add documentation for this stuff.
4691 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
4692 X509_*() to X509at_*() on the grounds that they don't handle X509
4693 structures and behave in an analagous way to the X509v3 functions:
4694 they shouldn't be called directly but wrapper functions should be used
4697 So we also now have some wrapper functions that call the X509at functions
4698 when passed certificate requests. (TO DO: similar things can be done with
4699 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
4700 things. Some of these need some d2i or i2d and print functionality
4701 because they handle more complex structures.)
4704 *) Add missing #ifndefs that caused missing symbols when building libssl
4705 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
4706 NO_RSA in ssl/s2*.c.
4707 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
4709 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
4710 has a return value which indicates the quality of the random data
4711 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
4712 error queue. New function RAND_pseudo_bytes() generates output that is
4713 guaranteed to be unique but not unpredictable. RAND_add is like
4714 RAND_seed, but takes an extra argument for an entropy estimate
4715 (RAND_seed always assumes full entropy).
4718 *) Do more iterations of Rabin-Miller probable prime test (specifically,
4719 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
4720 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
4721 in crypto/bn/bn_prime.c for the complete table). This guarantees a
4722 false-positive rate of at most 2^-80 for random input.
4725 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
4728 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
4729 in the 0.9.5 release), this returns the chain
4730 from an X509_CTX structure with a dup of the stack and all
4731 the X509 reference counts upped: so the stack will exist
4732 after X509_CTX_cleanup() has been called. Modify pkcs12.c
4735 Also make SSL_SESSION_print() print out the verify return
4739 *) Add manpage for the pkcs12 command. Also change the default
4740 behaviour so MAC iteration counts are used unless the new
4741 -nomaciter option is used. This improves file security and
4742 only older versions of MSIE (4.0 for example) need it.
4745 *) Honor the no-xxx Configure options when creating .DEF files.
4748 *) Add PKCS#10 attributes to field table: challengePassword,
4749 unstructuredName and unstructuredAddress. These are taken from
4750 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
4751 international characters are used.
4753 More changes to X509_ATTRIBUTE code: allow the setting of types
4754 based on strings. Remove the 'loc' parameter when adding
4755 attributes because these will be a SET OF encoding which is sorted
4759 *) Initial changes to the 'req' utility to allow request generation
4760 automation. This will allow an application to just generate a template
4761 file containing all the field values and have req construct the
4764 Initial support for X509_ATTRIBUTE handling. Stacks of these are
4765 used all over the place including certificate requests and PKCS#7
4766 structures. They are currently handled manually where necessary with
4767 some primitive wrappers for PKCS#7. The new functions behave in a
4768 manner analogous to the X509 extension functions: they allow
4769 attributes to be looked up by NID and added.
4771 Later something similar to the X509V3 code would be desirable to
4772 automatically handle the encoding, decoding and printing of the
4773 more complex types. The string types like challengePassword can
4774 be handled by the string table functions.
4776 Also modified the multi byte string table handling. Now there is
4777 a 'global mask' which masks out certain types. The table itself
4778 can use the flag STABLE_NO_MASK to ignore the mask setting: this
4779 is useful when for example there is only one permissible type
4780 (as in countryName) and using the mask might result in no valid
4784 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
4785 SSL_get_peer_finished to allow applications to obtain the latest
4786 Finished messages sent to the peer or expected from the peer,
4787 respectively. (SSL_get_peer_finished is usually the Finished message
4788 actually received from the peer, otherwise the protocol will be aborted.)
4790 As the Finished message are message digests of the complete handshake
4791 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
4792 be used for external authentication procedures when the authentication
4793 provided by SSL/TLS is not desired or is not enough.
4796 *) Enhanced support for Alpha Linux is added. Now ./config checks if
4797 the host supports BWX extension and if Compaq C is present on the
4798 $PATH. Just exploiting of the BWX extension results in 20-30%
4799 performance kick for some algorithms, e.g. DES and RC4 to mention
4800 a couple. Compaq C in turn generates ~20% faster code for MD5 and
4804 *) Add support for MS "fast SGC". This is arguably a violation of the
4805 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
4806 weak crypto and after checking the certificate is SGC a second one
4807 with strong crypto. MS SGC stops the first handshake after receiving
4808 the server certificate message and sends a second client hello. Since
4809 a server will typically do all the time consuming operations before
4810 expecting any further messages from the client (server key exchange
4811 is the most expensive) there is little difference between the two.
4813 To get OpenSSL to support MS SGC we have to permit a second client
4814 hello message after we have sent server done. In addition we have to
4815 reset the MAC if we do get this second client hello.
4818 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
4819 if a DER encoded private key is RSA or DSA traditional format. Changed
4820 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
4821 format DER encoded private key. Newer code should use PKCS#8 format which
4822 has the key type encoded in the ASN1 structure. Added DER private key
4823 support to pkcs8 application.
4826 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
4827 ciphersuites has been selected (as required by the SSL 3/TLS 1
4828 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
4829 is set, we interpret this as a request to violate the specification
4830 (the worst that can happen is a handshake failure, and 'correct'
4831 behaviour would result in a handshake failure anyway).
4834 *) In SSL_CTX_add_session, take into account that there might be multiple
4835 SSL_SESSION structures with the same session ID (e.g. when two threads
4836 concurrently obtain them from an external cache).
4837 The internal cache can handle only one SSL_SESSION with a given ID,
4838 so if there's a conflict, we now throw out the old one to achieve
4842 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
4843 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
4844 some routines that use cipher OIDs: some ciphers do not have OIDs
4845 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
4849 *) Simplify the trust setting structure and code. Now we just have
4850 two sequences of OIDs for trusted and rejected settings. These will
4851 typically have values the same as the extended key usage extension
4852 and any application specific purposes.
4854 The trust checking code now has a default behaviour: it will just
4855 check for an object with the same NID as the passed id. Functions can
4856 be provided to override either the default behaviour or the behaviour
4857 for a given id. SSL client, server and email already have functions
4858 in place for compatibility: they check the NID and also return "trusted"
4859 if the certificate is self signed.
4862 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
4863 traditional format into an EVP_PKEY structure.
4866 *) Add a password callback function PEM_cb() which either prompts for
4867 a password if usr_data is NULL or otherwise assumes it is a null
4868 terminated password. Allow passwords to be passed on command line
4869 environment or config files in a few more utilities.
4872 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
4873 keys. Add some short names for PKCS#8 PBE algorithms and allow them
4874 to be specified on the command line for the pkcs8 and pkcs12 utilities.
4875 Update documentation.
4878 *) Support for ASN1 "NULL" type. This could be handled before by using
4879 ASN1_TYPE but there wasn't any function that would try to read a NULL
4880 and produce an error if it couldn't. For compatibility we also have
4881 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
4882 don't allocate anything because they don't need to.
4885 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
4887 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
4889 *) Rebuild of the memory allocation routines used by OpenSSL code and
4890 possibly others as well. The purpose is to make an interface that
4891 provide hooks so anyone can build a separate set of allocation and
4892 deallocation routines to be used by OpenSSL, for example memory
4893 pool implementations, or something else, which was previously hard
4894 since Malloc(), Realloc() and Free() were defined as macros having
4895 the values malloc, realloc and free, respectively (except for Win32
4896 compilations). The same is provided for memory debugging code.
4897 OpenSSL already comes with functionality to find memory leaks, but
4898 this gives people a chance to debug other memory problems.
4900 With these changes, a new set of functions and macros have appeared:
4902 CRYPTO_set_mem_debug_functions() [F]
4903 CRYPTO_get_mem_debug_functions() [F]
4904 CRYPTO_dbg_set_options() [F]
4905 CRYPTO_dbg_get_options() [F]
4906 CRYPTO_malloc_debug_init() [M]
4908 The memory debug functions are NULL by default, unless the library
4909 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
4910 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
4911 gives the standard debugging functions that come with OpenSSL) or
4912 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
4913 provided by the library user) must be used. When the standard
4914 debugging functions are used, CRYPTO_dbg_set_options can be used to
4915 request additional information:
4916 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
4917 the CRYPTO_MDEBUG_xxx macro when compiling the library.
4919 Also, things like CRYPTO_set_mem_functions will always give the
4920 expected result (the new set of functions is used for allocation
4921 and deallocation) at all times, regardless of platform and compiler
4924 To finish it up, some functions that were never use in any other
4925 way than through macros have a new API and new semantic:
4928 CRYPTO_dbg_realloc()
4931 All macros of value have retained their old syntax.
4932 [Richard Levitte and Bodo Moeller]
4934 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
4935 ordering of SMIMECapabilities wasn't in "strength order" and there
4936 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
4940 *) Some ASN1 types with illegal zero length encoding (INTEGER,
4941 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
4942 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
4944 *) Merge in my S/MIME library for OpenSSL. This provides a simple
4945 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
4946 functionality to handle multipart/signed properly) and a utility
4947 called 'smime' to call all this stuff. This is based on code I
4948 originally wrote for Celo who have kindly allowed it to be
4949 included in OpenSSL.
4952 *) Add variants des_set_key_checked and des_set_key_unchecked of
4953 des_set_key (aka des_key_sched). Global variable des_check_key
4954 decides which of these is called by des_set_key; this way
4955 des_check_key behaves as it always did, but applications and
4956 the library itself, which was buggy for des_check_key == 1,
4957 have a cleaner way to pick the version they need.
4960 *) New function PKCS12_newpass() which changes the password of a
4964 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
4965 dynamic mix. In both cases the ids can be used as an index into the
4966 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
4967 functions so they accept a list of the field values and the
4968 application doesn't need to directly manipulate the X509_TRUST
4972 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
4976 *) Modify the way the V3 extension code looks up extensions. This now
4977 works in a similar way to the object code: we have some "standard"
4978 extensions in a static table which is searched with OBJ_bsearch()
4979 and the application can add dynamic ones if needed. The file
4980 crypto/x509v3/ext_dat.h now has the info: this file needs to be
4981 updated whenever a new extension is added to the core code and kept
4982 in ext_nid order. There is a simple program 'tabtest.c' which checks
4983 this. New extensions are not added too often so this file can readily
4984 be maintained manually.
4986 There are two big advantages in doing things this way. The extensions
4987 can be looked up immediately and no longer need to be "added" using
4988 X509V3_add_standard_extensions(): this function now does nothing.
4989 [Side note: I get *lots* of email saying the extension code doesn't
4990 work because people forget to call this function]
4991 Also no dynamic allocation is done unless new extensions are added:
4992 so if we don't add custom extensions there is no need to call
4993 X509V3_EXT_cleanup().
4996 *) Modify enc utility's salting as follows: make salting the default. Add a
4997 magic header, so unsalted files fail gracefully instead of just decrypting
4998 to garbage. This is because not salting is a big security hole, so people
4999 should be discouraged from doing it.
5002 *) Fixes and enhancements to the 'x509' utility. It allowed a message
5003 digest to be passed on the command line but it only used this
5004 parameter when signing a certificate. Modified so all relevant
5005 operations are affected by the digest parameter including the
5006 -fingerprint and -x509toreq options. Also -x509toreq choked if a
5007 DSA key was used because it didn't fix the digest.
5010 *) Initial certificate chain verify code. Currently tests the untrusted
5011 certificates for consistency with the verify purpose (which is set
5012 when the X509_STORE_CTX structure is set up) and checks the pathlength.
5014 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
5015 this is because it will reject chains with invalid extensions whereas
5016 every previous version of OpenSSL and SSLeay made no checks at all.
5018 Trust code: checks the root CA for the relevant trust settings. Trust
5019 settings have an initial value consistent with the verify purpose: e.g.
5020 if the verify purpose is for SSL client use it expects the CA to be
5021 trusted for SSL client use. However the default value can be changed to
5022 permit custom trust settings: one example of this would be to only trust
5023 certificates from a specific "secure" set of CAs.
5025 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
5026 which should be used for version portability: especially since the
5027 verify structure is likely to change more often now.
5029 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
5030 to set them. If not set then assume SSL clients will verify SSL servers
5033 Two new options to the verify program: -untrusted allows a set of
5034 untrusted certificates to be passed in and -purpose which sets the
5035 intended purpose of the certificate. If a purpose is set then the
5036 new chain verify code is used to check extension consistency.
5039 *) Support for the authority information access extension.
5042 *) Modify RSA and DSA PEM read routines to transparently handle
5043 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
5044 public keys in a format compatible with certificate
5045 SubjectPublicKeyInfo structures. Unfortunately there were already
5046 functions called *_PublicKey_* which used various odd formats so
5047 these are retained for compatibility: however the DSA variants were
5048 never in a public release so they have been deleted. Changed dsa/rsa
5049 utilities to handle the new format: note no releases ever handled public
5050 keys so we should be OK.
5052 The primary motivation for this change is to avoid the same fiasco
5053 that dogs private keys: there are several incompatible private key
5054 formats some of which are standard and some OpenSSL specific and
5055 require various evil hacks to allow partial transparent handling and
5056 even then it doesn't work with DER formats. Given the option anything
5057 other than PKCS#8 should be dumped: but the other formats have to
5058 stay in the name of compatibility.
5060 With public keys and the benefit of hindsight one standard format
5061 is used which works with EVP_PKEY, RSA or DSA structures: though
5062 it clearly returns an error if you try to read the wrong kind of key.
5064 Added a -pubkey option to the 'x509' utility to output the public key.
5065 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
5066 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
5067 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
5068 that do the same as the EVP_PKEY_assign_*() except they up the
5069 reference count of the added key (they don't "swallow" the
5073 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
5074 CRLs would fail if the file contained no certificates or no CRLs:
5075 added a new function to read in both types and return the number
5076 read: this means that if none are read it will be an error. The
5077 DER versions of the certificate and CRL reader would always fail
5078 because it isn't possible to mix certificates and CRLs in DER format
5079 without choking one or the other routine. Changed this to just read
5080 a certificate: this is the best we can do. Also modified the code
5081 in apps/verify.c to take notice of return codes: it was previously
5082 attempting to read in certificates from NULL pointers and ignoring
5083 any errors: this is one reason why the cert and CRL reader seemed
5084 to work. It doesn't check return codes from the default certificate
5085 routines: these may well fail if the certificates aren't installed.
5088 *) Code to support otherName option in GeneralName.
5091 *) First update to verify code. Change the verify utility
5092 so it warns if it is passed a self signed certificate:
5093 for consistency with the normal behaviour. X509_verify
5094 has been modified to it will now verify a self signed
5095 certificate if *exactly* the same certificate appears
5096 in the store: it was previously impossible to trust a
5097 single self signed certificate. This means that:
5098 openssl verify ss.pem
5099 now gives a warning about a self signed certificate but
5100 openssl verify -CAfile ss.pem ss.pem
5104 *) For servers, store verify_result in SSL_SESSION data structure
5105 (and add it to external session representation).
5106 This is needed when client certificate verifications fails,
5107 but an application-provided verification callback (set by
5108 SSL_CTX_set_cert_verify_callback) allows accepting the session
5109 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
5110 but returns 1): When the session is reused, we have to set
5111 ssl->verify_result to the appropriate error code to avoid
5113 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
5115 *) Fix a bug in the new PKCS#7 code: it didn't consider the
5116 case in PKCS7_dataInit() where the signed PKCS7 structure
5117 didn't contain any existing data because it was being created.
5118 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
5120 *) Add a salt to the key derivation routines in enc.c. This
5121 forms the first 8 bytes of the encrypted file. Also add a
5122 -S option to allow a salt to be input on the command line.
5125 *) New function X509_cmp(). Oddly enough there wasn't a function
5126 to compare two certificates. We do this by working out the SHA1
5127 hash and comparing that. X509_cmp() will be needed by the trust
5131 *) SSL_get1_session() is like SSL_get_session(), but increments
5132 the reference count in the SSL_SESSION returned.
5133 [Geoff Thorpe <geoff@eu.c2.net>]
5135 *) Fix for 'req': it was adding a null to request attributes.
5136 Also change the X509_LOOKUP and X509_INFO code to handle
5137 certificate auxiliary information.
5140 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
5144 *) Add the possibility to add extra information to the memory leak
5145 detecting output, to form tracebacks, showing from where each
5146 allocation was originated: CRYPTO_push_info("constant string") adds
5147 the string plus current file name and line number to a per-thread
5148 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
5149 is like calling CYRPTO_pop_info() until the stack is empty.
5150 Also updated memory leak detection code to be multi-thread-safe.
5153 *) Add options -text and -noout to pkcs7 utility and delete the
5154 encryption options which never did anything. Update docs.
5157 *) Add options to some of the utilities to allow the pass phrase
5158 to be included on either the command line (not recommended on
5159 OSes like Unix) or read from the environment. Update the
5160 manpages and fix a few bugs.
5163 *) Add a few manpages for some of the openssl commands.
5166 *) Fix the -revoke option in ca. It was freeing up memory twice,
5167 leaking and not finding already revoked certificates.
5170 *) Extensive changes to support certificate auxiliary information.
5171 This involves the use of X509_CERT_AUX structure and X509_AUX
5172 functions. An X509_AUX function such as PEM_read_X509_AUX()
5173 can still read in a certificate file in the usual way but it
5174 will also read in any additional "auxiliary information". By
5175 doing things this way a fair degree of compatibility can be
5176 retained: existing certificates can have this information added
5177 using the new 'x509' options.
5179 Current auxiliary information includes an "alias" and some trust
5180 settings. The trust settings will ultimately be used in enhanced
5181 certificate chain verification routines: currently a certificate
5182 can only be trusted if it is self signed and then it is trusted
5186 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
5187 The problem was that one of the replacement routines had not been working
5188 since SSLeay releases. For now the offending routine has been replaced
5189 with non-optimised assembler. Even so, this now gives around 95%
5190 performance improvement for 1024 bit RSA signs.
5193 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
5194 handling. Most clients have the effective key size in bits equal to
5195 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
5196 A few however don't do this and instead use the size of the decrypted key
5197 to determine the RC2 key length and the AlgorithmIdentifier to determine
5198 the effective key length. In this case the effective key length can still
5199 be 40 bits but the key length can be 168 bits for example. This is fixed
5200 by manually forcing an RC2 key into the EVP_PKEY structure because the
5201 EVP code can't currently handle unusual RC2 key sizes: it always assumes
5202 the key length and effective key length are equal.
5205 *) Add a bunch of functions that should simplify the creation of
5206 X509_NAME structures. Now you should be able to do:
5207 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
5208 and have it automatically work out the correct field type and fill in
5209 the structures. The more adventurous can try:
5210 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
5211 and it will (hopefully) work out the correct multibyte encoding.
5214 *) Change the 'req' utility to use the new field handling and multibyte
5215 copy routines. Before the DN field creation was handled in an ad hoc
5216 way in req, ca, and x509 which was rather broken and didn't support
5217 BMPStrings or UTF8Strings. Since some software doesn't implement
5218 BMPStrings or UTF8Strings yet, they can be enabled using the config file
5219 using the dirstring_type option. See the new comment in the default
5220 openssl.cnf for more info.
5223 *) Make crypto/rand/md_rand.c more robust:
5224 - Assure unique random numbers after fork().
5225 - Make sure that concurrent threads access the global counter and
5226 md serializably so that we never lose entropy in them
5227 or use exactly the same state in multiple threads.
5228 Access to the large state is not always serializable because
5229 the additional locking could be a performance killer, and
5230 md should be large enough anyway.
5233 *) New file apps/app_rand.c with commonly needed functionality
5234 for handling the random seed file.
5236 Use the random seed file in some applications that previously did not:
5238 dsaparam -genkey (which also ignored its '-rand' option),
5241 x509 (when signing).
5242 Except on systems with /dev/urandom, it is crucial to have a random
5243 seed file at least for key creation, DSA signing, and for DH exchanges;
5244 for RSA signatures we could do without one.
5246 gendh and gendsa (unlike genrsa) used to read only the first byte
5247 of each file listed in the '-rand' option. The function as previously
5248 found in genrsa is now in app_rand.c and is used by all programs
5249 that support '-rand'.
5252 *) In RAND_write_file, use mode 0600 for creating files;
5253 don't just chmod when it may be too late.
5256 *) Report an error from X509_STORE_load_locations
5257 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
5260 *) New function ASN1_mbstring_copy() this copies a string in either
5261 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
5262 into an ASN1_STRING type. A mask of permissible types is passed
5263 and it chooses the "minimal" type to use or an error if not type
5267 *) Add function equivalents to the various macros in asn1.h. The old
5268 macros are retained with an M_ prefix. Code inside the library can
5269 use the M_ macros. External code (including the openssl utility)
5270 should *NOT* in order to be "shared library friendly".
5273 *) Add various functions that can check a certificate's extensions
5274 to see if it usable for various purposes such as SSL client,
5275 server or S/MIME and CAs of these types. This is currently
5276 VERY EXPERIMENTAL but will ultimately be used for certificate chain
5277 verification. Also added a -purpose flag to x509 utility to
5278 print out all the purposes.
5281 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
5285 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
5286 for, obtain and decode and extension and obtain its critical flag.
5287 This allows all the necessary extension code to be handled in a
5288 single function call.
5291 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
5292 platforms. See crypto/rc4/rc4_enc.c for further details.
5295 *) New -noout option to asn1parse. This causes no output to be produced
5296 its main use is when combined with -strparse and -out to extract data
5297 from a file (which may not be in ASN.1 format).
5300 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
5301 when producing the local key id.
5302 [Richard Levitte <levitte@stacken.kth.se>]
5304 *) New option -dhparam in s_server. This allows a DH parameter file to be
5305 stated explicitly. If it is not stated then it tries the first server
5306 certificate file. The previous behaviour hard coded the filename
5310 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
5311 a public key to be input or output. For example:
5312 openssl rsa -in key.pem -pubout -out pubkey.pem
5313 Also added necessary DSA public key functions to handle this.
5316 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
5317 in the message. This was handled by allowing
5318 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
5319 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
5321 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
5322 to the end of the strings whereas this didn't. This would cause problems
5323 if strings read with d2i_ASN1_bytes() were later modified.
5324 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
5326 *) Fix for base64 decode bug. When a base64 bio reads only one line of
5327 data and it contains EOF it will end up returning an error. This is
5328 caused by input 46 bytes long. The cause is due to the way base64
5329 BIOs find the start of base64 encoded data. They do this by trying a
5330 trial decode on each line until they find one that works. When they
5331 do a flag is set and it starts again knowing it can pass all the
5332 data directly through the decoder. Unfortunately it doesn't reset
5333 the context it uses. This means that if EOF is reached an attempt
5334 is made to pass two EOFs through the context and this causes the
5335 resulting error. This can also cause other problems as well. As is
5336 usual with these problems it takes *ages* to find and the fix is
5337 trivial: move one line.
5338 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
5340 *) Ugly workaround to get s_client and s_server working under Windows. The
5341 old code wouldn't work because it needed to select() on sockets and the
5342 tty (for keypresses and to see if data could be written). Win32 only
5343 supports select() on sockets so we select() with a 1s timeout on the
5344 sockets and then see if any characters are waiting to be read, if none
5345 are present then we retry, we also assume we can always write data to
5346 the tty. This isn't nice because the code then blocks until we've
5347 received a complete line of data and it is effectively polling the
5348 keyboard at 1s intervals: however it's quite a bit better than not
5349 working at all :-) A dedicated Windows application might handle this
5350 with an event loop for example.
5353 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
5354 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
5355 will be called when RSA_sign() and RSA_verify() are used. This is useful
5356 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
5357 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
5358 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
5359 This necessitated the support of an extra signature type NID_md5_sha1
5360 for SSL signatures and modifications to the SSL library to use it instead
5361 of calling RSA_public_decrypt() and RSA_private_encrypt().
5364 *) Add new -verify -CAfile and -CApath options to the crl program, these
5365 will lookup a CRL issuers certificate and verify the signature in a
5366 similar way to the verify program. Tidy up the crl program so it
5367 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
5368 less strict. It will now permit CRL extensions even if it is not
5369 a V2 CRL: this will allow it to tolerate some broken CRLs.
5372 *) Initialize all non-automatic variables each time one of the openssl
5373 sub-programs is started (this is necessary as they may be started
5374 multiple times from the "OpenSSL>" prompt).
5375 [Lennart Bang, Bodo Moeller]
5377 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
5378 removing all other RSA functionality (this is what NO_RSA does). This
5379 is so (for example) those in the US can disable those operations covered
5380 by the RSA patent while allowing storage and parsing of RSA keys and RSA
5384 *) Non-copying interface to BIO pairs.
5385 (still largely untested)
5388 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
5389 ASCII string. This was handled independently in various places before.
5392 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
5393 UTF8 strings a character at a time.
5396 *) Use client_version from client hello to select the protocol
5397 (s23_srvr.c) and for RSA client key exchange verification
5398 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
5401 *) Add various utility functions to handle SPKACs, these were previously
5402 handled by poking round in the structure internals. Added new function
5403 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
5404 print, verify and generate SPKACs. Based on an original idea from
5405 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
5408 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
5411 *) Allow the config file extension section to be overwritten on the
5412 command line. Based on an original idea from Massimiliano Pala
5413 <madwolf@comune.modena.it>. The new option is called -extensions
5414 and can be applied to ca, req and x509. Also -reqexts to override
5415 the request extensions in req and -crlexts to override the crl extensions
5419 *) Add new feature to the SPKAC handling in ca. Now you can include
5420 the same field multiple times by preceding it by "XXXX." for example:
5423 this is the same syntax as used in the req config file.
5426 *) Allow certificate extensions to be added to certificate requests. These
5427 are specified in a 'req_extensions' option of the req section of the
5428 config file. They can be printed out with the -text option to req but
5429 are otherwise ignored at present.
5432 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
5433 data read consists of only the final block it would not decrypted because
5434 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
5435 A misplaced 'break' also meant the decrypted final block might not be
5436 copied until the next read.
5439 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
5440 a few extra parameters to the DH structure: these will be useful if
5441 for example we want the value of 'q' or implement X9.42 DH.
5444 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
5445 provides hooks that allow the default DSA functions or functions on a
5446 "per key" basis to be replaced. This allows hardware acceleration and
5447 hardware key storage to be handled without major modification to the
5448 library. Also added low level modexp hooks and CRYPTO_EX structure and
5449 associated functions.
5452 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
5453 as "read only": it can't be written to and the buffer it points to will
5454 not be freed. Reading from a read only BIO is much more efficient than
5455 a normal memory BIO. This was added because there are several times when
5456 an area of memory needs to be read from a BIO. The previous method was
5457 to create a memory BIO and write the data to it, this results in two
5458 copies of the data and an O(n^2) reading algorithm. There is a new
5459 function BIO_new_mem_buf() which creates a read only memory BIO from
5460 an area of memory. Also modified the PKCS#7 routines to use read only
5464 *) Bugfix: ssl23_get_client_hello did not work properly when called in
5465 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
5466 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
5467 but a retry condition occured while trying to read the rest.
5470 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
5471 NID_pkcs7_encrypted by default: this was wrong since this should almost
5472 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
5473 the encrypted data type: this is a more sensible place to put it and it
5474 allows the PKCS#12 code to be tidied up that duplicated this
5478 *) Changed obj_dat.pl script so it takes its input and output files on
5479 the command line. This should avoid shell escape redirection problems
5483 *) Initial support for certificate extension requests, these are included
5484 in things like Xenroll certificate requests. Included functions to allow
5485 extensions to be obtained and added.
5488 *) -crlf option to s_client and s_server for sending newlines as
5489 CRLF (as required by many protocols).
5492 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
5494 *) Install libRSAglue.a when OpenSSL is built with RSAref.
5495 [Ralf S. Engelschall]
5497 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
5498 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
5500 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
5504 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
5505 DH parameters/keys (q is lost during that conversion, but the resulting
5506 DH parameters contain its length).
5508 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
5509 much faster than DH_generate_parameters (which creates parameters
5510 where p = 2*q + 1), and also the smaller q makes DH computations
5511 much more efficient (160-bit exponentiation instead of 1024-bit
5512 exponentiation); so this provides a convenient way to support DHE
5513 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
5514 utter importance to use
5515 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5517 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5518 when such DH parameters are used, because otherwise small subgroup
5519 attacks may become possible!
5522 *) Avoid memory leak in i2d_DHparams.
5525 *) Allow the -k option to be used more than once in the enc program:
5526 this allows the same encrypted message to be read by multiple recipients.
5529 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
5530 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
5531 it will always use the numerical form of the OID, even if it has a short
5535 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
5536 method only got called if p,q,dmp1,dmq1,iqmp components were present,
5537 otherwise bn_mod_exp was called. In the case of hardware keys for example
5538 no private key components need be present and it might store extra data
5539 in the RSA structure, which cannot be accessed from bn_mod_exp.
5540 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
5541 private key operations.
5544 *) Added support for SPARC Linux.
5547 *) pem_password_cb function type incompatibly changed from
5548 typedef int pem_password_cb(char *buf, int size, int rwflag);
5550 ....(char *buf, int size, int rwflag, void *userdata);
5551 so that applications can pass data to their callbacks:
5552 The PEM[_ASN1]_{read,write}... functions and macros now take an
5553 additional void * argument, which is just handed through whenever
5554 the password callback is called.
5555 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
5557 New function SSL_CTX_set_default_passwd_cb_userdata.
5559 Compatibility note: As many C implementations push function arguments
5560 onto the stack in reverse order, the new library version is likely to
5561 interoperate with programs that have been compiled with the old
5562 pem_password_cb definition (PEM_whatever takes some data that
5563 happens to be on the stack as its last argument, and the callback
5564 just ignores this garbage); but there is no guarantee whatsoever that
5567 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
5568 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
5569 problems not only on Windows, but also on some Unix platforms.
5570 To avoid problematic command lines, these definitions are now in an
5571 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
5572 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
5575 *) MIPS III/IV assembler module is reimplemented.
5578 *) More DES library cleanups: remove references to srand/rand and
5579 delete an unused file.
5582 *) Add support for the the free Netwide assembler (NASM) under Win32,
5583 since not many people have MASM (ml) and it can be hard to obtain.
5584 This is currently experimental but it seems to work OK and pass all
5585 the tests. Check out INSTALL.W32 for info.
5588 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
5589 without temporary keys kept an extra copy of the server key,
5590 and connections with temporary keys did not free everything in case
5594 *) New function RSA_check_key and new openssl rsa option -check
5595 for verifying the consistency of RSA keys.
5596 [Ulf Moeller, Bodo Moeller]
5598 *) Various changes to make Win32 compile work:
5599 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
5600 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
5601 comparison" warnings.
5602 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
5605 *) Add a debugging option to PKCS#5 v2 key generation function: when
5606 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
5607 derived keys are printed to stderr.
5610 *) Copy the flags in ASN1_STRING_dup().
5611 [Roman E. Pavlov <pre@mo.msk.ru>]
5613 *) The x509 application mishandled signing requests containing DSA
5614 keys when the signing key was also DSA and the parameters didn't match.
5616 It was supposed to omit the parameters when they matched the signing key:
5617 the verifying software was then supposed to automatically use the CA's
5618 parameters if they were absent from the end user certificate.
5620 Omitting parameters is no longer recommended. The test was also
5621 the wrong way round! This was probably due to unusual behaviour in
5622 EVP_cmp_parameters() which returns 1 if the parameters match.
5623 This meant that parameters were omitted when they *didn't* match and
5624 the certificate was useless. Certificates signed with 'ca' didn't have
5626 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
5628 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
5629 The interface is as follows:
5630 Applications can use
5631 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
5632 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
5633 "off" is now the default.
5634 The library internally uses
5635 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
5636 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
5637 to disable memory-checking temporarily.
5639 Some inconsistent states that previously were possible (and were
5640 even the default) are now avoided.
5642 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
5643 with each memory chunk allocated; this is occasionally more helpful
5644 than just having a counter.
5646 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
5648 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
5652 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
5653 which largely parallels "options", but is for changing API behaviour,
5654 whereas "options" are about protocol behaviour.
5655 Initial "mode" flags are:
5657 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
5658 a single record has been written.
5659 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
5660 retries use the same buffer location.
5661 (But all of the contents must be
5665 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
5668 *) Fix problems with no-hmac etc.
5669 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
5671 *) New functions RSA_get_default_method(), RSA_set_method() and
5672 RSA_get_method(). These allows replacement of RSA_METHODs without having
5673 to mess around with the internals of an RSA structure.
5676 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
5677 Also really enable memory leak checks in openssl.c and in some
5679 [Chad C. Mulligan, Bodo Moeller]
5681 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
5682 up the length of negative integers. This has now been simplified to just
5683 store the length when it is first determined and use it later, rather
5684 than trying to keep track of where data is copied and updating it to
5686 [Steve Henson, reported by Brien Wheeler
5687 <bwheeler@authentica-security.com>]
5689 *) Add a new function PKCS7_signatureVerify. This allows the verification
5690 of a PKCS#7 signature but with the signing certificate passed to the
5691 function itself. This contrasts with PKCS7_dataVerify which assumes the
5692 certificate is present in the PKCS#7 structure. This isn't always the
5693 case: certificates can be omitted from a PKCS#7 structure and be
5694 distributed by "out of band" means (such as a certificate database).
5697 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
5698 function prototypes in pem.h, also change util/mkdef.pl to add the
5699 necessary function names.
5702 *) mk1mf.pl (used by Windows builds) did not properly read the
5703 options set by Configure in the top level Makefile, and Configure
5704 was not even able to write more than one option correctly.
5705 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
5708 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
5709 file to be loaded from a BIO or FILE pointer. The BIO version will
5710 for example allow memory BIOs to contain config info.
5713 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
5714 Whoever hopes to achieve shared-library compatibility across versions
5715 must use this, not the compile-time macro.
5716 (Exercise 0.9.4: Which is the minimum library version required by
5718 Note: All this applies only to multi-threaded programs, others don't
5722 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
5723 through a BIO pair triggered the default case, i.e.
5724 SSLerr(...,SSL_R_UNKNOWN_STATE).
5727 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
5728 can use the SSL library even if none of the specific BIOs is
5732 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
5733 for the encoded length.
5734 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
5736 *) Add initial documentation of the X509V3 functions.
5739 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
5740 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
5741 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
5742 secure PKCS#8 private key format with a high iteration count.
5745 *) Fix determination of Perl interpreter: A perl or perl5
5746 _directory_ in $PATH was also accepted as the interpreter.
5747 [Ralf S. Engelschall]
5749 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
5750 wrong with it but it was very old and did things like calling
5751 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
5755 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
5756 to use the new extension code.
5759 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
5760 with macros. This should make it easier to change their form, add extra
5761 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
5765 *) Add to configuration table a new entry that can specify an alternative
5766 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
5767 according to Mark Crispin <MRC@Panda.COM>.
5771 *) DES CBC did not update the IV. Weird.
5774 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
5775 Changing the behaviour of the former might break existing programs --
5776 where IV updating is needed, des_ncbc_encrypt can be used.
5779 *) When bntest is run from "make test" it drives bc to check its
5780 calculations, as well as internally checking them. If an internal check
5781 fails, it needs to cause bc to give a non-zero result or make test carries
5782 on without noticing the failure. Fixed.
5785 *) DES library cleanups.
5788 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
5789 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
5790 ciphers. NOTE: although the key derivation function has been verified
5791 against some published test vectors it has not been extensively tested
5792 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
5796 *) Instead of "mkdir -p", which is not fully portable, use new
5797 Perl script "util/mkdir-p.pl".
5800 *) Rewrite the way password based encryption (PBE) is handled. It used to
5801 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
5802 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
5803 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
5804 the 'parameter' field of the AlgorithmIdentifier is passed to the
5805 underlying key generation function so it must do its own ASN1 parsing.
5806 This has also changed the EVP_PBE_CipherInit() function which now has a
5807 'parameter' argument instead of literal salt and iteration count values
5808 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
5811 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
5812 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
5813 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
5814 KEY" because this clashed with PKCS#8 unencrypted string. Since this
5815 value was just used as a "magic string" and not used directly its
5816 value doesn't matter.
5819 *) Introduce some semblance of const correctness to BN. Shame C doesn't
5823 *) "linux-sparc64" configuration (ultrapenguin).
5824 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
5825 "linux-sparc" configuration.
5826 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
5828 *) config now generates no-xxx options for missing ciphers.
5831 *) Support the EBCDIC character set (work in progress).
5832 File ebcdic.c not yet included because it has a different license.
5833 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5835 *) Support BS2000/OSD-POSIX.
5836 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5838 *) Make callbacks for key generation use void * instead of char *.
5841 *) Make S/MIME samples compile (not yet tested).
5844 *) Additional typesafe stacks.
5847 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
5851 Changes between 0.9.3 and 0.9.3a [29 May 1999]
5853 *) New configuration variant "sco5-gcc".
5855 *) Updated some demos.
5856 [Sean O Riordain, Wade Scholine]
5858 *) Add missing BIO_free at exit of pkcs12 application.
5861 *) Fix memory leak in conf.c.
5864 *) Updates for Win32 to assembler version of MD5.
5867 *) Set #! path to perl in apps/der_chop to where we found it
5868 instead of using a fixed path.
5871 *) SHA library changes for irix64-mips4-cc.
5874 *) Improvements for VMS support.
5878 Changes between 0.9.2b and 0.9.3 [24 May 1999]
5880 *) Bignum library bug fix. IRIX 6 passes "make test" now!
5881 This also avoids the problems with SC4.2 and unpatched SC5.
5882 [Andy Polyakov <appro@fy.chalmers.se>]
5884 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
5885 These are required because of the typesafe stack would otherwise break
5886 existing code. If old code used a structure member which used to be STACK
5887 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
5888 sk_num or sk_value it would produce an error because the num, data members
5889 are not present in STACK_OF. Now it just produces a warning. sk_set
5890 replaces the old method of assigning a value to sk_value
5891 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
5892 that does this will no longer work (and should use sk_set instead) but
5893 this could be regarded as a "questionable" behaviour anyway.
5896 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
5897 correctly handle encrypted S/MIME data.
5900 *) Change type of various DES function arguments from des_cblock
5901 (which means, in function argument declarations, pointer to char)
5902 to des_cblock * (meaning pointer to array with 8 char elements),
5903 which allows the compiler to do more typechecking; it was like
5904 that back in SSLeay, but with lots of ugly casts.
5906 Introduce new type const_des_cblock.
5909 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
5910 problems: find RecipientInfo structure that matches recipient certificate
5911 and initialise the ASN1 structures properly based on passed cipher.
5914 *) Belatedly make the BN tests actually check the results.
5917 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
5918 to and from BNs: it was completely broken. New compilation option
5919 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
5920 key elements as negative integers.
5923 *) Reorganize and speed up MD5.
5924 [Andy Polyakov <appro@fy.chalmers.se>]
5927 [Richard Levitte <richard@levitte.org>]
5929 *) New option -out to asn1parse to allow the parsed structure to be
5930 output to a file. This is most useful when combined with the -strparse
5931 option to examine the output of things like OCTET STRINGS.
5934 *) Make SSL library a little more fool-proof by not requiring any longer
5935 that SSL_set_{accept,connect}_state be called before
5936 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
5937 in many applications because usually everything *appeared* to work as
5938 intended anyway -- now it really works as intended).
5941 *) Move openssl.cnf out of lib/.
5944 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
5945 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
5946 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
5947 [Ralf S. Engelschall]
5949 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
5950 handle PKCS#7 enveloped data properly.
5951 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
5953 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
5954 copying pointers. The cert_st handling is changed by this in
5955 various ways (and thus what used to be known as ctx->default_cert
5956 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
5957 any longer when s->cert does not give us what we need).
5958 ssl_cert_instantiate becomes obsolete by this change.
5959 As soon as we've got the new code right (possibly it already is?),
5960 we have solved a couple of bugs of the earlier code where s->cert
5961 was used as if it could not have been shared with other SSL structures.
5963 Note that using the SSL API in certain dirty ways now will result
5964 in different behaviour than observed with earlier library versions:
5965 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
5966 does not influence s as it used to.
5968 In order to clean up things more thoroughly, inside SSL_SESSION
5969 we don't use CERT any longer, but a new structure SESS_CERT
5970 that holds per-session data (if available); currently, this is
5971 the peer's certificate chain and, for clients, the server's certificate
5972 and temporary key. CERT holds only those values that can have
5973 meaningful defaults in an SSL_CTX.
5976 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
5977 from the internal representation. Various PKCS#7 fixes: remove some
5978 evil casts and set the enc_dig_alg field properly based on the signing
5982 *) Allow PKCS#12 password to be set from the command line or the
5983 environment. Let 'ca' get its config file name from the environment
5984 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
5988 *) Allow certificate policies extension to use an IA5STRING for the
5989 organization field. This is contrary to the PKIX definition but
5990 VeriSign uses it and IE5 only recognises this form. Document 'x509'
5994 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
5995 without disallowing inline assembler and the like for non-pedantic builds.
5998 *) Support Borland C++ builder.
5999 [Janez Jere <jj@void.si>, modified by Ulf Möller]
6004 *) SHA-1 cleanups and performance enhancements.
6005 [Andy Polyakov <appro@fy.chalmers.se>]
6007 *) Sparc v8plus assembler for the bignum library.
6008 [Andy Polyakov <appro@fy.chalmers.se>]
6010 *) Accept any -xxx and +xxx compiler options in Configure.
6013 *) Update HPUX configuration.
6016 *) Add missing sk_<type>_unshift() function to safestack.h
6017 [Ralf S. Engelschall]
6019 *) New function SSL_CTX_use_certificate_chain_file that sets the
6020 "extra_cert"s in addition to the certificate. (This makes sense
6021 only for "PEM" format files, as chains as a whole are not
6025 *) Support verify_depth from the SSL API.
6026 x509_vfy.c had what can be considered an off-by-one-error:
6027 Its depth (which was not part of the external interface)
6028 was actually counting the number of certificates in a chain;
6029 now it really counts the depth.
6032 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
6033 instead of X509err, which often resulted in confusing error
6034 messages since the error codes are not globally unique
6035 (e.g. an alleged error in ssl3_accept when a certificate
6036 didn't match the private key).
6038 *) New function SSL_CTX_set_session_id_context that allows to set a default
6039 value (so that you don't need SSL_set_session_id_context for each
6040 connection using the SSL_CTX).
6043 *) OAEP decoding bug fix.
6046 *) Support INSTALL_PREFIX for package builders, as proposed by
6050 *) New Configure options "threads" and "no-threads". For systems
6051 where the proper compiler options are known (currently Solaris
6052 and Linux), "threads" is the default.
6055 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
6058 *) Install various scripts to $(OPENSSLDIR)/misc, not to
6059 $(INSTALLTOP)/bin -- they shouldn't clutter directories
6060 such as /usr/local/bin.
6063 *) "make linux-shared" to build shared libraries.
6064 [Niels Poppe <niels@netbox.org>]
6066 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
6069 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
6070 extension adding in x509 utility.
6073 *) Remove NOPROTO sections and error code comments.
6076 *) Partial rewrite of the DEF file generator to now parse the ANSI
6080 *) New Configure options --prefix=DIR and --openssldir=DIR.
6083 *) Complete rewrite of the error code script(s). It is all now handled
6084 by one script at the top level which handles error code gathering,
6085 header rewriting and C source file generation. It should be much better
6086 than the old method: it now uses a modified version of Ulf's parser to
6087 read the ANSI prototypes in all header files (thus the old K&R definitions
6088 aren't needed for error creation any more) and do a better job of
6089 translating function codes into names. The old 'ASN1 error code imbedded
6090 in a comment' is no longer necessary and it doesn't use .err files which
6091 have now been deleted. Also the error code call doesn't have to appear all
6092 on one line (which resulted in some large lines...).
6095 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
6098 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
6099 0 (which usually indicates a closed connection), but continue reading.
6102 *) Fix some race conditions.
6105 *) Add support for CRL distribution points extension. Add Certificate
6106 Policies and CRL distribution points documentation.
6109 *) Move the autogenerated header file parts to crypto/opensslconf.h.
6112 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
6113 8 of keying material. Merlin has also confirmed interop with this fix
6114 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
6115 [Merlin Hughes <merlin@baltimore.ie>]
6117 *) Fix lots of warnings.
6118 [Richard Levitte <levitte@stacken.kth.se>]
6120 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
6121 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
6122 [Richard Levitte <levitte@stacken.kth.se>]
6124 *) Fix problems with sizeof(long) == 8.
6125 [Andy Polyakov <appro@fy.chalmers.se>]
6127 *) Change functions to ANSI C.
6130 *) Fix typos in error codes.
6131 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
6133 *) Remove defunct assembler files from Configure.
6136 *) SPARC v8 assembler BIGNUM implementation.
6137 [Andy Polyakov <appro@fy.chalmers.se>]
6139 *) Support for Certificate Policies extension: both print and set.
6140 Various additions to support the r2i method this uses.
6143 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
6144 return a const string when you are expecting an allocated buffer.
6147 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
6148 types DirectoryString and DisplayText.
6151 *) Add code to allow r2i extensions to access the configuration database,
6152 add an LHASH database driver and add several ctx helper functions.
6155 *) Fix an evil bug in bn_expand2() which caused various BN functions to
6156 fail when they extended the size of a BIGNUM.
6159 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
6160 support typesafe stack.
6163 *) Fix typo in SSL_[gs]et_options().
6164 [Nils Frostberg <nils@medcom.se>]
6166 *) Delete various functions and files that belonged to the (now obsolete)
6167 old X509V3 handling code.
6170 *) New Configure option "rsaref".
6173 *) Don't auto-generate pem.h.
6176 *) Introduce type-safe ASN.1 SETs.
6179 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
6180 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
6182 *) Introduce type-safe STACKs. This will almost certainly break lots of code
6183 that links with OpenSSL (well at least cause lots of warnings), but fear
6184 not: the conversion is trivial, and it eliminates loads of evil casts. A
6185 few STACKed things have been converted already. Feel free to convert more.
6186 In the fullness of time, I'll do away with the STACK type altogether.
6189 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
6190 specified in <certfile> by updating the entry in the index.txt file.
6191 This way one no longer has to edit the index.txt file manually for
6192 revoking a certificate. The -revoke option does the gory details now.
6193 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
6195 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
6196 `-text' option at all and this way the `-noout -text' combination was
6197 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
6198 [Ralf S. Engelschall]
6200 *) Make sure a corresponding plain text error message exists for the
6201 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
6202 verify callback function determined that a certificate was revoked.
6203 [Ralf S. Engelschall]
6205 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
6206 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
6207 all available cipers including rc5, which was forgotten until now.
6208 In order to let the testing shell script know which algorithms
6209 are available, a new (up to now undocumented) command
6210 "openssl list-cipher-commands" is used.
6213 *) Bugfix: s_client occasionally would sleep in select() when
6214 it should have checked SSL_pending() first.
6217 *) New functions DSA_do_sign and DSA_do_verify to provide access to
6218 the raw DSA values prior to ASN.1 encoding.
6221 *) Tweaks to Configure
6222 [Niels Poppe <niels@netbox.org>]
6224 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
6228 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
6231 *) New config option to avoid instructions that are illegal on the 80386.
6232 The default code is faster, but requires at least a 486.
6235 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
6236 SSL2_SERVER_VERSION (not used at all) macros, which are now the
6237 same as SSL2_VERSION anyway.
6240 *) New "-showcerts" option for s_client.
6243 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
6244 application. Various cleanups and fixes.
6247 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
6248 modify error routines to work internally. Add error codes and PBE init
6249 to library startup routines.
6252 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
6253 packing functions to asn1 and evp. Changed function names and error
6254 codes along the way.
6257 *) PKCS12 integration: and so it begins... First of several patches to
6258 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
6259 objects to objects.h
6262 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
6263 and display support for Thawte strong extranet extension.
6266 *) Add LinuxPPC support.
6267 [Jeff Dubrule <igor@pobox.org>]
6269 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
6270 bn_div_words in alpha.s.
6271 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
6273 *) Make sure the RSA OAEP test is skipped under -DRSAref because
6274 OAEP isn't supported when OpenSSL is built with RSAref.
6275 [Ulf Moeller <ulf@fitug.de>]
6277 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
6278 so they no longer are missing under -DNOPROTO.
6279 [Soren S. Jorvang <soren@t.dk>]
6282 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
6284 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
6285 doesn't work when the session is reused. Coming soon!
6288 *) Fix a security hole, that allows sessions to be reused in the wrong
6289 context thus bypassing client cert protection! All software that uses
6290 client certs and session caches in multiple contexts NEEDS PATCHING to
6291 allow session reuse! A fuller solution is in the works.
6292 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
6294 *) Some more source tree cleanups (removed obsolete files
6295 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
6296 permission on "config" script to be executable) and a fix for the INSTALL
6298 [Ulf Moeller <ulf@fitug.de>]
6300 *) Remove some legacy and erroneous uses of malloc, free instead of
6302 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
6304 *) Make rsa_oaep_test return non-zero on error.
6305 [Ulf Moeller <ulf@fitug.de>]
6307 *) Add support for native Solaris shared libraries. Configure
6308 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
6309 if someone would make that last step automatic.
6310 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
6312 *) ctx_size was not built with the right compiler during "make links". Fixed.
6315 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
6316 except NULL ciphers". This means the default cipher list will no longer
6317 enable NULL ciphers. They need to be specifically enabled e.g. with
6318 the string "DEFAULT:eNULL".
6321 *) Fix to RSA private encryption routines: if p < q then it would
6322 occasionally produce an invalid result. This will only happen with
6323 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
6326 *) Be less restrictive and allow also `perl util/perlpath.pl
6327 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
6328 because this way one can also use an interpreter named `perl5' (which is
6329 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
6330 installed as `perl').
6331 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6333 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
6334 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6336 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
6337 advapi32.lib to Win32 build and change the pem test comparision
6338 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
6339 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
6340 and crypto/des/ede_cbcm_enc.c.
6343 *) DES quad checksum was broken on big-endian architectures. Fixed.
6346 *) Comment out two functions in bio.h that aren't implemented. Fix up the
6347 Win32 test batch file so it (might) work again. The Win32 test batch file
6348 is horrible: I feel ill....
6351 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
6352 in e_os.h. Audit of header files to check ANSI and non ANSI
6353 sections: 10 functions were absent from non ANSI section and not exported
6354 from Windows DLLs. Fixed up libeay.num for new functions.
6357 *) Make `openssl version' output lines consistent.
6358 [Ralf S. Engelschall]
6360 *) Fix Win32 symbol export lists for BIO functions: Added
6361 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
6362 to ms/libeay{16,32}.def.
6363 [Ralf S. Engelschall]
6365 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
6366 fine under Unix and passes some trivial tests I've now added. But the
6367 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
6368 added to make sure no one expects that this stuff really works in the
6369 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
6370 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
6372 [Ralf S. Engelschall]
6374 *) Fix the generation of two part addresses in perl.
6375 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
6377 *) Add config entry for Linux on MIPS.
6378 [John Tobey <jtobey@channel1.com>]
6380 *) Make links whenever Configure is run, unless we are on Windoze.
6383 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
6384 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
6388 *) Add a useful kludge to allow package maintainers to specify compiler and
6389 other platforms details on the command line without having to patch the
6390 Configure script everytime: One now can use ``perl Configure
6391 <id>:<details>'', i.e. platform ids are allowed to have details appended
6392 to them (seperated by colons). This is treated as there would be a static
6393 pre-configured entry in Configure's %table under key <id> with value
6394 <details> and ``perl Configure <id>'' is called. So, when you want to
6395 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
6396 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
6397 now, which overrides the FreeBSD-elf entry on-the-fly.
6398 [Ralf S. Engelschall]
6400 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
6403 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
6404 on the `perl Configure ...' command line. This way one can compile
6405 OpenSSL libraries with Position Independent Code (PIC) which is needed
6406 for linking it into DSOs.
6407 [Ralf S. Engelschall]
6409 *) Remarkably, export ciphers were totally broken and no-one had noticed!
6413 *) Cleaned up the LICENSE document: The official contact for any license
6414 questions now is the OpenSSL core team under openssl-core@openssl.org.
6415 And add a paragraph about the dual-license situation to make sure people
6416 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
6417 to the OpenSSL toolkit.
6418 [Ralf S. Engelschall]
6420 *) General source tree makefile cleanups: Made `making xxx in yyy...'
6421 display consistent in the source tree and replaced `/bin/rm' by `rm'.
6422 Additonally cleaned up the `make links' target: Remove unnecessary
6423 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
6424 to speed processing and no longer clutter the display with confusing
6425 stuff. Instead only the actually done links are displayed.
6426 [Ralf S. Engelschall]
6428 *) Permit null encryption ciphersuites, used for authentication only. It used
6429 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
6430 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
6434 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
6435 signed attributes when verifying signatures (this would break them),
6436 the detached data encoding was wrong and public keys obtained using
6437 X509_get_pubkey() weren't freed.
6440 *) Add text documentation for the BUFFER functions. Also added a work around
6441 to a Win95 console bug. This was triggered by the password read stuff: the
6442 last character typed gets carried over to the next fread(). If you were
6443 generating a new cert request using 'req' for example then the last
6444 character of the passphrase would be CR which would then enter the first
6448 *) Added the new `Includes OpenSSL Cryptography Software' button as
6449 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
6450 button and can be used by applications based on OpenSSL to show the
6451 relationship to the OpenSSL project.
6452 [Ralf S. Engelschall]
6454 *) Remove confusing variables in function signatures in files
6455 ssl/ssl_lib.c and ssl/ssl.h.
6456 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6458 *) Don't install bss_file.c under PREFIX/include/
6459 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6461 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
6462 functions that return function pointers and has support for NT specific
6463 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
6464 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
6465 unsigned to signed types: this was killing the Win32 compile.
6468 *) Add new certificate file to stack functions,
6469 SSL_add_dir_cert_subjects_to_stack() and
6470 SSL_add_file_cert_subjects_to_stack(). These largely supplant
6471 SSL_load_client_CA_file(), and can be used to add multiple certs easily
6472 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
6473 This means that Apache-SSL and similar packages don't have to mess around
6474 to add as many CAs as they want to the preferred list.
6477 *) Experiment with doxygen documentation. Currently only partially applied to
6479 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
6480 openssl.doxy as the configuration file.
6483 *) Get rid of remaining C++-style comments which strict C compilers hate.
6484 [Ralf S. Engelschall, pointed out by Carlos Amengual]
6486 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
6487 compiled in by default: it has problems with large keys.
6490 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
6491 DH private keys and/or callback functions which directly correspond to
6492 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
6493 is needed for applications which have to configure certificates on a
6494 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
6496 For the RSA certificate situation is makes no difference, but
6497 for the DSA certificate situation this fixes the "no shared cipher"
6498 problem where the OpenSSL cipher selection procedure failed because the
6499 temporary keys were not overtaken from the context and the API provided
6500 no way to reconfigure them.
6501 The new functions now let applications reconfigure the stuff and they
6502 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
6503 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
6504 non-public-API function ssl_cert_instantiate() is used as a helper
6505 function and also to reduce code redundancy inside ssl_rsa.c.
6506 [Ralf S. Engelschall]
6508 *) Move s_server -dcert and -dkey options out of the undocumented feature
6509 area because they are useful for the DSA situation and should be
6510 recognized by the users.
6511 [Ralf S. Engelschall]
6513 *) Fix the cipher decision scheme for export ciphers: the export bits are
6514 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
6515 SSL_EXP_MASK. So, the original variable has to be used instead of the
6516 already masked variable.
6517 [Richard Levitte <levitte@stacken.kth.se>]
6519 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
6520 [Richard Levitte <levitte@stacken.kth.se>]
6522 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
6523 from `int' to `unsigned int' because it's a length and initialized by
6524 EVP_DigestFinal() which expects an `unsigned int *'.
6525 [Richard Levitte <levitte@stacken.kth.se>]
6527 *) Don't hard-code path to Perl interpreter on shebang line of Configure
6528 script. Instead use the usual Shell->Perl transition trick.
6529 [Ralf S. Engelschall]
6531 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
6532 (in addition to RSA certificates) to match the behaviour of `openssl dsa
6533 -noout -modulus' as it's already the case for `openssl rsa -noout
6534 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
6535 currently the public key is printed (a decision which was already done by
6536 `openssl dsa -modulus' in the past) which serves a similar purpose.
6537 Additionally the NO_RSA no longer completely removes the whole -modulus
6538 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
6540 [Ralf S. Engelschall]
6542 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
6543 BIO. See the source (crypto/evp/bio_ok.c) for more info.
6544 [Arne Ansper <arne@ats.cyber.ee>]
6546 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
6547 to be added. Now both 'req' and 'ca' can use new objects defined in the
6551 *) Add cool BIO that does syslog (or event log on NT).
6552 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
6554 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
6555 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
6556 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
6557 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
6560 *) Add preliminary config info for new extension code.
6563 *) Make RSA_NO_PADDING really use no padding.
6564 [Ulf Moeller <ulf@fitug.de>]
6566 *) Generate errors when private/public key check is done.
6569 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
6570 for some CRL extensions and new objects added.
6573 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
6574 key usage extension and fuller support for authority key id.
6577 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
6578 padding method for RSA, which is recommended for new applications in PKCS
6579 #1 v2.0 (RFC 2437, October 1998).
6580 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
6581 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
6582 against Bleichbacher's attack on RSA.
6583 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
6586 *) Updates to the new SSL compression code
6587 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6589 *) Fix so that the version number in the master secret, when passed
6590 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
6591 (because the server will not accept higher), that the version number
6592 is 0x03,0x01, not 0x03,0x00
6593 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6595 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
6596 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
6597 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
6600 *) Support for RAW extensions where an arbitrary extension can be
6601 created by including its DER encoding. See apps/openssl.cnf for
6605 *) Make sure latest Perl versions don't interpret some generated C array
6606 code as Perl array code in the crypto/err/err_genc.pl script.
6607 [Lars Weber <3weber@informatik.uni-hamburg.de>]
6609 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
6610 not many people have the assembler. Various Win32 compilation fixes and
6611 update to the INSTALL.W32 file with (hopefully) more accurate Win32
6615 *) Modify configure script 'Configure' to automatically create crypto/date.h
6616 file under Win32 and also build pem.h from pem.org. New script
6617 util/mkfiles.pl to create the MINFO file on environments that can't do a
6618 'make files': perl util/mkfiles.pl >MINFO should work.
6621 *) Major rework of DES function declarations, in the pursuit of correctness
6622 and purity. As a result, many evil casts evaporated, and some weirdness,
6623 too. You may find this causes warnings in your code. Zapping your evil
6624 casts will probably fix them. Mostly.
6627 *) Fix for a typo in asn1.h. Bug fix to object creation script
6628 obj_dat.pl. It considered a zero in an object definition to mean
6629 "end of object": none of the objects in objects.h have any zeros
6630 so it wasn't spotted.
6631 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
6633 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
6634 Masking (CBCM). In the absence of test vectors, the best I have been able
6635 to do is check that the decrypt undoes the encrypt, so far. Send me test
6636 vectors if you have them.
6639 *) Correct calculation of key length for export ciphers (too much space was
6640 allocated for null ciphers). This has not been tested!
6643 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
6644 message is now correct (it understands "crypto" and "ssl" on its
6645 command line). There is also now an "update" option. This will update
6646 the util/ssleay.num and util/libeay.num files with any new functions.
6648 perl util/mkdef.pl crypto ssl update
6649 it will update them.
6652 *) Overhauled the Perl interface (perl/*):
6653 - ported BN stuff to OpenSSL's different BN library
6654 - made the perl/ source tree CVS-aware
6655 - renamed the package from SSLeay to OpenSSL (the files still contain
6656 their history because I've copied them in the repository)
6657 - removed obsolete files (the test scripts will be replaced
6658 by better Test::Harness variants in the future)
6659 [Ralf S. Engelschall]
6661 *) First cut for a very conservative source tree cleanup:
6662 1. merge various obsolete readme texts into doc/ssleay.txt
6663 where we collect the old documents and readme texts.
6664 2. remove the first part of files where I'm already sure that we no
6665 longer need them because of three reasons: either they are just temporary
6666 files which were left by Eric or they are preserved original files where
6667 I've verified that the diff is also available in the CVS via "cvs diff
6668 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
6669 the crypto/md/ stuff).
6670 [Ralf S. Engelschall]
6672 *) More extension code. Incomplete support for subject and issuer alt
6673 name, issuer and authority key id. Change the i2v function parameters
6674 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
6675 what that's for :-) Fix to ASN1 macro which messed up
6676 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
6679 *) Preliminary support for ENUMERATED type. This is largely copied from the
6683 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
6684 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6686 *) Make sure `make rehash' target really finds the `openssl' program.
6687 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6689 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
6690 like to hear about it if this slows down other processors.
6693 *) Add CygWin32 platform information to Configure script.
6694 [Alan Batie <batie@aahz.jf.intel.com>]
6696 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
6697 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
6699 *) New program nseq to manipulate netscape certificate sequences
6702 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
6706 *) Fixes to BN code. Previously the default was to define BN_RECURSION
6707 but the BN code had some problems that would cause failures when
6708 doing certificate verification and some other functions.
6709 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6711 *) Add ASN1 and PEM code to support netscape certificate sequences.
6714 *) Add ASN1 and PEM code to support netscape certificate sequences.
6717 *) Add several PKIX and private extended key usage OIDs.
6720 *) Modify the 'ca' program to handle the new extension code. Modify
6721 openssl.cnf for new extension format, add comments.
6724 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
6725 and add a sample to openssl.cnf so req -x509 now adds appropriate
6729 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
6730 error code, add initial support to X509_print() and x509 application.
6733 *) Takes a deep breath and start addding X509 V3 extension support code. Add
6734 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
6735 stuff is currently isolated and isn't even compiled yet.
6738 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
6739 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
6740 Removed the versions check from X509 routines when loading extensions:
6741 this allows certain broken certificates that don't set the version
6742 properly to be processed.
6745 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
6746 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
6747 can still be regenerated with "make depend".
6750 *) Spelling mistake in C version of CAST-128.
6751 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
6753 *) Changes to the error generation code. The perl script err-code.pl
6754 now reads in the old error codes and retains the old numbers, only
6755 adding new ones if necessary. It also only changes the .err files if new
6756 codes are added. The makefiles have been modified to only insert errors
6757 when needed (to avoid needlessly modifying header files). This is done
6758 by only inserting errors if the .err file is newer than the auto generated
6759 C file. To rebuild all the error codes from scratch (the old behaviour)
6760 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
6761 or delete all the .err files.
6764 *) CAST-128 was incorrectly implemented for short keys. The C version has
6765 been fixed, but is untested. The assembler versions are also fixed, but
6766 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
6767 to regenerate it if needed.
6768 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
6769 Hagino <itojun@kame.net>]
6771 *) File was opened incorrectly in randfile.c.
6772 [Ulf Möller <ulf@fitug.de>]
6774 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
6775 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
6776 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
6777 al: it's just almost always a UTCTime. Note this patch adds new error
6778 codes so do a "make errors" if there are problems.
6781 *) Correct Linux 1 recognition in config.
6782 [Ulf Möller <ulf@fitug.de>]
6784 *) Remove pointless MD5 hash when using DSA keys in ca.
6785 [Anonymous <nobody@replay.com>]
6787 *) Generate an error if given an empty string as a cert directory. Also
6788 generate an error if handed NULL (previously returned 0 to indicate an
6789 error, but didn't set one).
6790 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
6792 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
6795 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
6796 parameters. This was causing a warning which killed off the Win32 compile.
6799 *) Remove C++ style comments from crypto/bn/bn_local.h.
6800 [Neil Costigan <neil.costigan@celocom.com>]
6802 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
6803 based on a text string, looking up short and long names and finally
6804 "dot" format. The "dot" format stuff didn't work. Added new function
6805 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
6806 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
6807 OID is not part of the table.
6810 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
6811 X509_LOOKUP_by_alias().
6814 *) Sort openssl functions by name.
6817 *) Get the gendsa program working (hopefully) and add it to app list. Remove
6818 encryption from sample DSA keys (in case anyone is interested the password
6822 *) Make _all_ *_free functions accept a NULL pointer.
6823 [Frans Heymans <fheymans@isaserver.be>]
6825 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
6827 [Anonymous <nobody@replay.com>]
6829 *) s_server should send the CAfile as acceptable CAs, not its own cert.
6830 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6832 *) Don't blow it for numeric -newkey arguments to apps/req.
6833 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6835 *) Temp key "for export" tests were wrong in s3_srvr.c.
6836 [Anonymous <nobody@replay.com>]
6838 *) Add prototype for temp key callback functions
6839 SSL_CTX_set_tmp_{rsa,dh}_callback().
6842 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
6843 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
6846 *) X509_name_add_entry() freed the wrong thing after an error.
6847 [Arne Ansper <arne@ats.cyber.ee>]
6849 *) rsa_eay.c would attempt to free a NULL context.
6850 [Arne Ansper <arne@ats.cyber.ee>]
6852 *) BIO_s_socket() had a broken should_retry() on Windoze.
6853 [Arne Ansper <arne@ats.cyber.ee>]
6855 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
6856 [Arne Ansper <arne@ats.cyber.ee>]
6858 *) Make sure the already existing X509_STORE->depth variable is initialized
6859 in X509_STORE_new(), but document the fact that this variable is still
6860 unused in the certificate verification process.
6861 [Ralf S. Engelschall]
6863 *) Fix the various library and apps files to free up pkeys obtained from
6864 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
6867 *) Fix reference counting in X509_PUBKEY_get(). This makes
6868 demos/maurice/example2.c work, amongst others, probably.
6869 [Steve Henson and Ben Laurie]
6871 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
6872 `openssl' and second, the shortcut symlinks for the `openssl <command>'
6873 are no longer created. This way we have a single and consistent command
6874 line interface `openssl <command>', similar to `cvs <command>'.
6875 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
6877 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
6878 BIT STRING wrapper always have zero unused bits.
6881 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
6884 *) Make the top-level INSTALL documentation easier to understand.
6887 *) Makefiles updated to exit if an error occurs in a sub-directory
6888 make (including if user presses ^C) [Paul Sutton]
6890 *) Make Montgomery context stuff explicit in RSA data structure.
6893 *) Fix build order of pem and err to allow for generated pem.h.
6896 *) Fix renumbering bug in X509_NAME_delete_entry().
6899 *) Enhanced the err-ins.pl script so it makes the error library number
6900 global and can add a library name. This is needed for external ASN1 and
6901 other error libraries.
6904 *) Fixed sk_insert which never worked properly.
6907 *) Fix ASN1 macros so they can handle indefinite length construted
6908 EXPLICIT tags. Some non standard certificates use these: they can now
6912 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
6913 into a single doc/ssleay.txt bundle. This way the information is still
6914 preserved but no longer messes up this directory. Now it's new room for
6915 the new set of documenation files.
6916 [Ralf S. Engelschall]
6918 *) SETs were incorrectly DER encoded. This was a major pain, because they
6919 shared code with SEQUENCEs, which aren't coded the same. This means that
6920 almost everything to do with SETs or SEQUENCEs has either changed name or
6921 number of arguments.
6922 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
6924 *) Fix test data to work with the above.
6927 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
6928 was already fixed by Eric for 0.9.1 it seems.
6929 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
6931 *) Autodetect FreeBSD3.
6934 *) Fix various bugs in Configure. This affects the following platforms:
6938 unixware-2.0-pentium
6942 *) Eliminate generated files from CVS. Reorder tests to regenerate files
6943 before they are needed.
6946 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
6950 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
6952 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
6953 changed SSLeay to OpenSSL in version strings.
6954 [Ralf S. Engelschall]
6956 *) Some fixups to the top-level documents.
6959 *) Fixed the nasty bug where rsaref.h was not found under compile-time
6960 because the symlink to include/ was missing.
6961 [Ralf S. Engelschall]
6963 *) Incorporated the popular no-RSA/DSA-only patches
6964 which allow to compile a RSA-free SSLeay.
6965 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
6967 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
6968 when "ssleay" is still not found.
6969 [Ralf S. Engelschall]
6971 *) Added more platforms to Configure: Cray T3E, HPUX 11,
6972 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
6974 *) Updated the README file.
6975 [Ralf S. Engelschall]
6977 *) Added various .cvsignore files in the CVS repository subdirs
6978 to make a "cvs update" really silent.
6979 [Ralf S. Engelschall]
6981 *) Recompiled the error-definition header files and added
6982 missing symbols to the Win32 linker tables.
6983 [Ralf S. Engelschall]
6985 *) Cleaned up the top-level documents;
6986 o new files: CHANGES and LICENSE
6987 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
6988 o merged COPYRIGHT into LICENSE
6989 o removed obsolete TODO file
6990 o renamed MICROSOFT to INSTALL.W32
6991 [Ralf S. Engelschall]
6993 *) Removed dummy files from the 0.9.1b source tree:
6994 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
6995 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
6996 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
6997 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
6998 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
6999 [Ralf S. Engelschall]
7001 *) Added various platform portability fixes.
7004 *) The Genesis of the OpenSSL rpject:
7005 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
7006 Young and Tim J. Hudson created while they were working for C2Net until
7008 [The OpenSSL Project]
7011 Changes between 0.9.0b and 0.9.1b [not released]
7013 *) Updated a few CA certificates under certs/
7016 *) Changed some BIGNUM api stuff.
7019 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
7020 DGUX x86, Linux Alpha, etc.
7023 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
7024 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
7028 *) Add -strparse option to asn1pars program which parses nested
7030 [Dr Stephen Henson <shenson@bigfoot.com>]
7032 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
7035 *) DSA fix for "ca" program.
7038 *) Added "-genkey" option to "dsaparam" program.
7041 *) Added RIPE MD160 (rmd160) message digest.
7044 *) Added -a (all) option to "ssleay version" command.
7047 *) Added PLATFORM define which is the id given to Configure.
7050 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
7053 *) Extended the ASN.1 parser routines.
7056 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
7059 *) Added a BN_CTX to the BN library.
7062 *) Fixed the weak key values in DES library
7065 *) Changed API in EVP library for cipher aliases.
7068 *) Added support for RC2/64bit cipher.
7071 *) Converted the lhash library to the crypto/mem.c functions.
7074 *) Added more recognized ASN.1 object ids.
7077 *) Added more RSA padding checks for SSL/TLS.
7080 *) Added BIO proxy/filter functionality.
7083 *) Added extra_certs to SSL_CTX which can be used
7084 send extra CA certificates to the client in the CA cert chain sending
7085 process. It can be configured with SSL_CTX_add_extra_chain_cert().
7088 *) Now Fortezza is denied in the authentication phase because
7089 this is key exchange mechanism is not supported by SSLeay at all.
7092 *) Additional PKCS1 checks.
7095 *) Support the string "TLSv1" for all TLS v1 ciphers.
7098 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
7099 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
7102 *) Fixed a few memory leaks.
7105 *) Fixed various code and comment typos.
7108 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
7109 bytes sent in the client random.
7110 [Edward Bishop <ebishop@spyglass.com>]