5 Changes between 0.9.7 and 0.9.8 [xx XXX 2002]
7 *) Add Makefile.shared, a helper makefile to build shared
8 libraries. Addapt Makefile.org.
11 *) Add version info to Win32 DLLs.
12 [Peter 'Luna' Runestig" <peter@runestig.com>]
14 *) Add new 'medium level' PKCS#12 API. Certificates and keys
15 can be added using this API to created arbitrary PKCS#12
16 files while avoiding the low level API.
18 New options to PKCS12_create(), key or cert can be NULL and
19 will then be omitted from the output file. The encryption
20 algorithm NIDs can be set to -1 for no encryption, the mac
21 iteration count can be set to 0 to omit the mac.
23 Enhance pkcs12 utility by making the -nokeys and -nocerts
24 options work when creating a PKCS#12 file. New option -nomac
25 to omit the mac, NONE can be set for an encryption algorithm.
26 New code is modified to use the enhanced PKCS12_create()
27 instead of the low level API.
30 *) Extend ASN1 encoder to support indefinite length constructed
31 encoding. This can output sequences tags and octet strings in
32 this form. Modify pk7_asn1.c to support indefinite length
33 encoding. This is experimental and needs additional code to
34 be useful, such as an ASN1 bio and some enhanced streaming
37 Extend template encode functionality so that tagging is passed
38 down to the template encoder.
41 *) Let 'openssl req' fail if an argument to '-newkey' is not
42 recognized instead of using RSA as a default.
45 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
46 As these are not official, they are not included in "ALL";
47 the "ECCdraft" ciphersuite group alias can be used to select them.
48 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
50 *) Add ECDH engine support.
51 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
53 *) Add ECDH in new directory crypto/ecdh/.
54 TODO: more general interface (return x coordinate, not its hash)
55 TODO: bug: pad x with leading zeros if necessary
56 [Douglas Stebila (Sun Microsystems Laboratories)]
58 *) Let BN_rand_range() abort with an error after 100 iterations
59 without success (which indicates a broken PRNG).
62 *) Change BN_mod_sqrt() so that it verifies that the input value
63 is really the square of the return value. (Previously,
64 BN_mod_sqrt would show GIGO behaviour.)
67 *) Add named elliptic curves over binary fields from X9.62, SECG,
68 and WAP/WTLS; add OIDs that were still missing.
70 [Sheueling Chang Shantz and Douglas Stebila
71 (Sun Microsystems Laboratories)]
73 *) Extend the EC library for elliptic curves over binary fields
74 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
81 EC_GROUP_new_curve_GF2m
82 EC_GROUP_set_curve_GF2m
83 EC_GROUP_get_curve_GF2m
84 EC_POINT_set_affine_coordinates_GF2m
85 EC_POINT_get_affine_coordinates_GF2m
86 EC_POINT_set_compressed_coordinates_GF2m
88 Point compression for binary fields is disabled by default for
89 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
92 As binary polynomials are represented as BIGNUMs, various members
93 of the EC_GROUP and EC_POINT data structures can be shared
94 between the implementations for prime fields and binary fields;
95 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
96 are essentially identical to their ..._GFp counterparts.
97 (For simplicity, the '..._GFp' prefix has been dropped from
98 various internal method names.)
100 An internal 'field_div' method (similar to 'field_mul' and
101 'field_sqr') has been added; this is used only for binary fields.
103 [Sheueling Chang Shantz and Douglas Stebila
104 (Sun Microsystems Laboratories)]
106 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
107 through methods ('mul', 'precompute_mult').
109 The generic implementations (now internally called 'ec_wNAF_mul'
110 and 'ec_wNAF_precomputed_mult') remain the default if these
111 methods are undefined.
113 [Sheueling Chang Shantz and Douglas Stebila
114 (Sun Microsystems Laboratories)]
116 *) New function EC_GROUP_get_degree, which is defined through
117 EC_METHOD. For curves over prime fields, this returns the bit
118 length of the modulus.
120 [Sheueling Chang Shantz and Douglas Stebila
121 (Sun Microsystems Laboratories)]
123 *) New functions EC_GROUP_dup, EC_POINT_dup.
124 (These simply call ..._new and ..._copy).
126 [Sheueling Chang Shantz and Douglas Stebila
127 (Sun Microsystems Laboratories)]
129 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
130 Polynomials are represented as BIGNUMs (where the sign bit is not
131 used) in the following functions [macros]:
134 BN_GF2m_sub [= BN_GF2m_add]
135 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
136 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
137 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
139 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
140 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
141 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
142 BN_GF2m_cmp [= BN_ucmp]
144 (Note that only the 'mod' functions are actually for fields GF(2^m).
145 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
147 For some functions, an the irreducible polynomial defining a
148 field can be given as an 'unsigned int[]' with strictly
149 decreasing elements giving the indices of those bits that are set;
150 i.e., p[] represents the polynomial
151 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
153 p[0] > p[1] > ... > p[k] = 0.
154 This applies to the following functions:
159 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
160 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
163 BN_GF2m_mod_solve_quad_arr
167 Conversion can be performed by the following functions:
172 bntest.c has additional tests for binary polynomial arithmetic.
174 Two implementations for BN_GF2m_mod_div() are available.
175 The default algorithm simply uses BN_GF2m_mod_inv() and
176 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
177 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
178 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
180 [Sheueling Chang Shantz and Douglas Stebila
181 (Sun Microsystems Laboratories)]
183 *) Add new error code 'ERR_R_DISABLED' that can be used when some
184 functionality is disabled at compile-time.
185 [Douglas Stebila <douglas.stebila@sun.com>]
187 *) Change default behaviour of 'openssl asn1parse' so that more
188 information is visible when viewing, e.g., a certificate:
190 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
191 mode the content of non-printable OCTET STRINGs is output in a
192 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
193 avoid the appearance of a printable string.
194 [Nils Larsch <nla@trustcenter.de>]
196 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
198 EC_GROUP_set_asn1_flag()
199 EC_GROUP_get_asn1_flag()
200 EC_GROUP_set_point_conversion_form()
201 EC_GROUP_get_point_conversion_form()
202 These control ASN1 encoding details:
203 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
204 has been set to OPENSSL_EC_NAMED_CURVE.
205 - Points are encoded in uncompressed form by default; options for
206 asn1_for are as for point2oct, namely
207 POINT_CONVERSION_COMPRESSED
208 POINT_CONVERSION_UNCOMPRESSED
209 POINT_CONVERSION_HYBRID
211 Also add 'seed' and 'seed_len' members to EC_GROUP with access
215 EC_GROUP_get_seed_len()
216 This is used only for ASN1 purposes (so far).
217 [Nils Larsch <nla@trustcenter.de>]
219 *) Add 'field_type' member to EC_METHOD, which holds the NID
220 of the appropriate field type OID. The new function
221 EC_METHOD_get_field_type() returns this value.
222 [Nils Larsch <nla@trustcenter.de>]
229 providing useful interfaces to EC_POINT_point2oct() and
230 EC_POINT_oct2point().
231 [Nils Larsch <nla@trustcenter.de>]
233 *) Change internals of the EC library so that the functions
234 EC_GROUP_set_generator()
235 EC_GROUP_get_generator()
237 EC_GROUP_get_cofactor()
238 are implemented directly in crypto/ec/ec_lib.c and not dispatched
239 to methods, which would lead to unnecessary code duplication when
240 adding different types of curves.
241 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
243 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
244 arithmetic, and such that modified wNAFs are generated
245 (which avoid length expansion in many cases).
248 *) Add a function EC_GROUP_check_discriminant() (defined via
249 EC_METHOD) that verifies that the curve discriminant is non-zero.
251 Add a function EC_GROUP_check() that makes some sanity tests
252 on a EC_GROUP, its generator and order. This includes
253 EC_GROUP_check_discriminant().
254 [Nils Larsch <nla@trustcenter.de>]
256 *) Add ECDSA in new directory crypto/ecdsa/.
258 Add applications 'openssl ecparam' and 'openssl ecdsa'
259 (these are based on 'openssl dsaparam' and 'openssl dsa').
261 ECDSA support is also included in various other files across the
262 library. Most notably,
263 - 'openssl req' now has a '-newkey ecdsa:file' option;
264 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
265 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
266 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
267 them suitable for ECDSA where domain parameters must be
268 extracted before the specific public key;
269 - ECDSA engine support has been added.
270 [Nils Larsch <nla@trustcenter.de>]
272 *) Include some named elliptic curves, and add OIDs from X9.62,
273 SECG, and WAP/WTLS. Each curve can be obtained from the new
275 EC_GROUP_new_by_nid(),
276 and the list of available named curves can be obtained with
277 EC_get_builtin_curves().
278 Also add a 'curve_name' member to EC_GROUP objects, which can be
282 [Nils Larsch <nla@trustcenter.de, Bodo Moeller]
284 Changes between 0.9.6h and 0.9.7 [XX xxx 2002]
286 *) Change from security patch (see 0.9.6e below) that did not affect
287 the 0.9.6 release series:
289 Remote buffer overflow in SSL3 protocol - an attacker could
290 supply an oversized master key in Kerberos-enabled versions.
294 *) Change the SSL kerb5 codes to match RFC 2712.
297 *) Make -nameopt work fully for req and add -reqopt switch.
298 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
300 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
301 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
303 *) Make sure tests can be performed even if the corresponding algorithms
304 have been removed entirely. This was also the last step to make
305 OpenSSL compilable with DJGPP under all reasonable conditions.
306 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
308 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
309 to allow version independent disabling of normally unselected ciphers,
310 which may be activated as a side-effect of selecting a single cipher.
312 (E.g., cipher list string "RSA" enables ciphersuites that are left
313 out of "ALL" because they do not provide symmetric encryption.
314 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
315 [Lutz Jaenicke, Bodo Moeller]
317 *) Add appropriate support for separate platform-dependent build
318 directories. The recommended way to make a platform-dependent
319 build directory is the following (tested on Linux), maybe with
322 # Place yourself outside of the OpenSSL source tree. In
323 # this example, the environment variable OPENSSL_SOURCE
324 # is assumed to contain the absolute OpenSSL source directory.
325 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
326 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
327 (cd $OPENSSL_SOURCE; find . -type f -o -type l) | while read F; do
328 mkdir -p `dirname $F`
329 ln -s $OPENSSL_SOURCE/$F $F
332 To be absolutely sure not to disturb the source tree, a "make clean"
333 is a good thing. If it isn't successfull, don't worry about it,
334 it probably means the source directory is very clean.
337 *) Make sure any ENGINE control commands make local copies of string
338 pointers passed to them whenever necessary. Otherwise it is possible
339 the caller may have overwritten (or deallocated) the original string
340 data when a later ENGINE operation tries to use the stored values.
341 [Götz Babin-Ebell <babinebell@trustcenter.de>]
343 *) Improve diagnostics in file reading and command-line digests.
344 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
346 *) Add AES modes CFB and OFB to the object database. Correct an
347 error in AES-CFB decryption.
350 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
351 allows existing EVP_CIPHER_CTX structures to be reused after
352 calling EVP_*Final(). This behaviour is used by encryption
353 BIOs and some applications. This has the side effect that
354 applications must explicitly clean up cipher contexts with
355 EVP_CIPHER_CTX_cleanup() or they will leak memory.
358 *) Check the values of dna and dnb in bn_mul_recursive before calling
359 bn_mul_comba (a non zero value means the a or b arrays do not contain
360 n2 elements) and fallback to bn_mul_normal if either is not zero.
363 *) Fix escaping of non-ASCII characters when using the -subj option
364 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
367 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
368 form for "surname", serialNumber has no short form.
369 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
370 therefore remove "mail" short name for "internet 7".
371 The OID for unique identifiers in X509 certificates is
372 x500UniqueIdentifier, not uniqueIdentifier.
373 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
376 *) Add an "init" command to the ENGINE config module and auto initialize
377 ENGINEs. Without any "init" command the ENGINE will be initialized
378 after all ctrl commands have been executed on it. If init=1 the
379 ENGINE is initailized at that point (ctrls before that point are run
380 on the uninitialized ENGINE and after on the initialized one). If
381 init=0 then the ENGINE will not be iniatialized at all.
384 *) Fix the 'app_verify_callback' interface so that the user-defined
385 argument is actually passed to the callback: In the
386 SSL_CTX_set_cert_verify_callback() prototype, the callback
387 declaration has been changed from
390 int (*cb)(X509_STORE_CTX *,void *);
391 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
392 i=s->ctx->app_verify_callback(&ctx)
393 has been changed into
394 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
396 To update applications using SSL_CTX_set_cert_verify_callback(),
397 a dummy argument can be added to their callback functions.
398 [D. K. Smetters <smetters@parc.xerox.com>]
400 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
401 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
403 *) Add and OPENSSL_LOAD_CONF define which will cause
404 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
405 This allows older applications to transparently support certain
406 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
407 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
408 load the config file and OPENSSL_add_all_algorithms_conf() which will
409 always load it have also been added.
412 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
413 Adjust NIDs and EVP layer.
414 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
416 *) Config modules support in openssl utility.
418 Most commands now load modules from the config file,
419 though in a few (such as version) this isn't done
420 because it couldn't be used for anything.
422 In the case of ca and req the config file used is
423 the same as the utility itself: that is the -config
424 command line option can be used to specify an
428 *) Move default behaviour from OPENSSL_config(). If appname is NULL
429 use "openssl_conf" if filename is NULL use default openssl config file.
432 *) Add an argument to OPENSSL_config() to allow the use of an alternative
433 config section name. Add a new flag to tolerate a missing config file
434 and move code to CONF_modules_load_file().
437 *) Support for crypto accelerator cards from Accelerated Encryption
438 Processing, www.aep.ie. (Use engine 'aep')
439 The support was copied from 0.9.6c [engine] and adapted/corrected
440 to work with the new engine framework.
441 [AEP Inc. and Richard Levitte]
443 *) Support for SureWare crypto accelerator cards from Baltimore
444 Technologies. (Use engine 'sureware')
445 The support was copied from 0.9.6c [engine] and adapted
446 to work with the new engine framework.
449 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
450 make the newer ENGINE framework commands for the CHIL engine work.
451 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
453 *) Make it possible to produce shared libraries on ReliantUNIX.
454 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
456 *) Add the configuration target debug-linux-ppro.
457 Make 'openssl rsa' use the general key loading routines
458 implemented in apps.c, and make those routines able to
459 handle the key format FORMAT_NETSCAPE and the variant
461 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
463 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
464 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
466 *) Add -keyform to rsautl, and document -engine.
467 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
469 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
470 BIO_R_NO_SUCH_FILE error code rather than the generic
471 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
476 ERR_peek_last_error_line
477 ERR_peek_last_error_line_data.
481 ERR_peek_error_line_data,
482 but report on the latest error recorded rather than the first one
483 still in the error queue.
484 [Ben Laurie, Bodo Moeller]
486 *) default_algorithms option in ENGINE config module. This allows things
488 default_algorithms = ALL
489 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
492 *) Prelminary ENGINE config module.
495 *) New experimental application configuration code.
498 *) Change the AES code to follow the same name structure as all other
499 symmetric ciphers, and behave the same way. Move everything to
500 the directory crypto/aes, thereby obsoleting crypto/rijndael.
501 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
503 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
504 [Ben Laurie and Theo de Raadt]
506 *) Add option to output public keys in req command.
507 [Massimiliano Pala madwolf@openca.org]
509 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
510 (up to about 10% better than before for P-192 and P-224).
513 *) New functions/macros
515 SSL_CTX_set_msg_callback(ctx, cb)
516 SSL_CTX_set_msg_callback_arg(ctx, arg)
517 SSL_set_msg_callback(ssl, cb)
518 SSL_set_msg_callback_arg(ssl, arg)
520 to request calling a callback function
522 void cb(int write_p, int version, int content_type,
523 const void *buf, size_t len, SSL *ssl, void *arg)
525 whenever a protocol message has been completely received
526 (write_p == 0) or sent (write_p == 1). Here 'version' is the
527 protocol version according to which the SSL library interprets
528 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
529 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
530 the content type as defined in the SSL 3.0/TLS 1.0 protocol
531 specification (change_cipher_spec(20), alert(21), handshake(22)).
532 'buf' and 'len' point to the actual message, 'ssl' to the
533 SSL object, and 'arg' is the application-defined value set by
534 SSL[_CTX]_set_msg_callback_arg().
536 'openssl s_client' and 'openssl s_server' have new '-msg' options
537 to enable a callback that displays all protocol messages.
540 *) Change the shared library support so shared libraries are built as
541 soon as the corresponding static library is finished, and thereby get
542 openssl and the test programs linked against the shared library.
543 This still only happens when the keyword "shard" has been given to
544 the configuration scripts.
546 NOTE: shared library support is still an experimental thing, and
547 backward binary compatibility is still not guaranteed.
548 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
550 *) Add support for Subject Information Access extension.
551 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
553 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
554 additional bytes when new memory had to be allocated, not just
555 when reusing an existing buffer.
558 *) New command line and configuration option 'utf8' for the req command.
559 This allows field values to be specified as UTF8 strings.
562 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
563 runs for the former and machine-readable output for the latter.
566 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
567 of the e-mail address in the DN (i.e., it will go into a certificate
568 extension only). The new configuration file option 'email_in_dn = no'
570 [Massimiliano Pala madwolf@openca.org]
572 *) Change all functions with names starting with des_ to be starting
573 with DES_ instead. Add wrappers that are compatible with libdes,
574 but are named _ossl_old_des_*. Finally, add macros that map the
575 des_* symbols to the corresponding _ossl_old_des_* if libdes
576 compatibility is desired. If OpenSSL 0.9.6c compatibility is
577 desired, the des_* symbols will be mapped to DES_*, with one
580 Since we provide two compatibility mappings, the user needs to
581 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
582 compatibility is desired. The default (i.e., when that macro
583 isn't defined) is OpenSSL 0.9.6c compatibility.
585 There are also macros that enable and disable the support of old
586 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
587 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
588 are defined, the default will apply: to support the old des routines.
590 In either case, one must include openssl/des.h to get the correct
591 definitions. Do not try to just include openssl/des_old.h, that
594 NOTE: This is a major break of an old API into a new one. Software
595 authors are encouraged to switch to the DES_ style functions. Some
596 time in the future, des_old.h and the libdes compatibility functions
597 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
598 default), and then completely removed.
601 *) Test for certificates which contain unsupported critical extensions.
602 If such a certificate is found during a verify operation it is
603 rejected by default: this behaviour can be overridden by either
604 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
605 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
606 X509_supported_extension() has also been added which returns 1 if a
607 particular extension is supported.
610 *) Modify the behaviour of EVP cipher functions in similar way to digests
611 to retain compatibility with existing code.
614 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
615 compatibility with existing code. In particular the 'ctx' parameter does
616 not have to be to be initialized before the call to EVP_DigestInit() and
617 it is tidied up after a call to EVP_DigestFinal(). New function
618 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
619 EVP_MD_CTX_copy() changed to not require the destination to be
620 initialized valid and new function EVP_MD_CTX_copy_ex() added which
621 requires the destination to be valid.
623 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
624 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
627 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
628 so that complete 'Handshake' protocol structures are kept in memory
629 instead of overwriting 'msg_type' and 'length' with 'body' data.
632 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
633 [Massimo Santin via Richard Levitte]
635 *) Major restructuring to the underlying ENGINE code. This includes
636 reduction of linker bloat, separation of pure "ENGINE" manipulation
637 (initialisation, etc) from functionality dealing with implementations
638 of specific crypto iterfaces. This change also introduces integrated
639 support for symmetric ciphers and digest implementations - so ENGINEs
640 can now accelerate these by providing EVP_CIPHER and EVP_MD
641 implementations of their own. This is detailed in crypto/engine/README
642 as it couldn't be adequately described here. However, there are a few
643 API changes worth noting - some RSA, DSA, DH, and RAND functions that
644 were changed in the original introduction of ENGINE code have now
645 reverted back - the hooking from this code to ENGINE is now a good
646 deal more passive and at run-time, operations deal directly with
647 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
648 dereferencing through an ENGINE pointer any more. Also, the ENGINE
649 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
650 they were not being used by the framework as there is no concept of a
651 BIGNUM_METHOD and they could not be generalised to the new
652 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
653 ENGINE_cpy() has been removed as it cannot be consistently defined in
657 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
660 *) Change mkdef.pl to sort symbols that get the same entry number,
661 and make sure the automatically generated functions ERR_load_*
662 become part of libeay.num as well.
665 *) New function SSL_renegotiate_pending(). This returns true once
666 renegotiation has been requested (either SSL_renegotiate() call
667 or HelloRequest/ClientHello receveived from the peer) and becomes
668 false once a handshake has been completed.
669 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
670 sends a HelloRequest, but does not ensure that a handshake takes
671 place. SSL_renegotiate_pending() is useful for checking if the
672 client has followed the request.)
675 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
676 By default, clients may request session resumption even during
677 renegotiation (if session ID contexts permit); with this option,
678 session resumption is possible only in the first handshake.
680 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
681 more bits available for options that should not be part of
682 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
685 *) Add some demos for certificate and certificate request creation.
688 *) Make maximum certificate chain size accepted from the peer application
689 settable (SSL*_get/set_max_cert_list()), as proposed by
690 "Douglas E. Engert" <deengert@anl.gov>.
693 *) Add support for shared libraries for Unixware-7
694 (Boyd Lynn Gerber <gerberb@zenez.com>).
697 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
698 be done prior to destruction. Use this to unload error strings from
699 ENGINEs that load their own error strings. NB: This adds two new API
700 functions to "get" and "set" this destroy handler in an ENGINE.
703 *) Alter all existing ENGINE implementations (except "openssl" and
704 "openbsd") to dynamically instantiate their own error strings. This
705 makes them more flexible to be built both as statically-linked ENGINEs
706 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
707 Also, add stub code to each that makes building them as self-contained
708 shared-libraries easier (see README.ENGINE).
711 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
712 implementations into applications that are completely implemented in
713 self-contained shared-libraries. The "dynamic" ENGINE exposes control
714 commands that can be used to configure what shared-library to load and
715 to control aspects of the way it is handled. Also, made an update to
716 the README.ENGINE file that brings its information up-to-date and
717 provides some information and instructions on the "dynamic" ENGINE
718 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
721 *) Make it possible to unload ranges of ERR strings with a new
722 "ERR_unload_strings" function.
725 *) Add a copy() function to EVP_MD.
728 *) Make EVP_MD routines take a context pointer instead of just the
729 md_data void pointer.
732 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
733 that the digest can only process a single chunk of data
734 (typically because it is provided by a piece of
735 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
736 is only going to provide a single chunk of data, and hence the
737 framework needn't accumulate the data for oneshot drivers.
740 *) As with "ERR", make it possible to replace the underlying "ex_data"
741 functions. This change also alters the storage and management of global
742 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
743 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
744 index counters. The API functions that use this state have been changed
745 to take a "class_index" rather than pointers to the class's local STACK
746 and counter, and there is now an API function to dynamically create new
747 classes. This centralisation allows us to (a) plug a lot of the
748 thread-safety problems that existed, and (b) makes it possible to clean
749 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
750 such data would previously have always leaked in application code and
751 workarounds were in place to make the memory debugging turn a blind eye
752 to it. Application code that doesn't use this new function will still
753 leak as before, but their memory debugging output will announce it now
754 rather than letting it slide.
756 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
757 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
758 has a return value to indicate success or failure.
761 *) Make it possible to replace the underlying "ERR" functions such that the
762 global state (2 LHASH tables and 2 locks) is only used by the "default"
763 implementation. This change also adds two functions to "get" and "set"
764 the implementation prior to it being automatically set the first time
765 any other ERR function takes place. Ie. an application can call "get",
766 pass the return value to a module it has just loaded, and that module
767 can call its own "set" function using that value. This means the
768 module's "ERR" operations will use (and modify) the error state in the
769 application and not in its own statically linked copy of OpenSSL code.
772 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
773 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
774 the operation, and provides a more encapsulated way for external code
775 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
776 to use these functions rather than manually incrementing the counts.
778 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
781 *) Add EVP test program.
784 *) Add symmetric cipher support to ENGINE. Expect the API to change!
787 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
788 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
789 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
790 These allow a CRL to be built without having to access X509_CRL fields
791 directly. Modify 'ca' application to use new functions.
794 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
795 bug workarounds. Rollback attack detection is a security feature.
796 The problem will only arise on OpenSSL servers when TLSv1 is not
797 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
798 Software authors not wanting to support TLSv1 will have special reasons
799 for their choice and can explicitly enable this option.
800 [Bodo Moeller, Lutz Jaenicke]
802 *) Rationalise EVP so it can be extended: don't include a union of
803 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
804 (similar to those existing for EVP_CIPHER_CTX).
809 EVP_MD_CTX_init(&md); /* new function call */
810 EVP_DigestInit(&md, EVP_sha1());
811 EVP_DigestUpdate(&md, in, len);
812 EVP_DigestFinal(&md, out, NULL);
813 EVP_MD_CTX_cleanup(&md); /* new function call */
817 *) Make DES key schedule conform to the usual scheme, as well as
818 correcting its structure. This means that calls to DES functions
819 now have to pass a pointer to a des_key_schedule instead of a
820 plain des_key_schedule (which was actually always a pointer
825 des_set_key_checked(..., &ks);
826 des_ncbc_encrypt(..., &ks, ...);
828 (Note that a later change renames 'des_...' into 'DES_...'.)
831 *) Initial reduction of linker bloat: the use of some functions, such as
832 PEM causes large amounts of unused functions to be linked in due to
833 poor organisation. For example pem_all.c contains every PEM function
834 which has a knock on effect of linking in large amounts of (unused)
835 ASN1 code. Grouping together similar functions and splitting unrelated
836 functions prevents this.
839 *) Cleanup of EVP macros.
842 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
846 *) Add initial OCSP responder support to ocsp application. The
847 revocation information is handled using the text based index
848 use by the ca application. The responder can either handle
849 requests generated internally, supplied in files (for example
850 via a CGI script) or using an internal minimal server.
853 *) Add configuration choices to get zlib compression for TLS.
856 *) Changes to Kerberos SSL for RFC 2712 compliance:
857 1. Implemented real KerberosWrapper, instead of just using
858 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
859 2. Implemented optional authenticator field of KerberosWrapper.
861 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
862 and authenticator structs; see crypto/krb5/.
864 Generalized Kerberos calls to support multiple Kerberos libraries.
865 [Vern Staats <staatsvr@asc.hpc.mil>,
866 Jeffrey Altman <jaltman@columbia.edu>
869 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
870 already does with RSA. testdsa.h now has 'priv_key/pub_key'
871 values for each of the key sizes rather than having just
872 parameters (and 'speed' generating keys each time).
875 *) Speed up EVP routines.
878 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
879 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
880 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
881 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
883 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
884 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
885 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
888 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
890 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
893 *) Added the OS2-EMX target.
894 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
896 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
897 to support NCONF routines in extension code. New function CONF_set_nconf()
898 to allow functions which take an NCONF to also handle the old LHASH
899 structure: this means that the old CONF compatible routines can be
900 retained (in particular wrt extensions) without having to duplicate the
901 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
904 *) Enhance the general user interface with mechanisms for inner control
905 and with possibilities to have yes/no kind of prompts.
908 *) Change all calls to low level digest routines in the library and
909 applications to use EVP. Add missing calls to HMAC_cleanup() and
910 don't assume HMAC_CTX can be copied using memcpy().
911 [Verdon Walker <VWalker@novell.com>, Steve Henson]
913 *) Add the possibility to control engines through control names but with
914 arbitrary arguments instead of just a string.
915 Change the key loaders to take a UI_METHOD instead of a callback
916 function pointer. NOTE: this breaks binary compatibility with earlier
917 versions of OpenSSL [engine].
918 Adapt the nCipher code for these new conditions and add a card insertion
922 *) Enhance the general user interface with mechanisms to better support
923 dialog box interfaces, application-defined prompts, the possibility
924 to use defaults (for example default passwords from somewhere else)
925 and interrupts/cancellations.
928 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
929 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
932 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
933 tidy up some unnecessarily weird code in 'sk_new()').
934 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
936 *) Change the key loading routines for ENGINEs to use the same kind
937 callback (pem_password_cb) as all other routines that need this
941 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
942 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
943 than this minimum value is recommended.
946 *) New random seeder for OpenVMS, using the system process statistics
947 that are easily reachable.
950 *) Windows apparently can't transparently handle global
951 variables defined in DLLs. Initialisations such as:
953 const ASN1_ITEM *it = &ASN1_INTEGER_it;
955 wont compile. This is used by the any applications that need to
956 declare their own ASN1 modules. This was fixed by adding the option
957 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
958 needed for static libraries under Win32.
961 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
962 setting of purpose and trust fields. New X509_STORE trust and
963 purpose functions and tidy up setting in other SSL functions.
966 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
967 structure. These are inherited by X509_STORE_CTX when it is
968 initialised. This allows various defaults to be set in the
969 X509_STORE structure (such as flags for CRL checking and custom
970 purpose or trust settings) for functions which only use X509_STORE_CTX
971 internally such as S/MIME.
973 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
974 trust settings if they are not set in X509_STORE. This allows X509_STORE
975 purposes and trust (in S/MIME for example) to override any set by default.
977 Add command line options for CRL checking to smime, s_client and s_server
981 *) Initial CRL based revocation checking. If the CRL checking flag(s)
982 are set then the CRL is looked up in the X509_STORE structure and
983 its validity and signature checked, then if the certificate is found
984 in the CRL the verify fails with a revoked error.
986 Various new CRL related callbacks added to X509_STORE_CTX structure.
988 Command line options added to 'verify' application to support this.
990 This needs some additional work, such as being able to handle multiple
991 CRLs with different times, extension based lookup (rather than just
992 by subject name) and ultimately more complete V2 CRL extension
996 *) Add a general user interface API (crypto/ui/). This is designed
997 to replace things like des_read_password and friends (backward
998 compatibility functions using this new API are provided).
999 The purpose is to remove prompting functions from the DES code
1000 section as well as provide for prompting through dialog boxes in
1001 a window system and the like.
1004 *) Add "ex_data" support to ENGINE so implementations can add state at a
1005 per-structure level rather than having to store it globally.
1008 *) Make it possible for ENGINE structures to be copied when retrieved by
1009 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
1010 This causes the "original" ENGINE structure to act like a template,
1011 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
1012 operational state can be localised to each ENGINE structure, despite the
1013 fact they all share the same "methods". New ENGINE structures returned in
1014 this case have no functional references and the return value is the single
1015 structural reference. This matches the single structural reference returned
1016 by ENGINE_by_id() normally, when it is incremented on the pre-existing
1020 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
1021 needs to match any other type at all we need to manually clear the
1025 *) Changes to the "openssl engine" utility to include;
1026 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
1027 about an ENGINE's available control commands.
1028 - executing control commands from command line arguments using the
1029 '-pre' and '-post' switches. '-post' is only used if '-t' is
1030 specified and the ENGINE is successfully initialised. The syntax for
1031 the individual commands are colon-separated, for example;
1032 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
1035 *) New dynamic control command support for ENGINEs. ENGINEs can now
1036 declare their own commands (numbers), names (strings), descriptions,
1037 and input types for run-time discovery by calling applications. A
1038 subset of these commands are implicitly classed as "executable"
1039 depending on their input type, and only these can be invoked through
1040 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
1041 can be based on user input, config files, etc). The distinction is
1042 that "executable" commands cannot return anything other than a boolean
1043 result and can only support numeric or string input, whereas some
1044 discoverable commands may only be for direct use through
1045 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
1046 pointers, or other custom uses. The "executable" commands are to
1047 support parameterisations of ENGINE behaviour that can be
1048 unambiguously defined by ENGINEs and used consistently across any
1049 OpenSSL-based application. Commands have been added to all the
1050 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
1051 control over shared-library paths without source code alterations.
1054 *) Changed all ENGINE implementations to dynamically allocate their
1055 ENGINEs rather than declaring them statically. Apart from this being
1056 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
1057 this also allows the implementations to compile without using the
1058 internal engine_int.h header.
1061 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
1062 'const' value. Any code that should be able to modify a RAND_METHOD
1063 should already have non-const pointers to it (ie. they should only
1064 modify their own ones).
1067 *) Made a variety of little tweaks to the ENGINE code.
1068 - "atalla" and "ubsec" string definitions were moved from header files
1069 to C code. "nuron" string definitions were placed in variables
1070 rather than hard-coded - allowing parameterisation of these values
1071 later on via ctrl() commands.
1072 - Removed unused "#if 0"'d code.
1073 - Fixed engine list iteration code so it uses ENGINE_free() to release
1074 structural references.
1075 - Constified the RAND_METHOD element of ENGINE structures.
1076 - Constified various get/set functions as appropriate and added
1077 missing functions (including a catch-all ENGINE_cpy that duplicates
1078 all ENGINE values onto a new ENGINE except reference counts/state).
1079 - Removed NULL parameter checks in get/set functions. Setting a method
1080 or function to NULL is a way of cancelling out a previously set
1081 value. Passing a NULL ENGINE parameter is just plain stupid anyway
1082 and doesn't justify the extra error symbols and code.
1083 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
1084 flags from engine_int.h to engine.h.
1085 - Changed prototypes for ENGINE handler functions (init(), finish(),
1086 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
1089 *) Implement binary inversion algorithm for BN_mod_inverse in addition
1090 to the algorithm using long division. The binary algorithm can be
1091 used only if the modulus is odd. On 32-bit systems, it is faster
1092 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
1093 roughly 5-15% for 256-bit moduli), so we use it only for moduli
1094 up to 450 bits. In 64-bit environments, the binary algorithm
1095 appears to be advantageous for much longer moduli; here we use it
1096 for moduli up to 2048 bits.
1099 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
1100 could not support the combine flag in choice fields.
1103 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
1104 extensions from a certificate request to the certificate.
1107 *) Allow multiple 'certopt' and 'nameopt' options to be separated
1108 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
1109 file: this allows the display of the certificate about to be
1110 signed to be customised, to allow certain fields to be included
1111 or excluded and extension details. The old system didn't display
1112 multicharacter strings properly, omitted fields not in the policy
1113 and couldn't display additional details such as extensions.
1116 *) Function EC_POINTs_mul for multiple scalar multiplication
1117 of an arbitrary number of elliptic curve points
1118 \sum scalars[i]*points[i],
1119 optionally including the generator defined for the EC_GROUP:
1120 scalar*generator + \sum scalars[i]*points[i].
1122 EC_POINT_mul is a simple wrapper function for the typical case
1123 that the point list has just one item (besides the optional
1127 *) First EC_METHODs for curves over GF(p):
1129 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
1130 operations and provides various method functions that can also
1131 operate with faster implementations of modular arithmetic.
1133 EC_GFp_mont_method() reuses most functions that are part of
1134 EC_GFp_simple_method, but uses Montgomery arithmetic.
1136 [Bodo Moeller; point addition and point doubling
1137 implementation directly derived from source code provided by
1138 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
1140 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
1141 crypto/ec/ec_lib.c):
1143 Curves are EC_GROUP objects (with an optional group generator)
1144 based on EC_METHODs that are built into the library.
1146 Points are EC_POINT objects based on EC_GROUP objects.
1148 Most of the framework would be able to handle curves over arbitrary
1149 finite fields, but as there are no obvious types for fields other
1150 than GF(p), some functions are limited to that for now.
1153 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
1154 that the file contains a complete HTTP response.
1157 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
1158 change the def and num file printf format specifier from "%-40sXXX"
1159 to "%-39s XXX". The latter will always guarantee a space after the
1160 field while the former will cause them to run together if the field
1161 is 40 of more characters long.
1164 *) Constify the cipher and digest 'method' functions and structures
1165 and modify related functions to take constant EVP_MD and EVP_CIPHER
1169 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
1170 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
1173 *) Modify EVP_Digest*() routines so they now return values. Although the
1174 internal software routines can never fail additional hardware versions
1178 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
1180 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
1181 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
1184 ERR_R_NESTED_ASN1_ERROR
1186 ERR_R_MISSING_ASN1_EOS
1187 were 4 .. 9, conflicting with
1188 ERR_LIB_RSA (= ERR_R_RSA_LIB)
1190 ERR_LIB_PEM (= ERR_R_PEM_LIB).
1191 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
1193 Add new error code 'ERR_R_INTERNAL_ERROR'.
1196 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
1200 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
1201 sets the subject name for a new request or supersedes the
1202 subject name in a given request. Formats that can be parsed are
1203 'CN=Some Name, OU=myOU, C=IT'
1205 'CN=Some Name/OU=myOU/C=IT'.
1207 Add options '-batch' and '-verbose' to 'openssl req'.
1208 [Massimiliano Pala <madwolf@hackmasters.net>]
1210 *) Introduce the possibility to access global variables through
1211 functions on platform were that's the best way to handle exporting
1212 global variables in shared libraries. To enable this functionality,
1213 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1214 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1215 is normally done by Configure or something similar).
1217 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1218 in the source file (foo.c) like this:
1220 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1221 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1223 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1224 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1226 OPENSSL_DECLARE_GLOBAL(int,foo);
1227 #define foo OPENSSL_GLOBAL_REF(foo)
1228 OPENSSL_DECLARE_GLOBAL(double,bar);
1229 #define bar OPENSSL_GLOBAL_REF(bar)
1231 The #defines are very important, and therefore so is including the
1232 header file everywhere where the defined globals are used.
1234 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1235 of ASN.1 items, but that structure is a bit different.
1237 The largest change is in util/mkdef.pl which has been enhanced with
1238 better and easier to understand logic to choose which symbols should
1239 go into the Windows .def files as well as a number of fixes and code
1240 cleanup (among others, algorithm keywords are now sorted
1241 lexicographically to avoid constant rewrites).
1244 *) In BN_div() keep a copy of the sign of 'num' before writing the
1245 result to 'rm' because if rm==num the value will be overwritten
1246 and produce the wrong result if 'num' is negative: this caused
1247 problems with BN_mod() and BN_nnmod().
1250 *) Function OCSP_request_verify(). This checks the signature on an
1251 OCSP request and verifies the signer certificate. The signer
1252 certificate is just checked for a generic purpose and OCSP request
1256 *) Add OCSP_check_validity() function to check the validity of OCSP
1257 responses. OCSP responses are prepared in real time and may only
1258 be a few seconds old. Simply checking that the current time lies
1259 between thisUpdate and nextUpdate max reject otherwise valid responses
1260 caused by either OCSP responder or client clock inaccuracy. Instead
1261 we allow thisUpdate and nextUpdate to fall within a certain period of
1262 the current time. The age of the response can also optionally be
1263 checked. Two new options -validity_period and -status_age added to
1267 *) If signature or public key algorithm is unrecognized print out its
1268 OID rather that just UNKNOWN.
1271 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1272 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1273 ID to be generated from the issuer certificate alone which can then be
1274 passed to OCSP_id_issuer_cmp().
1277 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1278 ASN1 modules to export functions returning ASN1_ITEM pointers
1279 instead of the ASN1_ITEM structures themselves. This adds several
1280 new macros which allow the underlying ASN1 function/structure to
1281 be accessed transparently. As a result code should not use ASN1_ITEM
1282 references directly (such as &X509_it) but instead use the relevant
1283 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1284 use of the new ASN1 code on platforms where exporting structures
1285 is problematical (for example in shared libraries) but exporting
1286 functions returning pointers to structures is not.
1289 *) Add support for overriding the generation of SSL/TLS session IDs.
1290 These callbacks can be registered either in an SSL_CTX or per SSL.
1291 The purpose of this is to allow applications to control, if they wish,
1292 the arbitrary values chosen for use as session IDs, particularly as it
1293 can be useful for session caching in multiple-server environments. A
1294 command-line switch for testing this (and any client code that wishes
1295 to use such a feature) has been added to "s_server".
1296 [Geoff Thorpe, Lutz Jaenicke]
1298 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
1299 of the form '#if defined(...) || defined(...) || ...' and
1300 '#if !defined(...) && !defined(...) && ...'. This also avoids
1301 the growing number of special cases it was previously handling.
1304 *) Make all configuration macros available for application by making
1305 sure they are available in opensslconf.h, by giving them names starting
1306 with "OPENSSL_" to avoid conflicts with other packages and by making
1307 sure e_os2.h will cover all platform-specific cases together with
1309 Additionally, it is now possible to define configuration/platform-
1310 specific names (called "system identities"). In the C code, these
1311 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1312 macro with the name beginning with "OPENSSL_SYS_", which is determined
1313 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1317 *) New option -set_serial to 'req' and 'x509' this allows the serial
1318 number to use to be specified on the command line. Previously self
1319 signed certificates were hard coded with serial number 0 and the
1320 CA options of 'x509' had to use a serial number in a file which was
1324 *) New options to 'ca' utility to support V2 CRL entry extensions.
1325 Currently CRL reason, invalidity date and hold instruction are
1326 supported. Add new CRL extensions to V3 code and some new objects.
1329 *) New function EVP_CIPHER_CTX_set_padding() this is used to
1330 disable standard block padding (aka PKCS#5 padding) in the EVP
1331 API, which was previously mandatory. This means that the data is
1332 not padded in any way and so the total length much be a multiple
1333 of the block size, otherwise an error occurs.
1336 *) Initial (incomplete) OCSP SSL support.
1339 *) New function OCSP_parse_url(). This splits up a URL into its host,
1340 port and path components: primarily to parse OCSP URLs. New -url
1341 option to ocsp utility.
1344 *) New nonce behavior. The return value of OCSP_check_nonce() now
1345 reflects the various checks performed. Applications can decide
1346 whether to tolerate certain situations such as an absent nonce
1347 in a response when one was present in a request: the ocsp application
1348 just prints out a warning. New function OCSP_add1_basic_nonce()
1349 this is to allow responders to include a nonce in a response even if
1350 the request is nonce-less.
1353 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1354 skipped when using openssl x509 multiple times on a single input file,
1355 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1358 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1359 set string type: to handle setting ASN1_TIME structures. Fix ca
1360 utility to correctly initialize revocation date of CRLs.
1363 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1364 the clients preferred ciphersuites and rather use its own preferences.
1365 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1366 Internet Explorer by ensuring unchanged hash method during stepup.
1367 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1370 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1371 to aes and add a new 'exist' option to print out symbols that don't
1375 *) Additional options to ocsp utility to allow flags to be set and
1376 additional certificates supplied.
1379 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
1380 OCSP client a number of certificate to only verify the response
1384 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1385 handle the new API. Currently only ECB, CBC modes supported. Add new
1388 Add TLS AES ciphersuites as described in RFC3268, "Advanced
1389 Encryption Standard (AES) Ciphersuites for Transport Layer
1390 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
1391 not enabled by default and were not part of the "ALL" ciphersuite
1392 alias because they were not yet official; they could be
1393 explicitly requested by specifying the "AESdraft" ciphersuite
1394 group alias. In the final release of OpenSSL 0.9.7, the group
1395 alias is called "AES" and is part of "ALL".)
1396 [Ben Laurie, Steve Henson, Bodo Moeller]
1398 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
1399 request to response.
1402 *) Functions for OCSP responders. OCSP_request_onereq_count(),
1403 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1404 extract information from a certificate request. OCSP_response_create()
1405 creates a response and optionally adds a basic response structure.
1406 OCSP_basic_add1_status() adds a complete single response to a basic
1407 response and returns the OCSP_SINGLERESP structure just added (to allow
1408 extensions to be included for example). OCSP_basic_add1_cert() adds a
1409 certificate to a basic response and OCSP_basic_sign() signs a basic
1410 response with various flags. New helper functions ASN1_TIME_check()
1411 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1412 (converts ASN1_TIME to GeneralizedTime).
1415 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1416 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1417 structure from a certificate. X509_pubkey_digest() digests the public_key
1418 contents: this is used in various key identifiers.
1421 *) Make sk_sort() tolerate a NULL argument.
1422 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1424 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1425 passed by the function are trusted implicitly. If any of them signed the
1426 response then it is assumed to be valid and is not verified.
1429 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1430 to data. This was previously part of the PKCS7 ASN1 code. This
1431 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1432 [Steve Henson, reported by Kenneth R. Robinette
1433 <support@securenetterm.com>]
1435 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1436 routines: without these tracing memory leaks is very painful.
1437 Fix leaks in PKCS12 and PKCS7 routines.
1440 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1441 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1442 effectively meant GeneralizedTime would never be used. Now it
1443 is initialised to -1 but X509_time_adj() now has to check the value
1444 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1445 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1446 [Steve Henson, reported by Kenneth R. Robinette
1447 <support@securenetterm.com>]
1449 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1450 result in a zero length in the ASN1_INTEGER structure which was
1451 not consistent with the structure when d2i_ASN1_INTEGER() was used
1452 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1453 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1454 where it did not print out a minus for negative ASN1_INTEGER.
1457 *) Add summary printout to ocsp utility. The various functions which
1458 convert status values to strings have been renamed to:
1459 OCSP_response_status_str(), OCSP_cert_status_str() and
1460 OCSP_crl_reason_str() and are no longer static. New options
1461 to verify nonce values and to disable verification. OCSP response
1462 printout format cleaned up.
1465 *) Add additional OCSP certificate checks. These are those specified
1466 in RFC2560. This consists of two separate checks: the CA of the
1467 certificate being checked must either be the OCSP signer certificate
1468 or the issuer of the OCSP signer certificate. In the latter case the
1469 OCSP signer certificate must contain the OCSP signing extended key
1470 usage. This check is performed by attempting to match the OCSP
1471 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1472 in the OCSP_CERTID structures of the response.
1475 *) Initial OCSP certificate verification added to OCSP_basic_verify()
1476 and related routines. This uses the standard OpenSSL certificate
1477 verify routines to perform initial checks (just CA validity) and
1478 to obtain the certificate chain. Then additional checks will be
1479 performed on the chain. Currently the root CA is checked to see
1480 if it is explicitly trusted for OCSP signing. This is used to set
1481 a root CA as a global signing root: that is any certificate that
1482 chains to that CA is an acceptable OCSP signing certificate.
1485 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1486 extensions from a separate configuration file.
1487 As when reading extensions from the main configuration file,
1488 the '-extensions ...' option may be used for specifying the
1490 [Massimiliano Pala <madwolf@comune.modena.it>]
1492 *) New OCSP utility. Allows OCSP requests to be generated or
1493 read. The request can be sent to a responder and the output
1494 parsed, outputed or printed in text form. Not complete yet:
1495 still needs to check the OCSP response validity.
1498 *) New subcommands for 'openssl ca':
1499 'openssl ca -status <serial>' prints the status of the cert with
1500 the given serial number (according to the index file).
1501 'openssl ca -updatedb' updates the expiry status of certificates
1503 [Massimiliano Pala <madwolf@comune.modena.it>]
1505 *) New '-newreq-nodes' command option to CA.pl. This is like
1506 '-newreq', but calls 'openssl req' with the '-nodes' option
1507 so that the resulting key is not encrypted.
1508 [Damien Miller <djm@mindrot.org>]
1510 *) New configuration for the GNU Hurd.
1511 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1513 *) Initial code to implement OCSP basic response verify. This
1514 is currently incomplete. Currently just finds the signer's
1515 certificate and verifies the signature on the response.
1518 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1519 value of OPENSSLDIR. This is available via the new '-d' option
1520 to 'openssl version', and is also included in 'openssl version -a'.
1523 *) Allowing defining memory allocation callbacks that will be given
1524 file name and line number information in additional arguments
1525 (a const char* and an int). The basic functionality remains, as
1526 well as the original possibility to just replace malloc(),
1527 realloc() and free() by functions that do not know about these
1528 additional arguments. To register and find out the current
1529 settings for extended allocation functions, the following
1530 functions are provided:
1532 CRYPTO_set_mem_ex_functions
1533 CRYPTO_set_locked_mem_ex_functions
1534 CRYPTO_get_mem_ex_functions
1535 CRYPTO_get_locked_mem_ex_functions
1537 These work the same way as CRYPTO_set_mem_functions and friends.
1538 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1539 extended allocation function is enabled.
1540 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1541 a conventional allocation function is enabled.
1542 [Richard Levitte, Bodo Moeller]
1544 *) Finish off removing the remaining LHASH function pointer casts.
1545 There should no longer be any prototype-casting required when using
1546 the LHASH abstraction, and any casts that remain are "bugs". See
1547 the callback types and macros at the head of lhash.h for details
1548 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1551 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1552 If /dev/[u]random devices are not available or do not return enough
1553 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1555 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1556 /etc/entropy will be queried once each in this sequence, quering stops
1557 when enough entropy was collected without querying more sockets.
1560 *) Change the Unix RAND_poll() variant to be able to poll several
1561 random devices, as specified by DEVRANDOM, until a sufficient amount
1562 of data has been collected. We spend at most 10 ms on each file
1563 (select timeout) and read in non-blocking mode. DEVRANDOM now
1564 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1565 (previously it was just the string "/dev/urandom"), so on typical
1566 platforms the 10 ms delay will never occur.
1567 Also separate out the Unix variant to its own file, rand_unix.c.
1568 For VMS, there's a currently-empty rand_vms.c.
1571 *) Move OCSP client related routines to ocsp_cl.c. These
1572 provide utility functions which an application needing
1573 to issue a request to an OCSP responder and analyse the
1574 response will typically need: as opposed to those which an
1575 OCSP responder itself would need which will be added later.
1577 OCSP_request_sign() signs an OCSP request with an API similar
1578 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1579 response. OCSP_response_get1_basic() extracts basic response
1580 from response. OCSP_resp_find_status(): finds and extracts status
1581 information from an OCSP_CERTID structure (which will be created
1582 when the request structure is built). These are built from lower
1583 level functions which work on OCSP_SINGLERESP structures but
1584 wont normally be used unless the application wishes to examine
1585 extensions in the OCSP response for example.
1587 Replace nonce routines with a pair of functions.
1588 OCSP_request_add1_nonce() adds a nonce value and optionally
1589 generates a random value. OCSP_check_nonce() checks the
1590 validity of the nonce in an OCSP response.
1593 *) Change function OCSP_request_add() to OCSP_request_add0_id().
1594 This doesn't copy the supplied OCSP_CERTID and avoids the
1595 need to free up the newly created id. Change return type
1596 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1597 This can then be used to add extensions to the request.
1598 Deleted OCSP_request_new(), since most of its functionality
1599 is now in OCSP_REQUEST_new() (and the case insensitive name
1600 clash) apart from the ability to set the request name which
1601 will be added elsewhere.
1604 *) Update OCSP API. Remove obsolete extensions argument from
1605 various functions. Extensions are now handled using the new
1606 OCSP extension code. New simple OCSP HTTP function which
1607 can be used to send requests and parse the response.
1610 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1611 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1612 uses the special reorder version of SET OF to sort the attributes
1613 and reorder them to match the encoded order. This resolves a long
1614 standing problem: a verify on a PKCS7 structure just after signing
1615 it used to fail because the attribute order did not match the
1616 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1617 it uses the received order. This is necessary to tolerate some broken
1618 software that does not order SET OF. This is handled by encoding
1619 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1620 to produce the required SET OF.
1623 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1624 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1625 files to get correct declarations of the ASN.1 item variables.
1628 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1629 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1630 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1631 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1632 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1633 ASN1_ITEM and no wrapper functions.
1636 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1637 replace the old function pointer based I/O routines. Change most of
1638 the *_d2i_bio() and *_d2i_fp() functions to use these.
1641 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1642 lines, recognice more "algorithms" that can be deselected, and make
1643 it complain about algorithm deselection that isn't recognised.
1646 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
1647 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1648 to use new functions. Add NO_ASN1_OLD which can be set to remove
1649 some old style ASN1 functions: this can be used to determine if old
1650 code will still work when these eventually go away.
1653 *) New extension functions for OCSP structures, these follow the
1654 same conventions as certificates and CRLs.
1657 *) New function X509V3_add1_i2d(). This automatically encodes and
1658 adds an extension. Its behaviour can be customised with various
1659 flags to append, replace or delete. Various wrappers added for
1660 certifcates and CRLs.
1663 *) Fix to avoid calling the underlying ASN1 print routine when
1664 an extension cannot be parsed. Correct a typo in the
1665 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1668 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1669 entries for variables.
1672 *) Add functionality to apps/openssl.c for detecting locking
1673 problems: As the program is single-threaded, all we have
1674 to do is register a locking callback using an array for
1675 storing which locks are currently held by the program.
1678 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
1679 SSL_get_ex_data_X509_STORE_idx(), which is used in
1680 ssl_verify_cert_chain() and thus can be called at any time
1681 during TLS/SSL handshakes so that thread-safety is essential.
1682 Unfortunately, the ex_data design is not at all suited
1683 for multi-threaded use, so it probably should be abolished.
1686 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
1687 [Broadcom, tweaked and integrated by Geoff Thorpe]
1689 *) Move common extension printing code to new function
1690 X509V3_print_extensions(). Reorganise OCSP print routines and
1691 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1694 *) New function X509_signature_print() to remove duplication in some
1698 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
1699 set (this was treated exactly the same as SET OF previously). This
1700 is used to reorder the STACK representing the structure to match the
1701 encoding. This will be used to get round a problem where a PKCS7
1702 structure which was signed could not be verified because the STACK
1703 order did not reflect the encoded order.
1706 *) Reimplement the OCSP ASN1 module using the new code.
1709 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1710 for its ASN1 operations. The old style function pointers still exist
1711 for now but they will eventually go away.
1714 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
1715 completely replaces the old ASN1 functionality with a table driven
1716 encoder and decoder which interprets an ASN1_ITEM structure describing
1717 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1718 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1719 has also been converted to the new form.
1722 *) Change BN_mod_exp_recp so that negative moduli are tolerated
1723 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1724 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1725 for negative moduli.
1728 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
1729 of not touching the result's sign bit.
1732 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1736 *) Changed the LHASH code to use prototypes for callbacks, and created
1737 macros to declare and implement thin (optionally static) functions
1738 that provide type-safety and avoid function pointer casting for the
1739 type-specific callbacks.
1742 *) Added Kerberos Cipher Suites to be used with TLS, as written in
1744 [Veers Staats <staatsvr@asc.hpc.mil>,
1745 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1747 *) Reformat the FAQ so the different questions and answers can be divided
1748 in sections depending on the subject.
1751 *) Have the zlib compression code load ZLIB.DLL dynamically under
1755 *) New function BN_mod_sqrt for computing square roots modulo a prime
1756 (using the probabilistic Tonelli-Shanks algorithm unless
1757 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1758 be handled deterministically).
1759 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1761 *) Make BN_mod_inverse faster by explicitly handling small quotients
1762 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1763 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1766 *) New function BN_kronecker.
1769 *) Fix BN_gcd so that it works on negative inputs; the result is
1770 positive unless both parameters are zero.
1771 Previously something reasonably close to an infinite loop was
1772 possible because numbers could be growing instead of shrinking
1773 in the implementation of Euclid's algorithm.
1776 *) Fix BN_is_word() and BN_is_one() macros to take into account the
1777 sign of the number in question.
1779 Fix BN_is_word(a,w) to work correctly for w == 0.
1781 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1782 because its test if the absolute value of 'a' equals 'w'.
1783 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1784 it exists mostly for use in the implementations of BN_is_zero(),
1785 BN_is_one(), and BN_is_word().
1788 *) New function BN_swap.
1791 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1792 the exponentiation functions are more likely to produce reasonable
1793 results on negative inputs.
1796 *) Change BN_mod_mul so that the result is always non-negative.
1797 Previously, it could be negative if one of the factors was negative;
1798 I don't think anyone really wanted that behaviour.
1801 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1802 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1803 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1804 and add new functions:
1813 BN_mod_lshift1_quick
1817 These functions always generate non-negative results.
1819 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
1820 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
1822 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
1823 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
1824 be reduced modulo m.
1825 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1827 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1828 was actually never needed) and in BN_mul(). The removal in BN_mul()
1829 required a small change in bn_mul_part_recursive() and the addition
1830 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1831 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1832 bn_sub_words() and bn_add_words() except they take arrays with
1836 *) In 'openssl passwd', verify passwords read from the terminal
1837 unless the '-salt' option is used (which usually means that
1838 verification would just waste user's time since the resulting
1839 hash is going to be compared with some given password hash)
1840 or the new '-noverify' option is used.
1842 This is an incompatible change, but it does not affect
1843 non-interactive use of 'openssl passwd' (passwords on the command
1844 line, '-stdin' option, '-in ...' option) and thus should not
1848 *) Remove all references to RSAref, since there's no more need for it.
1851 *) Make DSO load along a path given through an environment variable
1852 (SHLIB_PATH) with shl_load().
1855 *) Constify the ENGINE code as a result of BIGNUM constification.
1856 Also constify the RSA code and most things related to it. In a
1857 few places, most notable in the depth of the ASN.1 code, ugly
1858 casts back to non-const were required (to be solved at a later
1862 *) Make it so the openssl application has all engines loaded by default.
1865 *) Constify the BIGNUM routines a little more.
1868 *) Add the following functions:
1870 ENGINE_load_cswift()
1872 ENGINE_load_atalla()
1874 ENGINE_load_builtin_engines()
1876 That way, an application can itself choose if external engines that
1877 are built-in in OpenSSL shall ever be used or not. The benefit is
1878 that applications won't have to be linked with libdl or other dso
1879 libraries unless it's really needed.
1881 Changed 'openssl engine' to load all engines on demand.
1882 Changed the engine header files to avoid the duplication of some
1883 declarations (they differed!).
1886 *) 'openssl engine' can now list capabilities.
1889 *) Better error reporting in 'openssl engine'.
1892 *) Never call load_dh_param(NULL) in s_server.
1895 *) Add engine application. It can currently list engines by name and
1896 identity, and test if they are actually available.
1899 *) Improve RPM specification file by forcing symbolic linking and making
1900 sure the installed documentation is also owned by root.root.
1901 [Damien Miller <djm@mindrot.org>]
1903 *) Give the OpenSSL applications more possibilities to make use of
1904 keys (public as well as private) handled by engines.
1907 *) Add OCSP code that comes from CertCo.
1910 *) Add VMS support for the Rijndael code.
1913 *) Added untested support for Nuron crypto accelerator.
1916 *) Add support for external cryptographic devices. This code was
1917 previously distributed separately as the "engine" branch.
1918 [Geoff Thorpe, Richard Levitte]
1920 *) Rework the filename-translation in the DSO code. It is now possible to
1921 have far greater control over how a "name" is turned into a filename
1922 depending on the operating environment and any oddities about the
1923 different shared library filenames on each system.
1926 *) Support threads on FreeBSD-elf in Configure.
1929 *) Fix for SHA1 assembly problem with MASM: it produces
1930 warnings about corrupt line number information when assembling
1931 with debugging information. This is caused by the overlapping
1933 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
1936 NCONF_get_number() has no error checking at all. As a replacement,
1937 NCONF_get_number_e() is defined (_e for "error checking") and is
1938 promoted strongly. The old NCONF_get_number is kept around for
1939 binary backward compatibility.
1940 Make it possible for methods to load from something other than a BIO,
1941 by providing a function pointer that is given a name instead of a BIO.
1942 For example, this could be used to load configuration data from an
1946 *) Fix for non blocking accept BIOs. Added new I/O special reason
1947 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
1948 with non blocking I/O was not possible because no retry code was
1949 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
1953 *) Added the beginnings of Rijndael support.
1956 *) Fix for bug in DirectoryString mask setting. Add support for
1957 X509_NAME_print_ex() in 'req' and X509_print_ex() function
1958 to allow certificate printing to more controllable, additional
1959 'certopt' option to 'x509' to allow new printing options to be
1963 *) Clean old EAY MD5 hack from e_os.h.
1966 Changes between 0.9.6g and 0.9.6h [xx XXX xxxx]
1968 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
1969 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
1972 *) Fix initialization code race conditions in
1973 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
1974 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
1975 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
1976 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
1977 ssl2_get_cipher_by_char(),
1978 ssl3_get_cipher_by_char().
1979 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
1981 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
1982 the cached sessions are flushed, as the remove_cb() might use ex_data
1983 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
1984 (see [openssl.org #212]).
1985 [Geoff Thorpe, Lutz Jaenicke]
1987 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
1988 length, instead of the encoding length to d2i_ASN1_OBJECT.
1991 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
1993 *) [In 0.9.6g-engine release:]
1994 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
1995 [Lynn Gazis <lgazis@rainbow.com>]
1997 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
1999 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
2000 and get fix the header length calculation.
2001 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
2002 Alon Kantor <alonk@checkpoint.com> (and others),
2005 *) Use proper error handling instead of 'assertions' in buffer
2006 overflow checks added in 0.9.6e. This prevents DoS (the
2007 assertions could call abort()).
2008 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
2010 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
2012 *) Add various sanity checks to asn1_get_length() to reject
2013 the ASN1 length bytes if they exceed sizeof(long), will appear
2014 negative or the content length exceeds the length of the
2016 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2018 *) Fix cipher selection routines: ciphers without encryption had no flags
2019 for the cipher strength set and where therefore not handled correctly
2020 by the selection routines (PR #130).
2023 *) Fix EVP_dsa_sha macro.
2027 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
2028 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
2029 that was added in OpenSSL 0.9.6d.
2031 As the countermeasure turned out to be incompatible with some
2032 broken SSL implementations, the new option is part of SSL_OP_ALL.
2033 SSL_OP_ALL is usually employed when compatibility with weird SSL
2034 implementations is desired (e.g. '-bugs' option to 's_client' and
2035 's_server'), so the new option is automatically set in many
2039 *) Changes in security patch:
2041 Changes marked "(CHATS)" were sponsored by the Defense Advanced
2042 Research Projects Agency (DARPA) and Air Force Research Laboratory,
2043 Air Force Materiel Command, USAF, under agreement number
2046 *) Add various sanity checks to asn1_get_length() to reject
2047 the ASN1 length bytes if they exceed sizeof(long), will appear
2048 negative or the content length exceeds the length of the
2049 supplied buffer. (CAN-2002-0659)
2050 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2052 *) Assertions for various potential buffer overflows, not known to
2054 [Ben Laurie (CHATS)]
2056 *) Various temporary buffers to hold ASCII versions of integers were
2057 too small for 64 bit platforms. (CAN-2002-0655)
2058 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
2060 *) Remote buffer overflow in SSL3 protocol - an attacker could
2061 supply an oversized session ID to a client. (CAN-2002-0656)
2062 [Ben Laurie (CHATS)]
2064 *) Remote buffer overflow in SSL2 protocol - an attacker could
2065 supply an oversized client master key. (CAN-2002-0656)
2066 [Ben Laurie (CHATS)]
2068 Changes between 0.9.6c and 0.9.6d [9 May 2002]
2070 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
2071 encoded as NULL) with id-dsa-with-sha1.
2072 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
2074 *) Check various X509_...() return values in apps/req.c.
2075 [Nils Larsch <nla@trustcenter.de>]
2077 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
2078 an end-of-file condition would erronously be flagged, when the CRLF
2079 was just at the end of a processed block. The bug was discovered when
2080 processing data through a buffering memory BIO handing the data to a
2081 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
2082 <ptsekov@syntrex.com> and Nedelcho Stanev.
2085 *) Implement a countermeasure against a vulnerability recently found
2086 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
2087 before application data chunks to avoid the use of known IVs
2088 with data potentially chosen by the attacker.
2091 *) Fix length checks in ssl3_get_client_hello().
2094 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
2095 to prevent ssl3_read_internal() from incorrectly assuming that
2096 ssl3_read_bytes() found application data while handshake
2097 processing was enabled when in fact s->s3->in_read_app_data was
2098 merely automatically cleared during the initial handshake.
2099 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
2101 *) Fix object definitions for Private and Enterprise: they were not
2102 recognized in their shortname (=lowercase) representation. Extend
2103 obj_dat.pl to issue an error when using undefined keywords instead
2104 of silently ignoring the problem (Svenning Sorensen
2105 <sss@sss.dnsalias.net>).
2108 *) Fix DH_generate_parameters() so that it works for 'non-standard'
2109 generators, i.e. generators other than 2 and 5. (Previously, the
2110 code did not properly initialise the 'add' and 'rem' values to
2111 BN_generate_prime().)
2113 In the new general case, we do not insist that 'generator' is
2114 actually a primitive root: This requirement is rather pointless;
2115 a generator of the order-q subgroup is just as good, if not
2119 *) Map new X509 verification errors to alerts. Discovered and submitted by
2120 Tom Wu <tom@arcot.com>.
2123 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
2124 returning non-zero before the data has been completely received
2125 when using non-blocking I/O.
2126 [Bodo Moeller; problem pointed out by John Hughes]
2128 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
2129 [Ben Laurie, Lutz Jaenicke]
2131 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
2132 Yoram Zahavi <YoramZ@gilian.com>).
2135 *) Add information about CygWin 1.3 and on, and preserve proper
2136 configuration for the versions before that.
2137 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2139 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
2140 check whether we deal with a copy of a session and do not delete from
2141 the cache in this case. Problem reported by "Izhar Shoshani Levi"
2142 <izhar@checkpoint.com>.
2145 *) Do not store session data into the internal session cache, if it
2146 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
2147 flag is set). Proposed by Aslam <aslam@funk.com>.
2150 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
2154 *) [In 0.9.6d-engine release:]
2155 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2156 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2158 *) Add the configuration target linux-s390x.
2159 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
2161 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
2162 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
2163 variable as an indication that a ClientHello message has been
2164 received. As the flag value will be lost between multiple
2165 invocations of ssl3_accept when using non-blocking I/O, the
2166 function may not be aware that a handshake has actually taken
2167 place, thus preventing a new session from being added to the
2170 To avoid this problem, we now set s->new_session to 2 instead of
2171 using a local variable.
2172 [Lutz Jaenicke, Bodo Moeller]
2174 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
2175 if the SSL_R_LENGTH_MISMATCH error is detected.
2176 [Geoff Thorpe, Bodo Moeller]
2178 *) New 'shared_ldflag' column in Configure platform table.
2181 *) Fix EVP_CIPHER_mode macro.
2182 ["Dan S. Camper" <dan@bti.net>]
2184 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
2185 type, we must throw them away by setting rr->length to 0.
2186 [D P Chang <dpc@qualys.com>]
2188 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
2190 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
2191 <Dominikus.Scherkl@biodata.com>. (The previous implementation
2192 worked incorrectly for those cases where range = 10..._2 and
2193 3*range is two bits longer than range.)
2196 *) Only add signing time to PKCS7 structures if it is not already
2200 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
2201 OBJ_ld_ce should be OBJ_id_ce.
2202 Also some ip-pda OIDs in crypto/objects/objects.txt were
2203 incorrect (cf. RFC 3039).
2204 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
2206 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
2207 returns early because it has nothing to do.
2208 [Andy Schneider <andy.schneider@bjss.co.uk>]
2210 *) [In 0.9.6c-engine release:]
2211 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
2212 [Andy Schneider <andy.schneider@bjss.co.uk>]
2214 *) [In 0.9.6c-engine release:]
2215 Add support for Cryptographic Appliance's keyserver technology.
2216 (Use engine 'keyclient')
2217 [Cryptographic Appliances and Geoff Thorpe]
2219 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
2220 is called via tools/c89.sh because arguments have to be
2221 rearranged (all '-L' options must appear before the first object
2223 [Richard Shapiro <rshapiro@abinitio.com>]
2225 *) [In 0.9.6c-engine release:]
2226 Add support for Broadcom crypto accelerator cards, backported
2228 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
2230 *) [In 0.9.6c-engine release:]
2231 Add support for SureWare crypto accelerator cards from
2232 Baltimore Technologies. (Use engine 'sureware')
2233 [Baltimore Technologies and Mark Cox]
2235 *) [In 0.9.6c-engine release:]
2236 Add support for crypto accelerator cards from Accelerated
2237 Encryption Processing, www.aep.ie. (Use engine 'aep')
2238 [AEP Inc. and Mark Cox]
2240 *) Add a configuration entry for gcc on UnixWare.
2241 [Gary Benson <gbenson@redhat.com>]
2243 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
2244 messages are stored in a single piece (fixed-length part and
2245 variable-length part combined) and fix various bugs found on the way.
2248 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
2249 instead. BIO_gethostbyname() does not know what timeouts are
2250 appropriate, so entries would stay in cache even when they have
2252 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
2254 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
2255 faced with a pathologically small ClientHello fragment that does
2256 not contain client_version: Instead of aborting with an error,
2257 simply choose the highest available protocol version (i.e.,
2258 TLS 1.0 unless it is disabled). In practice, ClientHello
2259 messages are never sent like this, but this change gives us
2260 strictly correct behaviour at least for TLS.
2263 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
2264 never resets s->method to s->ctx->method when called from within
2265 one of the SSL handshake functions.
2266 [Bodo Moeller; problem pointed out by Niko Baric]
2268 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
2269 (sent using the client's version number) if client_version is
2270 smaller than the protocol version in use. Also change
2271 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
2272 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
2273 the client will at least see that alert.
2276 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
2280 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
2281 client receives HelloRequest while in a handshake.
2282 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
2284 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
2285 should end in 'break', not 'goto end' which circuments various
2286 cleanups done in state SSL_ST_OK. But session related stuff
2287 must be disabled for SSL_ST_OK in the case that we just sent a
2290 Also avoid some overhead by not calling ssl_init_wbio_buffer()
2291 before just sending a HelloRequest.
2292 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
2294 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
2295 reveal whether illegal block cipher padding was found or a MAC
2296 verification error occured. (Neither SSLerr() codes nor alerts
2297 are directly visible to potential attackers, but the information
2298 may leak via logfiles.)
2300 Similar changes are not required for the SSL 2.0 implementation
2301 because the number of padding bytes is sent in clear for SSL 2.0,
2302 and the extra bytes are just ignored. However ssl/s2_pkt.c
2303 failed to verify that the purported number of padding bytes is in
2307 *) Add OpenUNIX-8 support including shared libraries
2308 (Boyd Lynn Gerber <gerberb@zenez.com>).
2311 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
2312 'wristwatch attack' using huge encoding parameters (cf.
2313 James H. Manger's CRYPTO 2001 paper). Note that the
2314 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
2315 encoding parameters and hence was not vulnerable.
2318 *) BN_sqr() bug fix.
2319 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
2321 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
2322 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
2323 followed by modular reduction.
2324 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
2326 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
2327 equivalent based on BN_pseudo_rand() instead of BN_rand().
2330 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
2331 This function was broken, as the check for a new client hello message
2332 to handle SGC did not allow these large messages.
2333 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
2336 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
2339 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
2340 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
2343 *) Rework the configuration and shared library support for Tru64 Unix.
2344 The configuration part makes use of modern compiler features and
2345 still retains old compiler behavior for those that run older versions
2346 of the OS. The shared library support part includes a variant that
2347 uses the RPATH feature, and is available through the special
2348 configuration target "alpha-cc-rpath", which will never be selected
2350 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
2352 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
2353 with the same message size as in ssl3_get_certificate_request().
2354 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
2355 messages might inadvertently be reject as too long.
2356 [Petr Lampa <lampa@fee.vutbr.cz>]
2358 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
2361 *) Modified SSL library such that the verify_callback that has been set
2362 specificly for an SSL object with SSL_set_verify() is actually being
2363 used. Before the change, a verify_callback set with this function was
2364 ignored and the verify_callback() set in the SSL_CTX at the time of
2365 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
2366 to allow the necessary settings.
2369 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
2370 explicitly to NULL, as at least on Solaris 8 this seems not always to be
2371 done automatically (in contradiction to the requirements of the C
2372 standard). This made problems when used from OpenSSH.
2375 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
2376 dh->length and always used
2378 BN_rand_range(priv_key, dh->p).
2380 BN_rand_range() is not necessary for Diffie-Hellman, and this
2381 specific range makes Diffie-Hellman unnecessarily inefficient if
2382 dh->length (recommended exponent length) is much smaller than the
2383 length of dh->p. We could use BN_rand_range() if the order of
2384 the subgroup was stored in the DH structure, but we only have
2389 BN_rand(priv_key, l, ...)
2391 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
2397 RSA_eay_public_encrypt
2398 RSA_eay_private_decrypt
2399 RSA_eay_private_encrypt (signing)
2400 RSA_eay_public_decrypt (signature verification)
2402 (default implementations for RSA_public_encrypt,
2403 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
2404 always reject numbers >= n.
2407 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
2408 to synchronize access to 'locking_thread'. This is necessary on
2409 systems where access to 'locking_thread' (an 'unsigned long'
2410 variable) is not atomic.
2413 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
2414 *before* setting the 'crypto_lock_rand' flag. The previous code had
2415 a race condition if 0 is a valid thread ID.
2416 [Travis Vitek <vitek@roguewave.com>]
2418 *) Add support for shared libraries under Irix.
2419 [Albert Chin-A-Young <china@thewrittenword.com>]
2421 *) Add configuration option to build on Linux on both big-endian and
2423 [Ralf Baechle <ralf@uni-koblenz.de>]
2425 *) Add the possibility to create shared libraries on HP-UX.
2428 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
2430 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
2431 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
2432 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
2433 PRNG state recovery was possible based on the output of
2434 one PRNG request appropriately sized to gain knowledge on
2435 'md' followed by enough consecutive 1-byte PRNG requests
2436 to traverse all of 'state'.
2438 1. When updating 'md_local' (the current thread's copy of 'md')
2439 during PRNG output generation, hash all of the previous
2440 'md_local' value, not just the half used for PRNG output.
2442 2. Make the number of bytes from 'state' included into the hash
2443 independent from the number of PRNG bytes requested.
2445 The first measure alone would be sufficient to avoid
2446 Markku-Juhani's attack. (Actually it had never occurred
2447 to me that the half of 'md_local' used for chaining was the
2448 half from which PRNG output bytes were taken -- I had always
2449 assumed that the secret half would be used.) The second
2450 measure makes sure that additional data from 'state' is never
2451 mixed into 'md_local' in small portions; this heuristically
2452 further strengthens the PRNG.
2455 *) Fix crypto/bn/asm/mips3.s.
2458 *) When only the key is given to "enc", the IV is undefined. Print out
2459 an error message in this case.
2462 *) Handle special case when X509_NAME is empty in X509 printing routines.
2465 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
2466 positive and less than q.
2469 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
2470 used: it isn't thread safe and the add_lock_callback should handle
2472 [Paul Rose <Paul.Rose@bridge.com>]
2474 *) Verify that incoming data obeys the block size in
2475 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
2479 [Ulf Möller, Bodo Möller]
2481 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
2482 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
2483 when fixing the server behaviour for backwards-compatible 'client
2484 hello' messages. (Note that the attack is impractical against
2485 SSL 3.0 and TLS 1.0 anyway because length and version checking
2486 means that the probability of guessing a valid ciphertext is
2487 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
2490 Before 0.9.5, the countermeasure (hide the error by generating a
2491 random 'decryption result') did not work properly because
2492 ERR_clear_error() was missing, meaning that SSL_get_error() would
2493 detect the supposedly ignored error.
2495 Both problems are now fixed.
2498 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
2499 (previously it was 1024).
2502 *) Fix for compatibility mode trust settings: ignore trust settings
2503 unless some valid trust or reject settings are present.
2506 *) Fix for blowfish EVP: its a variable length cipher.
2509 *) Fix various bugs related to DSA S/MIME verification. Handle missing
2510 parameters in DSA public key structures and return an error in the
2511 DSA routines if parameters are absent.
2514 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
2515 in the current directory if neither $RANDFILE nor $HOME was set.
2516 RAND_file_name() in 0.9.6a returned NULL in this case. This has
2517 caused some confusion to Windows users who haven't defined $HOME.
2518 Thus RAND_file_name() is changed again: e_os.h can define a
2519 DEFAULT_HOME, which will be used if $HOME is not set.
2520 For Windows, we use "C:"; on other platforms, we still require
2521 environment variables.
2523 *) Move 'if (!initialized) RAND_poll()' into regions protected by
2524 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
2525 having multiple threads call RAND_poll() concurrently.
2528 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
2529 combination of a flag and a thread ID variable.
2530 Otherwise while one thread is in ssleay_rand_bytes (which sets the
2531 flag), *other* threads can enter ssleay_add_bytes without obeying
2532 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
2533 that they do not hold after the first thread unsets add_do_not_lock).
2536 *) Change bctest again: '-x' expressions are not available in all
2540 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
2542 *) Fix a couple of memory leaks in PKCS7_dataDecode()
2543 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
2545 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
2546 the default extension for executables, if any. Also, make the perl
2547 scripts that use symlink() to test if it really exists and use "cp"
2548 if it doesn't. All this made OpenSSL compilable and installable in
2552 *) Fix for asn1_GetSequence() for indefinite length constructed data.
2553 If SEQUENCE is length is indefinite just set c->slen to the total
2554 amount of data available.
2555 [Steve Henson, reported by shige@FreeBSD.org]
2556 [This change does not apply to 0.9.7.]
2558 *) Change bctest to avoid here-documents inside command substitution
2559 (workaround for FreeBSD /bin/sh bug).
2560 For compatibility with Ultrix, avoid shell functions (introduced
2561 in the bctest version that searches along $PATH).
2564 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
2565 with des_encrypt() defined on some operating systems, like Solaris
2569 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
2570 On the Importance of Eliminating Errors in Cryptographic
2571 Computations, J. Cryptology 14 (2001) 2, 101-119,
2572 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
2575 *) MIPS assembler BIGNUM division bug fix.
2578 *) Disabled incorrect Alpha assembler code.
2581 *) Fix PKCS#7 decode routines so they correctly update the length
2582 after reading an EOC for the EXPLICIT tag.
2584 [This change does not apply to 0.9.7.]
2586 *) Fix bug in PKCS#12 key generation routines. This was triggered
2587 if a 3DES key was generated with a 0 initial byte. Include
2588 PKCS12_BROKEN_KEYGEN compilation option to retain the old
2589 (but broken) behaviour.
2592 *) Enhance bctest to search for a working bc along $PATH and print
2594 [Tim Rice <tim@multitalents.net> via Richard Levitte]
2596 *) Fix memory leaks in err.c: free err_data string if necessary;
2597 don't write to the wrong index in ERR_set_error_data.
2600 *) Implement ssl23_peek (analogous to ssl23_read), which previously
2604 *) Replace rdtsc with _emit statements for VC++ version 5.
2605 [Jeremy Cooper <jeremy@baymoo.org>]
2607 *) Make it possible to reuse SSLv2 sessions.
2610 *) In copy_email() check for >= 0 as a return value for
2611 X509_NAME_get_index_by_NID() since 0 is a valid index.
2612 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
2614 *) Avoid coredump with unsupported or invalid public keys by checking if
2615 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
2616 PKCS7_verify() fails with non detached data.
2619 *) Don't use getenv in library functions when run as setuid/setgid.
2620 New function OPENSSL_issetugid().
2623 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
2624 due to incorrect handling of multi-threading:
2626 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
2628 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
2630 3. Count how many times MemCheck_off() has been called so that
2631 nested use can be treated correctly. This also avoids
2632 inband-signalling in the previous code (which relied on the
2633 assumption that thread ID 0 is impossible).
2636 *) Add "-rand" option also to s_client and s_server.
2639 *) Fix CPU detection on Irix 6.x.
2640 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
2641 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
2643 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
2646 [This change does not apply to 0.9.7.]
2648 *) Use the cached encoding of an X509_NAME structure rather than
2649 copying it. This is apparently the reason for the libsafe "errors"
2650 but the code is actually correct.
2653 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
2654 Bleichenbacher's DSA attack.
2655 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
2656 to be set and top=0 forces the highest bit to be set; top=-1 is new
2657 and leaves the highest bit random.
2658 [Ulf Moeller, Bodo Moeller]
2660 *) In the NCONF_...-based implementations for CONF_... queries
2661 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
2662 a temporary CONF structure with the data component set to NULL
2663 (which gives segmentation faults in lh_retrieve).
2664 Instead, use NULL for the CONF pointer in CONF_get_string and
2665 CONF_get_number (which may use environment variables) and directly
2666 return NULL from CONF_get_section.
2669 *) Fix potential buffer overrun for EBCDIC.
2672 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
2673 keyUsage if basicConstraints absent for a CA.
2676 *) Make SMIME_write_PKCS7() write mail header values with a format that
2677 is more generally accepted (no spaces before the semicolon), since
2678 some programs can't parse those values properly otherwise. Also make
2679 sure BIO's that break lines after each write do not create invalid
2683 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
2684 macros previously used would not encode an empty SEQUENCE OF
2685 and break the signature.
2687 [This change does not apply to 0.9.7.]
2689 *) Zero the premaster secret after deriving the master secret in
2693 *) Add some EVP_add_digest_alias registrations (as found in
2694 OpenSSL_add_all_digests()) to SSL_library_init()
2695 aka OpenSSL_add_ssl_algorithms(). This provides improved
2696 compatibility with peers using X.509 certificates
2697 with unconventional AlgorithmIdentifier OIDs.
2700 *) Fix for Irix with NO_ASM.
2701 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
2703 *) ./config script fixes.
2704 [Ulf Moeller, Richard Levitte]
2706 *) Fix 'openssl passwd -1'.
2709 *) Change PKCS12_key_gen_asc() so it can cope with non null
2710 terminated strings whose length is passed in the passlen
2711 parameter, for example from PEM callbacks. This was done
2712 by adding an extra length parameter to asc2uni().
2713 [Steve Henson, reported by <oddissey@samsung.co.kr>]
2715 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
2716 call failed, free the DSA structure.
2719 *) Fix to uni2asc() to cope with zero length Unicode strings.
2720 These are present in some PKCS#12 files.
2723 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
2724 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
2725 when writing a 32767 byte record.
2726 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
2728 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
2729 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
2731 (RSA objects have a reference count access to which is protected
2732 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
2733 so they are meant to be shared between threads.)
2734 [Bodo Moeller, Geoff Thorpe; original patch submitted by
2735 "Reddie, Steven" <Steven.Reddie@ca.com>]
2737 *) Fix a deadlock in CRYPTO_mem_leaks().
2740 *) Use better test patterns in bntest.
2743 *) rand_win.c fix for Borland C.
2746 *) BN_rshift bugfix for n == 0.
2749 *) Add a 'bctest' script that checks for some known 'bc' bugs
2750 so that 'make test' does not abort just because 'bc' is broken.
2753 *) Store verify_result within SSL_SESSION also for client side to
2754 avoid potential security hole. (Re-used sessions on the client side
2755 always resulted in verify_result==X509_V_OK, not using the original
2756 result of the server certificate verification.)
2759 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
2760 SSL3_RT_APPLICATION_DATA, return 0.
2761 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
2765 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
2766 releases, have been re-implemented by renaming the previous
2767 implementations of ssl2_read and ssl3_read to ssl2_read_internal
2768 and ssl3_read_internal, respectively, and adding 'peek' parameters
2769 to them. The new ssl[23]_{read,peek} functions are calls to
2770 ssl[23]_read_internal with the 'peek' flag set appropriately.
2771 A 'peek' parameter has also been added to ssl3_read_bytes, which
2772 does the actual work for ssl3_read_internal.
2775 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
2776 the method-specific "init()" handler. Also clean up ex_data after
2777 calling the method-specific "finish()" handler. Previously, this was
2778 happening the other way round.
2781 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
2782 The previous value, 12, was not always sufficient for BN_mod_exp().
2785 *) Make sure that shared libraries get the internal name engine with
2786 the full version number and not just 0. This should mark the
2787 shared libraries as not backward compatible. Of course, this should
2788 be changed again when we can guarantee backward binary compatibility.
2791 *) Fix typo in get_cert_by_subject() in by_dir.c
2792 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
2794 *) Rework the system to generate shared libraries:
2796 - Make note of the expected extension for the shared libraries and
2797 if there is a need for symbolic links from for example libcrypto.so.0
2798 to libcrypto.so.0.9.7. There is extended info in Configure for
2801 - Make as few rebuilds of the shared libraries as possible.
2803 - Still avoid linking the OpenSSL programs with the shared libraries.
2805 - When installing, install the shared libraries separately from the
2809 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
2811 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
2812 and not in SSL_clear because the latter is also used by the
2813 accept/connect functions; previously, the settings made by
2814 SSL_set_read_ahead would be lost during the handshake.
2815 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
2817 *) Correct util/mkdef.pl to be selective about disabled algorithms.
2818 Previously, it would create entries for disableed algorithms no
2822 *) Added several new manual pages for SSL_* function.
2825 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
2827 *) In ssl23_get_client_hello, generate an error message when faced
2828 with an initial SSL 3.0/TLS record that is too small to contain the
2829 first two bytes of the ClientHello message, i.e. client_version.
2830 (Note that this is a pathologic case that probably has never happened
2831 in real life.) The previous approach was to use the version number
2832 from the record header as a substitute; but our protocol choice
2833 should not depend on that one because it is not authenticated
2834 by the Finished messages.
2837 *) More robust randomness gathering functions for Windows.
2838 [Jeffrey Altman <jaltman@columbia.edu>]
2840 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
2841 not set then we don't setup the error code for issuer check errors
2842 to avoid possibly overwriting other errors which the callback does
2843 handle. If an application does set the flag then we assume it knows
2844 what it is doing and can handle the new informational codes
2848 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
2849 a general "ANY" type, as such it should be able to decode anything
2850 including tagged types. However it didn't check the class so it would
2851 wrongly interpret tagged types in the same way as their universal
2852 counterpart and unknown types were just rejected. Changed so that the
2853 tagged and unknown types are handled in the same way as a SEQUENCE:
2854 that is the encoding is stored intact. There is also a new type
2855 "V_ASN1_OTHER" which is used when the class is not universal, in this
2856 case we have no idea what the actual type is so we just lump them all
2860 *) On VMS, stdout may very well lead to a file that is written to
2861 in a record-oriented fashion. That means that every write() will
2862 write a separate record, which will be read separately by the
2863 programs trying to read from it. This can be very confusing.
2865 The solution is to put a BIO filter in the way that will buffer
2866 text until a linefeed is reached, and then write everything a
2867 line at a time, so every record written will be an actual line,
2868 not chunks of lines and not (usually doesn't happen, but I've
2869 seen it once) several lines in one record. BIO_f_linebuffer() is
2872 Currently, it's a VMS-only method, because that's where it has
2873 been tested well enough.
2876 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
2877 it can return incorrect results.
2878 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
2879 but it was in 0.9.6-beta[12].)
2882 *) Disable the check for content being present when verifying detached
2883 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
2884 include zero length content when signing messages.
2887 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
2888 BIO_ctrl (for BIO pairs).
2891 *) Add DSO method for VMS.
2894 *) Bug fix: Montgomery multiplication could produce results with the
2898 *) Add RPM specification openssl.spec and modify it to build three
2899 packages. The default package contains applications, application
2900 documentation and run-time libraries. The devel package contains
2901 include files, static libraries and function documentation. The
2902 doc package contains the contents of the doc directory. The original
2903 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
2906 *) Add a large number of documentation files for many SSL routines.
2907 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2909 *) Add a configuration entry for Sony News 4.
2910 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
2912 *) Don't set the two most significant bits to one when generating a
2913 random number < q in the DSA library.
2916 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
2917 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
2918 the underlying transport is blocking) if a handshake took place.
2919 (The default behaviour is needed by applications such as s_client
2920 and s_server that use select() to determine when to use SSL_read;
2921 but for applications that know in advance when to expect data, it
2922 just makes things more complicated.)
2925 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
2929 *) Add a few more EBCDIC conditionals that make `req' and `x509'
2930 work better on such systems.
2931 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2933 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
2934 Update PKCS12_parse() so it copies the friendlyName and the
2935 keyid to the certificates aux info.
2938 *) Fix bug in PKCS7_verify() which caused an infinite loop
2939 if there was more than one signature.
2940 [Sven Uszpelkat <su@celocom.de>]
2942 *) Major change in util/mkdef.pl to include extra information
2943 about each symbol, as well as presentig variables as well
2944 as functions. This change means that there's n more need
2945 to rebuild the .num files when some algorithms are excluded.
2948 *) Allow the verify time to be set by an application,
2949 rather than always using the current time.
2952 *) Phase 2 verify code reorganisation. The certificate
2953 verify code now looks up an issuer certificate by a
2954 number of criteria: subject name, authority key id
2955 and key usage. It also verifies self signed certificates
2956 by the same criteria. The main comparison function is
2957 X509_check_issued() which performs these checks.
2959 Lot of changes were necessary in order to support this
2960 without completely rewriting the lookup code.
2962 Authority and subject key identifier are now cached.
2964 The LHASH 'certs' is X509_STORE has now been replaced
2965 by a STACK_OF(X509_OBJECT). This is mainly because an
2966 LHASH can't store or retrieve multiple objects with
2967 the same hash value.
2969 As a result various functions (which were all internal
2970 use only) have changed to handle the new X509_STORE
2971 structure. This will break anything that messed round
2972 with X509_STORE internally.
2974 The functions X509_STORE_add_cert() now checks for an
2975 exact match, rather than just subject name.
2977 The X509_STORE API doesn't directly support the retrieval
2978 of multiple certificates matching a given criteria, however
2979 this can be worked round by performing a lookup first
2980 (which will fill the cache with candidate certificates)
2981 and then examining the cache for matches. This is probably
2982 the best we can do without throwing out X509_LOOKUP
2983 entirely (maybe later...).
2985 The X509_VERIFY_CTX structure has been enhanced considerably.
2987 All certificate lookup operations now go via a get_issuer()
2988 callback. Although this currently uses an X509_STORE it
2989 can be replaced by custom lookups. This is a simple way
2990 to bypass the X509_STORE hackery necessary to make this
2991 work and makes it possible to use more efficient techniques
2992 in future. A very simple version which uses a simple
2993 STACK for its trusted certificate store is also provided
2994 using X509_STORE_CTX_trusted_stack().
2996 The verify_cb() and verify() callbacks now have equivalents
2997 in the X509_STORE_CTX structure.
2999 X509_STORE_CTX also has a 'flags' field which can be used
3000 to customise the verify behaviour.
3003 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
3004 excludes S/MIME capabilities.
3007 *) When a certificate request is read in keep a copy of the
3008 original encoding of the signed data and use it when outputing
3009 again. Signatures then use the original encoding rather than
3010 a decoded, encoded version which may cause problems if the
3011 request is improperly encoded.
3014 *) For consistency with other BIO_puts implementations, call
3015 buffer_write(b, ...) directly in buffer_puts instead of calling
3018 In BIO_puts, increment b->num_write as in BIO_write.
3019 [Peter.Sylvester@EdelWeb.fr]
3021 *) Fix BN_mul_word for the case where the word is 0. (We have to use
3022 BN_zero, we may not return a BIGNUM with an array consisting of
3026 *) Avoid calling abort() from within the library when problems are
3027 detected, except if preprocessor symbols have been defined
3028 (such as REF_CHECK, BN_DEBUG etc.).
3031 *) New openssl application 'rsautl'. This utility can be
3032 used for low level RSA operations. DER public key
3033 BIO/fp routines also added.
3036 *) New Configure entry and patches for compiling on QNX 4.
3037 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
3039 *) A demo state-machine implementation was sponsored by
3040 Nuron (http://www.nuron.com/) and is now available in
3041 demos/state_machine.
3044 *) New options added to the 'dgst' utility for signature
3045 generation and verification.
3048 *) Unrecognized PKCS#7 content types are now handled via a
3049 catch all ASN1_TYPE structure. This allows unsupported
3050 types to be stored as a "blob" and an application can
3051 encode and decode it manually.
3054 *) Fix various signed/unsigned issues to make a_strex.c
3056 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
3058 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
3059 length if passed a buffer. ASN1_INTEGER_to_BN failed
3060 if passed a NULL BN and its argument was negative.
3061 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
3063 *) Modification to PKCS#7 encoding routines to output definite
3064 length encoding. Since currently the whole structures are in
3065 memory there's not real point in using indefinite length
3066 constructed encoding. However if OpenSSL is compiled with
3067 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
3070 *) Added BIO_vprintf() and BIO_vsnprintf().
3073 *) Added more prefixes to parse for in the the strings written
3074 through a logging bio, to cover all the levels that are available
3075 through syslog. The prefixes are now:
3077 PANIC, EMERG, EMR => LOG_EMERG
3078 ALERT, ALR => LOG_ALERT
3079 CRIT, CRI => LOG_CRIT
3080 ERROR, ERR => LOG_ERR
3081 WARNING, WARN, WAR => LOG_WARNING
3082 NOTICE, NOTE, NOT => LOG_NOTICE
3083 INFO, INF => LOG_INFO
3084 DEBUG, DBG => LOG_DEBUG
3086 and as before, if none of those prefixes are present at the
3087 beginning of the string, LOG_ERR is chosen.
3089 On Win32, the LOG_* levels are mapped according to this:
3091 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
3092 LOG_WARNING => EVENTLOG_WARNING_TYPE
3093 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
3097 *) Made it possible to reconfigure with just the configuration
3098 argument "reconf" or "reconfigure". The command line arguments
3099 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
3100 and are retrieved from there when reconfiguring.
3104 [Assar Westerlund <assar@sics.se>, Richard Levitte]
3106 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
3109 *) The obj_dat.pl script was messing up the sorting of object
3110 names. The reason was that it compared the quoted version
3111 of strings as a result "OCSP" > "OCSP Signing" because
3112 " > SPACE. Changed script to store unquoted versions of
3113 names and add quotes on output. It was also omitting some
3114 names from the lookup table if they were given a default
3115 value (that is if SN is missing it is given the same
3116 value as LN and vice versa), these are now added on the
3117 grounds that if an object has a name we should be able to
3118 look it up. Finally added warning output when duplicate
3119 short or long names are found.
3122 *) Changes needed for Tandem NSK.
3123 [Scott Uroff <scott@xypro.com>]
3125 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
3126 RSA_padding_check_SSLv23(), special padding was never detected
3127 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
3128 version rollback attacks was not effective.
3130 In s23_clnt.c, don't use special rollback-attack detection padding
3131 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
3132 client; similarly, in s23_srvr.c, don't do the rollback check if
3133 SSL 2.0 is the only protocol enabled in the server.
3136 *) Make it possible to get hexdumps of unprintable data with 'openssl
3137 asn1parse'. By implication, the functions ASN1_parse_dump() and
3138 BIO_dump_indent() are added.
3141 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
3142 these print out strings and name structures based on various
3143 flags including RFC2253 support and proper handling of
3144 multibyte characters. Added options to the 'x509' utility
3145 to allow the various flags to be set.
3148 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
3149 Also change the functions X509_cmp_current_time() and
3150 X509_gmtime_adj() work with an ASN1_TIME structure,
3151 this will enable certificates using GeneralizedTime in validity
3152 dates to be checked.
3155 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
3156 negative public key encodings) on by default,
3157 NO_NEG_PUBKEY_BUG can be set to disable it.
3160 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
3161 content octets. An i2c_ASN1_OBJECT is unnecessary because
3162 the encoding can be trivially obtained from the structure.
3165 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
3166 not read locks (CRYPTO_r_[un]lock).
3169 *) A first attempt at creating official support for shared
3170 libraries through configuration. I've kept it so the
3171 default is static libraries only, and the OpenSSL programs
3172 are always statically linked for now, but there are
3173 preparations for dynamic linking in place.
3174 This has been tested on Linux and Tru64.
3177 *) Randomness polling function for Win9x, as described in:
3178 Peter Gutmann, Software Generation of Practically Strong
3182 *) Fix so PRNG is seeded in req if using an already existing
3186 *) New options to smime application. -inform and -outform
3187 allow alternative formats for the S/MIME message including
3188 PEM and DER. The -content option allows the content to be
3189 specified separately. This should allow things like Netscape
3190 form signing output easier to verify.
3193 *) Fix the ASN1 encoding of tags using the 'long form'.
3196 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
3197 STRING types. These convert content octets to and from the
3198 underlying type. The actual tag and length octets are
3199 already assumed to have been read in and checked. These
3200 are needed because all other string types have virtually
3201 identical handling apart from the tag. By having versions
3202 of the ASN1 functions that just operate on content octets
3203 IMPLICIT tagging can be handled properly. It also allows
3204 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
3205 and ASN1_INTEGER are identical apart from the tag.
3208 *) Change the handling of OID objects as follows:
3210 - New object identifiers are inserted in objects.txt, following
3211 the syntax given in objects.README.
3212 - objects.pl is used to process obj_mac.num and create a new
3214 - obj_dat.pl is used to create a new obj_dat.h, using the data in
3217 This is currently kind of a hack, and the perl code in objects.pl
3218 isn't very elegant, but it works as I intended. The simplest way
3219 to check that it worked correctly is to look in obj_dat.h and
3220 check the array nid_objs and make sure the objects haven't moved
3221 around (this is important!). Additions are OK, as well as
3222 consistent name changes.
3225 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
3228 *) Addition of the command line parameter '-rand file' to 'openssl req'.
3229 The given file adds to whatever has already been seeded into the
3230 random pool through the RANDFILE configuration file option or
3231 environment variable, or the default random state file.
3234 *) mkstack.pl now sorts each macro group into lexical order.
3235 Previously the output order depended on the order the files
3236 appeared in the directory, resulting in needless rewriting
3240 *) Patches to make OpenSSL compile under Win32 again. Mostly
3241 work arounds for the VC++ problem that it treats func() as
3242 func(void). Also stripped out the parts of mkdef.pl that
3243 added extra typesafe functions: these no longer exist.
3246 *) Reorganisation of the stack code. The macros are now all
3247 collected in safestack.h . Each macro is defined in terms of
3248 a "stack macro" of the form SKM_<name>(type, a, b). The
3249 DEBUG_SAFESTACK is now handled in terms of function casts,
3250 this has the advantage of retaining type safety without the
3251 use of additional functions. If DEBUG_SAFESTACK is not defined
3252 then the non typesafe macros are used instead. Also modified the
3253 mkstack.pl script to handle the new form. Needs testing to see
3254 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
3255 the default if no major problems. Similar behaviour for ASN1_SET_OF
3256 and PKCS12_STACK_OF.
3259 *) When some versions of IIS use the 'NET' form of private key the
3260 key derivation algorithm is different. Normally MD5(password) is
3261 used as a 128 bit RC4 key. In the modified case
3262 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
3263 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
3264 as the old Netscape_RSA functions except they have an additional
3265 'sgckey' parameter which uses the modified algorithm. Also added
3266 an -sgckey command line option to the rsa utility. Thanks to
3267 Adrian Peck <bertie@ncipher.com> for posting details of the modified
3268 algorithm to openssl-dev.
3271 *) The evp_local.h macros were using 'c.##kname' which resulted in
3272 invalid expansion on some systems (SCO 5.0.5 for example).
3273 Corrected to 'c.kname'.
3274 [Phillip Porch <root@theporch.com>]
3276 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
3277 a STACK of email addresses from a certificate or request, these look
3278 in the subject name and the subject alternative name extensions and
3279 omit any duplicate addresses.
3282 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
3283 This makes DSA verification about 2 % faster.
3286 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
3287 (meaning that now 2^5 values will be precomputed, which is only 4 KB
3288 plus overhead for 1024 bit moduli).
3289 This makes exponentiations about 0.5 % faster for 1024 bit
3290 exponents (as measured by "openssl speed rsa2048").
3293 *) Rename memory handling macros to avoid conflicts with other
3295 Malloc => OPENSSL_malloc
3296 Malloc_locked => OPENSSL_malloc_locked
3297 Realloc => OPENSSL_realloc
3298 Free => OPENSSL_free
3301 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
3302 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
3305 *) CygWin32 support.
3306 [John Jarvie <jjarvie@newsguy.com>]
3308 *) The type-safe stack code has been rejigged. It is now only compiled
3309 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
3310 by default all type-specific stack functions are "#define"d back to
3311 standard stack functions. This results in more streamlined output
3312 but retains the type-safety checking possibilities of the original
3316 *) The STACK code has been cleaned up, and certain type declarations
3317 that didn't make a lot of sense have been brought in line. This has
3318 also involved a cleanup of sorts in safestack.h to more correctly
3319 map type-safe stack functions onto their plain stack counterparts.
3320 This work has also resulted in a variety of "const"ifications of
3321 lots of the code, especially "_cmp" operations which should normally
3322 be prototyped with "const" parameters anyway.
3325 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
3326 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
3327 (The PRNG state consists of two parts, the large pool 'state' and 'md',
3328 where all of 'md' is used each time the PRNG is used, but 'state'
3329 is used only indexed by a cyclic counter. As entropy may not be
3330 well distributed from the beginning, 'md' is important as a
3331 chaining variable. However, the output function chains only half
3332 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
3333 all of 'md', and seeding with STATE_SIZE dummy bytes will result
3334 in all of 'state' being rewritten, with the new values depending
3335 on virtually all of 'md'. This overcomes the 80 bit limitation.)
3338 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
3339 the handshake is continued after ssl_verify_cert_chain();
3340 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
3341 can lead to 'unexplainable' connection aborts later.
3342 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
3344 *) Major EVP API cipher revision.
3345 Add hooks for extra EVP features. This allows various cipher
3346 parameters to be set in the EVP interface. Support added for variable
3347 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
3348 setting of RC2 and RC5 parameters.
3350 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
3353 Remove lots of duplicated code from the EVP library. For example *every*
3354 cipher init() function handles the 'iv' in the same way according to the
3355 cipher mode. They also all do nothing if the 'key' parameter is NULL and
3356 for CFB and OFB modes they zero ctx->num.
3358 New functionality allows removal of S/MIME code RC2 hack.
3360 Most of the routines have the same form and so can be declared in terms
3363 By shifting this to the top level EVP_CipherInit() it can be removed from
3364 all individual ciphers. If the cipher wants to handle IVs or keys
3365 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
3368 Change lots of functions like EVP_EncryptUpdate() to now return a
3369 value: although software versions of the algorithms cannot fail
3370 any installed hardware versions can.
3373 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
3374 this option is set, tolerate broken clients that send the negotiated
3375 protocol version number instead of the requested protocol version
3379 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
3380 i.e. non-zero for export ciphersuites, zero otherwise.
3381 Previous versions had this flag inverted, inconsistent with
3382 rsa_tmp_cb (..._TMP_RSA_CB).
3383 [Bodo Moeller; problem reported by Amit Chopra]
3385 *) Add missing DSA library text string. Work around for some IIS
3386 key files with invalid SEQUENCE encoding.
3389 *) Add a document (doc/standards.txt) that list all kinds of standards
3390 and so on that are implemented in OpenSSL.
3393 *) Enhance c_rehash script. Old version would mishandle certificates
3394 with the same subject name hash and wouldn't handle CRLs at all.
3395 Added -fingerprint option to crl utility, to support new c_rehash
3399 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
3402 *) Fix for SSL server purpose checking. Server checking was
3403 rejecting certificates which had extended key usage present
3404 but no ssl client purpose.
3405 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
3407 *) Make PKCS#12 code work with no password. The PKCS#12 spec
3408 is a little unclear about how a blank password is handled.
3409 Since the password in encoded as a BMPString with terminating
3410 double NULL a zero length password would end up as just the
3411 double NULL. However no password at all is different and is
3412 handled differently in the PKCS#12 key generation code. NS
3413 treats a blank password as zero length. MSIE treats it as no
3414 password on export: but it will try both on import. We now do
3415 the same: PKCS12_parse() tries zero length and no password if
3416 the password is set to "" or NULL (NULL is now a valid password:
3417 it wasn't before) as does the pkcs12 application.
3420 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
3421 perror when PEM_read_bio_X509_REQ fails, the error message must
3422 be obtained from the error queue.
3425 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
3426 it in ERR_remove_state if appropriate, and change ERR_get_state
3427 accordingly to avoid race conditions (this is necessary because
3428 thread_hash is no longer constant once set).
3431 *) Bugfix for linux-elf makefile.one.
3434 *) RSA_get_default_method() will now cause a default
3435 RSA_METHOD to be chosen if one doesn't exist already.
3436 Previously this was only set during a call to RSA_new()
3437 or RSA_new_method(NULL) meaning it was possible for
3438 RSA_get_default_method() to return NULL.
3441 *) Added native name translation to the existing DSO code
3442 that will convert (if the flag to do so is set) filenames
3443 that are sufficiently small and have no path information
3444 into a canonical native form. Eg. "blah" converted to
3445 "libblah.so" or "blah.dll" etc.
3448 *) New function ERR_error_string_n(e, buf, len) which is like
3449 ERR_error_string(e, buf), but writes at most 'len' bytes
3450 including the 0 terminator. For ERR_error_string_n, 'buf'
3452 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
3454 *) CONF library reworked to become more general. A new CONF
3455 configuration file reader "class" is implemented as well as a
3456 new functions (NCONF_*, for "New CONF") to handle it. The now
3457 old CONF_* functions are still there, but are reimplemented to
3458 work in terms of the new functions. Also, a set of functions
3459 to handle the internal storage of the configuration data is
3460 provided to make it easier to write new configuration file
3461 reader "classes" (I can definitely see something reading a
3462 configuration file in XML format, for example), called _CONF_*,
3463 or "the configuration storage API"...
3465 The new configuration file reading functions are:
3467 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
3468 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
3470 NCONF_default, NCONF_WIN32
3472 NCONF_dump_fp, NCONF_dump_bio
3474 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
3475 NCONF_new creates a new CONF object. This works in the same way
3476 as other interfaces in OpenSSL, like the BIO interface.
3477 NCONF_dump_* dump the internal storage of the configuration file,
3478 which is useful for debugging. All other functions take the same
3479 arguments as the old CONF_* functions wth the exception of the
3480 first that must be a `CONF *' instead of a `LHASH *'.
3482 To make it easer to use the new classes with the old CONF_* functions,
3483 the function CONF_set_default_method is provided.
3486 *) Add '-tls1' option to 'openssl ciphers', which was already
3487 mentioned in the documentation but had not been implemented.
3488 (This option is not yet really useful because even the additional
3489 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
3492 *) Initial DSO code added into libcrypto for letting OpenSSL (and
3493 OpenSSL-based applications) load shared libraries and bind to
3494 them in a portable way.
3495 [Geoff Thorpe, with contributions from Richard Levitte]
3497 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
3499 *) Make sure _lrotl and _lrotr are only used with MSVC.
3501 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
3502 (the default implementation of RAND_status).
3504 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
3505 to '-clrext' (= clear extensions), as intended and documented.
3506 [Bodo Moeller; inconsistency pointed out by Michael Attili
3509 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
3510 was larger than the MD block size.
3511 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
3513 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
3514 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
3515 using the passed key: if the passed key was a private key the result
3516 of X509_print(), for example, would be to print out all the private key
3520 *) des_quad_cksum() byte order bug fix.
3521 [Ulf Möller, using the problem description in krb4-0.9.7, where
3522 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
3524 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
3526 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
3528 *) For easily testing in shell scripts whether some command
3529 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
3530 returns with exit code 0 iff no command of the given name is available.
3531 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
3532 the output goes to stdout and nothing is printed to stderr.
3533 Additional arguments are always ignored.
3535 Since for each cipher there is a command of the same name,
3536 the 'no-cipher' compilation switches can be tested this way.
3538 ('openssl no-XXX' is not able to detect pseudo-commands such
3539 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
3542 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
3545 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
3546 is set; it will be thrown away anyway because each handshake creates
3548 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
3549 to parameters -- in previous versions (since OpenSSL 0.9.3) the
3550 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
3551 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
3554 *) New s_client option -ign_eof: EOF at stdin is ignored, and
3555 'Q' and 'R' lose their special meanings (quit/renegotiate).
3556 This is part of what -quiet does; unlike -quiet, -ign_eof
3557 does not suppress any output.
3560 *) Add compatibility options to the purpose and trust code. The
3561 purpose X509_PURPOSE_ANY is "any purpose" which automatically
3562 accepts a certificate or CA, this was the previous behaviour,
3563 with all the associated security issues.
3565 X509_TRUST_COMPAT is the old trust behaviour: only and
3566 automatically trust self signed roots in certificate store. A
3567 new trust setting X509_TRUST_DEFAULT is used to specify that
3568 a purpose has no associated trust setting and it should instead
3569 use the value in the default purpose.
3572 *) Fix the PKCS#8 DSA private key code so it decodes keys again
3573 and fix a memory leak.
3576 *) In util/mkerr.pl (which implements 'make errors'), preserve
3577 reason strings from the previous version of the .c file, as
3578 the default to have only downcase letters (and digits) in
3579 automatically generated reasons codes is not always appropriate.
3582 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
3583 using strerror. Previously, ERR_reason_error_string() returned
3584 library names as reason strings for SYSerr; but SYSerr is a special
3585 case where small numbers are errno values, not library numbers.
3588 *) Add '-dsaparam' option to 'openssl dhparam' application. This
3589 converts DSA parameters into DH parameters. (When creating parameters,
3590 DSA_generate_parameters is used.)
3593 *) Include 'length' (recommended exponent length) in C code generated
3594 by 'openssl dhparam -C'.
3597 *) The second argument to set_label in perlasm was already being used
3598 so couldn't be used as a "file scope" flag. Moved to third argument
3602 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
3603 instead of RAND_bytes for encryption IVs and salts.
3606 *) Include RAND_status() into RAND_METHOD instead of implementing
3607 it only for md_rand.c Otherwise replacing the PRNG by calling
3608 RAND_set_rand_method would be impossible.
3611 *) Don't let DSA_generate_key() enter an infinite loop if the random
3612 number generation fails.
3615 *) New 'rand' application for creating pseudo-random output.
3618 *) Added configuration support for Linux/IA64
3619 [Rolf Haberrecker <rolf@suse.de>]
3621 *) Assembler module support for Mingw32.
3624 *) Shared library support for HPUX (in shlib/).
3625 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
3627 *) Shared library support for Solaris gcc.
3628 [Lutz Behnke <behnke@trustcenter.de>]
3630 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
3632 *) PKCS7_encrypt() was adding text MIME headers twice because they
3633 were added manually and by SMIME_crlf_copy().
3636 *) In bntest.c don't call BN_rand with zero bits argument.
3637 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
3639 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
3640 case was implemented. This caused BN_div_recp() to fail occasionally.
3643 *) Add an optional second argument to the set_label() in the perl
3644 assembly language builder. If this argument exists and is set
3645 to 1 it signals that the assembler should use a symbol whose
3646 scope is the entire file, not just the current function. This
3647 is needed with MASM which uses the format label:: for this scope.
3648 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
3650 *) Change the ASN1 types so they are typedefs by default. Before
3651 almost all types were #define'd to ASN1_STRING which was causing
3652 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
3656 *) Change names of new functions to the new get1/get0 naming
3657 convention: After 'get1', the caller owns a reference count
3658 and has to call ..._free; 'get0' returns a pointer to some
3659 data structure without incrementing reference counters.
3660 (Some of the existing 'get' functions increment a reference
3661 counter, some don't.)
3662 Similarly, 'set1' and 'add1' functions increase reference
3663 counters or duplicate objects.
3666 *) Allow for the possibility of temp RSA key generation failure:
3667 the code used to assume it always worked and crashed on failure.
3670 *) Fix potential buffer overrun problem in BIO_printf().
3671 [Ulf Möller, using public domain code by Patrick Powell; problem
3672 pointed out by David Sacerdote <das33@cornell.edu>]
3674 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
3675 RAND_egd() and RAND_status(). In the command line application,
3676 the EGD socket can be specified like a seed file using RANDFILE
3680 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
3681 Some CAs (e.g. Verisign) distribute certificates in this form.
3684 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
3685 list to exclude them. This means that no special compilation option
3686 is needed to use anonymous DH: it just needs to be included in the
3690 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
3691 EVP_MD_type. The old functionality is available in a new macro called
3692 EVP_MD_md(). Change code that uses it and update docs.
3695 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
3696 where the 'void *' argument is replaced by a function pointer argument.
3697 Previously 'void *' was abused to point to functions, which works on
3698 many platforms, but is not correct. As these functions are usually
3699 called by macros defined in OpenSSL header files, most source code
3700 should work without changes.
3703 *) <openssl/opensslconf.h> (which is created by Configure) now contains
3704 sections with information on -D... compiler switches used for
3705 compiling the library so that applications can see them. To enable
3706 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
3707 must be defined. E.g.,
3708 #define OPENSSL_ALGORITHM_DEFINES
3709 #include <openssl/opensslconf.h>
3710 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
3711 [Richard Levitte, Ulf and Bodo Möller]
3713 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
3717 *) Change the 'other' type in certificate aux info to a STACK_OF
3718 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
3719 the required ASN1 format: arbitrary types determined by an OID.
3722 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
3723 argument to 'req'. This is not because the function is newer or
3724 better than others it just uses the work 'NEW' in the certificate
3725 request header lines. Some software needs this.
3728 *) Reorganise password command line arguments: now passwords can be
3729 obtained from various sources. Delete the PEM_cb function and make
3730 it the default behaviour: i.e. if the callback is NULL and the
3731 usrdata argument is not NULL interpret it as a null terminated pass
3732 phrase. If usrdata and the callback are NULL then the pass phrase
3733 is prompted for as usual.
3736 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
3737 the support is automatically enabled. The resulting binaries will
3738 autodetect the card and use it if present.
3739 [Ben Laurie and Compaq Inc.]
3741 *) Work around for Netscape hang bug. This sends certificate request
3742 and server done in one record. Since this is perfectly legal in the
3743 SSL/TLS protocol it isn't a "bug" option and is on by default. See
3744 the bugs/SSLv3 entry for more info.
3747 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
3750 *) Add -rand argument to smime and pkcs12 applications and read/write
3754 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
3757 *) Add command line password options to the remaining applications.
3760 *) Bug fix for BN_div_recp() for numerators with an even number of
3764 *) More tests in bntest.c, and changed test_bn output.
3767 *) ./config recognizes MacOS X now.
3770 *) Bug fix for BN_div() when the first words of num and divsor are
3771 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
3774 *) Add support for various broken PKCS#8 formats, and command line
3775 options to produce them.
3778 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
3779 get temporary BIGNUMs from a BN_CTX.
3782 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
3786 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
3787 include a #define from the old name to the new. The original intent
3788 was that statically linked binaries could for example just call
3789 SSLeay_add_all_ciphers() to just add ciphers to the table and not
3790 link with digests. This never worked becayse SSLeay_add_all_digests()
3791 and SSLeay_add_all_ciphers() were in the same source file so calling
3792 one would link with the other. They are now in separate source files.
3795 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
3798 *) Use a less unusual form of the Miller-Rabin primality test (it used
3799 a binary algorithm for exponentiation integrated into the Miller-Rabin
3800 loop, our standard modexp algorithms are faster).
3803 *) Support for the EBCDIC character set completed.
3804 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
3806 *) Source code cleanups: use const where appropriate, eliminate casts,
3807 use void * instead of char * in lhash.
3810 *) Bugfix: ssl3_send_server_key_exchange was not restartable
3811 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
3812 this the server could overwrite ephemeral keys that the client
3816 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
3817 using 50 iterations of the Rabin-Miller test.
3819 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
3820 iterations of the Rabin-Miller test as required by the appendix
3821 to FIPS PUB 186[-1]) instead of DSA_is_prime.
3822 As BN_is_prime_fasttest includes trial division, DSA parameter
3823 generation becomes much faster.
3825 This implies a change for the callback functions in DSA_is_prime
3826 and DSA_generate_parameters: The callback function is called once
3827 for each positive witness in the Rabin-Miller test, not just
3828 occasionally in the inner loop; and the parameters to the
3829 callback function now provide an iteration count for the outer
3830 loop rather than for the current invocation of the inner loop.
3831 DSA_generate_parameters additionally can call the callback
3832 function with an 'iteration count' of -1, meaning that a
3833 candidate has passed the trial division test (when q is generated
3834 from an application-provided seed, trial division is skipped).
3837 *) New function BN_is_prime_fasttest that optionally does trial
3838 division before starting the Rabin-Miller test and has
3839 an additional BN_CTX * argument (whereas BN_is_prime always
3840 has to allocate at least one BN_CTX).
3841 'callback(1, -1, cb_arg)' is called when a number has passed the
3842 trial division stage.
3845 *) Fix for bug in CRL encoding. The validity dates weren't being handled
3849 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
3852 *) New function BN_pseudo_rand().
3855 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
3856 bignum version of BN_from_montgomery() with the working code from
3857 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
3861 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
3862 made it impossible to use the same SSL_SESSION data structure in
3863 SSL2 clients in multiple threads.
3866 *) The return value of RAND_load_file() no longer counts bytes obtained
3867 by stat(). RAND_load_file(..., -1) is new and uses the complete file
3868 to seed the PRNG (previously an explicit byte count was required).
3869 [Ulf Möller, Bodo Möller]
3871 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
3872 used (char *) instead of (void *) and had casts all over the place.
3875 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
3878 *) Retain source code compatibility for BN_prime_checks macro:
3879 BN_is_prime(..., BN_prime_checks, ...) now uses
3880 BN_prime_checks_for_size to determine the appropriate number of
3881 Rabin-Miller iterations.
3884 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
3885 DH_CHECK_P_NOT_SAFE_PRIME.
3886 (Check if this is true? OpenPGP calls them "strong".)
3889 *) Merge the functionality of "dh" and "gendh" programs into a new program
3890 "dhparam". The old programs are retained for now but will handle DH keys
3891 (instead of parameters) in future.
3894 *) Make the ciphers, s_server and s_client programs check the return values
3895 when a new cipher list is set.
3898 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
3899 ciphers. Before when the 56bit ciphers were enabled the sorting was
3902 The syntax for the cipher sorting has been extended to support sorting by
3903 cipher-strength (using the strength_bits hard coded in the tables).
3904 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
3906 Fix a bug in the cipher-command parser: when supplying a cipher command
3907 string with an "undefined" symbol (neither command nor alphanumeric
3908 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
3909 an error is flagged.
3911 Due to the strength-sorting extension, the code of the
3912 ssl_create_cipher_list() function was completely rearranged. I hope that
3913 the readability was also increased :-)
3914 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3916 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
3917 for the first serial number and places 2 in the serial number file. This
3918 avoids problems when the root CA is created with serial number zero and
3919 the first user certificate has the same issuer name and serial number
3923 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
3924 the new code. Add documentation for this stuff.
3927 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
3928 X509_*() to X509at_*() on the grounds that they don't handle X509
3929 structures and behave in an analagous way to the X509v3 functions:
3930 they shouldn't be called directly but wrapper functions should be used
3933 So we also now have some wrapper functions that call the X509at functions
3934 when passed certificate requests. (TO DO: similar things can be done with
3935 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
3936 things. Some of these need some d2i or i2d and print functionality
3937 because they handle more complex structures.)
3940 *) Add missing #ifndefs that caused missing symbols when building libssl
3941 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
3942 NO_RSA in ssl/s2*.c.
3943 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
3945 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
3946 has a return value which indicates the quality of the random data
3947 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
3948 error queue. New function RAND_pseudo_bytes() generates output that is
3949 guaranteed to be unique but not unpredictable. RAND_add is like
3950 RAND_seed, but takes an extra argument for an entropy estimate
3951 (RAND_seed always assumes full entropy).
3954 *) Do more iterations of Rabin-Miller probable prime test (specifically,
3955 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
3956 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
3957 in crypto/bn/bn_prime.c for the complete table). This guarantees a
3958 false-positive rate of at most 2^-80 for random input.
3961 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
3964 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
3965 in the 0.9.5 release), this returns the chain
3966 from an X509_CTX structure with a dup of the stack and all
3967 the X509 reference counts upped: so the stack will exist
3968 after X509_CTX_cleanup() has been called. Modify pkcs12.c
3971 Also make SSL_SESSION_print() print out the verify return
3975 *) Add manpage for the pkcs12 command. Also change the default
3976 behaviour so MAC iteration counts are used unless the new
3977 -nomaciter option is used. This improves file security and
3978 only older versions of MSIE (4.0 for example) need it.
3981 *) Honor the no-xxx Configure options when creating .DEF files.
3984 *) Add PKCS#10 attributes to field table: challengePassword,
3985 unstructuredName and unstructuredAddress. These are taken from
3986 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
3987 international characters are used.
3989 More changes to X509_ATTRIBUTE code: allow the setting of types
3990 based on strings. Remove the 'loc' parameter when adding
3991 attributes because these will be a SET OF encoding which is sorted
3995 *) Initial changes to the 'req' utility to allow request generation
3996 automation. This will allow an application to just generate a template
3997 file containing all the field values and have req construct the
4000 Initial support for X509_ATTRIBUTE handling. Stacks of these are
4001 used all over the place including certificate requests and PKCS#7
4002 structures. They are currently handled manually where necessary with
4003 some primitive wrappers for PKCS#7. The new functions behave in a
4004 manner analogous to the X509 extension functions: they allow
4005 attributes to be looked up by NID and added.
4007 Later something similar to the X509V3 code would be desirable to
4008 automatically handle the encoding, decoding and printing of the
4009 more complex types. The string types like challengePassword can
4010 be handled by the string table functions.
4012 Also modified the multi byte string table handling. Now there is
4013 a 'global mask' which masks out certain types. The table itself
4014 can use the flag STABLE_NO_MASK to ignore the mask setting: this
4015 is useful when for example there is only one permissible type
4016 (as in countryName) and using the mask might result in no valid
4020 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
4021 SSL_get_peer_finished to allow applications to obtain the latest
4022 Finished messages sent to the peer or expected from the peer,
4023 respectively. (SSL_get_peer_finished is usually the Finished message
4024 actually received from the peer, otherwise the protocol will be aborted.)
4026 As the Finished message are message digests of the complete handshake
4027 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
4028 be used for external authentication procedures when the authentication
4029 provided by SSL/TLS is not desired or is not enough.
4032 *) Enhanced support for Alpha Linux is added. Now ./config checks if
4033 the host supports BWX extension and if Compaq C is present on the
4034 $PATH. Just exploiting of the BWX extension results in 20-30%
4035 performance kick for some algorithms, e.g. DES and RC4 to mention
4036 a couple. Compaq C in turn generates ~20% faster code for MD5 and
4040 *) Add support for MS "fast SGC". This is arguably a violation of the
4041 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
4042 weak crypto and after checking the certificate is SGC a second one
4043 with strong crypto. MS SGC stops the first handshake after receiving
4044 the server certificate message and sends a second client hello. Since
4045 a server will typically do all the time consuming operations before
4046 expecting any further messages from the client (server key exchange
4047 is the most expensive) there is little difference between the two.
4049 To get OpenSSL to support MS SGC we have to permit a second client
4050 hello message after we have sent server done. In addition we have to
4051 reset the MAC if we do get this second client hello.
4054 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
4055 if a DER encoded private key is RSA or DSA traditional format. Changed
4056 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
4057 format DER encoded private key. Newer code should use PKCS#8 format which
4058 has the key type encoded in the ASN1 structure. Added DER private key
4059 support to pkcs8 application.
4062 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
4063 ciphersuites has been selected (as required by the SSL 3/TLS 1
4064 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
4065 is set, we interpret this as a request to violate the specification
4066 (the worst that can happen is a handshake failure, and 'correct'
4067 behaviour would result in a handshake failure anyway).
4070 *) In SSL_CTX_add_session, take into account that there might be multiple
4071 SSL_SESSION structures with the same session ID (e.g. when two threads
4072 concurrently obtain them from an external cache).
4073 The internal cache can handle only one SSL_SESSION with a given ID,
4074 so if there's a conflict, we now throw out the old one to achieve
4078 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
4079 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
4080 some routines that use cipher OIDs: some ciphers do not have OIDs
4081 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
4085 *) Simplify the trust setting structure and code. Now we just have
4086 two sequences of OIDs for trusted and rejected settings. These will
4087 typically have values the same as the extended key usage extension
4088 and any application specific purposes.
4090 The trust checking code now has a default behaviour: it will just
4091 check for an object with the same NID as the passed id. Functions can
4092 be provided to override either the default behaviour or the behaviour
4093 for a given id. SSL client, server and email already have functions
4094 in place for compatibility: they check the NID and also return "trusted"
4095 if the certificate is self signed.
4098 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
4099 traditional format into an EVP_PKEY structure.
4102 *) Add a password callback function PEM_cb() which either prompts for
4103 a password if usr_data is NULL or otherwise assumes it is a null
4104 terminated password. Allow passwords to be passed on command line
4105 environment or config files in a few more utilities.
4108 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
4109 keys. Add some short names for PKCS#8 PBE algorithms and allow them
4110 to be specified on the command line for the pkcs8 and pkcs12 utilities.
4111 Update documentation.
4114 *) Support for ASN1 "NULL" type. This could be handled before by using
4115 ASN1_TYPE but there wasn't any function that would try to read a NULL
4116 and produce an error if it couldn't. For compatibility we also have
4117 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
4118 don't allocate anything because they don't need to.
4121 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
4123 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
4125 *) Rebuild of the memory allocation routines used by OpenSSL code and
4126 possibly others as well. The purpose is to make an interface that
4127 provide hooks so anyone can build a separate set of allocation and
4128 deallocation routines to be used by OpenSSL, for example memory
4129 pool implementations, or something else, which was previously hard
4130 since Malloc(), Realloc() and Free() were defined as macros having
4131 the values malloc, realloc and free, respectively (except for Win32
4132 compilations). The same is provided for memory debugging code.
4133 OpenSSL already comes with functionality to find memory leaks, but
4134 this gives people a chance to debug other memory problems.
4136 With these changes, a new set of functions and macros have appeared:
4138 CRYPTO_set_mem_debug_functions() [F]
4139 CRYPTO_get_mem_debug_functions() [F]
4140 CRYPTO_dbg_set_options() [F]
4141 CRYPTO_dbg_get_options() [F]
4142 CRYPTO_malloc_debug_init() [M]
4144 The memory debug functions are NULL by default, unless the library
4145 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
4146 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
4147 gives the standard debugging functions that come with OpenSSL) or
4148 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
4149 provided by the library user) must be used. When the standard
4150 debugging functions are used, CRYPTO_dbg_set_options can be used to
4151 request additional information:
4152 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
4153 the CRYPTO_MDEBUG_xxx macro when compiling the library.
4155 Also, things like CRYPTO_set_mem_functions will always give the
4156 expected result (the new set of functions is used for allocation
4157 and deallocation) at all times, regardless of platform and compiler
4160 To finish it up, some functions that were never use in any other
4161 way than through macros have a new API and new semantic:
4164 CRYPTO_dbg_realloc()
4167 All macros of value have retained their old syntax.
4168 [Richard Levitte and Bodo Moeller]
4170 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
4171 ordering of SMIMECapabilities wasn't in "strength order" and there
4172 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
4176 *) Some ASN1 types with illegal zero length encoding (INTEGER,
4177 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
4178 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
4180 *) Merge in my S/MIME library for OpenSSL. This provides a simple
4181 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
4182 functionality to handle multipart/signed properly) and a utility
4183 called 'smime' to call all this stuff. This is based on code I
4184 originally wrote for Celo who have kindly allowed it to be
4185 included in OpenSSL.
4188 *) Add variants des_set_key_checked and des_set_key_unchecked of
4189 des_set_key (aka des_key_sched). Global variable des_check_key
4190 decides which of these is called by des_set_key; this way
4191 des_check_key behaves as it always did, but applications and
4192 the library itself, which was buggy for des_check_key == 1,
4193 have a cleaner way to pick the version they need.
4196 *) New function PKCS12_newpass() which changes the password of a
4200 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
4201 dynamic mix. In both cases the ids can be used as an index into the
4202 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
4203 functions so they accept a list of the field values and the
4204 application doesn't need to directly manipulate the X509_TRUST
4208 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
4212 *) Modify the way the V3 extension code looks up extensions. This now
4213 works in a similar way to the object code: we have some "standard"
4214 extensions in a static table which is searched with OBJ_bsearch()
4215 and the application can add dynamic ones if needed. The file
4216 crypto/x509v3/ext_dat.h now has the info: this file needs to be
4217 updated whenever a new extension is added to the core code and kept
4218 in ext_nid order. There is a simple program 'tabtest.c' which checks
4219 this. New extensions are not added too often so this file can readily
4220 be maintained manually.
4222 There are two big advantages in doing things this way. The extensions
4223 can be looked up immediately and no longer need to be "added" using
4224 X509V3_add_standard_extensions(): this function now does nothing.
4225 [Side note: I get *lots* of email saying the extension code doesn't
4226 work because people forget to call this function]
4227 Also no dynamic allocation is done unless new extensions are added:
4228 so if we don't add custom extensions there is no need to call
4229 X509V3_EXT_cleanup().
4232 *) Modify enc utility's salting as follows: make salting the default. Add a
4233 magic header, so unsalted files fail gracefully instead of just decrypting
4234 to garbage. This is because not salting is a big security hole, so people
4235 should be discouraged from doing it.
4238 *) Fixes and enhancements to the 'x509' utility. It allowed a message
4239 digest to be passed on the command line but it only used this
4240 parameter when signing a certificate. Modified so all relevant
4241 operations are affected by the digest parameter including the
4242 -fingerprint and -x509toreq options. Also -x509toreq choked if a
4243 DSA key was used because it didn't fix the digest.
4246 *) Initial certificate chain verify code. Currently tests the untrusted
4247 certificates for consistency with the verify purpose (which is set
4248 when the X509_STORE_CTX structure is set up) and checks the pathlength.
4250 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
4251 this is because it will reject chains with invalid extensions whereas
4252 every previous version of OpenSSL and SSLeay made no checks at all.
4254 Trust code: checks the root CA for the relevant trust settings. Trust
4255 settings have an initial value consistent with the verify purpose: e.g.
4256 if the verify purpose is for SSL client use it expects the CA to be
4257 trusted for SSL client use. However the default value can be changed to
4258 permit custom trust settings: one example of this would be to only trust
4259 certificates from a specific "secure" set of CAs.
4261 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
4262 which should be used for version portability: especially since the
4263 verify structure is likely to change more often now.
4265 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
4266 to set them. If not set then assume SSL clients will verify SSL servers
4269 Two new options to the verify program: -untrusted allows a set of
4270 untrusted certificates to be passed in and -purpose which sets the
4271 intended purpose of the certificate. If a purpose is set then the
4272 new chain verify code is used to check extension consistency.
4275 *) Support for the authority information access extension.
4278 *) Modify RSA and DSA PEM read routines to transparently handle
4279 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
4280 public keys in a format compatible with certificate
4281 SubjectPublicKeyInfo structures. Unfortunately there were already
4282 functions called *_PublicKey_* which used various odd formats so
4283 these are retained for compatibility: however the DSA variants were
4284 never in a public release so they have been deleted. Changed dsa/rsa
4285 utilities to handle the new format: note no releases ever handled public
4286 keys so we should be OK.
4288 The primary motivation for this change is to avoid the same fiasco
4289 that dogs private keys: there are several incompatible private key
4290 formats some of which are standard and some OpenSSL specific and
4291 require various evil hacks to allow partial transparent handling and
4292 even then it doesn't work with DER formats. Given the option anything
4293 other than PKCS#8 should be dumped: but the other formats have to
4294 stay in the name of compatibility.
4296 With public keys and the benefit of hindsight one standard format
4297 is used which works with EVP_PKEY, RSA or DSA structures: though
4298 it clearly returns an error if you try to read the wrong kind of key.
4300 Added a -pubkey option to the 'x509' utility to output the public key.
4301 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
4302 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
4303 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
4304 that do the same as the EVP_PKEY_assign_*() except they up the
4305 reference count of the added key (they don't "swallow" the
4309 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
4310 CRLs would fail if the file contained no certificates or no CRLs:
4311 added a new function to read in both types and return the number
4312 read: this means that if none are read it will be an error. The
4313 DER versions of the certificate and CRL reader would always fail
4314 because it isn't possible to mix certificates and CRLs in DER format
4315 without choking one or the other routine. Changed this to just read
4316 a certificate: this is the best we can do. Also modified the code
4317 in apps/verify.c to take notice of return codes: it was previously
4318 attempting to read in certificates from NULL pointers and ignoring
4319 any errors: this is one reason why the cert and CRL reader seemed
4320 to work. It doesn't check return codes from the default certificate
4321 routines: these may well fail if the certificates aren't installed.
4324 *) Code to support otherName option in GeneralName.
4327 *) First update to verify code. Change the verify utility
4328 so it warns if it is passed a self signed certificate:
4329 for consistency with the normal behaviour. X509_verify
4330 has been modified to it will now verify a self signed
4331 certificate if *exactly* the same certificate appears
4332 in the store: it was previously impossible to trust a
4333 single self signed certificate. This means that:
4334 openssl verify ss.pem
4335 now gives a warning about a self signed certificate but
4336 openssl verify -CAfile ss.pem ss.pem
4340 *) For servers, store verify_result in SSL_SESSION data structure
4341 (and add it to external session representation).
4342 This is needed when client certificate verifications fails,
4343 but an application-provided verification callback (set by
4344 SSL_CTX_set_cert_verify_callback) allows accepting the session
4345 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
4346 but returns 1): When the session is reused, we have to set
4347 ssl->verify_result to the appropriate error code to avoid
4349 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
4351 *) Fix a bug in the new PKCS#7 code: it didn't consider the
4352 case in PKCS7_dataInit() where the signed PKCS7 structure
4353 didn't contain any existing data because it was being created.
4354 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
4356 *) Add a salt to the key derivation routines in enc.c. This
4357 forms the first 8 bytes of the encrypted file. Also add a
4358 -S option to allow a salt to be input on the command line.
4361 *) New function X509_cmp(). Oddly enough there wasn't a function
4362 to compare two certificates. We do this by working out the SHA1
4363 hash and comparing that. X509_cmp() will be needed by the trust
4367 *) SSL_get1_session() is like SSL_get_session(), but increments
4368 the reference count in the SSL_SESSION returned.
4369 [Geoff Thorpe <geoff@eu.c2.net>]
4371 *) Fix for 'req': it was adding a null to request attributes.
4372 Also change the X509_LOOKUP and X509_INFO code to handle
4373 certificate auxiliary information.
4376 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
4380 *) Add the possibility to add extra information to the memory leak
4381 detecting output, to form tracebacks, showing from where each
4382 allocation was originated: CRYPTO_push_info("constant string") adds
4383 the string plus current file name and line number to a per-thread
4384 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
4385 is like calling CYRPTO_pop_info() until the stack is empty.
4386 Also updated memory leak detection code to be multi-thread-safe.
4389 *) Add options -text and -noout to pkcs7 utility and delete the
4390 encryption options which never did anything. Update docs.
4393 *) Add options to some of the utilities to allow the pass phrase
4394 to be included on either the command line (not recommended on
4395 OSes like Unix) or read from the environment. Update the
4396 manpages and fix a few bugs.
4399 *) Add a few manpages for some of the openssl commands.
4402 *) Fix the -revoke option in ca. It was freeing up memory twice,
4403 leaking and not finding already revoked certificates.
4406 *) Extensive changes to support certificate auxiliary information.
4407 This involves the use of X509_CERT_AUX structure and X509_AUX
4408 functions. An X509_AUX function such as PEM_read_X509_AUX()
4409 can still read in a certificate file in the usual way but it
4410 will also read in any additional "auxiliary information". By
4411 doing things this way a fair degree of compatibility can be
4412 retained: existing certificates can have this information added
4413 using the new 'x509' options.
4415 Current auxiliary information includes an "alias" and some trust
4416 settings. The trust settings will ultimately be used in enhanced
4417 certificate chain verification routines: currently a certificate
4418 can only be trusted if it is self signed and then it is trusted
4422 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
4423 The problem was that one of the replacement routines had not been working
4424 since SSLeay releases. For now the offending routine has been replaced
4425 with non-optimised assembler. Even so, this now gives around 95%
4426 performance improvement for 1024 bit RSA signs.
4429 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
4430 handling. Most clients have the effective key size in bits equal to
4431 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
4432 A few however don't do this and instead use the size of the decrypted key
4433 to determine the RC2 key length and the AlgorithmIdentifier to determine
4434 the effective key length. In this case the effective key length can still
4435 be 40 bits but the key length can be 168 bits for example. This is fixed
4436 by manually forcing an RC2 key into the EVP_PKEY structure because the
4437 EVP code can't currently handle unusual RC2 key sizes: it always assumes
4438 the key length and effective key length are equal.
4441 *) Add a bunch of functions that should simplify the creation of
4442 X509_NAME structures. Now you should be able to do:
4443 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
4444 and have it automatically work out the correct field type and fill in
4445 the structures. The more adventurous can try:
4446 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
4447 and it will (hopefully) work out the correct multibyte encoding.
4450 *) Change the 'req' utility to use the new field handling and multibyte
4451 copy routines. Before the DN field creation was handled in an ad hoc
4452 way in req, ca, and x509 which was rather broken and didn't support
4453 BMPStrings or UTF8Strings. Since some software doesn't implement
4454 BMPStrings or UTF8Strings yet, they can be enabled using the config file
4455 using the dirstring_type option. See the new comment in the default
4456 openssl.cnf for more info.
4459 *) Make crypto/rand/md_rand.c more robust:
4460 - Assure unique random numbers after fork().
4461 - Make sure that concurrent threads access the global counter and
4462 md serializably so that we never lose entropy in them
4463 or use exactly the same state in multiple threads.
4464 Access to the large state is not always serializable because
4465 the additional locking could be a performance killer, and
4466 md should be large enough anyway.
4469 *) New file apps/app_rand.c with commonly needed functionality
4470 for handling the random seed file.
4472 Use the random seed file in some applications that previously did not:
4474 dsaparam -genkey (which also ignored its '-rand' option),
4477 x509 (when signing).
4478 Except on systems with /dev/urandom, it is crucial to have a random
4479 seed file at least for key creation, DSA signing, and for DH exchanges;
4480 for RSA signatures we could do without one.
4482 gendh and gendsa (unlike genrsa) used to read only the first byte
4483 of each file listed in the '-rand' option. The function as previously
4484 found in genrsa is now in app_rand.c and is used by all programs
4485 that support '-rand'.
4488 *) In RAND_write_file, use mode 0600 for creating files;
4489 don't just chmod when it may be too late.
4492 *) Report an error from X509_STORE_load_locations
4493 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
4496 *) New function ASN1_mbstring_copy() this copies a string in either
4497 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
4498 into an ASN1_STRING type. A mask of permissible types is passed
4499 and it chooses the "minimal" type to use or an error if not type
4503 *) Add function equivalents to the various macros in asn1.h. The old
4504 macros are retained with an M_ prefix. Code inside the library can
4505 use the M_ macros. External code (including the openssl utility)
4506 should *NOT* in order to be "shared library friendly".
4509 *) Add various functions that can check a certificate's extensions
4510 to see if it usable for various purposes such as SSL client,
4511 server or S/MIME and CAs of these types. This is currently
4512 VERY EXPERIMENTAL but will ultimately be used for certificate chain
4513 verification. Also added a -purpose flag to x509 utility to
4514 print out all the purposes.
4517 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
4521 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
4522 for, obtain and decode and extension and obtain its critical flag.
4523 This allows all the necessary extension code to be handled in a
4524 single function call.
4527 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
4528 platforms. See crypto/rc4/rc4_enc.c for further details.
4531 *) New -noout option to asn1parse. This causes no output to be produced
4532 its main use is when combined with -strparse and -out to extract data
4533 from a file (which may not be in ASN.1 format).
4536 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
4537 when producing the local key id.
4538 [Richard Levitte <levitte@stacken.kth.se>]
4540 *) New option -dhparam in s_server. This allows a DH parameter file to be
4541 stated explicitly. If it is not stated then it tries the first server
4542 certificate file. The previous behaviour hard coded the filename
4546 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
4547 a public key to be input or output. For example:
4548 openssl rsa -in key.pem -pubout -out pubkey.pem
4549 Also added necessary DSA public key functions to handle this.
4552 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
4553 in the message. This was handled by allowing
4554 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
4555 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
4557 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
4558 to the end of the strings whereas this didn't. This would cause problems
4559 if strings read with d2i_ASN1_bytes() were later modified.
4560 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
4562 *) Fix for base64 decode bug. When a base64 bio reads only one line of
4563 data and it contains EOF it will end up returning an error. This is
4564 caused by input 46 bytes long. The cause is due to the way base64
4565 BIOs find the start of base64 encoded data. They do this by trying a
4566 trial decode on each line until they find one that works. When they
4567 do a flag is set and it starts again knowing it can pass all the
4568 data directly through the decoder. Unfortunately it doesn't reset
4569 the context it uses. This means that if EOF is reached an attempt
4570 is made to pass two EOFs through the context and this causes the
4571 resulting error. This can also cause other problems as well. As is
4572 usual with these problems it takes *ages* to find and the fix is
4573 trivial: move one line.
4574 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
4576 *) Ugly workaround to get s_client and s_server working under Windows. The
4577 old code wouldn't work because it needed to select() on sockets and the
4578 tty (for keypresses and to see if data could be written). Win32 only
4579 supports select() on sockets so we select() with a 1s timeout on the
4580 sockets and then see if any characters are waiting to be read, if none
4581 are present then we retry, we also assume we can always write data to
4582 the tty. This isn't nice because the code then blocks until we've
4583 received a complete line of data and it is effectively polling the
4584 keyboard at 1s intervals: however it's quite a bit better than not
4585 working at all :-) A dedicated Windows application might handle this
4586 with an event loop for example.
4589 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
4590 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
4591 will be called when RSA_sign() and RSA_verify() are used. This is useful
4592 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
4593 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
4594 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
4595 This necessitated the support of an extra signature type NID_md5_sha1
4596 for SSL signatures and modifications to the SSL library to use it instead
4597 of calling RSA_public_decrypt() and RSA_private_encrypt().
4600 *) Add new -verify -CAfile and -CApath options to the crl program, these
4601 will lookup a CRL issuers certificate and verify the signature in a
4602 similar way to the verify program. Tidy up the crl program so it
4603 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
4604 less strict. It will now permit CRL extensions even if it is not
4605 a V2 CRL: this will allow it to tolerate some broken CRLs.
4608 *) Initialize all non-automatic variables each time one of the openssl
4609 sub-programs is started (this is necessary as they may be started
4610 multiple times from the "OpenSSL>" prompt).
4611 [Lennart Bang, Bodo Moeller]
4613 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
4614 removing all other RSA functionality (this is what NO_RSA does). This
4615 is so (for example) those in the US can disable those operations covered
4616 by the RSA patent while allowing storage and parsing of RSA keys and RSA
4620 *) Non-copying interface to BIO pairs.
4621 (still largely untested)
4624 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
4625 ASCII string. This was handled independently in various places before.
4628 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
4629 UTF8 strings a character at a time.
4632 *) Use client_version from client hello to select the protocol
4633 (s23_srvr.c) and for RSA client key exchange verification
4634 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
4637 *) Add various utility functions to handle SPKACs, these were previously
4638 handled by poking round in the structure internals. Added new function
4639 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
4640 print, verify and generate SPKACs. Based on an original idea from
4641 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
4644 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
4647 *) Allow the config file extension section to be overwritten on the
4648 command line. Based on an original idea from Massimiliano Pala
4649 <madwolf@comune.modena.it>. The new option is called -extensions
4650 and can be applied to ca, req and x509. Also -reqexts to override
4651 the request extensions in req and -crlexts to override the crl extensions
4655 *) Add new feature to the SPKAC handling in ca. Now you can include
4656 the same field multiple times by preceding it by "XXXX." for example:
4659 this is the same syntax as used in the req config file.
4662 *) Allow certificate extensions to be added to certificate requests. These
4663 are specified in a 'req_extensions' option of the req section of the
4664 config file. They can be printed out with the -text option to req but
4665 are otherwise ignored at present.
4668 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
4669 data read consists of only the final block it would not decrypted because
4670 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
4671 A misplaced 'break' also meant the decrypted final block might not be
4672 copied until the next read.
4675 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
4676 a few extra parameters to the DH structure: these will be useful if
4677 for example we want the value of 'q' or implement X9.42 DH.
4680 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
4681 provides hooks that allow the default DSA functions or functions on a
4682 "per key" basis to be replaced. This allows hardware acceleration and
4683 hardware key storage to be handled without major modification to the
4684 library. Also added low level modexp hooks and CRYPTO_EX structure and
4685 associated functions.
4688 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
4689 as "read only": it can't be written to and the buffer it points to will
4690 not be freed. Reading from a read only BIO is much more efficient than
4691 a normal memory BIO. This was added because there are several times when
4692 an area of memory needs to be read from a BIO. The previous method was
4693 to create a memory BIO and write the data to it, this results in two
4694 copies of the data and an O(n^2) reading algorithm. There is a new
4695 function BIO_new_mem_buf() which creates a read only memory BIO from
4696 an area of memory. Also modified the PKCS#7 routines to use read only
4700 *) Bugfix: ssl23_get_client_hello did not work properly when called in
4701 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
4702 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
4703 but a retry condition occured while trying to read the rest.
4706 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
4707 NID_pkcs7_encrypted by default: this was wrong since this should almost
4708 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
4709 the encrypted data type: this is a more sensible place to put it and it
4710 allows the PKCS#12 code to be tidied up that duplicated this
4714 *) Changed obj_dat.pl script so it takes its input and output files on
4715 the command line. This should avoid shell escape redirection problems
4719 *) Initial support for certificate extension requests, these are included
4720 in things like Xenroll certificate requests. Included functions to allow
4721 extensions to be obtained and added.
4724 *) -crlf option to s_client and s_server for sending newlines as
4725 CRLF (as required by many protocols).
4728 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
4730 *) Install libRSAglue.a when OpenSSL is built with RSAref.
4731 [Ralf S. Engelschall]
4733 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
4734 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
4736 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
4740 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
4741 DH parameters/keys (q is lost during that conversion, but the resulting
4742 DH parameters contain its length).
4744 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
4745 much faster than DH_generate_parameters (which creates parameters
4746 where p = 2*q + 1), and also the smaller q makes DH computations
4747 much more efficient (160-bit exponentiation instead of 1024-bit
4748 exponentiation); so this provides a convenient way to support DHE
4749 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
4750 utter importance to use
4751 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4753 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4754 when such DH parameters are used, because otherwise small subgroup
4755 attacks may become possible!
4758 *) Avoid memory leak in i2d_DHparams.
4761 *) Allow the -k option to be used more than once in the enc program:
4762 this allows the same encrypted message to be read by multiple recipients.
4765 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
4766 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
4767 it will always use the numerical form of the OID, even if it has a short
4771 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
4772 method only got called if p,q,dmp1,dmq1,iqmp components were present,
4773 otherwise bn_mod_exp was called. In the case of hardware keys for example
4774 no private key components need be present and it might store extra data
4775 in the RSA structure, which cannot be accessed from bn_mod_exp.
4776 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
4777 private key operations.
4780 *) Added support for SPARC Linux.
4783 *) pem_password_cb function type incompatibly changed from
4784 typedef int pem_password_cb(char *buf, int size, int rwflag);
4786 ....(char *buf, int size, int rwflag, void *userdata);
4787 so that applications can pass data to their callbacks:
4788 The PEM[_ASN1]_{read,write}... functions and macros now take an
4789 additional void * argument, which is just handed through whenever
4790 the password callback is called.
4791 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
4793 New function SSL_CTX_set_default_passwd_cb_userdata.
4795 Compatibility note: As many C implementations push function arguments
4796 onto the stack in reverse order, the new library version is likely to
4797 interoperate with programs that have been compiled with the old
4798 pem_password_cb definition (PEM_whatever takes some data that
4799 happens to be on the stack as its last argument, and the callback
4800 just ignores this garbage); but there is no guarantee whatsoever that
4803 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
4804 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
4805 problems not only on Windows, but also on some Unix platforms.
4806 To avoid problematic command lines, these definitions are now in an
4807 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
4808 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
4811 *) MIPS III/IV assembler module is reimplemented.
4814 *) More DES library cleanups: remove references to srand/rand and
4815 delete an unused file.
4818 *) Add support for the the free Netwide assembler (NASM) under Win32,
4819 since not many people have MASM (ml) and it can be hard to obtain.
4820 This is currently experimental but it seems to work OK and pass all
4821 the tests. Check out INSTALL.W32 for info.
4824 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
4825 without temporary keys kept an extra copy of the server key,
4826 and connections with temporary keys did not free everything in case
4830 *) New function RSA_check_key and new openssl rsa option -check
4831 for verifying the consistency of RSA keys.
4832 [Ulf Moeller, Bodo Moeller]
4834 *) Various changes to make Win32 compile work:
4835 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
4836 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
4837 comparison" warnings.
4838 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
4841 *) Add a debugging option to PKCS#5 v2 key generation function: when
4842 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
4843 derived keys are printed to stderr.
4846 *) Copy the flags in ASN1_STRING_dup().
4847 [Roman E. Pavlov <pre@mo.msk.ru>]
4849 *) The x509 application mishandled signing requests containing DSA
4850 keys when the signing key was also DSA and the parameters didn't match.
4852 It was supposed to omit the parameters when they matched the signing key:
4853 the verifying software was then supposed to automatically use the CA's
4854 parameters if they were absent from the end user certificate.
4856 Omitting parameters is no longer recommended. The test was also
4857 the wrong way round! This was probably due to unusual behaviour in
4858 EVP_cmp_parameters() which returns 1 if the parameters match.
4859 This meant that parameters were omitted when they *didn't* match and
4860 the certificate was useless. Certificates signed with 'ca' didn't have
4862 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
4864 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
4865 The interface is as follows:
4866 Applications can use
4867 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
4868 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
4869 "off" is now the default.
4870 The library internally uses
4871 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
4872 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
4873 to disable memory-checking temporarily.
4875 Some inconsistent states that previously were possible (and were
4876 even the default) are now avoided.
4878 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
4879 with each memory chunk allocated; this is occasionally more helpful
4880 than just having a counter.
4882 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
4884 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
4888 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
4889 which largely parallels "options", but is for changing API behaviour,
4890 whereas "options" are about protocol behaviour.
4891 Initial "mode" flags are:
4893 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
4894 a single record has been written.
4895 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
4896 retries use the same buffer location.
4897 (But all of the contents must be
4901 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
4904 *) Fix problems with no-hmac etc.
4905 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
4907 *) New functions RSA_get_default_method(), RSA_set_method() and
4908 RSA_get_method(). These allows replacement of RSA_METHODs without having
4909 to mess around with the internals of an RSA structure.
4912 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
4913 Also really enable memory leak checks in openssl.c and in some
4915 [Chad C. Mulligan, Bodo Moeller]
4917 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
4918 up the length of negative integers. This has now been simplified to just
4919 store the length when it is first determined and use it later, rather
4920 than trying to keep track of where data is copied and updating it to
4922 [Steve Henson, reported by Brien Wheeler
4923 <bwheeler@authentica-security.com>]
4925 *) Add a new function PKCS7_signatureVerify. This allows the verification
4926 of a PKCS#7 signature but with the signing certificate passed to the
4927 function itself. This contrasts with PKCS7_dataVerify which assumes the
4928 certificate is present in the PKCS#7 structure. This isn't always the
4929 case: certificates can be omitted from a PKCS#7 structure and be
4930 distributed by "out of band" means (such as a certificate database).
4933 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
4934 function prototypes in pem.h, also change util/mkdef.pl to add the
4935 necessary function names.
4938 *) mk1mf.pl (used by Windows builds) did not properly read the
4939 options set by Configure in the top level Makefile, and Configure
4940 was not even able to write more than one option correctly.
4941 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
4944 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
4945 file to be loaded from a BIO or FILE pointer. The BIO version will
4946 for example allow memory BIOs to contain config info.
4949 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
4950 Whoever hopes to achieve shared-library compatibility across versions
4951 must use this, not the compile-time macro.
4952 (Exercise 0.9.4: Which is the minimum library version required by
4954 Note: All this applies only to multi-threaded programs, others don't
4958 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
4959 through a BIO pair triggered the default case, i.e.
4960 SSLerr(...,SSL_R_UNKNOWN_STATE).
4963 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
4964 can use the SSL library even if none of the specific BIOs is
4968 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
4969 for the encoded length.
4970 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
4972 *) Add initial documentation of the X509V3 functions.
4975 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
4976 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
4977 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
4978 secure PKCS#8 private key format with a high iteration count.
4981 *) Fix determination of Perl interpreter: A perl or perl5
4982 _directory_ in $PATH was also accepted as the interpreter.
4983 [Ralf S. Engelschall]
4985 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
4986 wrong with it but it was very old and did things like calling
4987 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
4991 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
4992 to use the new extension code.
4995 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
4996 with macros. This should make it easier to change their form, add extra
4997 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
5001 *) Add to configuration table a new entry that can specify an alternative
5002 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
5003 according to Mark Crispin <MRC@Panda.COM>.
5007 *) DES CBC did not update the IV. Weird.
5010 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
5011 Changing the behaviour of the former might break existing programs --
5012 where IV updating is needed, des_ncbc_encrypt can be used.
5015 *) When bntest is run from "make test" it drives bc to check its
5016 calculations, as well as internally checking them. If an internal check
5017 fails, it needs to cause bc to give a non-zero result or make test carries
5018 on without noticing the failure. Fixed.
5021 *) DES library cleanups.
5024 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
5025 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
5026 ciphers. NOTE: although the key derivation function has been verified
5027 against some published test vectors it has not been extensively tested
5028 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
5032 *) Instead of "mkdir -p", which is not fully portable, use new
5033 Perl script "util/mkdir-p.pl".
5036 *) Rewrite the way password based encryption (PBE) is handled. It used to
5037 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
5038 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
5039 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
5040 the 'parameter' field of the AlgorithmIdentifier is passed to the
5041 underlying key generation function so it must do its own ASN1 parsing.
5042 This has also changed the EVP_PBE_CipherInit() function which now has a
5043 'parameter' argument instead of literal salt and iteration count values
5044 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
5047 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
5048 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
5049 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
5050 KEY" because this clashed with PKCS#8 unencrypted string. Since this
5051 value was just used as a "magic string" and not used directly its
5052 value doesn't matter.
5055 *) Introduce some semblance of const correctness to BN. Shame C doesn't
5059 *) "linux-sparc64" configuration (ultrapenguin).
5060 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
5061 "linux-sparc" configuration.
5062 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
5064 *) config now generates no-xxx options for missing ciphers.
5067 *) Support the EBCDIC character set (work in progress).
5068 File ebcdic.c not yet included because it has a different license.
5069 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5071 *) Support BS2000/OSD-POSIX.
5072 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5074 *) Make callbacks for key generation use void * instead of char *.
5077 *) Make S/MIME samples compile (not yet tested).
5080 *) Additional typesafe stacks.
5083 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
5087 Changes between 0.9.3 and 0.9.3a [29 May 1999]
5089 *) New configuration variant "sco5-gcc".
5091 *) Updated some demos.
5092 [Sean O Riordain, Wade Scholine]
5094 *) Add missing BIO_free at exit of pkcs12 application.
5097 *) Fix memory leak in conf.c.
5100 *) Updates for Win32 to assembler version of MD5.
5103 *) Set #! path to perl in apps/der_chop to where we found it
5104 instead of using a fixed path.
5107 *) SHA library changes for irix64-mips4-cc.
5110 *) Improvements for VMS support.
5114 Changes between 0.9.2b and 0.9.3 [24 May 1999]
5116 *) Bignum library bug fix. IRIX 6 passes "make test" now!
5117 This also avoids the problems with SC4.2 and unpatched SC5.
5118 [Andy Polyakov <appro@fy.chalmers.se>]
5120 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
5121 These are required because of the typesafe stack would otherwise break
5122 existing code. If old code used a structure member which used to be STACK
5123 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
5124 sk_num or sk_value it would produce an error because the num, data members
5125 are not present in STACK_OF. Now it just produces a warning. sk_set
5126 replaces the old method of assigning a value to sk_value
5127 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
5128 that does this will no longer work (and should use sk_set instead) but
5129 this could be regarded as a "questionable" behaviour anyway.
5132 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
5133 correctly handle encrypted S/MIME data.
5136 *) Change type of various DES function arguments from des_cblock
5137 (which means, in function argument declarations, pointer to char)
5138 to des_cblock * (meaning pointer to array with 8 char elements),
5139 which allows the compiler to do more typechecking; it was like
5140 that back in SSLeay, but with lots of ugly casts.
5142 Introduce new type const_des_cblock.
5145 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
5146 problems: find RecipientInfo structure that matches recipient certificate
5147 and initialise the ASN1 structures properly based on passed cipher.
5150 *) Belatedly make the BN tests actually check the results.
5153 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
5154 to and from BNs: it was completely broken. New compilation option
5155 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
5156 key elements as negative integers.
5159 *) Reorganize and speed up MD5.
5160 [Andy Polyakov <appro@fy.chalmers.se>]
5163 [Richard Levitte <richard@levitte.org>]
5165 *) New option -out to asn1parse to allow the parsed structure to be
5166 output to a file. This is most useful when combined with the -strparse
5167 option to examine the output of things like OCTET STRINGS.
5170 *) Make SSL library a little more fool-proof by not requiring any longer
5171 that SSL_set_{accept,connect}_state be called before
5172 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
5173 in many applications because usually everything *appeared* to work as
5174 intended anyway -- now it really works as intended).
5177 *) Move openssl.cnf out of lib/.
5180 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
5181 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
5182 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
5183 [Ralf S. Engelschall]
5185 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
5186 handle PKCS#7 enveloped data properly.
5187 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
5189 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
5190 copying pointers. The cert_st handling is changed by this in
5191 various ways (and thus what used to be known as ctx->default_cert
5192 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
5193 any longer when s->cert does not give us what we need).
5194 ssl_cert_instantiate becomes obsolete by this change.
5195 As soon as we've got the new code right (possibly it already is?),
5196 we have solved a couple of bugs of the earlier code where s->cert
5197 was used as if it could not have been shared with other SSL structures.
5199 Note that using the SSL API in certain dirty ways now will result
5200 in different behaviour than observed with earlier library versions:
5201 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
5202 does not influence s as it used to.
5204 In order to clean up things more thoroughly, inside SSL_SESSION
5205 we don't use CERT any longer, but a new structure SESS_CERT
5206 that holds per-session data (if available); currently, this is
5207 the peer's certificate chain and, for clients, the server's certificate
5208 and temporary key. CERT holds only those values that can have
5209 meaningful defaults in an SSL_CTX.
5212 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
5213 from the internal representation. Various PKCS#7 fixes: remove some
5214 evil casts and set the enc_dig_alg field properly based on the signing
5218 *) Allow PKCS#12 password to be set from the command line or the
5219 environment. Let 'ca' get its config file name from the environment
5220 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
5224 *) Allow certificate policies extension to use an IA5STRING for the
5225 organization field. This is contrary to the PKIX definition but
5226 VeriSign uses it and IE5 only recognises this form. Document 'x509'
5230 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
5231 without disallowing inline assembler and the like for non-pedantic builds.
5234 *) Support Borland C++ builder.
5235 [Janez Jere <jj@void.si>, modified by Ulf Möller]
5240 *) SHA-1 cleanups and performance enhancements.
5241 [Andy Polyakov <appro@fy.chalmers.se>]
5243 *) Sparc v8plus assembler for the bignum library.
5244 [Andy Polyakov <appro@fy.chalmers.se>]
5246 *) Accept any -xxx and +xxx compiler options in Configure.
5249 *) Update HPUX configuration.
5252 *) Add missing sk_<type>_unshift() function to safestack.h
5253 [Ralf S. Engelschall]
5255 *) New function SSL_CTX_use_certificate_chain_file that sets the
5256 "extra_cert"s in addition to the certificate. (This makes sense
5257 only for "PEM" format files, as chains as a whole are not
5261 *) Support verify_depth from the SSL API.
5262 x509_vfy.c had what can be considered an off-by-one-error:
5263 Its depth (which was not part of the external interface)
5264 was actually counting the number of certificates in a chain;
5265 now it really counts the depth.
5268 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
5269 instead of X509err, which often resulted in confusing error
5270 messages since the error codes are not globally unique
5271 (e.g. an alleged error in ssl3_accept when a certificate
5272 didn't match the private key).
5274 *) New function SSL_CTX_set_session_id_context that allows to set a default
5275 value (so that you don't need SSL_set_session_id_context for each
5276 connection using the SSL_CTX).
5279 *) OAEP decoding bug fix.
5282 *) Support INSTALL_PREFIX for package builders, as proposed by
5286 *) New Configure options "threads" and "no-threads". For systems
5287 where the proper compiler options are known (currently Solaris
5288 and Linux), "threads" is the default.
5291 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
5294 *) Install various scripts to $(OPENSSLDIR)/misc, not to
5295 $(INSTALLTOP)/bin -- they shouldn't clutter directories
5296 such as /usr/local/bin.
5299 *) "make linux-shared" to build shared libraries.
5300 [Niels Poppe <niels@netbox.org>]
5302 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
5305 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
5306 extension adding in x509 utility.
5309 *) Remove NOPROTO sections and error code comments.
5312 *) Partial rewrite of the DEF file generator to now parse the ANSI
5316 *) New Configure options --prefix=DIR and --openssldir=DIR.
5319 *) Complete rewrite of the error code script(s). It is all now handled
5320 by one script at the top level which handles error code gathering,
5321 header rewriting and C source file generation. It should be much better
5322 than the old method: it now uses a modified version of Ulf's parser to
5323 read the ANSI prototypes in all header files (thus the old K&R definitions
5324 aren't needed for error creation any more) and do a better job of
5325 translating function codes into names. The old 'ASN1 error code imbedded
5326 in a comment' is no longer necessary and it doesn't use .err files which
5327 have now been deleted. Also the error code call doesn't have to appear all
5328 on one line (which resulted in some large lines...).
5331 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
5334 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
5335 0 (which usually indicates a closed connection), but continue reading.
5338 *) Fix some race conditions.
5341 *) Add support for CRL distribution points extension. Add Certificate
5342 Policies and CRL distribution points documentation.
5345 *) Move the autogenerated header file parts to crypto/opensslconf.h.
5348 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
5349 8 of keying material. Merlin has also confirmed interop with this fix
5350 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
5351 [Merlin Hughes <merlin@baltimore.ie>]
5353 *) Fix lots of warnings.
5354 [Richard Levitte <levitte@stacken.kth.se>]
5356 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
5357 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
5358 [Richard Levitte <levitte@stacken.kth.se>]
5360 *) Fix problems with sizeof(long) == 8.
5361 [Andy Polyakov <appro@fy.chalmers.se>]
5363 *) Change functions to ANSI C.
5366 *) Fix typos in error codes.
5367 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
5369 *) Remove defunct assembler files from Configure.
5372 *) SPARC v8 assembler BIGNUM implementation.
5373 [Andy Polyakov <appro@fy.chalmers.se>]
5375 *) Support for Certificate Policies extension: both print and set.
5376 Various additions to support the r2i method this uses.
5379 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
5380 return a const string when you are expecting an allocated buffer.
5383 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
5384 types DirectoryString and DisplayText.
5387 *) Add code to allow r2i extensions to access the configuration database,
5388 add an LHASH database driver and add several ctx helper functions.
5391 *) Fix an evil bug in bn_expand2() which caused various BN functions to
5392 fail when they extended the size of a BIGNUM.
5395 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
5396 support typesafe stack.
5399 *) Fix typo in SSL_[gs]et_options().
5400 [Nils Frostberg <nils@medcom.se>]
5402 *) Delete various functions and files that belonged to the (now obsolete)
5403 old X509V3 handling code.
5406 *) New Configure option "rsaref".
5409 *) Don't auto-generate pem.h.
5412 *) Introduce type-safe ASN.1 SETs.
5415 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
5416 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
5418 *) Introduce type-safe STACKs. This will almost certainly break lots of code
5419 that links with OpenSSL (well at least cause lots of warnings), but fear
5420 not: the conversion is trivial, and it eliminates loads of evil casts. A
5421 few STACKed things have been converted already. Feel free to convert more.
5422 In the fullness of time, I'll do away with the STACK type altogether.
5425 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
5426 specified in <certfile> by updating the entry in the index.txt file.
5427 This way one no longer has to edit the index.txt file manually for
5428 revoking a certificate. The -revoke option does the gory details now.
5429 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
5431 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
5432 `-text' option at all and this way the `-noout -text' combination was
5433 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
5434 [Ralf S. Engelschall]
5436 *) Make sure a corresponding plain text error message exists for the
5437 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
5438 verify callback function determined that a certificate was revoked.
5439 [Ralf S. Engelschall]
5441 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
5442 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
5443 all available cipers including rc5, which was forgotten until now.
5444 In order to let the testing shell script know which algorithms
5445 are available, a new (up to now undocumented) command
5446 "openssl list-cipher-commands" is used.
5449 *) Bugfix: s_client occasionally would sleep in select() when
5450 it should have checked SSL_pending() first.
5453 *) New functions DSA_do_sign and DSA_do_verify to provide access to
5454 the raw DSA values prior to ASN.1 encoding.
5457 *) Tweaks to Configure
5458 [Niels Poppe <niels@netbox.org>]
5460 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
5464 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
5467 *) New config option to avoid instructions that are illegal on the 80386.
5468 The default code is faster, but requires at least a 486.
5471 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
5472 SSL2_SERVER_VERSION (not used at all) macros, which are now the
5473 same as SSL2_VERSION anyway.
5476 *) New "-showcerts" option for s_client.
5479 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
5480 application. Various cleanups and fixes.
5483 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
5484 modify error routines to work internally. Add error codes and PBE init
5485 to library startup routines.
5488 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
5489 packing functions to asn1 and evp. Changed function names and error
5490 codes along the way.
5493 *) PKCS12 integration: and so it begins... First of several patches to
5494 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
5495 objects to objects.h
5498 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
5499 and display support for Thawte strong extranet extension.
5502 *) Add LinuxPPC support.
5503 [Jeff Dubrule <igor@pobox.org>]
5505 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
5506 bn_div_words in alpha.s.
5507 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
5509 *) Make sure the RSA OAEP test is skipped under -DRSAref because
5510 OAEP isn't supported when OpenSSL is built with RSAref.
5511 [Ulf Moeller <ulf@fitug.de>]
5513 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
5514 so they no longer are missing under -DNOPROTO.
5515 [Soren S. Jorvang <soren@t.dk>]
5518 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
5520 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
5521 doesn't work when the session is reused. Coming soon!
5524 *) Fix a security hole, that allows sessions to be reused in the wrong
5525 context thus bypassing client cert protection! All software that uses
5526 client certs and session caches in multiple contexts NEEDS PATCHING to
5527 allow session reuse! A fuller solution is in the works.
5528 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
5530 *) Some more source tree cleanups (removed obsolete files
5531 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
5532 permission on "config" script to be executable) and a fix for the INSTALL
5534 [Ulf Moeller <ulf@fitug.de>]
5536 *) Remove some legacy and erroneous uses of malloc, free instead of
5538 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
5540 *) Make rsa_oaep_test return non-zero on error.
5541 [Ulf Moeller <ulf@fitug.de>]
5543 *) Add support for native Solaris shared libraries. Configure
5544 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
5545 if someone would make that last step automatic.
5546 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
5548 *) ctx_size was not built with the right compiler during "make links". Fixed.
5551 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
5552 except NULL ciphers". This means the default cipher list will no longer
5553 enable NULL ciphers. They need to be specifically enabled e.g. with
5554 the string "DEFAULT:eNULL".
5557 *) Fix to RSA private encryption routines: if p < q then it would
5558 occasionally produce an invalid result. This will only happen with
5559 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
5562 *) Be less restrictive and allow also `perl util/perlpath.pl
5563 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
5564 because this way one can also use an interpreter named `perl5' (which is
5565 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
5566 installed as `perl').
5567 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5569 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
5570 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5572 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
5573 advapi32.lib to Win32 build and change the pem test comparision
5574 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
5575 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
5576 and crypto/des/ede_cbcm_enc.c.
5579 *) DES quad checksum was broken on big-endian architectures. Fixed.
5582 *) Comment out two functions in bio.h that aren't implemented. Fix up the
5583 Win32 test batch file so it (might) work again. The Win32 test batch file
5584 is horrible: I feel ill....
5587 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
5588 in e_os.h. Audit of header files to check ANSI and non ANSI
5589 sections: 10 functions were absent from non ANSI section and not exported
5590 from Windows DLLs. Fixed up libeay.num for new functions.
5593 *) Make `openssl version' output lines consistent.
5594 [Ralf S. Engelschall]
5596 *) Fix Win32 symbol export lists for BIO functions: Added
5597 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
5598 to ms/libeay{16,32}.def.
5599 [Ralf S. Engelschall]
5601 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
5602 fine under Unix and passes some trivial tests I've now added. But the
5603 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
5604 added to make sure no one expects that this stuff really works in the
5605 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
5606 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
5608 [Ralf S. Engelschall]
5610 *) Fix the generation of two part addresses in perl.
5611 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
5613 *) Add config entry for Linux on MIPS.
5614 [John Tobey <jtobey@channel1.com>]
5616 *) Make links whenever Configure is run, unless we are on Windoze.
5619 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
5620 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
5624 *) Add a useful kludge to allow package maintainers to specify compiler and
5625 other platforms details on the command line without having to patch the
5626 Configure script everytime: One now can use ``perl Configure
5627 <id>:<details>'', i.e. platform ids are allowed to have details appended
5628 to them (seperated by colons). This is treated as there would be a static
5629 pre-configured entry in Configure's %table under key <id> with value
5630 <details> and ``perl Configure <id>'' is called. So, when you want to
5631 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
5632 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
5633 now, which overrides the FreeBSD-elf entry on-the-fly.
5634 [Ralf S. Engelschall]
5636 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
5639 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
5640 on the `perl Configure ...' command line. This way one can compile
5641 OpenSSL libraries with Position Independent Code (PIC) which is needed
5642 for linking it into DSOs.
5643 [Ralf S. Engelschall]
5645 *) Remarkably, export ciphers were totally broken and no-one had noticed!
5649 *) Cleaned up the LICENSE document: The official contact for any license
5650 questions now is the OpenSSL core team under openssl-core@openssl.org.
5651 And add a paragraph about the dual-license situation to make sure people
5652 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
5653 to the OpenSSL toolkit.
5654 [Ralf S. Engelschall]
5656 *) General source tree makefile cleanups: Made `making xxx in yyy...'
5657 display consistent in the source tree and replaced `/bin/rm' by `rm'.
5658 Additonally cleaned up the `make links' target: Remove unnecessary
5659 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
5660 to speed processing and no longer clutter the display with confusing
5661 stuff. Instead only the actually done links are displayed.
5662 [Ralf S. Engelschall]
5664 *) Permit null encryption ciphersuites, used for authentication only. It used
5665 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
5666 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
5670 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
5671 signed attributes when verifying signatures (this would break them),
5672 the detached data encoding was wrong and public keys obtained using
5673 X509_get_pubkey() weren't freed.
5676 *) Add text documentation for the BUFFER functions. Also added a work around
5677 to a Win95 console bug. This was triggered by the password read stuff: the
5678 last character typed gets carried over to the next fread(). If you were
5679 generating a new cert request using 'req' for example then the last
5680 character of the passphrase would be CR which would then enter the first
5684 *) Added the new `Includes OpenSSL Cryptography Software' button as
5685 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
5686 button and can be used by applications based on OpenSSL to show the
5687 relationship to the OpenSSL project.
5688 [Ralf S. Engelschall]
5690 *) Remove confusing variables in function signatures in files
5691 ssl/ssl_lib.c and ssl/ssl.h.
5692 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5694 *) Don't install bss_file.c under PREFIX/include/
5695 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5697 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
5698 functions that return function pointers and has support for NT specific
5699 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
5700 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
5701 unsigned to signed types: this was killing the Win32 compile.
5704 *) Add new certificate file to stack functions,
5705 SSL_add_dir_cert_subjects_to_stack() and
5706 SSL_add_file_cert_subjects_to_stack(). These largely supplant
5707 SSL_load_client_CA_file(), and can be used to add multiple certs easily
5708 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
5709 This means that Apache-SSL and similar packages don't have to mess around
5710 to add as many CAs as they want to the preferred list.
5713 *) Experiment with doxygen documentation. Currently only partially applied to
5715 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
5716 openssl.doxy as the configuration file.
5719 *) Get rid of remaining C++-style comments which strict C compilers hate.
5720 [Ralf S. Engelschall, pointed out by Carlos Amengual]
5722 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
5723 compiled in by default: it has problems with large keys.
5726 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
5727 DH private keys and/or callback functions which directly correspond to
5728 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
5729 is needed for applications which have to configure certificates on a
5730 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
5732 For the RSA certificate situation is makes no difference, but
5733 for the DSA certificate situation this fixes the "no shared cipher"
5734 problem where the OpenSSL cipher selection procedure failed because the
5735 temporary keys were not overtaken from the context and the API provided
5736 no way to reconfigure them.
5737 The new functions now let applications reconfigure the stuff and they
5738 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
5739 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
5740 non-public-API function ssl_cert_instantiate() is used as a helper
5741 function and also to reduce code redundancy inside ssl_rsa.c.
5742 [Ralf S. Engelschall]
5744 *) Move s_server -dcert and -dkey options out of the undocumented feature
5745 area because they are useful for the DSA situation and should be
5746 recognized by the users.
5747 [Ralf S. Engelschall]
5749 *) Fix the cipher decision scheme for export ciphers: the export bits are
5750 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
5751 SSL_EXP_MASK. So, the original variable has to be used instead of the
5752 already masked variable.
5753 [Richard Levitte <levitte@stacken.kth.se>]
5755 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
5756 [Richard Levitte <levitte@stacken.kth.se>]
5758 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
5759 from `int' to `unsigned int' because it's a length and initialized by
5760 EVP_DigestFinal() which expects an `unsigned int *'.
5761 [Richard Levitte <levitte@stacken.kth.se>]
5763 *) Don't hard-code path to Perl interpreter on shebang line of Configure
5764 script. Instead use the usual Shell->Perl transition trick.
5765 [Ralf S. Engelschall]
5767 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
5768 (in addition to RSA certificates) to match the behaviour of `openssl dsa
5769 -noout -modulus' as it's already the case for `openssl rsa -noout
5770 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
5771 currently the public key is printed (a decision which was already done by
5772 `openssl dsa -modulus' in the past) which serves a similar purpose.
5773 Additionally the NO_RSA no longer completely removes the whole -modulus
5774 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
5776 [Ralf S. Engelschall]
5778 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
5779 BIO. See the source (crypto/evp/bio_ok.c) for more info.
5780 [Arne Ansper <arne@ats.cyber.ee>]
5782 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
5783 to be added. Now both 'req' and 'ca' can use new objects defined in the
5787 *) Add cool BIO that does syslog (or event log on NT).
5788 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
5790 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
5791 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
5792 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
5793 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
5796 *) Add preliminary config info for new extension code.
5799 *) Make RSA_NO_PADDING really use no padding.
5800 [Ulf Moeller <ulf@fitug.de>]
5802 *) Generate errors when private/public key check is done.
5805 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
5806 for some CRL extensions and new objects added.
5809 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
5810 key usage extension and fuller support for authority key id.
5813 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
5814 padding method for RSA, which is recommended for new applications in PKCS
5815 #1 v2.0 (RFC 2437, October 1998).
5816 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
5817 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
5818 against Bleichbacher's attack on RSA.
5819 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
5822 *) Updates to the new SSL compression code
5823 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5825 *) Fix so that the version number in the master secret, when passed
5826 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
5827 (because the server will not accept higher), that the version number
5828 is 0x03,0x01, not 0x03,0x00
5829 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5831 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
5832 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
5833 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
5836 *) Support for RAW extensions where an arbitrary extension can be
5837 created by including its DER encoding. See apps/openssl.cnf for
5841 *) Make sure latest Perl versions don't interpret some generated C array
5842 code as Perl array code in the crypto/err/err_genc.pl script.
5843 [Lars Weber <3weber@informatik.uni-hamburg.de>]
5845 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
5846 not many people have the assembler. Various Win32 compilation fixes and
5847 update to the INSTALL.W32 file with (hopefully) more accurate Win32
5851 *) Modify configure script 'Configure' to automatically create crypto/date.h
5852 file under Win32 and also build pem.h from pem.org. New script
5853 util/mkfiles.pl to create the MINFO file on environments that can't do a
5854 'make files': perl util/mkfiles.pl >MINFO should work.
5857 *) Major rework of DES function declarations, in the pursuit of correctness
5858 and purity. As a result, many evil casts evaporated, and some weirdness,
5859 too. You may find this causes warnings in your code. Zapping your evil
5860 casts will probably fix them. Mostly.
5863 *) Fix for a typo in asn1.h. Bug fix to object creation script
5864 obj_dat.pl. It considered a zero in an object definition to mean
5865 "end of object": none of the objects in objects.h have any zeros
5866 so it wasn't spotted.
5867 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
5869 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
5870 Masking (CBCM). In the absence of test vectors, the best I have been able
5871 to do is check that the decrypt undoes the encrypt, so far. Send me test
5872 vectors if you have them.
5875 *) Correct calculation of key length for export ciphers (too much space was
5876 allocated for null ciphers). This has not been tested!
5879 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
5880 message is now correct (it understands "crypto" and "ssl" on its
5881 command line). There is also now an "update" option. This will update
5882 the util/ssleay.num and util/libeay.num files with any new functions.
5884 perl util/mkdef.pl crypto ssl update
5885 it will update them.
5888 *) Overhauled the Perl interface (perl/*):
5889 - ported BN stuff to OpenSSL's different BN library
5890 - made the perl/ source tree CVS-aware
5891 - renamed the package from SSLeay to OpenSSL (the files still contain
5892 their history because I've copied them in the repository)
5893 - removed obsolete files (the test scripts will be replaced
5894 by better Test::Harness variants in the future)
5895 [Ralf S. Engelschall]
5897 *) First cut for a very conservative source tree cleanup:
5898 1. merge various obsolete readme texts into doc/ssleay.txt
5899 where we collect the old documents and readme texts.
5900 2. remove the first part of files where I'm already sure that we no
5901 longer need them because of three reasons: either they are just temporary
5902 files which were left by Eric or they are preserved original files where
5903 I've verified that the diff is also available in the CVS via "cvs diff
5904 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
5905 the crypto/md/ stuff).
5906 [Ralf S. Engelschall]
5908 *) More extension code. Incomplete support for subject and issuer alt
5909 name, issuer and authority key id. Change the i2v function parameters
5910 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
5911 what that's for :-) Fix to ASN1 macro which messed up
5912 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
5915 *) Preliminary support for ENUMERATED type. This is largely copied from the
5919 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
5920 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5922 *) Make sure `make rehash' target really finds the `openssl' program.
5923 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5925 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
5926 like to hear about it if this slows down other processors.
5929 *) Add CygWin32 platform information to Configure script.
5930 [Alan Batie <batie@aahz.jf.intel.com>]
5932 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
5933 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
5935 *) New program nseq to manipulate netscape certificate sequences
5938 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
5942 *) Fixes to BN code. Previously the default was to define BN_RECURSION
5943 but the BN code had some problems that would cause failures when
5944 doing certificate verification and some other functions.
5945 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5947 *) Add ASN1 and PEM code to support netscape certificate sequences.
5950 *) Add ASN1 and PEM code to support netscape certificate sequences.
5953 *) Add several PKIX and private extended key usage OIDs.
5956 *) Modify the 'ca' program to handle the new extension code. Modify
5957 openssl.cnf for new extension format, add comments.
5960 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
5961 and add a sample to openssl.cnf so req -x509 now adds appropriate
5965 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
5966 error code, add initial support to X509_print() and x509 application.
5969 *) Takes a deep breath and start addding X509 V3 extension support code. Add
5970 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
5971 stuff is currently isolated and isn't even compiled yet.
5974 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
5975 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
5976 Removed the versions check from X509 routines when loading extensions:
5977 this allows certain broken certificates that don't set the version
5978 properly to be processed.
5981 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
5982 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
5983 can still be regenerated with "make depend".
5986 *) Spelling mistake in C version of CAST-128.
5987 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
5989 *) Changes to the error generation code. The perl script err-code.pl
5990 now reads in the old error codes and retains the old numbers, only
5991 adding new ones if necessary. It also only changes the .err files if new
5992 codes are added. The makefiles have been modified to only insert errors
5993 when needed (to avoid needlessly modifying header files). This is done
5994 by only inserting errors if the .err file is newer than the auto generated
5995 C file. To rebuild all the error codes from scratch (the old behaviour)
5996 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
5997 or delete all the .err files.
6000 *) CAST-128 was incorrectly implemented for short keys. The C version has
6001 been fixed, but is untested. The assembler versions are also fixed, but
6002 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
6003 to regenerate it if needed.
6004 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
6005 Hagino <itojun@kame.net>]
6007 *) File was opened incorrectly in randfile.c.
6008 [Ulf Möller <ulf@fitug.de>]
6010 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
6011 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
6012 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
6013 al: it's just almost always a UTCTime. Note this patch adds new error
6014 codes so do a "make errors" if there are problems.
6017 *) Correct Linux 1 recognition in config.
6018 [Ulf Möller <ulf@fitug.de>]
6020 *) Remove pointless MD5 hash when using DSA keys in ca.
6021 [Anonymous <nobody@replay.com>]
6023 *) Generate an error if given an empty string as a cert directory. Also
6024 generate an error if handed NULL (previously returned 0 to indicate an
6025 error, but didn't set one).
6026 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
6028 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
6031 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
6032 parameters. This was causing a warning which killed off the Win32 compile.
6035 *) Remove C++ style comments from crypto/bn/bn_local.h.
6036 [Neil Costigan <neil.costigan@celocom.com>]
6038 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
6039 based on a text string, looking up short and long names and finally
6040 "dot" format. The "dot" format stuff didn't work. Added new function
6041 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
6042 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
6043 OID is not part of the table.
6046 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
6047 X509_LOOKUP_by_alias().
6050 *) Sort openssl functions by name.
6053 *) Get the gendsa program working (hopefully) and add it to app list. Remove
6054 encryption from sample DSA keys (in case anyone is interested the password
6058 *) Make _all_ *_free functions accept a NULL pointer.
6059 [Frans Heymans <fheymans@isaserver.be>]
6061 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
6063 [Anonymous <nobody@replay.com>]
6065 *) s_server should send the CAfile as acceptable CAs, not its own cert.
6066 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6068 *) Don't blow it for numeric -newkey arguments to apps/req.
6069 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6071 *) Temp key "for export" tests were wrong in s3_srvr.c.
6072 [Anonymous <nobody@replay.com>]
6074 *) Add prototype for temp key callback functions
6075 SSL_CTX_set_tmp_{rsa,dh}_callback().
6078 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
6079 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
6082 *) X509_name_add_entry() freed the wrong thing after an error.
6083 [Arne Ansper <arne@ats.cyber.ee>]
6085 *) rsa_eay.c would attempt to free a NULL context.
6086 [Arne Ansper <arne@ats.cyber.ee>]
6088 *) BIO_s_socket() had a broken should_retry() on Windoze.
6089 [Arne Ansper <arne@ats.cyber.ee>]
6091 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
6092 [Arne Ansper <arne@ats.cyber.ee>]
6094 *) Make sure the already existing X509_STORE->depth variable is initialized
6095 in X509_STORE_new(), but document the fact that this variable is still
6096 unused in the certificate verification process.
6097 [Ralf S. Engelschall]
6099 *) Fix the various library and apps files to free up pkeys obtained from
6100 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
6103 *) Fix reference counting in X509_PUBKEY_get(). This makes
6104 demos/maurice/example2.c work, amongst others, probably.
6105 [Steve Henson and Ben Laurie]
6107 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
6108 `openssl' and second, the shortcut symlinks for the `openssl <command>'
6109 are no longer created. This way we have a single and consistent command
6110 line interface `openssl <command>', similar to `cvs <command>'.
6111 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
6113 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
6114 BIT STRING wrapper always have zero unused bits.
6117 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
6120 *) Make the top-level INSTALL documentation easier to understand.
6123 *) Makefiles updated to exit if an error occurs in a sub-directory
6124 make (including if user presses ^C) [Paul Sutton]
6126 *) Make Montgomery context stuff explicit in RSA data structure.
6129 *) Fix build order of pem and err to allow for generated pem.h.
6132 *) Fix renumbering bug in X509_NAME_delete_entry().
6135 *) Enhanced the err-ins.pl script so it makes the error library number
6136 global and can add a library name. This is needed for external ASN1 and
6137 other error libraries.
6140 *) Fixed sk_insert which never worked properly.
6143 *) Fix ASN1 macros so they can handle indefinite length construted
6144 EXPLICIT tags. Some non standard certificates use these: they can now
6148 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
6149 into a single doc/ssleay.txt bundle. This way the information is still
6150 preserved but no longer messes up this directory. Now it's new room for
6151 the new set of documenation files.
6152 [Ralf S. Engelschall]
6154 *) SETs were incorrectly DER encoded. This was a major pain, because they
6155 shared code with SEQUENCEs, which aren't coded the same. This means that
6156 almost everything to do with SETs or SEQUENCEs has either changed name or
6157 number of arguments.
6158 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
6160 *) Fix test data to work with the above.
6163 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
6164 was already fixed by Eric for 0.9.1 it seems.
6165 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
6167 *) Autodetect FreeBSD3.
6170 *) Fix various bugs in Configure. This affects the following platforms:
6174 unixware-2.0-pentium
6178 *) Eliminate generated files from CVS. Reorder tests to regenerate files
6179 before they are needed.
6182 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
6186 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
6188 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
6189 changed SSLeay to OpenSSL in version strings.
6190 [Ralf S. Engelschall]
6192 *) Some fixups to the top-level documents.
6195 *) Fixed the nasty bug where rsaref.h was not found under compile-time
6196 because the symlink to include/ was missing.
6197 [Ralf S. Engelschall]
6199 *) Incorporated the popular no-RSA/DSA-only patches
6200 which allow to compile a RSA-free SSLeay.
6201 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
6203 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
6204 when "ssleay" is still not found.
6205 [Ralf S. Engelschall]
6207 *) Added more platforms to Configure: Cray T3E, HPUX 11,
6208 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
6210 *) Updated the README file.
6211 [Ralf S. Engelschall]
6213 *) Added various .cvsignore files in the CVS repository subdirs
6214 to make a "cvs update" really silent.
6215 [Ralf S. Engelschall]
6217 *) Recompiled the error-definition header files and added
6218 missing symbols to the Win32 linker tables.
6219 [Ralf S. Engelschall]
6221 *) Cleaned up the top-level documents;
6222 o new files: CHANGES and LICENSE
6223 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
6224 o merged COPYRIGHT into LICENSE
6225 o removed obsolete TODO file
6226 o renamed MICROSOFT to INSTALL.W32
6227 [Ralf S. Engelschall]
6229 *) Removed dummy files from the 0.9.1b source tree:
6230 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
6231 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
6232 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
6233 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
6234 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
6235 [Ralf S. Engelschall]
6237 *) Added various platform portability fixes.
6240 *) The Genesis of the OpenSSL rpject:
6241 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
6242 Young and Tim J. Hudson created while they were working for C2Net until
6244 [The OpenSSL Project]
6247 Changes between 0.9.0b and 0.9.1b [not released]
6249 *) Updated a few CA certificates under certs/
6252 *) Changed some BIGNUM api stuff.
6255 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
6256 DGUX x86, Linux Alpha, etc.
6259 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
6260 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
6264 *) Add -strparse option to asn1pars program which parses nested
6266 [Dr Stephen Henson <shenson@bigfoot.com>]
6268 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
6271 *) DSA fix for "ca" program.
6274 *) Added "-genkey" option to "dsaparam" program.
6277 *) Added RIPE MD160 (rmd160) message digest.
6280 *) Added -a (all) option to "ssleay version" command.
6283 *) Added PLATFORM define which is the id given to Configure.
6286 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
6289 *) Extended the ASN.1 parser routines.
6292 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
6295 *) Added a BN_CTX to the BN library.
6298 *) Fixed the weak key values in DES library
6301 *) Changed API in EVP library for cipher aliases.
6304 *) Added support for RC2/64bit cipher.
6307 *) Converted the lhash library to the crypto/mem.c functions.
6310 *) Added more recognized ASN.1 object ids.
6313 *) Added more RSA padding checks for SSL/TLS.
6316 *) Added BIO proxy/filter functionality.
6319 *) Added extra_certs to SSL_CTX which can be used
6320 send extra CA certificates to the client in the CA cert chain sending
6321 process. It can be configured with SSL_CTX_add_extra_chain_cert().
6324 *) Now Fortezza is denied in the authentication phase because
6325 this is key exchange mechanism is not supported by SSLeay at all.
6328 *) Additional PKCS1 checks.
6331 *) Support the string "TLSv1" for all TLS v1 ciphers.
6334 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
6335 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
6338 *) Fixed a few memory leaks.
6341 *) Fixed various code and comment typos.
6344 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
6345 bytes sent in the client random.
6346 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob