5 Changes between 0.9.7 and 0.9.8 [xx XXX 2002]
7 *) Change the ENGINE framework to automatically load engines
8 dynamically from specific directories unless they could be
9 found to already be built in or loaded. Move all the
10 current engines except for the cryptodev one to a new
12 The engines in engines/ are built as shared libraries if
13 the "shared" options was given to ./Configure or ./config.
14 Otherwise, they are inserted in libcrypto.a.
15 /usr/local/ssl/engines is the default directory for dynamic
16 engines, but that can be overriden with the environment
17 variable OPENSSL_ENGINES.
18 [Geoff Thorpe and Richard Levitte]
20 *) Add Makefile.shared, a helper makefile to build shared
21 libraries. Addapt Makefile.org.
24 *) Add version info to Win32 DLLs.
25 [Peter 'Luna' Runestig" <peter@runestig.com>]
27 *) Add new 'medium level' PKCS#12 API. Certificates and keys
28 can be added using this API to created arbitrary PKCS#12
29 files while avoiding the low level API.
31 New options to PKCS12_create(), key or cert can be NULL and
32 will then be omitted from the output file. The encryption
33 algorithm NIDs can be set to -1 for no encryption, the mac
34 iteration count can be set to 0 to omit the mac.
36 Enhance pkcs12 utility by making the -nokeys and -nocerts
37 options work when creating a PKCS#12 file. New option -nomac
38 to omit the mac, NONE can be set for an encryption algorithm.
39 New code is modified to use the enhanced PKCS12_create()
40 instead of the low level API.
43 *) Extend ASN1 encoder to support indefinite length constructed
44 encoding. This can output sequences tags and octet strings in
45 this form. Modify pk7_asn1.c to support indefinite length
46 encoding. This is experimental and needs additional code to
47 be useful, such as an ASN1 bio and some enhanced streaming
50 Extend template encode functionality so that tagging is passed
51 down to the template encoder.
54 *) Let 'openssl req' fail if an argument to '-newkey' is not
55 recognized instead of using RSA as a default.
58 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
59 As these are not official, they are not included in "ALL";
60 the "ECCdraft" ciphersuite group alias can be used to select them.
61 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
63 *) Add ECDH engine support.
64 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
66 *) Add ECDH in new directory crypto/ecdh/.
67 TODO: more general interface (return x coordinate, not its hash)
68 TODO: bug: pad x with leading zeros if necessary
69 [Douglas Stebila (Sun Microsystems Laboratories)]
71 *) Let BN_rand_range() abort with an error after 100 iterations
72 without success (which indicates a broken PRNG).
75 *) Change BN_mod_sqrt() so that it verifies that the input value
76 is really the square of the return value. (Previously,
77 BN_mod_sqrt would show GIGO behaviour.)
80 *) Add named elliptic curves over binary fields from X9.62, SECG,
81 and WAP/WTLS; add OIDs that were still missing.
83 [Sheueling Chang Shantz and Douglas Stebila
84 (Sun Microsystems Laboratories)]
86 *) Extend the EC library for elliptic curves over binary fields
87 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
94 EC_GROUP_new_curve_GF2m
95 EC_GROUP_set_curve_GF2m
96 EC_GROUP_get_curve_GF2m
97 EC_POINT_set_affine_coordinates_GF2m
98 EC_POINT_get_affine_coordinates_GF2m
99 EC_POINT_set_compressed_coordinates_GF2m
101 Point compression for binary fields is disabled by default for
102 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
105 As binary polynomials are represented as BIGNUMs, various members
106 of the EC_GROUP and EC_POINT data structures can be shared
107 between the implementations for prime fields and binary fields;
108 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
109 are essentially identical to their ..._GFp counterparts.
110 (For simplicity, the '..._GFp' prefix has been dropped from
111 various internal method names.)
113 An internal 'field_div' method (similar to 'field_mul' and
114 'field_sqr') has been added; this is used only for binary fields.
116 [Sheueling Chang Shantz and Douglas Stebila
117 (Sun Microsystems Laboratories)]
119 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
120 through methods ('mul', 'precompute_mult').
122 The generic implementations (now internally called 'ec_wNAF_mul'
123 and 'ec_wNAF_precomputed_mult') remain the default if these
124 methods are undefined.
126 [Sheueling Chang Shantz and Douglas Stebila
127 (Sun Microsystems Laboratories)]
129 *) New function EC_GROUP_get_degree, which is defined through
130 EC_METHOD. For curves over prime fields, this returns the bit
131 length of the modulus.
133 [Sheueling Chang Shantz and Douglas Stebila
134 (Sun Microsystems Laboratories)]
136 *) New functions EC_GROUP_dup, EC_POINT_dup.
137 (These simply call ..._new and ..._copy).
139 [Sheueling Chang Shantz and Douglas Stebila
140 (Sun Microsystems Laboratories)]
142 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
143 Polynomials are represented as BIGNUMs (where the sign bit is not
144 used) in the following functions [macros]:
147 BN_GF2m_sub [= BN_GF2m_add]
148 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
149 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
150 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
152 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
153 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
154 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
155 BN_GF2m_cmp [= BN_ucmp]
157 (Note that only the 'mod' functions are actually for fields GF(2^m).
158 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
160 For some functions, an the irreducible polynomial defining a
161 field can be given as an 'unsigned int[]' with strictly
162 decreasing elements giving the indices of those bits that are set;
163 i.e., p[] represents the polynomial
164 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
166 p[0] > p[1] > ... > p[k] = 0.
167 This applies to the following functions:
172 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
173 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
176 BN_GF2m_mod_solve_quad_arr
180 Conversion can be performed by the following functions:
185 bntest.c has additional tests for binary polynomial arithmetic.
187 Two implementations for BN_GF2m_mod_div() are available.
188 The default algorithm simply uses BN_GF2m_mod_inv() and
189 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
190 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
191 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
193 [Sheueling Chang Shantz and Douglas Stebila
194 (Sun Microsystems Laboratories)]
196 *) Add new error code 'ERR_R_DISABLED' that can be used when some
197 functionality is disabled at compile-time.
198 [Douglas Stebila <douglas.stebila@sun.com>]
200 *) Change default behaviour of 'openssl asn1parse' so that more
201 information is visible when viewing, e.g., a certificate:
203 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
204 mode the content of non-printable OCTET STRINGs is output in a
205 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
206 avoid the appearance of a printable string.
207 [Nils Larsch <nla@trustcenter.de>]
209 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
211 EC_GROUP_set_asn1_flag()
212 EC_GROUP_get_asn1_flag()
213 EC_GROUP_set_point_conversion_form()
214 EC_GROUP_get_point_conversion_form()
215 These control ASN1 encoding details:
216 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
217 has been set to OPENSSL_EC_NAMED_CURVE.
218 - Points are encoded in uncompressed form by default; options for
219 asn1_for are as for point2oct, namely
220 POINT_CONVERSION_COMPRESSED
221 POINT_CONVERSION_UNCOMPRESSED
222 POINT_CONVERSION_HYBRID
224 Also add 'seed' and 'seed_len' members to EC_GROUP with access
228 EC_GROUP_get_seed_len()
229 This is used only for ASN1 purposes (so far).
230 [Nils Larsch <nla@trustcenter.de>]
232 *) Add 'field_type' member to EC_METHOD, which holds the NID
233 of the appropriate field type OID. The new function
234 EC_METHOD_get_field_type() returns this value.
235 [Nils Larsch <nla@trustcenter.de>]
242 providing useful interfaces to EC_POINT_point2oct() and
243 EC_POINT_oct2point().
244 [Nils Larsch <nla@trustcenter.de>]
246 *) Change internals of the EC library so that the functions
247 EC_GROUP_set_generator()
248 EC_GROUP_get_generator()
250 EC_GROUP_get_cofactor()
251 are implemented directly in crypto/ec/ec_lib.c and not dispatched
252 to methods, which would lead to unnecessary code duplication when
253 adding different types of curves.
254 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
256 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
257 arithmetic, and such that modified wNAFs are generated
258 (which avoid length expansion in many cases).
261 *) Add a function EC_GROUP_check_discriminant() (defined via
262 EC_METHOD) that verifies that the curve discriminant is non-zero.
264 Add a function EC_GROUP_check() that makes some sanity tests
265 on a EC_GROUP, its generator and order. This includes
266 EC_GROUP_check_discriminant().
267 [Nils Larsch <nla@trustcenter.de>]
269 *) Add ECDSA in new directory crypto/ecdsa/.
271 Add applications 'openssl ecparam' and 'openssl ecdsa'
272 (these are based on 'openssl dsaparam' and 'openssl dsa').
274 ECDSA support is also included in various other files across the
275 library. Most notably,
276 - 'openssl req' now has a '-newkey ecdsa:file' option;
277 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
278 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
279 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
280 them suitable for ECDSA where domain parameters must be
281 extracted before the specific public key;
282 - ECDSA engine support has been added.
283 [Nils Larsch <nla@trustcenter.de>]
285 *) Include some named elliptic curves, and add OIDs from X9.62,
286 SECG, and WAP/WTLS. Each curve can be obtained from the new
288 EC_GROUP_new_by_nid(),
289 and the list of available named curves can be obtained with
290 EC_get_builtin_curves().
291 Also add a 'curve_name' member to EC_GROUP objects, which can be
295 [Nils Larsch <nla@trustcenter.de, Bodo Moeller]
297 Changes between 0.9.6h and 0.9.7 [XX xxx 2002]
299 *) Change from security patch (see 0.9.6e below) that did not affect
300 the 0.9.6 release series:
302 Remote buffer overflow in SSL3 protocol - an attacker could
303 supply an oversized master key in Kerberos-enabled versions.
307 *) Change the SSL kerb5 codes to match RFC 2712.
310 *) Make -nameopt work fully for req and add -reqopt switch.
311 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
313 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
314 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
316 *) Make sure tests can be performed even if the corresponding algorithms
317 have been removed entirely. This was also the last step to make
318 OpenSSL compilable with DJGPP under all reasonable conditions.
319 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
321 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
322 to allow version independent disabling of normally unselected ciphers,
323 which may be activated as a side-effect of selecting a single cipher.
325 (E.g., cipher list string "RSA" enables ciphersuites that are left
326 out of "ALL" because they do not provide symmetric encryption.
327 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
328 [Lutz Jaenicke, Bodo Moeller]
330 *) Add appropriate support for separate platform-dependent build
331 directories. The recommended way to make a platform-dependent
332 build directory is the following (tested on Linux), maybe with
335 # Place yourself outside of the OpenSSL source tree. In
336 # this example, the environment variable OPENSSL_SOURCE
337 # is assumed to contain the absolute OpenSSL source directory.
338 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
339 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
340 (cd $OPENSSL_SOURCE; find . -type f -o -type l) | while read F; do
341 mkdir -p `dirname $F`
342 ln -s $OPENSSL_SOURCE/$F $F
345 To be absolutely sure not to disturb the source tree, a "make clean"
346 is a good thing. If it isn't successfull, don't worry about it,
347 it probably means the source directory is very clean.
350 *) Make sure any ENGINE control commands make local copies of string
351 pointers passed to them whenever necessary. Otherwise it is possible
352 the caller may have overwritten (or deallocated) the original string
353 data when a later ENGINE operation tries to use the stored values.
354 [Götz Babin-Ebell <babinebell@trustcenter.de>]
356 *) Improve diagnostics in file reading and command-line digests.
357 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
359 *) Add AES modes CFB and OFB to the object database. Correct an
360 error in AES-CFB decryption.
363 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
364 allows existing EVP_CIPHER_CTX structures to be reused after
365 calling EVP_*Final(). This behaviour is used by encryption
366 BIOs and some applications. This has the side effect that
367 applications must explicitly clean up cipher contexts with
368 EVP_CIPHER_CTX_cleanup() or they will leak memory.
371 *) Check the values of dna and dnb in bn_mul_recursive before calling
372 bn_mul_comba (a non zero value means the a or b arrays do not contain
373 n2 elements) and fallback to bn_mul_normal if either is not zero.
376 *) Fix escaping of non-ASCII characters when using the -subj option
377 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
380 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
381 form for "surname", serialNumber has no short form.
382 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
383 therefore remove "mail" short name for "internet 7".
384 The OID for unique identifiers in X509 certificates is
385 x500UniqueIdentifier, not uniqueIdentifier.
386 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
389 *) Add an "init" command to the ENGINE config module and auto initialize
390 ENGINEs. Without any "init" command the ENGINE will be initialized
391 after all ctrl commands have been executed on it. If init=1 the
392 ENGINE is initailized at that point (ctrls before that point are run
393 on the uninitialized ENGINE and after on the initialized one). If
394 init=0 then the ENGINE will not be iniatialized at all.
397 *) Fix the 'app_verify_callback' interface so that the user-defined
398 argument is actually passed to the callback: In the
399 SSL_CTX_set_cert_verify_callback() prototype, the callback
400 declaration has been changed from
403 int (*cb)(X509_STORE_CTX *,void *);
404 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
405 i=s->ctx->app_verify_callback(&ctx)
406 has been changed into
407 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
409 To update applications using SSL_CTX_set_cert_verify_callback(),
410 a dummy argument can be added to their callback functions.
411 [D. K. Smetters <smetters@parc.xerox.com>]
413 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
414 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
416 *) Add and OPENSSL_LOAD_CONF define which will cause
417 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
418 This allows older applications to transparently support certain
419 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
420 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
421 load the config file and OPENSSL_add_all_algorithms_conf() which will
422 always load it have also been added.
425 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
426 Adjust NIDs and EVP layer.
427 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
429 *) Config modules support in openssl utility.
431 Most commands now load modules from the config file,
432 though in a few (such as version) this isn't done
433 because it couldn't be used for anything.
435 In the case of ca and req the config file used is
436 the same as the utility itself: that is the -config
437 command line option can be used to specify an
441 *) Move default behaviour from OPENSSL_config(). If appname is NULL
442 use "openssl_conf" if filename is NULL use default openssl config file.
445 *) Add an argument to OPENSSL_config() to allow the use of an alternative
446 config section name. Add a new flag to tolerate a missing config file
447 and move code to CONF_modules_load_file().
450 *) Support for crypto accelerator cards from Accelerated Encryption
451 Processing, www.aep.ie. (Use engine 'aep')
452 The support was copied from 0.9.6c [engine] and adapted/corrected
453 to work with the new engine framework.
454 [AEP Inc. and Richard Levitte]
456 *) Support for SureWare crypto accelerator cards from Baltimore
457 Technologies. (Use engine 'sureware')
458 The support was copied from 0.9.6c [engine] and adapted
459 to work with the new engine framework.
462 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
463 make the newer ENGINE framework commands for the CHIL engine work.
464 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
466 *) Make it possible to produce shared libraries on ReliantUNIX.
467 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
469 *) Add the configuration target debug-linux-ppro.
470 Make 'openssl rsa' use the general key loading routines
471 implemented in apps.c, and make those routines able to
472 handle the key format FORMAT_NETSCAPE and the variant
474 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
476 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
477 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
479 *) Add -keyform to rsautl, and document -engine.
480 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
482 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
483 BIO_R_NO_SUCH_FILE error code rather than the generic
484 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
489 ERR_peek_last_error_line
490 ERR_peek_last_error_line_data.
494 ERR_peek_error_line_data,
495 but report on the latest error recorded rather than the first one
496 still in the error queue.
497 [Ben Laurie, Bodo Moeller]
499 *) default_algorithms option in ENGINE config module. This allows things
501 default_algorithms = ALL
502 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
505 *) Prelminary ENGINE config module.
508 *) New experimental application configuration code.
511 *) Change the AES code to follow the same name structure as all other
512 symmetric ciphers, and behave the same way. Move everything to
513 the directory crypto/aes, thereby obsoleting crypto/rijndael.
514 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
516 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
517 [Ben Laurie and Theo de Raadt]
519 *) Add option to output public keys in req command.
520 [Massimiliano Pala madwolf@openca.org]
522 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
523 (up to about 10% better than before for P-192 and P-224).
526 *) New functions/macros
528 SSL_CTX_set_msg_callback(ctx, cb)
529 SSL_CTX_set_msg_callback_arg(ctx, arg)
530 SSL_set_msg_callback(ssl, cb)
531 SSL_set_msg_callback_arg(ssl, arg)
533 to request calling a callback function
535 void cb(int write_p, int version, int content_type,
536 const void *buf, size_t len, SSL *ssl, void *arg)
538 whenever a protocol message has been completely received
539 (write_p == 0) or sent (write_p == 1). Here 'version' is the
540 protocol version according to which the SSL library interprets
541 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
542 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
543 the content type as defined in the SSL 3.0/TLS 1.0 protocol
544 specification (change_cipher_spec(20), alert(21), handshake(22)).
545 'buf' and 'len' point to the actual message, 'ssl' to the
546 SSL object, and 'arg' is the application-defined value set by
547 SSL[_CTX]_set_msg_callback_arg().
549 'openssl s_client' and 'openssl s_server' have new '-msg' options
550 to enable a callback that displays all protocol messages.
553 *) Change the shared library support so shared libraries are built as
554 soon as the corresponding static library is finished, and thereby get
555 openssl and the test programs linked against the shared library.
556 This still only happens when the keyword "shard" has been given to
557 the configuration scripts.
559 NOTE: shared library support is still an experimental thing, and
560 backward binary compatibility is still not guaranteed.
561 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
563 *) Add support for Subject Information Access extension.
564 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
566 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
567 additional bytes when new memory had to be allocated, not just
568 when reusing an existing buffer.
571 *) New command line and configuration option 'utf8' for the req command.
572 This allows field values to be specified as UTF8 strings.
575 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
576 runs for the former and machine-readable output for the latter.
579 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
580 of the e-mail address in the DN (i.e., it will go into a certificate
581 extension only). The new configuration file option 'email_in_dn = no'
583 [Massimiliano Pala madwolf@openca.org]
585 *) Change all functions with names starting with des_ to be starting
586 with DES_ instead. Add wrappers that are compatible with libdes,
587 but are named _ossl_old_des_*. Finally, add macros that map the
588 des_* symbols to the corresponding _ossl_old_des_* if libdes
589 compatibility is desired. If OpenSSL 0.9.6c compatibility is
590 desired, the des_* symbols will be mapped to DES_*, with one
593 Since we provide two compatibility mappings, the user needs to
594 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
595 compatibility is desired. The default (i.e., when that macro
596 isn't defined) is OpenSSL 0.9.6c compatibility.
598 There are also macros that enable and disable the support of old
599 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
600 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
601 are defined, the default will apply: to support the old des routines.
603 In either case, one must include openssl/des.h to get the correct
604 definitions. Do not try to just include openssl/des_old.h, that
607 NOTE: This is a major break of an old API into a new one. Software
608 authors are encouraged to switch to the DES_ style functions. Some
609 time in the future, des_old.h and the libdes compatibility functions
610 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
611 default), and then completely removed.
614 *) Test for certificates which contain unsupported critical extensions.
615 If such a certificate is found during a verify operation it is
616 rejected by default: this behaviour can be overridden by either
617 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
618 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
619 X509_supported_extension() has also been added which returns 1 if a
620 particular extension is supported.
623 *) Modify the behaviour of EVP cipher functions in similar way to digests
624 to retain compatibility with existing code.
627 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
628 compatibility with existing code. In particular the 'ctx' parameter does
629 not have to be to be initialized before the call to EVP_DigestInit() and
630 it is tidied up after a call to EVP_DigestFinal(). New function
631 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
632 EVP_MD_CTX_copy() changed to not require the destination to be
633 initialized valid and new function EVP_MD_CTX_copy_ex() added which
634 requires the destination to be valid.
636 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
637 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
640 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
641 so that complete 'Handshake' protocol structures are kept in memory
642 instead of overwriting 'msg_type' and 'length' with 'body' data.
645 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
646 [Massimo Santin via Richard Levitte]
648 *) Major restructuring to the underlying ENGINE code. This includes
649 reduction of linker bloat, separation of pure "ENGINE" manipulation
650 (initialisation, etc) from functionality dealing with implementations
651 of specific crypto iterfaces. This change also introduces integrated
652 support for symmetric ciphers and digest implementations - so ENGINEs
653 can now accelerate these by providing EVP_CIPHER and EVP_MD
654 implementations of their own. This is detailed in crypto/engine/README
655 as it couldn't be adequately described here. However, there are a few
656 API changes worth noting - some RSA, DSA, DH, and RAND functions that
657 were changed in the original introduction of ENGINE code have now
658 reverted back - the hooking from this code to ENGINE is now a good
659 deal more passive and at run-time, operations deal directly with
660 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
661 dereferencing through an ENGINE pointer any more. Also, the ENGINE
662 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
663 they were not being used by the framework as there is no concept of a
664 BIGNUM_METHOD and they could not be generalised to the new
665 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
666 ENGINE_cpy() has been removed as it cannot be consistently defined in
670 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
673 *) Change mkdef.pl to sort symbols that get the same entry number,
674 and make sure the automatically generated functions ERR_load_*
675 become part of libeay.num as well.
678 *) New function SSL_renegotiate_pending(). This returns true once
679 renegotiation has been requested (either SSL_renegotiate() call
680 or HelloRequest/ClientHello receveived from the peer) and becomes
681 false once a handshake has been completed.
682 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
683 sends a HelloRequest, but does not ensure that a handshake takes
684 place. SSL_renegotiate_pending() is useful for checking if the
685 client has followed the request.)
688 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
689 By default, clients may request session resumption even during
690 renegotiation (if session ID contexts permit); with this option,
691 session resumption is possible only in the first handshake.
693 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
694 more bits available for options that should not be part of
695 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
698 *) Add some demos for certificate and certificate request creation.
701 *) Make maximum certificate chain size accepted from the peer application
702 settable (SSL*_get/set_max_cert_list()), as proposed by
703 "Douglas E. Engert" <deengert@anl.gov>.
706 *) Add support for shared libraries for Unixware-7
707 (Boyd Lynn Gerber <gerberb@zenez.com>).
710 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
711 be done prior to destruction. Use this to unload error strings from
712 ENGINEs that load their own error strings. NB: This adds two new API
713 functions to "get" and "set" this destroy handler in an ENGINE.
716 *) Alter all existing ENGINE implementations (except "openssl" and
717 "openbsd") to dynamically instantiate their own error strings. This
718 makes them more flexible to be built both as statically-linked ENGINEs
719 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
720 Also, add stub code to each that makes building them as self-contained
721 shared-libraries easier (see README.ENGINE).
724 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
725 implementations into applications that are completely implemented in
726 self-contained shared-libraries. The "dynamic" ENGINE exposes control
727 commands that can be used to configure what shared-library to load and
728 to control aspects of the way it is handled. Also, made an update to
729 the README.ENGINE file that brings its information up-to-date and
730 provides some information and instructions on the "dynamic" ENGINE
731 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
734 *) Make it possible to unload ranges of ERR strings with a new
735 "ERR_unload_strings" function.
738 *) Add a copy() function to EVP_MD.
741 *) Make EVP_MD routines take a context pointer instead of just the
742 md_data void pointer.
745 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
746 that the digest can only process a single chunk of data
747 (typically because it is provided by a piece of
748 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
749 is only going to provide a single chunk of data, and hence the
750 framework needn't accumulate the data for oneshot drivers.
753 *) As with "ERR", make it possible to replace the underlying "ex_data"
754 functions. This change also alters the storage and management of global
755 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
756 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
757 index counters. The API functions that use this state have been changed
758 to take a "class_index" rather than pointers to the class's local STACK
759 and counter, and there is now an API function to dynamically create new
760 classes. This centralisation allows us to (a) plug a lot of the
761 thread-safety problems that existed, and (b) makes it possible to clean
762 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
763 such data would previously have always leaked in application code and
764 workarounds were in place to make the memory debugging turn a blind eye
765 to it. Application code that doesn't use this new function will still
766 leak as before, but their memory debugging output will announce it now
767 rather than letting it slide.
769 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
770 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
771 has a return value to indicate success or failure.
774 *) Make it possible to replace the underlying "ERR" functions such that the
775 global state (2 LHASH tables and 2 locks) is only used by the "default"
776 implementation. This change also adds two functions to "get" and "set"
777 the implementation prior to it being automatically set the first time
778 any other ERR function takes place. Ie. an application can call "get",
779 pass the return value to a module it has just loaded, and that module
780 can call its own "set" function using that value. This means the
781 module's "ERR" operations will use (and modify) the error state in the
782 application and not in its own statically linked copy of OpenSSL code.
785 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
786 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
787 the operation, and provides a more encapsulated way for external code
788 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
789 to use these functions rather than manually incrementing the counts.
791 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
794 *) Add EVP test program.
797 *) Add symmetric cipher support to ENGINE. Expect the API to change!
800 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
801 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
802 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
803 These allow a CRL to be built without having to access X509_CRL fields
804 directly. Modify 'ca' application to use new functions.
807 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
808 bug workarounds. Rollback attack detection is a security feature.
809 The problem will only arise on OpenSSL servers when TLSv1 is not
810 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
811 Software authors not wanting to support TLSv1 will have special reasons
812 for their choice and can explicitly enable this option.
813 [Bodo Moeller, Lutz Jaenicke]
815 *) Rationalise EVP so it can be extended: don't include a union of
816 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
817 (similar to those existing for EVP_CIPHER_CTX).
822 EVP_MD_CTX_init(&md); /* new function call */
823 EVP_DigestInit(&md, EVP_sha1());
824 EVP_DigestUpdate(&md, in, len);
825 EVP_DigestFinal(&md, out, NULL);
826 EVP_MD_CTX_cleanup(&md); /* new function call */
830 *) Make DES key schedule conform to the usual scheme, as well as
831 correcting its structure. This means that calls to DES functions
832 now have to pass a pointer to a des_key_schedule instead of a
833 plain des_key_schedule (which was actually always a pointer
838 des_set_key_checked(..., &ks);
839 des_ncbc_encrypt(..., &ks, ...);
841 (Note that a later change renames 'des_...' into 'DES_...'.)
844 *) Initial reduction of linker bloat: the use of some functions, such as
845 PEM causes large amounts of unused functions to be linked in due to
846 poor organisation. For example pem_all.c contains every PEM function
847 which has a knock on effect of linking in large amounts of (unused)
848 ASN1 code. Grouping together similar functions and splitting unrelated
849 functions prevents this.
852 *) Cleanup of EVP macros.
855 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
859 *) Add initial OCSP responder support to ocsp application. The
860 revocation information is handled using the text based index
861 use by the ca application. The responder can either handle
862 requests generated internally, supplied in files (for example
863 via a CGI script) or using an internal minimal server.
866 *) Add configuration choices to get zlib compression for TLS.
869 *) Changes to Kerberos SSL for RFC 2712 compliance:
870 1. Implemented real KerberosWrapper, instead of just using
871 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
872 2. Implemented optional authenticator field of KerberosWrapper.
874 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
875 and authenticator structs; see crypto/krb5/.
877 Generalized Kerberos calls to support multiple Kerberos libraries.
878 [Vern Staats <staatsvr@asc.hpc.mil>,
879 Jeffrey Altman <jaltman@columbia.edu>
882 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
883 already does with RSA. testdsa.h now has 'priv_key/pub_key'
884 values for each of the key sizes rather than having just
885 parameters (and 'speed' generating keys each time).
888 *) Speed up EVP routines.
891 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
892 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
893 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
894 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
896 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
897 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
898 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
901 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
903 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
906 *) Added the OS2-EMX target.
907 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
909 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
910 to support NCONF routines in extension code. New function CONF_set_nconf()
911 to allow functions which take an NCONF to also handle the old LHASH
912 structure: this means that the old CONF compatible routines can be
913 retained (in particular wrt extensions) without having to duplicate the
914 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
917 *) Enhance the general user interface with mechanisms for inner control
918 and with possibilities to have yes/no kind of prompts.
921 *) Change all calls to low level digest routines in the library and
922 applications to use EVP. Add missing calls to HMAC_cleanup() and
923 don't assume HMAC_CTX can be copied using memcpy().
924 [Verdon Walker <VWalker@novell.com>, Steve Henson]
926 *) Add the possibility to control engines through control names but with
927 arbitrary arguments instead of just a string.
928 Change the key loaders to take a UI_METHOD instead of a callback
929 function pointer. NOTE: this breaks binary compatibility with earlier
930 versions of OpenSSL [engine].
931 Adapt the nCipher code for these new conditions and add a card insertion
935 *) Enhance the general user interface with mechanisms to better support
936 dialog box interfaces, application-defined prompts, the possibility
937 to use defaults (for example default passwords from somewhere else)
938 and interrupts/cancellations.
941 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
942 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
945 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
946 tidy up some unnecessarily weird code in 'sk_new()').
947 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
949 *) Change the key loading routines for ENGINEs to use the same kind
950 callback (pem_password_cb) as all other routines that need this
954 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
955 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
956 than this minimum value is recommended.
959 *) New random seeder for OpenVMS, using the system process statistics
960 that are easily reachable.
963 *) Windows apparently can't transparently handle global
964 variables defined in DLLs. Initialisations such as:
966 const ASN1_ITEM *it = &ASN1_INTEGER_it;
968 wont compile. This is used by the any applications that need to
969 declare their own ASN1 modules. This was fixed by adding the option
970 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
971 needed for static libraries under Win32.
974 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
975 setting of purpose and trust fields. New X509_STORE trust and
976 purpose functions and tidy up setting in other SSL functions.
979 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
980 structure. These are inherited by X509_STORE_CTX when it is
981 initialised. This allows various defaults to be set in the
982 X509_STORE structure (such as flags for CRL checking and custom
983 purpose or trust settings) for functions which only use X509_STORE_CTX
984 internally such as S/MIME.
986 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
987 trust settings if they are not set in X509_STORE. This allows X509_STORE
988 purposes and trust (in S/MIME for example) to override any set by default.
990 Add command line options for CRL checking to smime, s_client and s_server
994 *) Initial CRL based revocation checking. If the CRL checking flag(s)
995 are set then the CRL is looked up in the X509_STORE structure and
996 its validity and signature checked, then if the certificate is found
997 in the CRL the verify fails with a revoked error.
999 Various new CRL related callbacks added to X509_STORE_CTX structure.
1001 Command line options added to 'verify' application to support this.
1003 This needs some additional work, such as being able to handle multiple
1004 CRLs with different times, extension based lookup (rather than just
1005 by subject name) and ultimately more complete V2 CRL extension
1009 *) Add a general user interface API (crypto/ui/). This is designed
1010 to replace things like des_read_password and friends (backward
1011 compatibility functions using this new API are provided).
1012 The purpose is to remove prompting functions from the DES code
1013 section as well as provide for prompting through dialog boxes in
1014 a window system and the like.
1017 *) Add "ex_data" support to ENGINE so implementations can add state at a
1018 per-structure level rather than having to store it globally.
1021 *) Make it possible for ENGINE structures to be copied when retrieved by
1022 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
1023 This causes the "original" ENGINE structure to act like a template,
1024 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
1025 operational state can be localised to each ENGINE structure, despite the
1026 fact they all share the same "methods". New ENGINE structures returned in
1027 this case have no functional references and the return value is the single
1028 structural reference. This matches the single structural reference returned
1029 by ENGINE_by_id() normally, when it is incremented on the pre-existing
1033 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
1034 needs to match any other type at all we need to manually clear the
1038 *) Changes to the "openssl engine" utility to include;
1039 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
1040 about an ENGINE's available control commands.
1041 - executing control commands from command line arguments using the
1042 '-pre' and '-post' switches. '-post' is only used if '-t' is
1043 specified and the ENGINE is successfully initialised. The syntax for
1044 the individual commands are colon-separated, for example;
1045 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
1048 *) New dynamic control command support for ENGINEs. ENGINEs can now
1049 declare their own commands (numbers), names (strings), descriptions,
1050 and input types for run-time discovery by calling applications. A
1051 subset of these commands are implicitly classed as "executable"
1052 depending on their input type, and only these can be invoked through
1053 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
1054 can be based on user input, config files, etc). The distinction is
1055 that "executable" commands cannot return anything other than a boolean
1056 result and can only support numeric or string input, whereas some
1057 discoverable commands may only be for direct use through
1058 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
1059 pointers, or other custom uses. The "executable" commands are to
1060 support parameterisations of ENGINE behaviour that can be
1061 unambiguously defined by ENGINEs and used consistently across any
1062 OpenSSL-based application. Commands have been added to all the
1063 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
1064 control over shared-library paths without source code alterations.
1067 *) Changed all ENGINE implementations to dynamically allocate their
1068 ENGINEs rather than declaring them statically. Apart from this being
1069 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
1070 this also allows the implementations to compile without using the
1071 internal engine_int.h header.
1074 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
1075 'const' value. Any code that should be able to modify a RAND_METHOD
1076 should already have non-const pointers to it (ie. they should only
1077 modify their own ones).
1080 *) Made a variety of little tweaks to the ENGINE code.
1081 - "atalla" and "ubsec" string definitions were moved from header files
1082 to C code. "nuron" string definitions were placed in variables
1083 rather than hard-coded - allowing parameterisation of these values
1084 later on via ctrl() commands.
1085 - Removed unused "#if 0"'d code.
1086 - Fixed engine list iteration code so it uses ENGINE_free() to release
1087 structural references.
1088 - Constified the RAND_METHOD element of ENGINE structures.
1089 - Constified various get/set functions as appropriate and added
1090 missing functions (including a catch-all ENGINE_cpy that duplicates
1091 all ENGINE values onto a new ENGINE except reference counts/state).
1092 - Removed NULL parameter checks in get/set functions. Setting a method
1093 or function to NULL is a way of cancelling out a previously set
1094 value. Passing a NULL ENGINE parameter is just plain stupid anyway
1095 and doesn't justify the extra error symbols and code.
1096 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
1097 flags from engine_int.h to engine.h.
1098 - Changed prototypes for ENGINE handler functions (init(), finish(),
1099 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
1102 *) Implement binary inversion algorithm for BN_mod_inverse in addition
1103 to the algorithm using long division. The binary algorithm can be
1104 used only if the modulus is odd. On 32-bit systems, it is faster
1105 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
1106 roughly 5-15% for 256-bit moduli), so we use it only for moduli
1107 up to 450 bits. In 64-bit environments, the binary algorithm
1108 appears to be advantageous for much longer moduli; here we use it
1109 for moduli up to 2048 bits.
1112 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
1113 could not support the combine flag in choice fields.
1116 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
1117 extensions from a certificate request to the certificate.
1120 *) Allow multiple 'certopt' and 'nameopt' options to be separated
1121 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
1122 file: this allows the display of the certificate about to be
1123 signed to be customised, to allow certain fields to be included
1124 or excluded and extension details. The old system didn't display
1125 multicharacter strings properly, omitted fields not in the policy
1126 and couldn't display additional details such as extensions.
1129 *) Function EC_POINTs_mul for multiple scalar multiplication
1130 of an arbitrary number of elliptic curve points
1131 \sum scalars[i]*points[i],
1132 optionally including the generator defined for the EC_GROUP:
1133 scalar*generator + \sum scalars[i]*points[i].
1135 EC_POINT_mul is a simple wrapper function for the typical case
1136 that the point list has just one item (besides the optional
1140 *) First EC_METHODs for curves over GF(p):
1142 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
1143 operations and provides various method functions that can also
1144 operate with faster implementations of modular arithmetic.
1146 EC_GFp_mont_method() reuses most functions that are part of
1147 EC_GFp_simple_method, but uses Montgomery arithmetic.
1149 [Bodo Moeller; point addition and point doubling
1150 implementation directly derived from source code provided by
1151 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
1153 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
1154 crypto/ec/ec_lib.c):
1156 Curves are EC_GROUP objects (with an optional group generator)
1157 based on EC_METHODs that are built into the library.
1159 Points are EC_POINT objects based on EC_GROUP objects.
1161 Most of the framework would be able to handle curves over arbitrary
1162 finite fields, but as there are no obvious types for fields other
1163 than GF(p), some functions are limited to that for now.
1166 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
1167 that the file contains a complete HTTP response.
1170 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
1171 change the def and num file printf format specifier from "%-40sXXX"
1172 to "%-39s XXX". The latter will always guarantee a space after the
1173 field while the former will cause them to run together if the field
1174 is 40 of more characters long.
1177 *) Constify the cipher and digest 'method' functions and structures
1178 and modify related functions to take constant EVP_MD and EVP_CIPHER
1182 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
1183 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
1186 *) Modify EVP_Digest*() routines so they now return values. Although the
1187 internal software routines can never fail additional hardware versions
1191 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
1193 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
1194 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
1197 ERR_R_NESTED_ASN1_ERROR
1199 ERR_R_MISSING_ASN1_EOS
1200 were 4 .. 9, conflicting with
1201 ERR_LIB_RSA (= ERR_R_RSA_LIB)
1203 ERR_LIB_PEM (= ERR_R_PEM_LIB).
1204 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
1206 Add new error code 'ERR_R_INTERNAL_ERROR'.
1209 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
1213 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
1214 sets the subject name for a new request or supersedes the
1215 subject name in a given request. Formats that can be parsed are
1216 'CN=Some Name, OU=myOU, C=IT'
1218 'CN=Some Name/OU=myOU/C=IT'.
1220 Add options '-batch' and '-verbose' to 'openssl req'.
1221 [Massimiliano Pala <madwolf@hackmasters.net>]
1223 *) Introduce the possibility to access global variables through
1224 functions on platform were that's the best way to handle exporting
1225 global variables in shared libraries. To enable this functionality,
1226 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1227 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1228 is normally done by Configure or something similar).
1230 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1231 in the source file (foo.c) like this:
1233 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1234 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1236 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1237 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1239 OPENSSL_DECLARE_GLOBAL(int,foo);
1240 #define foo OPENSSL_GLOBAL_REF(foo)
1241 OPENSSL_DECLARE_GLOBAL(double,bar);
1242 #define bar OPENSSL_GLOBAL_REF(bar)
1244 The #defines are very important, and therefore so is including the
1245 header file everywhere where the defined globals are used.
1247 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1248 of ASN.1 items, but that structure is a bit different.
1250 The largest change is in util/mkdef.pl which has been enhanced with
1251 better and easier to understand logic to choose which symbols should
1252 go into the Windows .def files as well as a number of fixes and code
1253 cleanup (among others, algorithm keywords are now sorted
1254 lexicographically to avoid constant rewrites).
1257 *) In BN_div() keep a copy of the sign of 'num' before writing the
1258 result to 'rm' because if rm==num the value will be overwritten
1259 and produce the wrong result if 'num' is negative: this caused
1260 problems with BN_mod() and BN_nnmod().
1263 *) Function OCSP_request_verify(). This checks the signature on an
1264 OCSP request and verifies the signer certificate. The signer
1265 certificate is just checked for a generic purpose and OCSP request
1269 *) Add OCSP_check_validity() function to check the validity of OCSP
1270 responses. OCSP responses are prepared in real time and may only
1271 be a few seconds old. Simply checking that the current time lies
1272 between thisUpdate and nextUpdate max reject otherwise valid responses
1273 caused by either OCSP responder or client clock inaccuracy. Instead
1274 we allow thisUpdate and nextUpdate to fall within a certain period of
1275 the current time. The age of the response can also optionally be
1276 checked. Two new options -validity_period and -status_age added to
1280 *) If signature or public key algorithm is unrecognized print out its
1281 OID rather that just UNKNOWN.
1284 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1285 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1286 ID to be generated from the issuer certificate alone which can then be
1287 passed to OCSP_id_issuer_cmp().
1290 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1291 ASN1 modules to export functions returning ASN1_ITEM pointers
1292 instead of the ASN1_ITEM structures themselves. This adds several
1293 new macros which allow the underlying ASN1 function/structure to
1294 be accessed transparently. As a result code should not use ASN1_ITEM
1295 references directly (such as &X509_it) but instead use the relevant
1296 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1297 use of the new ASN1 code on platforms where exporting structures
1298 is problematical (for example in shared libraries) but exporting
1299 functions returning pointers to structures is not.
1302 *) Add support for overriding the generation of SSL/TLS session IDs.
1303 These callbacks can be registered either in an SSL_CTX or per SSL.
1304 The purpose of this is to allow applications to control, if they wish,
1305 the arbitrary values chosen for use as session IDs, particularly as it
1306 can be useful for session caching in multiple-server environments. A
1307 command-line switch for testing this (and any client code that wishes
1308 to use such a feature) has been added to "s_server".
1309 [Geoff Thorpe, Lutz Jaenicke]
1311 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
1312 of the form '#if defined(...) || defined(...) || ...' and
1313 '#if !defined(...) && !defined(...) && ...'. This also avoids
1314 the growing number of special cases it was previously handling.
1317 *) Make all configuration macros available for application by making
1318 sure they are available in opensslconf.h, by giving them names starting
1319 with "OPENSSL_" to avoid conflicts with other packages and by making
1320 sure e_os2.h will cover all platform-specific cases together with
1322 Additionally, it is now possible to define configuration/platform-
1323 specific names (called "system identities"). In the C code, these
1324 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1325 macro with the name beginning with "OPENSSL_SYS_", which is determined
1326 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1330 *) New option -set_serial to 'req' and 'x509' this allows the serial
1331 number to use to be specified on the command line. Previously self
1332 signed certificates were hard coded with serial number 0 and the
1333 CA options of 'x509' had to use a serial number in a file which was
1337 *) New options to 'ca' utility to support V2 CRL entry extensions.
1338 Currently CRL reason, invalidity date and hold instruction are
1339 supported. Add new CRL extensions to V3 code and some new objects.
1342 *) New function EVP_CIPHER_CTX_set_padding() this is used to
1343 disable standard block padding (aka PKCS#5 padding) in the EVP
1344 API, which was previously mandatory. This means that the data is
1345 not padded in any way and so the total length much be a multiple
1346 of the block size, otherwise an error occurs.
1349 *) Initial (incomplete) OCSP SSL support.
1352 *) New function OCSP_parse_url(). This splits up a URL into its host,
1353 port and path components: primarily to parse OCSP URLs. New -url
1354 option to ocsp utility.
1357 *) New nonce behavior. The return value of OCSP_check_nonce() now
1358 reflects the various checks performed. Applications can decide
1359 whether to tolerate certain situations such as an absent nonce
1360 in a response when one was present in a request: the ocsp application
1361 just prints out a warning. New function OCSP_add1_basic_nonce()
1362 this is to allow responders to include a nonce in a response even if
1363 the request is nonce-less.
1366 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1367 skipped when using openssl x509 multiple times on a single input file,
1368 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1371 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1372 set string type: to handle setting ASN1_TIME structures. Fix ca
1373 utility to correctly initialize revocation date of CRLs.
1376 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1377 the clients preferred ciphersuites and rather use its own preferences.
1378 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1379 Internet Explorer by ensuring unchanged hash method during stepup.
1380 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1383 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1384 to aes and add a new 'exist' option to print out symbols that don't
1388 *) Additional options to ocsp utility to allow flags to be set and
1389 additional certificates supplied.
1392 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
1393 OCSP client a number of certificate to only verify the response
1397 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1398 handle the new API. Currently only ECB, CBC modes supported. Add new
1401 Add TLS AES ciphersuites as described in RFC3268, "Advanced
1402 Encryption Standard (AES) Ciphersuites for Transport Layer
1403 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
1404 not enabled by default and were not part of the "ALL" ciphersuite
1405 alias because they were not yet official; they could be
1406 explicitly requested by specifying the "AESdraft" ciphersuite
1407 group alias. In the final release of OpenSSL 0.9.7, the group
1408 alias is called "AES" and is part of "ALL".)
1409 [Ben Laurie, Steve Henson, Bodo Moeller]
1411 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
1412 request to response.
1415 *) Functions for OCSP responders. OCSP_request_onereq_count(),
1416 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1417 extract information from a certificate request. OCSP_response_create()
1418 creates a response and optionally adds a basic response structure.
1419 OCSP_basic_add1_status() adds a complete single response to a basic
1420 response and returns the OCSP_SINGLERESP structure just added (to allow
1421 extensions to be included for example). OCSP_basic_add1_cert() adds a
1422 certificate to a basic response and OCSP_basic_sign() signs a basic
1423 response with various flags. New helper functions ASN1_TIME_check()
1424 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1425 (converts ASN1_TIME to GeneralizedTime).
1428 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1429 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1430 structure from a certificate. X509_pubkey_digest() digests the public_key
1431 contents: this is used in various key identifiers.
1434 *) Make sk_sort() tolerate a NULL argument.
1435 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1437 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1438 passed by the function are trusted implicitly. If any of them signed the
1439 response then it is assumed to be valid and is not verified.
1442 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1443 to data. This was previously part of the PKCS7 ASN1 code. This
1444 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1445 [Steve Henson, reported by Kenneth R. Robinette
1446 <support@securenetterm.com>]
1448 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1449 routines: without these tracing memory leaks is very painful.
1450 Fix leaks in PKCS12 and PKCS7 routines.
1453 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1454 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1455 effectively meant GeneralizedTime would never be used. Now it
1456 is initialised to -1 but X509_time_adj() now has to check the value
1457 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1458 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1459 [Steve Henson, reported by Kenneth R. Robinette
1460 <support@securenetterm.com>]
1462 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1463 result in a zero length in the ASN1_INTEGER structure which was
1464 not consistent with the structure when d2i_ASN1_INTEGER() was used
1465 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1466 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1467 where it did not print out a minus for negative ASN1_INTEGER.
1470 *) Add summary printout to ocsp utility. The various functions which
1471 convert status values to strings have been renamed to:
1472 OCSP_response_status_str(), OCSP_cert_status_str() and
1473 OCSP_crl_reason_str() and are no longer static. New options
1474 to verify nonce values and to disable verification. OCSP response
1475 printout format cleaned up.
1478 *) Add additional OCSP certificate checks. These are those specified
1479 in RFC2560. This consists of two separate checks: the CA of the
1480 certificate being checked must either be the OCSP signer certificate
1481 or the issuer of the OCSP signer certificate. In the latter case the
1482 OCSP signer certificate must contain the OCSP signing extended key
1483 usage. This check is performed by attempting to match the OCSP
1484 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1485 in the OCSP_CERTID structures of the response.
1488 *) Initial OCSP certificate verification added to OCSP_basic_verify()
1489 and related routines. This uses the standard OpenSSL certificate
1490 verify routines to perform initial checks (just CA validity) and
1491 to obtain the certificate chain. Then additional checks will be
1492 performed on the chain. Currently the root CA is checked to see
1493 if it is explicitly trusted for OCSP signing. This is used to set
1494 a root CA as a global signing root: that is any certificate that
1495 chains to that CA is an acceptable OCSP signing certificate.
1498 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1499 extensions from a separate configuration file.
1500 As when reading extensions from the main configuration file,
1501 the '-extensions ...' option may be used for specifying the
1503 [Massimiliano Pala <madwolf@comune.modena.it>]
1505 *) New OCSP utility. Allows OCSP requests to be generated or
1506 read. The request can be sent to a responder and the output
1507 parsed, outputed or printed in text form. Not complete yet:
1508 still needs to check the OCSP response validity.
1511 *) New subcommands for 'openssl ca':
1512 'openssl ca -status <serial>' prints the status of the cert with
1513 the given serial number (according to the index file).
1514 'openssl ca -updatedb' updates the expiry status of certificates
1516 [Massimiliano Pala <madwolf@comune.modena.it>]
1518 *) New '-newreq-nodes' command option to CA.pl. This is like
1519 '-newreq', but calls 'openssl req' with the '-nodes' option
1520 so that the resulting key is not encrypted.
1521 [Damien Miller <djm@mindrot.org>]
1523 *) New configuration for the GNU Hurd.
1524 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1526 *) Initial code to implement OCSP basic response verify. This
1527 is currently incomplete. Currently just finds the signer's
1528 certificate and verifies the signature on the response.
1531 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1532 value of OPENSSLDIR. This is available via the new '-d' option
1533 to 'openssl version', and is also included in 'openssl version -a'.
1536 *) Allowing defining memory allocation callbacks that will be given
1537 file name and line number information in additional arguments
1538 (a const char* and an int). The basic functionality remains, as
1539 well as the original possibility to just replace malloc(),
1540 realloc() and free() by functions that do not know about these
1541 additional arguments. To register and find out the current
1542 settings for extended allocation functions, the following
1543 functions are provided:
1545 CRYPTO_set_mem_ex_functions
1546 CRYPTO_set_locked_mem_ex_functions
1547 CRYPTO_get_mem_ex_functions
1548 CRYPTO_get_locked_mem_ex_functions
1550 These work the same way as CRYPTO_set_mem_functions and friends.
1551 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1552 extended allocation function is enabled.
1553 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1554 a conventional allocation function is enabled.
1555 [Richard Levitte, Bodo Moeller]
1557 *) Finish off removing the remaining LHASH function pointer casts.
1558 There should no longer be any prototype-casting required when using
1559 the LHASH abstraction, and any casts that remain are "bugs". See
1560 the callback types and macros at the head of lhash.h for details
1561 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1564 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1565 If /dev/[u]random devices are not available or do not return enough
1566 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1568 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1569 /etc/entropy will be queried once each in this sequence, quering stops
1570 when enough entropy was collected without querying more sockets.
1573 *) Change the Unix RAND_poll() variant to be able to poll several
1574 random devices, as specified by DEVRANDOM, until a sufficient amount
1575 of data has been collected. We spend at most 10 ms on each file
1576 (select timeout) and read in non-blocking mode. DEVRANDOM now
1577 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1578 (previously it was just the string "/dev/urandom"), so on typical
1579 platforms the 10 ms delay will never occur.
1580 Also separate out the Unix variant to its own file, rand_unix.c.
1581 For VMS, there's a currently-empty rand_vms.c.
1584 *) Move OCSP client related routines to ocsp_cl.c. These
1585 provide utility functions which an application needing
1586 to issue a request to an OCSP responder and analyse the
1587 response will typically need: as opposed to those which an
1588 OCSP responder itself would need which will be added later.
1590 OCSP_request_sign() signs an OCSP request with an API similar
1591 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1592 response. OCSP_response_get1_basic() extracts basic response
1593 from response. OCSP_resp_find_status(): finds and extracts status
1594 information from an OCSP_CERTID structure (which will be created
1595 when the request structure is built). These are built from lower
1596 level functions which work on OCSP_SINGLERESP structures but
1597 wont normally be used unless the application wishes to examine
1598 extensions in the OCSP response for example.
1600 Replace nonce routines with a pair of functions.
1601 OCSP_request_add1_nonce() adds a nonce value and optionally
1602 generates a random value. OCSP_check_nonce() checks the
1603 validity of the nonce in an OCSP response.
1606 *) Change function OCSP_request_add() to OCSP_request_add0_id().
1607 This doesn't copy the supplied OCSP_CERTID and avoids the
1608 need to free up the newly created id. Change return type
1609 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1610 This can then be used to add extensions to the request.
1611 Deleted OCSP_request_new(), since most of its functionality
1612 is now in OCSP_REQUEST_new() (and the case insensitive name
1613 clash) apart from the ability to set the request name which
1614 will be added elsewhere.
1617 *) Update OCSP API. Remove obsolete extensions argument from
1618 various functions. Extensions are now handled using the new
1619 OCSP extension code. New simple OCSP HTTP function which
1620 can be used to send requests and parse the response.
1623 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1624 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1625 uses the special reorder version of SET OF to sort the attributes
1626 and reorder them to match the encoded order. This resolves a long
1627 standing problem: a verify on a PKCS7 structure just after signing
1628 it used to fail because the attribute order did not match the
1629 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1630 it uses the received order. This is necessary to tolerate some broken
1631 software that does not order SET OF. This is handled by encoding
1632 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1633 to produce the required SET OF.
1636 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1637 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1638 files to get correct declarations of the ASN.1 item variables.
1641 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1642 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1643 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1644 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1645 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1646 ASN1_ITEM and no wrapper functions.
1649 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1650 replace the old function pointer based I/O routines. Change most of
1651 the *_d2i_bio() and *_d2i_fp() functions to use these.
1654 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1655 lines, recognice more "algorithms" that can be deselected, and make
1656 it complain about algorithm deselection that isn't recognised.
1659 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
1660 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1661 to use new functions. Add NO_ASN1_OLD which can be set to remove
1662 some old style ASN1 functions: this can be used to determine if old
1663 code will still work when these eventually go away.
1666 *) New extension functions for OCSP structures, these follow the
1667 same conventions as certificates and CRLs.
1670 *) New function X509V3_add1_i2d(). This automatically encodes and
1671 adds an extension. Its behaviour can be customised with various
1672 flags to append, replace or delete. Various wrappers added for
1673 certifcates and CRLs.
1676 *) Fix to avoid calling the underlying ASN1 print routine when
1677 an extension cannot be parsed. Correct a typo in the
1678 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1681 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1682 entries for variables.
1685 *) Add functionality to apps/openssl.c for detecting locking
1686 problems: As the program is single-threaded, all we have
1687 to do is register a locking callback using an array for
1688 storing which locks are currently held by the program.
1691 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
1692 SSL_get_ex_data_X509_STORE_idx(), which is used in
1693 ssl_verify_cert_chain() and thus can be called at any time
1694 during TLS/SSL handshakes so that thread-safety is essential.
1695 Unfortunately, the ex_data design is not at all suited
1696 for multi-threaded use, so it probably should be abolished.
1699 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
1700 [Broadcom, tweaked and integrated by Geoff Thorpe]
1702 *) Move common extension printing code to new function
1703 X509V3_print_extensions(). Reorganise OCSP print routines and
1704 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1707 *) New function X509_signature_print() to remove duplication in some
1711 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
1712 set (this was treated exactly the same as SET OF previously). This
1713 is used to reorder the STACK representing the structure to match the
1714 encoding. This will be used to get round a problem where a PKCS7
1715 structure which was signed could not be verified because the STACK
1716 order did not reflect the encoded order.
1719 *) Reimplement the OCSP ASN1 module using the new code.
1722 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1723 for its ASN1 operations. The old style function pointers still exist
1724 for now but they will eventually go away.
1727 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
1728 completely replaces the old ASN1 functionality with a table driven
1729 encoder and decoder which interprets an ASN1_ITEM structure describing
1730 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1731 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1732 has also been converted to the new form.
1735 *) Change BN_mod_exp_recp so that negative moduli are tolerated
1736 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1737 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1738 for negative moduli.
1741 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
1742 of not touching the result's sign bit.
1745 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1749 *) Changed the LHASH code to use prototypes for callbacks, and created
1750 macros to declare and implement thin (optionally static) functions
1751 that provide type-safety and avoid function pointer casting for the
1752 type-specific callbacks.
1755 *) Added Kerberos Cipher Suites to be used with TLS, as written in
1757 [Veers Staats <staatsvr@asc.hpc.mil>,
1758 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1760 *) Reformat the FAQ so the different questions and answers can be divided
1761 in sections depending on the subject.
1764 *) Have the zlib compression code load ZLIB.DLL dynamically under
1768 *) New function BN_mod_sqrt for computing square roots modulo a prime
1769 (using the probabilistic Tonelli-Shanks algorithm unless
1770 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1771 be handled deterministically).
1772 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1774 *) Make BN_mod_inverse faster by explicitly handling small quotients
1775 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1776 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1779 *) New function BN_kronecker.
1782 *) Fix BN_gcd so that it works on negative inputs; the result is
1783 positive unless both parameters are zero.
1784 Previously something reasonably close to an infinite loop was
1785 possible because numbers could be growing instead of shrinking
1786 in the implementation of Euclid's algorithm.
1789 *) Fix BN_is_word() and BN_is_one() macros to take into account the
1790 sign of the number in question.
1792 Fix BN_is_word(a,w) to work correctly for w == 0.
1794 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1795 because its test if the absolute value of 'a' equals 'w'.
1796 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1797 it exists mostly for use in the implementations of BN_is_zero(),
1798 BN_is_one(), and BN_is_word().
1801 *) New function BN_swap.
1804 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1805 the exponentiation functions are more likely to produce reasonable
1806 results on negative inputs.
1809 *) Change BN_mod_mul so that the result is always non-negative.
1810 Previously, it could be negative if one of the factors was negative;
1811 I don't think anyone really wanted that behaviour.
1814 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1815 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1816 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1817 and add new functions:
1826 BN_mod_lshift1_quick
1830 These functions always generate non-negative results.
1832 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
1833 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
1835 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
1836 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
1837 be reduced modulo m.
1838 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1840 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1841 was actually never needed) and in BN_mul(). The removal in BN_mul()
1842 required a small change in bn_mul_part_recursive() and the addition
1843 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1844 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1845 bn_sub_words() and bn_add_words() except they take arrays with
1849 *) In 'openssl passwd', verify passwords read from the terminal
1850 unless the '-salt' option is used (which usually means that
1851 verification would just waste user's time since the resulting
1852 hash is going to be compared with some given password hash)
1853 or the new '-noverify' option is used.
1855 This is an incompatible change, but it does not affect
1856 non-interactive use of 'openssl passwd' (passwords on the command
1857 line, '-stdin' option, '-in ...' option) and thus should not
1861 *) Remove all references to RSAref, since there's no more need for it.
1864 *) Make DSO load along a path given through an environment variable
1865 (SHLIB_PATH) with shl_load().
1868 *) Constify the ENGINE code as a result of BIGNUM constification.
1869 Also constify the RSA code and most things related to it. In a
1870 few places, most notable in the depth of the ASN.1 code, ugly
1871 casts back to non-const were required (to be solved at a later
1875 *) Make it so the openssl application has all engines loaded by default.
1878 *) Constify the BIGNUM routines a little more.
1881 *) Add the following functions:
1883 ENGINE_load_cswift()
1885 ENGINE_load_atalla()
1887 ENGINE_load_builtin_engines()
1889 That way, an application can itself choose if external engines that
1890 are built-in in OpenSSL shall ever be used or not. The benefit is
1891 that applications won't have to be linked with libdl or other dso
1892 libraries unless it's really needed.
1894 Changed 'openssl engine' to load all engines on demand.
1895 Changed the engine header files to avoid the duplication of some
1896 declarations (they differed!).
1899 *) 'openssl engine' can now list capabilities.
1902 *) Better error reporting in 'openssl engine'.
1905 *) Never call load_dh_param(NULL) in s_server.
1908 *) Add engine application. It can currently list engines by name and
1909 identity, and test if they are actually available.
1912 *) Improve RPM specification file by forcing symbolic linking and making
1913 sure the installed documentation is also owned by root.root.
1914 [Damien Miller <djm@mindrot.org>]
1916 *) Give the OpenSSL applications more possibilities to make use of
1917 keys (public as well as private) handled by engines.
1920 *) Add OCSP code that comes from CertCo.
1923 *) Add VMS support for the Rijndael code.
1926 *) Added untested support for Nuron crypto accelerator.
1929 *) Add support for external cryptographic devices. This code was
1930 previously distributed separately as the "engine" branch.
1931 [Geoff Thorpe, Richard Levitte]
1933 *) Rework the filename-translation in the DSO code. It is now possible to
1934 have far greater control over how a "name" is turned into a filename
1935 depending on the operating environment and any oddities about the
1936 different shared library filenames on each system.
1939 *) Support threads on FreeBSD-elf in Configure.
1942 *) Fix for SHA1 assembly problem with MASM: it produces
1943 warnings about corrupt line number information when assembling
1944 with debugging information. This is caused by the overlapping
1946 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
1949 NCONF_get_number() has no error checking at all. As a replacement,
1950 NCONF_get_number_e() is defined (_e for "error checking") and is
1951 promoted strongly. The old NCONF_get_number is kept around for
1952 binary backward compatibility.
1953 Make it possible for methods to load from something other than a BIO,
1954 by providing a function pointer that is given a name instead of a BIO.
1955 For example, this could be used to load configuration data from an
1959 *) Fix for non blocking accept BIOs. Added new I/O special reason
1960 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
1961 with non blocking I/O was not possible because no retry code was
1962 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
1966 *) Added the beginnings of Rijndael support.
1969 *) Fix for bug in DirectoryString mask setting. Add support for
1970 X509_NAME_print_ex() in 'req' and X509_print_ex() function
1971 to allow certificate printing to more controllable, additional
1972 'certopt' option to 'x509' to allow new printing options to be
1976 *) Clean old EAY MD5 hack from e_os.h.
1979 Changes between 0.9.6g and 0.9.6h [xx XXX xxxx]
1981 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
1982 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
1985 *) Fix initialization code race conditions in
1986 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
1987 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
1988 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
1989 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
1990 ssl2_get_cipher_by_char(),
1991 ssl3_get_cipher_by_char().
1992 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
1994 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
1995 the cached sessions are flushed, as the remove_cb() might use ex_data
1996 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
1997 (see [openssl.org #212]).
1998 [Geoff Thorpe, Lutz Jaenicke]
2000 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
2001 length, instead of the encoding length to d2i_ASN1_OBJECT.
2004 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
2006 *) [In 0.9.6g-engine release:]
2007 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
2008 [Lynn Gazis <lgazis@rainbow.com>]
2010 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
2012 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
2013 and get fix the header length calculation.
2014 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
2015 Alon Kantor <alonk@checkpoint.com> (and others),
2018 *) Use proper error handling instead of 'assertions' in buffer
2019 overflow checks added in 0.9.6e. This prevents DoS (the
2020 assertions could call abort()).
2021 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
2023 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
2025 *) Add various sanity checks to asn1_get_length() to reject
2026 the ASN1 length bytes if they exceed sizeof(long), will appear
2027 negative or the content length exceeds the length of the
2029 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2031 *) Fix cipher selection routines: ciphers without encryption had no flags
2032 for the cipher strength set and where therefore not handled correctly
2033 by the selection routines (PR #130).
2036 *) Fix EVP_dsa_sha macro.
2040 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
2041 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
2042 that was added in OpenSSL 0.9.6d.
2044 As the countermeasure turned out to be incompatible with some
2045 broken SSL implementations, the new option is part of SSL_OP_ALL.
2046 SSL_OP_ALL is usually employed when compatibility with weird SSL
2047 implementations is desired (e.g. '-bugs' option to 's_client' and
2048 's_server'), so the new option is automatically set in many
2052 *) Changes in security patch:
2054 Changes marked "(CHATS)" were sponsored by the Defense Advanced
2055 Research Projects Agency (DARPA) and Air Force Research Laboratory,
2056 Air Force Materiel Command, USAF, under agreement number
2059 *) Add various sanity checks to asn1_get_length() to reject
2060 the ASN1 length bytes if they exceed sizeof(long), will appear
2061 negative or the content length exceeds the length of the
2062 supplied buffer. (CAN-2002-0659)
2063 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2065 *) Assertions for various potential buffer overflows, not known to
2067 [Ben Laurie (CHATS)]
2069 *) Various temporary buffers to hold ASCII versions of integers were
2070 too small for 64 bit platforms. (CAN-2002-0655)
2071 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
2073 *) Remote buffer overflow in SSL3 protocol - an attacker could
2074 supply an oversized session ID to a client. (CAN-2002-0656)
2075 [Ben Laurie (CHATS)]
2077 *) Remote buffer overflow in SSL2 protocol - an attacker could
2078 supply an oversized client master key. (CAN-2002-0656)
2079 [Ben Laurie (CHATS)]
2081 Changes between 0.9.6c and 0.9.6d [9 May 2002]
2083 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
2084 encoded as NULL) with id-dsa-with-sha1.
2085 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
2087 *) Check various X509_...() return values in apps/req.c.
2088 [Nils Larsch <nla@trustcenter.de>]
2090 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
2091 an end-of-file condition would erronously be flagged, when the CRLF
2092 was just at the end of a processed block. The bug was discovered when
2093 processing data through a buffering memory BIO handing the data to a
2094 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
2095 <ptsekov@syntrex.com> and Nedelcho Stanev.
2098 *) Implement a countermeasure against a vulnerability recently found
2099 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
2100 before application data chunks to avoid the use of known IVs
2101 with data potentially chosen by the attacker.
2104 *) Fix length checks in ssl3_get_client_hello().
2107 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
2108 to prevent ssl3_read_internal() from incorrectly assuming that
2109 ssl3_read_bytes() found application data while handshake
2110 processing was enabled when in fact s->s3->in_read_app_data was
2111 merely automatically cleared during the initial handshake.
2112 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
2114 *) Fix object definitions for Private and Enterprise: they were not
2115 recognized in their shortname (=lowercase) representation. Extend
2116 obj_dat.pl to issue an error when using undefined keywords instead
2117 of silently ignoring the problem (Svenning Sorensen
2118 <sss@sss.dnsalias.net>).
2121 *) Fix DH_generate_parameters() so that it works for 'non-standard'
2122 generators, i.e. generators other than 2 and 5. (Previously, the
2123 code did not properly initialise the 'add' and 'rem' values to
2124 BN_generate_prime().)
2126 In the new general case, we do not insist that 'generator' is
2127 actually a primitive root: This requirement is rather pointless;
2128 a generator of the order-q subgroup is just as good, if not
2132 *) Map new X509 verification errors to alerts. Discovered and submitted by
2133 Tom Wu <tom@arcot.com>.
2136 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
2137 returning non-zero before the data has been completely received
2138 when using non-blocking I/O.
2139 [Bodo Moeller; problem pointed out by John Hughes]
2141 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
2142 [Ben Laurie, Lutz Jaenicke]
2144 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
2145 Yoram Zahavi <YoramZ@gilian.com>).
2148 *) Add information about CygWin 1.3 and on, and preserve proper
2149 configuration for the versions before that.
2150 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2152 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
2153 check whether we deal with a copy of a session and do not delete from
2154 the cache in this case. Problem reported by "Izhar Shoshani Levi"
2155 <izhar@checkpoint.com>.
2158 *) Do not store session data into the internal session cache, if it
2159 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
2160 flag is set). Proposed by Aslam <aslam@funk.com>.
2163 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
2167 *) [In 0.9.6d-engine release:]
2168 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2169 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2171 *) Add the configuration target linux-s390x.
2172 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
2174 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
2175 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
2176 variable as an indication that a ClientHello message has been
2177 received. As the flag value will be lost between multiple
2178 invocations of ssl3_accept when using non-blocking I/O, the
2179 function may not be aware that a handshake has actually taken
2180 place, thus preventing a new session from being added to the
2183 To avoid this problem, we now set s->new_session to 2 instead of
2184 using a local variable.
2185 [Lutz Jaenicke, Bodo Moeller]
2187 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
2188 if the SSL_R_LENGTH_MISMATCH error is detected.
2189 [Geoff Thorpe, Bodo Moeller]
2191 *) New 'shared_ldflag' column in Configure platform table.
2194 *) Fix EVP_CIPHER_mode macro.
2195 ["Dan S. Camper" <dan@bti.net>]
2197 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
2198 type, we must throw them away by setting rr->length to 0.
2199 [D P Chang <dpc@qualys.com>]
2201 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
2203 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
2204 <Dominikus.Scherkl@biodata.com>. (The previous implementation
2205 worked incorrectly for those cases where range = 10..._2 and
2206 3*range is two bits longer than range.)
2209 *) Only add signing time to PKCS7 structures if it is not already
2213 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
2214 OBJ_ld_ce should be OBJ_id_ce.
2215 Also some ip-pda OIDs in crypto/objects/objects.txt were
2216 incorrect (cf. RFC 3039).
2217 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
2219 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
2220 returns early because it has nothing to do.
2221 [Andy Schneider <andy.schneider@bjss.co.uk>]
2223 *) [In 0.9.6c-engine release:]
2224 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
2225 [Andy Schneider <andy.schneider@bjss.co.uk>]
2227 *) [In 0.9.6c-engine release:]
2228 Add support for Cryptographic Appliance's keyserver technology.
2229 (Use engine 'keyclient')
2230 [Cryptographic Appliances and Geoff Thorpe]
2232 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
2233 is called via tools/c89.sh because arguments have to be
2234 rearranged (all '-L' options must appear before the first object
2236 [Richard Shapiro <rshapiro@abinitio.com>]
2238 *) [In 0.9.6c-engine release:]
2239 Add support for Broadcom crypto accelerator cards, backported
2241 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
2243 *) [In 0.9.6c-engine release:]
2244 Add support for SureWare crypto accelerator cards from
2245 Baltimore Technologies. (Use engine 'sureware')
2246 [Baltimore Technologies and Mark Cox]
2248 *) [In 0.9.6c-engine release:]
2249 Add support for crypto accelerator cards from Accelerated
2250 Encryption Processing, www.aep.ie. (Use engine 'aep')
2251 [AEP Inc. and Mark Cox]
2253 *) Add a configuration entry for gcc on UnixWare.
2254 [Gary Benson <gbenson@redhat.com>]
2256 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
2257 messages are stored in a single piece (fixed-length part and
2258 variable-length part combined) and fix various bugs found on the way.
2261 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
2262 instead. BIO_gethostbyname() does not know what timeouts are
2263 appropriate, so entries would stay in cache even when they have
2265 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
2267 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
2268 faced with a pathologically small ClientHello fragment that does
2269 not contain client_version: Instead of aborting with an error,
2270 simply choose the highest available protocol version (i.e.,
2271 TLS 1.0 unless it is disabled). In practice, ClientHello
2272 messages are never sent like this, but this change gives us
2273 strictly correct behaviour at least for TLS.
2276 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
2277 never resets s->method to s->ctx->method when called from within
2278 one of the SSL handshake functions.
2279 [Bodo Moeller; problem pointed out by Niko Baric]
2281 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
2282 (sent using the client's version number) if client_version is
2283 smaller than the protocol version in use. Also change
2284 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
2285 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
2286 the client will at least see that alert.
2289 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
2293 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
2294 client receives HelloRequest while in a handshake.
2295 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
2297 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
2298 should end in 'break', not 'goto end' which circuments various
2299 cleanups done in state SSL_ST_OK. But session related stuff
2300 must be disabled for SSL_ST_OK in the case that we just sent a
2303 Also avoid some overhead by not calling ssl_init_wbio_buffer()
2304 before just sending a HelloRequest.
2305 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
2307 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
2308 reveal whether illegal block cipher padding was found or a MAC
2309 verification error occured. (Neither SSLerr() codes nor alerts
2310 are directly visible to potential attackers, but the information
2311 may leak via logfiles.)
2313 Similar changes are not required for the SSL 2.0 implementation
2314 because the number of padding bytes is sent in clear for SSL 2.0,
2315 and the extra bytes are just ignored. However ssl/s2_pkt.c
2316 failed to verify that the purported number of padding bytes is in
2320 *) Add OpenUNIX-8 support including shared libraries
2321 (Boyd Lynn Gerber <gerberb@zenez.com>).
2324 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
2325 'wristwatch attack' using huge encoding parameters (cf.
2326 James H. Manger's CRYPTO 2001 paper). Note that the
2327 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
2328 encoding parameters and hence was not vulnerable.
2331 *) BN_sqr() bug fix.
2332 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
2334 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
2335 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
2336 followed by modular reduction.
2337 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
2339 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
2340 equivalent based on BN_pseudo_rand() instead of BN_rand().
2343 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
2344 This function was broken, as the check for a new client hello message
2345 to handle SGC did not allow these large messages.
2346 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
2349 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
2352 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
2353 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
2356 *) Rework the configuration and shared library support for Tru64 Unix.
2357 The configuration part makes use of modern compiler features and
2358 still retains old compiler behavior for those that run older versions
2359 of the OS. The shared library support part includes a variant that
2360 uses the RPATH feature, and is available through the special
2361 configuration target "alpha-cc-rpath", which will never be selected
2363 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
2365 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
2366 with the same message size as in ssl3_get_certificate_request().
2367 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
2368 messages might inadvertently be reject as too long.
2369 [Petr Lampa <lampa@fee.vutbr.cz>]
2371 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
2374 *) Modified SSL library such that the verify_callback that has been set
2375 specificly for an SSL object with SSL_set_verify() is actually being
2376 used. Before the change, a verify_callback set with this function was
2377 ignored and the verify_callback() set in the SSL_CTX at the time of
2378 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
2379 to allow the necessary settings.
2382 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
2383 explicitly to NULL, as at least on Solaris 8 this seems not always to be
2384 done automatically (in contradiction to the requirements of the C
2385 standard). This made problems when used from OpenSSH.
2388 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
2389 dh->length and always used
2391 BN_rand_range(priv_key, dh->p).
2393 BN_rand_range() is not necessary for Diffie-Hellman, and this
2394 specific range makes Diffie-Hellman unnecessarily inefficient if
2395 dh->length (recommended exponent length) is much smaller than the
2396 length of dh->p. We could use BN_rand_range() if the order of
2397 the subgroup was stored in the DH structure, but we only have
2402 BN_rand(priv_key, l, ...)
2404 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
2410 RSA_eay_public_encrypt
2411 RSA_eay_private_decrypt
2412 RSA_eay_private_encrypt (signing)
2413 RSA_eay_public_decrypt (signature verification)
2415 (default implementations for RSA_public_encrypt,
2416 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
2417 always reject numbers >= n.
2420 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
2421 to synchronize access to 'locking_thread'. This is necessary on
2422 systems where access to 'locking_thread' (an 'unsigned long'
2423 variable) is not atomic.
2426 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
2427 *before* setting the 'crypto_lock_rand' flag. The previous code had
2428 a race condition if 0 is a valid thread ID.
2429 [Travis Vitek <vitek@roguewave.com>]
2431 *) Add support for shared libraries under Irix.
2432 [Albert Chin-A-Young <china@thewrittenword.com>]
2434 *) Add configuration option to build on Linux on both big-endian and
2436 [Ralf Baechle <ralf@uni-koblenz.de>]
2438 *) Add the possibility to create shared libraries on HP-UX.
2441 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
2443 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
2444 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
2445 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
2446 PRNG state recovery was possible based on the output of
2447 one PRNG request appropriately sized to gain knowledge on
2448 'md' followed by enough consecutive 1-byte PRNG requests
2449 to traverse all of 'state'.
2451 1. When updating 'md_local' (the current thread's copy of 'md')
2452 during PRNG output generation, hash all of the previous
2453 'md_local' value, not just the half used for PRNG output.
2455 2. Make the number of bytes from 'state' included into the hash
2456 independent from the number of PRNG bytes requested.
2458 The first measure alone would be sufficient to avoid
2459 Markku-Juhani's attack. (Actually it had never occurred
2460 to me that the half of 'md_local' used for chaining was the
2461 half from which PRNG output bytes were taken -- I had always
2462 assumed that the secret half would be used.) The second
2463 measure makes sure that additional data from 'state' is never
2464 mixed into 'md_local' in small portions; this heuristically
2465 further strengthens the PRNG.
2468 *) Fix crypto/bn/asm/mips3.s.
2471 *) When only the key is given to "enc", the IV is undefined. Print out
2472 an error message in this case.
2475 *) Handle special case when X509_NAME is empty in X509 printing routines.
2478 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
2479 positive and less than q.
2482 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
2483 used: it isn't thread safe and the add_lock_callback should handle
2485 [Paul Rose <Paul.Rose@bridge.com>]
2487 *) Verify that incoming data obeys the block size in
2488 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
2492 [Ulf Möller, Bodo Möller]
2494 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
2495 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
2496 when fixing the server behaviour for backwards-compatible 'client
2497 hello' messages. (Note that the attack is impractical against
2498 SSL 3.0 and TLS 1.0 anyway because length and version checking
2499 means that the probability of guessing a valid ciphertext is
2500 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
2503 Before 0.9.5, the countermeasure (hide the error by generating a
2504 random 'decryption result') did not work properly because
2505 ERR_clear_error() was missing, meaning that SSL_get_error() would
2506 detect the supposedly ignored error.
2508 Both problems are now fixed.
2511 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
2512 (previously it was 1024).
2515 *) Fix for compatibility mode trust settings: ignore trust settings
2516 unless some valid trust or reject settings are present.
2519 *) Fix for blowfish EVP: its a variable length cipher.
2522 *) Fix various bugs related to DSA S/MIME verification. Handle missing
2523 parameters in DSA public key structures and return an error in the
2524 DSA routines if parameters are absent.
2527 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
2528 in the current directory if neither $RANDFILE nor $HOME was set.
2529 RAND_file_name() in 0.9.6a returned NULL in this case. This has
2530 caused some confusion to Windows users who haven't defined $HOME.
2531 Thus RAND_file_name() is changed again: e_os.h can define a
2532 DEFAULT_HOME, which will be used if $HOME is not set.
2533 For Windows, we use "C:"; on other platforms, we still require
2534 environment variables.
2536 *) Move 'if (!initialized) RAND_poll()' into regions protected by
2537 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
2538 having multiple threads call RAND_poll() concurrently.
2541 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
2542 combination of a flag and a thread ID variable.
2543 Otherwise while one thread is in ssleay_rand_bytes (which sets the
2544 flag), *other* threads can enter ssleay_add_bytes without obeying
2545 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
2546 that they do not hold after the first thread unsets add_do_not_lock).
2549 *) Change bctest again: '-x' expressions are not available in all
2553 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
2555 *) Fix a couple of memory leaks in PKCS7_dataDecode()
2556 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
2558 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
2559 the default extension for executables, if any. Also, make the perl
2560 scripts that use symlink() to test if it really exists and use "cp"
2561 if it doesn't. All this made OpenSSL compilable and installable in
2565 *) Fix for asn1_GetSequence() for indefinite length constructed data.
2566 If SEQUENCE is length is indefinite just set c->slen to the total
2567 amount of data available.
2568 [Steve Henson, reported by shige@FreeBSD.org]
2569 [This change does not apply to 0.9.7.]
2571 *) Change bctest to avoid here-documents inside command substitution
2572 (workaround for FreeBSD /bin/sh bug).
2573 For compatibility with Ultrix, avoid shell functions (introduced
2574 in the bctest version that searches along $PATH).
2577 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
2578 with des_encrypt() defined on some operating systems, like Solaris
2582 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
2583 On the Importance of Eliminating Errors in Cryptographic
2584 Computations, J. Cryptology 14 (2001) 2, 101-119,
2585 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
2588 *) MIPS assembler BIGNUM division bug fix.
2591 *) Disabled incorrect Alpha assembler code.
2594 *) Fix PKCS#7 decode routines so they correctly update the length
2595 after reading an EOC for the EXPLICIT tag.
2597 [This change does not apply to 0.9.7.]
2599 *) Fix bug in PKCS#12 key generation routines. This was triggered
2600 if a 3DES key was generated with a 0 initial byte. Include
2601 PKCS12_BROKEN_KEYGEN compilation option to retain the old
2602 (but broken) behaviour.
2605 *) Enhance bctest to search for a working bc along $PATH and print
2607 [Tim Rice <tim@multitalents.net> via Richard Levitte]
2609 *) Fix memory leaks in err.c: free err_data string if necessary;
2610 don't write to the wrong index in ERR_set_error_data.
2613 *) Implement ssl23_peek (analogous to ssl23_read), which previously
2617 *) Replace rdtsc with _emit statements for VC++ version 5.
2618 [Jeremy Cooper <jeremy@baymoo.org>]
2620 *) Make it possible to reuse SSLv2 sessions.
2623 *) In copy_email() check for >= 0 as a return value for
2624 X509_NAME_get_index_by_NID() since 0 is a valid index.
2625 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
2627 *) Avoid coredump with unsupported or invalid public keys by checking if
2628 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
2629 PKCS7_verify() fails with non detached data.
2632 *) Don't use getenv in library functions when run as setuid/setgid.
2633 New function OPENSSL_issetugid().
2636 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
2637 due to incorrect handling of multi-threading:
2639 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
2641 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
2643 3. Count how many times MemCheck_off() has been called so that
2644 nested use can be treated correctly. This also avoids
2645 inband-signalling in the previous code (which relied on the
2646 assumption that thread ID 0 is impossible).
2649 *) Add "-rand" option also to s_client and s_server.
2652 *) Fix CPU detection on Irix 6.x.
2653 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
2654 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
2656 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
2659 [This change does not apply to 0.9.7.]
2661 *) Use the cached encoding of an X509_NAME structure rather than
2662 copying it. This is apparently the reason for the libsafe "errors"
2663 but the code is actually correct.
2666 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
2667 Bleichenbacher's DSA attack.
2668 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
2669 to be set and top=0 forces the highest bit to be set; top=-1 is new
2670 and leaves the highest bit random.
2671 [Ulf Moeller, Bodo Moeller]
2673 *) In the NCONF_...-based implementations for CONF_... queries
2674 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
2675 a temporary CONF structure with the data component set to NULL
2676 (which gives segmentation faults in lh_retrieve).
2677 Instead, use NULL for the CONF pointer in CONF_get_string and
2678 CONF_get_number (which may use environment variables) and directly
2679 return NULL from CONF_get_section.
2682 *) Fix potential buffer overrun for EBCDIC.
2685 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
2686 keyUsage if basicConstraints absent for a CA.
2689 *) Make SMIME_write_PKCS7() write mail header values with a format that
2690 is more generally accepted (no spaces before the semicolon), since
2691 some programs can't parse those values properly otherwise. Also make
2692 sure BIO's that break lines after each write do not create invalid
2696 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
2697 macros previously used would not encode an empty SEQUENCE OF
2698 and break the signature.
2700 [This change does not apply to 0.9.7.]
2702 *) Zero the premaster secret after deriving the master secret in
2706 *) Add some EVP_add_digest_alias registrations (as found in
2707 OpenSSL_add_all_digests()) to SSL_library_init()
2708 aka OpenSSL_add_ssl_algorithms(). This provides improved
2709 compatibility with peers using X.509 certificates
2710 with unconventional AlgorithmIdentifier OIDs.
2713 *) Fix for Irix with NO_ASM.
2714 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
2716 *) ./config script fixes.
2717 [Ulf Moeller, Richard Levitte]
2719 *) Fix 'openssl passwd -1'.
2722 *) Change PKCS12_key_gen_asc() so it can cope with non null
2723 terminated strings whose length is passed in the passlen
2724 parameter, for example from PEM callbacks. This was done
2725 by adding an extra length parameter to asc2uni().
2726 [Steve Henson, reported by <oddissey@samsung.co.kr>]
2728 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
2729 call failed, free the DSA structure.
2732 *) Fix to uni2asc() to cope with zero length Unicode strings.
2733 These are present in some PKCS#12 files.
2736 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
2737 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
2738 when writing a 32767 byte record.
2739 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
2741 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
2742 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
2744 (RSA objects have a reference count access to which is protected
2745 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
2746 so they are meant to be shared between threads.)
2747 [Bodo Moeller, Geoff Thorpe; original patch submitted by
2748 "Reddie, Steven" <Steven.Reddie@ca.com>]
2750 *) Fix a deadlock in CRYPTO_mem_leaks().
2753 *) Use better test patterns in bntest.
2756 *) rand_win.c fix for Borland C.
2759 *) BN_rshift bugfix for n == 0.
2762 *) Add a 'bctest' script that checks for some known 'bc' bugs
2763 so that 'make test' does not abort just because 'bc' is broken.
2766 *) Store verify_result within SSL_SESSION also for client side to
2767 avoid potential security hole. (Re-used sessions on the client side
2768 always resulted in verify_result==X509_V_OK, not using the original
2769 result of the server certificate verification.)
2772 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
2773 SSL3_RT_APPLICATION_DATA, return 0.
2774 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
2778 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
2779 releases, have been re-implemented by renaming the previous
2780 implementations of ssl2_read and ssl3_read to ssl2_read_internal
2781 and ssl3_read_internal, respectively, and adding 'peek' parameters
2782 to them. The new ssl[23]_{read,peek} functions are calls to
2783 ssl[23]_read_internal with the 'peek' flag set appropriately.
2784 A 'peek' parameter has also been added to ssl3_read_bytes, which
2785 does the actual work for ssl3_read_internal.
2788 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
2789 the method-specific "init()" handler. Also clean up ex_data after
2790 calling the method-specific "finish()" handler. Previously, this was
2791 happening the other way round.
2794 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
2795 The previous value, 12, was not always sufficient for BN_mod_exp().
2798 *) Make sure that shared libraries get the internal name engine with
2799 the full version number and not just 0. This should mark the
2800 shared libraries as not backward compatible. Of course, this should
2801 be changed again when we can guarantee backward binary compatibility.
2804 *) Fix typo in get_cert_by_subject() in by_dir.c
2805 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
2807 *) Rework the system to generate shared libraries:
2809 - Make note of the expected extension for the shared libraries and
2810 if there is a need for symbolic links from for example libcrypto.so.0
2811 to libcrypto.so.0.9.7. There is extended info in Configure for
2814 - Make as few rebuilds of the shared libraries as possible.
2816 - Still avoid linking the OpenSSL programs with the shared libraries.
2818 - When installing, install the shared libraries separately from the
2822 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
2824 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
2825 and not in SSL_clear because the latter is also used by the
2826 accept/connect functions; previously, the settings made by
2827 SSL_set_read_ahead would be lost during the handshake.
2828 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
2830 *) Correct util/mkdef.pl to be selective about disabled algorithms.
2831 Previously, it would create entries for disableed algorithms no
2835 *) Added several new manual pages for SSL_* function.
2838 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
2840 *) In ssl23_get_client_hello, generate an error message when faced
2841 with an initial SSL 3.0/TLS record that is too small to contain the
2842 first two bytes of the ClientHello message, i.e. client_version.
2843 (Note that this is a pathologic case that probably has never happened
2844 in real life.) The previous approach was to use the version number
2845 from the record header as a substitute; but our protocol choice
2846 should not depend on that one because it is not authenticated
2847 by the Finished messages.
2850 *) More robust randomness gathering functions for Windows.
2851 [Jeffrey Altman <jaltman@columbia.edu>]
2853 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
2854 not set then we don't setup the error code for issuer check errors
2855 to avoid possibly overwriting other errors which the callback does
2856 handle. If an application does set the flag then we assume it knows
2857 what it is doing and can handle the new informational codes
2861 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
2862 a general "ANY" type, as such it should be able to decode anything
2863 including tagged types. However it didn't check the class so it would
2864 wrongly interpret tagged types in the same way as their universal
2865 counterpart and unknown types were just rejected. Changed so that the
2866 tagged and unknown types are handled in the same way as a SEQUENCE:
2867 that is the encoding is stored intact. There is also a new type
2868 "V_ASN1_OTHER" which is used when the class is not universal, in this
2869 case we have no idea what the actual type is so we just lump them all
2873 *) On VMS, stdout may very well lead to a file that is written to
2874 in a record-oriented fashion. That means that every write() will
2875 write a separate record, which will be read separately by the
2876 programs trying to read from it. This can be very confusing.
2878 The solution is to put a BIO filter in the way that will buffer
2879 text until a linefeed is reached, and then write everything a
2880 line at a time, so every record written will be an actual line,
2881 not chunks of lines and not (usually doesn't happen, but I've
2882 seen it once) several lines in one record. BIO_f_linebuffer() is
2885 Currently, it's a VMS-only method, because that's where it has
2886 been tested well enough.
2889 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
2890 it can return incorrect results.
2891 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
2892 but it was in 0.9.6-beta[12].)
2895 *) Disable the check for content being present when verifying detached
2896 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
2897 include zero length content when signing messages.
2900 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
2901 BIO_ctrl (for BIO pairs).
2904 *) Add DSO method for VMS.
2907 *) Bug fix: Montgomery multiplication could produce results with the
2911 *) Add RPM specification openssl.spec and modify it to build three
2912 packages. The default package contains applications, application
2913 documentation and run-time libraries. The devel package contains
2914 include files, static libraries and function documentation. The
2915 doc package contains the contents of the doc directory. The original
2916 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
2919 *) Add a large number of documentation files for many SSL routines.
2920 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2922 *) Add a configuration entry for Sony News 4.
2923 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
2925 *) Don't set the two most significant bits to one when generating a
2926 random number < q in the DSA library.
2929 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
2930 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
2931 the underlying transport is blocking) if a handshake took place.
2932 (The default behaviour is needed by applications such as s_client
2933 and s_server that use select() to determine when to use SSL_read;
2934 but for applications that know in advance when to expect data, it
2935 just makes things more complicated.)
2938 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
2942 *) Add a few more EBCDIC conditionals that make `req' and `x509'
2943 work better on such systems.
2944 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2946 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
2947 Update PKCS12_parse() so it copies the friendlyName and the
2948 keyid to the certificates aux info.
2951 *) Fix bug in PKCS7_verify() which caused an infinite loop
2952 if there was more than one signature.
2953 [Sven Uszpelkat <su@celocom.de>]
2955 *) Major change in util/mkdef.pl to include extra information
2956 about each symbol, as well as presentig variables as well
2957 as functions. This change means that there's n more need
2958 to rebuild the .num files when some algorithms are excluded.
2961 *) Allow the verify time to be set by an application,
2962 rather than always using the current time.
2965 *) Phase 2 verify code reorganisation. The certificate
2966 verify code now looks up an issuer certificate by a
2967 number of criteria: subject name, authority key id
2968 and key usage. It also verifies self signed certificates
2969 by the same criteria. The main comparison function is
2970 X509_check_issued() which performs these checks.
2972 Lot of changes were necessary in order to support this
2973 without completely rewriting the lookup code.
2975 Authority and subject key identifier are now cached.
2977 The LHASH 'certs' is X509_STORE has now been replaced
2978 by a STACK_OF(X509_OBJECT). This is mainly because an
2979 LHASH can't store or retrieve multiple objects with
2980 the same hash value.
2982 As a result various functions (which were all internal
2983 use only) have changed to handle the new X509_STORE
2984 structure. This will break anything that messed round
2985 with X509_STORE internally.
2987 The functions X509_STORE_add_cert() now checks for an
2988 exact match, rather than just subject name.
2990 The X509_STORE API doesn't directly support the retrieval
2991 of multiple certificates matching a given criteria, however
2992 this can be worked round by performing a lookup first
2993 (which will fill the cache with candidate certificates)
2994 and then examining the cache for matches. This is probably
2995 the best we can do without throwing out X509_LOOKUP
2996 entirely (maybe later...).
2998 The X509_VERIFY_CTX structure has been enhanced considerably.
3000 All certificate lookup operations now go via a get_issuer()
3001 callback. Although this currently uses an X509_STORE it
3002 can be replaced by custom lookups. This is a simple way
3003 to bypass the X509_STORE hackery necessary to make this
3004 work and makes it possible to use more efficient techniques
3005 in future. A very simple version which uses a simple
3006 STACK for its trusted certificate store is also provided
3007 using X509_STORE_CTX_trusted_stack().
3009 The verify_cb() and verify() callbacks now have equivalents
3010 in the X509_STORE_CTX structure.
3012 X509_STORE_CTX also has a 'flags' field which can be used
3013 to customise the verify behaviour.
3016 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
3017 excludes S/MIME capabilities.
3020 *) When a certificate request is read in keep a copy of the
3021 original encoding of the signed data and use it when outputing
3022 again. Signatures then use the original encoding rather than
3023 a decoded, encoded version which may cause problems if the
3024 request is improperly encoded.
3027 *) For consistency with other BIO_puts implementations, call
3028 buffer_write(b, ...) directly in buffer_puts instead of calling
3031 In BIO_puts, increment b->num_write as in BIO_write.
3032 [Peter.Sylvester@EdelWeb.fr]
3034 *) Fix BN_mul_word for the case where the word is 0. (We have to use
3035 BN_zero, we may not return a BIGNUM with an array consisting of
3039 *) Avoid calling abort() from within the library when problems are
3040 detected, except if preprocessor symbols have been defined
3041 (such as REF_CHECK, BN_DEBUG etc.).
3044 *) New openssl application 'rsautl'. This utility can be
3045 used for low level RSA operations. DER public key
3046 BIO/fp routines also added.
3049 *) New Configure entry and patches for compiling on QNX 4.
3050 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
3052 *) A demo state-machine implementation was sponsored by
3053 Nuron (http://www.nuron.com/) and is now available in
3054 demos/state_machine.
3057 *) New options added to the 'dgst' utility for signature
3058 generation and verification.
3061 *) Unrecognized PKCS#7 content types are now handled via a
3062 catch all ASN1_TYPE structure. This allows unsupported
3063 types to be stored as a "blob" and an application can
3064 encode and decode it manually.
3067 *) Fix various signed/unsigned issues to make a_strex.c
3069 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
3071 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
3072 length if passed a buffer. ASN1_INTEGER_to_BN failed
3073 if passed a NULL BN and its argument was negative.
3074 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
3076 *) Modification to PKCS#7 encoding routines to output definite
3077 length encoding. Since currently the whole structures are in
3078 memory there's not real point in using indefinite length
3079 constructed encoding. However if OpenSSL is compiled with
3080 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
3083 *) Added BIO_vprintf() and BIO_vsnprintf().
3086 *) Added more prefixes to parse for in the the strings written
3087 through a logging bio, to cover all the levels that are available
3088 through syslog. The prefixes are now:
3090 PANIC, EMERG, EMR => LOG_EMERG
3091 ALERT, ALR => LOG_ALERT
3092 CRIT, CRI => LOG_CRIT
3093 ERROR, ERR => LOG_ERR
3094 WARNING, WARN, WAR => LOG_WARNING
3095 NOTICE, NOTE, NOT => LOG_NOTICE
3096 INFO, INF => LOG_INFO
3097 DEBUG, DBG => LOG_DEBUG
3099 and as before, if none of those prefixes are present at the
3100 beginning of the string, LOG_ERR is chosen.
3102 On Win32, the LOG_* levels are mapped according to this:
3104 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
3105 LOG_WARNING => EVENTLOG_WARNING_TYPE
3106 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
3110 *) Made it possible to reconfigure with just the configuration
3111 argument "reconf" or "reconfigure". The command line arguments
3112 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
3113 and are retrieved from there when reconfiguring.
3117 [Assar Westerlund <assar@sics.se>, Richard Levitte]
3119 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
3122 *) The obj_dat.pl script was messing up the sorting of object
3123 names. The reason was that it compared the quoted version
3124 of strings as a result "OCSP" > "OCSP Signing" because
3125 " > SPACE. Changed script to store unquoted versions of
3126 names and add quotes on output. It was also omitting some
3127 names from the lookup table if they were given a default
3128 value (that is if SN is missing it is given the same
3129 value as LN and vice versa), these are now added on the
3130 grounds that if an object has a name we should be able to
3131 look it up. Finally added warning output when duplicate
3132 short or long names are found.
3135 *) Changes needed for Tandem NSK.
3136 [Scott Uroff <scott@xypro.com>]
3138 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
3139 RSA_padding_check_SSLv23(), special padding was never detected
3140 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
3141 version rollback attacks was not effective.
3143 In s23_clnt.c, don't use special rollback-attack detection padding
3144 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
3145 client; similarly, in s23_srvr.c, don't do the rollback check if
3146 SSL 2.0 is the only protocol enabled in the server.
3149 *) Make it possible to get hexdumps of unprintable data with 'openssl
3150 asn1parse'. By implication, the functions ASN1_parse_dump() and
3151 BIO_dump_indent() are added.
3154 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
3155 these print out strings and name structures based on various
3156 flags including RFC2253 support and proper handling of
3157 multibyte characters. Added options to the 'x509' utility
3158 to allow the various flags to be set.
3161 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
3162 Also change the functions X509_cmp_current_time() and
3163 X509_gmtime_adj() work with an ASN1_TIME structure,
3164 this will enable certificates using GeneralizedTime in validity
3165 dates to be checked.
3168 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
3169 negative public key encodings) on by default,
3170 NO_NEG_PUBKEY_BUG can be set to disable it.
3173 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
3174 content octets. An i2c_ASN1_OBJECT is unnecessary because
3175 the encoding can be trivially obtained from the structure.
3178 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
3179 not read locks (CRYPTO_r_[un]lock).
3182 *) A first attempt at creating official support for shared
3183 libraries through configuration. I've kept it so the
3184 default is static libraries only, and the OpenSSL programs
3185 are always statically linked for now, but there are
3186 preparations for dynamic linking in place.
3187 This has been tested on Linux and Tru64.
3190 *) Randomness polling function for Win9x, as described in:
3191 Peter Gutmann, Software Generation of Practically Strong
3195 *) Fix so PRNG is seeded in req if using an already existing
3199 *) New options to smime application. -inform and -outform
3200 allow alternative formats for the S/MIME message including
3201 PEM and DER. The -content option allows the content to be
3202 specified separately. This should allow things like Netscape
3203 form signing output easier to verify.
3206 *) Fix the ASN1 encoding of tags using the 'long form'.
3209 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
3210 STRING types. These convert content octets to and from the
3211 underlying type. The actual tag and length octets are
3212 already assumed to have been read in and checked. These
3213 are needed because all other string types have virtually
3214 identical handling apart from the tag. By having versions
3215 of the ASN1 functions that just operate on content octets
3216 IMPLICIT tagging can be handled properly. It also allows
3217 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
3218 and ASN1_INTEGER are identical apart from the tag.
3221 *) Change the handling of OID objects as follows:
3223 - New object identifiers are inserted in objects.txt, following
3224 the syntax given in objects.README.
3225 - objects.pl is used to process obj_mac.num and create a new
3227 - obj_dat.pl is used to create a new obj_dat.h, using the data in
3230 This is currently kind of a hack, and the perl code in objects.pl
3231 isn't very elegant, but it works as I intended. The simplest way
3232 to check that it worked correctly is to look in obj_dat.h and
3233 check the array nid_objs and make sure the objects haven't moved
3234 around (this is important!). Additions are OK, as well as
3235 consistent name changes.
3238 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
3241 *) Addition of the command line parameter '-rand file' to 'openssl req'.
3242 The given file adds to whatever has already been seeded into the
3243 random pool through the RANDFILE configuration file option or
3244 environment variable, or the default random state file.
3247 *) mkstack.pl now sorts each macro group into lexical order.
3248 Previously the output order depended on the order the files
3249 appeared in the directory, resulting in needless rewriting
3253 *) Patches to make OpenSSL compile under Win32 again. Mostly
3254 work arounds for the VC++ problem that it treats func() as
3255 func(void). Also stripped out the parts of mkdef.pl that
3256 added extra typesafe functions: these no longer exist.
3259 *) Reorganisation of the stack code. The macros are now all
3260 collected in safestack.h . Each macro is defined in terms of
3261 a "stack macro" of the form SKM_<name>(type, a, b). The
3262 DEBUG_SAFESTACK is now handled in terms of function casts,
3263 this has the advantage of retaining type safety without the
3264 use of additional functions. If DEBUG_SAFESTACK is not defined
3265 then the non typesafe macros are used instead. Also modified the
3266 mkstack.pl script to handle the new form. Needs testing to see
3267 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
3268 the default if no major problems. Similar behaviour for ASN1_SET_OF
3269 and PKCS12_STACK_OF.
3272 *) When some versions of IIS use the 'NET' form of private key the
3273 key derivation algorithm is different. Normally MD5(password) is
3274 used as a 128 bit RC4 key. In the modified case
3275 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
3276 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
3277 as the old Netscape_RSA functions except they have an additional
3278 'sgckey' parameter which uses the modified algorithm. Also added
3279 an -sgckey command line option to the rsa utility. Thanks to
3280 Adrian Peck <bertie@ncipher.com> for posting details of the modified
3281 algorithm to openssl-dev.
3284 *) The evp_local.h macros were using 'c.##kname' which resulted in
3285 invalid expansion on some systems (SCO 5.0.5 for example).
3286 Corrected to 'c.kname'.
3287 [Phillip Porch <root@theporch.com>]
3289 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
3290 a STACK of email addresses from a certificate or request, these look
3291 in the subject name and the subject alternative name extensions and
3292 omit any duplicate addresses.
3295 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
3296 This makes DSA verification about 2 % faster.
3299 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
3300 (meaning that now 2^5 values will be precomputed, which is only 4 KB
3301 plus overhead for 1024 bit moduli).
3302 This makes exponentiations about 0.5 % faster for 1024 bit
3303 exponents (as measured by "openssl speed rsa2048").
3306 *) Rename memory handling macros to avoid conflicts with other
3308 Malloc => OPENSSL_malloc
3309 Malloc_locked => OPENSSL_malloc_locked
3310 Realloc => OPENSSL_realloc
3311 Free => OPENSSL_free
3314 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
3315 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
3318 *) CygWin32 support.
3319 [John Jarvie <jjarvie@newsguy.com>]
3321 *) The type-safe stack code has been rejigged. It is now only compiled
3322 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
3323 by default all type-specific stack functions are "#define"d back to
3324 standard stack functions. This results in more streamlined output
3325 but retains the type-safety checking possibilities of the original
3329 *) The STACK code has been cleaned up, and certain type declarations
3330 that didn't make a lot of sense have been brought in line. This has
3331 also involved a cleanup of sorts in safestack.h to more correctly
3332 map type-safe stack functions onto their plain stack counterparts.
3333 This work has also resulted in a variety of "const"ifications of
3334 lots of the code, especially "_cmp" operations which should normally
3335 be prototyped with "const" parameters anyway.
3338 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
3339 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
3340 (The PRNG state consists of two parts, the large pool 'state' and 'md',
3341 where all of 'md' is used each time the PRNG is used, but 'state'
3342 is used only indexed by a cyclic counter. As entropy may not be
3343 well distributed from the beginning, 'md' is important as a
3344 chaining variable. However, the output function chains only half
3345 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
3346 all of 'md', and seeding with STATE_SIZE dummy bytes will result
3347 in all of 'state' being rewritten, with the new values depending
3348 on virtually all of 'md'. This overcomes the 80 bit limitation.)
3351 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
3352 the handshake is continued after ssl_verify_cert_chain();
3353 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
3354 can lead to 'unexplainable' connection aborts later.
3355 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
3357 *) Major EVP API cipher revision.
3358 Add hooks for extra EVP features. This allows various cipher
3359 parameters to be set in the EVP interface. Support added for variable
3360 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
3361 setting of RC2 and RC5 parameters.
3363 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
3366 Remove lots of duplicated code from the EVP library. For example *every*
3367 cipher init() function handles the 'iv' in the same way according to the
3368 cipher mode. They also all do nothing if the 'key' parameter is NULL and
3369 for CFB and OFB modes they zero ctx->num.
3371 New functionality allows removal of S/MIME code RC2 hack.
3373 Most of the routines have the same form and so can be declared in terms
3376 By shifting this to the top level EVP_CipherInit() it can be removed from
3377 all individual ciphers. If the cipher wants to handle IVs or keys
3378 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
3381 Change lots of functions like EVP_EncryptUpdate() to now return a
3382 value: although software versions of the algorithms cannot fail
3383 any installed hardware versions can.
3386 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
3387 this option is set, tolerate broken clients that send the negotiated
3388 protocol version number instead of the requested protocol version
3392 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
3393 i.e. non-zero for export ciphersuites, zero otherwise.
3394 Previous versions had this flag inverted, inconsistent with
3395 rsa_tmp_cb (..._TMP_RSA_CB).
3396 [Bodo Moeller; problem reported by Amit Chopra]
3398 *) Add missing DSA library text string. Work around for some IIS
3399 key files with invalid SEQUENCE encoding.
3402 *) Add a document (doc/standards.txt) that list all kinds of standards
3403 and so on that are implemented in OpenSSL.
3406 *) Enhance c_rehash script. Old version would mishandle certificates
3407 with the same subject name hash and wouldn't handle CRLs at all.
3408 Added -fingerprint option to crl utility, to support new c_rehash
3412 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
3415 *) Fix for SSL server purpose checking. Server checking was
3416 rejecting certificates which had extended key usage present
3417 but no ssl client purpose.
3418 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
3420 *) Make PKCS#12 code work with no password. The PKCS#12 spec
3421 is a little unclear about how a blank password is handled.
3422 Since the password in encoded as a BMPString with terminating
3423 double NULL a zero length password would end up as just the
3424 double NULL. However no password at all is different and is
3425 handled differently in the PKCS#12 key generation code. NS
3426 treats a blank password as zero length. MSIE treats it as no
3427 password on export: but it will try both on import. We now do
3428 the same: PKCS12_parse() tries zero length and no password if
3429 the password is set to "" or NULL (NULL is now a valid password:
3430 it wasn't before) as does the pkcs12 application.
3433 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
3434 perror when PEM_read_bio_X509_REQ fails, the error message must
3435 be obtained from the error queue.
3438 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
3439 it in ERR_remove_state if appropriate, and change ERR_get_state
3440 accordingly to avoid race conditions (this is necessary because
3441 thread_hash is no longer constant once set).
3444 *) Bugfix for linux-elf makefile.one.
3447 *) RSA_get_default_method() will now cause a default
3448 RSA_METHOD to be chosen if one doesn't exist already.
3449 Previously this was only set during a call to RSA_new()
3450 or RSA_new_method(NULL) meaning it was possible for
3451 RSA_get_default_method() to return NULL.
3454 *) Added native name translation to the existing DSO code
3455 that will convert (if the flag to do so is set) filenames
3456 that are sufficiently small and have no path information
3457 into a canonical native form. Eg. "blah" converted to
3458 "libblah.so" or "blah.dll" etc.
3461 *) New function ERR_error_string_n(e, buf, len) which is like
3462 ERR_error_string(e, buf), but writes at most 'len' bytes
3463 including the 0 terminator. For ERR_error_string_n, 'buf'
3465 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
3467 *) CONF library reworked to become more general. A new CONF
3468 configuration file reader "class" is implemented as well as a
3469 new functions (NCONF_*, for "New CONF") to handle it. The now
3470 old CONF_* functions are still there, but are reimplemented to
3471 work in terms of the new functions. Also, a set of functions
3472 to handle the internal storage of the configuration data is
3473 provided to make it easier to write new configuration file
3474 reader "classes" (I can definitely see something reading a
3475 configuration file in XML format, for example), called _CONF_*,
3476 or "the configuration storage API"...
3478 The new configuration file reading functions are:
3480 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
3481 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
3483 NCONF_default, NCONF_WIN32
3485 NCONF_dump_fp, NCONF_dump_bio
3487 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
3488 NCONF_new creates a new CONF object. This works in the same way
3489 as other interfaces in OpenSSL, like the BIO interface.
3490 NCONF_dump_* dump the internal storage of the configuration file,
3491 which is useful for debugging. All other functions take the same
3492 arguments as the old CONF_* functions wth the exception of the
3493 first that must be a `CONF *' instead of a `LHASH *'.
3495 To make it easer to use the new classes with the old CONF_* functions,
3496 the function CONF_set_default_method is provided.
3499 *) Add '-tls1' option to 'openssl ciphers', which was already
3500 mentioned in the documentation but had not been implemented.
3501 (This option is not yet really useful because even the additional
3502 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
3505 *) Initial DSO code added into libcrypto for letting OpenSSL (and
3506 OpenSSL-based applications) load shared libraries and bind to
3507 them in a portable way.
3508 [Geoff Thorpe, with contributions from Richard Levitte]
3510 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
3512 *) Make sure _lrotl and _lrotr are only used with MSVC.
3514 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
3515 (the default implementation of RAND_status).
3517 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
3518 to '-clrext' (= clear extensions), as intended and documented.
3519 [Bodo Moeller; inconsistency pointed out by Michael Attili
3522 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
3523 was larger than the MD block size.
3524 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
3526 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
3527 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
3528 using the passed key: if the passed key was a private key the result
3529 of X509_print(), for example, would be to print out all the private key
3533 *) des_quad_cksum() byte order bug fix.
3534 [Ulf Möller, using the problem description in krb4-0.9.7, where
3535 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
3537 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
3539 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
3541 *) For easily testing in shell scripts whether some command
3542 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
3543 returns with exit code 0 iff no command of the given name is available.
3544 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
3545 the output goes to stdout and nothing is printed to stderr.
3546 Additional arguments are always ignored.
3548 Since for each cipher there is a command of the same name,
3549 the 'no-cipher' compilation switches can be tested this way.
3551 ('openssl no-XXX' is not able to detect pseudo-commands such
3552 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
3555 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
3558 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
3559 is set; it will be thrown away anyway because each handshake creates
3561 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
3562 to parameters -- in previous versions (since OpenSSL 0.9.3) the
3563 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
3564 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
3567 *) New s_client option -ign_eof: EOF at stdin is ignored, and
3568 'Q' and 'R' lose their special meanings (quit/renegotiate).
3569 This is part of what -quiet does; unlike -quiet, -ign_eof
3570 does not suppress any output.
3573 *) Add compatibility options to the purpose and trust code. The
3574 purpose X509_PURPOSE_ANY is "any purpose" which automatically
3575 accepts a certificate or CA, this was the previous behaviour,
3576 with all the associated security issues.
3578 X509_TRUST_COMPAT is the old trust behaviour: only and
3579 automatically trust self signed roots in certificate store. A
3580 new trust setting X509_TRUST_DEFAULT is used to specify that
3581 a purpose has no associated trust setting and it should instead
3582 use the value in the default purpose.
3585 *) Fix the PKCS#8 DSA private key code so it decodes keys again
3586 and fix a memory leak.
3589 *) In util/mkerr.pl (which implements 'make errors'), preserve
3590 reason strings from the previous version of the .c file, as
3591 the default to have only downcase letters (and digits) in
3592 automatically generated reasons codes is not always appropriate.
3595 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
3596 using strerror. Previously, ERR_reason_error_string() returned
3597 library names as reason strings for SYSerr; but SYSerr is a special
3598 case where small numbers are errno values, not library numbers.
3601 *) Add '-dsaparam' option to 'openssl dhparam' application. This
3602 converts DSA parameters into DH parameters. (When creating parameters,
3603 DSA_generate_parameters is used.)
3606 *) Include 'length' (recommended exponent length) in C code generated
3607 by 'openssl dhparam -C'.
3610 *) The second argument to set_label in perlasm was already being used
3611 so couldn't be used as a "file scope" flag. Moved to third argument
3615 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
3616 instead of RAND_bytes for encryption IVs and salts.
3619 *) Include RAND_status() into RAND_METHOD instead of implementing
3620 it only for md_rand.c Otherwise replacing the PRNG by calling
3621 RAND_set_rand_method would be impossible.
3624 *) Don't let DSA_generate_key() enter an infinite loop if the random
3625 number generation fails.
3628 *) New 'rand' application for creating pseudo-random output.
3631 *) Added configuration support for Linux/IA64
3632 [Rolf Haberrecker <rolf@suse.de>]
3634 *) Assembler module support for Mingw32.
3637 *) Shared library support for HPUX (in shlib/).
3638 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
3640 *) Shared library support for Solaris gcc.
3641 [Lutz Behnke <behnke@trustcenter.de>]
3643 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
3645 *) PKCS7_encrypt() was adding text MIME headers twice because they
3646 were added manually and by SMIME_crlf_copy().
3649 *) In bntest.c don't call BN_rand with zero bits argument.
3650 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
3652 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
3653 case was implemented. This caused BN_div_recp() to fail occasionally.
3656 *) Add an optional second argument to the set_label() in the perl
3657 assembly language builder. If this argument exists and is set
3658 to 1 it signals that the assembler should use a symbol whose
3659 scope is the entire file, not just the current function. This
3660 is needed with MASM which uses the format label:: for this scope.
3661 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
3663 *) Change the ASN1 types so they are typedefs by default. Before
3664 almost all types were #define'd to ASN1_STRING which was causing
3665 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
3669 *) Change names of new functions to the new get1/get0 naming
3670 convention: After 'get1', the caller owns a reference count
3671 and has to call ..._free; 'get0' returns a pointer to some
3672 data structure without incrementing reference counters.
3673 (Some of the existing 'get' functions increment a reference
3674 counter, some don't.)
3675 Similarly, 'set1' and 'add1' functions increase reference
3676 counters or duplicate objects.
3679 *) Allow for the possibility of temp RSA key generation failure:
3680 the code used to assume it always worked and crashed on failure.
3683 *) Fix potential buffer overrun problem in BIO_printf().
3684 [Ulf Möller, using public domain code by Patrick Powell; problem
3685 pointed out by David Sacerdote <das33@cornell.edu>]
3687 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
3688 RAND_egd() and RAND_status(). In the command line application,
3689 the EGD socket can be specified like a seed file using RANDFILE
3693 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
3694 Some CAs (e.g. Verisign) distribute certificates in this form.
3697 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
3698 list to exclude them. This means that no special compilation option
3699 is needed to use anonymous DH: it just needs to be included in the
3703 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
3704 EVP_MD_type. The old functionality is available in a new macro called
3705 EVP_MD_md(). Change code that uses it and update docs.
3708 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
3709 where the 'void *' argument is replaced by a function pointer argument.
3710 Previously 'void *' was abused to point to functions, which works on
3711 many platforms, but is not correct. As these functions are usually
3712 called by macros defined in OpenSSL header files, most source code
3713 should work without changes.
3716 *) <openssl/opensslconf.h> (which is created by Configure) now contains
3717 sections with information on -D... compiler switches used for
3718 compiling the library so that applications can see them. To enable
3719 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
3720 must be defined. E.g.,
3721 #define OPENSSL_ALGORITHM_DEFINES
3722 #include <openssl/opensslconf.h>
3723 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
3724 [Richard Levitte, Ulf and Bodo Möller]
3726 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
3730 *) Change the 'other' type in certificate aux info to a STACK_OF
3731 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
3732 the required ASN1 format: arbitrary types determined by an OID.
3735 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
3736 argument to 'req'. This is not because the function is newer or
3737 better than others it just uses the work 'NEW' in the certificate
3738 request header lines. Some software needs this.
3741 *) Reorganise password command line arguments: now passwords can be
3742 obtained from various sources. Delete the PEM_cb function and make
3743 it the default behaviour: i.e. if the callback is NULL and the
3744 usrdata argument is not NULL interpret it as a null terminated pass
3745 phrase. If usrdata and the callback are NULL then the pass phrase
3746 is prompted for as usual.
3749 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
3750 the support is automatically enabled. The resulting binaries will
3751 autodetect the card and use it if present.
3752 [Ben Laurie and Compaq Inc.]
3754 *) Work around for Netscape hang bug. This sends certificate request
3755 and server done in one record. Since this is perfectly legal in the
3756 SSL/TLS protocol it isn't a "bug" option and is on by default. See
3757 the bugs/SSLv3 entry for more info.
3760 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
3763 *) Add -rand argument to smime and pkcs12 applications and read/write
3767 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
3770 *) Add command line password options to the remaining applications.
3773 *) Bug fix for BN_div_recp() for numerators with an even number of
3777 *) More tests in bntest.c, and changed test_bn output.
3780 *) ./config recognizes MacOS X now.
3783 *) Bug fix for BN_div() when the first words of num and divsor are
3784 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
3787 *) Add support for various broken PKCS#8 formats, and command line
3788 options to produce them.
3791 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
3792 get temporary BIGNUMs from a BN_CTX.
3795 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
3799 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
3800 include a #define from the old name to the new. The original intent
3801 was that statically linked binaries could for example just call
3802 SSLeay_add_all_ciphers() to just add ciphers to the table and not
3803 link with digests. This never worked becayse SSLeay_add_all_digests()
3804 and SSLeay_add_all_ciphers() were in the same source file so calling
3805 one would link with the other. They are now in separate source files.
3808 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
3811 *) Use a less unusual form of the Miller-Rabin primality test (it used
3812 a binary algorithm for exponentiation integrated into the Miller-Rabin
3813 loop, our standard modexp algorithms are faster).
3816 *) Support for the EBCDIC character set completed.
3817 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
3819 *) Source code cleanups: use const where appropriate, eliminate casts,
3820 use void * instead of char * in lhash.
3823 *) Bugfix: ssl3_send_server_key_exchange was not restartable
3824 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
3825 this the server could overwrite ephemeral keys that the client
3829 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
3830 using 50 iterations of the Rabin-Miller test.
3832 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
3833 iterations of the Rabin-Miller test as required by the appendix
3834 to FIPS PUB 186[-1]) instead of DSA_is_prime.
3835 As BN_is_prime_fasttest includes trial division, DSA parameter
3836 generation becomes much faster.
3838 This implies a change for the callback functions in DSA_is_prime
3839 and DSA_generate_parameters: The callback function is called once
3840 for each positive witness in the Rabin-Miller test, not just
3841 occasionally in the inner loop; and the parameters to the
3842 callback function now provide an iteration count for the outer
3843 loop rather than for the current invocation of the inner loop.
3844 DSA_generate_parameters additionally can call the callback
3845 function with an 'iteration count' of -1, meaning that a
3846 candidate has passed the trial division test (when q is generated
3847 from an application-provided seed, trial division is skipped).
3850 *) New function BN_is_prime_fasttest that optionally does trial
3851 division before starting the Rabin-Miller test and has
3852 an additional BN_CTX * argument (whereas BN_is_prime always
3853 has to allocate at least one BN_CTX).
3854 'callback(1, -1, cb_arg)' is called when a number has passed the
3855 trial division stage.
3858 *) Fix for bug in CRL encoding. The validity dates weren't being handled
3862 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
3865 *) New function BN_pseudo_rand().
3868 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
3869 bignum version of BN_from_montgomery() with the working code from
3870 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
3874 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
3875 made it impossible to use the same SSL_SESSION data structure in
3876 SSL2 clients in multiple threads.
3879 *) The return value of RAND_load_file() no longer counts bytes obtained
3880 by stat(). RAND_load_file(..., -1) is new and uses the complete file
3881 to seed the PRNG (previously an explicit byte count was required).
3882 [Ulf Möller, Bodo Möller]
3884 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
3885 used (char *) instead of (void *) and had casts all over the place.
3888 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
3891 *) Retain source code compatibility for BN_prime_checks macro:
3892 BN_is_prime(..., BN_prime_checks, ...) now uses
3893 BN_prime_checks_for_size to determine the appropriate number of
3894 Rabin-Miller iterations.
3897 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
3898 DH_CHECK_P_NOT_SAFE_PRIME.
3899 (Check if this is true? OpenPGP calls them "strong".)
3902 *) Merge the functionality of "dh" and "gendh" programs into a new program
3903 "dhparam". The old programs are retained for now but will handle DH keys
3904 (instead of parameters) in future.
3907 *) Make the ciphers, s_server and s_client programs check the return values
3908 when a new cipher list is set.
3911 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
3912 ciphers. Before when the 56bit ciphers were enabled the sorting was
3915 The syntax for the cipher sorting has been extended to support sorting by
3916 cipher-strength (using the strength_bits hard coded in the tables).
3917 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
3919 Fix a bug in the cipher-command parser: when supplying a cipher command
3920 string with an "undefined" symbol (neither command nor alphanumeric
3921 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
3922 an error is flagged.
3924 Due to the strength-sorting extension, the code of the
3925 ssl_create_cipher_list() function was completely rearranged. I hope that
3926 the readability was also increased :-)
3927 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3929 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
3930 for the first serial number and places 2 in the serial number file. This
3931 avoids problems when the root CA is created with serial number zero and
3932 the first user certificate has the same issuer name and serial number
3936 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
3937 the new code. Add documentation for this stuff.
3940 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
3941 X509_*() to X509at_*() on the grounds that they don't handle X509
3942 structures and behave in an analagous way to the X509v3 functions:
3943 they shouldn't be called directly but wrapper functions should be used
3946 So we also now have some wrapper functions that call the X509at functions
3947 when passed certificate requests. (TO DO: similar things can be done with
3948 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
3949 things. Some of these need some d2i or i2d and print functionality
3950 because they handle more complex structures.)
3953 *) Add missing #ifndefs that caused missing symbols when building libssl
3954 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
3955 NO_RSA in ssl/s2*.c.
3956 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
3958 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
3959 has a return value which indicates the quality of the random data
3960 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
3961 error queue. New function RAND_pseudo_bytes() generates output that is
3962 guaranteed to be unique but not unpredictable. RAND_add is like
3963 RAND_seed, but takes an extra argument for an entropy estimate
3964 (RAND_seed always assumes full entropy).
3967 *) Do more iterations of Rabin-Miller probable prime test (specifically,
3968 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
3969 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
3970 in crypto/bn/bn_prime.c for the complete table). This guarantees a
3971 false-positive rate of at most 2^-80 for random input.
3974 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
3977 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
3978 in the 0.9.5 release), this returns the chain
3979 from an X509_CTX structure with a dup of the stack and all
3980 the X509 reference counts upped: so the stack will exist
3981 after X509_CTX_cleanup() has been called. Modify pkcs12.c
3984 Also make SSL_SESSION_print() print out the verify return
3988 *) Add manpage for the pkcs12 command. Also change the default
3989 behaviour so MAC iteration counts are used unless the new
3990 -nomaciter option is used. This improves file security and
3991 only older versions of MSIE (4.0 for example) need it.
3994 *) Honor the no-xxx Configure options when creating .DEF files.
3997 *) Add PKCS#10 attributes to field table: challengePassword,
3998 unstructuredName and unstructuredAddress. These are taken from
3999 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
4000 international characters are used.
4002 More changes to X509_ATTRIBUTE code: allow the setting of types
4003 based on strings. Remove the 'loc' parameter when adding
4004 attributes because these will be a SET OF encoding which is sorted
4008 *) Initial changes to the 'req' utility to allow request generation
4009 automation. This will allow an application to just generate a template
4010 file containing all the field values and have req construct the
4013 Initial support for X509_ATTRIBUTE handling. Stacks of these are
4014 used all over the place including certificate requests and PKCS#7
4015 structures. They are currently handled manually where necessary with
4016 some primitive wrappers for PKCS#7. The new functions behave in a
4017 manner analogous to the X509 extension functions: they allow
4018 attributes to be looked up by NID and added.
4020 Later something similar to the X509V3 code would be desirable to
4021 automatically handle the encoding, decoding and printing of the
4022 more complex types. The string types like challengePassword can
4023 be handled by the string table functions.
4025 Also modified the multi byte string table handling. Now there is
4026 a 'global mask' which masks out certain types. The table itself
4027 can use the flag STABLE_NO_MASK to ignore the mask setting: this
4028 is useful when for example there is only one permissible type
4029 (as in countryName) and using the mask might result in no valid
4033 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
4034 SSL_get_peer_finished to allow applications to obtain the latest
4035 Finished messages sent to the peer or expected from the peer,
4036 respectively. (SSL_get_peer_finished is usually the Finished message
4037 actually received from the peer, otherwise the protocol will be aborted.)
4039 As the Finished message are message digests of the complete handshake
4040 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
4041 be used for external authentication procedures when the authentication
4042 provided by SSL/TLS is not desired or is not enough.
4045 *) Enhanced support for Alpha Linux is added. Now ./config checks if
4046 the host supports BWX extension and if Compaq C is present on the
4047 $PATH. Just exploiting of the BWX extension results in 20-30%
4048 performance kick for some algorithms, e.g. DES and RC4 to mention
4049 a couple. Compaq C in turn generates ~20% faster code for MD5 and
4053 *) Add support for MS "fast SGC". This is arguably a violation of the
4054 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
4055 weak crypto and after checking the certificate is SGC a second one
4056 with strong crypto. MS SGC stops the first handshake after receiving
4057 the server certificate message and sends a second client hello. Since
4058 a server will typically do all the time consuming operations before
4059 expecting any further messages from the client (server key exchange
4060 is the most expensive) there is little difference between the two.
4062 To get OpenSSL to support MS SGC we have to permit a second client
4063 hello message after we have sent server done. In addition we have to
4064 reset the MAC if we do get this second client hello.
4067 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
4068 if a DER encoded private key is RSA or DSA traditional format. Changed
4069 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
4070 format DER encoded private key. Newer code should use PKCS#8 format which
4071 has the key type encoded in the ASN1 structure. Added DER private key
4072 support to pkcs8 application.
4075 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
4076 ciphersuites has been selected (as required by the SSL 3/TLS 1
4077 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
4078 is set, we interpret this as a request to violate the specification
4079 (the worst that can happen is a handshake failure, and 'correct'
4080 behaviour would result in a handshake failure anyway).
4083 *) In SSL_CTX_add_session, take into account that there might be multiple
4084 SSL_SESSION structures with the same session ID (e.g. when two threads
4085 concurrently obtain them from an external cache).
4086 The internal cache can handle only one SSL_SESSION with a given ID,
4087 so if there's a conflict, we now throw out the old one to achieve
4091 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
4092 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
4093 some routines that use cipher OIDs: some ciphers do not have OIDs
4094 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
4098 *) Simplify the trust setting structure and code. Now we just have
4099 two sequences of OIDs for trusted and rejected settings. These will
4100 typically have values the same as the extended key usage extension
4101 and any application specific purposes.
4103 The trust checking code now has a default behaviour: it will just
4104 check for an object with the same NID as the passed id. Functions can
4105 be provided to override either the default behaviour or the behaviour
4106 for a given id. SSL client, server and email already have functions
4107 in place for compatibility: they check the NID and also return "trusted"
4108 if the certificate is self signed.
4111 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
4112 traditional format into an EVP_PKEY structure.
4115 *) Add a password callback function PEM_cb() which either prompts for
4116 a password if usr_data is NULL or otherwise assumes it is a null
4117 terminated password. Allow passwords to be passed on command line
4118 environment or config files in a few more utilities.
4121 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
4122 keys. Add some short names for PKCS#8 PBE algorithms and allow them
4123 to be specified on the command line for the pkcs8 and pkcs12 utilities.
4124 Update documentation.
4127 *) Support for ASN1 "NULL" type. This could be handled before by using
4128 ASN1_TYPE but there wasn't any function that would try to read a NULL
4129 and produce an error if it couldn't. For compatibility we also have
4130 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
4131 don't allocate anything because they don't need to.
4134 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
4136 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
4138 *) Rebuild of the memory allocation routines used by OpenSSL code and
4139 possibly others as well. The purpose is to make an interface that
4140 provide hooks so anyone can build a separate set of allocation and
4141 deallocation routines to be used by OpenSSL, for example memory
4142 pool implementations, or something else, which was previously hard
4143 since Malloc(), Realloc() and Free() were defined as macros having
4144 the values malloc, realloc and free, respectively (except for Win32
4145 compilations). The same is provided for memory debugging code.
4146 OpenSSL already comes with functionality to find memory leaks, but
4147 this gives people a chance to debug other memory problems.
4149 With these changes, a new set of functions and macros have appeared:
4151 CRYPTO_set_mem_debug_functions() [F]
4152 CRYPTO_get_mem_debug_functions() [F]
4153 CRYPTO_dbg_set_options() [F]
4154 CRYPTO_dbg_get_options() [F]
4155 CRYPTO_malloc_debug_init() [M]
4157 The memory debug functions are NULL by default, unless the library
4158 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
4159 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
4160 gives the standard debugging functions that come with OpenSSL) or
4161 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
4162 provided by the library user) must be used. When the standard
4163 debugging functions are used, CRYPTO_dbg_set_options can be used to
4164 request additional information:
4165 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
4166 the CRYPTO_MDEBUG_xxx macro when compiling the library.
4168 Also, things like CRYPTO_set_mem_functions will always give the
4169 expected result (the new set of functions is used for allocation
4170 and deallocation) at all times, regardless of platform and compiler
4173 To finish it up, some functions that were never use in any other
4174 way than through macros have a new API and new semantic:
4177 CRYPTO_dbg_realloc()
4180 All macros of value have retained their old syntax.
4181 [Richard Levitte and Bodo Moeller]
4183 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
4184 ordering of SMIMECapabilities wasn't in "strength order" and there
4185 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
4189 *) Some ASN1 types with illegal zero length encoding (INTEGER,
4190 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
4191 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
4193 *) Merge in my S/MIME library for OpenSSL. This provides a simple
4194 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
4195 functionality to handle multipart/signed properly) and a utility
4196 called 'smime' to call all this stuff. This is based on code I
4197 originally wrote for Celo who have kindly allowed it to be
4198 included in OpenSSL.
4201 *) Add variants des_set_key_checked and des_set_key_unchecked of
4202 des_set_key (aka des_key_sched). Global variable des_check_key
4203 decides which of these is called by des_set_key; this way
4204 des_check_key behaves as it always did, but applications and
4205 the library itself, which was buggy for des_check_key == 1,
4206 have a cleaner way to pick the version they need.
4209 *) New function PKCS12_newpass() which changes the password of a
4213 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
4214 dynamic mix. In both cases the ids can be used as an index into the
4215 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
4216 functions so they accept a list of the field values and the
4217 application doesn't need to directly manipulate the X509_TRUST
4221 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
4225 *) Modify the way the V3 extension code looks up extensions. This now
4226 works in a similar way to the object code: we have some "standard"
4227 extensions in a static table which is searched with OBJ_bsearch()
4228 and the application can add dynamic ones if needed. The file
4229 crypto/x509v3/ext_dat.h now has the info: this file needs to be
4230 updated whenever a new extension is added to the core code and kept
4231 in ext_nid order. There is a simple program 'tabtest.c' which checks
4232 this. New extensions are not added too often so this file can readily
4233 be maintained manually.
4235 There are two big advantages in doing things this way. The extensions
4236 can be looked up immediately and no longer need to be "added" using
4237 X509V3_add_standard_extensions(): this function now does nothing.
4238 [Side note: I get *lots* of email saying the extension code doesn't
4239 work because people forget to call this function]
4240 Also no dynamic allocation is done unless new extensions are added:
4241 so if we don't add custom extensions there is no need to call
4242 X509V3_EXT_cleanup().
4245 *) Modify enc utility's salting as follows: make salting the default. Add a
4246 magic header, so unsalted files fail gracefully instead of just decrypting
4247 to garbage. This is because not salting is a big security hole, so people
4248 should be discouraged from doing it.
4251 *) Fixes and enhancements to the 'x509' utility. It allowed a message
4252 digest to be passed on the command line but it only used this
4253 parameter when signing a certificate. Modified so all relevant
4254 operations are affected by the digest parameter including the
4255 -fingerprint and -x509toreq options. Also -x509toreq choked if a
4256 DSA key was used because it didn't fix the digest.
4259 *) Initial certificate chain verify code. Currently tests the untrusted
4260 certificates for consistency with the verify purpose (which is set
4261 when the X509_STORE_CTX structure is set up) and checks the pathlength.
4263 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
4264 this is because it will reject chains with invalid extensions whereas
4265 every previous version of OpenSSL and SSLeay made no checks at all.
4267 Trust code: checks the root CA for the relevant trust settings. Trust
4268 settings have an initial value consistent with the verify purpose: e.g.
4269 if the verify purpose is for SSL client use it expects the CA to be
4270 trusted for SSL client use. However the default value can be changed to
4271 permit custom trust settings: one example of this would be to only trust
4272 certificates from a specific "secure" set of CAs.
4274 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
4275 which should be used for version portability: especially since the
4276 verify structure is likely to change more often now.
4278 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
4279 to set them. If not set then assume SSL clients will verify SSL servers
4282 Two new options to the verify program: -untrusted allows a set of
4283 untrusted certificates to be passed in and -purpose which sets the
4284 intended purpose of the certificate. If a purpose is set then the
4285 new chain verify code is used to check extension consistency.
4288 *) Support for the authority information access extension.
4291 *) Modify RSA and DSA PEM read routines to transparently handle
4292 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
4293 public keys in a format compatible with certificate
4294 SubjectPublicKeyInfo structures. Unfortunately there were already
4295 functions called *_PublicKey_* which used various odd formats so
4296 these are retained for compatibility: however the DSA variants were
4297 never in a public release so they have been deleted. Changed dsa/rsa
4298 utilities to handle the new format: note no releases ever handled public
4299 keys so we should be OK.
4301 The primary motivation for this change is to avoid the same fiasco
4302 that dogs private keys: there are several incompatible private key
4303 formats some of which are standard and some OpenSSL specific and
4304 require various evil hacks to allow partial transparent handling and
4305 even then it doesn't work with DER formats. Given the option anything
4306 other than PKCS#8 should be dumped: but the other formats have to
4307 stay in the name of compatibility.
4309 With public keys and the benefit of hindsight one standard format
4310 is used which works with EVP_PKEY, RSA or DSA structures: though
4311 it clearly returns an error if you try to read the wrong kind of key.
4313 Added a -pubkey option to the 'x509' utility to output the public key.
4314 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
4315 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
4316 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
4317 that do the same as the EVP_PKEY_assign_*() except they up the
4318 reference count of the added key (they don't "swallow" the
4322 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
4323 CRLs would fail if the file contained no certificates or no CRLs:
4324 added a new function to read in both types and return the number
4325 read: this means that if none are read it will be an error. The
4326 DER versions of the certificate and CRL reader would always fail
4327 because it isn't possible to mix certificates and CRLs in DER format
4328 without choking one or the other routine. Changed this to just read
4329 a certificate: this is the best we can do. Also modified the code
4330 in apps/verify.c to take notice of return codes: it was previously
4331 attempting to read in certificates from NULL pointers and ignoring
4332 any errors: this is one reason why the cert and CRL reader seemed
4333 to work. It doesn't check return codes from the default certificate
4334 routines: these may well fail if the certificates aren't installed.
4337 *) Code to support otherName option in GeneralName.
4340 *) First update to verify code. Change the verify utility
4341 so it warns if it is passed a self signed certificate:
4342 for consistency with the normal behaviour. X509_verify
4343 has been modified to it will now verify a self signed
4344 certificate if *exactly* the same certificate appears
4345 in the store: it was previously impossible to trust a
4346 single self signed certificate. This means that:
4347 openssl verify ss.pem
4348 now gives a warning about a self signed certificate but
4349 openssl verify -CAfile ss.pem ss.pem
4353 *) For servers, store verify_result in SSL_SESSION data structure
4354 (and add it to external session representation).
4355 This is needed when client certificate verifications fails,
4356 but an application-provided verification callback (set by
4357 SSL_CTX_set_cert_verify_callback) allows accepting the session
4358 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
4359 but returns 1): When the session is reused, we have to set
4360 ssl->verify_result to the appropriate error code to avoid
4362 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
4364 *) Fix a bug in the new PKCS#7 code: it didn't consider the
4365 case in PKCS7_dataInit() where the signed PKCS7 structure
4366 didn't contain any existing data because it was being created.
4367 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
4369 *) Add a salt to the key derivation routines in enc.c. This
4370 forms the first 8 bytes of the encrypted file. Also add a
4371 -S option to allow a salt to be input on the command line.
4374 *) New function X509_cmp(). Oddly enough there wasn't a function
4375 to compare two certificates. We do this by working out the SHA1
4376 hash and comparing that. X509_cmp() will be needed by the trust
4380 *) SSL_get1_session() is like SSL_get_session(), but increments
4381 the reference count in the SSL_SESSION returned.
4382 [Geoff Thorpe <geoff@eu.c2.net>]
4384 *) Fix for 'req': it was adding a null to request attributes.
4385 Also change the X509_LOOKUP and X509_INFO code to handle
4386 certificate auxiliary information.
4389 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
4393 *) Add the possibility to add extra information to the memory leak
4394 detecting output, to form tracebacks, showing from where each
4395 allocation was originated: CRYPTO_push_info("constant string") adds
4396 the string plus current file name and line number to a per-thread
4397 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
4398 is like calling CYRPTO_pop_info() until the stack is empty.
4399 Also updated memory leak detection code to be multi-thread-safe.
4402 *) Add options -text and -noout to pkcs7 utility and delete the
4403 encryption options which never did anything. Update docs.
4406 *) Add options to some of the utilities to allow the pass phrase
4407 to be included on either the command line (not recommended on
4408 OSes like Unix) or read from the environment. Update the
4409 manpages and fix a few bugs.
4412 *) Add a few manpages for some of the openssl commands.
4415 *) Fix the -revoke option in ca. It was freeing up memory twice,
4416 leaking and not finding already revoked certificates.
4419 *) Extensive changes to support certificate auxiliary information.
4420 This involves the use of X509_CERT_AUX structure and X509_AUX
4421 functions. An X509_AUX function such as PEM_read_X509_AUX()
4422 can still read in a certificate file in the usual way but it
4423 will also read in any additional "auxiliary information". By
4424 doing things this way a fair degree of compatibility can be
4425 retained: existing certificates can have this information added
4426 using the new 'x509' options.
4428 Current auxiliary information includes an "alias" and some trust
4429 settings. The trust settings will ultimately be used in enhanced
4430 certificate chain verification routines: currently a certificate
4431 can only be trusted if it is self signed and then it is trusted
4435 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
4436 The problem was that one of the replacement routines had not been working
4437 since SSLeay releases. For now the offending routine has been replaced
4438 with non-optimised assembler. Even so, this now gives around 95%
4439 performance improvement for 1024 bit RSA signs.
4442 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
4443 handling. Most clients have the effective key size in bits equal to
4444 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
4445 A few however don't do this and instead use the size of the decrypted key
4446 to determine the RC2 key length and the AlgorithmIdentifier to determine
4447 the effective key length. In this case the effective key length can still
4448 be 40 bits but the key length can be 168 bits for example. This is fixed
4449 by manually forcing an RC2 key into the EVP_PKEY structure because the
4450 EVP code can't currently handle unusual RC2 key sizes: it always assumes
4451 the key length and effective key length are equal.
4454 *) Add a bunch of functions that should simplify the creation of
4455 X509_NAME structures. Now you should be able to do:
4456 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
4457 and have it automatically work out the correct field type and fill in
4458 the structures. The more adventurous can try:
4459 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
4460 and it will (hopefully) work out the correct multibyte encoding.
4463 *) Change the 'req' utility to use the new field handling and multibyte
4464 copy routines. Before the DN field creation was handled in an ad hoc
4465 way in req, ca, and x509 which was rather broken and didn't support
4466 BMPStrings or UTF8Strings. Since some software doesn't implement
4467 BMPStrings or UTF8Strings yet, they can be enabled using the config file
4468 using the dirstring_type option. See the new comment in the default
4469 openssl.cnf for more info.
4472 *) Make crypto/rand/md_rand.c more robust:
4473 - Assure unique random numbers after fork().
4474 - Make sure that concurrent threads access the global counter and
4475 md serializably so that we never lose entropy in them
4476 or use exactly the same state in multiple threads.
4477 Access to the large state is not always serializable because
4478 the additional locking could be a performance killer, and
4479 md should be large enough anyway.
4482 *) New file apps/app_rand.c with commonly needed functionality
4483 for handling the random seed file.
4485 Use the random seed file in some applications that previously did not:
4487 dsaparam -genkey (which also ignored its '-rand' option),
4490 x509 (when signing).
4491 Except on systems with /dev/urandom, it is crucial to have a random
4492 seed file at least for key creation, DSA signing, and for DH exchanges;
4493 for RSA signatures we could do without one.
4495 gendh and gendsa (unlike genrsa) used to read only the first byte
4496 of each file listed in the '-rand' option. The function as previously
4497 found in genrsa is now in app_rand.c and is used by all programs
4498 that support '-rand'.
4501 *) In RAND_write_file, use mode 0600 for creating files;
4502 don't just chmod when it may be too late.
4505 *) Report an error from X509_STORE_load_locations
4506 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
4509 *) New function ASN1_mbstring_copy() this copies a string in either
4510 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
4511 into an ASN1_STRING type. A mask of permissible types is passed
4512 and it chooses the "minimal" type to use or an error if not type
4516 *) Add function equivalents to the various macros in asn1.h. The old
4517 macros are retained with an M_ prefix. Code inside the library can
4518 use the M_ macros. External code (including the openssl utility)
4519 should *NOT* in order to be "shared library friendly".
4522 *) Add various functions that can check a certificate's extensions
4523 to see if it usable for various purposes such as SSL client,
4524 server or S/MIME and CAs of these types. This is currently
4525 VERY EXPERIMENTAL but will ultimately be used for certificate chain
4526 verification. Also added a -purpose flag to x509 utility to
4527 print out all the purposes.
4530 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
4534 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
4535 for, obtain and decode and extension and obtain its critical flag.
4536 This allows all the necessary extension code to be handled in a
4537 single function call.
4540 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
4541 platforms. See crypto/rc4/rc4_enc.c for further details.
4544 *) New -noout option to asn1parse. This causes no output to be produced
4545 its main use is when combined with -strparse and -out to extract data
4546 from a file (which may not be in ASN.1 format).
4549 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
4550 when producing the local key id.
4551 [Richard Levitte <levitte@stacken.kth.se>]
4553 *) New option -dhparam in s_server. This allows a DH parameter file to be
4554 stated explicitly. If it is not stated then it tries the first server
4555 certificate file. The previous behaviour hard coded the filename
4559 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
4560 a public key to be input or output. For example:
4561 openssl rsa -in key.pem -pubout -out pubkey.pem
4562 Also added necessary DSA public key functions to handle this.
4565 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
4566 in the message. This was handled by allowing
4567 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
4568 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
4570 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
4571 to the end of the strings whereas this didn't. This would cause problems
4572 if strings read with d2i_ASN1_bytes() were later modified.
4573 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
4575 *) Fix for base64 decode bug. When a base64 bio reads only one line of
4576 data and it contains EOF it will end up returning an error. This is
4577 caused by input 46 bytes long. The cause is due to the way base64
4578 BIOs find the start of base64 encoded data. They do this by trying a
4579 trial decode on each line until they find one that works. When they
4580 do a flag is set and it starts again knowing it can pass all the
4581 data directly through the decoder. Unfortunately it doesn't reset
4582 the context it uses. This means that if EOF is reached an attempt
4583 is made to pass two EOFs through the context and this causes the
4584 resulting error. This can also cause other problems as well. As is
4585 usual with these problems it takes *ages* to find and the fix is
4586 trivial: move one line.
4587 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
4589 *) Ugly workaround to get s_client and s_server working under Windows. The
4590 old code wouldn't work because it needed to select() on sockets and the
4591 tty (for keypresses and to see if data could be written). Win32 only
4592 supports select() on sockets so we select() with a 1s timeout on the
4593 sockets and then see if any characters are waiting to be read, if none
4594 are present then we retry, we also assume we can always write data to
4595 the tty. This isn't nice because the code then blocks until we've
4596 received a complete line of data and it is effectively polling the
4597 keyboard at 1s intervals: however it's quite a bit better than not
4598 working at all :-) A dedicated Windows application might handle this
4599 with an event loop for example.
4602 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
4603 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
4604 will be called when RSA_sign() and RSA_verify() are used. This is useful
4605 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
4606 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
4607 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
4608 This necessitated the support of an extra signature type NID_md5_sha1
4609 for SSL signatures and modifications to the SSL library to use it instead
4610 of calling RSA_public_decrypt() and RSA_private_encrypt().
4613 *) Add new -verify -CAfile and -CApath options to the crl program, these
4614 will lookup a CRL issuers certificate and verify the signature in a
4615 similar way to the verify program. Tidy up the crl program so it
4616 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
4617 less strict. It will now permit CRL extensions even if it is not
4618 a V2 CRL: this will allow it to tolerate some broken CRLs.
4621 *) Initialize all non-automatic variables each time one of the openssl
4622 sub-programs is started (this is necessary as they may be started
4623 multiple times from the "OpenSSL>" prompt).
4624 [Lennart Bang, Bodo Moeller]
4626 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
4627 removing all other RSA functionality (this is what NO_RSA does). This
4628 is so (for example) those in the US can disable those operations covered
4629 by the RSA patent while allowing storage and parsing of RSA keys and RSA
4633 *) Non-copying interface to BIO pairs.
4634 (still largely untested)
4637 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
4638 ASCII string. This was handled independently in various places before.
4641 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
4642 UTF8 strings a character at a time.
4645 *) Use client_version from client hello to select the protocol
4646 (s23_srvr.c) and for RSA client key exchange verification
4647 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
4650 *) Add various utility functions to handle SPKACs, these were previously
4651 handled by poking round in the structure internals. Added new function
4652 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
4653 print, verify and generate SPKACs. Based on an original idea from
4654 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
4657 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
4660 *) Allow the config file extension section to be overwritten on the
4661 command line. Based on an original idea from Massimiliano Pala
4662 <madwolf@comune.modena.it>. The new option is called -extensions
4663 and can be applied to ca, req and x509. Also -reqexts to override
4664 the request extensions in req and -crlexts to override the crl extensions
4668 *) Add new feature to the SPKAC handling in ca. Now you can include
4669 the same field multiple times by preceding it by "XXXX." for example:
4672 this is the same syntax as used in the req config file.
4675 *) Allow certificate extensions to be added to certificate requests. These
4676 are specified in a 'req_extensions' option of the req section of the
4677 config file. They can be printed out with the -text option to req but
4678 are otherwise ignored at present.
4681 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
4682 data read consists of only the final block it would not decrypted because
4683 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
4684 A misplaced 'break' also meant the decrypted final block might not be
4685 copied until the next read.
4688 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
4689 a few extra parameters to the DH structure: these will be useful if
4690 for example we want the value of 'q' or implement X9.42 DH.
4693 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
4694 provides hooks that allow the default DSA functions or functions on a
4695 "per key" basis to be replaced. This allows hardware acceleration and
4696 hardware key storage to be handled without major modification to the
4697 library. Also added low level modexp hooks and CRYPTO_EX structure and
4698 associated functions.
4701 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
4702 as "read only": it can't be written to and the buffer it points to will
4703 not be freed. Reading from a read only BIO is much more efficient than
4704 a normal memory BIO. This was added because there are several times when
4705 an area of memory needs to be read from a BIO. The previous method was
4706 to create a memory BIO and write the data to it, this results in two
4707 copies of the data and an O(n^2) reading algorithm. There is a new
4708 function BIO_new_mem_buf() which creates a read only memory BIO from
4709 an area of memory. Also modified the PKCS#7 routines to use read only
4713 *) Bugfix: ssl23_get_client_hello did not work properly when called in
4714 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
4715 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
4716 but a retry condition occured while trying to read the rest.
4719 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
4720 NID_pkcs7_encrypted by default: this was wrong since this should almost
4721 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
4722 the encrypted data type: this is a more sensible place to put it and it
4723 allows the PKCS#12 code to be tidied up that duplicated this
4727 *) Changed obj_dat.pl script so it takes its input and output files on
4728 the command line. This should avoid shell escape redirection problems
4732 *) Initial support for certificate extension requests, these are included
4733 in things like Xenroll certificate requests. Included functions to allow
4734 extensions to be obtained and added.
4737 *) -crlf option to s_client and s_server for sending newlines as
4738 CRLF (as required by many protocols).
4741 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
4743 *) Install libRSAglue.a when OpenSSL is built with RSAref.
4744 [Ralf S. Engelschall]
4746 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
4747 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
4749 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
4753 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
4754 DH parameters/keys (q is lost during that conversion, but the resulting
4755 DH parameters contain its length).
4757 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
4758 much faster than DH_generate_parameters (which creates parameters
4759 where p = 2*q + 1), and also the smaller q makes DH computations
4760 much more efficient (160-bit exponentiation instead of 1024-bit
4761 exponentiation); so this provides a convenient way to support DHE
4762 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
4763 utter importance to use
4764 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4766 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4767 when such DH parameters are used, because otherwise small subgroup
4768 attacks may become possible!
4771 *) Avoid memory leak in i2d_DHparams.
4774 *) Allow the -k option to be used more than once in the enc program:
4775 this allows the same encrypted message to be read by multiple recipients.
4778 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
4779 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
4780 it will always use the numerical form of the OID, even if it has a short
4784 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
4785 method only got called if p,q,dmp1,dmq1,iqmp components were present,
4786 otherwise bn_mod_exp was called. In the case of hardware keys for example
4787 no private key components need be present and it might store extra data
4788 in the RSA structure, which cannot be accessed from bn_mod_exp.
4789 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
4790 private key operations.
4793 *) Added support for SPARC Linux.
4796 *) pem_password_cb function type incompatibly changed from
4797 typedef int pem_password_cb(char *buf, int size, int rwflag);
4799 ....(char *buf, int size, int rwflag, void *userdata);
4800 so that applications can pass data to their callbacks:
4801 The PEM[_ASN1]_{read,write}... functions and macros now take an
4802 additional void * argument, which is just handed through whenever
4803 the password callback is called.
4804 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
4806 New function SSL_CTX_set_default_passwd_cb_userdata.
4808 Compatibility note: As many C implementations push function arguments
4809 onto the stack in reverse order, the new library version is likely to
4810 interoperate with programs that have been compiled with the old
4811 pem_password_cb definition (PEM_whatever takes some data that
4812 happens to be on the stack as its last argument, and the callback
4813 just ignores this garbage); but there is no guarantee whatsoever that
4816 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
4817 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
4818 problems not only on Windows, but also on some Unix platforms.
4819 To avoid problematic command lines, these definitions are now in an
4820 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
4821 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
4824 *) MIPS III/IV assembler module is reimplemented.
4827 *) More DES library cleanups: remove references to srand/rand and
4828 delete an unused file.
4831 *) Add support for the the free Netwide assembler (NASM) under Win32,
4832 since not many people have MASM (ml) and it can be hard to obtain.
4833 This is currently experimental but it seems to work OK and pass all
4834 the tests. Check out INSTALL.W32 for info.
4837 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
4838 without temporary keys kept an extra copy of the server key,
4839 and connections with temporary keys did not free everything in case
4843 *) New function RSA_check_key and new openssl rsa option -check
4844 for verifying the consistency of RSA keys.
4845 [Ulf Moeller, Bodo Moeller]
4847 *) Various changes to make Win32 compile work:
4848 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
4849 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
4850 comparison" warnings.
4851 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
4854 *) Add a debugging option to PKCS#5 v2 key generation function: when
4855 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
4856 derived keys are printed to stderr.
4859 *) Copy the flags in ASN1_STRING_dup().
4860 [Roman E. Pavlov <pre@mo.msk.ru>]
4862 *) The x509 application mishandled signing requests containing DSA
4863 keys when the signing key was also DSA and the parameters didn't match.
4865 It was supposed to omit the parameters when they matched the signing key:
4866 the verifying software was then supposed to automatically use the CA's
4867 parameters if they were absent from the end user certificate.
4869 Omitting parameters is no longer recommended. The test was also
4870 the wrong way round! This was probably due to unusual behaviour in
4871 EVP_cmp_parameters() which returns 1 if the parameters match.
4872 This meant that parameters were omitted when they *didn't* match and
4873 the certificate was useless. Certificates signed with 'ca' didn't have
4875 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
4877 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
4878 The interface is as follows:
4879 Applications can use
4880 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
4881 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
4882 "off" is now the default.
4883 The library internally uses
4884 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
4885 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
4886 to disable memory-checking temporarily.
4888 Some inconsistent states that previously were possible (and were
4889 even the default) are now avoided.
4891 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
4892 with each memory chunk allocated; this is occasionally more helpful
4893 than just having a counter.
4895 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
4897 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
4901 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
4902 which largely parallels "options", but is for changing API behaviour,
4903 whereas "options" are about protocol behaviour.
4904 Initial "mode" flags are:
4906 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
4907 a single record has been written.
4908 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
4909 retries use the same buffer location.
4910 (But all of the contents must be
4914 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
4917 *) Fix problems with no-hmac etc.
4918 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
4920 *) New functions RSA_get_default_method(), RSA_set_method() and
4921 RSA_get_method(). These allows replacement of RSA_METHODs without having
4922 to mess around with the internals of an RSA structure.
4925 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
4926 Also really enable memory leak checks in openssl.c and in some
4928 [Chad C. Mulligan, Bodo Moeller]
4930 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
4931 up the length of negative integers. This has now been simplified to just
4932 store the length when it is first determined and use it later, rather
4933 than trying to keep track of where data is copied and updating it to
4935 [Steve Henson, reported by Brien Wheeler
4936 <bwheeler@authentica-security.com>]
4938 *) Add a new function PKCS7_signatureVerify. This allows the verification
4939 of a PKCS#7 signature but with the signing certificate passed to the
4940 function itself. This contrasts with PKCS7_dataVerify which assumes the
4941 certificate is present in the PKCS#7 structure. This isn't always the
4942 case: certificates can be omitted from a PKCS#7 structure and be
4943 distributed by "out of band" means (such as a certificate database).
4946 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
4947 function prototypes in pem.h, also change util/mkdef.pl to add the
4948 necessary function names.
4951 *) mk1mf.pl (used by Windows builds) did not properly read the
4952 options set by Configure in the top level Makefile, and Configure
4953 was not even able to write more than one option correctly.
4954 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
4957 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
4958 file to be loaded from a BIO or FILE pointer. The BIO version will
4959 for example allow memory BIOs to contain config info.
4962 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
4963 Whoever hopes to achieve shared-library compatibility across versions
4964 must use this, not the compile-time macro.
4965 (Exercise 0.9.4: Which is the minimum library version required by
4967 Note: All this applies only to multi-threaded programs, others don't
4971 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
4972 through a BIO pair triggered the default case, i.e.
4973 SSLerr(...,SSL_R_UNKNOWN_STATE).
4976 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
4977 can use the SSL library even if none of the specific BIOs is
4981 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
4982 for the encoded length.
4983 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
4985 *) Add initial documentation of the X509V3 functions.
4988 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
4989 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
4990 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
4991 secure PKCS#8 private key format with a high iteration count.
4994 *) Fix determination of Perl interpreter: A perl or perl5
4995 _directory_ in $PATH was also accepted as the interpreter.
4996 [Ralf S. Engelschall]
4998 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
4999 wrong with it but it was very old and did things like calling
5000 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
5004 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
5005 to use the new extension code.
5008 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
5009 with macros. This should make it easier to change their form, add extra
5010 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
5014 *) Add to configuration table a new entry that can specify an alternative
5015 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
5016 according to Mark Crispin <MRC@Panda.COM>.
5020 *) DES CBC did not update the IV. Weird.
5023 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
5024 Changing the behaviour of the former might break existing programs --
5025 where IV updating is needed, des_ncbc_encrypt can be used.
5028 *) When bntest is run from "make test" it drives bc to check its
5029 calculations, as well as internally checking them. If an internal check
5030 fails, it needs to cause bc to give a non-zero result or make test carries
5031 on without noticing the failure. Fixed.
5034 *) DES library cleanups.
5037 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
5038 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
5039 ciphers. NOTE: although the key derivation function has been verified
5040 against some published test vectors it has not been extensively tested
5041 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
5045 *) Instead of "mkdir -p", which is not fully portable, use new
5046 Perl script "util/mkdir-p.pl".
5049 *) Rewrite the way password based encryption (PBE) is handled. It used to
5050 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
5051 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
5052 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
5053 the 'parameter' field of the AlgorithmIdentifier is passed to the
5054 underlying key generation function so it must do its own ASN1 parsing.
5055 This has also changed the EVP_PBE_CipherInit() function which now has a
5056 'parameter' argument instead of literal salt and iteration count values
5057 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
5060 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
5061 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
5062 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
5063 KEY" because this clashed with PKCS#8 unencrypted string. Since this
5064 value was just used as a "magic string" and not used directly its
5065 value doesn't matter.
5068 *) Introduce some semblance of const correctness to BN. Shame C doesn't
5072 *) "linux-sparc64" configuration (ultrapenguin).
5073 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
5074 "linux-sparc" configuration.
5075 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
5077 *) config now generates no-xxx options for missing ciphers.
5080 *) Support the EBCDIC character set (work in progress).
5081 File ebcdic.c not yet included because it has a different license.
5082 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5084 *) Support BS2000/OSD-POSIX.
5085 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5087 *) Make callbacks for key generation use void * instead of char *.
5090 *) Make S/MIME samples compile (not yet tested).
5093 *) Additional typesafe stacks.
5096 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
5100 Changes between 0.9.3 and 0.9.3a [29 May 1999]
5102 *) New configuration variant "sco5-gcc".
5104 *) Updated some demos.
5105 [Sean O Riordain, Wade Scholine]
5107 *) Add missing BIO_free at exit of pkcs12 application.
5110 *) Fix memory leak in conf.c.
5113 *) Updates for Win32 to assembler version of MD5.
5116 *) Set #! path to perl in apps/der_chop to where we found it
5117 instead of using a fixed path.
5120 *) SHA library changes for irix64-mips4-cc.
5123 *) Improvements for VMS support.
5127 Changes between 0.9.2b and 0.9.3 [24 May 1999]
5129 *) Bignum library bug fix. IRIX 6 passes "make test" now!
5130 This also avoids the problems with SC4.2 and unpatched SC5.
5131 [Andy Polyakov <appro@fy.chalmers.se>]
5133 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
5134 These are required because of the typesafe stack would otherwise break
5135 existing code. If old code used a structure member which used to be STACK
5136 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
5137 sk_num or sk_value it would produce an error because the num, data members
5138 are not present in STACK_OF. Now it just produces a warning. sk_set
5139 replaces the old method of assigning a value to sk_value
5140 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
5141 that does this will no longer work (and should use sk_set instead) but
5142 this could be regarded as a "questionable" behaviour anyway.
5145 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
5146 correctly handle encrypted S/MIME data.
5149 *) Change type of various DES function arguments from des_cblock
5150 (which means, in function argument declarations, pointer to char)
5151 to des_cblock * (meaning pointer to array with 8 char elements),
5152 which allows the compiler to do more typechecking; it was like
5153 that back in SSLeay, but with lots of ugly casts.
5155 Introduce new type const_des_cblock.
5158 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
5159 problems: find RecipientInfo structure that matches recipient certificate
5160 and initialise the ASN1 structures properly based on passed cipher.
5163 *) Belatedly make the BN tests actually check the results.
5166 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
5167 to and from BNs: it was completely broken. New compilation option
5168 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
5169 key elements as negative integers.
5172 *) Reorganize and speed up MD5.
5173 [Andy Polyakov <appro@fy.chalmers.se>]
5176 [Richard Levitte <richard@levitte.org>]
5178 *) New option -out to asn1parse to allow the parsed structure to be
5179 output to a file. This is most useful when combined with the -strparse
5180 option to examine the output of things like OCTET STRINGS.
5183 *) Make SSL library a little more fool-proof by not requiring any longer
5184 that SSL_set_{accept,connect}_state be called before
5185 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
5186 in many applications because usually everything *appeared* to work as
5187 intended anyway -- now it really works as intended).
5190 *) Move openssl.cnf out of lib/.
5193 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
5194 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
5195 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
5196 [Ralf S. Engelschall]
5198 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
5199 handle PKCS#7 enveloped data properly.
5200 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
5202 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
5203 copying pointers. The cert_st handling is changed by this in
5204 various ways (and thus what used to be known as ctx->default_cert
5205 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
5206 any longer when s->cert does not give us what we need).
5207 ssl_cert_instantiate becomes obsolete by this change.
5208 As soon as we've got the new code right (possibly it already is?),
5209 we have solved a couple of bugs of the earlier code where s->cert
5210 was used as if it could not have been shared with other SSL structures.
5212 Note that using the SSL API in certain dirty ways now will result
5213 in different behaviour than observed with earlier library versions:
5214 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
5215 does not influence s as it used to.
5217 In order to clean up things more thoroughly, inside SSL_SESSION
5218 we don't use CERT any longer, but a new structure SESS_CERT
5219 that holds per-session data (if available); currently, this is
5220 the peer's certificate chain and, for clients, the server's certificate
5221 and temporary key. CERT holds only those values that can have
5222 meaningful defaults in an SSL_CTX.
5225 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
5226 from the internal representation. Various PKCS#7 fixes: remove some
5227 evil casts and set the enc_dig_alg field properly based on the signing
5231 *) Allow PKCS#12 password to be set from the command line or the
5232 environment. Let 'ca' get its config file name from the environment
5233 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
5237 *) Allow certificate policies extension to use an IA5STRING for the
5238 organization field. This is contrary to the PKIX definition but
5239 VeriSign uses it and IE5 only recognises this form. Document 'x509'
5243 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
5244 without disallowing inline assembler and the like for non-pedantic builds.
5247 *) Support Borland C++ builder.
5248 [Janez Jere <jj@void.si>, modified by Ulf Möller]
5253 *) SHA-1 cleanups and performance enhancements.
5254 [Andy Polyakov <appro@fy.chalmers.se>]
5256 *) Sparc v8plus assembler for the bignum library.
5257 [Andy Polyakov <appro@fy.chalmers.se>]
5259 *) Accept any -xxx and +xxx compiler options in Configure.
5262 *) Update HPUX configuration.
5265 *) Add missing sk_<type>_unshift() function to safestack.h
5266 [Ralf S. Engelschall]
5268 *) New function SSL_CTX_use_certificate_chain_file that sets the
5269 "extra_cert"s in addition to the certificate. (This makes sense
5270 only for "PEM" format files, as chains as a whole are not
5274 *) Support verify_depth from the SSL API.
5275 x509_vfy.c had what can be considered an off-by-one-error:
5276 Its depth (which was not part of the external interface)
5277 was actually counting the number of certificates in a chain;
5278 now it really counts the depth.
5281 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
5282 instead of X509err, which often resulted in confusing error
5283 messages since the error codes are not globally unique
5284 (e.g. an alleged error in ssl3_accept when a certificate
5285 didn't match the private key).
5287 *) New function SSL_CTX_set_session_id_context that allows to set a default
5288 value (so that you don't need SSL_set_session_id_context for each
5289 connection using the SSL_CTX).
5292 *) OAEP decoding bug fix.
5295 *) Support INSTALL_PREFIX for package builders, as proposed by
5299 *) New Configure options "threads" and "no-threads". For systems
5300 where the proper compiler options are known (currently Solaris
5301 and Linux), "threads" is the default.
5304 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
5307 *) Install various scripts to $(OPENSSLDIR)/misc, not to
5308 $(INSTALLTOP)/bin -- they shouldn't clutter directories
5309 such as /usr/local/bin.
5312 *) "make linux-shared" to build shared libraries.
5313 [Niels Poppe <niels@netbox.org>]
5315 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
5318 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
5319 extension adding in x509 utility.
5322 *) Remove NOPROTO sections and error code comments.
5325 *) Partial rewrite of the DEF file generator to now parse the ANSI
5329 *) New Configure options --prefix=DIR and --openssldir=DIR.
5332 *) Complete rewrite of the error code script(s). It is all now handled
5333 by one script at the top level which handles error code gathering,
5334 header rewriting and C source file generation. It should be much better
5335 than the old method: it now uses a modified version of Ulf's parser to
5336 read the ANSI prototypes in all header files (thus the old K&R definitions
5337 aren't needed for error creation any more) and do a better job of
5338 translating function codes into names. The old 'ASN1 error code imbedded
5339 in a comment' is no longer necessary and it doesn't use .err files which
5340 have now been deleted. Also the error code call doesn't have to appear all
5341 on one line (which resulted in some large lines...).
5344 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
5347 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
5348 0 (which usually indicates a closed connection), but continue reading.
5351 *) Fix some race conditions.
5354 *) Add support for CRL distribution points extension. Add Certificate
5355 Policies and CRL distribution points documentation.
5358 *) Move the autogenerated header file parts to crypto/opensslconf.h.
5361 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
5362 8 of keying material. Merlin has also confirmed interop with this fix
5363 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
5364 [Merlin Hughes <merlin@baltimore.ie>]
5366 *) Fix lots of warnings.
5367 [Richard Levitte <levitte@stacken.kth.se>]
5369 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
5370 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
5371 [Richard Levitte <levitte@stacken.kth.se>]
5373 *) Fix problems with sizeof(long) == 8.
5374 [Andy Polyakov <appro@fy.chalmers.se>]
5376 *) Change functions to ANSI C.
5379 *) Fix typos in error codes.
5380 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
5382 *) Remove defunct assembler files from Configure.
5385 *) SPARC v8 assembler BIGNUM implementation.
5386 [Andy Polyakov <appro@fy.chalmers.se>]
5388 *) Support for Certificate Policies extension: both print and set.
5389 Various additions to support the r2i method this uses.
5392 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
5393 return a const string when you are expecting an allocated buffer.
5396 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
5397 types DirectoryString and DisplayText.
5400 *) Add code to allow r2i extensions to access the configuration database,
5401 add an LHASH database driver and add several ctx helper functions.
5404 *) Fix an evil bug in bn_expand2() which caused various BN functions to
5405 fail when they extended the size of a BIGNUM.
5408 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
5409 support typesafe stack.
5412 *) Fix typo in SSL_[gs]et_options().
5413 [Nils Frostberg <nils@medcom.se>]
5415 *) Delete various functions and files that belonged to the (now obsolete)
5416 old X509V3 handling code.
5419 *) New Configure option "rsaref".
5422 *) Don't auto-generate pem.h.
5425 *) Introduce type-safe ASN.1 SETs.
5428 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
5429 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
5431 *) Introduce type-safe STACKs. This will almost certainly break lots of code
5432 that links with OpenSSL (well at least cause lots of warnings), but fear
5433 not: the conversion is trivial, and it eliminates loads of evil casts. A
5434 few STACKed things have been converted already. Feel free to convert more.
5435 In the fullness of time, I'll do away with the STACK type altogether.
5438 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
5439 specified in <certfile> by updating the entry in the index.txt file.
5440 This way one no longer has to edit the index.txt file manually for
5441 revoking a certificate. The -revoke option does the gory details now.
5442 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
5444 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
5445 `-text' option at all and this way the `-noout -text' combination was
5446 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
5447 [Ralf S. Engelschall]
5449 *) Make sure a corresponding plain text error message exists for the
5450 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
5451 verify callback function determined that a certificate was revoked.
5452 [Ralf S. Engelschall]
5454 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
5455 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
5456 all available cipers including rc5, which was forgotten until now.
5457 In order to let the testing shell script know which algorithms
5458 are available, a new (up to now undocumented) command
5459 "openssl list-cipher-commands" is used.
5462 *) Bugfix: s_client occasionally would sleep in select() when
5463 it should have checked SSL_pending() first.
5466 *) New functions DSA_do_sign and DSA_do_verify to provide access to
5467 the raw DSA values prior to ASN.1 encoding.
5470 *) Tweaks to Configure
5471 [Niels Poppe <niels@netbox.org>]
5473 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
5477 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
5480 *) New config option to avoid instructions that are illegal on the 80386.
5481 The default code is faster, but requires at least a 486.
5484 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
5485 SSL2_SERVER_VERSION (not used at all) macros, which are now the
5486 same as SSL2_VERSION anyway.
5489 *) New "-showcerts" option for s_client.
5492 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
5493 application. Various cleanups and fixes.
5496 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
5497 modify error routines to work internally. Add error codes and PBE init
5498 to library startup routines.
5501 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
5502 packing functions to asn1 and evp. Changed function names and error
5503 codes along the way.
5506 *) PKCS12 integration: and so it begins... First of several patches to
5507 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
5508 objects to objects.h
5511 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
5512 and display support for Thawte strong extranet extension.
5515 *) Add LinuxPPC support.
5516 [Jeff Dubrule <igor@pobox.org>]
5518 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
5519 bn_div_words in alpha.s.
5520 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
5522 *) Make sure the RSA OAEP test is skipped under -DRSAref because
5523 OAEP isn't supported when OpenSSL is built with RSAref.
5524 [Ulf Moeller <ulf@fitug.de>]
5526 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
5527 so they no longer are missing under -DNOPROTO.
5528 [Soren S. Jorvang <soren@t.dk>]
5531 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
5533 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
5534 doesn't work when the session is reused. Coming soon!
5537 *) Fix a security hole, that allows sessions to be reused in the wrong
5538 context thus bypassing client cert protection! All software that uses
5539 client certs and session caches in multiple contexts NEEDS PATCHING to
5540 allow session reuse! A fuller solution is in the works.
5541 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
5543 *) Some more source tree cleanups (removed obsolete files
5544 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
5545 permission on "config" script to be executable) and a fix for the INSTALL
5547 [Ulf Moeller <ulf@fitug.de>]
5549 *) Remove some legacy and erroneous uses of malloc, free instead of
5551 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
5553 *) Make rsa_oaep_test return non-zero on error.
5554 [Ulf Moeller <ulf@fitug.de>]
5556 *) Add support for native Solaris shared libraries. Configure
5557 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
5558 if someone would make that last step automatic.
5559 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
5561 *) ctx_size was not built with the right compiler during "make links". Fixed.
5564 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
5565 except NULL ciphers". This means the default cipher list will no longer
5566 enable NULL ciphers. They need to be specifically enabled e.g. with
5567 the string "DEFAULT:eNULL".
5570 *) Fix to RSA private encryption routines: if p < q then it would
5571 occasionally produce an invalid result. This will only happen with
5572 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
5575 *) Be less restrictive and allow also `perl util/perlpath.pl
5576 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
5577 because this way one can also use an interpreter named `perl5' (which is
5578 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
5579 installed as `perl').
5580 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5582 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
5583 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5585 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
5586 advapi32.lib to Win32 build and change the pem test comparision
5587 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
5588 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
5589 and crypto/des/ede_cbcm_enc.c.
5592 *) DES quad checksum was broken on big-endian architectures. Fixed.
5595 *) Comment out two functions in bio.h that aren't implemented. Fix up the
5596 Win32 test batch file so it (might) work again. The Win32 test batch file
5597 is horrible: I feel ill....
5600 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
5601 in e_os.h. Audit of header files to check ANSI and non ANSI
5602 sections: 10 functions were absent from non ANSI section and not exported
5603 from Windows DLLs. Fixed up libeay.num for new functions.
5606 *) Make `openssl version' output lines consistent.
5607 [Ralf S. Engelschall]
5609 *) Fix Win32 symbol export lists for BIO functions: Added
5610 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
5611 to ms/libeay{16,32}.def.
5612 [Ralf S. Engelschall]
5614 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
5615 fine under Unix and passes some trivial tests I've now added. But the
5616 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
5617 added to make sure no one expects that this stuff really works in the
5618 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
5619 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
5621 [Ralf S. Engelschall]
5623 *) Fix the generation of two part addresses in perl.
5624 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
5626 *) Add config entry for Linux on MIPS.
5627 [John Tobey <jtobey@channel1.com>]
5629 *) Make links whenever Configure is run, unless we are on Windoze.
5632 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
5633 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
5637 *) Add a useful kludge to allow package maintainers to specify compiler and
5638 other platforms details on the command line without having to patch the
5639 Configure script everytime: One now can use ``perl Configure
5640 <id>:<details>'', i.e. platform ids are allowed to have details appended
5641 to them (seperated by colons). This is treated as there would be a static
5642 pre-configured entry in Configure's %table under key <id> with value
5643 <details> and ``perl Configure <id>'' is called. So, when you want to
5644 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
5645 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
5646 now, which overrides the FreeBSD-elf entry on-the-fly.
5647 [Ralf S. Engelschall]
5649 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
5652 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
5653 on the `perl Configure ...' command line. This way one can compile
5654 OpenSSL libraries with Position Independent Code (PIC) which is needed
5655 for linking it into DSOs.
5656 [Ralf S. Engelschall]
5658 *) Remarkably, export ciphers were totally broken and no-one had noticed!
5662 *) Cleaned up the LICENSE document: The official contact for any license
5663 questions now is the OpenSSL core team under openssl-core@openssl.org.
5664 And add a paragraph about the dual-license situation to make sure people
5665 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
5666 to the OpenSSL toolkit.
5667 [Ralf S. Engelschall]
5669 *) General source tree makefile cleanups: Made `making xxx in yyy...'
5670 display consistent in the source tree and replaced `/bin/rm' by `rm'.
5671 Additonally cleaned up the `make links' target: Remove unnecessary
5672 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
5673 to speed processing and no longer clutter the display with confusing
5674 stuff. Instead only the actually done links are displayed.
5675 [Ralf S. Engelschall]
5677 *) Permit null encryption ciphersuites, used for authentication only. It used
5678 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
5679 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
5683 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
5684 signed attributes when verifying signatures (this would break them),
5685 the detached data encoding was wrong and public keys obtained using
5686 X509_get_pubkey() weren't freed.
5689 *) Add text documentation for the BUFFER functions. Also added a work around
5690 to a Win95 console bug. This was triggered by the password read stuff: the
5691 last character typed gets carried over to the next fread(). If you were
5692 generating a new cert request using 'req' for example then the last
5693 character of the passphrase would be CR which would then enter the first
5697 *) Added the new `Includes OpenSSL Cryptography Software' button as
5698 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
5699 button and can be used by applications based on OpenSSL to show the
5700 relationship to the OpenSSL project.
5701 [Ralf S. Engelschall]
5703 *) Remove confusing variables in function signatures in files
5704 ssl/ssl_lib.c and ssl/ssl.h.
5705 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5707 *) Don't install bss_file.c under PREFIX/include/
5708 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5710 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
5711 functions that return function pointers and has support for NT specific
5712 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
5713 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
5714 unsigned to signed types: this was killing the Win32 compile.
5717 *) Add new certificate file to stack functions,
5718 SSL_add_dir_cert_subjects_to_stack() and
5719 SSL_add_file_cert_subjects_to_stack(). These largely supplant
5720 SSL_load_client_CA_file(), and can be used to add multiple certs easily
5721 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
5722 This means that Apache-SSL and similar packages don't have to mess around
5723 to add as many CAs as they want to the preferred list.
5726 *) Experiment with doxygen documentation. Currently only partially applied to
5728 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
5729 openssl.doxy as the configuration file.
5732 *) Get rid of remaining C++-style comments which strict C compilers hate.
5733 [Ralf S. Engelschall, pointed out by Carlos Amengual]
5735 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
5736 compiled in by default: it has problems with large keys.
5739 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
5740 DH private keys and/or callback functions which directly correspond to
5741 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
5742 is needed for applications which have to configure certificates on a
5743 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
5745 For the RSA certificate situation is makes no difference, but
5746 for the DSA certificate situation this fixes the "no shared cipher"
5747 problem where the OpenSSL cipher selection procedure failed because the
5748 temporary keys were not overtaken from the context and the API provided
5749 no way to reconfigure them.
5750 The new functions now let applications reconfigure the stuff and they
5751 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
5752 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
5753 non-public-API function ssl_cert_instantiate() is used as a helper
5754 function and also to reduce code redundancy inside ssl_rsa.c.
5755 [Ralf S. Engelschall]
5757 *) Move s_server -dcert and -dkey options out of the undocumented feature
5758 area because they are useful for the DSA situation and should be
5759 recognized by the users.
5760 [Ralf S. Engelschall]
5762 *) Fix the cipher decision scheme for export ciphers: the export bits are
5763 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
5764 SSL_EXP_MASK. So, the original variable has to be used instead of the
5765 already masked variable.
5766 [Richard Levitte <levitte@stacken.kth.se>]
5768 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
5769 [Richard Levitte <levitte@stacken.kth.se>]
5771 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
5772 from `int' to `unsigned int' because it's a length and initialized by
5773 EVP_DigestFinal() which expects an `unsigned int *'.
5774 [Richard Levitte <levitte@stacken.kth.se>]
5776 *) Don't hard-code path to Perl interpreter on shebang line of Configure
5777 script. Instead use the usual Shell->Perl transition trick.
5778 [Ralf S. Engelschall]
5780 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
5781 (in addition to RSA certificates) to match the behaviour of `openssl dsa
5782 -noout -modulus' as it's already the case for `openssl rsa -noout
5783 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
5784 currently the public key is printed (a decision which was already done by
5785 `openssl dsa -modulus' in the past) which serves a similar purpose.
5786 Additionally the NO_RSA no longer completely removes the whole -modulus
5787 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
5789 [Ralf S. Engelschall]
5791 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
5792 BIO. See the source (crypto/evp/bio_ok.c) for more info.
5793 [Arne Ansper <arne@ats.cyber.ee>]
5795 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
5796 to be added. Now both 'req' and 'ca' can use new objects defined in the
5800 *) Add cool BIO that does syslog (or event log on NT).
5801 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
5803 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
5804 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
5805 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
5806 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
5809 *) Add preliminary config info for new extension code.
5812 *) Make RSA_NO_PADDING really use no padding.
5813 [Ulf Moeller <ulf@fitug.de>]
5815 *) Generate errors when private/public key check is done.
5818 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
5819 for some CRL extensions and new objects added.
5822 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
5823 key usage extension and fuller support for authority key id.
5826 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
5827 padding method for RSA, which is recommended for new applications in PKCS
5828 #1 v2.0 (RFC 2437, October 1998).
5829 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
5830 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
5831 against Bleichbacher's attack on RSA.
5832 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
5835 *) Updates to the new SSL compression code
5836 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5838 *) Fix so that the version number in the master secret, when passed
5839 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
5840 (because the server will not accept higher), that the version number
5841 is 0x03,0x01, not 0x03,0x00
5842 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5844 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
5845 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
5846 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
5849 *) Support for RAW extensions where an arbitrary extension can be
5850 created by including its DER encoding. See apps/openssl.cnf for
5854 *) Make sure latest Perl versions don't interpret some generated C array
5855 code as Perl array code in the crypto/err/err_genc.pl script.
5856 [Lars Weber <3weber@informatik.uni-hamburg.de>]
5858 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
5859 not many people have the assembler. Various Win32 compilation fixes and
5860 update to the INSTALL.W32 file with (hopefully) more accurate Win32
5864 *) Modify configure script 'Configure' to automatically create crypto/date.h
5865 file under Win32 and also build pem.h from pem.org. New script
5866 util/mkfiles.pl to create the MINFO file on environments that can't do a
5867 'make files': perl util/mkfiles.pl >MINFO should work.
5870 *) Major rework of DES function declarations, in the pursuit of correctness
5871 and purity. As a result, many evil casts evaporated, and some weirdness,
5872 too. You may find this causes warnings in your code. Zapping your evil
5873 casts will probably fix them. Mostly.
5876 *) Fix for a typo in asn1.h. Bug fix to object creation script
5877 obj_dat.pl. It considered a zero in an object definition to mean
5878 "end of object": none of the objects in objects.h have any zeros
5879 so it wasn't spotted.
5880 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
5882 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
5883 Masking (CBCM). In the absence of test vectors, the best I have been able
5884 to do is check that the decrypt undoes the encrypt, so far. Send me test
5885 vectors if you have them.
5888 *) Correct calculation of key length for export ciphers (too much space was
5889 allocated for null ciphers). This has not been tested!
5892 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
5893 message is now correct (it understands "crypto" and "ssl" on its
5894 command line). There is also now an "update" option. This will update
5895 the util/ssleay.num and util/libeay.num files with any new functions.
5897 perl util/mkdef.pl crypto ssl update
5898 it will update them.
5901 *) Overhauled the Perl interface (perl/*):
5902 - ported BN stuff to OpenSSL's different BN library
5903 - made the perl/ source tree CVS-aware
5904 - renamed the package from SSLeay to OpenSSL (the files still contain
5905 their history because I've copied them in the repository)
5906 - removed obsolete files (the test scripts will be replaced
5907 by better Test::Harness variants in the future)
5908 [Ralf S. Engelschall]
5910 *) First cut for a very conservative source tree cleanup:
5911 1. merge various obsolete readme texts into doc/ssleay.txt
5912 where we collect the old documents and readme texts.
5913 2. remove the first part of files where I'm already sure that we no
5914 longer need them because of three reasons: either they are just temporary
5915 files which were left by Eric or they are preserved original files where
5916 I've verified that the diff is also available in the CVS via "cvs diff
5917 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
5918 the crypto/md/ stuff).
5919 [Ralf S. Engelschall]
5921 *) More extension code. Incomplete support for subject and issuer alt
5922 name, issuer and authority key id. Change the i2v function parameters
5923 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
5924 what that's for :-) Fix to ASN1 macro which messed up
5925 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
5928 *) Preliminary support for ENUMERATED type. This is largely copied from the
5932 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
5933 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5935 *) Make sure `make rehash' target really finds the `openssl' program.
5936 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5938 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
5939 like to hear about it if this slows down other processors.
5942 *) Add CygWin32 platform information to Configure script.
5943 [Alan Batie <batie@aahz.jf.intel.com>]
5945 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
5946 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
5948 *) New program nseq to manipulate netscape certificate sequences
5951 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
5955 *) Fixes to BN code. Previously the default was to define BN_RECURSION
5956 but the BN code had some problems that would cause failures when
5957 doing certificate verification and some other functions.
5958 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5960 *) Add ASN1 and PEM code to support netscape certificate sequences.
5963 *) Add ASN1 and PEM code to support netscape certificate sequences.
5966 *) Add several PKIX and private extended key usage OIDs.
5969 *) Modify the 'ca' program to handle the new extension code. Modify
5970 openssl.cnf for new extension format, add comments.
5973 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
5974 and add a sample to openssl.cnf so req -x509 now adds appropriate
5978 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
5979 error code, add initial support to X509_print() and x509 application.
5982 *) Takes a deep breath and start addding X509 V3 extension support code. Add
5983 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
5984 stuff is currently isolated and isn't even compiled yet.
5987 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
5988 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
5989 Removed the versions check from X509 routines when loading extensions:
5990 this allows certain broken certificates that don't set the version
5991 properly to be processed.
5994 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
5995 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
5996 can still be regenerated with "make depend".
5999 *) Spelling mistake in C version of CAST-128.
6000 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
6002 *) Changes to the error generation code. The perl script err-code.pl
6003 now reads in the old error codes and retains the old numbers, only
6004 adding new ones if necessary. It also only changes the .err files if new
6005 codes are added. The makefiles have been modified to only insert errors
6006 when needed (to avoid needlessly modifying header files). This is done
6007 by only inserting errors if the .err file is newer than the auto generated
6008 C file. To rebuild all the error codes from scratch (the old behaviour)
6009 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
6010 or delete all the .err files.
6013 *) CAST-128 was incorrectly implemented for short keys. The C version has
6014 been fixed, but is untested. The assembler versions are also fixed, but
6015 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
6016 to regenerate it if needed.
6017 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
6018 Hagino <itojun@kame.net>]
6020 *) File was opened incorrectly in randfile.c.
6021 [Ulf Möller <ulf@fitug.de>]
6023 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
6024 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
6025 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
6026 al: it's just almost always a UTCTime. Note this patch adds new error
6027 codes so do a "make errors" if there are problems.
6030 *) Correct Linux 1 recognition in config.
6031 [Ulf Möller <ulf@fitug.de>]
6033 *) Remove pointless MD5 hash when using DSA keys in ca.
6034 [Anonymous <nobody@replay.com>]
6036 *) Generate an error if given an empty string as a cert directory. Also
6037 generate an error if handed NULL (previously returned 0 to indicate an
6038 error, but didn't set one).
6039 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
6041 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
6044 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
6045 parameters. This was causing a warning which killed off the Win32 compile.
6048 *) Remove C++ style comments from crypto/bn/bn_local.h.
6049 [Neil Costigan <neil.costigan@celocom.com>]
6051 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
6052 based on a text string, looking up short and long names and finally
6053 "dot" format. The "dot" format stuff didn't work. Added new function
6054 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
6055 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
6056 OID is not part of the table.
6059 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
6060 X509_LOOKUP_by_alias().
6063 *) Sort openssl functions by name.
6066 *) Get the gendsa program working (hopefully) and add it to app list. Remove
6067 encryption from sample DSA keys (in case anyone is interested the password
6071 *) Make _all_ *_free functions accept a NULL pointer.
6072 [Frans Heymans <fheymans@isaserver.be>]
6074 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
6076 [Anonymous <nobody@replay.com>]
6078 *) s_server should send the CAfile as acceptable CAs, not its own cert.
6079 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6081 *) Don't blow it for numeric -newkey arguments to apps/req.
6082 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6084 *) Temp key "for export" tests were wrong in s3_srvr.c.
6085 [Anonymous <nobody@replay.com>]
6087 *) Add prototype for temp key callback functions
6088 SSL_CTX_set_tmp_{rsa,dh}_callback().
6091 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
6092 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
6095 *) X509_name_add_entry() freed the wrong thing after an error.
6096 [Arne Ansper <arne@ats.cyber.ee>]
6098 *) rsa_eay.c would attempt to free a NULL context.
6099 [Arne Ansper <arne@ats.cyber.ee>]
6101 *) BIO_s_socket() had a broken should_retry() on Windoze.
6102 [Arne Ansper <arne@ats.cyber.ee>]
6104 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
6105 [Arne Ansper <arne@ats.cyber.ee>]
6107 *) Make sure the already existing X509_STORE->depth variable is initialized
6108 in X509_STORE_new(), but document the fact that this variable is still
6109 unused in the certificate verification process.
6110 [Ralf S. Engelschall]
6112 *) Fix the various library and apps files to free up pkeys obtained from
6113 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
6116 *) Fix reference counting in X509_PUBKEY_get(). This makes
6117 demos/maurice/example2.c work, amongst others, probably.
6118 [Steve Henson and Ben Laurie]
6120 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
6121 `openssl' and second, the shortcut symlinks for the `openssl <command>'
6122 are no longer created. This way we have a single and consistent command
6123 line interface `openssl <command>', similar to `cvs <command>'.
6124 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
6126 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
6127 BIT STRING wrapper always have zero unused bits.
6130 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
6133 *) Make the top-level INSTALL documentation easier to understand.
6136 *) Makefiles updated to exit if an error occurs in a sub-directory
6137 make (including if user presses ^C) [Paul Sutton]
6139 *) Make Montgomery context stuff explicit in RSA data structure.
6142 *) Fix build order of pem and err to allow for generated pem.h.
6145 *) Fix renumbering bug in X509_NAME_delete_entry().
6148 *) Enhanced the err-ins.pl script so it makes the error library number
6149 global and can add a library name. This is needed for external ASN1 and
6150 other error libraries.
6153 *) Fixed sk_insert which never worked properly.
6156 *) Fix ASN1 macros so they can handle indefinite length construted
6157 EXPLICIT tags. Some non standard certificates use these: they can now
6161 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
6162 into a single doc/ssleay.txt bundle. This way the information is still
6163 preserved but no longer messes up this directory. Now it's new room for
6164 the new set of documenation files.
6165 [Ralf S. Engelschall]
6167 *) SETs were incorrectly DER encoded. This was a major pain, because they
6168 shared code with SEQUENCEs, which aren't coded the same. This means that
6169 almost everything to do with SETs or SEQUENCEs has either changed name or
6170 number of arguments.
6171 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
6173 *) Fix test data to work with the above.
6176 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
6177 was already fixed by Eric for 0.9.1 it seems.
6178 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
6180 *) Autodetect FreeBSD3.
6183 *) Fix various bugs in Configure. This affects the following platforms:
6187 unixware-2.0-pentium
6191 *) Eliminate generated files from CVS. Reorder tests to regenerate files
6192 before they are needed.
6195 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
6199 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
6201 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
6202 changed SSLeay to OpenSSL in version strings.
6203 [Ralf S. Engelschall]
6205 *) Some fixups to the top-level documents.
6208 *) Fixed the nasty bug where rsaref.h was not found under compile-time
6209 because the symlink to include/ was missing.
6210 [Ralf S. Engelschall]
6212 *) Incorporated the popular no-RSA/DSA-only patches
6213 which allow to compile a RSA-free SSLeay.
6214 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
6216 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
6217 when "ssleay" is still not found.
6218 [Ralf S. Engelschall]
6220 *) Added more platforms to Configure: Cray T3E, HPUX 11,
6221 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
6223 *) Updated the README file.
6224 [Ralf S. Engelschall]
6226 *) Added various .cvsignore files in the CVS repository subdirs
6227 to make a "cvs update" really silent.
6228 [Ralf S. Engelschall]
6230 *) Recompiled the error-definition header files and added
6231 missing symbols to the Win32 linker tables.
6232 [Ralf S. Engelschall]
6234 *) Cleaned up the top-level documents;
6235 o new files: CHANGES and LICENSE
6236 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
6237 o merged COPYRIGHT into LICENSE
6238 o removed obsolete TODO file
6239 o renamed MICROSOFT to INSTALL.W32
6240 [Ralf S. Engelschall]
6242 *) Removed dummy files from the 0.9.1b source tree:
6243 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
6244 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
6245 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
6246 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
6247 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
6248 [Ralf S. Engelschall]
6250 *) Added various platform portability fixes.
6253 *) The Genesis of the OpenSSL rpject:
6254 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
6255 Young and Tim J. Hudson created while they were working for C2Net until
6257 [The OpenSSL Project]
6260 Changes between 0.9.0b and 0.9.1b [not released]
6262 *) Updated a few CA certificates under certs/
6265 *) Changed some BIGNUM api stuff.
6268 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
6269 DGUX x86, Linux Alpha, etc.
6272 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
6273 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
6277 *) Add -strparse option to asn1pars program which parses nested
6279 [Dr Stephen Henson <shenson@bigfoot.com>]
6281 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
6284 *) DSA fix for "ca" program.
6287 *) Added "-genkey" option to "dsaparam" program.
6290 *) Added RIPE MD160 (rmd160) message digest.
6293 *) Added -a (all) option to "ssleay version" command.
6296 *) Added PLATFORM define which is the id given to Configure.
6299 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
6302 *) Extended the ASN.1 parser routines.
6305 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
6308 *) Added a BN_CTX to the BN library.
6311 *) Fixed the weak key values in DES library
6314 *) Changed API in EVP library for cipher aliases.
6317 *) Added support for RC2/64bit cipher.
6320 *) Converted the lhash library to the crypto/mem.c functions.
6323 *) Added more recognized ASN.1 object ids.
6326 *) Added more RSA padding checks for SSL/TLS.
6329 *) Added BIO proxy/filter functionality.
6332 *) Added extra_certs to SSL_CTX which can be used
6333 send extra CA certificates to the client in the CA cert chain sending
6334 process. It can be configured with SSL_CTX_add_extra_chain_cert().
6337 *) Now Fortezza is denied in the authentication phase because
6338 this is key exchange mechanism is not supported by SSLeay at all.
6341 *) Additional PKCS1 checks.
6344 *) Support the string "TLSv1" for all TLS v1 ciphers.
6347 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
6348 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
6351 *) Fixed a few memory leaks.
6354 *) Fixed various code and comment typos.
6357 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
6358 bytes sent in the client random.
6359 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob