5 Changes between 0.9.7c and 0.9.8 [xx XXX xxxx]
7 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
8 that can only be obtained through BN_CTX_new() (which implicitly
9 initialises it). The presence of this function only made it possible
10 to overwrite an existing structure (and cause memory leaks).
13 *) Because of the callback-based approach for implementing LHASH as a
14 template type, lh_insert() adds opaque objects to hash-tables and
15 lh_doall() or lh_doall_arg() are typically used with a destructor callback
16 to clean up those corresponding objects before destroying the hash table
17 (and losing the object pointers). So some over-zealous constifications in
18 LHASH have been relaxed so that lh_insert() does not take (nor store) the
19 objects as "const" and the lh_doall[_arg] callback wrappers are not
20 prototyped to have "const" restrictions on the object pointers they are
21 given (and so aren't required to cast them away any more).
24 *) The tmdiff.h API was so ugly and minimal that our own timing utility
25 (speed) prefers to use its own implementation. The two implementations
26 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
27 its object type properly exposed (MS_TM) instead of casting to/from "char
28 *". This may still change yet if someone realises MS_TM and "ms_time_***"
29 aren't necessarily the greatest nomenclatures - but this is what was used
30 internally to the implementation so I've used that for now.
33 *) Ensure that deprecated functions do not get compiled when
34 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
35 the self-tests were still using deprecated key-generation functions so
36 these have been updated also.
39 *) Reorganise PKCS#7 code to separate the digest location functionality
40 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
41 New function PKCS7_set_digest() to set the digest type for PKCS#7
42 digestedData type. Add additional code to correctly generate the
43 digestedData type and add support for this type in PKCS7 initialization
47 *) New function PKCS7_set0_type_other() this initializes a PKCS7
48 structure of type "other".
51 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
52 sure the loop does correctly stop and breaking ("division by zero")
53 modulus operations are not performed. The (pre-generated) prime
54 table crypto/bn/bn_prime.h was already correct, but it could not be
55 re-generated on some platforms because of the "division by zero"
56 situation in the script.
59 *) Update support for ECC-based TLS ciphersuites according to
60 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
61 SHA-1 now is only used for "small" curves (where the
62 representation of a field element takes up to 24 bytes); for
63 larger curves, the field element resulting from ECDH is directly
64 used as premaster secret.
65 [Douglas Stebila (Sun Microsystems Laboratories)]
67 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
68 curve secp160r1 to the tests.
69 [Douglas Stebila (Sun Microsystems Laboratories)]
71 *) Add the possibility to load symbols globally with DSO.
72 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
74 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
75 control of the error stack.
78 *) Add support for STORE in ENGINE.
81 *) Add the STORE type. The intention is to provide a common interface
82 to certificate and key stores, be they simple file-based stores, or
83 HSM-type store, or LDAP stores, or...
84 NOTE: The code is currently UNTESTED and isn't really used anywhere.
87 *) Add a generic structure called OPENSSL_ITEM. This can be used to
88 pass a list of arguments to any function as well as provide a way
89 for a function to pass data back to the caller.
92 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
93 works like BUF_strdup() but can be used to duplicate a portion of
94 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
98 *) Add the function sk_find_ex() which works like sk_find(), but will
99 return an index to an element even if an exact match couldn't be
100 found. The index is guaranteed to point at the element where the
101 searched-for key would be inserted to preserve sorting order.
104 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
105 takes an extra flags argument for optional functionality. Currently,
106 the following flags are defined:
108 OBJ_BSEARCH_VALUE_ON_NOMATCH
109 This one gets OBJ_bsearch_ex() to return a pointer to the first
110 element where the comparing function returns a negative or zero
113 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
114 This one gets OBJ_bsearch_ex() to return a pointer to the first
115 element where the comparing function returns zero. This is useful
116 if there are more than one element where the comparing function
120 *) Make it possible to create self-signed certificates with 'openssl ca'
121 in such a way that the self-signed certificate becomes part of the
122 CA database and uses the same mechanisms for serial number generation
123 as all other certificate signing. The new flag '-selfsign' enables
124 this functionality. Adapt CA.sh and CA.pl.in.
127 *) Add functionality to check the public key of a certificate request
128 against a given private. This is useful to check that a certificate
129 request can be signed by that key (self-signing).
132 *) Make it possible to have multiple active certificates with the same
133 subject in the CA index file. This is done only if the keyword
134 'unique_subject' is set to 'no' in the main CA section (default
135 if 'CA_default') of the configuration file. The value is saved
136 with the database itself in a separate index attribute file,
137 named like the index file with '.attr' appended to the name.
140 *) Generate muti valued AVAs using '+' notation in config files for
144 *) Support for nameConstraints certificate extension.
147 *) Support for policyConstraints certificate extension.
150 *) Support for policyMappings certificate extension.
153 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
154 ENGINE as defaults for all supported algorithms irrespective of
155 the 'flags' parameter. 'flags' is now honoured, so applications
156 should make sure they are passing it correctly.
159 *) Make sure the default DSA_METHOD implementation only uses its
160 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
161 and change its own handlers to be NULL so as to remove unnecessary
162 indirection. This lets alternative implementations fallback to the
163 default implementation more easily.
166 *) Support for directoryName in GeneralName related extensions
170 *) Make it possible to link applications using Makefile.shared.
171 Make that possible even when linking against static libraries!
174 *) Support for single pass processing for S/MIME signing. This now
175 means that S/MIME signing can be done from a pipe, in addition
176 cleartext signing (multipart/signed type) is effectively streaming
177 and the signed data does not need to be all held in memory.
179 This is done with a new flag PKCS7_STREAM. When this flag is set
180 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
181 is done after the data is output (and digests calculated) in
185 *) Add full support for -rpath/-R, both in shared libraries and
186 applications, at least on the platforms where it's known how
190 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
191 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
192 will now compute a table of multiples of the generator that
193 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
194 faster (notably in the case of a single point multiplication,
196 [Nils Larsch, Bodo Moeller]
198 *) IPv6 support for certificate extensions. The various extensions
199 which use the IP:a.b.c.d can now take IPv6 addresses using the
200 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
204 *) Added an ENGINE that implements RSA by performing private key
205 exponentiations with the GMP library. The conversions to and from
206 GMP's mpz_t format aren't optimised nor are any montgomery forms
207 cached, and on x86 it appears OpenSSL's own performance has caught up.
208 However there are likely to be other architectures where GMP could
209 provide a boost. This ENGINE is not built in by default, but it can be
210 specified at Configure time and should be accompanied by the necessary
211 linker additions, eg;
212 ./config -DOPENSSL_USE_GMP -lgmp
215 *) "openssl engine" will not display ENGINE/DSO load failure errors when
216 testing availability of engines with "-t" - the old behaviour is
217 produced by increasing the feature's verbosity with "-tt".
220 *) ECDSA routines: under certain error conditions uninitialized BN objects
221 could be freed. Solution: make sure initialization is performed early
222 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
226 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
227 and DH_METHOD (eg. by ENGINE implementations) to override the normal
228 software implementations. For DSA and DH, parameter generation can
229 also be overriden by providing the appropriate method callbacks.
232 *) Change the "progress" mechanism used in key-generation and
233 primality testing to functions that take a new BN_GENCB pointer in
234 place of callback/argument pairs. The new API functions have "_ex"
235 postfixes and the older functions are reimplemented as wrappers for
236 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
237 declarations of the old functions to help (graceful) attempts to
238 migrate to the new functions. Also, the new key-generation API
239 functions operate on a caller-supplied key-structure and return
240 success/failure rather than returning a key or NULL - this is to
241 help make "keygen" another member function of RSA_METHOD etc.
243 Example for using the new callback interface:
245 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
249 BN_GENCB_set(&my_cb, my_callback, my_arg);
251 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
252 /* For the meaning of a, b in calls to my_callback(), see the
253 * documentation of the function that calls the callback.
254 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
255 * my_callback should return 1 if it wants BN_is_prime_ex()
256 * to continue, or 0 to stop.
261 *) Change the ZLIB compression method to be stateful, and make it
262 available to TLS with the number defined in
263 draft-ietf-tls-compression-04.txt.
266 *) Add the ASN.1 structures and functions for CertificatePair, which
267 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
269 CertificatePair ::= SEQUENCE {
270 forward [0] Certificate OPTIONAL,
271 reverse [1] Certificate OPTIONAL,
272 -- at least one of the pair shall be present -- }
274 Also implement the PEM functions to read and write certificate
275 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
277 This needed to be defined, mostly for the sake of the LDAP
278 attribute crossCertificatePair, but may prove useful elsewhere as
282 *) Make it possible to inhibit symlinking of shared libraries in
283 Makefile.shared, for Cygwin's sake.
286 *) Extend the BIGNUM API by creating new macros that behave like
289 void BN_set_sign(BIGNUM *a, int neg);
290 int BN_get_sign(const BIGNUM *a);
292 and avoid the need to access 'a->neg' directly in applications.
293 [Nils Larsch <nla@trustcenter.de>]
295 *) Implement fast modular reduction for pseudo-Mersenne primes
296 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
297 EC_GROUP_new_curve_GFp() will now automatically use this
299 [Nils Larsch <nla@trustcenter.de>]
301 *) Add new lock type (CRYPTO_LOCK_BN).
304 *) Change the ENGINE framework to automatically load engines
305 dynamically from specific directories unless they could be
306 found to already be built in or loaded. Move all the
307 current engines except for the cryptodev one to a new
309 The engines in engines/ are built as shared libraries if
310 the "shared" options was given to ./Configure or ./config.
311 Otherwise, they are inserted in libcrypto.a.
312 /usr/local/ssl/engines is the default directory for dynamic
313 engines, but that can be overriden at configure time through
314 the usual use of --prefix and/or --openssldir, and at run
315 time with the environment variable OPENSSL_ENGINES.
316 [Geoff Thorpe and Richard Levitte]
318 *) Add Makefile.shared, a helper makefile to build shared
319 libraries. Addapt Makefile.org.
322 *) Add version info to Win32 DLLs.
323 [Peter 'Luna' Runestig" <peter@runestig.com>]
325 *) Add new 'medium level' PKCS#12 API. Certificates and keys
326 can be added using this API to created arbitrary PKCS#12
327 files while avoiding the low level API.
329 New options to PKCS12_create(), key or cert can be NULL and
330 will then be omitted from the output file. The encryption
331 algorithm NIDs can be set to -1 for no encryption, the mac
332 iteration count can be set to 0 to omit the mac.
334 Enhance pkcs12 utility by making the -nokeys and -nocerts
335 options work when creating a PKCS#12 file. New option -nomac
336 to omit the mac, NONE can be set for an encryption algorithm.
337 New code is modified to use the enhanced PKCS12_create()
338 instead of the low level API.
341 *) Extend ASN1 encoder to support indefinite length constructed
342 encoding. This can output sequences tags and octet strings in
343 this form. Modify pk7_asn1.c to support indefinite length
344 encoding. This is experimental and needs additional code to
345 be useful, such as an ASN1 bio and some enhanced streaming
348 Extend template encode functionality so that tagging is passed
349 down to the template encoder.
352 *) Let 'openssl req' fail if an argument to '-newkey' is not
353 recognized instead of using RSA as a default.
356 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
357 As these are not official, they are not included in "ALL";
358 the "ECCdraft" ciphersuite group alias can be used to select them.
359 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
361 *) Add ECDH engine support.
362 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
364 *) Add ECDH in new directory crypto/ecdh/.
365 [Douglas Stebila (Sun Microsystems Laboratories)]
367 *) Let BN_rand_range() abort with an error after 100 iterations
368 without success (which indicates a broken PRNG).
371 *) Change BN_mod_sqrt() so that it verifies that the input value
372 is really the square of the return value. (Previously,
373 BN_mod_sqrt would show GIGO behaviour.)
376 *) Add named elliptic curves over binary fields from X9.62, SECG,
377 and WAP/WTLS; add OIDs that were still missing.
379 [Sheueling Chang Shantz and Douglas Stebila
380 (Sun Microsystems Laboratories)]
382 *) Extend the EC library for elliptic curves over binary fields
383 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
386 EC_GF2m_simple_method
390 EC_GROUP_new_curve_GF2m
391 EC_GROUP_set_curve_GF2m
392 EC_GROUP_get_curve_GF2m
393 EC_POINT_set_affine_coordinates_GF2m
394 EC_POINT_get_affine_coordinates_GF2m
395 EC_POINT_set_compressed_coordinates_GF2m
397 Point compression for binary fields is disabled by default for
398 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
401 As binary polynomials are represented as BIGNUMs, various members
402 of the EC_GROUP and EC_POINT data structures can be shared
403 between the implementations for prime fields and binary fields;
404 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
405 are essentially identical to their ..._GFp counterparts.
406 (For simplicity, the '..._GFp' prefix has been dropped from
407 various internal method names.)
409 An internal 'field_div' method (similar to 'field_mul' and
410 'field_sqr') has been added; this is used only for binary fields.
412 [Sheueling Chang Shantz and Douglas Stebila
413 (Sun Microsystems Laboratories)]
415 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
416 through methods ('mul', 'precompute_mult').
418 The generic implementations (now internally called 'ec_wNAF_mul'
419 and 'ec_wNAF_precomputed_mult') remain the default if these
420 methods are undefined.
422 [Sheueling Chang Shantz and Douglas Stebila
423 (Sun Microsystems Laboratories)]
425 *) New function EC_GROUP_get_degree, which is defined through
426 EC_METHOD. For curves over prime fields, this returns the bit
427 length of the modulus.
429 [Sheueling Chang Shantz and Douglas Stebila
430 (Sun Microsystems Laboratories)]
432 *) New functions EC_GROUP_dup, EC_POINT_dup.
433 (These simply call ..._new and ..._copy).
435 [Sheueling Chang Shantz and Douglas Stebila
436 (Sun Microsystems Laboratories)]
438 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
439 Polynomials are represented as BIGNUMs (where the sign bit is not
440 used) in the following functions [macros]:
443 BN_GF2m_sub [= BN_GF2m_add]
444 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
445 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
446 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
448 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
449 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
450 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
451 BN_GF2m_cmp [= BN_ucmp]
453 (Note that only the 'mod' functions are actually for fields GF(2^m).
454 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
456 For some functions, an the irreducible polynomial defining a
457 field can be given as an 'unsigned int[]' with strictly
458 decreasing elements giving the indices of those bits that are set;
459 i.e., p[] represents the polynomial
460 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
462 p[0] > p[1] > ... > p[k] = 0.
463 This applies to the following functions:
468 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
469 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
472 BN_GF2m_mod_solve_quad_arr
476 Conversion can be performed by the following functions:
481 bntest.c has additional tests for binary polynomial arithmetic.
483 Two implementations for BN_GF2m_mod_div() are available.
484 The default algorithm simply uses BN_GF2m_mod_inv() and
485 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
486 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
487 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
489 [Sheueling Chang Shantz and Douglas Stebila
490 (Sun Microsystems Laboratories)]
492 *) Add new error code 'ERR_R_DISABLED' that can be used when some
493 functionality is disabled at compile-time.
494 [Douglas Stebila <douglas.stebila@sun.com>]
496 *) Change default behaviour of 'openssl asn1parse' so that more
497 information is visible when viewing, e.g., a certificate:
499 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
500 mode the content of non-printable OCTET STRINGs is output in a
501 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
502 avoid the appearance of a printable string.
503 [Nils Larsch <nla@trustcenter.de>]
505 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
507 EC_GROUP_set_asn1_flag()
508 EC_GROUP_get_asn1_flag()
509 EC_GROUP_set_point_conversion_form()
510 EC_GROUP_get_point_conversion_form()
511 These control ASN1 encoding details:
512 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
513 has been set to OPENSSL_EC_NAMED_CURVE.
514 - Points are encoded in uncompressed form by default; options for
515 asn1_for are as for point2oct, namely
516 POINT_CONVERSION_COMPRESSED
517 POINT_CONVERSION_UNCOMPRESSED
518 POINT_CONVERSION_HYBRID
520 Also add 'seed' and 'seed_len' members to EC_GROUP with access
524 EC_GROUP_get_seed_len()
525 This is used only for ASN1 purposes (so far).
526 [Nils Larsch <nla@trustcenter.de>]
528 *) Add 'field_type' member to EC_METHOD, which holds the NID
529 of the appropriate field type OID. The new function
530 EC_METHOD_get_field_type() returns this value.
531 [Nils Larsch <nla@trustcenter.de>]
538 providing useful interfaces to EC_POINT_point2oct() and
539 EC_POINT_oct2point().
540 [Nils Larsch <nla@trustcenter.de>]
542 *) Change internals of the EC library so that the functions
543 EC_GROUP_set_generator()
544 EC_GROUP_get_generator()
546 EC_GROUP_get_cofactor()
547 are implemented directly in crypto/ec/ec_lib.c and not dispatched
548 to methods, which would lead to unnecessary code duplication when
549 adding different types of curves.
550 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
552 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
553 arithmetic, and such that modified wNAFs are generated
554 (which avoid length expansion in many cases).
557 *) Add a function EC_GROUP_check_discriminant() (defined via
558 EC_METHOD) that verifies that the curve discriminant is non-zero.
560 Add a function EC_GROUP_check() that makes some sanity tests
561 on a EC_GROUP, its generator and order. This includes
562 EC_GROUP_check_discriminant().
563 [Nils Larsch <nla@trustcenter.de>]
565 *) Add ECDSA in new directory crypto/ecdsa/.
567 Add applications 'openssl ecparam' and 'openssl ecdsa'
568 (these are based on 'openssl dsaparam' and 'openssl dsa').
570 ECDSA support is also included in various other files across the
571 library. Most notably,
572 - 'openssl req' now has a '-newkey ecdsa:file' option;
573 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
574 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
575 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
576 them suitable for ECDSA where domain parameters must be
577 extracted before the specific public key;
578 - ECDSA engine support has been added.
579 [Nils Larsch <nla@trustcenter.de>]
581 *) Include some named elliptic curves, and add OIDs from X9.62,
582 SECG, and WAP/WTLS. Each curve can be obtained from the new
584 EC_GROUP_new_by_nid(),
585 and the list of available named curves can be obtained with
586 EC_get_builtin_curves().
587 Also add a 'curve_name' member to EC_GROUP objects, which can be
591 [Nils Larsch <nla@trustcenter.de, Bodo Moeller]
593 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
594 was actually never needed) and in BN_mul(). The removal in BN_mul()
595 required a small change in bn_mul_part_recursive() and the addition
596 of the functions bn_cmp_part_words(), bn_sub_part_words() and
597 bn_add_part_words(), which do the same thing as bn_cmp_words(),
598 bn_sub_words() and bn_add_words() except they take arrays with
602 Changes between 0.9.7c and 0.9.7d [xx XXX XXXX]
603 *) Use the correct content when signing type "other".
606 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
608 *) Fix various bugs revealed by running the NISCC test suite:
610 Stop out of bounds reads in the ASN1 code when presented with
611 invalid tags (CAN-2003-0543 and CAN-2003-0544).
613 Free up ASN1_TYPE correctly if ANY type is invalid (CAN-2003-0545).
615 If verify callback ignores invalid public key errors don't try to check
616 certificate signature with the NULL public key.
620 *) New -ignore_err option in ocsp application to stop the server
621 exiting on the first error in a request.
624 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
625 if the server requested one: as stated in TLS 1.0 and SSL 3.0
629 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
630 extra data after the compression methods not only for TLS 1.0
631 but also for SSL 3.0 (as required by the specification).
632 [Bodo Moeller; problem pointed out by Matthias Loepfe]
634 *) Change X509_certificate_type() to mark the key as exported/exportable
635 when it's 512 *bits* long, not 512 bytes.
638 *) Change AES_cbc_encrypt() so it outputs exact multiple of
639 blocks during encryption.
642 *) Various fixes to base64 BIO and non blocking I/O. On write
643 flushes were not handled properly if the BIO retried. On read
644 data was not being buffered properly and had various logic bugs.
645 This also affects blocking I/O when the data being decoded is a
649 *) Various S/MIME bugfixes and compatibility changes:
650 output correct application/pkcs7 MIME type if
651 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
652 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
653 of files as .eml work). Correctly handle very long lines in MIME
657 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
659 *) Countermeasure against the Klima-Pokorny-Rosa extension of
660 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
661 a protocol version number mismatch like a decryption error
662 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
665 *) Turn on RSA blinding by default in the default implementation
666 to avoid a timing attack. Applications that don't want it can call
667 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
668 They would be ill-advised to do so in most cases.
669 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
671 *) Change RSA blinding code so that it works when the PRNG is not
672 seeded (in this case, the secret RSA exponent is abused as
673 an unpredictable seed -- if it is not unpredictable, there
674 is no point in blinding anyway). Make RSA blinding thread-safe
675 by remembering the creator's thread ID in rsa->blinding and
676 having all other threads use local one-time blinding factors
677 (this requires more computation than sharing rsa->blinding, but
678 avoids excessive locking; and if an RSA object is not shared
679 between threads, blinding will still be very fast).
682 yet to be integrated into this CVS branch:
683 - Geoff's ENGINE_set_default() fix
685 *) Target "mingw" now allows native Windows code to be generated in
686 the Cygwin environment as well as with the MinGW compiler.
689 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
691 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
692 via timing by performing a MAC computation even if incorrrect
693 block cipher padding has been found. This is a countermeasure
694 against active attacks where the attacker has to distinguish
695 between bad padding and a MAC verification error. (CAN-2003-0078)
697 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
698 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
699 Martin Vuagnoux (EPFL, Ilion)]
701 *) Make the no-err option work as intended. The intention with no-err
702 is not to have the whole error stack handling routines removed from
703 libcrypto, it's only intended to remove all the function name and
704 reason texts, thereby removing some of the footprint that may not
705 be interesting if those errors aren't displayed anyway.
707 NOTE: it's still possible for any application or module to have it's
708 own set of error texts inserted. The routines are there, just not
709 used by default when no-err is given.
712 *) Add support for FreeBSD on IA64.
713 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
715 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
716 Kerberos function mit_des_cbc_cksum(). Before this change,
717 the value returned by DES_cbc_cksum() was like the one from
718 mit_des_cbc_cksum(), except the bytes were swapped.
719 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
721 *) Allow an application to disable the automatic SSL chain building.
722 Before this a rather primitive chain build was always performed in
723 ssl3_output_cert_chain(): an application had no way to send the
724 correct chain if the automatic operation produced an incorrect result.
726 Now the chain builder is disabled if either:
728 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
730 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
732 The reasoning behind this is that an application would not want the
733 auto chain building to take place if extra chain certificates are
734 present and it might also want a means of sending no additional
735 certificates (for example the chain has two certificates and the
739 *) Add the possibility to build without the ENGINE framework.
740 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
742 *) Under Win32 gmtime() can return NULL: check return value in
743 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
746 *) DSA routines: under certain error conditions uninitialized BN objects
747 could be freed. Solution: make sure initialization is performed early
748 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
749 Nils Larsch <nla@trustcenter.de> via PR#459)
752 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
753 checked on reconnect on the client side, therefore session resumption
754 could still fail with a "ssl session id is different" error. This
755 behaviour is masked when SSL_OP_ALL is used due to
756 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
757 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
761 *) IA-32 assembler support enhancements: unified ELF targets, support
762 for SCO/Caldera platforms, fix for Cygwin shared build.
765 *) Add support for FreeBSD on sparc64. As a consequence, support for
766 FreeBSD on non-x86 processors is separate from x86 processors on
767 the config script, much like the NetBSD support.
768 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
770 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
772 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
775 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
776 code (06) was taken as the first octet of the session ID and the last
777 octet was ignored consequently. As a result SSLv2 client side session
778 caching could not have worked due to the session ID mismatch between
780 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
784 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
785 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
789 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
790 seems that in spite of existing for more than a year, many application
791 author have done nothing to provide the necessary callbacks, which
792 means that this particular engine will not work properly anywhere.
793 This is a very unfortunate situation which forces us, in the name
794 of usability, to give the hw_ncipher.c a static lock, which is part
796 NOTE: This is for the 0.9.7 series ONLY. This hack will never
797 appear in 0.9.8 or later. We EXPECT application authors to have
798 dealt properly with this when 0.9.8 is released (unless we actually
799 make such changes in the libcrypto locking code that changes will
800 have to be made anyway).
803 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
804 octets have been read, EOF or an error occurs. Without this change
805 some truncated ASN1 structures will not produce an error.
808 *) Disable Heimdal support, since it hasn't been fully implemented.
809 Still give the possibility to force the use of Heimdal, but with
810 warnings and a request that patches get sent to openssl-dev.
813 *) Add the VC-CE target, introduce the WINCE sysname, and add
814 INSTALL.WCE and appropriate conditionals to make it build.
815 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
817 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
818 cygssl-x.y.z.dll, where x, y and z are the major, minor and
819 edit numbers of the version.
820 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
822 *) Introduce safe string copy and catenation functions
823 (BUF_strlcpy() and BUF_strlcat()).
824 [Ben Laurie (CHATS) and Richard Levitte]
826 *) Avoid using fixed-size buffers for one-line DNs.
829 *) Add BUF_MEM_grow_clean() to avoid information leakage when
830 resizing buffers containing secrets, and use where appropriate.
833 *) Avoid using fixed size buffers for configuration file location.
836 *) Avoid filename truncation for various CA files.
839 *) Use sizeof in preference to magic numbers.
842 *) Avoid filename truncation in cert requests.
845 *) Add assertions to check for (supposedly impossible) buffer
849 *) Don't cache truncated DNS entries in the local cache (this could
850 potentially lead to a spoofing attack).
853 *) Fix various buffers to be large enough for hex/decimal
854 representations in a platform independent manner.
857 *) Add CRYPTO_realloc_clean() to avoid information leakage when
858 resizing buffers containing secrets, and use where appropriate.
861 *) Add BIO_indent() to avoid much slightly worrying code to do
865 *) Convert sprintf()/BIO_puts() to BIO_printf().
868 *) buffer_gets() could terminate with the buffer only half
872 *) Add assertions to prevent user-supplied crypto functions from
873 overflowing internal buffers by having large block sizes, etc.
876 *) New OPENSSL_assert() macro (similar to assert(), but enabled
880 *) Eliminate unused copy of key in RC4.
883 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
886 *) Fix off-by-one error in EGD path.
889 *) If RANDFILE path is too long, ignore instead of truncating.
892 *) Eliminate unused and incorrectly sized X.509 structure
896 *) Eliminate unused and dangerous function knumber().
899 *) Eliminate unused and dangerous structure, KSSL_ERR.
902 *) Protect against overlong session ID context length in an encoded
903 session object. Since these are local, this does not appear to be
907 *) Change from security patch (see 0.9.6e below) that did not affect
908 the 0.9.6 release series:
910 Remote buffer overflow in SSL3 protocol - an attacker could
911 supply an oversized master key in Kerberos-enabled versions.
915 *) Change the SSL kerb5 codes to match RFC 2712.
918 *) Make -nameopt work fully for req and add -reqopt switch.
919 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
921 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
922 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
924 *) Make sure tests can be performed even if the corresponding algorithms
925 have been removed entirely. This was also the last step to make
926 OpenSSL compilable with DJGPP under all reasonable conditions.
927 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
929 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
930 to allow version independent disabling of normally unselected ciphers,
931 which may be activated as a side-effect of selecting a single cipher.
933 (E.g., cipher list string "RSA" enables ciphersuites that are left
934 out of "ALL" because they do not provide symmetric encryption.
935 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
936 [Lutz Jaenicke, Bodo Moeller]
938 *) Add appropriate support for separate platform-dependent build
939 directories. The recommended way to make a platform-dependent
940 build directory is the following (tested on Linux), maybe with
943 # Place yourself outside of the OpenSSL source tree. In
944 # this example, the environment variable OPENSSL_SOURCE
945 # is assumed to contain the absolute OpenSSL source directory.
946 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
947 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
948 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
949 mkdir -p `dirname $F`
950 ln -s $OPENSSL_SOURCE/$F $F
953 To be absolutely sure not to disturb the source tree, a "make clean"
954 is a good thing. If it isn't successfull, don't worry about it,
955 it probably means the source directory is very clean.
958 *) Make sure any ENGINE control commands make local copies of string
959 pointers passed to them whenever necessary. Otherwise it is possible
960 the caller may have overwritten (or deallocated) the original string
961 data when a later ENGINE operation tries to use the stored values.
962 [Götz Babin-Ebell <babinebell@trustcenter.de>]
964 *) Improve diagnostics in file reading and command-line digests.
965 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
967 *) Add AES modes CFB and OFB to the object database. Correct an
968 error in AES-CFB decryption.
971 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
972 allows existing EVP_CIPHER_CTX structures to be reused after
973 calling EVP_*Final(). This behaviour is used by encryption
974 BIOs and some applications. This has the side effect that
975 applications must explicitly clean up cipher contexts with
976 EVP_CIPHER_CTX_cleanup() or they will leak memory.
979 *) Check the values of dna and dnb in bn_mul_recursive before calling
980 bn_mul_comba (a non zero value means the a or b arrays do not contain
981 n2 elements) and fallback to bn_mul_normal if either is not zero.
984 *) Fix escaping of non-ASCII characters when using the -subj option
985 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
988 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
989 form for "surname", serialNumber has no short form.
990 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
991 therefore remove "mail" short name for "internet 7".
992 The OID for unique identifiers in X509 certificates is
993 x500UniqueIdentifier, not uniqueIdentifier.
994 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
997 *) Add an "init" command to the ENGINE config module and auto initialize
998 ENGINEs. Without any "init" command the ENGINE will be initialized
999 after all ctrl commands have been executed on it. If init=1 the
1000 ENGINE is initailized at that point (ctrls before that point are run
1001 on the uninitialized ENGINE and after on the initialized one). If
1002 init=0 then the ENGINE will not be iniatialized at all.
1005 *) Fix the 'app_verify_callback' interface so that the user-defined
1006 argument is actually passed to the callback: In the
1007 SSL_CTX_set_cert_verify_callback() prototype, the callback
1008 declaration has been changed from
1011 int (*cb)(X509_STORE_CTX *,void *);
1012 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
1013 i=s->ctx->app_verify_callback(&ctx)
1014 has been changed into
1015 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
1017 To update applications using SSL_CTX_set_cert_verify_callback(),
1018 a dummy argument can be added to their callback functions.
1019 [D. K. Smetters <smetters@parc.xerox.com>]
1021 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
1022 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
1024 *) Add and OPENSSL_LOAD_CONF define which will cause
1025 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
1026 This allows older applications to transparently support certain
1027 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
1028 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
1029 load the config file and OPENSSL_add_all_algorithms_conf() which will
1030 always load it have also been added.
1033 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
1034 Adjust NIDs and EVP layer.
1035 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1037 *) Config modules support in openssl utility.
1039 Most commands now load modules from the config file,
1040 though in a few (such as version) this isn't done
1041 because it couldn't be used for anything.
1043 In the case of ca and req the config file used is
1044 the same as the utility itself: that is the -config
1045 command line option can be used to specify an
1049 *) Move default behaviour from OPENSSL_config(). If appname is NULL
1050 use "openssl_conf" if filename is NULL use default openssl config file.
1053 *) Add an argument to OPENSSL_config() to allow the use of an alternative
1054 config section name. Add a new flag to tolerate a missing config file
1055 and move code to CONF_modules_load_file().
1058 *) Support for crypto accelerator cards from Accelerated Encryption
1059 Processing, www.aep.ie. (Use engine 'aep')
1060 The support was copied from 0.9.6c [engine] and adapted/corrected
1061 to work with the new engine framework.
1062 [AEP Inc. and Richard Levitte]
1064 *) Support for SureWare crypto accelerator cards from Baltimore
1065 Technologies. (Use engine 'sureware')
1066 The support was copied from 0.9.6c [engine] and adapted
1067 to work with the new engine framework.
1070 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
1071 make the newer ENGINE framework commands for the CHIL engine work.
1072 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
1074 *) Make it possible to produce shared libraries on ReliantUNIX.
1075 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
1077 *) Add the configuration target debug-linux-ppro.
1078 Make 'openssl rsa' use the general key loading routines
1079 implemented in apps.c, and make those routines able to
1080 handle the key format FORMAT_NETSCAPE and the variant
1082 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1084 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
1085 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1087 *) Add -keyform to rsautl, and document -engine.
1088 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
1090 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
1091 BIO_R_NO_SUCH_FILE error code rather than the generic
1092 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
1095 *) Add new functions
1097 ERR_peek_last_error_line
1098 ERR_peek_last_error_line_data.
1099 These are similar to
1102 ERR_peek_error_line_data,
1103 but report on the latest error recorded rather than the first one
1104 still in the error queue.
1105 [Ben Laurie, Bodo Moeller]
1107 *) default_algorithms option in ENGINE config module. This allows things
1109 default_algorithms = ALL
1110 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
1113 *) Prelminary ENGINE config module.
1116 *) New experimental application configuration code.
1119 *) Change the AES code to follow the same name structure as all other
1120 symmetric ciphers, and behave the same way. Move everything to
1121 the directory crypto/aes, thereby obsoleting crypto/rijndael.
1122 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1124 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
1125 [Ben Laurie and Theo de Raadt]
1127 *) Add option to output public keys in req command.
1128 [Massimiliano Pala madwolf@openca.org]
1130 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
1131 (up to about 10% better than before for P-192 and P-224).
1134 *) New functions/macros
1136 SSL_CTX_set_msg_callback(ctx, cb)
1137 SSL_CTX_set_msg_callback_arg(ctx, arg)
1138 SSL_set_msg_callback(ssl, cb)
1139 SSL_set_msg_callback_arg(ssl, arg)
1141 to request calling a callback function
1143 void cb(int write_p, int version, int content_type,
1144 const void *buf, size_t len, SSL *ssl, void *arg)
1146 whenever a protocol message has been completely received
1147 (write_p == 0) or sent (write_p == 1). Here 'version' is the
1148 protocol version according to which the SSL library interprets
1149 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
1150 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
1151 the content type as defined in the SSL 3.0/TLS 1.0 protocol
1152 specification (change_cipher_spec(20), alert(21), handshake(22)).
1153 'buf' and 'len' point to the actual message, 'ssl' to the
1154 SSL object, and 'arg' is the application-defined value set by
1155 SSL[_CTX]_set_msg_callback_arg().
1157 'openssl s_client' and 'openssl s_server' have new '-msg' options
1158 to enable a callback that displays all protocol messages.
1161 *) Change the shared library support so shared libraries are built as
1162 soon as the corresponding static library is finished, and thereby get
1163 openssl and the test programs linked against the shared library.
1164 This still only happens when the keyword "shard" has been given to
1165 the configuration scripts.
1167 NOTE: shared library support is still an experimental thing, and
1168 backward binary compatibility is still not guaranteed.
1169 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
1171 *) Add support for Subject Information Access extension.
1172 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1174 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
1175 additional bytes when new memory had to be allocated, not just
1176 when reusing an existing buffer.
1179 *) New command line and configuration option 'utf8' for the req command.
1180 This allows field values to be specified as UTF8 strings.
1183 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
1184 runs for the former and machine-readable output for the latter.
1187 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
1188 of the e-mail address in the DN (i.e., it will go into a certificate
1189 extension only). The new configuration file option 'email_in_dn = no'
1190 has the same effect.
1191 [Massimiliano Pala madwolf@openca.org]
1193 *) Change all functions with names starting with des_ to be starting
1194 with DES_ instead. Add wrappers that are compatible with libdes,
1195 but are named _ossl_old_des_*. Finally, add macros that map the
1196 des_* symbols to the corresponding _ossl_old_des_* if libdes
1197 compatibility is desired. If OpenSSL 0.9.6c compatibility is
1198 desired, the des_* symbols will be mapped to DES_*, with one
1201 Since we provide two compatibility mappings, the user needs to
1202 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
1203 compatibility is desired. The default (i.e., when that macro
1204 isn't defined) is OpenSSL 0.9.6c compatibility.
1206 There are also macros that enable and disable the support of old
1207 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
1208 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
1209 are defined, the default will apply: to support the old des routines.
1211 In either case, one must include openssl/des.h to get the correct
1212 definitions. Do not try to just include openssl/des_old.h, that
1215 NOTE: This is a major break of an old API into a new one. Software
1216 authors are encouraged to switch to the DES_ style functions. Some
1217 time in the future, des_old.h and the libdes compatibility functions
1218 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
1219 default), and then completely removed.
1222 *) Test for certificates which contain unsupported critical extensions.
1223 If such a certificate is found during a verify operation it is
1224 rejected by default: this behaviour can be overridden by either
1225 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
1226 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
1227 X509_supported_extension() has also been added which returns 1 if a
1228 particular extension is supported.
1231 *) Modify the behaviour of EVP cipher functions in similar way to digests
1232 to retain compatibility with existing code.
1235 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
1236 compatibility with existing code. In particular the 'ctx' parameter does
1237 not have to be to be initialized before the call to EVP_DigestInit() and
1238 it is tidied up after a call to EVP_DigestFinal(). New function
1239 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
1240 EVP_MD_CTX_copy() changed to not require the destination to be
1241 initialized valid and new function EVP_MD_CTX_copy_ex() added which
1242 requires the destination to be valid.
1244 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
1245 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
1248 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
1249 so that complete 'Handshake' protocol structures are kept in memory
1250 instead of overwriting 'msg_type' and 'length' with 'body' data.
1253 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
1254 [Massimo Santin via Richard Levitte]
1256 *) Major restructuring to the underlying ENGINE code. This includes
1257 reduction of linker bloat, separation of pure "ENGINE" manipulation
1258 (initialisation, etc) from functionality dealing with implementations
1259 of specific crypto iterfaces. This change also introduces integrated
1260 support for symmetric ciphers and digest implementations - so ENGINEs
1261 can now accelerate these by providing EVP_CIPHER and EVP_MD
1262 implementations of their own. This is detailed in crypto/engine/README
1263 as it couldn't be adequately described here. However, there are a few
1264 API changes worth noting - some RSA, DSA, DH, and RAND functions that
1265 were changed in the original introduction of ENGINE code have now
1266 reverted back - the hooking from this code to ENGINE is now a good
1267 deal more passive and at run-time, operations deal directly with
1268 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
1269 dereferencing through an ENGINE pointer any more. Also, the ENGINE
1270 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
1271 they were not being used by the framework as there is no concept of a
1272 BIGNUM_METHOD and they could not be generalised to the new
1273 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
1274 ENGINE_cpy() has been removed as it cannot be consistently defined in
1278 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
1281 *) Change mkdef.pl to sort symbols that get the same entry number,
1282 and make sure the automatically generated functions ERR_load_*
1283 become part of libeay.num as well.
1286 *) New function SSL_renegotiate_pending(). This returns true once
1287 renegotiation has been requested (either SSL_renegotiate() call
1288 or HelloRequest/ClientHello receveived from the peer) and becomes
1289 false once a handshake has been completed.
1290 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
1291 sends a HelloRequest, but does not ensure that a handshake takes
1292 place. SSL_renegotiate_pending() is useful for checking if the
1293 client has followed the request.)
1296 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
1297 By default, clients may request session resumption even during
1298 renegotiation (if session ID contexts permit); with this option,
1299 session resumption is possible only in the first handshake.
1301 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
1302 more bits available for options that should not be part of
1303 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
1306 *) Add some demos for certificate and certificate request creation.
1309 *) Make maximum certificate chain size accepted from the peer application
1310 settable (SSL*_get/set_max_cert_list()), as proposed by
1311 "Douglas E. Engert" <deengert@anl.gov>.
1314 *) Add support for shared libraries for Unixware-7
1315 (Boyd Lynn Gerber <gerberb@zenez.com>).
1318 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
1319 be done prior to destruction. Use this to unload error strings from
1320 ENGINEs that load their own error strings. NB: This adds two new API
1321 functions to "get" and "set" this destroy handler in an ENGINE.
1324 *) Alter all existing ENGINE implementations (except "openssl" and
1325 "openbsd") to dynamically instantiate their own error strings. This
1326 makes them more flexible to be built both as statically-linked ENGINEs
1327 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
1328 Also, add stub code to each that makes building them as self-contained
1329 shared-libraries easier (see README.ENGINE).
1332 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
1333 implementations into applications that are completely implemented in
1334 self-contained shared-libraries. The "dynamic" ENGINE exposes control
1335 commands that can be used to configure what shared-library to load and
1336 to control aspects of the way it is handled. Also, made an update to
1337 the README.ENGINE file that brings its information up-to-date and
1338 provides some information and instructions on the "dynamic" ENGINE
1339 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
1342 *) Make it possible to unload ranges of ERR strings with a new
1343 "ERR_unload_strings" function.
1346 *) Add a copy() function to EVP_MD.
1349 *) Make EVP_MD routines take a context pointer instead of just the
1350 md_data void pointer.
1353 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
1354 that the digest can only process a single chunk of data
1355 (typically because it is provided by a piece of
1356 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
1357 is only going to provide a single chunk of data, and hence the
1358 framework needn't accumulate the data for oneshot drivers.
1361 *) As with "ERR", make it possible to replace the underlying "ex_data"
1362 functions. This change also alters the storage and management of global
1363 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
1364 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
1365 index counters. The API functions that use this state have been changed
1366 to take a "class_index" rather than pointers to the class's local STACK
1367 and counter, and there is now an API function to dynamically create new
1368 classes. This centralisation allows us to (a) plug a lot of the
1369 thread-safety problems that existed, and (b) makes it possible to clean
1370 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
1371 such data would previously have always leaked in application code and
1372 workarounds were in place to make the memory debugging turn a blind eye
1373 to it. Application code that doesn't use this new function will still
1374 leak as before, but their memory debugging output will announce it now
1375 rather than letting it slide.
1377 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
1378 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
1379 has a return value to indicate success or failure.
1382 *) Make it possible to replace the underlying "ERR" functions such that the
1383 global state (2 LHASH tables and 2 locks) is only used by the "default"
1384 implementation. This change also adds two functions to "get" and "set"
1385 the implementation prior to it being automatically set the first time
1386 any other ERR function takes place. Ie. an application can call "get",
1387 pass the return value to a module it has just loaded, and that module
1388 can call its own "set" function using that value. This means the
1389 module's "ERR" operations will use (and modify) the error state in the
1390 application and not in its own statically linked copy of OpenSSL code.
1393 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
1394 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
1395 the operation, and provides a more encapsulated way for external code
1396 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
1397 to use these functions rather than manually incrementing the counts.
1399 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
1402 *) Add EVP test program.
1405 *) Add symmetric cipher support to ENGINE. Expect the API to change!
1408 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
1409 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
1410 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
1411 These allow a CRL to be built without having to access X509_CRL fields
1412 directly. Modify 'ca' application to use new functions.
1415 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
1416 bug workarounds. Rollback attack detection is a security feature.
1417 The problem will only arise on OpenSSL servers when TLSv1 is not
1418 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
1419 Software authors not wanting to support TLSv1 will have special reasons
1420 for their choice and can explicitly enable this option.
1421 [Bodo Moeller, Lutz Jaenicke]
1423 *) Rationalise EVP so it can be extended: don't include a union of
1424 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
1425 (similar to those existing for EVP_CIPHER_CTX).
1430 EVP_MD_CTX_init(&md); /* new function call */
1431 EVP_DigestInit(&md, EVP_sha1());
1432 EVP_DigestUpdate(&md, in, len);
1433 EVP_DigestFinal(&md, out, NULL);
1434 EVP_MD_CTX_cleanup(&md); /* new function call */
1438 *) Make DES key schedule conform to the usual scheme, as well as
1439 correcting its structure. This means that calls to DES functions
1440 now have to pass a pointer to a des_key_schedule instead of a
1441 plain des_key_schedule (which was actually always a pointer
1444 des_key_schedule ks;
1446 des_set_key_checked(..., &ks);
1447 des_ncbc_encrypt(..., &ks, ...);
1449 (Note that a later change renames 'des_...' into 'DES_...'.)
1452 *) Initial reduction of linker bloat: the use of some functions, such as
1453 PEM causes large amounts of unused functions to be linked in due to
1454 poor organisation. For example pem_all.c contains every PEM function
1455 which has a knock on effect of linking in large amounts of (unused)
1456 ASN1 code. Grouping together similar functions and splitting unrelated
1457 functions prevents this.
1460 *) Cleanup of EVP macros.
1463 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
1464 correct _ecb suffix.
1467 *) Add initial OCSP responder support to ocsp application. The
1468 revocation information is handled using the text based index
1469 use by the ca application. The responder can either handle
1470 requests generated internally, supplied in files (for example
1471 via a CGI script) or using an internal minimal server.
1474 *) Add configuration choices to get zlib compression for TLS.
1477 *) Changes to Kerberos SSL for RFC 2712 compliance:
1478 1. Implemented real KerberosWrapper, instead of just using
1479 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
1480 2. Implemented optional authenticator field of KerberosWrapper.
1482 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
1483 and authenticator structs; see crypto/krb5/.
1485 Generalized Kerberos calls to support multiple Kerberos libraries.
1486 [Vern Staats <staatsvr@asc.hpc.mil>,
1487 Jeffrey Altman <jaltman@columbia.edu>
1488 via Richard Levitte]
1490 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
1491 already does with RSA. testdsa.h now has 'priv_key/pub_key'
1492 values for each of the key sizes rather than having just
1493 parameters (and 'speed' generating keys each time).
1496 *) Speed up EVP routines.
1499 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
1500 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
1501 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
1502 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
1504 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
1505 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
1506 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
1509 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
1511 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
1514 *) Added the OS2-EMX target.
1515 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
1517 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
1518 to support NCONF routines in extension code. New function CONF_set_nconf()
1519 to allow functions which take an NCONF to also handle the old LHASH
1520 structure: this means that the old CONF compatible routines can be
1521 retained (in particular wrt extensions) without having to duplicate the
1522 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
1525 *) Enhance the general user interface with mechanisms for inner control
1526 and with possibilities to have yes/no kind of prompts.
1529 *) Change all calls to low level digest routines in the library and
1530 applications to use EVP. Add missing calls to HMAC_cleanup() and
1531 don't assume HMAC_CTX can be copied using memcpy().
1532 [Verdon Walker <VWalker@novell.com>, Steve Henson]
1534 *) Add the possibility to control engines through control names but with
1535 arbitrary arguments instead of just a string.
1536 Change the key loaders to take a UI_METHOD instead of a callback
1537 function pointer. NOTE: this breaks binary compatibility with earlier
1538 versions of OpenSSL [engine].
1539 Adapt the nCipher code for these new conditions and add a card insertion
1543 *) Enhance the general user interface with mechanisms to better support
1544 dialog box interfaces, application-defined prompts, the possibility
1545 to use defaults (for example default passwords from somewhere else)
1546 and interrupts/cancellations.
1549 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
1550 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
1553 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
1554 tidy up some unnecessarily weird code in 'sk_new()').
1555 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
1557 *) Change the key loading routines for ENGINEs to use the same kind
1558 callback (pem_password_cb) as all other routines that need this
1562 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
1563 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
1564 than this minimum value is recommended.
1567 *) New random seeder for OpenVMS, using the system process statistics
1568 that are easily reachable.
1571 *) Windows apparently can't transparently handle global
1572 variables defined in DLLs. Initialisations such as:
1574 const ASN1_ITEM *it = &ASN1_INTEGER_it;
1576 wont compile. This is used by the any applications that need to
1577 declare their own ASN1 modules. This was fixed by adding the option
1578 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
1579 needed for static libraries under Win32.
1582 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
1583 setting of purpose and trust fields. New X509_STORE trust and
1584 purpose functions and tidy up setting in other SSL functions.
1587 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
1588 structure. These are inherited by X509_STORE_CTX when it is
1589 initialised. This allows various defaults to be set in the
1590 X509_STORE structure (such as flags for CRL checking and custom
1591 purpose or trust settings) for functions which only use X509_STORE_CTX
1592 internally such as S/MIME.
1594 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
1595 trust settings if they are not set in X509_STORE. This allows X509_STORE
1596 purposes and trust (in S/MIME for example) to override any set by default.
1598 Add command line options for CRL checking to smime, s_client and s_server
1602 *) Initial CRL based revocation checking. If the CRL checking flag(s)
1603 are set then the CRL is looked up in the X509_STORE structure and
1604 its validity and signature checked, then if the certificate is found
1605 in the CRL the verify fails with a revoked error.
1607 Various new CRL related callbacks added to X509_STORE_CTX structure.
1609 Command line options added to 'verify' application to support this.
1611 This needs some additional work, such as being able to handle multiple
1612 CRLs with different times, extension based lookup (rather than just
1613 by subject name) and ultimately more complete V2 CRL extension
1617 *) Add a general user interface API (crypto/ui/). This is designed
1618 to replace things like des_read_password and friends (backward
1619 compatibility functions using this new API are provided).
1620 The purpose is to remove prompting functions from the DES code
1621 section as well as provide for prompting through dialog boxes in
1622 a window system and the like.
1625 *) Add "ex_data" support to ENGINE so implementations can add state at a
1626 per-structure level rather than having to store it globally.
1629 *) Make it possible for ENGINE structures to be copied when retrieved by
1630 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
1631 This causes the "original" ENGINE structure to act like a template,
1632 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
1633 operational state can be localised to each ENGINE structure, despite the
1634 fact they all share the same "methods". New ENGINE structures returned in
1635 this case have no functional references and the return value is the single
1636 structural reference. This matches the single structural reference returned
1637 by ENGINE_by_id() normally, when it is incremented on the pre-existing
1641 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
1642 needs to match any other type at all we need to manually clear the
1646 *) Changes to the "openssl engine" utility to include;
1647 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
1648 about an ENGINE's available control commands.
1649 - executing control commands from command line arguments using the
1650 '-pre' and '-post' switches. '-post' is only used if '-t' is
1651 specified and the ENGINE is successfully initialised. The syntax for
1652 the individual commands are colon-separated, for example;
1653 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
1656 *) New dynamic control command support for ENGINEs. ENGINEs can now
1657 declare their own commands (numbers), names (strings), descriptions,
1658 and input types for run-time discovery by calling applications. A
1659 subset of these commands are implicitly classed as "executable"
1660 depending on their input type, and only these can be invoked through
1661 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
1662 can be based on user input, config files, etc). The distinction is
1663 that "executable" commands cannot return anything other than a boolean
1664 result and can only support numeric or string input, whereas some
1665 discoverable commands may only be for direct use through
1666 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
1667 pointers, or other custom uses. The "executable" commands are to
1668 support parameterisations of ENGINE behaviour that can be
1669 unambiguously defined by ENGINEs and used consistently across any
1670 OpenSSL-based application. Commands have been added to all the
1671 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
1672 control over shared-library paths without source code alterations.
1675 *) Changed all ENGINE implementations to dynamically allocate their
1676 ENGINEs rather than declaring them statically. Apart from this being
1677 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
1678 this also allows the implementations to compile without using the
1679 internal engine_int.h header.
1682 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
1683 'const' value. Any code that should be able to modify a RAND_METHOD
1684 should already have non-const pointers to it (ie. they should only
1685 modify their own ones).
1688 *) Made a variety of little tweaks to the ENGINE code.
1689 - "atalla" and "ubsec" string definitions were moved from header files
1690 to C code. "nuron" string definitions were placed in variables
1691 rather than hard-coded - allowing parameterisation of these values
1692 later on via ctrl() commands.
1693 - Removed unused "#if 0"'d code.
1694 - Fixed engine list iteration code so it uses ENGINE_free() to release
1695 structural references.
1696 - Constified the RAND_METHOD element of ENGINE structures.
1697 - Constified various get/set functions as appropriate and added
1698 missing functions (including a catch-all ENGINE_cpy that duplicates
1699 all ENGINE values onto a new ENGINE except reference counts/state).
1700 - Removed NULL parameter checks in get/set functions. Setting a method
1701 or function to NULL is a way of cancelling out a previously set
1702 value. Passing a NULL ENGINE parameter is just plain stupid anyway
1703 and doesn't justify the extra error symbols and code.
1704 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
1705 flags from engine_int.h to engine.h.
1706 - Changed prototypes for ENGINE handler functions (init(), finish(),
1707 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
1710 *) Implement binary inversion algorithm for BN_mod_inverse in addition
1711 to the algorithm using long division. The binary algorithm can be
1712 used only if the modulus is odd. On 32-bit systems, it is faster
1713 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
1714 roughly 5-15% for 256-bit moduli), so we use it only for moduli
1715 up to 450 bits. In 64-bit environments, the binary algorithm
1716 appears to be advantageous for much longer moduli; here we use it
1717 for moduli up to 2048 bits.
1720 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
1721 could not support the combine flag in choice fields.
1724 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
1725 extensions from a certificate request to the certificate.
1728 *) Allow multiple 'certopt' and 'nameopt' options to be separated
1729 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
1730 file: this allows the display of the certificate about to be
1731 signed to be customised, to allow certain fields to be included
1732 or excluded and extension details. The old system didn't display
1733 multicharacter strings properly, omitted fields not in the policy
1734 and couldn't display additional details such as extensions.
1737 *) Function EC_POINTs_mul for multiple scalar multiplication
1738 of an arbitrary number of elliptic curve points
1739 \sum scalars[i]*points[i],
1740 optionally including the generator defined for the EC_GROUP:
1741 scalar*generator + \sum scalars[i]*points[i].
1743 EC_POINT_mul is a simple wrapper function for the typical case
1744 that the point list has just one item (besides the optional
1748 *) First EC_METHODs for curves over GF(p):
1750 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
1751 operations and provides various method functions that can also
1752 operate with faster implementations of modular arithmetic.
1754 EC_GFp_mont_method() reuses most functions that are part of
1755 EC_GFp_simple_method, but uses Montgomery arithmetic.
1757 [Bodo Moeller; point addition and point doubling
1758 implementation directly derived from source code provided by
1759 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
1761 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
1762 crypto/ec/ec_lib.c):
1764 Curves are EC_GROUP objects (with an optional group generator)
1765 based on EC_METHODs that are built into the library.
1767 Points are EC_POINT objects based on EC_GROUP objects.
1769 Most of the framework would be able to handle curves over arbitrary
1770 finite fields, but as there are no obvious types for fields other
1771 than GF(p), some functions are limited to that for now.
1774 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
1775 that the file contains a complete HTTP response.
1778 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
1779 change the def and num file printf format specifier from "%-40sXXX"
1780 to "%-39s XXX". The latter will always guarantee a space after the
1781 field while the former will cause them to run together if the field
1782 is 40 of more characters long.
1785 *) Constify the cipher and digest 'method' functions and structures
1786 and modify related functions to take constant EVP_MD and EVP_CIPHER
1790 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
1791 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
1794 *) Modify EVP_Digest*() routines so they now return values. Although the
1795 internal software routines can never fail additional hardware versions
1799 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
1801 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
1802 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
1805 ERR_R_NESTED_ASN1_ERROR
1807 ERR_R_MISSING_ASN1_EOS
1808 were 4 .. 9, conflicting with
1809 ERR_LIB_RSA (= ERR_R_RSA_LIB)
1811 ERR_LIB_PEM (= ERR_R_PEM_LIB).
1812 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
1814 Add new error code 'ERR_R_INTERNAL_ERROR'.
1817 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
1821 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
1822 sets the subject name for a new request or supersedes the
1823 subject name in a given request. Formats that can be parsed are
1824 'CN=Some Name, OU=myOU, C=IT'
1826 'CN=Some Name/OU=myOU/C=IT'.
1828 Add options '-batch' and '-verbose' to 'openssl req'.
1829 [Massimiliano Pala <madwolf@hackmasters.net>]
1831 *) Introduce the possibility to access global variables through
1832 functions on platform were that's the best way to handle exporting
1833 global variables in shared libraries. To enable this functionality,
1834 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1835 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1836 is normally done by Configure or something similar).
1838 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1839 in the source file (foo.c) like this:
1841 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1842 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1844 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1845 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1847 OPENSSL_DECLARE_GLOBAL(int,foo);
1848 #define foo OPENSSL_GLOBAL_REF(foo)
1849 OPENSSL_DECLARE_GLOBAL(double,bar);
1850 #define bar OPENSSL_GLOBAL_REF(bar)
1852 The #defines are very important, and therefore so is including the
1853 header file everywhere where the defined globals are used.
1855 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1856 of ASN.1 items, but that structure is a bit different.
1858 The largest change is in util/mkdef.pl which has been enhanced with
1859 better and easier to understand logic to choose which symbols should
1860 go into the Windows .def files as well as a number of fixes and code
1861 cleanup (among others, algorithm keywords are now sorted
1862 lexicographically to avoid constant rewrites).
1865 *) In BN_div() keep a copy of the sign of 'num' before writing the
1866 result to 'rm' because if rm==num the value will be overwritten
1867 and produce the wrong result if 'num' is negative: this caused
1868 problems with BN_mod() and BN_nnmod().
1871 *) Function OCSP_request_verify(). This checks the signature on an
1872 OCSP request and verifies the signer certificate. The signer
1873 certificate is just checked for a generic purpose and OCSP request
1877 *) Add OCSP_check_validity() function to check the validity of OCSP
1878 responses. OCSP responses are prepared in real time and may only
1879 be a few seconds old. Simply checking that the current time lies
1880 between thisUpdate and nextUpdate max reject otherwise valid responses
1881 caused by either OCSP responder or client clock inaccuracy. Instead
1882 we allow thisUpdate and nextUpdate to fall within a certain period of
1883 the current time. The age of the response can also optionally be
1884 checked. Two new options -validity_period and -status_age added to
1888 *) If signature or public key algorithm is unrecognized print out its
1889 OID rather that just UNKNOWN.
1892 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1893 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1894 ID to be generated from the issuer certificate alone which can then be
1895 passed to OCSP_id_issuer_cmp().
1898 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1899 ASN1 modules to export functions returning ASN1_ITEM pointers
1900 instead of the ASN1_ITEM structures themselves. This adds several
1901 new macros which allow the underlying ASN1 function/structure to
1902 be accessed transparently. As a result code should not use ASN1_ITEM
1903 references directly (such as &X509_it) but instead use the relevant
1904 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1905 use of the new ASN1 code on platforms where exporting structures
1906 is problematical (for example in shared libraries) but exporting
1907 functions returning pointers to structures is not.
1910 *) Add support for overriding the generation of SSL/TLS session IDs.
1911 These callbacks can be registered either in an SSL_CTX or per SSL.
1912 The purpose of this is to allow applications to control, if they wish,
1913 the arbitrary values chosen for use as session IDs, particularly as it
1914 can be useful for session caching in multiple-server environments. A
1915 command-line switch for testing this (and any client code that wishes
1916 to use such a feature) has been added to "s_server".
1917 [Geoff Thorpe, Lutz Jaenicke]
1919 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
1920 of the form '#if defined(...) || defined(...) || ...' and
1921 '#if !defined(...) && !defined(...) && ...'. This also avoids
1922 the growing number of special cases it was previously handling.
1925 *) Make all configuration macros available for application by making
1926 sure they are available in opensslconf.h, by giving them names starting
1927 with "OPENSSL_" to avoid conflicts with other packages and by making
1928 sure e_os2.h will cover all platform-specific cases together with
1930 Additionally, it is now possible to define configuration/platform-
1931 specific names (called "system identities"). In the C code, these
1932 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1933 macro with the name beginning with "OPENSSL_SYS_", which is determined
1934 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1938 *) New option -set_serial to 'req' and 'x509' this allows the serial
1939 number to use to be specified on the command line. Previously self
1940 signed certificates were hard coded with serial number 0 and the
1941 CA options of 'x509' had to use a serial number in a file which was
1945 *) New options to 'ca' utility to support V2 CRL entry extensions.
1946 Currently CRL reason, invalidity date and hold instruction are
1947 supported. Add new CRL extensions to V3 code and some new objects.
1950 *) New function EVP_CIPHER_CTX_set_padding() this is used to
1951 disable standard block padding (aka PKCS#5 padding) in the EVP
1952 API, which was previously mandatory. This means that the data is
1953 not padded in any way and so the total length much be a multiple
1954 of the block size, otherwise an error occurs.
1957 *) Initial (incomplete) OCSP SSL support.
1960 *) New function OCSP_parse_url(). This splits up a URL into its host,
1961 port and path components: primarily to parse OCSP URLs. New -url
1962 option to ocsp utility.
1965 *) New nonce behavior. The return value of OCSP_check_nonce() now
1966 reflects the various checks performed. Applications can decide
1967 whether to tolerate certain situations such as an absent nonce
1968 in a response when one was present in a request: the ocsp application
1969 just prints out a warning. New function OCSP_add1_basic_nonce()
1970 this is to allow responders to include a nonce in a response even if
1971 the request is nonce-less.
1974 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1975 skipped when using openssl x509 multiple times on a single input file,
1976 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1979 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1980 set string type: to handle setting ASN1_TIME structures. Fix ca
1981 utility to correctly initialize revocation date of CRLs.
1984 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1985 the clients preferred ciphersuites and rather use its own preferences.
1986 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1987 Internet Explorer by ensuring unchanged hash method during stepup.
1988 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1991 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1992 to aes and add a new 'exist' option to print out symbols that don't
1996 *) Additional options to ocsp utility to allow flags to be set and
1997 additional certificates supplied.
2000 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
2001 OCSP client a number of certificate to only verify the response
2005 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
2006 handle the new API. Currently only ECB, CBC modes supported. Add new
2009 Add TLS AES ciphersuites as described in RFC3268, "Advanced
2010 Encryption Standard (AES) Ciphersuites for Transport Layer
2011 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
2012 not enabled by default and were not part of the "ALL" ciphersuite
2013 alias because they were not yet official; they could be
2014 explicitly requested by specifying the "AESdraft" ciphersuite
2015 group alias. In the final release of OpenSSL 0.9.7, the group
2016 alias is called "AES" and is part of "ALL".)
2017 [Ben Laurie, Steve Henson, Bodo Moeller]
2019 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
2020 request to response.
2023 *) Functions for OCSP responders. OCSP_request_onereq_count(),
2024 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
2025 extract information from a certificate request. OCSP_response_create()
2026 creates a response and optionally adds a basic response structure.
2027 OCSP_basic_add1_status() adds a complete single response to a basic
2028 response and returns the OCSP_SINGLERESP structure just added (to allow
2029 extensions to be included for example). OCSP_basic_add1_cert() adds a
2030 certificate to a basic response and OCSP_basic_sign() signs a basic
2031 response with various flags. New helper functions ASN1_TIME_check()
2032 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
2033 (converts ASN1_TIME to GeneralizedTime).
2036 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
2037 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
2038 structure from a certificate. X509_pubkey_digest() digests the public_key
2039 contents: this is used in various key identifiers.
2042 *) Make sk_sort() tolerate a NULL argument.
2043 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
2045 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
2046 passed by the function are trusted implicitly. If any of them signed the
2047 response then it is assumed to be valid and is not verified.
2050 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
2051 to data. This was previously part of the PKCS7 ASN1 code. This
2052 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
2053 [Steve Henson, reported by Kenneth R. Robinette
2054 <support@securenetterm.com>]
2056 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
2057 routines: without these tracing memory leaks is very painful.
2058 Fix leaks in PKCS12 and PKCS7 routines.
2061 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
2062 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
2063 effectively meant GeneralizedTime would never be used. Now it
2064 is initialised to -1 but X509_time_adj() now has to check the value
2065 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
2066 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
2067 [Steve Henson, reported by Kenneth R. Robinette
2068 <support@securenetterm.com>]
2070 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
2071 result in a zero length in the ASN1_INTEGER structure which was
2072 not consistent with the structure when d2i_ASN1_INTEGER() was used
2073 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
2074 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
2075 where it did not print out a minus for negative ASN1_INTEGER.
2078 *) Add summary printout to ocsp utility. The various functions which
2079 convert status values to strings have been renamed to:
2080 OCSP_response_status_str(), OCSP_cert_status_str() and
2081 OCSP_crl_reason_str() and are no longer static. New options
2082 to verify nonce values and to disable verification. OCSP response
2083 printout format cleaned up.
2086 *) Add additional OCSP certificate checks. These are those specified
2087 in RFC2560. This consists of two separate checks: the CA of the
2088 certificate being checked must either be the OCSP signer certificate
2089 or the issuer of the OCSP signer certificate. In the latter case the
2090 OCSP signer certificate must contain the OCSP signing extended key
2091 usage. This check is performed by attempting to match the OCSP
2092 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
2093 in the OCSP_CERTID structures of the response.
2096 *) Initial OCSP certificate verification added to OCSP_basic_verify()
2097 and related routines. This uses the standard OpenSSL certificate
2098 verify routines to perform initial checks (just CA validity) and
2099 to obtain the certificate chain. Then additional checks will be
2100 performed on the chain. Currently the root CA is checked to see
2101 if it is explicitly trusted for OCSP signing. This is used to set
2102 a root CA as a global signing root: that is any certificate that
2103 chains to that CA is an acceptable OCSP signing certificate.
2106 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
2107 extensions from a separate configuration file.
2108 As when reading extensions from the main configuration file,
2109 the '-extensions ...' option may be used for specifying the
2111 [Massimiliano Pala <madwolf@comune.modena.it>]
2113 *) New OCSP utility. Allows OCSP requests to be generated or
2114 read. The request can be sent to a responder and the output
2115 parsed, outputed or printed in text form. Not complete yet:
2116 still needs to check the OCSP response validity.
2119 *) New subcommands for 'openssl ca':
2120 'openssl ca -status <serial>' prints the status of the cert with
2121 the given serial number (according to the index file).
2122 'openssl ca -updatedb' updates the expiry status of certificates
2124 [Massimiliano Pala <madwolf@comune.modena.it>]
2126 *) New '-newreq-nodes' command option to CA.pl. This is like
2127 '-newreq', but calls 'openssl req' with the '-nodes' option
2128 so that the resulting key is not encrypted.
2129 [Damien Miller <djm@mindrot.org>]
2131 *) New configuration for the GNU Hurd.
2132 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
2134 *) Initial code to implement OCSP basic response verify. This
2135 is currently incomplete. Currently just finds the signer's
2136 certificate and verifies the signature on the response.
2139 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
2140 value of OPENSSLDIR. This is available via the new '-d' option
2141 to 'openssl version', and is also included in 'openssl version -a'.
2144 *) Allowing defining memory allocation callbacks that will be given
2145 file name and line number information in additional arguments
2146 (a const char* and an int). The basic functionality remains, as
2147 well as the original possibility to just replace malloc(),
2148 realloc() and free() by functions that do not know about these
2149 additional arguments. To register and find out the current
2150 settings for extended allocation functions, the following
2151 functions are provided:
2153 CRYPTO_set_mem_ex_functions
2154 CRYPTO_set_locked_mem_ex_functions
2155 CRYPTO_get_mem_ex_functions
2156 CRYPTO_get_locked_mem_ex_functions
2158 These work the same way as CRYPTO_set_mem_functions and friends.
2159 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
2160 extended allocation function is enabled.
2161 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
2162 a conventional allocation function is enabled.
2163 [Richard Levitte, Bodo Moeller]
2165 *) Finish off removing the remaining LHASH function pointer casts.
2166 There should no longer be any prototype-casting required when using
2167 the LHASH abstraction, and any casts that remain are "bugs". See
2168 the callback types and macros at the head of lhash.h for details
2169 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
2172 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
2173 If /dev/[u]random devices are not available or do not return enough
2174 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
2176 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
2177 /etc/entropy will be queried once each in this sequence, quering stops
2178 when enough entropy was collected without querying more sockets.
2181 *) Change the Unix RAND_poll() variant to be able to poll several
2182 random devices, as specified by DEVRANDOM, until a sufficient amount
2183 of data has been collected. We spend at most 10 ms on each file
2184 (select timeout) and read in non-blocking mode. DEVRANDOM now
2185 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
2186 (previously it was just the string "/dev/urandom"), so on typical
2187 platforms the 10 ms delay will never occur.
2188 Also separate out the Unix variant to its own file, rand_unix.c.
2189 For VMS, there's a currently-empty rand_vms.c.
2192 *) Move OCSP client related routines to ocsp_cl.c. These
2193 provide utility functions which an application needing
2194 to issue a request to an OCSP responder and analyse the
2195 response will typically need: as opposed to those which an
2196 OCSP responder itself would need which will be added later.
2198 OCSP_request_sign() signs an OCSP request with an API similar
2199 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
2200 response. OCSP_response_get1_basic() extracts basic response
2201 from response. OCSP_resp_find_status(): finds and extracts status
2202 information from an OCSP_CERTID structure (which will be created
2203 when the request structure is built). These are built from lower
2204 level functions which work on OCSP_SINGLERESP structures but
2205 wont normally be used unless the application wishes to examine
2206 extensions in the OCSP response for example.
2208 Replace nonce routines with a pair of functions.
2209 OCSP_request_add1_nonce() adds a nonce value and optionally
2210 generates a random value. OCSP_check_nonce() checks the
2211 validity of the nonce in an OCSP response.
2214 *) Change function OCSP_request_add() to OCSP_request_add0_id().
2215 This doesn't copy the supplied OCSP_CERTID and avoids the
2216 need to free up the newly created id. Change return type
2217 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
2218 This can then be used to add extensions to the request.
2219 Deleted OCSP_request_new(), since most of its functionality
2220 is now in OCSP_REQUEST_new() (and the case insensitive name
2221 clash) apart from the ability to set the request name which
2222 will be added elsewhere.
2225 *) Update OCSP API. Remove obsolete extensions argument from
2226 various functions. Extensions are now handled using the new
2227 OCSP extension code. New simple OCSP HTTP function which
2228 can be used to send requests and parse the response.
2231 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
2232 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
2233 uses the special reorder version of SET OF to sort the attributes
2234 and reorder them to match the encoded order. This resolves a long
2235 standing problem: a verify on a PKCS7 structure just after signing
2236 it used to fail because the attribute order did not match the
2237 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
2238 it uses the received order. This is necessary to tolerate some broken
2239 software that does not order SET OF. This is handled by encoding
2240 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
2241 to produce the required SET OF.
2244 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
2245 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
2246 files to get correct declarations of the ASN.1 item variables.
2249 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
2250 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
2251 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
2252 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
2253 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
2254 ASN1_ITEM and no wrapper functions.
2257 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
2258 replace the old function pointer based I/O routines. Change most of
2259 the *_d2i_bio() and *_d2i_fp() functions to use these.
2262 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
2263 lines, recognice more "algorithms" that can be deselected, and make
2264 it complain about algorithm deselection that isn't recognised.
2267 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
2268 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
2269 to use new functions. Add NO_ASN1_OLD which can be set to remove
2270 some old style ASN1 functions: this can be used to determine if old
2271 code will still work when these eventually go away.
2274 *) New extension functions for OCSP structures, these follow the
2275 same conventions as certificates and CRLs.
2278 *) New function X509V3_add1_i2d(). This automatically encodes and
2279 adds an extension. Its behaviour can be customised with various
2280 flags to append, replace or delete. Various wrappers added for
2281 certifcates and CRLs.
2284 *) Fix to avoid calling the underlying ASN1 print routine when
2285 an extension cannot be parsed. Correct a typo in the
2286 OCSP_SERVICELOC extension. Tidy up print OCSP format.
2289 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
2290 entries for variables.
2293 *) Add functionality to apps/openssl.c for detecting locking
2294 problems: As the program is single-threaded, all we have
2295 to do is register a locking callback using an array for
2296 storing which locks are currently held by the program.
2299 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
2300 SSL_get_ex_data_X509_STORE_idx(), which is used in
2301 ssl_verify_cert_chain() and thus can be called at any time
2302 during TLS/SSL handshakes so that thread-safety is essential.
2303 Unfortunately, the ex_data design is not at all suited
2304 for multi-threaded use, so it probably should be abolished.
2307 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
2308 [Broadcom, tweaked and integrated by Geoff Thorpe]
2310 *) Move common extension printing code to new function
2311 X509V3_print_extensions(). Reorganise OCSP print routines and
2312 implement some needed OCSP ASN1 functions. Add OCSP extensions.
2315 *) New function X509_signature_print() to remove duplication in some
2319 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
2320 set (this was treated exactly the same as SET OF previously). This
2321 is used to reorder the STACK representing the structure to match the
2322 encoding. This will be used to get round a problem where a PKCS7
2323 structure which was signed could not be verified because the STACK
2324 order did not reflect the encoded order.
2327 *) Reimplement the OCSP ASN1 module using the new code.
2330 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
2331 for its ASN1 operations. The old style function pointers still exist
2332 for now but they will eventually go away.
2335 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
2336 completely replaces the old ASN1 functionality with a table driven
2337 encoder and decoder which interprets an ASN1_ITEM structure describing
2338 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
2339 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
2340 has also been converted to the new form.
2343 *) Change BN_mod_exp_recp so that negative moduli are tolerated
2344 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
2345 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
2346 for negative moduli.
2349 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
2350 of not touching the result's sign bit.
2353 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
2357 *) Changed the LHASH code to use prototypes for callbacks, and created
2358 macros to declare and implement thin (optionally static) functions
2359 that provide type-safety and avoid function pointer casting for the
2360 type-specific callbacks.
2363 *) Added Kerberos Cipher Suites to be used with TLS, as written in
2365 [Veers Staats <staatsvr@asc.hpc.mil>,
2366 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
2368 *) Reformat the FAQ so the different questions and answers can be divided
2369 in sections depending on the subject.
2372 *) Have the zlib compression code load ZLIB.DLL dynamically under
2376 *) New function BN_mod_sqrt for computing square roots modulo a prime
2377 (using the probabilistic Tonelli-Shanks algorithm unless
2378 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
2379 be handled deterministically).
2380 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2382 *) Make BN_mod_inverse faster by explicitly handling small quotients
2383 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
2384 512 bits], about 30% for larger ones [1024 or 2048 bits].)
2387 *) New function BN_kronecker.
2390 *) Fix BN_gcd so that it works on negative inputs; the result is
2391 positive unless both parameters are zero.
2392 Previously something reasonably close to an infinite loop was
2393 possible because numbers could be growing instead of shrinking
2394 in the implementation of Euclid's algorithm.
2397 *) Fix BN_is_word() and BN_is_one() macros to take into account the
2398 sign of the number in question.
2400 Fix BN_is_word(a,w) to work correctly for w == 0.
2402 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
2403 because its test if the absolute value of 'a' equals 'w'.
2404 Note that BN_abs_is_word does *not* handle w == 0 reliably;
2405 it exists mostly for use in the implementations of BN_is_zero(),
2406 BN_is_one(), and BN_is_word().
2409 *) New function BN_swap.
2412 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
2413 the exponentiation functions are more likely to produce reasonable
2414 results on negative inputs.
2417 *) Change BN_mod_mul so that the result is always non-negative.
2418 Previously, it could be negative if one of the factors was negative;
2419 I don't think anyone really wanted that behaviour.
2422 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
2423 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
2424 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
2425 and add new functions:
2434 BN_mod_lshift1_quick
2438 These functions always generate non-negative results.
2440 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
2441 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
2443 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2444 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2445 be reduced modulo m.
2446 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2449 The following entry accidentily appeared in the CHANGES file
2450 distributed with OpenSSL 0.9.7. The modifications described in
2451 it do *not* apply to OpenSSL 0.9.7.
2453 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2454 was actually never needed) and in BN_mul(). The removal in BN_mul()
2455 required a small change in bn_mul_part_recursive() and the addition
2456 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2457 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2458 bn_sub_words() and bn_add_words() except they take arrays with
2463 *) In 'openssl passwd', verify passwords read from the terminal
2464 unless the '-salt' option is used (which usually means that
2465 verification would just waste user's time since the resulting
2466 hash is going to be compared with some given password hash)
2467 or the new '-noverify' option is used.
2469 This is an incompatible change, but it does not affect
2470 non-interactive use of 'openssl passwd' (passwords on the command
2471 line, '-stdin' option, '-in ...' option) and thus should not
2475 *) Remove all references to RSAref, since there's no more need for it.
2478 *) Make DSO load along a path given through an environment variable
2479 (SHLIB_PATH) with shl_load().
2482 *) Constify the ENGINE code as a result of BIGNUM constification.
2483 Also constify the RSA code and most things related to it. In a
2484 few places, most notable in the depth of the ASN.1 code, ugly
2485 casts back to non-const were required (to be solved at a later
2489 *) Make it so the openssl application has all engines loaded by default.
2492 *) Constify the BIGNUM routines a little more.
2495 *) Add the following functions:
2497 ENGINE_load_cswift()
2499 ENGINE_load_atalla()
2501 ENGINE_load_builtin_engines()
2503 That way, an application can itself choose if external engines that
2504 are built-in in OpenSSL shall ever be used or not. The benefit is
2505 that applications won't have to be linked with libdl or other dso
2506 libraries unless it's really needed.
2508 Changed 'openssl engine' to load all engines on demand.
2509 Changed the engine header files to avoid the duplication of some
2510 declarations (they differed!).
2513 *) 'openssl engine' can now list capabilities.
2516 *) Better error reporting in 'openssl engine'.
2519 *) Never call load_dh_param(NULL) in s_server.
2522 *) Add engine application. It can currently list engines by name and
2523 identity, and test if they are actually available.
2526 *) Improve RPM specification file by forcing symbolic linking and making
2527 sure the installed documentation is also owned by root.root.
2528 [Damien Miller <djm@mindrot.org>]
2530 *) Give the OpenSSL applications more possibilities to make use of
2531 keys (public as well as private) handled by engines.
2534 *) Add OCSP code that comes from CertCo.
2537 *) Add VMS support for the Rijndael code.
2540 *) Added untested support for Nuron crypto accelerator.
2543 *) Add support for external cryptographic devices. This code was
2544 previously distributed separately as the "engine" branch.
2545 [Geoff Thorpe, Richard Levitte]
2547 *) Rework the filename-translation in the DSO code. It is now possible to
2548 have far greater control over how a "name" is turned into a filename
2549 depending on the operating environment and any oddities about the
2550 different shared library filenames on each system.
2553 *) Support threads on FreeBSD-elf in Configure.
2556 *) Fix for SHA1 assembly problem with MASM: it produces
2557 warnings about corrupt line number information when assembling
2558 with debugging information. This is caused by the overlapping
2560 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2563 NCONF_get_number() has no error checking at all. As a replacement,
2564 NCONF_get_number_e() is defined (_e for "error checking") and is
2565 promoted strongly. The old NCONF_get_number is kept around for
2566 binary backward compatibility.
2567 Make it possible for methods to load from something other than a BIO,
2568 by providing a function pointer that is given a name instead of a BIO.
2569 For example, this could be used to load configuration data from an
2573 *) Fix for non blocking accept BIOs. Added new I/O special reason
2574 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2575 with non blocking I/O was not possible because no retry code was
2576 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2580 *) Added the beginnings of Rijndael support.
2583 *) Fix for bug in DirectoryString mask setting. Add support for
2584 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2585 to allow certificate printing to more controllable, additional
2586 'certopt' option to 'x509' to allow new printing options to be
2590 *) Clean old EAY MD5 hack from e_os.h.
2593 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
2595 *) Fix various bugs revealed by running the NISCC test suite:
2597 Stop out of bounds reads in the ASN1 code when presented with
2598 invalid tags (CAN-2003-0543 and CAN-2003-0544).
2600 If verify callback ignores invalid public key errors don't try to check
2601 certificate signature with the NULL public key.
2605 *) Fix various bugs revealed by running the NISCC test suite:
2607 Stop out of bounds reads in the ASN1 code when presented with
2608 invalid tags (CAN-2003-0543 and CAN-2003-0544).
2610 If verify callback ignores invalid public key errors don't try to check
2611 certificate signature with the NULL public key.
2615 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2616 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2620 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2621 extra data after the compression methods not only for TLS 1.0
2622 but also for SSL 3.0 (as required by the specification).
2623 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2625 *) Change X509_certificate_type() to mark the key as exported/exportable
2626 when it's 512 *bits* long, not 512 bytes.
2629 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
2631 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2632 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2633 a protocol version number mismatch like a decryption error
2634 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2637 *) Turn on RSA blinding by default in the default implementation
2638 to avoid a timing attack. Applications that don't want it can call
2639 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2640 They would be ill-advised to do so in most cases.
2641 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2643 *) Change RSA blinding code so that it works when the PRNG is not
2644 seeded (in this case, the secret RSA exponent is abused as
2645 an unpredictable seed -- if it is not unpredictable, there
2646 is no point in blinding anyway). Make RSA blinding thread-safe
2647 by remembering the creator's thread ID in rsa->blinding and
2648 having all other threads use local one-time blinding factors
2649 (this requires more computation than sharing rsa->blinding, but
2650 avoids excessive locking; and if an RSA object is not shared
2651 between threads, blinding will still be very fast).
2654 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
2656 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2657 via timing by performing a MAC computation even if incorrrect
2658 block cipher padding has been found. This is a countermeasure
2659 against active attacks where the attacker has to distinguish
2660 between bad padding and a MAC verification error. (CAN-2003-0078)
2662 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2663 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2664 Martin Vuagnoux (EPFL, Ilion)]
2666 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
2668 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
2669 memory from it's contents. This is done with a counter that will
2670 place alternating values in each byte. This can be used to solve
2671 two issues: 1) the removal of calls to memset() by highly optimizing
2672 compilers, and 2) cleansing with other values than 0, since those can
2673 be read through on certain media, for example a swap space on disk.
2676 *) Bugfix: client side session caching did not work with external caching,
2677 because the session->cipher setting was not restored when reloading
2678 from the external cache. This problem was masked, when
2679 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
2680 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
2683 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
2684 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
2685 [Zeev Lieber <zeev-l@yahoo.com>]
2687 *) Undo an undocumented change introduced in 0.9.6e which caused
2688 repeated calls to OpenSSL_add_all_ciphers() and
2689 OpenSSL_add_all_digests() to be ignored, even after calling
2693 *) Change the default configuration reader to deal with last line not
2694 being properly terminated.
2697 *) Change X509_NAME_cmp() so it applies the special rules on handling
2698 DN values that are of type PrintableString, as well as RDNs of type
2699 emailAddress where the value has the type ia5String.
2700 [stefank@valicert.com via Richard Levitte]
2702 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
2703 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
2704 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
2705 the bitwise-OR of the two for use by the majority of applications
2706 wanting this behaviour, and update the docs. The documented
2707 behaviour and actual behaviour were inconsistent and had been
2708 changing anyway, so this is more a bug-fix than a behavioural
2710 [Geoff Thorpe, diagnosed by Nadav Har'El]
2712 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
2713 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
2716 *) Fix initialization code race conditions in
2717 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
2718 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
2719 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
2720 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
2721 ssl2_get_cipher_by_char(),
2722 ssl3_get_cipher_by_char().
2723 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
2725 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
2726 the cached sessions are flushed, as the remove_cb() might use ex_data
2727 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
2728 (see [openssl.org #212]).
2729 [Geoff Thorpe, Lutz Jaenicke]
2731 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
2732 length, instead of the encoding length to d2i_ASN1_OBJECT.
2735 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
2737 *) [In 0.9.6g-engine release:]
2738 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
2739 [Lynn Gazis <lgazis@rainbow.com>]
2741 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
2743 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
2744 and get fix the header length calculation.
2745 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
2746 Alon Kantor <alonk@checkpoint.com> (and others),
2749 *) Use proper error handling instead of 'assertions' in buffer
2750 overflow checks added in 0.9.6e. This prevents DoS (the
2751 assertions could call abort()).
2752 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
2754 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
2756 *) Add various sanity checks to asn1_get_length() to reject
2757 the ASN1 length bytes if they exceed sizeof(long), will appear
2758 negative or the content length exceeds the length of the
2760 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2762 *) Fix cipher selection routines: ciphers without encryption had no flags
2763 for the cipher strength set and where therefore not handled correctly
2764 by the selection routines (PR #130).
2767 *) Fix EVP_dsa_sha macro.
2771 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
2772 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
2773 that was added in OpenSSL 0.9.6d.
2775 As the countermeasure turned out to be incompatible with some
2776 broken SSL implementations, the new option is part of SSL_OP_ALL.
2777 SSL_OP_ALL is usually employed when compatibility with weird SSL
2778 implementations is desired (e.g. '-bugs' option to 's_client' and
2779 's_server'), so the new option is automatically set in many
2783 *) Changes in security patch:
2785 Changes marked "(CHATS)" were sponsored by the Defense Advanced
2786 Research Projects Agency (DARPA) and Air Force Research Laboratory,
2787 Air Force Materiel Command, USAF, under agreement number
2790 *) Add various sanity checks to asn1_get_length() to reject
2791 the ASN1 length bytes if they exceed sizeof(long), will appear
2792 negative or the content length exceeds the length of the
2793 supplied buffer. (CAN-2002-0659)
2794 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2796 *) Assertions for various potential buffer overflows, not known to
2798 [Ben Laurie (CHATS)]
2800 *) Various temporary buffers to hold ASCII versions of integers were
2801 too small for 64 bit platforms. (CAN-2002-0655)
2802 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
2804 *) Remote buffer overflow in SSL3 protocol - an attacker could
2805 supply an oversized session ID to a client. (CAN-2002-0656)
2806 [Ben Laurie (CHATS)]
2808 *) Remote buffer overflow in SSL2 protocol - an attacker could
2809 supply an oversized client master key. (CAN-2002-0656)
2810 [Ben Laurie (CHATS)]
2812 Changes between 0.9.6c and 0.9.6d [9 May 2002]
2814 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
2815 encoded as NULL) with id-dsa-with-sha1.
2816 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
2818 *) Check various X509_...() return values in apps/req.c.
2819 [Nils Larsch <nla@trustcenter.de>]
2821 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
2822 an end-of-file condition would erronously be flagged, when the CRLF
2823 was just at the end of a processed block. The bug was discovered when
2824 processing data through a buffering memory BIO handing the data to a
2825 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
2826 <ptsekov@syntrex.com> and Nedelcho Stanev.
2829 *) Implement a countermeasure against a vulnerability recently found
2830 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
2831 before application data chunks to avoid the use of known IVs
2832 with data potentially chosen by the attacker.
2835 *) Fix length checks in ssl3_get_client_hello().
2838 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
2839 to prevent ssl3_read_internal() from incorrectly assuming that
2840 ssl3_read_bytes() found application data while handshake
2841 processing was enabled when in fact s->s3->in_read_app_data was
2842 merely automatically cleared during the initial handshake.
2843 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
2845 *) Fix object definitions for Private and Enterprise: they were not
2846 recognized in their shortname (=lowercase) representation. Extend
2847 obj_dat.pl to issue an error when using undefined keywords instead
2848 of silently ignoring the problem (Svenning Sorensen
2849 <sss@sss.dnsalias.net>).
2852 *) Fix DH_generate_parameters() so that it works for 'non-standard'
2853 generators, i.e. generators other than 2 and 5. (Previously, the
2854 code did not properly initialise the 'add' and 'rem' values to
2855 BN_generate_prime().)
2857 In the new general case, we do not insist that 'generator' is
2858 actually a primitive root: This requirement is rather pointless;
2859 a generator of the order-q subgroup is just as good, if not
2863 *) Map new X509 verification errors to alerts. Discovered and submitted by
2864 Tom Wu <tom@arcot.com>.
2867 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
2868 returning non-zero before the data has been completely received
2869 when using non-blocking I/O.
2870 [Bodo Moeller; problem pointed out by John Hughes]
2872 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
2873 [Ben Laurie, Lutz Jaenicke]
2875 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
2876 Yoram Zahavi <YoramZ@gilian.com>).
2879 *) Add information about CygWin 1.3 and on, and preserve proper
2880 configuration for the versions before that.
2881 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2883 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
2884 check whether we deal with a copy of a session and do not delete from
2885 the cache in this case. Problem reported by "Izhar Shoshani Levi"
2886 <izhar@checkpoint.com>.
2889 *) Do not store session data into the internal session cache, if it
2890 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
2891 flag is set). Proposed by Aslam <aslam@funk.com>.
2894 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
2898 *) [In 0.9.6d-engine release:]
2899 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2900 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2902 *) Add the configuration target linux-s390x.
2903 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
2905 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
2906 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
2907 variable as an indication that a ClientHello message has been
2908 received. As the flag value will be lost between multiple
2909 invocations of ssl3_accept when using non-blocking I/O, the
2910 function may not be aware that a handshake has actually taken
2911 place, thus preventing a new session from being added to the
2914 To avoid this problem, we now set s->new_session to 2 instead of
2915 using a local variable.
2916 [Lutz Jaenicke, Bodo Moeller]
2918 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
2919 if the SSL_R_LENGTH_MISMATCH error is detected.
2920 [Geoff Thorpe, Bodo Moeller]
2922 *) New 'shared_ldflag' column in Configure platform table.
2925 *) Fix EVP_CIPHER_mode macro.
2926 ["Dan S. Camper" <dan@bti.net>]
2928 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
2929 type, we must throw them away by setting rr->length to 0.
2930 [D P Chang <dpc@qualys.com>]
2932 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
2934 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
2935 <Dominikus.Scherkl@biodata.com>. (The previous implementation
2936 worked incorrectly for those cases where range = 10..._2 and
2937 3*range is two bits longer than range.)
2940 *) Only add signing time to PKCS7 structures if it is not already
2944 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
2945 OBJ_ld_ce should be OBJ_id_ce.
2946 Also some ip-pda OIDs in crypto/objects/objects.txt were
2947 incorrect (cf. RFC 3039).
2948 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
2950 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
2951 returns early because it has nothing to do.
2952 [Andy Schneider <andy.schneider@bjss.co.uk>]
2954 *) [In 0.9.6c-engine release:]
2955 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
2956 [Andy Schneider <andy.schneider@bjss.co.uk>]
2958 *) [In 0.9.6c-engine release:]
2959 Add support for Cryptographic Appliance's keyserver technology.
2960 (Use engine 'keyclient')
2961 [Cryptographic Appliances and Geoff Thorpe]
2963 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
2964 is called via tools/c89.sh because arguments have to be
2965 rearranged (all '-L' options must appear before the first object
2967 [Richard Shapiro <rshapiro@abinitio.com>]
2969 *) [In 0.9.6c-engine release:]
2970 Add support for Broadcom crypto accelerator cards, backported
2972 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
2974 *) [In 0.9.6c-engine release:]
2975 Add support for SureWare crypto accelerator cards from
2976 Baltimore Technologies. (Use engine 'sureware')
2977 [Baltimore Technologies and Mark Cox]
2979 *) [In 0.9.6c-engine release:]
2980 Add support for crypto accelerator cards from Accelerated
2981 Encryption Processing, www.aep.ie. (Use engine 'aep')
2982 [AEP Inc. and Mark Cox]
2984 *) Add a configuration entry for gcc on UnixWare.
2985 [Gary Benson <gbenson@redhat.com>]
2987 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
2988 messages are stored in a single piece (fixed-length part and
2989 variable-length part combined) and fix various bugs found on the way.
2992 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
2993 instead. BIO_gethostbyname() does not know what timeouts are
2994 appropriate, so entries would stay in cache even when they have
2996 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
2998 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
2999 faced with a pathologically small ClientHello fragment that does
3000 not contain client_version: Instead of aborting with an error,
3001 simply choose the highest available protocol version (i.e.,
3002 TLS 1.0 unless it is disabled). In practice, ClientHello
3003 messages are never sent like this, but this change gives us
3004 strictly correct behaviour at least for TLS.
3007 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
3008 never resets s->method to s->ctx->method when called from within
3009 one of the SSL handshake functions.
3010 [Bodo Moeller; problem pointed out by Niko Baric]
3012 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
3013 (sent using the client's version number) if client_version is
3014 smaller than the protocol version in use. Also change
3015 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
3016 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
3017 the client will at least see that alert.
3020 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
3024 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
3025 client receives HelloRequest while in a handshake.
3026 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
3028 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
3029 should end in 'break', not 'goto end' which circuments various
3030 cleanups done in state SSL_ST_OK. But session related stuff
3031 must be disabled for SSL_ST_OK in the case that we just sent a
3034 Also avoid some overhead by not calling ssl_init_wbio_buffer()
3035 before just sending a HelloRequest.
3036 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
3038 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
3039 reveal whether illegal block cipher padding was found or a MAC
3040 verification error occured. (Neither SSLerr() codes nor alerts
3041 are directly visible to potential attackers, but the information
3042 may leak via logfiles.)
3044 Similar changes are not required for the SSL 2.0 implementation
3045 because the number of padding bytes is sent in clear for SSL 2.0,
3046 and the extra bytes are just ignored. However ssl/s2_pkt.c
3047 failed to verify that the purported number of padding bytes is in
3051 *) Add OpenUNIX-8 support including shared libraries
3052 (Boyd Lynn Gerber <gerberb@zenez.com>).
3055 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
3056 'wristwatch attack' using huge encoding parameters (cf.
3057 James H. Manger's CRYPTO 2001 paper). Note that the
3058 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
3059 encoding parameters and hence was not vulnerable.
3062 *) BN_sqr() bug fix.
3063 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
3065 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
3066 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
3067 followed by modular reduction.
3068 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
3070 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
3071 equivalent based on BN_pseudo_rand() instead of BN_rand().
3074 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
3075 This function was broken, as the check for a new client hello message
3076 to handle SGC did not allow these large messages.
3077 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
3080 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
3083 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
3084 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
3087 *) Rework the configuration and shared library support for Tru64 Unix.
3088 The configuration part makes use of modern compiler features and
3089 still retains old compiler behavior for those that run older versions
3090 of the OS. The shared library support part includes a variant that
3091 uses the RPATH feature, and is available through the special
3092 configuration target "alpha-cc-rpath", which will never be selected
3094 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
3096 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
3097 with the same message size as in ssl3_get_certificate_request().
3098 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
3099 messages might inadvertently be reject as too long.
3100 [Petr Lampa <lampa@fee.vutbr.cz>]
3102 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
3105 *) Modified SSL library such that the verify_callback that has been set
3106 specificly for an SSL object with SSL_set_verify() is actually being
3107 used. Before the change, a verify_callback set with this function was
3108 ignored and the verify_callback() set in the SSL_CTX at the time of
3109 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
3110 to allow the necessary settings.
3113 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
3114 explicitly to NULL, as at least on Solaris 8 this seems not always to be
3115 done automatically (in contradiction to the requirements of the C
3116 standard). This made problems when used from OpenSSH.
3119 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
3120 dh->length and always used
3122 BN_rand_range(priv_key, dh->p).
3124 BN_rand_range() is not necessary for Diffie-Hellman, and this
3125 specific range makes Diffie-Hellman unnecessarily inefficient if
3126 dh->length (recommended exponent length) is much smaller than the
3127 length of dh->p. We could use BN_rand_range() if the order of
3128 the subgroup was stored in the DH structure, but we only have
3133 BN_rand(priv_key, l, ...)
3135 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
3141 RSA_eay_public_encrypt
3142 RSA_eay_private_decrypt
3143 RSA_eay_private_encrypt (signing)
3144 RSA_eay_public_decrypt (signature verification)
3146 (default implementations for RSA_public_encrypt,
3147 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
3148 always reject numbers >= n.
3151 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
3152 to synchronize access to 'locking_thread'. This is necessary on
3153 systems where access to 'locking_thread' (an 'unsigned long'
3154 variable) is not atomic.
3157 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
3158 *before* setting the 'crypto_lock_rand' flag. The previous code had
3159 a race condition if 0 is a valid thread ID.
3160 [Travis Vitek <vitek@roguewave.com>]
3162 *) Add support for shared libraries under Irix.
3163 [Albert Chin-A-Young <china@thewrittenword.com>]
3165 *) Add configuration option to build on Linux on both big-endian and
3167 [Ralf Baechle <ralf@uni-koblenz.de>]
3169 *) Add the possibility to create shared libraries on HP-UX.
3172 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
3174 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
3175 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
3176 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
3177 PRNG state recovery was possible based on the output of
3178 one PRNG request appropriately sized to gain knowledge on
3179 'md' followed by enough consecutive 1-byte PRNG requests
3180 to traverse all of 'state'.
3182 1. When updating 'md_local' (the current thread's copy of 'md')
3183 during PRNG output generation, hash all of the previous
3184 'md_local' value, not just the half used for PRNG output.
3186 2. Make the number of bytes from 'state' included into the hash
3187 independent from the number of PRNG bytes requested.
3189 The first measure alone would be sufficient to avoid
3190 Markku-Juhani's attack. (Actually it had never occurred
3191 to me that the half of 'md_local' used for chaining was the
3192 half from which PRNG output bytes were taken -- I had always
3193 assumed that the secret half would be used.) The second
3194 measure makes sure that additional data from 'state' is never
3195 mixed into 'md_local' in small portions; this heuristically
3196 further strengthens the PRNG.
3199 *) Fix crypto/bn/asm/mips3.s.
3202 *) When only the key is given to "enc", the IV is undefined. Print out
3203 an error message in this case.
3206 *) Handle special case when X509_NAME is empty in X509 printing routines.
3209 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
3210 positive and less than q.
3213 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
3214 used: it isn't thread safe and the add_lock_callback should handle
3216 [Paul Rose <Paul.Rose@bridge.com>]
3218 *) Verify that incoming data obeys the block size in
3219 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
3223 [Ulf Möller, Bodo Möller]
3225 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
3226 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
3227 when fixing the server behaviour for backwards-compatible 'client
3228 hello' messages. (Note that the attack is impractical against
3229 SSL 3.0 and TLS 1.0 anyway because length and version checking
3230 means that the probability of guessing a valid ciphertext is
3231 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
3234 Before 0.9.5, the countermeasure (hide the error by generating a
3235 random 'decryption result') did not work properly because
3236 ERR_clear_error() was missing, meaning that SSL_get_error() would
3237 detect the supposedly ignored error.
3239 Both problems are now fixed.
3242 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
3243 (previously it was 1024).
3246 *) Fix for compatibility mode trust settings: ignore trust settings
3247 unless some valid trust or reject settings are present.
3250 *) Fix for blowfish EVP: its a variable length cipher.
3253 *) Fix various bugs related to DSA S/MIME verification. Handle missing
3254 parameters in DSA public key structures and return an error in the
3255 DSA routines if parameters are absent.
3258 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
3259 in the current directory if neither $RANDFILE nor $HOME was set.
3260 RAND_file_name() in 0.9.6a returned NULL in this case. This has
3261 caused some confusion to Windows users who haven't defined $HOME.
3262 Thus RAND_file_name() is changed again: e_os.h can define a
3263 DEFAULT_HOME, which will be used if $HOME is not set.
3264 For Windows, we use "C:"; on other platforms, we still require
3265 environment variables.
3267 *) Move 'if (!initialized) RAND_poll()' into regions protected by
3268 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
3269 having multiple threads call RAND_poll() concurrently.
3272 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
3273 combination of a flag and a thread ID variable.
3274 Otherwise while one thread is in ssleay_rand_bytes (which sets the
3275 flag), *other* threads can enter ssleay_add_bytes without obeying
3276 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
3277 that they do not hold after the first thread unsets add_do_not_lock).
3280 *) Change bctest again: '-x' expressions are not available in all
3284 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
3286 *) Fix a couple of memory leaks in PKCS7_dataDecode()
3287 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
3289 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
3290 the default extension for executables, if any. Also, make the perl
3291 scripts that use symlink() to test if it really exists and use "cp"
3292 if it doesn't. All this made OpenSSL compilable and installable in
3296 *) Fix for asn1_GetSequence() for indefinite length constructed data.
3297 If SEQUENCE is length is indefinite just set c->slen to the total
3298 amount of data available.
3299 [Steve Henson, reported by shige@FreeBSD.org]
3300 [This change does not apply to 0.9.7.]
3302 *) Change bctest to avoid here-documents inside command substitution
3303 (workaround for FreeBSD /bin/sh bug).
3304 For compatibility with Ultrix, avoid shell functions (introduced
3305 in the bctest version that searches along $PATH).
3308 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
3309 with des_encrypt() defined on some operating systems, like Solaris
3313 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
3314 On the Importance of Eliminating Errors in Cryptographic
3315 Computations, J. Cryptology 14 (2001) 2, 101-119,
3316 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
3319 *) MIPS assembler BIGNUM division bug fix.
3322 *) Disabled incorrect Alpha assembler code.
3325 *) Fix PKCS#7 decode routines so they correctly update the length
3326 after reading an EOC for the EXPLICIT tag.
3328 [This change does not apply to 0.9.7.]
3330 *) Fix bug in PKCS#12 key generation routines. This was triggered
3331 if a 3DES key was generated with a 0 initial byte. Include
3332 PKCS12_BROKEN_KEYGEN compilation option to retain the old
3333 (but broken) behaviour.
3336 *) Enhance bctest to search for a working bc along $PATH and print
3338 [Tim Rice <tim@multitalents.net> via Richard Levitte]
3340 *) Fix memory leaks in err.c: free err_data string if necessary;
3341 don't write to the wrong index in ERR_set_error_data.
3344 *) Implement ssl23_peek (analogous to ssl23_read), which previously
3348 *) Replace rdtsc with _emit statements for VC++ version 5.
3349 [Jeremy Cooper <jeremy@baymoo.org>]
3351 *) Make it possible to reuse SSLv2 sessions.
3354 *) In copy_email() check for >= 0 as a return value for
3355 X509_NAME_get_index_by_NID() since 0 is a valid index.
3356 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
3358 *) Avoid coredump with unsupported or invalid public keys by checking if
3359 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
3360 PKCS7_verify() fails with non detached data.
3363 *) Don't use getenv in library functions when run as setuid/setgid.
3364 New function OPENSSL_issetugid().
3367 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
3368 due to incorrect handling of multi-threading:
3370 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
3372 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
3374 3. Count how many times MemCheck_off() has been called so that
3375 nested use can be treated correctly. This also avoids
3376 inband-signalling in the previous code (which relied on the
3377 assumption that thread ID 0 is impossible).
3380 *) Add "-rand" option also to s_client and s_server.
3383 *) Fix CPU detection on Irix 6.x.
3384 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
3385 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3387 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
3390 [This change does not apply to 0.9.7.]
3392 *) Use the cached encoding of an X509_NAME structure rather than
3393 copying it. This is apparently the reason for the libsafe "errors"
3394 but the code is actually correct.
3397 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
3398 Bleichenbacher's DSA attack.
3399 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
3400 to be set and top=0 forces the highest bit to be set; top=-1 is new
3401 and leaves the highest bit random.
3402 [Ulf Moeller, Bodo Moeller]
3404 *) In the NCONF_...-based implementations for CONF_... queries
3405 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
3406 a temporary CONF structure with the data component set to NULL
3407 (which gives segmentation faults in lh_retrieve).
3408 Instead, use NULL for the CONF pointer in CONF_get_string and
3409 CONF_get_number (which may use environment variables) and directly
3410 return NULL from CONF_get_section.
3413 *) Fix potential buffer overrun for EBCDIC.
3416 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
3417 keyUsage if basicConstraints absent for a CA.
3420 *) Make SMIME_write_PKCS7() write mail header values with a format that
3421 is more generally accepted (no spaces before the semicolon), since
3422 some programs can't parse those values properly otherwise. Also make
3423 sure BIO's that break lines after each write do not create invalid
3427 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
3428 macros previously used would not encode an empty SEQUENCE OF
3429 and break the signature.
3431 [This change does not apply to 0.9.7.]
3433 *) Zero the premaster secret after deriving the master secret in
3437 *) Add some EVP_add_digest_alias registrations (as found in
3438 OpenSSL_add_all_digests()) to SSL_library_init()
3439 aka OpenSSL_add_ssl_algorithms(). This provides improved
3440 compatibility with peers using X.509 certificates
3441 with unconventional AlgorithmIdentifier OIDs.
3444 *) Fix for Irix with NO_ASM.
3445 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3447 *) ./config script fixes.
3448 [Ulf Moeller, Richard Levitte]
3450 *) Fix 'openssl passwd -1'.
3453 *) Change PKCS12_key_gen_asc() so it can cope with non null
3454 terminated strings whose length is passed in the passlen
3455 parameter, for example from PEM callbacks. This was done
3456 by adding an extra length parameter to asc2uni().
3457 [Steve Henson, reported by <oddissey@samsung.co.kr>]
3459 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
3460 call failed, free the DSA structure.
3463 *) Fix to uni2asc() to cope with zero length Unicode strings.
3464 These are present in some PKCS#12 files.
3467 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
3468 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
3469 when writing a 32767 byte record.
3470 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
3472 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
3473 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
3475 (RSA objects have a reference count access to which is protected
3476 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
3477 so they are meant to be shared between threads.)
3478 [Bodo Moeller, Geoff Thorpe; original patch submitted by
3479 "Reddie, Steven" <Steven.Reddie@ca.com>]
3481 *) Fix a deadlock in CRYPTO_mem_leaks().
3484 *) Use better test patterns in bntest.
3487 *) rand_win.c fix for Borland C.
3490 *) BN_rshift bugfix for n == 0.
3493 *) Add a 'bctest' script that checks for some known 'bc' bugs
3494 so that 'make test' does not abort just because 'bc' is broken.
3497 *) Store verify_result within SSL_SESSION also for client side to
3498 avoid potential security hole. (Re-used sessions on the client side
3499 always resulted in verify_result==X509_V_OK, not using the original
3500 result of the server certificate verification.)
3503 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
3504 SSL3_RT_APPLICATION_DATA, return 0.
3505 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
3509 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
3510 releases, have been re-implemented by renaming the previous
3511 implementations of ssl2_read and ssl3_read to ssl2_read_internal
3512 and ssl3_read_internal, respectively, and adding 'peek' parameters
3513 to them. The new ssl[23]_{read,peek} functions are calls to
3514 ssl[23]_read_internal with the 'peek' flag set appropriately.
3515 A 'peek' parameter has also been added to ssl3_read_bytes, which
3516 does the actual work for ssl3_read_internal.
3519 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
3520 the method-specific "init()" handler. Also clean up ex_data after
3521 calling the method-specific "finish()" handler. Previously, this was
3522 happening the other way round.
3525 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
3526 The previous value, 12, was not always sufficient for BN_mod_exp().
3529 *) Make sure that shared libraries get the internal name engine with
3530 the full version number and not just 0. This should mark the
3531 shared libraries as not backward compatible. Of course, this should
3532 be changed again when we can guarantee backward binary compatibility.
3535 *) Fix typo in get_cert_by_subject() in by_dir.c
3536 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
3538 *) Rework the system to generate shared libraries:
3540 - Make note of the expected extension for the shared libraries and
3541 if there is a need for symbolic links from for example libcrypto.so.0
3542 to libcrypto.so.0.9.7. There is extended info in Configure for
3545 - Make as few rebuilds of the shared libraries as possible.
3547 - Still avoid linking the OpenSSL programs with the shared libraries.
3549 - When installing, install the shared libraries separately from the
3553 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
3555 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
3556 and not in SSL_clear because the latter is also used by the
3557 accept/connect functions; previously, the settings made by
3558 SSL_set_read_ahead would be lost during the handshake.
3559 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
3561 *) Correct util/mkdef.pl to be selective about disabled algorithms.
3562 Previously, it would create entries for disableed algorithms no
3566 *) Added several new manual pages for SSL_* function.
3569 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
3571 *) In ssl23_get_client_hello, generate an error message when faced
3572 with an initial SSL 3.0/TLS record that is too small to contain the
3573 first two bytes of the ClientHello message, i.e. client_version.
3574 (Note that this is a pathologic case that probably has never happened
3575 in real life.) The previous approach was to use the version number
3576 from the record header as a substitute; but our protocol choice
3577 should not depend on that one because it is not authenticated
3578 by the Finished messages.
3581 *) More robust randomness gathering functions for Windows.
3582 [Jeffrey Altman <jaltman@columbia.edu>]
3584 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
3585 not set then we don't setup the error code for issuer check errors
3586 to avoid possibly overwriting other errors which the callback does
3587 handle. If an application does set the flag then we assume it knows
3588 what it is doing and can handle the new informational codes
3592 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
3593 a general "ANY" type, as such it should be able to decode anything
3594 including tagged types. However it didn't check the class so it would
3595 wrongly interpret tagged types in the same way as their universal
3596 counterpart and unknown types were just rejected. Changed so that the
3597 tagged and unknown types are handled in the same way as a SEQUENCE:
3598 that is the encoding is stored intact. There is also a new type
3599 "V_ASN1_OTHER" which is used when the class is not universal, in this
3600 case we have no idea what the actual type is so we just lump them all
3604 *) On VMS, stdout may very well lead to a file that is written to
3605 in a record-oriented fashion. That means that every write() will
3606 write a separate record, which will be read separately by the
3607 programs trying to read from it. This can be very confusing.
3609 The solution is to put a BIO filter in the way that will buffer
3610 text until a linefeed is reached, and then write everything a
3611 line at a time, so every record written will be an actual line,
3612 not chunks of lines and not (usually doesn't happen, but I've
3613 seen it once) several lines in one record. BIO_f_linebuffer() is
3616 Currently, it's a VMS-only method, because that's where it has
3617 been tested well enough.
3620 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
3621 it can return incorrect results.
3622 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
3623 but it was in 0.9.6-beta[12].)
3626 *) Disable the check for content being present when verifying detached
3627 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
3628 include zero length content when signing messages.
3631 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
3632 BIO_ctrl (for BIO pairs).
3635 *) Add DSO method for VMS.
3638 *) Bug fix: Montgomery multiplication could produce results with the
3642 *) Add RPM specification openssl.spec and modify it to build three
3643 packages. The default package contains applications, application
3644 documentation and run-time libraries. The devel package contains
3645 include files, static libraries and function documentation. The
3646 doc package contains the contents of the doc directory. The original
3647 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
3650 *) Add a large number of documentation files for many SSL routines.
3651 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3653 *) Add a configuration entry for Sony News 4.
3654 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
3656 *) Don't set the two most significant bits to one when generating a
3657 random number < q in the DSA library.
3660 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
3661 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
3662 the underlying transport is blocking) if a handshake took place.
3663 (The default behaviour is needed by applications such as s_client
3664 and s_server that use select() to determine when to use SSL_read;
3665 but for applications that know in advance when to expect data, it
3666 just makes things more complicated.)
3669 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
3673 *) Add a few more EBCDIC conditionals that make `req' and `x509'
3674 work better on such systems.
3675 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3677 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
3678 Update PKCS12_parse() so it copies the friendlyName and the
3679 keyid to the certificates aux info.
3682 *) Fix bug in PKCS7_verify() which caused an infinite loop
3683 if there was more than one signature.
3684 [Sven Uszpelkat <su@celocom.de>]
3686 *) Major change in util/mkdef.pl to include extra information
3687 about each symbol, as well as presentig variables as well
3688 as functions. This change means that there's n more need
3689 to rebuild the .num files when some algorithms are excluded.
3692 *) Allow the verify time to be set by an application,
3693 rather than always using the current time.
3696 *) Phase 2 verify code reorganisation. The certificate
3697 verify code now looks up an issuer certificate by a
3698 number of criteria: subject name, authority key id
3699 and key usage. It also verifies self signed certificates
3700 by the same criteria. The main comparison function is
3701 X509_check_issued() which performs these checks.
3703 Lot of changes were necessary in order to support this
3704 without completely rewriting the lookup code.
3706 Authority and subject key identifier are now cached.
3708 The LHASH 'certs' is X509_STORE has now been replaced
3709 by a STACK_OF(X509_OBJECT). This is mainly because an
3710 LHASH can't store or retrieve multiple objects with
3711 the same hash value.
3713 As a result various functions (which were all internal
3714 use only) have changed to handle the new X509_STORE
3715 structure. This will break anything that messed round
3716 with X509_STORE internally.
3718 The functions X509_STORE_add_cert() now checks for an
3719 exact match, rather than just subject name.
3721 The X509_STORE API doesn't directly support the retrieval
3722 of multiple certificates matching a given criteria, however
3723 this can be worked round by performing a lookup first
3724 (which will fill the cache with candidate certificates)
3725 and then examining the cache for matches. This is probably
3726 the best we can do without throwing out X509_LOOKUP
3727 entirely (maybe later...).
3729 The X509_VERIFY_CTX structure has been enhanced considerably.
3731 All certificate lookup operations now go via a get_issuer()
3732 callback. Although this currently uses an X509_STORE it
3733 can be replaced by custom lookups. This is a simple way
3734 to bypass the X509_STORE hackery necessary to make this
3735 work and makes it possible to use more efficient techniques
3736 in future. A very simple version which uses a simple
3737 STACK for its trusted certificate store is also provided
3738 using X509_STORE_CTX_trusted_stack().
3740 The verify_cb() and verify() callbacks now have equivalents
3741 in the X509_STORE_CTX structure.
3743 X509_STORE_CTX also has a 'flags' field which can be used
3744 to customise the verify behaviour.
3747 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
3748 excludes S/MIME capabilities.
3751 *) When a certificate request is read in keep a copy of the
3752 original encoding of the signed data and use it when outputing
3753 again. Signatures then use the original encoding rather than
3754 a decoded, encoded version which may cause problems if the
3755 request is improperly encoded.
3758 *) For consistency with other BIO_puts implementations, call
3759 buffer_write(b, ...) directly in buffer_puts instead of calling
3762 In BIO_puts, increment b->num_write as in BIO_write.
3763 [Peter.Sylvester@EdelWeb.fr]
3765 *) Fix BN_mul_word for the case where the word is 0. (We have to use
3766 BN_zero, we may not return a BIGNUM with an array consisting of
3770 *) Avoid calling abort() from within the library when problems are
3771 detected, except if preprocessor symbols have been defined
3772 (such as REF_CHECK, BN_DEBUG etc.).
3775 *) New openssl application 'rsautl'. This utility can be
3776 used for low level RSA operations. DER public key
3777 BIO/fp routines also added.
3780 *) New Configure entry and patches for compiling on QNX 4.
3781 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
3783 *) A demo state-machine implementation was sponsored by
3784 Nuron (http://www.nuron.com/) and is now available in
3785 demos/state_machine.
3788 *) New options added to the 'dgst' utility for signature
3789 generation and verification.
3792 *) Unrecognized PKCS#7 content types are now handled via a
3793 catch all ASN1_TYPE structure. This allows unsupported
3794 types to be stored as a "blob" and an application can
3795 encode and decode it manually.
3798 *) Fix various signed/unsigned issues to make a_strex.c
3800 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
3802 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
3803 length if passed a buffer. ASN1_INTEGER_to_BN failed
3804 if passed a NULL BN and its argument was negative.
3805 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
3807 *) Modification to PKCS#7 encoding routines to output definite
3808 length encoding. Since currently the whole structures are in
3809 memory there's not real point in using indefinite length
3810 constructed encoding. However if OpenSSL is compiled with
3811 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
3814 *) Added BIO_vprintf() and BIO_vsnprintf().
3817 *) Added more prefixes to parse for in the the strings written
3818 through a logging bio, to cover all the levels that are available
3819 through syslog. The prefixes are now:
3821 PANIC, EMERG, EMR => LOG_EMERG
3822 ALERT, ALR => LOG_ALERT
3823 CRIT, CRI => LOG_CRIT
3824 ERROR, ERR => LOG_ERR
3825 WARNING, WARN, WAR => LOG_WARNING
3826 NOTICE, NOTE, NOT => LOG_NOTICE
3827 INFO, INF => LOG_INFO
3828 DEBUG, DBG => LOG_DEBUG
3830 and as before, if none of those prefixes are present at the
3831 beginning of the string, LOG_ERR is chosen.
3833 On Win32, the LOG_* levels are mapped according to this:
3835 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
3836 LOG_WARNING => EVENTLOG_WARNING_TYPE
3837 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
3841 *) Made it possible to reconfigure with just the configuration
3842 argument "reconf" or "reconfigure". The command line arguments
3843 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
3844 and are retrieved from there when reconfiguring.
3848 [Assar Westerlund <assar@sics.se>, Richard Levitte]
3850 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
3853 *) The obj_dat.pl script was messing up the sorting of object
3854 names. The reason was that it compared the quoted version
3855 of strings as a result "OCSP" > "OCSP Signing" because
3856 " > SPACE. Changed script to store unquoted versions of
3857 names and add quotes on output. It was also omitting some
3858 names from the lookup table if they were given a default
3859 value (that is if SN is missing it is given the same
3860 value as LN and vice versa), these are now added on the
3861 grounds that if an object has a name we should be able to
3862 look it up. Finally added warning output when duplicate
3863 short or long names are found.
3866 *) Changes needed for Tandem NSK.
3867 [Scott Uroff <scott@xypro.com>]
3869 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
3870 RSA_padding_check_SSLv23(), special padding was never detected
3871 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
3872 version rollback attacks was not effective.
3874 In s23_clnt.c, don't use special rollback-attack detection padding
3875 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
3876 client; similarly, in s23_srvr.c, don't do the rollback check if
3877 SSL 2.0 is the only protocol enabled in the server.
3880 *) Make it possible to get hexdumps of unprintable data with 'openssl
3881 asn1parse'. By implication, the functions ASN1_parse_dump() and
3882 BIO_dump_indent() are added.
3885 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
3886 these print out strings and name structures based on various
3887 flags including RFC2253 support and proper handling of
3888 multibyte characters. Added options to the 'x509' utility
3889 to allow the various flags to be set.
3892 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
3893 Also change the functions X509_cmp_current_time() and
3894 X509_gmtime_adj() work with an ASN1_TIME structure,
3895 this will enable certificates using GeneralizedTime in validity
3896 dates to be checked.
3899 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
3900 negative public key encodings) on by default,
3901 NO_NEG_PUBKEY_BUG can be set to disable it.
3904 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
3905 content octets. An i2c_ASN1_OBJECT is unnecessary because
3906 the encoding can be trivially obtained from the structure.
3909 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
3910 not read locks (CRYPTO_r_[un]lock).
3913 *) A first attempt at creating official support for shared
3914 libraries through configuration. I've kept it so the
3915 default is static libraries only, and the OpenSSL programs
3916 are always statically linked for now, but there are
3917 preparations for dynamic linking in place.
3918 This has been tested on Linux and Tru64.
3921 *) Randomness polling function for Win9x, as described in:
3922 Peter Gutmann, Software Generation of Practically Strong
3926 *) Fix so PRNG is seeded in req if using an already existing
3930 *) New options to smime application. -inform and -outform
3931 allow alternative formats for the S/MIME message including
3932 PEM and DER. The -content option allows the content to be
3933 specified separately. This should allow things like Netscape
3934 form signing output easier to verify.
3937 *) Fix the ASN1 encoding of tags using the 'long form'.
3940 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
3941 STRING types. These convert content octets to and from the
3942 underlying type. The actual tag and length octets are
3943 already assumed to have been read in and checked. These
3944 are needed because all other string types have virtually
3945 identical handling apart from the tag. By having versions
3946 of the ASN1 functions that just operate on content octets
3947 IMPLICIT tagging can be handled properly. It also allows
3948 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
3949 and ASN1_INTEGER are identical apart from the tag.
3952 *) Change the handling of OID objects as follows:
3954 - New object identifiers are inserted in objects.txt, following
3955 the syntax given in objects.README.
3956 - objects.pl is used to process obj_mac.num and create a new
3958 - obj_dat.pl is used to create a new obj_dat.h, using the data in
3961 This is currently kind of a hack, and the perl code in objects.pl
3962 isn't very elegant, but it works as I intended. The simplest way
3963 to check that it worked correctly is to look in obj_dat.h and
3964 check the array nid_objs and make sure the objects haven't moved
3965 around (this is important!). Additions are OK, as well as
3966 consistent name changes.
3969 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
3972 *) Addition of the command line parameter '-rand file' to 'openssl req'.
3973 The given file adds to whatever has already been seeded into the
3974 random pool through the RANDFILE configuration file option or
3975 environment variable, or the default random state file.
3978 *) mkstack.pl now sorts each macro group into lexical order.
3979 Previously the output order depended on the order the files
3980 appeared in the directory, resulting in needless rewriting
3984 *) Patches to make OpenSSL compile under Win32 again. Mostly
3985 work arounds for the VC++ problem that it treats func() as
3986 func(void). Also stripped out the parts of mkdef.pl that
3987 added extra typesafe functions: these no longer exist.
3990 *) Reorganisation of the stack code. The macros are now all
3991 collected in safestack.h . Each macro is defined in terms of
3992 a "stack macro" of the form SKM_<name>(type, a, b). The
3993 DEBUG_SAFESTACK is now handled in terms of function casts,
3994 this has the advantage of retaining type safety without the
3995 use of additional functions. If DEBUG_SAFESTACK is not defined
3996 then the non typesafe macros are used instead. Also modified the
3997 mkstack.pl script to handle the new form. Needs testing to see
3998 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
3999 the default if no major problems. Similar behaviour for ASN1_SET_OF
4000 and PKCS12_STACK_OF.
4003 *) When some versions of IIS use the 'NET' form of private key the
4004 key derivation algorithm is different. Normally MD5(password) is
4005 used as a 128 bit RC4 key. In the modified case
4006 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
4007 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
4008 as the old Netscape_RSA functions except they have an additional
4009 'sgckey' parameter which uses the modified algorithm. Also added
4010 an -sgckey command line option to the rsa utility. Thanks to
4011 Adrian Peck <bertie@ncipher.com> for posting details of the modified
4012 algorithm to openssl-dev.
4015 *) The evp_local.h macros were using 'c.##kname' which resulted in
4016 invalid expansion on some systems (SCO 5.0.5 for example).
4017 Corrected to 'c.kname'.
4018 [Phillip Porch <root@theporch.com>]
4020 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
4021 a STACK of email addresses from a certificate or request, these look
4022 in the subject name and the subject alternative name extensions and
4023 omit any duplicate addresses.
4026 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
4027 This makes DSA verification about 2 % faster.
4030 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
4031 (meaning that now 2^5 values will be precomputed, which is only 4 KB
4032 plus overhead for 1024 bit moduli).
4033 This makes exponentiations about 0.5 % faster for 1024 bit
4034 exponents (as measured by "openssl speed rsa2048").
4037 *) Rename memory handling macros to avoid conflicts with other
4039 Malloc => OPENSSL_malloc
4040 Malloc_locked => OPENSSL_malloc_locked
4041 Realloc => OPENSSL_realloc
4042 Free => OPENSSL_free
4045 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
4046 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
4049 *) CygWin32 support.
4050 [John Jarvie <jjarvie@newsguy.com>]
4052 *) The type-safe stack code has been rejigged. It is now only compiled
4053 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
4054 by default all type-specific stack functions are "#define"d back to
4055 standard stack functions. This results in more streamlined output
4056 but retains the type-safety checking possibilities of the original
4060 *) The STACK code has been cleaned up, and certain type declarations
4061 that didn't make a lot of sense have been brought in line. This has
4062 also involved a cleanup of sorts in safestack.h to more correctly
4063 map type-safe stack functions onto their plain stack counterparts.
4064 This work has also resulted in a variety of "const"ifications of
4065 lots of the code, especially "_cmp" operations which should normally
4066 be prototyped with "const" parameters anyway.
4069 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
4070 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
4071 (The PRNG state consists of two parts, the large pool 'state' and 'md',
4072 where all of 'md' is used each time the PRNG is used, but 'state'
4073 is used only indexed by a cyclic counter. As entropy may not be
4074 well distributed from the beginning, 'md' is important as a
4075 chaining variable. However, the output function chains only half
4076 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
4077 all of 'md', and seeding with STATE_SIZE dummy bytes will result
4078 in all of 'state' being rewritten, with the new values depending
4079 on virtually all of 'md'. This overcomes the 80 bit limitation.)
4082 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
4083 the handshake is continued after ssl_verify_cert_chain();
4084 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
4085 can lead to 'unexplainable' connection aborts later.
4086 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
4088 *) Major EVP API cipher revision.
4089 Add hooks for extra EVP features. This allows various cipher
4090 parameters to be set in the EVP interface. Support added for variable
4091 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
4092 setting of RC2 and RC5 parameters.
4094 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
4097 Remove lots of duplicated code from the EVP library. For example *every*
4098 cipher init() function handles the 'iv' in the same way according to the
4099 cipher mode. They also all do nothing if the 'key' parameter is NULL and
4100 for CFB and OFB modes they zero ctx->num.
4102 New functionality allows removal of S/MIME code RC2 hack.
4104 Most of the routines have the same form and so can be declared in terms
4107 By shifting this to the top level EVP_CipherInit() it can be removed from
4108 all individual ciphers. If the cipher wants to handle IVs or keys
4109 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
4112 Change lots of functions like EVP_EncryptUpdate() to now return a
4113 value: although software versions of the algorithms cannot fail
4114 any installed hardware versions can.
4117 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
4118 this option is set, tolerate broken clients that send the negotiated
4119 protocol version number instead of the requested protocol version
4123 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
4124 i.e. non-zero for export ciphersuites, zero otherwise.
4125 Previous versions had this flag inverted, inconsistent with
4126 rsa_tmp_cb (..._TMP_RSA_CB).
4127 [Bodo Moeller; problem reported by Amit Chopra]
4129 *) Add missing DSA library text string. Work around for some IIS
4130 key files with invalid SEQUENCE encoding.
4133 *) Add a document (doc/standards.txt) that list all kinds of standards
4134 and so on that are implemented in OpenSSL.
4137 *) Enhance c_rehash script. Old version would mishandle certificates
4138 with the same subject name hash and wouldn't handle CRLs at all.
4139 Added -fingerprint option to crl utility, to support new c_rehash
4143 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
4146 *) Fix for SSL server purpose checking. Server checking was
4147 rejecting certificates which had extended key usage present
4148 but no ssl client purpose.
4149 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
4151 *) Make PKCS#12 code work with no password. The PKCS#12 spec
4152 is a little unclear about how a blank password is handled.
4153 Since the password in encoded as a BMPString with terminating
4154 double NULL a zero length password would end up as just the
4155 double NULL. However no password at all is different and is
4156 handled differently in the PKCS#12 key generation code. NS
4157 treats a blank password as zero length. MSIE treats it as no
4158 password on export: but it will try both on import. We now do
4159 the same: PKCS12_parse() tries zero length and no password if
4160 the password is set to "" or NULL (NULL is now a valid password:
4161 it wasn't before) as does the pkcs12 application.
4164 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
4165 perror when PEM_read_bio_X509_REQ fails, the error message must
4166 be obtained from the error queue.
4169 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
4170 it in ERR_remove_state if appropriate, and change ERR_get_state
4171 accordingly to avoid race conditions (this is necessary because
4172 thread_hash is no longer constant once set).
4175 *) Bugfix for linux-elf makefile.one.
4178 *) RSA_get_default_method() will now cause a default
4179 RSA_METHOD to be chosen if one doesn't exist already.
4180 Previously this was only set during a call to RSA_new()
4181 or RSA_new_method(NULL) meaning it was possible for
4182 RSA_get_default_method() to return NULL.
4185 *) Added native name translation to the existing DSO code
4186 that will convert (if the flag to do so is set) filenames
4187 that are sufficiently small and have no path information
4188 into a canonical native form. Eg. "blah" converted to
4189 "libblah.so" or "blah.dll" etc.
4192 *) New function ERR_error_string_n(e, buf, len) which is like
4193 ERR_error_string(e, buf), but writes at most 'len' bytes
4194 including the 0 terminator. For ERR_error_string_n, 'buf'
4196 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
4198 *) CONF library reworked to become more general. A new CONF
4199 configuration file reader "class" is implemented as well as a
4200 new functions (NCONF_*, for "New CONF") to handle it. The now
4201 old CONF_* functions are still there, but are reimplemented to
4202 work in terms of the new functions. Also, a set of functions
4203 to handle the internal storage of the configuration data is
4204 provided to make it easier to write new configuration file
4205 reader "classes" (I can definitely see something reading a
4206 configuration file in XML format, for example), called _CONF_*,
4207 or "the configuration storage API"...
4209 The new configuration file reading functions are:
4211 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
4212 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
4214 NCONF_default, NCONF_WIN32
4216 NCONF_dump_fp, NCONF_dump_bio
4218 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
4219 NCONF_new creates a new CONF object. This works in the same way
4220 as other interfaces in OpenSSL, like the BIO interface.
4221 NCONF_dump_* dump the internal storage of the configuration file,
4222 which is useful for debugging. All other functions take the same
4223 arguments as the old CONF_* functions wth the exception of the
4224 first that must be a `CONF *' instead of a `LHASH *'.
4226 To make it easer to use the new classes with the old CONF_* functions,
4227 the function CONF_set_default_method is provided.
4230 *) Add '-tls1' option to 'openssl ciphers', which was already
4231 mentioned in the documentation but had not been implemented.
4232 (This option is not yet really useful because even the additional
4233 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
4236 *) Initial DSO code added into libcrypto for letting OpenSSL (and
4237 OpenSSL-based applications) load shared libraries and bind to
4238 them in a portable way.
4239 [Geoff Thorpe, with contributions from Richard Levitte]
4241 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
4243 *) Make sure _lrotl and _lrotr are only used with MSVC.
4245 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
4246 (the default implementation of RAND_status).
4248 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
4249 to '-clrext' (= clear extensions), as intended and documented.
4250 [Bodo Moeller; inconsistency pointed out by Michael Attili
4253 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
4254 was larger than the MD block size.
4255 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
4257 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
4258 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
4259 using the passed key: if the passed key was a private key the result
4260 of X509_print(), for example, would be to print out all the private key
4264 *) des_quad_cksum() byte order bug fix.
4265 [Ulf Möller, using the problem description in krb4-0.9.7, where
4266 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
4268 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
4270 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
4272 *) For easily testing in shell scripts whether some command
4273 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
4274 returns with exit code 0 iff no command of the given name is available.
4275 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
4276 the output goes to stdout and nothing is printed to stderr.
4277 Additional arguments are always ignored.
4279 Since for each cipher there is a command of the same name,
4280 the 'no-cipher' compilation switches can be tested this way.
4282 ('openssl no-XXX' is not able to detect pseudo-commands such
4283 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
4286 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
4289 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
4290 is set; it will be thrown away anyway because each handshake creates
4292 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
4293 to parameters -- in previous versions (since OpenSSL 0.9.3) the
4294 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
4295 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
4298 *) New s_client option -ign_eof: EOF at stdin is ignored, and
4299 'Q' and 'R' lose their special meanings (quit/renegotiate).
4300 This is part of what -quiet does; unlike -quiet, -ign_eof
4301 does not suppress any output.
4304 *) Add compatibility options to the purpose and trust code. The
4305 purpose X509_PURPOSE_ANY is "any purpose" which automatically
4306 accepts a certificate or CA, this was the previous behaviour,
4307 with all the associated security issues.
4309 X509_TRUST_COMPAT is the old trust behaviour: only and
4310 automatically trust self signed roots in certificate store. A
4311 new trust setting X509_TRUST_DEFAULT is used to specify that
4312 a purpose has no associated trust setting and it should instead
4313 use the value in the default purpose.
4316 *) Fix the PKCS#8 DSA private key code so it decodes keys again
4317 and fix a memory leak.
4320 *) In util/mkerr.pl (which implements 'make errors'), preserve
4321 reason strings from the previous version of the .c file, as
4322 the default to have only downcase letters (and digits) in
4323 automatically generated reasons codes is not always appropriate.
4326 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
4327 using strerror. Previously, ERR_reason_error_string() returned
4328 library names as reason strings for SYSerr; but SYSerr is a special
4329 case where small numbers are errno values, not library numbers.
4332 *) Add '-dsaparam' option to 'openssl dhparam' application. This
4333 converts DSA parameters into DH parameters. (When creating parameters,
4334 DSA_generate_parameters is used.)
4337 *) Include 'length' (recommended exponent length) in C code generated
4338 by 'openssl dhparam -C'.
4341 *) The second argument to set_label in perlasm was already being used
4342 so couldn't be used as a "file scope" flag. Moved to third argument
4346 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
4347 instead of RAND_bytes for encryption IVs and salts.
4350 *) Include RAND_status() into RAND_METHOD instead of implementing
4351 it only for md_rand.c Otherwise replacing the PRNG by calling
4352 RAND_set_rand_method would be impossible.
4355 *) Don't let DSA_generate_key() enter an infinite loop if the random
4356 number generation fails.
4359 *) New 'rand' application for creating pseudo-random output.
4362 *) Added configuration support for Linux/IA64
4363 [Rolf Haberrecker <rolf@suse.de>]
4365 *) Assembler module support for Mingw32.
4368 *) Shared library support for HPUX (in shlib/).
4369 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
4371 *) Shared library support for Solaris gcc.
4372 [Lutz Behnke <behnke@trustcenter.de>]
4374 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
4376 *) PKCS7_encrypt() was adding text MIME headers twice because they
4377 were added manually and by SMIME_crlf_copy().
4380 *) In bntest.c don't call BN_rand with zero bits argument.
4381 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
4383 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
4384 case was implemented. This caused BN_div_recp() to fail occasionally.
4387 *) Add an optional second argument to the set_label() in the perl
4388 assembly language builder. If this argument exists and is set
4389 to 1 it signals that the assembler should use a symbol whose
4390 scope is the entire file, not just the current function. This
4391 is needed with MASM which uses the format label:: for this scope.
4392 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
4394 *) Change the ASN1 types so they are typedefs by default. Before
4395 almost all types were #define'd to ASN1_STRING which was causing
4396 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
4400 *) Change names of new functions to the new get1/get0 naming
4401 convention: After 'get1', the caller owns a reference count
4402 and has to call ..._free; 'get0' returns a pointer to some
4403 data structure without incrementing reference counters.
4404 (Some of the existing 'get' functions increment a reference
4405 counter, some don't.)
4406 Similarly, 'set1' and 'add1' functions increase reference
4407 counters or duplicate objects.
4410 *) Allow for the possibility of temp RSA key generation failure:
4411 the code used to assume it always worked and crashed on failure.
4414 *) Fix potential buffer overrun problem in BIO_printf().
4415 [Ulf Möller, using public domain code by Patrick Powell; problem
4416 pointed out by David Sacerdote <das33@cornell.edu>]
4418 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
4419 RAND_egd() and RAND_status(). In the command line application,
4420 the EGD socket can be specified like a seed file using RANDFILE
4424 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
4425 Some CAs (e.g. Verisign) distribute certificates in this form.
4428 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
4429 list to exclude them. This means that no special compilation option
4430 is needed to use anonymous DH: it just needs to be included in the
4434 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
4435 EVP_MD_type. The old functionality is available in a new macro called
4436 EVP_MD_md(). Change code that uses it and update docs.
4439 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
4440 where the 'void *' argument is replaced by a function pointer argument.
4441 Previously 'void *' was abused to point to functions, which works on
4442 many platforms, but is not correct. As these functions are usually
4443 called by macros defined in OpenSSL header files, most source code
4444 should work without changes.
4447 *) <openssl/opensslconf.h> (which is created by Configure) now contains
4448 sections with information on -D... compiler switches used for
4449 compiling the library so that applications can see them. To enable
4450 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
4451 must be defined. E.g.,
4452 #define OPENSSL_ALGORITHM_DEFINES
4453 #include <openssl/opensslconf.h>
4454 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
4455 [Richard Levitte, Ulf and Bodo Möller]
4457 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
4461 *) Change the 'other' type in certificate aux info to a STACK_OF
4462 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
4463 the required ASN1 format: arbitrary types determined by an OID.
4466 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
4467 argument to 'req'. This is not because the function is newer or
4468 better than others it just uses the work 'NEW' in the certificate
4469 request header lines. Some software needs this.
4472 *) Reorganise password command line arguments: now passwords can be
4473 obtained from various sources. Delete the PEM_cb function and make
4474 it the default behaviour: i.e. if the callback is NULL and the
4475 usrdata argument is not NULL interpret it as a null terminated pass
4476 phrase. If usrdata and the callback are NULL then the pass phrase
4477 is prompted for as usual.
4480 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
4481 the support is automatically enabled. The resulting binaries will
4482 autodetect the card and use it if present.
4483 [Ben Laurie and Compaq Inc.]
4485 *) Work around for Netscape hang bug. This sends certificate request
4486 and server done in one record. Since this is perfectly legal in the
4487 SSL/TLS protocol it isn't a "bug" option and is on by default. See
4488 the bugs/SSLv3 entry for more info.
4491 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
4494 *) Add -rand argument to smime and pkcs12 applications and read/write
4498 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
4501 *) Add command line password options to the remaining applications.
4504 *) Bug fix for BN_div_recp() for numerators with an even number of
4508 *) More tests in bntest.c, and changed test_bn output.
4511 *) ./config recognizes MacOS X now.
4514 *) Bug fix for BN_div() when the first words of num and divsor are
4515 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
4518 *) Add support for various broken PKCS#8 formats, and command line
4519 options to produce them.
4522 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
4523 get temporary BIGNUMs from a BN_CTX.
4526 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
4530 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
4531 include a #define from the old name to the new. The original intent
4532 was that statically linked binaries could for example just call
4533 SSLeay_add_all_ciphers() to just add ciphers to the table and not
4534 link with digests. This never worked becayse SSLeay_add_all_digests()
4535 and SSLeay_add_all_ciphers() were in the same source file so calling
4536 one would link with the other. They are now in separate source files.
4539 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
4542 *) Use a less unusual form of the Miller-Rabin primality test (it used
4543 a binary algorithm for exponentiation integrated into the Miller-Rabin
4544 loop, our standard modexp algorithms are faster).
4547 *) Support for the EBCDIC character set completed.
4548 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
4550 *) Source code cleanups: use const where appropriate, eliminate casts,
4551 use void * instead of char * in lhash.
4554 *) Bugfix: ssl3_send_server_key_exchange was not restartable
4555 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
4556 this the server could overwrite ephemeral keys that the client
4560 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
4561 using 50 iterations of the Rabin-Miller test.
4563 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
4564 iterations of the Rabin-Miller test as required by the appendix
4565 to FIPS PUB 186[-1]) instead of DSA_is_prime.
4566 As BN_is_prime_fasttest includes trial division, DSA parameter
4567 generation becomes much faster.
4569 This implies a change for the callback functions in DSA_is_prime
4570 and DSA_generate_parameters: The callback function is called once
4571 for each positive witness in the Rabin-Miller test, not just
4572 occasionally in the inner loop; and the parameters to the
4573 callback function now provide an iteration count for the outer
4574 loop rather than for the current invocation of the inner loop.
4575 DSA_generate_parameters additionally can call the callback
4576 function with an 'iteration count' of -1, meaning that a
4577 candidate has passed the trial division test (when q is generated
4578 from an application-provided seed, trial division is skipped).
4581 *) New function BN_is_prime_fasttest that optionally does trial
4582 division before starting the Rabin-Miller test and has
4583 an additional BN_CTX * argument (whereas BN_is_prime always
4584 has to allocate at least one BN_CTX).
4585 'callback(1, -1, cb_arg)' is called when a number has passed the
4586 trial division stage.
4589 *) Fix for bug in CRL encoding. The validity dates weren't being handled
4593 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
4596 *) New function BN_pseudo_rand().
4599 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
4600 bignum version of BN_from_montgomery() with the working code from
4601 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
4605 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
4606 made it impossible to use the same SSL_SESSION data structure in
4607 SSL2 clients in multiple threads.
4610 *) The return value of RAND_load_file() no longer counts bytes obtained
4611 by stat(). RAND_load_file(..., -1) is new and uses the complete file
4612 to seed the PRNG (previously an explicit byte count was required).
4613 [Ulf Möller, Bodo Möller]
4615 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
4616 used (char *) instead of (void *) and had casts all over the place.
4619 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
4622 *) Retain source code compatibility for BN_prime_checks macro:
4623 BN_is_prime(..., BN_prime_checks, ...) now uses
4624 BN_prime_checks_for_size to determine the appropriate number of
4625 Rabin-Miller iterations.
4628 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
4629 DH_CHECK_P_NOT_SAFE_PRIME.
4630 (Check if this is true? OpenPGP calls them "strong".)
4633 *) Merge the functionality of "dh" and "gendh" programs into a new program
4634 "dhparam". The old programs are retained for now but will handle DH keys
4635 (instead of parameters) in future.
4638 *) Make the ciphers, s_server and s_client programs check the return values
4639 when a new cipher list is set.
4642 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
4643 ciphers. Before when the 56bit ciphers were enabled the sorting was
4646 The syntax for the cipher sorting has been extended to support sorting by
4647 cipher-strength (using the strength_bits hard coded in the tables).
4648 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
4650 Fix a bug in the cipher-command parser: when supplying a cipher command
4651 string with an "undefined" symbol (neither command nor alphanumeric
4652 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
4653 an error is flagged.
4655 Due to the strength-sorting extension, the code of the
4656 ssl_create_cipher_list() function was completely rearranged. I hope that
4657 the readability was also increased :-)
4658 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4660 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
4661 for the first serial number and places 2 in the serial number file. This
4662 avoids problems when the root CA is created with serial number zero and
4663 the first user certificate has the same issuer name and serial number
4667 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
4668 the new code. Add documentation for this stuff.
4671 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
4672 X509_*() to X509at_*() on the grounds that they don't handle X509
4673 structures and behave in an analagous way to the X509v3 functions:
4674 they shouldn't be called directly but wrapper functions should be used
4677 So we also now have some wrapper functions that call the X509at functions
4678 when passed certificate requests. (TO DO: similar things can be done with
4679 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
4680 things. Some of these need some d2i or i2d and print functionality
4681 because they handle more complex structures.)
4684 *) Add missing #ifndefs that caused missing symbols when building libssl
4685 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
4686 NO_RSA in ssl/s2*.c.
4687 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
4689 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
4690 has a return value which indicates the quality of the random data
4691 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
4692 error queue. New function RAND_pseudo_bytes() generates output that is
4693 guaranteed to be unique but not unpredictable. RAND_add is like
4694 RAND_seed, but takes an extra argument for an entropy estimate
4695 (RAND_seed always assumes full entropy).
4698 *) Do more iterations of Rabin-Miller probable prime test (specifically,
4699 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
4700 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
4701 in crypto/bn/bn_prime.c for the complete table). This guarantees a
4702 false-positive rate of at most 2^-80 for random input.
4705 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
4708 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
4709 in the 0.9.5 release), this returns the chain
4710 from an X509_CTX structure with a dup of the stack and all
4711 the X509 reference counts upped: so the stack will exist
4712 after X509_CTX_cleanup() has been called. Modify pkcs12.c
4715 Also make SSL_SESSION_print() print out the verify return
4719 *) Add manpage for the pkcs12 command. Also change the default
4720 behaviour so MAC iteration counts are used unless the new
4721 -nomaciter option is used. This improves file security and
4722 only older versions of MSIE (4.0 for example) need it.
4725 *) Honor the no-xxx Configure options when creating .DEF files.
4728 *) Add PKCS#10 attributes to field table: challengePassword,
4729 unstructuredName and unstructuredAddress. These are taken from
4730 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
4731 international characters are used.
4733 More changes to X509_ATTRIBUTE code: allow the setting of types
4734 based on strings. Remove the 'loc' parameter when adding
4735 attributes because these will be a SET OF encoding which is sorted
4739 *) Initial changes to the 'req' utility to allow request generation
4740 automation. This will allow an application to just generate a template
4741 file containing all the field values and have req construct the
4744 Initial support for X509_ATTRIBUTE handling. Stacks of these are
4745 used all over the place including certificate requests and PKCS#7
4746 structures. They are currently handled manually where necessary with
4747 some primitive wrappers for PKCS#7. The new functions behave in a
4748 manner analogous to the X509 extension functions: they allow
4749 attributes to be looked up by NID and added.
4751 Later something similar to the X509V3 code would be desirable to
4752 automatically handle the encoding, decoding and printing of the
4753 more complex types. The string types like challengePassword can
4754 be handled by the string table functions.
4756 Also modified the multi byte string table handling. Now there is
4757 a 'global mask' which masks out certain types. The table itself
4758 can use the flag STABLE_NO_MASK to ignore the mask setting: this
4759 is useful when for example there is only one permissible type
4760 (as in countryName) and using the mask might result in no valid
4764 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
4765 SSL_get_peer_finished to allow applications to obtain the latest
4766 Finished messages sent to the peer or expected from the peer,
4767 respectively. (SSL_get_peer_finished is usually the Finished message
4768 actually received from the peer, otherwise the protocol will be aborted.)
4770 As the Finished message are message digests of the complete handshake
4771 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
4772 be used for external authentication procedures when the authentication
4773 provided by SSL/TLS is not desired or is not enough.
4776 *) Enhanced support for Alpha Linux is added. Now ./config checks if
4777 the host supports BWX extension and if Compaq C is present on the
4778 $PATH. Just exploiting of the BWX extension results in 20-30%
4779 performance kick for some algorithms, e.g. DES and RC4 to mention
4780 a couple. Compaq C in turn generates ~20% faster code for MD5 and
4784 *) Add support for MS "fast SGC". This is arguably a violation of the
4785 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
4786 weak crypto and after checking the certificate is SGC a second one
4787 with strong crypto. MS SGC stops the first handshake after receiving
4788 the server certificate message and sends a second client hello. Since
4789 a server will typically do all the time consuming operations before
4790 expecting any further messages from the client (server key exchange
4791 is the most expensive) there is little difference between the two.
4793 To get OpenSSL to support MS SGC we have to permit a second client
4794 hello message after we have sent server done. In addition we have to
4795 reset the MAC if we do get this second client hello.
4798 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
4799 if a DER encoded private key is RSA or DSA traditional format. Changed
4800 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
4801 format DER encoded private key. Newer code should use PKCS#8 format which
4802 has the key type encoded in the ASN1 structure. Added DER private key
4803 support to pkcs8 application.
4806 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
4807 ciphersuites has been selected (as required by the SSL 3/TLS 1
4808 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
4809 is set, we interpret this as a request to violate the specification
4810 (the worst that can happen is a handshake failure, and 'correct'
4811 behaviour would result in a handshake failure anyway).
4814 *) In SSL_CTX_add_session, take into account that there might be multiple
4815 SSL_SESSION structures with the same session ID (e.g. when two threads
4816 concurrently obtain them from an external cache).
4817 The internal cache can handle only one SSL_SESSION with a given ID,
4818 so if there's a conflict, we now throw out the old one to achieve
4822 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
4823 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
4824 some routines that use cipher OIDs: some ciphers do not have OIDs
4825 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
4829 *) Simplify the trust setting structure and code. Now we just have
4830 two sequences of OIDs for trusted and rejected settings. These will
4831 typically have values the same as the extended key usage extension
4832 and any application specific purposes.
4834 The trust checking code now has a default behaviour: it will just
4835 check for an object with the same NID as the passed id. Functions can
4836 be provided to override either the default behaviour or the behaviour
4837 for a given id. SSL client, server and email already have functions
4838 in place for compatibility: they check the NID and also return "trusted"
4839 if the certificate is self signed.
4842 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
4843 traditional format into an EVP_PKEY structure.
4846 *) Add a password callback function PEM_cb() which either prompts for
4847 a password if usr_data is NULL or otherwise assumes it is a null
4848 terminated password. Allow passwords to be passed on command line
4849 environment or config files in a few more utilities.
4852 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
4853 keys. Add some short names for PKCS#8 PBE algorithms and allow them
4854 to be specified on the command line for the pkcs8 and pkcs12 utilities.
4855 Update documentation.
4858 *) Support for ASN1 "NULL" type. This could be handled before by using
4859 ASN1_TYPE but there wasn't any function that would try to read a NULL
4860 and produce an error if it couldn't. For compatibility we also have
4861 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
4862 don't allocate anything because they don't need to.
4865 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
4867 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
4869 *) Rebuild of the memory allocation routines used by OpenSSL code and
4870 possibly others as well. The purpose is to make an interface that
4871 provide hooks so anyone can build a separate set of allocation and
4872 deallocation routines to be used by OpenSSL, for example memory
4873 pool implementations, or something else, which was previously hard
4874 since Malloc(), Realloc() and Free() were defined as macros having
4875 the values malloc, realloc and free, respectively (except for Win32
4876 compilations). The same is provided for memory debugging code.
4877 OpenSSL already comes with functionality to find memory leaks, but
4878 this gives people a chance to debug other memory problems.
4880 With these changes, a new set of functions and macros have appeared:
4882 CRYPTO_set_mem_debug_functions() [F]
4883 CRYPTO_get_mem_debug_functions() [F]
4884 CRYPTO_dbg_set_options() [F]
4885 CRYPTO_dbg_get_options() [F]
4886 CRYPTO_malloc_debug_init() [M]
4888 The memory debug functions are NULL by default, unless the library
4889 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
4890 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
4891 gives the standard debugging functions that come with OpenSSL) or
4892 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
4893 provided by the library user) must be used. When the standard
4894 debugging functions are used, CRYPTO_dbg_set_options can be used to
4895 request additional information:
4896 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
4897 the CRYPTO_MDEBUG_xxx macro when compiling the library.
4899 Also, things like CRYPTO_set_mem_functions will always give the
4900 expected result (the new set of functions is used for allocation
4901 and deallocation) at all times, regardless of platform and compiler
4904 To finish it up, some functions that were never use in any other
4905 way than through macros have a new API and new semantic:
4908 CRYPTO_dbg_realloc()
4911 All macros of value have retained their old syntax.
4912 [Richard Levitte and Bodo Moeller]
4914 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
4915 ordering of SMIMECapabilities wasn't in "strength order" and there
4916 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
4920 *) Some ASN1 types with illegal zero length encoding (INTEGER,
4921 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
4922 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
4924 *) Merge in my S/MIME library for OpenSSL. This provides a simple
4925 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
4926 functionality to handle multipart/signed properly) and a utility
4927 called 'smime' to call all this stuff. This is based on code I
4928 originally wrote for Celo who have kindly allowed it to be
4929 included in OpenSSL.
4932 *) Add variants des_set_key_checked and des_set_key_unchecked of
4933 des_set_key (aka des_key_sched). Global variable des_check_key
4934 decides which of these is called by des_set_key; this way
4935 des_check_key behaves as it always did, but applications and
4936 the library itself, which was buggy for des_check_key == 1,
4937 have a cleaner way to pick the version they need.
4940 *) New function PKCS12_newpass() which changes the password of a
4944 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
4945 dynamic mix. In both cases the ids can be used as an index into the
4946 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
4947 functions so they accept a list of the field values and the
4948 application doesn't need to directly manipulate the X509_TRUST
4952 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
4956 *) Modify the way the V3 extension code looks up extensions. This now
4957 works in a similar way to the object code: we have some "standard"
4958 extensions in a static table which is searched with OBJ_bsearch()
4959 and the application can add dynamic ones if needed. The file
4960 crypto/x509v3/ext_dat.h now has the info: this file needs to be
4961 updated whenever a new extension is added to the core code and kept
4962 in ext_nid order. There is a simple program 'tabtest.c' which checks
4963 this. New extensions are not added too often so this file can readily
4964 be maintained manually.
4966 There are two big advantages in doing things this way. The extensions
4967 can be looked up immediately and no longer need to be "added" using
4968 X509V3_add_standard_extensions(): this function now does nothing.
4969 [Side note: I get *lots* of email saying the extension code doesn't
4970 work because people forget to call this function]
4971 Also no dynamic allocation is done unless new extensions are added:
4972 so if we don't add custom extensions there is no need to call
4973 X509V3_EXT_cleanup().
4976 *) Modify enc utility's salting as follows: make salting the default. Add a
4977 magic header, so unsalted files fail gracefully instead of just decrypting
4978 to garbage. This is because not salting is a big security hole, so people
4979 should be discouraged from doing it.
4982 *) Fixes and enhancements to the 'x509' utility. It allowed a message
4983 digest to be passed on the command line but it only used this
4984 parameter when signing a certificate. Modified so all relevant
4985 operations are affected by the digest parameter including the
4986 -fingerprint and -x509toreq options. Also -x509toreq choked if a
4987 DSA key was used because it didn't fix the digest.
4990 *) Initial certificate chain verify code. Currently tests the untrusted
4991 certificates for consistency with the verify purpose (which is set
4992 when the X509_STORE_CTX structure is set up) and checks the pathlength.
4994 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
4995 this is because it will reject chains with invalid extensions whereas
4996 every previous version of OpenSSL and SSLeay made no checks at all.
4998 Trust code: checks the root CA for the relevant trust settings. Trust
4999 settings have an initial value consistent with the verify purpose: e.g.
5000 if the verify purpose is for SSL client use it expects the CA to be
5001 trusted for SSL client use. However the default value can be changed to
5002 permit custom trust settings: one example of this would be to only trust
5003 certificates from a specific "secure" set of CAs.
5005 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
5006 which should be used for version portability: especially since the
5007 verify structure is likely to change more often now.
5009 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
5010 to set them. If not set then assume SSL clients will verify SSL servers
5013 Two new options to the verify program: -untrusted allows a set of
5014 untrusted certificates to be passed in and -purpose which sets the
5015 intended purpose of the certificate. If a purpose is set then the
5016 new chain verify code is used to check extension consistency.
5019 *) Support for the authority information access extension.
5022 *) Modify RSA and DSA PEM read routines to transparently handle
5023 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
5024 public keys in a format compatible with certificate
5025 SubjectPublicKeyInfo structures. Unfortunately there were already
5026 functions called *_PublicKey_* which used various odd formats so
5027 these are retained for compatibility: however the DSA variants were
5028 never in a public release so they have been deleted. Changed dsa/rsa
5029 utilities to handle the new format: note no releases ever handled public
5030 keys so we should be OK.
5032 The primary motivation for this change is to avoid the same fiasco
5033 that dogs private keys: there are several incompatible private key
5034 formats some of which are standard and some OpenSSL specific and
5035 require various evil hacks to allow partial transparent handling and
5036 even then it doesn't work with DER formats. Given the option anything
5037 other than PKCS#8 should be dumped: but the other formats have to
5038 stay in the name of compatibility.
5040 With public keys and the benefit of hindsight one standard format
5041 is used which works with EVP_PKEY, RSA or DSA structures: though
5042 it clearly returns an error if you try to read the wrong kind of key.
5044 Added a -pubkey option to the 'x509' utility to output the public key.
5045 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
5046 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
5047 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
5048 that do the same as the EVP_PKEY_assign_*() except they up the
5049 reference count of the added key (they don't "swallow" the
5053 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
5054 CRLs would fail if the file contained no certificates or no CRLs:
5055 added a new function to read in both types and return the number
5056 read: this means that if none are read it will be an error. The
5057 DER versions of the certificate and CRL reader would always fail
5058 because it isn't possible to mix certificates and CRLs in DER format
5059 without choking one or the other routine. Changed this to just read
5060 a certificate: this is the best we can do. Also modified the code
5061 in apps/verify.c to take notice of return codes: it was previously
5062 attempting to read in certificates from NULL pointers and ignoring
5063 any errors: this is one reason why the cert and CRL reader seemed
5064 to work. It doesn't check return codes from the default certificate
5065 routines: these may well fail if the certificates aren't installed.
5068 *) Code to support otherName option in GeneralName.
5071 *) First update to verify code. Change the verify utility
5072 so it warns if it is passed a self signed certificate:
5073 for consistency with the normal behaviour. X509_verify
5074 has been modified to it will now verify a self signed
5075 certificate if *exactly* the same certificate appears
5076 in the store: it was previously impossible to trust a
5077 single self signed certificate. This means that:
5078 openssl verify ss.pem
5079 now gives a warning about a self signed certificate but
5080 openssl verify -CAfile ss.pem ss.pem
5084 *) For servers, store verify_result in SSL_SESSION data structure
5085 (and add it to external session representation).
5086 This is needed when client certificate verifications fails,
5087 but an application-provided verification callback (set by
5088 SSL_CTX_set_cert_verify_callback) allows accepting the session
5089 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
5090 but returns 1): When the session is reused, we have to set
5091 ssl->verify_result to the appropriate error code to avoid
5093 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
5095 *) Fix a bug in the new PKCS#7 code: it didn't consider the
5096 case in PKCS7_dataInit() where the signed PKCS7 structure
5097 didn't contain any existing data because it was being created.
5098 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
5100 *) Add a salt to the key derivation routines in enc.c. This
5101 forms the first 8 bytes of the encrypted file. Also add a
5102 -S option to allow a salt to be input on the command line.
5105 *) New function X509_cmp(). Oddly enough there wasn't a function
5106 to compare two certificates. We do this by working out the SHA1
5107 hash and comparing that. X509_cmp() will be needed by the trust
5111 *) SSL_get1_session() is like SSL_get_session(), but increments
5112 the reference count in the SSL_SESSION returned.
5113 [Geoff Thorpe <geoff@eu.c2.net>]
5115 *) Fix for 'req': it was adding a null to request attributes.
5116 Also change the X509_LOOKUP and X509_INFO code to handle
5117 certificate auxiliary information.
5120 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
5124 *) Add the possibility to add extra information to the memory leak
5125 detecting output, to form tracebacks, showing from where each
5126 allocation was originated: CRYPTO_push_info("constant string") adds
5127 the string plus current file name and line number to a per-thread
5128 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
5129 is like calling CYRPTO_pop_info() until the stack is empty.
5130 Also updated memory leak detection code to be multi-thread-safe.
5133 *) Add options -text and -noout to pkcs7 utility and delete the
5134 encryption options which never did anything. Update docs.
5137 *) Add options to some of the utilities to allow the pass phrase
5138 to be included on either the command line (not recommended on
5139 OSes like Unix) or read from the environment. Update the
5140 manpages and fix a few bugs.
5143 *) Add a few manpages for some of the openssl commands.
5146 *) Fix the -revoke option in ca. It was freeing up memory twice,
5147 leaking and not finding already revoked certificates.
5150 *) Extensive changes to support certificate auxiliary information.
5151 This involves the use of X509_CERT_AUX structure and X509_AUX
5152 functions. An X509_AUX function such as PEM_read_X509_AUX()
5153 can still read in a certificate file in the usual way but it
5154 will also read in any additional "auxiliary information". By
5155 doing things this way a fair degree of compatibility can be
5156 retained: existing certificates can have this information added
5157 using the new 'x509' options.
5159 Current auxiliary information includes an "alias" and some trust
5160 settings. The trust settings will ultimately be used in enhanced
5161 certificate chain verification routines: currently a certificate
5162 can only be trusted if it is self signed and then it is trusted
5166 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
5167 The problem was that one of the replacement routines had not been working
5168 since SSLeay releases. For now the offending routine has been replaced
5169 with non-optimised assembler. Even so, this now gives around 95%
5170 performance improvement for 1024 bit RSA signs.
5173 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
5174 handling. Most clients have the effective key size in bits equal to
5175 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
5176 A few however don't do this and instead use the size of the decrypted key
5177 to determine the RC2 key length and the AlgorithmIdentifier to determine
5178 the effective key length. In this case the effective key length can still
5179 be 40 bits but the key length can be 168 bits for example. This is fixed
5180 by manually forcing an RC2 key into the EVP_PKEY structure because the
5181 EVP code can't currently handle unusual RC2 key sizes: it always assumes
5182 the key length and effective key length are equal.
5185 *) Add a bunch of functions that should simplify the creation of
5186 X509_NAME structures. Now you should be able to do:
5187 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
5188 and have it automatically work out the correct field type and fill in
5189 the structures. The more adventurous can try:
5190 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
5191 and it will (hopefully) work out the correct multibyte encoding.
5194 *) Change the 'req' utility to use the new field handling and multibyte
5195 copy routines. Before the DN field creation was handled in an ad hoc
5196 way in req, ca, and x509 which was rather broken and didn't support
5197 BMPStrings or UTF8Strings. Since some software doesn't implement
5198 BMPStrings or UTF8Strings yet, they can be enabled using the config file
5199 using the dirstring_type option. See the new comment in the default
5200 openssl.cnf for more info.
5203 *) Make crypto/rand/md_rand.c more robust:
5204 - Assure unique random numbers after fork().
5205 - Make sure that concurrent threads access the global counter and
5206 md serializably so that we never lose entropy in them
5207 or use exactly the same state in multiple threads.
5208 Access to the large state is not always serializable because
5209 the additional locking could be a performance killer, and
5210 md should be large enough anyway.
5213 *) New file apps/app_rand.c with commonly needed functionality
5214 for handling the random seed file.
5216 Use the random seed file in some applications that previously did not:
5218 dsaparam -genkey (which also ignored its '-rand' option),
5221 x509 (when signing).
5222 Except on systems with /dev/urandom, it is crucial to have a random
5223 seed file at least for key creation, DSA signing, and for DH exchanges;
5224 for RSA signatures we could do without one.
5226 gendh and gendsa (unlike genrsa) used to read only the first byte
5227 of each file listed in the '-rand' option. The function as previously
5228 found in genrsa is now in app_rand.c and is used by all programs
5229 that support '-rand'.
5232 *) In RAND_write_file, use mode 0600 for creating files;
5233 don't just chmod when it may be too late.
5236 *) Report an error from X509_STORE_load_locations
5237 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
5240 *) New function ASN1_mbstring_copy() this copies a string in either
5241 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
5242 into an ASN1_STRING type. A mask of permissible types is passed
5243 and it chooses the "minimal" type to use or an error if not type
5247 *) Add function equivalents to the various macros in asn1.h. The old
5248 macros are retained with an M_ prefix. Code inside the library can
5249 use the M_ macros. External code (including the openssl utility)
5250 should *NOT* in order to be "shared library friendly".
5253 *) Add various functions that can check a certificate's extensions
5254 to see if it usable for various purposes such as SSL client,
5255 server or S/MIME and CAs of these types. This is currently
5256 VERY EXPERIMENTAL but will ultimately be used for certificate chain
5257 verification. Also added a -purpose flag to x509 utility to
5258 print out all the purposes.
5261 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
5265 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
5266 for, obtain and decode and extension and obtain its critical flag.
5267 This allows all the necessary extension code to be handled in a
5268 single function call.
5271 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
5272 platforms. See crypto/rc4/rc4_enc.c for further details.
5275 *) New -noout option to asn1parse. This causes no output to be produced
5276 its main use is when combined with -strparse and -out to extract data
5277 from a file (which may not be in ASN.1 format).
5280 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
5281 when producing the local key id.
5282 [Richard Levitte <levitte@stacken.kth.se>]
5284 *) New option -dhparam in s_server. This allows a DH parameter file to be
5285 stated explicitly. If it is not stated then it tries the first server
5286 certificate file. The previous behaviour hard coded the filename
5290 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
5291 a public key to be input or output. For example:
5292 openssl rsa -in key.pem -pubout -out pubkey.pem
5293 Also added necessary DSA public key functions to handle this.
5296 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
5297 in the message. This was handled by allowing
5298 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
5299 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
5301 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
5302 to the end of the strings whereas this didn't. This would cause problems
5303 if strings read with d2i_ASN1_bytes() were later modified.
5304 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
5306 *) Fix for base64 decode bug. When a base64 bio reads only one line of
5307 data and it contains EOF it will end up returning an error. This is
5308 caused by input 46 bytes long. The cause is due to the way base64
5309 BIOs find the start of base64 encoded data. They do this by trying a
5310 trial decode on each line until they find one that works. When they
5311 do a flag is set and it starts again knowing it can pass all the
5312 data directly through the decoder. Unfortunately it doesn't reset
5313 the context it uses. This means that if EOF is reached an attempt
5314 is made to pass two EOFs through the context and this causes the
5315 resulting error. This can also cause other problems as well. As is
5316 usual with these problems it takes *ages* to find and the fix is
5317 trivial: move one line.
5318 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
5320 *) Ugly workaround to get s_client and s_server working under Windows. The
5321 old code wouldn't work because it needed to select() on sockets and the
5322 tty (for keypresses and to see if data could be written). Win32 only
5323 supports select() on sockets so we select() with a 1s timeout on the
5324 sockets and then see if any characters are waiting to be read, if none
5325 are present then we retry, we also assume we can always write data to
5326 the tty. This isn't nice because the code then blocks until we've
5327 received a complete line of data and it is effectively polling the
5328 keyboard at 1s intervals: however it's quite a bit better than not
5329 working at all :-) A dedicated Windows application might handle this
5330 with an event loop for example.
5333 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
5334 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
5335 will be called when RSA_sign() and RSA_verify() are used. This is useful
5336 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
5337 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
5338 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
5339 This necessitated the support of an extra signature type NID_md5_sha1
5340 for SSL signatures and modifications to the SSL library to use it instead
5341 of calling RSA_public_decrypt() and RSA_private_encrypt().
5344 *) Add new -verify -CAfile and -CApath options to the crl program, these
5345 will lookup a CRL issuers certificate and verify the signature in a
5346 similar way to the verify program. Tidy up the crl program so it
5347 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
5348 less strict. It will now permit CRL extensions even if it is not
5349 a V2 CRL: this will allow it to tolerate some broken CRLs.
5352 *) Initialize all non-automatic variables each time one of the openssl
5353 sub-programs is started (this is necessary as they may be started
5354 multiple times from the "OpenSSL>" prompt).
5355 [Lennart Bang, Bodo Moeller]
5357 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
5358 removing all other RSA functionality (this is what NO_RSA does). This
5359 is so (for example) those in the US can disable those operations covered
5360 by the RSA patent while allowing storage and parsing of RSA keys and RSA
5364 *) Non-copying interface to BIO pairs.
5365 (still largely untested)
5368 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
5369 ASCII string. This was handled independently in various places before.
5372 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
5373 UTF8 strings a character at a time.
5376 *) Use client_version from client hello to select the protocol
5377 (s23_srvr.c) and for RSA client key exchange verification
5378 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
5381 *) Add various utility functions to handle SPKACs, these were previously
5382 handled by poking round in the structure internals. Added new function
5383 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
5384 print, verify and generate SPKACs. Based on an original idea from
5385 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
5388 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
5391 *) Allow the config file extension section to be overwritten on the
5392 command line. Based on an original idea from Massimiliano Pala
5393 <madwolf@comune.modena.it>. The new option is called -extensions
5394 and can be applied to ca, req and x509. Also -reqexts to override
5395 the request extensions in req and -crlexts to override the crl extensions
5399 *) Add new feature to the SPKAC handling in ca. Now you can include
5400 the same field multiple times by preceding it by "XXXX." for example:
5403 this is the same syntax as used in the req config file.
5406 *) Allow certificate extensions to be added to certificate requests. These
5407 are specified in a 'req_extensions' option of the req section of the
5408 config file. They can be printed out with the -text option to req but
5409 are otherwise ignored at present.
5412 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
5413 data read consists of only the final block it would not decrypted because
5414 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
5415 A misplaced 'break' also meant the decrypted final block might not be
5416 copied until the next read.
5419 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
5420 a few extra parameters to the DH structure: these will be useful if
5421 for example we want the value of 'q' or implement X9.42 DH.
5424 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
5425 provides hooks that allow the default DSA functions or functions on a
5426 "per key" basis to be replaced. This allows hardware acceleration and
5427 hardware key storage to be handled without major modification to the
5428 library. Also added low level modexp hooks and CRYPTO_EX structure and
5429 associated functions.
5432 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
5433 as "read only": it can't be written to and the buffer it points to will
5434 not be freed. Reading from a read only BIO is much more efficient than
5435 a normal memory BIO. This was added because there are several times when
5436 an area of memory needs to be read from a BIO. The previous method was
5437 to create a memory BIO and write the data to it, this results in two
5438 copies of the data and an O(n^2) reading algorithm. There is a new
5439 function BIO_new_mem_buf() which creates a read only memory BIO from
5440 an area of memory. Also modified the PKCS#7 routines to use read only
5444 *) Bugfix: ssl23_get_client_hello did not work properly when called in
5445 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
5446 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
5447 but a retry condition occured while trying to read the rest.
5450 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
5451 NID_pkcs7_encrypted by default: this was wrong since this should almost
5452 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
5453 the encrypted data type: this is a more sensible place to put it and it
5454 allows the PKCS#12 code to be tidied up that duplicated this
5458 *) Changed obj_dat.pl script so it takes its input and output files on
5459 the command line. This should avoid shell escape redirection problems
5463 *) Initial support for certificate extension requests, these are included
5464 in things like Xenroll certificate requests. Included functions to allow
5465 extensions to be obtained and added.
5468 *) -crlf option to s_client and s_server for sending newlines as
5469 CRLF (as required by many protocols).
5472 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
5474 *) Install libRSAglue.a when OpenSSL is built with RSAref.
5475 [Ralf S. Engelschall]
5477 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
5478 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
5480 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
5484 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
5485 DH parameters/keys (q is lost during that conversion, but the resulting
5486 DH parameters contain its length).
5488 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
5489 much faster than DH_generate_parameters (which creates parameters
5490 where p = 2*q + 1), and also the smaller q makes DH computations
5491 much more efficient (160-bit exponentiation instead of 1024-bit
5492 exponentiation); so this provides a convenient way to support DHE
5493 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
5494 utter importance to use
5495 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5497 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5498 when such DH parameters are used, because otherwise small subgroup
5499 attacks may become possible!
5502 *) Avoid memory leak in i2d_DHparams.
5505 *) Allow the -k option to be used more than once in the enc program:
5506 this allows the same encrypted message to be read by multiple recipients.
5509 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
5510 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
5511 it will always use the numerical form of the OID, even if it has a short
5515 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
5516 method only got called if p,q,dmp1,dmq1,iqmp components were present,
5517 otherwise bn_mod_exp was called. In the case of hardware keys for example
5518 no private key components need be present and it might store extra data
5519 in the RSA structure, which cannot be accessed from bn_mod_exp.
5520 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
5521 private key operations.
5524 *) Added support for SPARC Linux.
5527 *) pem_password_cb function type incompatibly changed from
5528 typedef int pem_password_cb(char *buf, int size, int rwflag);
5530 ....(char *buf, int size, int rwflag, void *userdata);
5531 so that applications can pass data to their callbacks:
5532 The PEM[_ASN1]_{read,write}... functions and macros now take an
5533 additional void * argument, which is just handed through whenever
5534 the password callback is called.
5535 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
5537 New function SSL_CTX_set_default_passwd_cb_userdata.
5539 Compatibility note: As many C implementations push function arguments
5540 onto the stack in reverse order, the new library version is likely to
5541 interoperate with programs that have been compiled with the old
5542 pem_password_cb definition (PEM_whatever takes some data that
5543 happens to be on the stack as its last argument, and the callback
5544 just ignores this garbage); but there is no guarantee whatsoever that
5547 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
5548 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
5549 problems not only on Windows, but also on some Unix platforms.
5550 To avoid problematic command lines, these definitions are now in an
5551 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
5552 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
5555 *) MIPS III/IV assembler module is reimplemented.
5558 *) More DES library cleanups: remove references to srand/rand and
5559 delete an unused file.
5562 *) Add support for the the free Netwide assembler (NASM) under Win32,
5563 since not many people have MASM (ml) and it can be hard to obtain.
5564 This is currently experimental but it seems to work OK and pass all
5565 the tests. Check out INSTALL.W32 for info.
5568 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
5569 without temporary keys kept an extra copy of the server key,
5570 and connections with temporary keys did not free everything in case
5574 *) New function RSA_check_key and new openssl rsa option -check
5575 for verifying the consistency of RSA keys.
5576 [Ulf Moeller, Bodo Moeller]
5578 *) Various changes to make Win32 compile work:
5579 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
5580 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
5581 comparison" warnings.
5582 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
5585 *) Add a debugging option to PKCS#5 v2 key generation function: when
5586 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
5587 derived keys are printed to stderr.
5590 *) Copy the flags in ASN1_STRING_dup().
5591 [Roman E. Pavlov <pre@mo.msk.ru>]
5593 *) The x509 application mishandled signing requests containing DSA
5594 keys when the signing key was also DSA and the parameters didn't match.
5596 It was supposed to omit the parameters when they matched the signing key:
5597 the verifying software was then supposed to automatically use the CA's
5598 parameters if they were absent from the end user certificate.
5600 Omitting parameters is no longer recommended. The test was also
5601 the wrong way round! This was probably due to unusual behaviour in
5602 EVP_cmp_parameters() which returns 1 if the parameters match.
5603 This meant that parameters were omitted when they *didn't* match and
5604 the certificate was useless. Certificates signed with 'ca' didn't have
5606 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
5608 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
5609 The interface is as follows:
5610 Applications can use
5611 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
5612 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
5613 "off" is now the default.
5614 The library internally uses
5615 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
5616 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
5617 to disable memory-checking temporarily.
5619 Some inconsistent states that previously were possible (and were
5620 even the default) are now avoided.
5622 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
5623 with each memory chunk allocated; this is occasionally more helpful
5624 than just having a counter.
5626 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
5628 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
5632 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
5633 which largely parallels "options", but is for changing API behaviour,
5634 whereas "options" are about protocol behaviour.
5635 Initial "mode" flags are:
5637 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
5638 a single record has been written.
5639 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
5640 retries use the same buffer location.
5641 (But all of the contents must be
5645 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
5648 *) Fix problems with no-hmac etc.
5649 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
5651 *) New functions RSA_get_default_method(), RSA_set_method() and
5652 RSA_get_method(). These allows replacement of RSA_METHODs without having
5653 to mess around with the internals of an RSA structure.
5656 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
5657 Also really enable memory leak checks in openssl.c and in some
5659 [Chad C. Mulligan, Bodo Moeller]
5661 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
5662 up the length of negative integers. This has now been simplified to just
5663 store the length when it is first determined and use it later, rather
5664 than trying to keep track of where data is copied and updating it to
5666 [Steve Henson, reported by Brien Wheeler
5667 <bwheeler@authentica-security.com>]
5669 *) Add a new function PKCS7_signatureVerify. This allows the verification
5670 of a PKCS#7 signature but with the signing certificate passed to the
5671 function itself. This contrasts with PKCS7_dataVerify which assumes the
5672 certificate is present in the PKCS#7 structure. This isn't always the
5673 case: certificates can be omitted from a PKCS#7 structure and be
5674 distributed by "out of band" means (such as a certificate database).
5677 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
5678 function prototypes in pem.h, also change util/mkdef.pl to add the
5679 necessary function names.
5682 *) mk1mf.pl (used by Windows builds) did not properly read the
5683 options set by Configure in the top level Makefile, and Configure
5684 was not even able to write more than one option correctly.
5685 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
5688 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
5689 file to be loaded from a BIO or FILE pointer. The BIO version will
5690 for example allow memory BIOs to contain config info.
5693 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
5694 Whoever hopes to achieve shared-library compatibility across versions
5695 must use this, not the compile-time macro.
5696 (Exercise 0.9.4: Which is the minimum library version required by
5698 Note: All this applies only to multi-threaded programs, others don't
5702 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
5703 through a BIO pair triggered the default case, i.e.
5704 SSLerr(...,SSL_R_UNKNOWN_STATE).
5707 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
5708 can use the SSL library even if none of the specific BIOs is
5712 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
5713 for the encoded length.
5714 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
5716 *) Add initial documentation of the X509V3 functions.
5719 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
5720 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
5721 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
5722 secure PKCS#8 private key format with a high iteration count.
5725 *) Fix determination of Perl interpreter: A perl or perl5
5726 _directory_ in $PATH was also accepted as the interpreter.
5727 [Ralf S. Engelschall]
5729 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
5730 wrong with it but it was very old and did things like calling
5731 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
5735 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
5736 to use the new extension code.
5739 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
5740 with macros. This should make it easier to change their form, add extra
5741 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
5745 *) Add to configuration table a new entry that can specify an alternative
5746 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
5747 according to Mark Crispin <MRC@Panda.COM>.
5751 *) DES CBC did not update the IV. Weird.
5754 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
5755 Changing the behaviour of the former might break existing programs --
5756 where IV updating is needed, des_ncbc_encrypt can be used.
5759 *) When bntest is run from "make test" it drives bc to check its
5760 calculations, as well as internally checking them. If an internal check
5761 fails, it needs to cause bc to give a non-zero result or make test carries
5762 on without noticing the failure. Fixed.
5765 *) DES library cleanups.
5768 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
5769 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
5770 ciphers. NOTE: although the key derivation function has been verified
5771 against some published test vectors it has not been extensively tested
5772 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
5776 *) Instead of "mkdir -p", which is not fully portable, use new
5777 Perl script "util/mkdir-p.pl".
5780 *) Rewrite the way password based encryption (PBE) is handled. It used to
5781 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
5782 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
5783 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
5784 the 'parameter' field of the AlgorithmIdentifier is passed to the
5785 underlying key generation function so it must do its own ASN1 parsing.
5786 This has also changed the EVP_PBE_CipherInit() function which now has a
5787 'parameter' argument instead of literal salt and iteration count values
5788 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
5791 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
5792 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
5793 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
5794 KEY" because this clashed with PKCS#8 unencrypted string. Since this
5795 value was just used as a "magic string" and not used directly its
5796 value doesn't matter.
5799 *) Introduce some semblance of const correctness to BN. Shame C doesn't
5803 *) "linux-sparc64" configuration (ultrapenguin).
5804 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
5805 "linux-sparc" configuration.
5806 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
5808 *) config now generates no-xxx options for missing ciphers.
5811 *) Support the EBCDIC character set (work in progress).
5812 File ebcdic.c not yet included because it has a different license.
5813 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5815 *) Support BS2000/OSD-POSIX.
5816 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5818 *) Make callbacks for key generation use void * instead of char *.
5821 *) Make S/MIME samples compile (not yet tested).
5824 *) Additional typesafe stacks.
5827 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
5831 Changes between 0.9.3 and 0.9.3a [29 May 1999]
5833 *) New configuration variant "sco5-gcc".
5835 *) Updated some demos.
5836 [Sean O Riordain, Wade Scholine]
5838 *) Add missing BIO_free at exit of pkcs12 application.
5841 *) Fix memory leak in conf.c.
5844 *) Updates for Win32 to assembler version of MD5.
5847 *) Set #! path to perl in apps/der_chop to where we found it
5848 instead of using a fixed path.
5851 *) SHA library changes for irix64-mips4-cc.
5854 *) Improvements for VMS support.
5858 Changes between 0.9.2b and 0.9.3 [24 May 1999]
5860 *) Bignum library bug fix. IRIX 6 passes "make test" now!
5861 This also avoids the problems with SC4.2 and unpatched SC5.
5862 [Andy Polyakov <appro@fy.chalmers.se>]
5864 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
5865 These are required because of the typesafe stack would otherwise break
5866 existing code. If old code used a structure member which used to be STACK
5867 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
5868 sk_num or sk_value it would produce an error because the num, data members
5869 are not present in STACK_OF. Now it just produces a warning. sk_set
5870 replaces the old method of assigning a value to sk_value
5871 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
5872 that does this will no longer work (and should use sk_set instead) but
5873 this could be regarded as a "questionable" behaviour anyway.
5876 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
5877 correctly handle encrypted S/MIME data.
5880 *) Change type of various DES function arguments from des_cblock
5881 (which means, in function argument declarations, pointer to char)
5882 to des_cblock * (meaning pointer to array with 8 char elements),
5883 which allows the compiler to do more typechecking; it was like
5884 that back in SSLeay, but with lots of ugly casts.
5886 Introduce new type const_des_cblock.
5889 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
5890 problems: find RecipientInfo structure that matches recipient certificate
5891 and initialise the ASN1 structures properly based on passed cipher.
5894 *) Belatedly make the BN tests actually check the results.
5897 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
5898 to and from BNs: it was completely broken. New compilation option
5899 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
5900 key elements as negative integers.
5903 *) Reorganize and speed up MD5.
5904 [Andy Polyakov <appro@fy.chalmers.se>]
5907 [Richard Levitte <richard@levitte.org>]
5909 *) New option -out to asn1parse to allow the parsed structure to be
5910 output to a file. This is most useful when combined with the -strparse
5911 option to examine the output of things like OCTET STRINGS.
5914 *) Make SSL library a little more fool-proof by not requiring any longer
5915 that SSL_set_{accept,connect}_state be called before
5916 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
5917 in many applications because usually everything *appeared* to work as
5918 intended anyway -- now it really works as intended).
5921 *) Move openssl.cnf out of lib/.
5924 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
5925 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
5926 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
5927 [Ralf S. Engelschall]
5929 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
5930 handle PKCS#7 enveloped data properly.
5931 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
5933 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
5934 copying pointers. The cert_st handling is changed by this in
5935 various ways (and thus what used to be known as ctx->default_cert
5936 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
5937 any longer when s->cert does not give us what we need).
5938 ssl_cert_instantiate becomes obsolete by this change.
5939 As soon as we've got the new code right (possibly it already is?),
5940 we have solved a couple of bugs of the earlier code where s->cert
5941 was used as if it could not have been shared with other SSL structures.
5943 Note that using the SSL API in certain dirty ways now will result
5944 in different behaviour than observed with earlier library versions:
5945 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
5946 does not influence s as it used to.
5948 In order to clean up things more thoroughly, inside SSL_SESSION
5949 we don't use CERT any longer, but a new structure SESS_CERT
5950 that holds per-session data (if available); currently, this is
5951 the peer's certificate chain and, for clients, the server's certificate
5952 and temporary key. CERT holds only those values that can have
5953 meaningful defaults in an SSL_CTX.
5956 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
5957 from the internal representation. Various PKCS#7 fixes: remove some
5958 evil casts and set the enc_dig_alg field properly based on the signing
5962 *) Allow PKCS#12 password to be set from the command line or the
5963 environment. Let 'ca' get its config file name from the environment
5964 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
5968 *) Allow certificate policies extension to use an IA5STRING for the
5969 organization field. This is contrary to the PKIX definition but
5970 VeriSign uses it and IE5 only recognises this form. Document 'x509'
5974 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
5975 without disallowing inline assembler and the like for non-pedantic builds.
5978 *) Support Borland C++ builder.
5979 [Janez Jere <jj@void.si>, modified by Ulf Möller]
5984 *) SHA-1 cleanups and performance enhancements.
5985 [Andy Polyakov <appro@fy.chalmers.se>]
5987 *) Sparc v8plus assembler for the bignum library.
5988 [Andy Polyakov <appro@fy.chalmers.se>]
5990 *) Accept any -xxx and +xxx compiler options in Configure.
5993 *) Update HPUX configuration.
5996 *) Add missing sk_<type>_unshift() function to safestack.h
5997 [Ralf S. Engelschall]
5999 *) New function SSL_CTX_use_certificate_chain_file that sets the
6000 "extra_cert"s in addition to the certificate. (This makes sense
6001 only for "PEM" format files, as chains as a whole are not
6005 *) Support verify_depth from the SSL API.
6006 x509_vfy.c had what can be considered an off-by-one-error:
6007 Its depth (which was not part of the external interface)
6008 was actually counting the number of certificates in a chain;
6009 now it really counts the depth.
6012 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
6013 instead of X509err, which often resulted in confusing error
6014 messages since the error codes are not globally unique
6015 (e.g. an alleged error in ssl3_accept when a certificate
6016 didn't match the private key).
6018 *) New function SSL_CTX_set_session_id_context that allows to set a default
6019 value (so that you don't need SSL_set_session_id_context for each
6020 connection using the SSL_CTX).
6023 *) OAEP decoding bug fix.
6026 *) Support INSTALL_PREFIX for package builders, as proposed by
6030 *) New Configure options "threads" and "no-threads". For systems
6031 where the proper compiler options are known (currently Solaris
6032 and Linux), "threads" is the default.
6035 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
6038 *) Install various scripts to $(OPENSSLDIR)/misc, not to
6039 $(INSTALLTOP)/bin -- they shouldn't clutter directories
6040 such as /usr/local/bin.
6043 *) "make linux-shared" to build shared libraries.
6044 [Niels Poppe <niels@netbox.org>]
6046 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
6049 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
6050 extension adding in x509 utility.
6053 *) Remove NOPROTO sections and error code comments.
6056 *) Partial rewrite of the DEF file generator to now parse the ANSI
6060 *) New Configure options --prefix=DIR and --openssldir=DIR.
6063 *) Complete rewrite of the error code script(s). It is all now handled
6064 by one script at the top level which handles error code gathering,
6065 header rewriting and C source file generation. It should be much better
6066 than the old method: it now uses a modified version of Ulf's parser to
6067 read the ANSI prototypes in all header files (thus the old K&R definitions
6068 aren't needed for error creation any more) and do a better job of
6069 translating function codes into names. The old 'ASN1 error code imbedded
6070 in a comment' is no longer necessary and it doesn't use .err files which
6071 have now been deleted. Also the error code call doesn't have to appear all
6072 on one line (which resulted in some large lines...).
6075 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
6078 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
6079 0 (which usually indicates a closed connection), but continue reading.
6082 *) Fix some race conditions.
6085 *) Add support for CRL distribution points extension. Add Certificate
6086 Policies and CRL distribution points documentation.
6089 *) Move the autogenerated header file parts to crypto/opensslconf.h.
6092 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
6093 8 of keying material. Merlin has also confirmed interop with this fix
6094 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
6095 [Merlin Hughes <merlin@baltimore.ie>]
6097 *) Fix lots of warnings.
6098 [Richard Levitte <levitte@stacken.kth.se>]
6100 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
6101 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
6102 [Richard Levitte <levitte@stacken.kth.se>]
6104 *) Fix problems with sizeof(long) == 8.
6105 [Andy Polyakov <appro@fy.chalmers.se>]
6107 *) Change functions to ANSI C.
6110 *) Fix typos in error codes.
6111 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
6113 *) Remove defunct assembler files from Configure.
6116 *) SPARC v8 assembler BIGNUM implementation.
6117 [Andy Polyakov <appro@fy.chalmers.se>]
6119 *) Support for Certificate Policies extension: both print and set.
6120 Various additions to support the r2i method this uses.
6123 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
6124 return a const string when you are expecting an allocated buffer.
6127 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
6128 types DirectoryString and DisplayText.
6131 *) Add code to allow r2i extensions to access the configuration database,
6132 add an LHASH database driver and add several ctx helper functions.
6135 *) Fix an evil bug in bn_expand2() which caused various BN functions to
6136 fail when they extended the size of a BIGNUM.
6139 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
6140 support typesafe stack.
6143 *) Fix typo in SSL_[gs]et_options().
6144 [Nils Frostberg <nils@medcom.se>]
6146 *) Delete various functions and files that belonged to the (now obsolete)
6147 old X509V3 handling code.
6150 *) New Configure option "rsaref".
6153 *) Don't auto-generate pem.h.
6156 *) Introduce type-safe ASN.1 SETs.
6159 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
6160 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
6162 *) Introduce type-safe STACKs. This will almost certainly break lots of code
6163 that links with OpenSSL (well at least cause lots of warnings), but fear
6164 not: the conversion is trivial, and it eliminates loads of evil casts. A
6165 few STACKed things have been converted already. Feel free to convert more.
6166 In the fullness of time, I'll do away with the STACK type altogether.
6169 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
6170 specified in <certfile> by updating the entry in the index.txt file.
6171 This way one no longer has to edit the index.txt file manually for
6172 revoking a certificate. The -revoke option does the gory details now.
6173 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
6175 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
6176 `-text' option at all and this way the `-noout -text' combination was
6177 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
6178 [Ralf S. Engelschall]
6180 *) Make sure a corresponding plain text error message exists for the
6181 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
6182 verify callback function determined that a certificate was revoked.
6183 [Ralf S. Engelschall]
6185 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
6186 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
6187 all available cipers including rc5, which was forgotten until now.
6188 In order to let the testing shell script know which algorithms
6189 are available, a new (up to now undocumented) command
6190 "openssl list-cipher-commands" is used.
6193 *) Bugfix: s_client occasionally would sleep in select() when
6194 it should have checked SSL_pending() first.
6197 *) New functions DSA_do_sign and DSA_do_verify to provide access to
6198 the raw DSA values prior to ASN.1 encoding.
6201 *) Tweaks to Configure
6202 [Niels Poppe <niels@netbox.org>]
6204 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
6208 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
6211 *) New config option to avoid instructions that are illegal on the 80386.
6212 The default code is faster, but requires at least a 486.
6215 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
6216 SSL2_SERVER_VERSION (not used at all) macros, which are now the
6217 same as SSL2_VERSION anyway.
6220 *) New "-showcerts" option for s_client.
6223 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
6224 application. Various cleanups and fixes.
6227 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
6228 modify error routines to work internally. Add error codes and PBE init
6229 to library startup routines.
6232 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
6233 packing functions to asn1 and evp. Changed function names and error
6234 codes along the way.
6237 *) PKCS12 integration: and so it begins... First of several patches to
6238 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
6239 objects to objects.h
6242 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
6243 and display support for Thawte strong extranet extension.
6246 *) Add LinuxPPC support.
6247 [Jeff Dubrule <igor@pobox.org>]
6249 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
6250 bn_div_words in alpha.s.
6251 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
6253 *) Make sure the RSA OAEP test is skipped under -DRSAref because
6254 OAEP isn't supported when OpenSSL is built with RSAref.
6255 [Ulf Moeller <ulf@fitug.de>]
6257 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
6258 so they no longer are missing under -DNOPROTO.
6259 [Soren S. Jorvang <soren@t.dk>]
6262 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
6264 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
6265 doesn't work when the session is reused. Coming soon!
6268 *) Fix a security hole, that allows sessions to be reused in the wrong
6269 context thus bypassing client cert protection! All software that uses
6270 client certs and session caches in multiple contexts NEEDS PATCHING to
6271 allow session reuse! A fuller solution is in the works.
6272 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
6274 *) Some more source tree cleanups (removed obsolete files
6275 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
6276 permission on "config" script to be executable) and a fix for the INSTALL
6278 [Ulf Moeller <ulf@fitug.de>]
6280 *) Remove some legacy and erroneous uses of malloc, free instead of
6282 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
6284 *) Make rsa_oaep_test return non-zero on error.
6285 [Ulf Moeller <ulf@fitug.de>]
6287 *) Add support for native Solaris shared libraries. Configure
6288 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
6289 if someone would make that last step automatic.
6290 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
6292 *) ctx_size was not built with the right compiler during "make links". Fixed.
6295 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
6296 except NULL ciphers". This means the default cipher list will no longer
6297 enable NULL ciphers. They need to be specifically enabled e.g. with
6298 the string "DEFAULT:eNULL".
6301 *) Fix to RSA private encryption routines: if p < q then it would
6302 occasionally produce an invalid result. This will only happen with
6303 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
6306 *) Be less restrictive and allow also `perl util/perlpath.pl
6307 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
6308 because this way one can also use an interpreter named `perl5' (which is
6309 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
6310 installed as `perl').
6311 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6313 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
6314 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6316 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
6317 advapi32.lib to Win32 build and change the pem test comparision
6318 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
6319 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
6320 and crypto/des/ede_cbcm_enc.c.
6323 *) DES quad checksum was broken on big-endian architectures. Fixed.
6326 *) Comment out two functions in bio.h that aren't implemented. Fix up the
6327 Win32 test batch file so it (might) work again. The Win32 test batch file
6328 is horrible: I feel ill....
6331 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
6332 in e_os.h. Audit of header files to check ANSI and non ANSI
6333 sections: 10 functions were absent from non ANSI section and not exported
6334 from Windows DLLs. Fixed up libeay.num for new functions.
6337 *) Make `openssl version' output lines consistent.
6338 [Ralf S. Engelschall]
6340 *) Fix Win32 symbol export lists for BIO functions: Added
6341 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
6342 to ms/libeay{16,32}.def.
6343 [Ralf S. Engelschall]
6345 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
6346 fine under Unix and passes some trivial tests I've now added. But the
6347 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
6348 added to make sure no one expects that this stuff really works in the
6349 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
6350 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
6352 [Ralf S. Engelschall]
6354 *) Fix the generation of two part addresses in perl.
6355 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
6357 *) Add config entry for Linux on MIPS.
6358 [John Tobey <jtobey@channel1.com>]
6360 *) Make links whenever Configure is run, unless we are on Windoze.
6363 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
6364 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
6368 *) Add a useful kludge to allow package maintainers to specify compiler and
6369 other platforms details on the command line without having to patch the
6370 Configure script everytime: One now can use ``perl Configure
6371 <id>:<details>'', i.e. platform ids are allowed to have details appended
6372 to them (seperated by colons). This is treated as there would be a static
6373 pre-configured entry in Configure's %table under key <id> with value
6374 <details> and ``perl Configure <id>'' is called. So, when you want to
6375 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
6376 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
6377 now, which overrides the FreeBSD-elf entry on-the-fly.
6378 [Ralf S. Engelschall]
6380 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
6383 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
6384 on the `perl Configure ...' command line. This way one can compile
6385 OpenSSL libraries with Position Independent Code (PIC) which is needed
6386 for linking it into DSOs.
6387 [Ralf S. Engelschall]
6389 *) Remarkably, export ciphers were totally broken and no-one had noticed!
6393 *) Cleaned up the LICENSE document: The official contact for any license
6394 questions now is the OpenSSL core team under openssl-core@openssl.org.
6395 And add a paragraph about the dual-license situation to make sure people
6396 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
6397 to the OpenSSL toolkit.
6398 [Ralf S. Engelschall]
6400 *) General source tree makefile cleanups: Made `making xxx in yyy...'
6401 display consistent in the source tree and replaced `/bin/rm' by `rm'.
6402 Additonally cleaned up the `make links' target: Remove unnecessary
6403 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
6404 to speed processing and no longer clutter the display with confusing
6405 stuff. Instead only the actually done links are displayed.
6406 [Ralf S. Engelschall]
6408 *) Permit null encryption ciphersuites, used for authentication only. It used
6409 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
6410 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
6414 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
6415 signed attributes when verifying signatures (this would break them),
6416 the detached data encoding was wrong and public keys obtained using
6417 X509_get_pubkey() weren't freed.
6420 *) Add text documentation for the BUFFER functions. Also added a work around
6421 to a Win95 console bug. This was triggered by the password read stuff: the
6422 last character typed gets carried over to the next fread(). If you were
6423 generating a new cert request using 'req' for example then the last
6424 character of the passphrase would be CR which would then enter the first
6428 *) Added the new `Includes OpenSSL Cryptography Software' button as
6429 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
6430 button and can be used by applications based on OpenSSL to show the
6431 relationship to the OpenSSL project.
6432 [Ralf S. Engelschall]
6434 *) Remove confusing variables in function signatures in files
6435 ssl/ssl_lib.c and ssl/ssl.h.
6436 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6438 *) Don't install bss_file.c under PREFIX/include/
6439 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6441 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
6442 functions that return function pointers and has support for NT specific
6443 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
6444 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
6445 unsigned to signed types: this was killing the Win32 compile.
6448 *) Add new certificate file to stack functions,
6449 SSL_add_dir_cert_subjects_to_stack() and
6450 SSL_add_file_cert_subjects_to_stack(). These largely supplant
6451 SSL_load_client_CA_file(), and can be used to add multiple certs easily
6452 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
6453 This means that Apache-SSL and similar packages don't have to mess around
6454 to add as many CAs as they want to the preferred list.
6457 *) Experiment with doxygen documentation. Currently only partially applied to
6459 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
6460 openssl.doxy as the configuration file.
6463 *) Get rid of remaining C++-style comments which strict C compilers hate.
6464 [Ralf S. Engelschall, pointed out by Carlos Amengual]
6466 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
6467 compiled in by default: it has problems with large keys.
6470 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
6471 DH private keys and/or callback functions which directly correspond to
6472 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
6473 is needed for applications which have to configure certificates on a
6474 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
6476 For the RSA certificate situation is makes no difference, but
6477 for the DSA certificate situation this fixes the "no shared cipher"
6478 problem where the OpenSSL cipher selection procedure failed because the
6479 temporary keys were not overtaken from the context and the API provided
6480 no way to reconfigure them.
6481 The new functions now let applications reconfigure the stuff and they
6482 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
6483 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
6484 non-public-API function ssl_cert_instantiate() is used as a helper
6485 function and also to reduce code redundancy inside ssl_rsa.c.
6486 [Ralf S. Engelschall]
6488 *) Move s_server -dcert and -dkey options out of the undocumented feature
6489 area because they are useful for the DSA situation and should be
6490 recognized by the users.
6491 [Ralf S. Engelschall]
6493 *) Fix the cipher decision scheme for export ciphers: the export bits are
6494 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
6495 SSL_EXP_MASK. So, the original variable has to be used instead of the
6496 already masked variable.
6497 [Richard Levitte <levitte@stacken.kth.se>]
6499 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
6500 [Richard Levitte <levitte@stacken.kth.se>]
6502 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
6503 from `int' to `unsigned int' because it's a length and initialized by
6504 EVP_DigestFinal() which expects an `unsigned int *'.
6505 [Richard Levitte <levitte@stacken.kth.se>]
6507 *) Don't hard-code path to Perl interpreter on shebang line of Configure
6508 script. Instead use the usual Shell->Perl transition trick.
6509 [Ralf S. Engelschall]
6511 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
6512 (in addition to RSA certificates) to match the behaviour of `openssl dsa
6513 -noout -modulus' as it's already the case for `openssl rsa -noout
6514 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
6515 currently the public key is printed (a decision which was already done by
6516 `openssl dsa -modulus' in the past) which serves a similar purpose.
6517 Additionally the NO_RSA no longer completely removes the whole -modulus
6518 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
6520 [Ralf S. Engelschall]
6522 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
6523 BIO. See the source (crypto/evp/bio_ok.c) for more info.
6524 [Arne Ansper <arne@ats.cyber.ee>]
6526 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
6527 to be added. Now both 'req' and 'ca' can use new objects defined in the
6531 *) Add cool BIO that does syslog (or event log on NT).
6532 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
6534 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
6535 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
6536 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
6537 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
6540 *) Add preliminary config info for new extension code.
6543 *) Make RSA_NO_PADDING really use no padding.
6544 [Ulf Moeller <ulf@fitug.de>]
6546 *) Generate errors when private/public key check is done.
6549 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
6550 for some CRL extensions and new objects added.
6553 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
6554 key usage extension and fuller support for authority key id.
6557 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
6558 padding method for RSA, which is recommended for new applications in PKCS
6559 #1 v2.0 (RFC 2437, October 1998).
6560 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
6561 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
6562 against Bleichbacher's attack on RSA.
6563 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
6566 *) Updates to the new SSL compression code
6567 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6569 *) Fix so that the version number in the master secret, when passed
6570 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
6571 (because the server will not accept higher), that the version number
6572 is 0x03,0x01, not 0x03,0x00
6573 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6575 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
6576 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
6577 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
6580 *) Support for RAW extensions where an arbitrary extension can be
6581 created by including its DER encoding. See apps/openssl.cnf for
6585 *) Make sure latest Perl versions don't interpret some generated C array
6586 code as Perl array code in the crypto/err/err_genc.pl script.
6587 [Lars Weber <3weber@informatik.uni-hamburg.de>]
6589 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
6590 not many people have the assembler. Various Win32 compilation fixes and
6591 update to the INSTALL.W32 file with (hopefully) more accurate Win32
6595 *) Modify configure script 'Configure' to automatically create crypto/date.h
6596 file under Win32 and also build pem.h from pem.org. New script
6597 util/mkfiles.pl to create the MINFO file on environments that can't do a
6598 'make files': perl util/mkfiles.pl >MINFO should work.
6601 *) Major rework of DES function declarations, in the pursuit of correctness
6602 and purity. As a result, many evil casts evaporated, and some weirdness,
6603 too. You may find this causes warnings in your code. Zapping your evil
6604 casts will probably fix them. Mostly.
6607 *) Fix for a typo in asn1.h. Bug fix to object creation script
6608 obj_dat.pl. It considered a zero in an object definition to mean
6609 "end of object": none of the objects in objects.h have any zeros
6610 so it wasn't spotted.
6611 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
6613 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
6614 Masking (CBCM). In the absence of test vectors, the best I have been able
6615 to do is check that the decrypt undoes the encrypt, so far. Send me test
6616 vectors if you have them.
6619 *) Correct calculation of key length for export ciphers (too much space was
6620 allocated for null ciphers). This has not been tested!
6623 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
6624 message is now correct (it understands "crypto" and "ssl" on its
6625 command line). There is also now an "update" option. This will update
6626 the util/ssleay.num and util/libeay.num files with any new functions.
6628 perl util/mkdef.pl crypto ssl update
6629 it will update them.
6632 *) Overhauled the Perl interface (perl/*):
6633 - ported BN stuff to OpenSSL's different BN library
6634 - made the perl/ source tree CVS-aware
6635 - renamed the package from SSLeay to OpenSSL (the files still contain
6636 their history because I've copied them in the repository)
6637 - removed obsolete files (the test scripts will be replaced
6638 by better Test::Harness variants in the future)
6639 [Ralf S. Engelschall]
6641 *) First cut for a very conservative source tree cleanup:
6642 1. merge various obsolete readme texts into doc/ssleay.txt
6643 where we collect the old documents and readme texts.
6644 2. remove the first part of files where I'm already sure that we no
6645 longer need them because of three reasons: either they are just temporary
6646 files which were left by Eric or they are preserved original files where
6647 I've verified that the diff is also available in the CVS via "cvs diff
6648 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
6649 the crypto/md/ stuff).
6650 [Ralf S. Engelschall]
6652 *) More extension code. Incomplete support for subject and issuer alt
6653 name, issuer and authority key id. Change the i2v function parameters
6654 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
6655 what that's for :-) Fix to ASN1 macro which messed up
6656 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
6659 *) Preliminary support for ENUMERATED type. This is largely copied from the
6663 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
6664 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6666 *) Make sure `make rehash' target really finds the `openssl' program.
6667 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6669 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
6670 like to hear about it if this slows down other processors.
6673 *) Add CygWin32 platform information to Configure script.
6674 [Alan Batie <batie@aahz.jf.intel.com>]
6676 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
6677 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
6679 *) New program nseq to manipulate netscape certificate sequences
6682 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
6686 *) Fixes to BN code. Previously the default was to define BN_RECURSION
6687 but the BN code had some problems that would cause failures when
6688 doing certificate verification and some other functions.
6689 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6691 *) Add ASN1 and PEM code to support netscape certificate sequences.
6694 *) Add ASN1 and PEM code to support netscape certificate sequences.
6697 *) Add several PKIX and private extended key usage OIDs.
6700 *) Modify the 'ca' program to handle the new extension code. Modify
6701 openssl.cnf for new extension format, add comments.
6704 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
6705 and add a sample to openssl.cnf so req -x509 now adds appropriate
6709 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
6710 error code, add initial support to X509_print() and x509 application.
6713 *) Takes a deep breath and start addding X509 V3 extension support code. Add
6714 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
6715 stuff is currently isolated and isn't even compiled yet.
6718 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
6719 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
6720 Removed the versions check from X509 routines when loading extensions:
6721 this allows certain broken certificates that don't set the version
6722 properly to be processed.
6725 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
6726 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
6727 can still be regenerated with "make depend".
6730 *) Spelling mistake in C version of CAST-128.
6731 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
6733 *) Changes to the error generation code. The perl script err-code.pl
6734 now reads in the old error codes and retains the old numbers, only
6735 adding new ones if necessary. It also only changes the .err files if new
6736 codes are added. The makefiles have been modified to only insert errors
6737 when needed (to avoid needlessly modifying header files). This is done
6738 by only inserting errors if the .err file is newer than the auto generated
6739 C file. To rebuild all the error codes from scratch (the old behaviour)
6740 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
6741 or delete all the .err files.
6744 *) CAST-128 was incorrectly implemented for short keys. The C version has
6745 been fixed, but is untested. The assembler versions are also fixed, but
6746 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
6747 to regenerate it if needed.
6748 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
6749 Hagino <itojun@kame.net>]
6751 *) File was opened incorrectly in randfile.c.
6752 [Ulf Möller <ulf@fitug.de>]
6754 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
6755 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
6756 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
6757 al: it's just almost always a UTCTime. Note this patch adds new error
6758 codes so do a "make errors" if there are problems.
6761 *) Correct Linux 1 recognition in config.
6762 [Ulf Möller <ulf@fitug.de>]
6764 *) Remove pointless MD5 hash when using DSA keys in ca.
6765 [Anonymous <nobody@replay.com>]
6767 *) Generate an error if given an empty string as a cert directory. Also
6768 generate an error if handed NULL (previously returned 0 to indicate an
6769 error, but didn't set one).
6770 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
6772 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
6775 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
6776 parameters. This was causing a warning which killed off the Win32 compile.
6779 *) Remove C++ style comments from crypto/bn/bn_local.h.
6780 [Neil Costigan <neil.costigan@celocom.com>]
6782 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
6783 based on a text string, looking up short and long names and finally
6784 "dot" format. The "dot" format stuff didn't work. Added new function
6785 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
6786 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
6787 OID is not part of the table.
6790 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
6791 X509_LOOKUP_by_alias().
6794 *) Sort openssl functions by name.
6797 *) Get the gendsa program working (hopefully) and add it to app list. Remove
6798 encryption from sample DSA keys (in case anyone is interested the password
6802 *) Make _all_ *_free functions accept a NULL pointer.
6803 [Frans Heymans <fheymans@isaserver.be>]
6805 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
6807 [Anonymous <nobody@replay.com>]
6809 *) s_server should send the CAfile as acceptable CAs, not its own cert.
6810 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6812 *) Don't blow it for numeric -newkey arguments to apps/req.
6813 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6815 *) Temp key "for export" tests were wrong in s3_srvr.c.
6816 [Anonymous <nobody@replay.com>]
6818 *) Add prototype for temp key callback functions
6819 SSL_CTX_set_tmp_{rsa,dh}_callback().
6822 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
6823 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
6826 *) X509_name_add_entry() freed the wrong thing after an error.
6827 [Arne Ansper <arne@ats.cyber.ee>]
6829 *) rsa_eay.c would attempt to free a NULL context.
6830 [Arne Ansper <arne@ats.cyber.ee>]
6832 *) BIO_s_socket() had a broken should_retry() on Windoze.
6833 [Arne Ansper <arne@ats.cyber.ee>]
6835 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
6836 [Arne Ansper <arne@ats.cyber.ee>]
6838 *) Make sure the already existing X509_STORE->depth variable is initialized
6839 in X509_STORE_new(), but document the fact that this variable is still
6840 unused in the certificate verification process.
6841 [Ralf S. Engelschall]
6843 *) Fix the various library and apps files to free up pkeys obtained from
6844 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
6847 *) Fix reference counting in X509_PUBKEY_get(). This makes
6848 demos/maurice/example2.c work, amongst others, probably.
6849 [Steve Henson and Ben Laurie]
6851 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
6852 `openssl' and second, the shortcut symlinks for the `openssl <command>'
6853 are no longer created. This way we have a single and consistent command
6854 line interface `openssl <command>', similar to `cvs <command>'.
6855 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
6857 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
6858 BIT STRING wrapper always have zero unused bits.
6861 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
6864 *) Make the top-level INSTALL documentation easier to understand.
6867 *) Makefiles updated to exit if an error occurs in a sub-directory
6868 make (including if user presses ^C) [Paul Sutton]
6870 *) Make Montgomery context stuff explicit in RSA data structure.
6873 *) Fix build order of pem and err to allow for generated pem.h.
6876 *) Fix renumbering bug in X509_NAME_delete_entry().
6879 *) Enhanced the err-ins.pl script so it makes the error library number
6880 global and can add a library name. This is needed for external ASN1 and
6881 other error libraries.
6884 *) Fixed sk_insert which never worked properly.
6887 *) Fix ASN1 macros so they can handle indefinite length construted
6888 EXPLICIT tags. Some non standard certificates use these: they can now
6892 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
6893 into a single doc/ssleay.txt bundle. This way the information is still
6894 preserved but no longer messes up this directory. Now it's new room for
6895 the new set of documenation files.
6896 [Ralf S. Engelschall]
6898 *) SETs were incorrectly DER encoded. This was a major pain, because they
6899 shared code with SEQUENCEs, which aren't coded the same. This means that
6900 almost everything to do with SETs or SEQUENCEs has either changed name or
6901 number of arguments.
6902 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
6904 *) Fix test data to work with the above.
6907 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
6908 was already fixed by Eric for 0.9.1 it seems.
6909 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
6911 *) Autodetect FreeBSD3.
6914 *) Fix various bugs in Configure. This affects the following platforms:
6918 unixware-2.0-pentium
6922 *) Eliminate generated files from CVS. Reorder tests to regenerate files
6923 before they are needed.
6926 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
6930 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
6932 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
6933 changed SSLeay to OpenSSL in version strings.
6934 [Ralf S. Engelschall]
6936 *) Some fixups to the top-level documents.
6939 *) Fixed the nasty bug where rsaref.h was not found under compile-time
6940 because the symlink to include/ was missing.
6941 [Ralf S. Engelschall]
6943 *) Incorporated the popular no-RSA/DSA-only patches
6944 which allow to compile a RSA-free SSLeay.
6945 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
6947 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
6948 when "ssleay" is still not found.
6949 [Ralf S. Engelschall]
6951 *) Added more platforms to Configure: Cray T3E, HPUX 11,
6952 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
6954 *) Updated the README file.
6955 [Ralf S. Engelschall]
6957 *) Added various .cvsignore files in the CVS repository subdirs
6958 to make a "cvs update" really silent.
6959 [Ralf S. Engelschall]
6961 *) Recompiled the error-definition header files and added
6962 missing symbols to the Win32 linker tables.
6963 [Ralf S. Engelschall]
6965 *) Cleaned up the top-level documents;
6966 o new files: CHANGES and LICENSE
6967 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
6968 o merged COPYRIGHT into LICENSE
6969 o removed obsolete TODO file
6970 o renamed MICROSOFT to INSTALL.W32
6971 [Ralf S. Engelschall]
6973 *) Removed dummy files from the 0.9.1b source tree:
6974 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
6975 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
6976 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
6977 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
6978 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
6979 [Ralf S. Engelschall]
6981 *) Added various platform portability fixes.
6984 *) The Genesis of the OpenSSL rpject:
6985 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
6986 Young and Tim J. Hudson created while they were working for C2Net until
6988 [The OpenSSL Project]
6991 Changes between 0.9.0b and 0.9.1b [not released]
6993 *) Updated a few CA certificates under certs/
6996 *) Changed some BIGNUM api stuff.
6999 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
7000 DGUX x86, Linux Alpha, etc.
7003 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
7004 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
7008 *) Add -strparse option to asn1pars program which parses nested
7010 [Dr Stephen Henson <shenson@bigfoot.com>]
7012 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
7015 *) DSA fix for "ca" program.
7018 *) Added "-genkey" option to "dsaparam" program.
7021 *) Added RIPE MD160 (rmd160) message digest.
7024 *) Added -a (all) option to "ssleay version" command.
7027 *) Added PLATFORM define which is the id given to Configure.
7030 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
7033 *) Extended the ASN.1 parser routines.
7036 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
7039 *) Added a BN_CTX to the BN library.
7042 *) Fixed the weak key values in DES library
7045 *) Changed API in EVP library for cipher aliases.
7048 *) Added support for RC2/64bit cipher.
7051 *) Converted the lhash library to the crypto/mem.c functions.
7054 *) Added more recognized ASN.1 object ids.
7057 *) Added more RSA padding checks for SSL/TLS.
7060 *) Added BIO proxy/filter functionality.
7063 *) Added extra_certs to SSL_CTX which can be used
7064 send extra CA certificates to the client in the CA cert chain sending
7065 process. It can be configured with SSL_CTX_add_extra_chain_cert().
7068 *) Now Fortezza is denied in the authentication phase because
7069 this is key exchange mechanism is not supported by SSLeay at all.
7072 *) Additional PKCS1 checks.
7075 *) Support the string "TLSv1" for all TLS v1 ciphers.
7078 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
7079 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
7082 *) Fixed a few memory leaks.
7085 *) Fixed various code and comment typos.
7088 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
7089 bytes sent in the client random.
7090 [Edward Bishop <ebishop@spyglass.com>]