]> git.ipfire.org Git - thirdparty/openssl.git/blob - CHANGES
Update from stable branch.
[thirdparty/openssl.git] / CHANGES
1
2 OpenSSL CHANGES
3 _______________
4
5 Changes between 0.9.8j and 0.9.9 [xx XXX xxxx]
6
7 *) Update Gost ENGINE to support parameter files.
8 [Victor B. Wagner <vitus@cryptocom.ru>]
9
10 *) Support GeneralizedTime in ca utility.
11 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
12
13 *) Enhance the hash format used for certificate directory links. The new
14 form uses the canonical encoding (meaning equivalent names will work
15 even if they aren't identical) and uses SHA1 instead of MD5. This form
16 is incompatible with the older format and as a result c_rehash should
17 be used to rebuild symbolic links.
18 [Steve Henson]
19
20 *) Make PKCS#8 the default write format for private keys, replacing the
21 traditional format. This form is standardised, more secure and doesn't
22 include an implicit MD5 dependency.
23 [Steve Henson]
24
25 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
26 committed to OpenSSL should pass this lot as a minimum.
27 [Steve Henson]
28
29 *) Add session ticket override functionality for use by EAP-FAST.
30 [Jouni Malinen <j@w1.fi>]
31
32 *) Modify HMAC functions to return a value. Since these can be implemented
33 in an ENGINE errors can occur.
34 [Steve Henson]
35
36 *) Type-checked OBJ_bsearch_ex.
37 [Ben Laurie]
38
39 *) Type-checked OBJ_bsearch. Also some constification necessitated
40 by type-checking. Still to come: TXT_DB, bsearch(?),
41 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
42 CONF_VALUE.
43 [Ben Laurie]
44
45 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
46 seconds to a tm structure directly, instead of going through OS
47 specific date routines. This avoids any issues with OS routines such
48 as the year 2038 bug. New *_adj() functions for ASN1 time structures
49 and X509_time_adj_ex() to cover the extended range. The existing
50 X509_time_adj() is still usable and will no longer have any date issues.
51 [Steve Henson]
52
53 *) Delta CRL support. New use deltas option which will attempt to locate
54 and search any appropriate delta CRLs available.
55
56 This work was sponsored by Google.
57 [Steve Henson]
58
59 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
60 code and add additional score elements. Validate alternate CRL paths
61 as part of the CRL checking and indicate a new error "CRL path validation
62 error" in this case. Applications wanting additional details can use
63 the verify callback and check the new "parent" field. If this is not
64 NULL CRL path validation is taking place. Existing applications wont
65 see this because it requires extended CRL support which is off by
66 default.
67
68 This work was sponsored by Google.
69 [Steve Henson]
70
71 *) Support for freshest CRL extension.
72
73 This work was sponsored by Google.
74 [Steve Henson]
75
76 *) Initial indirect CRL support. Currently only supported in the CRLs
77 passed directly and not via lookup. Process certificate issuer
78 CRL entry extension and lookup CRL entries by bother issuer name
79 and serial number. Check and process CRL issuer entry in IDP extension.
80
81 This work was sponsored by Google.
82 [Steve Henson]
83
84 *) Add support for distinct certificate and CRL paths. The CRL issuer
85 certificate is validated separately in this case. Only enabled if
86 an extended CRL support flag is set: this flag will enable additional
87 CRL functionality in future.
88
89 This work was sponsored by Google.
90 [Steve Henson]
91
92 *) Add support for policy mappings extension.
93
94 This work was sponsored by Google.
95 [Steve Henson]
96
97 *) Fixes to pathlength constraint, self issued certificate handling,
98 policy processing to align with RFC3280 and PKITS tests.
99
100 This work was sponsored by Google.
101 [Steve Henson]
102
103 *) Support for name constraints certificate extension. DN, email, DNS
104 and URI types are currently supported.
105
106 This work was sponsored by Google.
107 [Steve Henson]
108
109 *) To cater for systems that provide a pointer-based thread ID rather
110 than numeric, deprecate the current numeric thread ID mechanism and
111 replace it with a structure and associated callback type. This
112 mechanism allows a numeric "hash" to be extracted from a thread ID in
113 either case, and on platforms where pointers are larger than 'long',
114 mixing is done to help ensure the numeric 'hash' is usable even if it
115 can't be guaranteed unique. The default mechanism is to use "&errno"
116 as a pointer-based thread ID to distinguish between threads.
117
118 Applications that want to provide their own thread IDs should now use
119 CRYPTO_THREADID_set_callback() to register a callback that will call
120 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
121
122 Note that ERR_remove_state() is now deprecated, because it is tied
123 to the assumption that thread IDs are numeric. ERR_remove_state(0)
124 to free the current thread's error state should be replaced by
125 ERR_remove_thread_state(NULL).
126
127 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
128 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
129 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
130 application was previously providing a numeric thread callback that
131 was inappropriate for distinguishing threads, then uniqueness might
132 have been obtained with &errno that happened immediately in the
133 intermediate development versions of OpenSSL; this is no longer the
134 case, the numeric thread callback will now override the automatic use
135 of &errno.)
136 [Geoff Thorpe, with help from Bodo Moeller]
137
138 *) Initial support for different CRL issuing certificates. This covers a
139 simple case where the self issued certificates in the chain exist and
140 the real CRL issuer is higher in the existing chain.
141
142 This work was sponsored by Google.
143 [Steve Henson]
144
145 *) Removed effectively defunct crypto/store from the build.
146 [Ben Laurie]
147
148 *) Revamp of STACK to provide stronger type-checking. Still to come:
149 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
150 ASN1_STRING, CONF_VALUE.
151 [Ben Laurie]
152
153 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
154 RAM on SSL connections. This option can save about 34k per idle SSL.
155 [Nick Mathewson]
156
157 *) Revamp of LHASH to provide stronger type-checking. Still to come:
158 STACK, TXT_DB, bsearch, qsort.
159 [Ben Laurie]
160
161 *) Initial support for Cryptographic Message Syntax (aka CMS) based
162 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
163 support for data, signedData, compressedData, digestedData and
164 encryptedData, envelopedData types included. Scripts to check against
165 RFC4134 examples draft and interop and consistency checks of many
166 content types and variants.
167 [Steve Henson]
168
169 *) Add options to enc utility to support use of zlib compression BIO.
170 [Steve Henson]
171
172 *) Extend mk1mf to support importing of options and assembly language
173 files from Configure script, currently only included in VC-WIN32.
174 The assembly language rules can now optionally generate the source
175 files from the associated perl scripts.
176 [Steve Henson]
177
178 *) Implement remaining functionality needed to support GOST ciphersuites.
179 Interop testing has been performed using CryptoPro implementations.
180 [Victor B. Wagner <vitus@cryptocom.ru>]
181
182 *) s390x assembler pack.
183 [Andy Polyakov]
184
185 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
186 "family."
187 [Andy Polyakov]
188
189 *) Implement Opaque PRF Input TLS extension as specified in
190 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
191 official specification yet and no extension type assignment by
192 IANA exists, this extension (for now) will have to be explicitly
193 enabled when building OpenSSL by providing the extension number
194 to use. For example, specify an option
195
196 -DTLSEXT_TYPE_opaque_prf_input=0x9527
197
198 to the "config" or "Configure" script to enable the extension,
199 assuming extension number 0x9527 (which is a completely arbitrary
200 and unofficial assignment based on the MD5 hash of the Internet
201 Draft). Note that by doing so, you potentially lose
202 interoperability with other TLS implementations since these might
203 be using the same extension number for other purposes.
204
205 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
206 opaque PRF input value to use in the handshake. This will create
207 an interal copy of the length-'len' string at 'src', and will
208 return non-zero for success.
209
210 To get more control and flexibility, provide a callback function
211 by using
212
213 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
214 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
215
216 where
217
218 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
219 void *arg;
220
221 Callback function 'cb' will be called in handshakes, and is
222 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
223 Argument 'arg' is for application purposes (the value as given to
224 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
225 be provided to the callback function). The callback function
226 has to return non-zero to report success: usually 1 to use opaque
227 PRF input just if possible, or 2 to enforce use of the opaque PRF
228 input. In the latter case, the library will abort the handshake
229 if opaque PRF input is not successfully negotiated.
230
231 Arguments 'peerinput' and 'len' given to the callback function
232 will always be NULL and 0 in the case of a client. A server will
233 see the client's opaque PRF input through these variables if
234 available (NULL and 0 otherwise). Note that if the server
235 provides an opaque PRF input, the length must be the same as the
236 length of the client's opaque PRF input.
237
238 Note that the callback function will only be called when creating
239 a new session (session resumption can resume whatever was
240 previously negotiated), and will not be called in SSL 2.0
241 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
242 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
243 for applications that need to enforce opaque PRF input.
244
245 [Bodo Moeller]
246
247 *) Update ssl code to support digests other than SHA1+MD5 for handshake
248 MAC.
249
250 [Victor B. Wagner <vitus@cryptocom.ru>]
251
252 *) Add RFC4507 support to OpenSSL. This includes the corrections in
253 RFC4507bis. The encrypted ticket format is an encrypted encoded
254 SSL_SESSION structure, that way new session features are automatically
255 supported.
256
257 If a client application caches session in an SSL_SESSION structure
258 support is transparent because tickets are now stored in the encoded
259 SSL_SESSION.
260
261 The SSL_CTX structure automatically generates keys for ticket
262 protection in servers so again support should be possible
263 with no application modification.
264
265 If a client or server wishes to disable RFC4507 support then the option
266 SSL_OP_NO_TICKET can be set.
267
268 Add a TLS extension debugging callback to allow the contents of any client
269 or server extensions to be examined.
270
271 This work was sponsored by Google.
272 [Steve Henson]
273
274 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
275 OpenSSL should now compile cleanly on gcc 4.2
276 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
277
278 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
279 support including streaming MAC support: this is required for GOST
280 ciphersuite support.
281 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
282
283 *) Add option -stream to use PKCS#7 streaming in smime utility. New
284 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
285 to output in BER and PEM format.
286 [Steve Henson]
287
288 *) Experimental support for use of HMAC via EVP_PKEY interface. This
289 allows HMAC to be handled via the EVP_DigestSign*() interface. The
290 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
291 ENGINE support for HMAC keys which are unextractable. New -mac and
292 -macopt options to dgst utility.
293 [Steve Henson]
294
295 *) New option -sigopt to dgst utility. Update dgst to use
296 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
297 alternative signing paramaters such as X9.31 or PSS in the dgst
298 utility.
299 [Steve Henson]
300
301 *) Change ssl_cipher_apply_rule(), the internal function that does
302 the work each time a ciphersuite string requests enabling
303 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
304 removing ("!foo+bar") a class of ciphersuites: Now it maintains
305 the order of disabled ciphersuites such that those ciphersuites
306 that most recently went from enabled to disabled not only stay
307 in order with respect to each other, but also have higher priority
308 than other disabled ciphersuites the next time ciphersuites are
309 enabled again.
310
311 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
312 the same ciphersuites as with "HIGH" alone, but in a specific
313 order where the PSK ciphersuites come first (since they are the
314 most recently disabled ciphersuites when "HIGH" is parsed).
315
316 Also, change ssl_create_cipher_list() (using this new
317 funcionality) such that between otherwise identical
318 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
319 the default order.
320 [Bodo Moeller]
321
322 *) Change ssl_create_cipher_list() so that it automatically
323 arranges the ciphersuites in reasonable order before starting
324 to process the rule string. Thus, the definition for "DEFAULT"
325 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
326 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
327 This makes it much easier to arrive at a reasonable default order
328 in applications for which anonymous ciphers are OK (meaning
329 that you can't actually use DEFAULT).
330 [Bodo Moeller; suggested by Victor Duchovni]
331
332 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
333 processing) into multiple integers instead of setting
334 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
335 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
336 (These masks as well as the individual bit definitions are hidden
337 away into the non-exported interface ssl/ssl_locl.h, so this
338 change to the definition of the SSL_CIPHER structure shouldn't
339 affect applications.) This give us more bits for each of these
340 categories, so there is no longer a need to coagulate AES128 and
341 AES256 into a single algorithm bit, and to coagulate Camellia128
342 and Camellia256 into a single algorithm bit, which has led to all
343 kinds of kludges.
344
345 Thus, among other things, the kludge introduced in 0.9.7m and
346 0.9.8e for masking out AES256 independently of AES128 or masking
347 out Camellia256 independently of AES256 is not needed here in 0.9.9.
348
349 With the change, we also introduce new ciphersuite aliases that
350 so far were missing: "AES128", "AES256", "CAMELLIA128", and
351 "CAMELLIA256".
352 [Bodo Moeller]
353
354 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
355 Use the leftmost N bytes of the signature input if the input is
356 larger than the prime q (with N being the size in bytes of q).
357 [Nils Larsch]
358
359 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
360 it yet and it is largely untested.
361 [Steve Henson]
362
363 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
364 [Nils Larsch]
365
366 *) Initial incomplete changes to avoid need for function casts in OpenSSL
367 some compilers (gcc 4.2 and later) reject their use. Safestack is
368 reimplemented. Update ASN1 to avoid use of legacy functions.
369 [Steve Henson]
370
371 *) Win32/64 targets are linked with Winsock2.
372 [Andy Polyakov]
373
374 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
375 to external functions. This can be used to increase CRL handling
376 efficiency especially when CRLs are very large by (for example) storing
377 the CRL revoked certificates in a database.
378 [Steve Henson]
379
380 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
381 new CRLs added to a directory can be used. New command line option
382 -verify_return_error to s_client and s_server. This causes real errors
383 to be returned by the verify callback instead of carrying on no matter
384 what. This reflects the way a "real world" verify callback would behave.
385 [Steve Henson]
386
387 *) GOST engine, supporting several GOST algorithms and public key formats.
388 Kindly donated by Cryptocom.
389 [Cryptocom]
390
391 *) Partial support for Issuing Distribution Point CRL extension. CRLs
392 partitioned by DP are handled but no indirect CRL or reason partitioning
393 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
394 selected via a scoring technique which handles IDP and AKID in CRLs.
395 [Steve Henson]
396
397 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
398 will ultimately be used for all verify operations: this will remove the
399 X509_STORE dependency on certificate verification and allow alternative
400 lookup methods. X509_STORE based implementations of these two callbacks.
401 [Steve Henson]
402
403 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
404 Modify get_crl() to find a valid (unexpired) CRL if possible.
405 [Steve Henson]
406
407 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
408 this would be called X509_CRL_cmp() but that name is already used by
409 a function that just compares CRL issuer names. Cache several CRL
410 extensions in X509_CRL structure and cache CRLDP in X509.
411 [Steve Henson]
412
413 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
414 this maps equivalent X509_NAME structures into a consistent structure.
415 Name comparison can then be performed rapidly using memcmp().
416 [Steve Henson]
417
418 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
419 utility.
420 [Steve Henson]
421
422 *) Allow digests to supply their own micalg string for S/MIME type using
423 the ctrl EVP_MD_CTRL_MICALG.
424 [Steve Henson]
425
426 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
427 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
428 ctrl. It can then customise the structure before and/or after signing
429 if necessary.
430 [Steve Henson]
431
432 *) New function OBJ_add_sigid() to allow application defined signature OIDs
433 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
434 to free up any added signature OIDs.
435 [Steve Henson]
436
437 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
438 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
439 digest and cipher tables. New options added to openssl utility:
440 list-message-digest-algorithms and list-cipher-algorithms.
441 [Steve Henson]
442
443 *) Change the array representation of binary polynomials: the list
444 of degrees of non-zero coefficients is now terminated with -1.
445 Previously it was terminated with 0, which was also part of the
446 value; thus, the array representation was not applicable to
447 polynomials where t^0 has coefficient zero. This change makes
448 the array representation useful in a more general context.
449 [Douglas Stebila]
450
451 *) Various modifications and fixes to SSL/TLS cipher string
452 handling. For ECC, the code now distinguishes between fixed ECDH
453 with RSA certificates on the one hand and with ECDSA certificates
454 on the other hand, since these are separate ciphersuites. The
455 unused code for Fortezza ciphersuites has been removed.
456
457 For consistency with EDH, ephemeral ECDH is now called "EECDH"
458 (not "ECDHE"). For consistency with the code for DH
459 certificates, use of ECDH certificates is now considered ECDH
460 authentication, not RSA or ECDSA authentication (the latter is
461 merely the CA's signing algorithm and not actively used in the
462 protocol).
463
464 The temporary ciphersuite alias "ECCdraft" is no longer
465 available, and ECC ciphersuites are no longer excluded from "ALL"
466 and "DEFAULT". The following aliases now exist for RFC 4492
467 ciphersuites, most of these by analogy with the DH case:
468
469 kECDHr - ECDH cert, signed with RSA
470 kECDHe - ECDH cert, signed with ECDSA
471 kECDH - ECDH cert (signed with either RSA or ECDSA)
472 kEECDH - ephemeral ECDH
473 ECDH - ECDH cert or ephemeral ECDH
474
475 aECDH - ECDH cert
476 aECDSA - ECDSA cert
477 ECDSA - ECDSA cert
478
479 AECDH - anonymous ECDH
480 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
481
482 [Bodo Moeller]
483
484 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
485 Use correct micalg parameters depending on digest(s) in signed message.
486 [Steve Henson]
487
488 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
489 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
490 [Steve Henson]
491
492 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
493 an engine to register a method. Add ENGINE lookups for methods and
494 functional reference processing.
495 [Steve Henson]
496
497 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
498 EVP_{Sign,Verify}* which allow an application to customise the signature
499 process.
500 [Steve Henson]
501
502 *) New -resign option to smime utility. This adds one or more signers
503 to an existing PKCS#7 signedData structure. Also -md option to use an
504 alternative message digest algorithm for signing.
505 [Steve Henson]
506
507 *) Tidy up PKCS#7 routines and add new functions to make it easier to
508 create PKCS7 structures containing multiple signers. Update smime
509 application to support multiple signers.
510 [Steve Henson]
511
512 *) New -macalg option to pkcs12 utility to allow setting of an alternative
513 digest MAC.
514 [Steve Henson]
515
516 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
517 Reorganize PBE internals to lookup from a static table using NIDs,
518 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
519 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
520 PRF which will be automatically used with PBES2.
521 [Steve Henson]
522
523 *) Replace the algorithm specific calls to generate keys in "req" with the
524 new API.
525 [Steve Henson]
526
527 *) Update PKCS#7 enveloped data routines to use new API. This is now
528 supported by any public key method supporting the encrypt operation. A
529 ctrl is added to allow the public key algorithm to examine or modify
530 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
531 a no op.
532 [Steve Henson]
533
534 *) Add a ctrl to asn1 method to allow a public key algorithm to express
535 a default digest type to use. In most cases this will be SHA1 but some
536 algorithms (such as GOST) need to specify an alternative digest. The
537 return value indicates how strong the prefernce is 1 means optional and
538 2 is mandatory (that is it is the only supported type). Modify
539 ASN1_item_sign() to accept a NULL digest argument to indicate it should
540 use the default md. Update openssl utilities to use the default digest
541 type for signing if it is not explicitly indicated.
542 [Steve Henson]
543
544 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
545 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
546 signing method from the key type. This effectively removes the link
547 between digests and public key types.
548 [Steve Henson]
549
550 *) Add an OID cross reference table and utility functions. Its purpose is to
551 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
552 rsaEncryption. This will allow some of the algorithm specific hackery
553 needed to use the correct OID to be removed.
554 [Steve Henson]
555
556 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
557 structures for PKCS7_sign(). They are now set up by the relevant public
558 key ASN1 method.
559 [Steve Henson]
560
561 *) Add provisional EC pkey method with support for ECDSA and ECDH.
562 [Steve Henson]
563
564 *) Add support for key derivation (agreement) in the API, DH method and
565 pkeyutl.
566 [Steve Henson]
567
568 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
569 public and private key formats. As a side effect these add additional
570 command line functionality not previously available: DSA signatures can be
571 generated and verified using pkeyutl and DH key support and generation in
572 pkey, genpkey.
573 [Steve Henson]
574
575 *) BeOS support.
576 [Oliver Tappe <zooey@hirschkaefer.de>]
577
578 *) New make target "install_html_docs" installs HTML renditions of the
579 manual pages.
580 [Oliver Tappe <zooey@hirschkaefer.de>]
581
582 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
583 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
584 support key and parameter generation and add initial key generation
585 functionality for RSA.
586 [Steve Henson]
587
588 *) Add functions for main EVP_PKEY_method operations. The undocumented
589 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
590 EVP_PKEY_{encrypt,decrypt}_old.
591 [Steve Henson]
592
593 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
594 key API, doesn't do much yet.
595 [Steve Henson]
596
597 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
598 public key algorithms. New option to openssl utility:
599 "list-public-key-algorithms" to print out info.
600 [Steve Henson]
601
602 *) Implement the Supported Elliptic Curves Extension for
603 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
604 [Douglas Stebila]
605
606 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
607 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
608 [Steve Henson]
609
610 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
611 utilities such as rsa, dsa, dsaparam etc except they process any key
612 type.
613 [Steve Henson]
614
615 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
616 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
617 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
618 structure.
619 [Steve Henson]
620
621 *) Initial support for pluggable public key ASN1.
622 De-spaghettify the public key ASN1 handling. Move public and private
623 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
624 algorithm specific handling to a single module within the relevant
625 algorithm directory. Add functions to allow (near) opaque processing
626 of public and private key structures.
627 [Steve Henson]
628
629 *) Implement the Supported Point Formats Extension for
630 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
631 [Douglas Stebila]
632
633 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
634 for the psk identity [hint] and the psk callback functions to the
635 SSL_SESSION, SSL and SSL_CTX structure.
636
637 New ciphersuites:
638 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
639 PSK-AES256-CBC-SHA
640
641 New functions:
642 SSL_CTX_use_psk_identity_hint
643 SSL_get_psk_identity_hint
644 SSL_get_psk_identity
645 SSL_use_psk_identity_hint
646
647 [Mika Kousa and Pasi Eronen of Nokia Corporation]
648
649 *) Add RFC 3161 compliant time stamp request creation, response generation
650 and response verification functionality.
651 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
652
653 *) Add initial support for TLS extensions, specifically for the server_name
654 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
655 have new members for a host name. The SSL data structure has an
656 additional member SSL_CTX *initial_ctx so that new sessions can be
657 stored in that context to allow for session resumption, even after the
658 SSL has been switched to a new SSL_CTX in reaction to a client's
659 server_name extension.
660
661 New functions (subject to change):
662
663 SSL_get_servername()
664 SSL_get_servername_type()
665 SSL_set_SSL_CTX()
666
667 New CTRL codes and macros (subject to change):
668
669 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
670 - SSL_CTX_set_tlsext_servername_callback()
671 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
672 - SSL_CTX_set_tlsext_servername_arg()
673 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
674
675 openssl s_client has a new '-servername ...' option.
676
677 openssl s_server has new options '-servername_host ...', '-cert2 ...',
678 '-key2 ...', '-servername_fatal' (subject to change). This allows
679 testing the HostName extension for a specific single host name ('-cert'
680 and '-key' remain fallbacks for handshakes without HostName
681 negotiation). If the unrecogninzed_name alert has to be sent, this by
682 default is a warning; it becomes fatal with the '-servername_fatal'
683 option.
684
685 [Peter Sylvester, Remy Allais, Christophe Renou]
686
687 *) Whirlpool hash implementation is added.
688 [Andy Polyakov]
689
690 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
691 bn(64,32). Because of instruction set limitations it doesn't have
692 any negative impact on performance. This was done mostly in order
693 to make it possible to share assembler modules, such as bn_mul_mont
694 implementations, between 32- and 64-bit builds without hassle.
695 [Andy Polyakov]
696
697 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
698 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
699 macro.
700 [Bodo Moeller]
701
702 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
703 dedicated Montgomery multiplication procedure, is introduced.
704 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
705 "64-bit" performance on certain 32-bit targets.
706 [Andy Polyakov]
707
708 *) New option SSL_OP_NO_COMP to disable use of compression selectively
709 in SSL structures. New SSL ctrl to set maximum send fragment size.
710 Save memory by seeting the I/O buffer sizes dynamically instead of
711 using the maximum available value.
712 [Steve Henson]
713
714 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
715 in addition to the text details.
716 [Bodo Moeller]
717
718 *) Very, very preliminary EXPERIMENTAL support for printing of general
719 ASN1 structures. This currently produces rather ugly output and doesn't
720 handle several customised structures at all.
721 [Steve Henson]
722
723 *) Integrated support for PVK file format and some related formats such
724 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
725 these in the 'rsa' and 'dsa' utilities.
726 [Steve Henson]
727
728 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
729 [Steve Henson]
730
731 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
732 place for the (very old) "NETSCAPE" format certificates which are now
733 handled using new ASN1 code equivalents.
734 [Steve Henson]
735
736 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
737 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
738 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
739 [Nils Larsch]
740
741 *) Modify CRL distribution points extension code to print out previously
742 unsupported fields. Enhance extension setting code to allow setting of
743 all fields.
744 [Steve Henson]
745
746 *) Add print and set support for Issuing Distribution Point CRL extension.
747 [Steve Henson]
748
749 *) Change 'Configure' script to enable Camellia by default.
750 [NTT]
751
752 Changes between 0.9.8j and 0.9.8k [xx XXX xxxx]
753
754 *) Reject UniversalString and BMPString types with invalid lengths. This
755 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
756 a legal length. (CVE-2009-0590)
757 [Steve Henson]
758
759 *) Set S/MIME signing as the default purpose rather than setting it
760 unconditionally. This allows applications to override it at the store
761 level.
762 [Steve Henson]
763
764 *) Permit restricted recursion of ASN1 strings. This is needed in practice
765 to handle some structures.
766 [Steve Henson]
767
768 *) Improve efficiency of mem_gets: don't search whole buffer each time
769 for a '\n'
770 [Jeremy Shapiro <jnshapir@us.ibm.com>]
771
772 *) New -hex option for openssl rand.
773 [Matthieu Herrb]
774
775 *) Print out UTF8String and NumericString when parsing ASN1.
776 [Steve Henson]
777
778 *) Support NumericString type for name components.
779 [Steve Henson]
780
781 *) Allow CC in the environment to override the automatically chosen
782 compiler. Note that nothing is done to ensure flags work with the
783 chosen compiler.
784 [Ben Laurie]
785
786 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
787
788 *) Properly check EVP_VerifyFinal() and similar return values
789 (CVE-2008-5077).
790 [Ben Laurie, Bodo Moeller, Google Security Team]
791
792 *) Enable TLS extensions by default.
793 [Ben Laurie]
794
795 *) Allow the CHIL engine to be loaded, whether the application is
796 multithreaded or not. (This does not release the developer from the
797 obligation to set up the dynamic locking callbacks.)
798 [Sander Temme <sander@temme.net>]
799
800 *) Use correct exit code if there is an error in dgst command.
801 [Steve Henson; problem pointed out by Roland Dirlewanger]
802
803 *) Tweak Configure so that you need to say "experimental-jpake" to enable
804 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
805 [Bodo Moeller]
806
807 *) Add experimental JPAKE support, including demo authentication in
808 s_client and s_server.
809 [Ben Laurie]
810
811 *) Set the comparison function in v3_addr_canonize().
812 [Rob Austein <sra@hactrn.net>]
813
814 *) Add support for XMPP STARTTLS in s_client.
815 [Philip Paeps <philip@freebsd.org>]
816
817 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
818 to ensure that even with this option, only ciphersuites in the
819 server's preference list will be accepted. (Note that the option
820 applies only when resuming a session, so the earlier behavior was
821 just about the algorithm choice for symmetric cryptography.)
822 [Bodo Moeller]
823
824 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
825
826 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
827 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
828 [Nagendra Modadugu]
829
830 *) The fix in 0.9.8c that supposedly got rid of unsafe
831 double-checked locking was incomplete for RSA blinding,
832 addressing just one layer of what turns out to have been
833 doubly unsafe triple-checked locking.
834
835 So now fix this for real by retiring the MONT_HELPER macro
836 in crypto/rsa/rsa_eay.c.
837
838 [Bodo Moeller; problem pointed out by Marius Schilder]
839
840 *) Various precautionary measures:
841
842 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
843
844 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
845 (NB: This would require knowledge of the secret session ticket key
846 to exploit, in which case you'd be SOL either way.)
847
848 - Change bn_nist.c so that it will properly handle input BIGNUMs
849 outside the expected range.
850
851 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
852 builds.
853
854 [Neel Mehta, Bodo Moeller]
855
856 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
857 the load fails. Useful for distros.
858 [Ben Laurie and the FreeBSD team]
859
860 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
861 [Steve Henson]
862
863 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
864 [Huang Ying]
865
866 *) Expand ENGINE to support engine supplied SSL client certificate functions.
867
868 This work was sponsored by Logica.
869 [Steve Henson]
870
871 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
872 keystores. Support for SSL/TLS client authentication too.
873 Not compiled unless enable-capieng specified to Configure.
874
875 This work was sponsored by Logica.
876 [Steve Henson]
877
878 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
879 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
880 attribute creation routines such as certifcate requests and PKCS#12
881 files.
882 [Steve Henson]
883
884 Changes between 0.9.8g and 0.9.8h [28 May 2008]
885
886 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
887 handshake which could lead to a cilent crash as found using the
888 Codenomicon TLS test suite (CVE-2008-1672)
889 [Steve Henson, Mark Cox]
890
891 *) Fix double free in TLS server name extensions which could lead to
892 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
893 [Joe Orton]
894
895 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
896
897 Clear the error queue to ensure that error entries left from
898 older function calls do not interfere with the correct operation.
899 [Lutz Jaenicke, Erik de Castro Lopo]
900
901 *) Remove root CA certificates of commercial CAs:
902
903 The OpenSSL project does not recommend any specific CA and does not
904 have any policy with respect to including or excluding any CA.
905 Therefore it does not make any sense to ship an arbitrary selection
906 of root CA certificates with the OpenSSL software.
907 [Lutz Jaenicke]
908
909 *) RSA OAEP patches to fix two separate invalid memory reads.
910 The first one involves inputs when 'lzero' is greater than
911 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
912 before the beginning of from). The second one involves inputs where
913 the 'db' section contains nothing but zeroes (there is a one-byte
914 invalid read after the end of 'db').
915 [Ivan Nestlerode <inestlerode@us.ibm.com>]
916
917 *) Partial backport from 0.9.9-dev:
918
919 Introduce bn_mul_mont (dedicated Montgomery multiplication
920 procedure) as a candidate for BIGNUM assembler implementation.
921 While 0.9.9-dev uses assembler for various architectures, only
922 x86_64 is available by default here in the 0.9.8 branch, and
923 32-bit x86 is available through a compile-time setting.
924
925 To try the 32-bit x86 assembler implementation, use Configure
926 option "enable-montasm" (which exists only for this backport).
927
928 As "enable-montasm" for 32-bit x86 disclaims code stability
929 anyway, in this constellation we activate additional code
930 backported from 0.9.9-dev for further performance improvements,
931 namely BN_from_montgomery_word. (To enable this otherwise,
932 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
933
934 [Andy Polyakov (backport partially by Bodo Moeller)]
935
936 *) Add TLS session ticket callback. This allows an application to set
937 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
938 values. This is useful for key rollover for example where several key
939 sets may exist with different names.
940 [Steve Henson]
941
942 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
943 This was broken until now in 0.9.8 releases, such that the only way
944 a registered ENGINE could be used (assuming it initialises
945 successfully on the host) was to explicitly set it as the default
946 for the relevant algorithms. This is in contradiction with 0.9.7
947 behaviour and the documentation. With this fix, when an ENGINE is
948 registered into a given algorithm's table of implementations, the
949 'uptodate' flag is reset so that auto-discovery will be used next
950 time a new context for that algorithm attempts to select an
951 implementation.
952 [Ian Lister (tweaked by Geoff Thorpe)]
953
954 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
955 implemention in the following ways:
956
957 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
958 hard coded.
959
960 Lack of BER streaming support means one pass streaming processing is
961 only supported if data is detached: setting the streaming flag is
962 ignored for embedded content.
963
964 CMS support is disabled by default and must be explicitly enabled
965 with the enable-cms configuration option.
966 [Steve Henson]
967
968 *) Update the GMP engine glue to do direct copies between BIGNUM and
969 mpz_t when openssl and GMP use the same limb size. Otherwise the
970 existing "conversion via a text string export" trick is still used.
971 [Paul Sheer <paulsheer@gmail.com>]
972
973 *) Zlib compression BIO. This is a filter BIO which compressed and
974 uncompresses any data passed through it.
975 [Steve Henson]
976
977 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
978 RFC3394 compatible AES key wrapping.
979 [Steve Henson]
980
981 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
982 sets string data without copying. X509_ALGOR_set0() and
983 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
984 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
985 from an X509_ATTRIBUTE structure optionally checking it occurs only
986 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
987 data.
988 [Steve Henson]
989
990 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
991 to get the expected BN_FLG_CONSTTIME behavior.
992 [Bodo Moeller (Google)]
993
994 *) Netware support:
995
996 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
997 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
998 - added some more tests to do_tests.pl
999 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1000 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1001 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1002 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1003 - various changes to netware.pl to enable gcc-cross builds on Win32
1004 platform
1005 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1006 - various changes to fix missing prototype warnings
1007 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1008 - added AES, WHIRLPOOL and CPUID assembler code to build files
1009 - added missing AES assembler make rules to mk1mf.pl
1010 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1011 [Guenter Knauf <eflash@gmx.net>]
1012
1013 *) Implement certificate status request TLS extension defined in RFC3546.
1014 A client can set the appropriate parameters and receive the encoded
1015 OCSP response via a callback. A server can query the supplied parameters
1016 and set the encoded OCSP response in the callback. Add simplified examples
1017 to s_client and s_server.
1018 [Steve Henson]
1019
1020 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1021
1022 *) Fix various bugs:
1023 + Binary incompatibility of ssl_ctx_st structure
1024 + DTLS interoperation with non-compliant servers
1025 + Don't call get_session_cb() without proposed session
1026 + Fix ia64 assembler code
1027 [Andy Polyakov, Steve Henson]
1028
1029 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1030
1031 *) DTLS Handshake overhaul. There were longstanding issues with
1032 OpenSSL DTLS implementation, which were making it impossible for
1033 RFC 4347 compliant client to communicate with OpenSSL server.
1034 Unfortunately just fixing these incompatibilities would "cut off"
1035 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1036 server keeps tolerating non RFC compliant syntax. The opposite is
1037 not true, 0.9.8f client can not communicate with earlier server.
1038 This update even addresses CVE-2007-4995.
1039 [Andy Polyakov]
1040
1041 *) Changes to avoid need for function casts in OpenSSL: some compilers
1042 (gcc 4.2 and later) reject their use.
1043 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1044 Steve Henson]
1045
1046 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1047 RFC4507bis. The encrypted ticket format is an encrypted encoded
1048 SSL_SESSION structure, that way new session features are automatically
1049 supported.
1050
1051 If a client application caches session in an SSL_SESSION structure
1052 support is transparent because tickets are now stored in the encoded
1053 SSL_SESSION.
1054
1055 The SSL_CTX structure automatically generates keys for ticket
1056 protection in servers so again support should be possible
1057 with no application modification.
1058
1059 If a client or server wishes to disable RFC4507 support then the option
1060 SSL_OP_NO_TICKET can be set.
1061
1062 Add a TLS extension debugging callback to allow the contents of any client
1063 or server extensions to be examined.
1064
1065 This work was sponsored by Google.
1066 [Steve Henson]
1067
1068 *) Add initial support for TLS extensions, specifically for the server_name
1069 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1070 have new members for a host name. The SSL data structure has an
1071 additional member SSL_CTX *initial_ctx so that new sessions can be
1072 stored in that context to allow for session resumption, even after the
1073 SSL has been switched to a new SSL_CTX in reaction to a client's
1074 server_name extension.
1075
1076 New functions (subject to change):
1077
1078 SSL_get_servername()
1079 SSL_get_servername_type()
1080 SSL_set_SSL_CTX()
1081
1082 New CTRL codes and macros (subject to change):
1083
1084 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1085 - SSL_CTX_set_tlsext_servername_callback()
1086 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1087 - SSL_CTX_set_tlsext_servername_arg()
1088 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1089
1090 openssl s_client has a new '-servername ...' option.
1091
1092 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1093 '-key2 ...', '-servername_fatal' (subject to change). This allows
1094 testing the HostName extension for a specific single host name ('-cert'
1095 and '-key' remain fallbacks for handshakes without HostName
1096 negotiation). If the unrecogninzed_name alert has to be sent, this by
1097 default is a warning; it becomes fatal with the '-servername_fatal'
1098 option.
1099
1100 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1101
1102 *) Add AES and SSE2 assembly language support to VC++ build.
1103 [Steve Henson]
1104
1105 *) Mitigate attack on final subtraction in Montgomery reduction.
1106 [Andy Polyakov]
1107
1108 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1109 (which previously caused an internal error).
1110 [Bodo Moeller]
1111
1112 *) Squeeze another 10% out of IGE mode when in != out.
1113 [Ben Laurie]
1114
1115 *) AES IGE mode speedup.
1116 [Dean Gaudet (Google)]
1117
1118 *) Add the Korean symmetric 128-bit cipher SEED (see
1119 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1120 add SEED ciphersuites from RFC 4162:
1121
1122 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1123 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1124 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1125 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1126
1127 To minimize changes between patchlevels in the OpenSSL 0.9.8
1128 series, SEED remains excluded from compilation unless OpenSSL
1129 is configured with 'enable-seed'.
1130 [KISA, Bodo Moeller]
1131
1132 *) Mitigate branch prediction attacks, which can be practical if a
1133 single processor is shared, allowing a spy process to extract
1134 information. For detailed background information, see
1135 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1136 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
1137 and Necessary Software Countermeasures"). The core of the change
1138 are new versions BN_div_no_branch() and
1139 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
1140 respectively, which are slower, but avoid the security-relevant
1141 conditional branches. These are automatically called by BN_div()
1142 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
1143 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
1144 remove a conditional branch.
1145
1146 BN_FLG_CONSTTIME is the new name for the previous
1147 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
1148 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
1149 in the exponent causes BN_mod_exp_mont() to use the alternative
1150 implementation in BN_mod_exp_mont_consttime().) The old name
1151 remains as a deprecated alias.
1152
1153 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
1154 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
1155 constant-time implementations for more than just exponentiation.
1156 Here too the old name is kept as a deprecated alias.
1157
1158 BN_BLINDING_new() will now use BN_dup() for the modulus so that
1159 the BN_BLINDING structure gets an independent copy of the
1160 modulus. This means that the previous "BIGNUM *m" argument to
1161 BN_BLINDING_new() and to BN_BLINDING_create_param() now
1162 essentially becomes "const BIGNUM *m", although we can't actually
1163 change this in the header file before 0.9.9. It allows
1164 RSA_setup_blinding() to use BN_with_flags() on the modulus to
1165 enable BN_FLG_CONSTTIME.
1166
1167 [Matthew D Wood (Intel Corp)]
1168
1169 *) In the SSL/TLS server implementation, be strict about session ID
1170 context matching (which matters if an application uses a single
1171 external cache for different purposes). Previously,
1172 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1173 set. This did ensure strict client verification, but meant that,
1174 with applications using a single external cache for quite
1175 different requirements, clients could circumvent ciphersuite
1176 restrictions for a given session ID context by starting a session
1177 in a different context.
1178 [Bodo Moeller]
1179
1180 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1181 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1182 authentication-only ciphersuites.
1183 [Bodo Moeller]
1184
1185 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
1186 not complete and could lead to a possible single byte overflow
1187 (CVE-2007-5135) [Ben Laurie]
1188
1189 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
1190
1191 *) Since AES128 and AES256 (and similarly Camellia128 and
1192 Camellia256) share a single mask bit in the logic of
1193 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1194 kludge to work properly if AES128 is available and AES256 isn't
1195 (or if Camellia128 is available and Camellia256 isn't).
1196 [Victor Duchovni]
1197
1198 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
1199 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
1200 When a point or a seed is encoded in a BIT STRING, we need to
1201 prevent the removal of trailing zero bits to get the proper DER
1202 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
1203 of a NamedBitList, for which trailing 0 bits need to be removed.)
1204 [Bodo Moeller]
1205
1206 *) Have SSL/TLS server implementation tolerate "mismatched" record
1207 protocol version while receiving ClientHello even if the
1208 ClientHello is fragmented. (The server can't insist on the
1209 particular protocol version it has chosen before the ServerHello
1210 message has informed the client about his choice.)
1211 [Bodo Moeller]
1212
1213 *) Add RFC 3779 support.
1214 [Rob Austein for ARIN, Ben Laurie]
1215
1216 *) Load error codes if they are not already present instead of using a
1217 static variable. This allows them to be cleanly unloaded and reloaded.
1218 Improve header file function name parsing.
1219 [Steve Henson]
1220
1221 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
1222 or CAPABILITY handshake as required by RFCs.
1223 [Goetz Babin-Ebell]
1224
1225 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
1226
1227 *) Introduce limits to prevent malicious keys being able to
1228 cause a denial of service. (CVE-2006-2940)
1229 [Steve Henson, Bodo Moeller]
1230
1231 *) Fix ASN.1 parsing of certain invalid structures that can result
1232 in a denial of service. (CVE-2006-2937) [Steve Henson]
1233
1234 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1235 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1236
1237 *) Fix SSL client code which could crash if connecting to a
1238 malicious SSLv2 server. (CVE-2006-4343)
1239 [Tavis Ormandy and Will Drewry, Google Security Team]
1240
1241 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
1242 match only those. Before that, "AES256-SHA" would be interpreted
1243 as a pattern and match "AES128-SHA" too (since AES128-SHA got
1244 the same strength classification in 0.9.7h) as we currently only
1245 have a single AES bit in the ciphersuite description bitmap.
1246 That change, however, also applied to ciphersuite strings such as
1247 "RC4-MD5" that intentionally matched multiple ciphersuites --
1248 namely, SSL 2.0 ciphersuites in addition to the more common ones
1249 from SSL 3.0/TLS 1.0.
1250
1251 So we change the selection algorithm again: Naming an explicit
1252 ciphersuite selects this one ciphersuite, and any other similar
1253 ciphersuite (same bitmap) from *other* protocol versions.
1254 Thus, "RC4-MD5" again will properly select both the SSL 2.0
1255 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
1256
1257 Since SSL 2.0 does not have any ciphersuites for which the
1258 128/256 bit distinction would be relevant, this works for now.
1259 The proper fix will be to use different bits for AES128 and
1260 AES256, which would have avoided the problems from the beginning;
1261 however, bits are scarce, so we can only do this in a new release
1262 (not just a patchlevel) when we can change the SSL_CIPHER
1263 definition to split the single 'unsigned long mask' bitmap into
1264 multiple values to extend the available space.
1265
1266 [Bodo Moeller]
1267
1268 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
1269
1270 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1271 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1272
1273 *) Add AES IGE and biIGE modes.
1274 [Ben Laurie]
1275
1276 *) Change the Unix randomness entropy gathering to use poll() when
1277 possible instead of select(), since the latter has some
1278 undesirable limitations.
1279 [Darryl Miles via Richard Levitte and Bodo Moeller]
1280
1281 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
1282 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
1283 cannot be implicitly activated as part of, e.g., the "AES" alias.
1284 However, please upgrade to OpenSSL 0.9.9[-dev] for
1285 non-experimental use of the ECC ciphersuites to get TLS extension
1286 support, which is required for curve and point format negotiation
1287 to avoid potential handshake problems.
1288 [Bodo Moeller]
1289
1290 *) Disable rogue ciphersuites:
1291
1292 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1293 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1294 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1295
1296 The latter two were purportedly from
1297 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1298 appear there.
1299
1300 Also deactivate the remaining ciphersuites from
1301 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1302 unofficial, and the ID has long expired.
1303 [Bodo Moeller]
1304
1305 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1306 dual-core machines) and other potential thread-safety issues.
1307 [Bodo Moeller]
1308
1309 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
1310 versions), which is now available for royalty-free use
1311 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
1312 Also, add Camellia TLS ciphersuites from RFC 4132.
1313
1314 To minimize changes between patchlevels in the OpenSSL 0.9.8
1315 series, Camellia remains excluded from compilation unless OpenSSL
1316 is configured with 'enable-camellia'.
1317 [NTT]
1318
1319 *) Disable the padding bug check when compression is in use. The padding
1320 bug check assumes the first packet is of even length, this is not
1321 necessarily true if compresssion is enabled and can result in false
1322 positives causing handshake failure. The actual bug test is ancient
1323 code so it is hoped that implementations will either have fixed it by
1324 now or any which still have the bug do not support compression.
1325 [Steve Henson]
1326
1327 Changes between 0.9.8a and 0.9.8b [04 May 2006]
1328
1329 *) When applying a cipher rule check to see if string match is an explicit
1330 cipher suite and only match that one cipher suite if it is.
1331 [Steve Henson]
1332
1333 *) Link in manifests for VC++ if needed.
1334 [Austin Ziegler <halostatue@gmail.com>]
1335
1336 *) Update support for ECC-based TLS ciphersuites according to
1337 draft-ietf-tls-ecc-12.txt with proposed changes (but without
1338 TLS extensions, which are supported starting with the 0.9.9
1339 branch, not in the OpenSSL 0.9.8 branch).
1340 [Douglas Stebila]
1341
1342 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
1343 opaque EVP_CIPHER_CTX handling.
1344 [Steve Henson]
1345
1346 *) Fixes and enhancements to zlib compression code. We now only use
1347 "zlib1.dll" and use the default __cdecl calling convention on Win32
1348 to conform with the standards mentioned here:
1349 http://www.zlib.net/DLL_FAQ.txt
1350 Static zlib linking now works on Windows and the new --with-zlib-include
1351 --with-zlib-lib options to Configure can be used to supply the location
1352 of the headers and library. Gracefully handle case where zlib library
1353 can't be loaded.
1354 [Steve Henson]
1355
1356 *) Several fixes and enhancements to the OID generation code. The old code
1357 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
1358 handle numbers larger than ULONG_MAX, truncated printing and had a
1359 non standard OBJ_obj2txt() behaviour.
1360 [Steve Henson]
1361
1362 *) Add support for building of engines under engine/ as shared libraries
1363 under VC++ build system.
1364 [Steve Henson]
1365
1366 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
1367 Hopefully, we will not see any false combination of paths any more.
1368 [Richard Levitte]
1369
1370 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
1371
1372 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1373 (part of SSL_OP_ALL). This option used to disable the
1374 countermeasure against man-in-the-middle protocol-version
1375 rollback in the SSL 2.0 server implementation, which is a bad
1376 idea. (CVE-2005-2969)
1377
1378 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1379 for Information Security, National Institute of Advanced Industrial
1380 Science and Technology [AIST], Japan)]
1381
1382 *) Add two function to clear and return the verify parameter flags.
1383 [Steve Henson]
1384
1385 *) Keep cipherlists sorted in the source instead of sorting them at
1386 runtime, thus removing the need for a lock.
1387 [Nils Larsch]
1388
1389 *) Avoid some small subgroup attacks in Diffie-Hellman.
1390 [Nick Mathewson and Ben Laurie]
1391
1392 *) Add functions for well-known primes.
1393 [Nick Mathewson]
1394
1395 *) Extended Windows CE support.
1396 [Satoshi Nakamura and Andy Polyakov]
1397
1398 *) Initialize SSL_METHOD structures at compile time instead of during
1399 runtime, thus removing the need for a lock.
1400 [Steve Henson]
1401
1402 *) Make PKCS7_decrypt() work even if no certificate is supplied by
1403 attempting to decrypt each encrypted key in turn. Add support to
1404 smime utility.
1405 [Steve Henson]
1406
1407 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
1408
1409 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
1410 OpenSSL 0.9.8.]
1411
1412 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
1413 [Richard Levitte]
1414
1415 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
1416 key into the same file any more.
1417 [Richard Levitte]
1418
1419 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
1420 [Andy Polyakov]
1421
1422 *) Add -utf8 command line and config file option to 'ca'.
1423 [Stefan <stf@udoma.org]
1424
1425 *) Removed the macro des_crypt(), as it seems to conflict with some
1426 libraries. Use DES_crypt().
1427 [Richard Levitte]
1428
1429 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
1430 involves renaming the source and generated shared-libs for
1431 both. The engines will accept the corrected or legacy ids
1432 ('ncipher' and '4758_cca' respectively) when binding. NB,
1433 this only applies when building 'shared'.
1434 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
1435
1436 *) Add attribute functions to EVP_PKEY structure. Modify
1437 PKCS12_create() to recognize a CSP name attribute and
1438 use it. Make -CSP option work again in pkcs12 utility.
1439 [Steve Henson]
1440
1441 *) Add new functionality to the bn blinding code:
1442 - automatic re-creation of the BN_BLINDING parameters after
1443 a fixed number of uses (currently 32)
1444 - add new function for parameter creation
1445 - introduce flags to control the update behaviour of the
1446 BN_BLINDING parameters
1447 - hide BN_BLINDING structure
1448 Add a second BN_BLINDING slot to the RSA structure to improve
1449 performance when a single RSA object is shared among several
1450 threads.
1451 [Nils Larsch]
1452
1453 *) Add support for DTLS.
1454 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
1455
1456 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
1457 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
1458 [Walter Goulet]
1459
1460 *) Remove buggy and incompletet DH cert support from
1461 ssl/ssl_rsa.c and ssl/s3_both.c
1462 [Nils Larsch]
1463
1464 *) Use SHA-1 instead of MD5 as the default digest algorithm for
1465 the apps/openssl applications.
1466 [Nils Larsch]
1467
1468 *) Compile clean with "-Wall -Wmissing-prototypes
1469 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
1470 DEBUG_SAFESTACK must also be set.
1471 [Ben Laurie]
1472
1473 *) Change ./Configure so that certain algorithms can be disabled by default.
1474 The new counterpiece to "no-xxx" is "enable-xxx".
1475
1476 The patented RC5 and MDC2 algorithms will now be disabled unless
1477 "enable-rc5" and "enable-mdc2", respectively, are specified.
1478
1479 (IDEA remains enabled despite being patented. This is because IDEA
1480 is frequently required for interoperability, and there is no license
1481 fee for non-commercial use. As before, "no-idea" can be used to
1482 avoid this algorithm.)
1483
1484 [Bodo Moeller]
1485
1486 *) Add processing of proxy certificates (see RFC 3820). This work was
1487 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
1488 EGEE (Enabling Grids for E-science in Europe).
1489 [Richard Levitte]
1490
1491 *) RC4 performance overhaul on modern architectures/implementations, such
1492 as Intel P4, IA-64 and AMD64.
1493 [Andy Polyakov]
1494
1495 *) New utility extract-section.pl. This can be used specify an alternative
1496 section number in a pod file instead of having to treat each file as
1497 a separate case in Makefile. This can be done by adding two lines to the
1498 pod file:
1499
1500 =for comment openssl_section:XXX
1501
1502 The blank line is mandatory.
1503
1504 [Steve Henson]
1505
1506 *) New arguments -certform, -keyform and -pass for s_client and s_server
1507 to allow alternative format key and certificate files and passphrase
1508 sources.
1509 [Steve Henson]
1510
1511 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
1512 update associated structures and add various utility functions.
1513
1514 Add new policy related verify parameters, include policy checking in
1515 standard verify code. Enhance 'smime' application with extra parameters
1516 to support policy checking and print out.
1517 [Steve Henson]
1518
1519 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
1520 Nehemiah processors. These extensions support AES encryption in hardware
1521 as well as RNG (though RNG support is currently disabled).
1522 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
1523
1524 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
1525 [Geoff Thorpe]
1526
1527 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
1528 [Andy Polyakov and a number of other people]
1529
1530 *) Improved PowerPC platform support. Most notably BIGNUM assembler
1531 implementation contributed by IBM.
1532 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
1533
1534 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
1535 exponent rather than 'unsigned long'. There is a corresponding change to
1536 the new 'rsa_keygen' element of the RSA_METHOD structure.
1537 [Jelte Jansen, Geoff Thorpe]
1538
1539 *) Functionality for creating the initial serial number file is now
1540 moved from CA.pl to the 'ca' utility with a new option -create_serial.
1541
1542 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
1543 number file to 1, which is bound to cause problems. To avoid
1544 the problems while respecting compatibility between different 0.9.7
1545 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
1546 CA.pl for serial number initialization. With the new release 0.9.8,
1547 we can fix the problem directly in the 'ca' utility.)
1548 [Steve Henson]
1549
1550 *) Reduced header interdepencies by declaring more opaque objects in
1551 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
1552 give fewer recursive includes, which could break lazy source code - so
1553 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
1554 developers should define this symbol when building and using openssl to
1555 ensure they track the recommended behaviour, interfaces, [etc], but
1556 backwards-compatible behaviour prevails when this isn't defined.
1557 [Geoff Thorpe]
1558
1559 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
1560 [Steve Henson]
1561
1562 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
1563 This will generate a random key of the appropriate length based on the
1564 cipher context. The EVP_CIPHER can provide its own random key generation
1565 routine to support keys of a specific form. This is used in the des and
1566 3des routines to generate a key of the correct parity. Update S/MIME
1567 code to use new functions and hence generate correct parity DES keys.
1568 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
1569 valid (weak or incorrect parity).
1570 [Steve Henson]
1571
1572 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
1573 as looking them up. This is useful when the verified structure may contain
1574 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
1575 present unless the new PKCS7_NO_CRL flag is asserted.
1576 [Steve Henson]
1577
1578 *) Extend ASN1 oid configuration module. It now additionally accepts the
1579 syntax:
1580
1581 shortName = some long name, 1.2.3.4
1582 [Steve Henson]
1583
1584 *) Reimplemented the BN_CTX implementation. There is now no more static
1585 limitation on the number of variables it can handle nor the depth of the
1586 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
1587 information can now expand as required, and rather than having a single
1588 static array of bignums, BN_CTX now uses a linked-list of such arrays
1589 allowing it to expand on demand whilst maintaining the usefulness of
1590 BN_CTX's "bundling".
1591 [Geoff Thorpe]
1592
1593 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
1594 to allow all RSA operations to function using a single BN_CTX.
1595 [Geoff Thorpe]
1596
1597 *) Preliminary support for certificate policy evaluation and checking. This
1598 is initially intended to pass the tests outlined in "Conformance Testing
1599 of Relying Party Client Certificate Path Processing Logic" v1.07.
1600 [Steve Henson]
1601
1602 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1603 remained unused and not that useful. A variety of other little bignum
1604 tweaks and fixes have also been made continuing on from the audit (see
1605 below).
1606 [Geoff Thorpe]
1607
1608 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1609 associated ASN1, EVP and SSL functions and old ASN1 macros.
1610 [Richard Levitte]
1611
1612 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1613 and this should never fail. So the return value from the use of
1614 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1615 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1616 [Geoff Thorpe]
1617
1618 *) BN_CTX_get() should return zero-valued bignums, providing the same
1619 initialised value as BN_new().
1620 [Geoff Thorpe, suggested by Ulf Möller]
1621
1622 *) Support for inhibitAnyPolicy certificate extension.
1623 [Steve Henson]
1624
1625 *) An audit of the BIGNUM code is underway, for which debugging code is
1626 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1627 is considered valid when processing BIGNUMs, and causes execution to
1628 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1629 further steps are taken to deliberately pollute unused data in BIGNUM
1630 structures to try and expose faulty code further on. For now, openssl will
1631 (in its default mode of operation) continue to tolerate the inconsistent
1632 forms that it has tolerated in the past, but authors and packagers should
1633 consider trying openssl and their own applications when compiled with
1634 these debugging symbols defined. It will help highlight potential bugs in
1635 their own code, and will improve the test coverage for OpenSSL itself. At
1636 some point, these tighter rules will become openssl's default to improve
1637 maintainability, though the assert()s and other overheads will remain only
1638 in debugging configurations. See bn.h for more details.
1639 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1640
1641 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1642 that can only be obtained through BN_CTX_new() (which implicitly
1643 initialises it). The presence of this function only made it possible
1644 to overwrite an existing structure (and cause memory leaks).
1645 [Geoff Thorpe]
1646
1647 *) Because of the callback-based approach for implementing LHASH as a
1648 template type, lh_insert() adds opaque objects to hash-tables and
1649 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1650 to clean up those corresponding objects before destroying the hash table
1651 (and losing the object pointers). So some over-zealous constifications in
1652 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1653 objects as "const" and the lh_doall[_arg] callback wrappers are not
1654 prototyped to have "const" restrictions on the object pointers they are
1655 given (and so aren't required to cast them away any more).
1656 [Geoff Thorpe]
1657
1658 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1659 (speed) prefers to use its own implementation. The two implementations
1660 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1661 its object type properly exposed (MS_TM) instead of casting to/from "char
1662 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1663 aren't necessarily the greatest nomenclatures - but this is what was used
1664 internally to the implementation so I've used that for now.
1665 [Geoff Thorpe]
1666
1667 *) Ensure that deprecated functions do not get compiled when
1668 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1669 the self-tests were still using deprecated key-generation functions so
1670 these have been updated also.
1671 [Geoff Thorpe]
1672
1673 *) Reorganise PKCS#7 code to separate the digest location functionality
1674 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1675 New function PKCS7_set_digest() to set the digest type for PKCS#7
1676 digestedData type. Add additional code to correctly generate the
1677 digestedData type and add support for this type in PKCS7 initialization
1678 functions.
1679 [Steve Henson]
1680
1681 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1682 structure of type "other".
1683 [Steve Henson]
1684
1685 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1686 sure the loop does correctly stop and breaking ("division by zero")
1687 modulus operations are not performed. The (pre-generated) prime
1688 table crypto/bn/bn_prime.h was already correct, but it could not be
1689 re-generated on some platforms because of the "division by zero"
1690 situation in the script.
1691 [Ralf S. Engelschall]
1692
1693 *) Update support for ECC-based TLS ciphersuites according to
1694 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1695 SHA-1 now is only used for "small" curves (where the
1696 representation of a field element takes up to 24 bytes); for
1697 larger curves, the field element resulting from ECDH is directly
1698 used as premaster secret.
1699 [Douglas Stebila (Sun Microsystems Laboratories)]
1700
1701 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1702 curve secp160r1 to the tests.
1703 [Douglas Stebila (Sun Microsystems Laboratories)]
1704
1705 *) Add the possibility to load symbols globally with DSO.
1706 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1707
1708 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1709 control of the error stack.
1710 [Richard Levitte]
1711
1712 *) Add support for STORE in ENGINE.
1713 [Richard Levitte]
1714
1715 *) Add the STORE type. The intention is to provide a common interface
1716 to certificate and key stores, be they simple file-based stores, or
1717 HSM-type store, or LDAP stores, or...
1718 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1719 [Richard Levitte]
1720
1721 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1722 pass a list of arguments to any function as well as provide a way
1723 for a function to pass data back to the caller.
1724 [Richard Levitte]
1725
1726 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1727 works like BUF_strdup() but can be used to duplicate a portion of
1728 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1729 a memory area.
1730 [Richard Levitte]
1731
1732 *) Add the function sk_find_ex() which works like sk_find(), but will
1733 return an index to an element even if an exact match couldn't be
1734 found. The index is guaranteed to point at the element where the
1735 searched-for key would be inserted to preserve sorting order.
1736 [Richard Levitte]
1737
1738 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1739 takes an extra flags argument for optional functionality. Currently,
1740 the following flags are defined:
1741
1742 OBJ_BSEARCH_VALUE_ON_NOMATCH
1743 This one gets OBJ_bsearch_ex() to return a pointer to the first
1744 element where the comparing function returns a negative or zero
1745 number.
1746
1747 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1748 This one gets OBJ_bsearch_ex() to return a pointer to the first
1749 element where the comparing function returns zero. This is useful
1750 if there are more than one element where the comparing function
1751 returns zero.
1752 [Richard Levitte]
1753
1754 *) Make it possible to create self-signed certificates with 'openssl ca'
1755 in such a way that the self-signed certificate becomes part of the
1756 CA database and uses the same mechanisms for serial number generation
1757 as all other certificate signing. The new flag '-selfsign' enables
1758 this functionality. Adapt CA.sh and CA.pl.in.
1759 [Richard Levitte]
1760
1761 *) Add functionality to check the public key of a certificate request
1762 against a given private. This is useful to check that a certificate
1763 request can be signed by that key (self-signing).
1764 [Richard Levitte]
1765
1766 *) Make it possible to have multiple active certificates with the same
1767 subject in the CA index file. This is done only if the keyword
1768 'unique_subject' is set to 'no' in the main CA section (default
1769 if 'CA_default') of the configuration file. The value is saved
1770 with the database itself in a separate index attribute file,
1771 named like the index file with '.attr' appended to the name.
1772 [Richard Levitte]
1773
1774 *) Generate muti valued AVAs using '+' notation in config files for
1775 req and dirName.
1776 [Steve Henson]
1777
1778 *) Support for nameConstraints certificate extension.
1779 [Steve Henson]
1780
1781 *) Support for policyConstraints certificate extension.
1782 [Steve Henson]
1783
1784 *) Support for policyMappings certificate extension.
1785 [Steve Henson]
1786
1787 *) Make sure the default DSA_METHOD implementation only uses its
1788 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
1789 and change its own handlers to be NULL so as to remove unnecessary
1790 indirection. This lets alternative implementations fallback to the
1791 default implementation more easily.
1792 [Geoff Thorpe]
1793
1794 *) Support for directoryName in GeneralName related extensions
1795 in config files.
1796 [Steve Henson]
1797
1798 *) Make it possible to link applications using Makefile.shared.
1799 Make that possible even when linking against static libraries!
1800 [Richard Levitte]
1801
1802 *) Support for single pass processing for S/MIME signing. This now
1803 means that S/MIME signing can be done from a pipe, in addition
1804 cleartext signing (multipart/signed type) is effectively streaming
1805 and the signed data does not need to be all held in memory.
1806
1807 This is done with a new flag PKCS7_STREAM. When this flag is set
1808 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1809 is done after the data is output (and digests calculated) in
1810 SMIME_write_PKCS7().
1811 [Steve Henson]
1812
1813 *) Add full support for -rpath/-R, both in shared libraries and
1814 applications, at least on the platforms where it's known how
1815 to do it.
1816 [Richard Levitte]
1817
1818 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1819 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1820 will now compute a table of multiples of the generator that
1821 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1822 faster (notably in the case of a single point multiplication,
1823 scalar * generator).
1824 [Nils Larsch, Bodo Moeller]
1825
1826 *) IPv6 support for certificate extensions. The various extensions
1827 which use the IP:a.b.c.d can now take IPv6 addresses using the
1828 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1829 correctly.
1830 [Steve Henson]
1831
1832 *) Added an ENGINE that implements RSA by performing private key
1833 exponentiations with the GMP library. The conversions to and from
1834 GMP's mpz_t format aren't optimised nor are any montgomery forms
1835 cached, and on x86 it appears OpenSSL's own performance has caught up.
1836 However there are likely to be other architectures where GMP could
1837 provide a boost. This ENGINE is not built in by default, but it can be
1838 specified at Configure time and should be accompanied by the necessary
1839 linker additions, eg;
1840 ./config -DOPENSSL_USE_GMP -lgmp
1841 [Geoff Thorpe]
1842
1843 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1844 testing availability of engines with "-t" - the old behaviour is
1845 produced by increasing the feature's verbosity with "-tt".
1846 [Geoff Thorpe]
1847
1848 *) ECDSA routines: under certain error conditions uninitialized BN objects
1849 could be freed. Solution: make sure initialization is performed early
1850 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1851 via PR#459)
1852 [Lutz Jaenicke]
1853
1854 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1855 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1856 software implementations. For DSA and DH, parameter generation can
1857 also be overriden by providing the appropriate method callbacks.
1858 [Geoff Thorpe]
1859
1860 *) Change the "progress" mechanism used in key-generation and
1861 primality testing to functions that take a new BN_GENCB pointer in
1862 place of callback/argument pairs. The new API functions have "_ex"
1863 postfixes and the older functions are reimplemented as wrappers for
1864 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1865 declarations of the old functions to help (graceful) attempts to
1866 migrate to the new functions. Also, the new key-generation API
1867 functions operate on a caller-supplied key-structure and return
1868 success/failure rather than returning a key or NULL - this is to
1869 help make "keygen" another member function of RSA_METHOD etc.
1870
1871 Example for using the new callback interface:
1872
1873 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1874 void *my_arg = ...;
1875 BN_GENCB my_cb;
1876
1877 BN_GENCB_set(&my_cb, my_callback, my_arg);
1878
1879 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1880 /* For the meaning of a, b in calls to my_callback(), see the
1881 * documentation of the function that calls the callback.
1882 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1883 * my_callback should return 1 if it wants BN_is_prime_ex()
1884 * to continue, or 0 to stop.
1885 */
1886
1887 [Geoff Thorpe]
1888
1889 *) Change the ZLIB compression method to be stateful, and make it
1890 available to TLS with the number defined in
1891 draft-ietf-tls-compression-04.txt.
1892 [Richard Levitte]
1893
1894 *) Add the ASN.1 structures and functions for CertificatePair, which
1895 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1896
1897 CertificatePair ::= SEQUENCE {
1898 forward [0] Certificate OPTIONAL,
1899 reverse [1] Certificate OPTIONAL,
1900 -- at least one of the pair shall be present -- }
1901
1902 Also implement the PEM functions to read and write certificate
1903 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1904
1905 This needed to be defined, mostly for the sake of the LDAP
1906 attribute crossCertificatePair, but may prove useful elsewhere as
1907 well.
1908 [Richard Levitte]
1909
1910 *) Make it possible to inhibit symlinking of shared libraries in
1911 Makefile.shared, for Cygwin's sake.
1912 [Richard Levitte]
1913
1914 *) Extend the BIGNUM API by creating a function
1915 void BN_set_negative(BIGNUM *a, int neg);
1916 and a macro that behave like
1917 int BN_is_negative(const BIGNUM *a);
1918
1919 to avoid the need to access 'a->neg' directly in applications.
1920 [Nils Larsch]
1921
1922 *) Implement fast modular reduction for pseudo-Mersenne primes
1923 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1924 EC_GROUP_new_curve_GFp() will now automatically use this
1925 if applicable.
1926 [Nils Larsch <nla@trustcenter.de>]
1927
1928 *) Add new lock type (CRYPTO_LOCK_BN).
1929 [Bodo Moeller]
1930
1931 *) Change the ENGINE framework to automatically load engines
1932 dynamically from specific directories unless they could be
1933 found to already be built in or loaded. Move all the
1934 current engines except for the cryptodev one to a new
1935 directory engines/.
1936 The engines in engines/ are built as shared libraries if
1937 the "shared" options was given to ./Configure or ./config.
1938 Otherwise, they are inserted in libcrypto.a.
1939 /usr/local/ssl/engines is the default directory for dynamic
1940 engines, but that can be overriden at configure time through
1941 the usual use of --prefix and/or --openssldir, and at run
1942 time with the environment variable OPENSSL_ENGINES.
1943 [Geoff Thorpe and Richard Levitte]
1944
1945 *) Add Makefile.shared, a helper makefile to build shared
1946 libraries. Addapt Makefile.org.
1947 [Richard Levitte]
1948
1949 *) Add version info to Win32 DLLs.
1950 [Peter 'Luna' Runestig" <peter@runestig.com>]
1951
1952 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1953 can be added using this API to created arbitrary PKCS#12
1954 files while avoiding the low level API.
1955
1956 New options to PKCS12_create(), key or cert can be NULL and
1957 will then be omitted from the output file. The encryption
1958 algorithm NIDs can be set to -1 for no encryption, the mac
1959 iteration count can be set to 0 to omit the mac.
1960
1961 Enhance pkcs12 utility by making the -nokeys and -nocerts
1962 options work when creating a PKCS#12 file. New option -nomac
1963 to omit the mac, NONE can be set for an encryption algorithm.
1964 New code is modified to use the enhanced PKCS12_create()
1965 instead of the low level API.
1966 [Steve Henson]
1967
1968 *) Extend ASN1 encoder to support indefinite length constructed
1969 encoding. This can output sequences tags and octet strings in
1970 this form. Modify pk7_asn1.c to support indefinite length
1971 encoding. This is experimental and needs additional code to
1972 be useful, such as an ASN1 bio and some enhanced streaming
1973 PKCS#7 code.
1974
1975 Extend template encode functionality so that tagging is passed
1976 down to the template encoder.
1977 [Steve Henson]
1978
1979 *) Let 'openssl req' fail if an argument to '-newkey' is not
1980 recognized instead of using RSA as a default.
1981 [Bodo Moeller]
1982
1983 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1984 As these are not official, they are not included in "ALL";
1985 the "ECCdraft" ciphersuite group alias can be used to select them.
1986 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1987
1988 *) Add ECDH engine support.
1989 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1990
1991 *) Add ECDH in new directory crypto/ecdh/.
1992 [Douglas Stebila (Sun Microsystems Laboratories)]
1993
1994 *) Let BN_rand_range() abort with an error after 100 iterations
1995 without success (which indicates a broken PRNG).
1996 [Bodo Moeller]
1997
1998 *) Change BN_mod_sqrt() so that it verifies that the input value
1999 is really the square of the return value. (Previously,
2000 BN_mod_sqrt would show GIGO behaviour.)
2001 [Bodo Moeller]
2002
2003 *) Add named elliptic curves over binary fields from X9.62, SECG,
2004 and WAP/WTLS; add OIDs that were still missing.
2005
2006 [Sheueling Chang Shantz and Douglas Stebila
2007 (Sun Microsystems Laboratories)]
2008
2009 *) Extend the EC library for elliptic curves over binary fields
2010 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2011 New EC_METHOD:
2012
2013 EC_GF2m_simple_method
2014
2015 New API functions:
2016
2017 EC_GROUP_new_curve_GF2m
2018 EC_GROUP_set_curve_GF2m
2019 EC_GROUP_get_curve_GF2m
2020 EC_POINT_set_affine_coordinates_GF2m
2021 EC_POINT_get_affine_coordinates_GF2m
2022 EC_POINT_set_compressed_coordinates_GF2m
2023
2024 Point compression for binary fields is disabled by default for
2025 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2026 enable it).
2027
2028 As binary polynomials are represented as BIGNUMs, various members
2029 of the EC_GROUP and EC_POINT data structures can be shared
2030 between the implementations for prime fields and binary fields;
2031 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2032 are essentially identical to their ..._GFp counterparts.
2033 (For simplicity, the '..._GFp' prefix has been dropped from
2034 various internal method names.)
2035
2036 An internal 'field_div' method (similar to 'field_mul' and
2037 'field_sqr') has been added; this is used only for binary fields.
2038
2039 [Sheueling Chang Shantz and Douglas Stebila
2040 (Sun Microsystems Laboratories)]
2041
2042 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2043 through methods ('mul', 'precompute_mult').
2044
2045 The generic implementations (now internally called 'ec_wNAF_mul'
2046 and 'ec_wNAF_precomputed_mult') remain the default if these
2047 methods are undefined.
2048
2049 [Sheueling Chang Shantz and Douglas Stebila
2050 (Sun Microsystems Laboratories)]
2051
2052 *) New function EC_GROUP_get_degree, which is defined through
2053 EC_METHOD. For curves over prime fields, this returns the bit
2054 length of the modulus.
2055
2056 [Sheueling Chang Shantz and Douglas Stebila
2057 (Sun Microsystems Laboratories)]
2058
2059 *) New functions EC_GROUP_dup, EC_POINT_dup.
2060 (These simply call ..._new and ..._copy).
2061
2062 [Sheueling Chang Shantz and Douglas Stebila
2063 (Sun Microsystems Laboratories)]
2064
2065 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2066 Polynomials are represented as BIGNUMs (where the sign bit is not
2067 used) in the following functions [macros]:
2068
2069 BN_GF2m_add
2070 BN_GF2m_sub [= BN_GF2m_add]
2071 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2072 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2073 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2074 BN_GF2m_mod_inv
2075 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2076 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2077 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2078 BN_GF2m_cmp [= BN_ucmp]
2079
2080 (Note that only the 'mod' functions are actually for fields GF(2^m).
2081 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2082
2083 For some functions, an the irreducible polynomial defining a
2084 field can be given as an 'unsigned int[]' with strictly
2085 decreasing elements giving the indices of those bits that are set;
2086 i.e., p[] represents the polynomial
2087 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2088 where
2089 p[0] > p[1] > ... > p[k] = 0.
2090 This applies to the following functions:
2091
2092 BN_GF2m_mod_arr
2093 BN_GF2m_mod_mul_arr
2094 BN_GF2m_mod_sqr_arr
2095 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2096 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2097 BN_GF2m_mod_exp_arr
2098 BN_GF2m_mod_sqrt_arr
2099 BN_GF2m_mod_solve_quad_arr
2100 BN_GF2m_poly2arr
2101 BN_GF2m_arr2poly
2102
2103 Conversion can be performed by the following functions:
2104
2105 BN_GF2m_poly2arr
2106 BN_GF2m_arr2poly
2107
2108 bntest.c has additional tests for binary polynomial arithmetic.
2109
2110 Two implementations for BN_GF2m_mod_div() are available.
2111 The default algorithm simply uses BN_GF2m_mod_inv() and
2112 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2113 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2114 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2115
2116 [Sheueling Chang Shantz and Douglas Stebila
2117 (Sun Microsystems Laboratories)]
2118
2119 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2120 functionality is disabled at compile-time.
2121 [Douglas Stebila <douglas.stebila@sun.com>]
2122
2123 *) Change default behaviour of 'openssl asn1parse' so that more
2124 information is visible when viewing, e.g., a certificate:
2125
2126 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2127 mode the content of non-printable OCTET STRINGs is output in a
2128 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2129 avoid the appearance of a printable string.
2130 [Nils Larsch <nla@trustcenter.de>]
2131
2132 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2133 functions
2134 EC_GROUP_set_asn1_flag()
2135 EC_GROUP_get_asn1_flag()
2136 EC_GROUP_set_point_conversion_form()
2137 EC_GROUP_get_point_conversion_form()
2138 These control ASN1 encoding details:
2139 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
2140 has been set to OPENSSL_EC_NAMED_CURVE.
2141 - Points are encoded in uncompressed form by default; options for
2142 asn1_for are as for point2oct, namely
2143 POINT_CONVERSION_COMPRESSED
2144 POINT_CONVERSION_UNCOMPRESSED
2145 POINT_CONVERSION_HYBRID
2146
2147 Also add 'seed' and 'seed_len' members to EC_GROUP with access
2148 functions
2149 EC_GROUP_set_seed()
2150 EC_GROUP_get0_seed()
2151 EC_GROUP_get_seed_len()
2152 This is used only for ASN1 purposes (so far).
2153 [Nils Larsch <nla@trustcenter.de>]
2154
2155 *) Add 'field_type' member to EC_METHOD, which holds the NID
2156 of the appropriate field type OID. The new function
2157 EC_METHOD_get_field_type() returns this value.
2158 [Nils Larsch <nla@trustcenter.de>]
2159
2160 *) Add functions
2161 EC_POINT_point2bn()
2162 EC_POINT_bn2point()
2163 EC_POINT_point2hex()
2164 EC_POINT_hex2point()
2165 providing useful interfaces to EC_POINT_point2oct() and
2166 EC_POINT_oct2point().
2167 [Nils Larsch <nla@trustcenter.de>]
2168
2169 *) Change internals of the EC library so that the functions
2170 EC_GROUP_set_generator()
2171 EC_GROUP_get_generator()
2172 EC_GROUP_get_order()
2173 EC_GROUP_get_cofactor()
2174 are implemented directly in crypto/ec/ec_lib.c and not dispatched
2175 to methods, which would lead to unnecessary code duplication when
2176 adding different types of curves.
2177 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
2178
2179 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
2180 arithmetic, and such that modified wNAFs are generated
2181 (which avoid length expansion in many cases).
2182 [Bodo Moeller]
2183
2184 *) Add a function EC_GROUP_check_discriminant() (defined via
2185 EC_METHOD) that verifies that the curve discriminant is non-zero.
2186
2187 Add a function EC_GROUP_check() that makes some sanity tests
2188 on a EC_GROUP, its generator and order. This includes
2189 EC_GROUP_check_discriminant().
2190 [Nils Larsch <nla@trustcenter.de>]
2191
2192 *) Add ECDSA in new directory crypto/ecdsa/.
2193
2194 Add applications 'openssl ecparam' and 'openssl ecdsa'
2195 (these are based on 'openssl dsaparam' and 'openssl dsa').
2196
2197 ECDSA support is also included in various other files across the
2198 library. Most notably,
2199 - 'openssl req' now has a '-newkey ecdsa:file' option;
2200 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
2201 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
2202 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
2203 them suitable for ECDSA where domain parameters must be
2204 extracted before the specific public key;
2205 - ECDSA engine support has been added.
2206 [Nils Larsch <nla@trustcenter.de>]
2207
2208 *) Include some named elliptic curves, and add OIDs from X9.62,
2209 SECG, and WAP/WTLS. Each curve can be obtained from the new
2210 function
2211 EC_GROUP_new_by_curve_name(),
2212 and the list of available named curves can be obtained with
2213 EC_get_builtin_curves().
2214 Also add a 'curve_name' member to EC_GROUP objects, which can be
2215 accessed via
2216 EC_GROUP_set_curve_name()
2217 EC_GROUP_get_curve_name()
2218 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
2219
2220 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2221 was actually never needed) and in BN_mul(). The removal in BN_mul()
2222 required a small change in bn_mul_part_recursive() and the addition
2223 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2224 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2225 bn_sub_words() and bn_add_words() except they take arrays with
2226 differing sizes.
2227 [Richard Levitte]
2228
2229 Changes between 0.9.7m and 0.9.7n [xx XXX xxxx]
2230
2231 *) In the SSL/TLS server implementation, be strict about session ID
2232 context matching (which matters if an application uses a single
2233 external cache for different purposes). Previously,
2234 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2235 set. This did ensure strict client verification, but meant that,
2236 with applications using a single external cache for quite
2237 different requirements, clients could circumvent ciphersuite
2238 restrictions for a given session ID context by starting a session
2239 in a different context.
2240 [Bodo Moeller]
2241
2242 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
2243
2244 *) Cleanse PEM buffers before freeing them since they may contain
2245 sensitive data.
2246 [Benjamin Bennett <ben@psc.edu>]
2247
2248 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2249 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2250 authentication-only ciphersuites.
2251 [Bodo Moeller]
2252
2253 *) Since AES128 and AES256 share a single mask bit in the logic of
2254 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2255 kludge to work properly if AES128 is available and AES256 isn't.
2256 [Victor Duchovni]
2257
2258 *) Expand security boundary to match 1.1.1 module.
2259 [Steve Henson]
2260
2261 *) Remove redundant features: hash file source, editing of test vectors
2262 modify fipsld to use external fips_premain.c signature.
2263 [Steve Henson]
2264
2265 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
2266 run algorithm test programs.
2267 [Steve Henson]
2268
2269 *) Make algorithm test programs more tolerant of whitespace.
2270 [Steve Henson]
2271
2272 *) Have SSL/TLS server implementation tolerate "mismatched" record
2273 protocol version while receiving ClientHello even if the
2274 ClientHello is fragmented. (The server can't insist on the
2275 particular protocol version it has chosen before the ServerHello
2276 message has informed the client about his choice.)
2277 [Bodo Moeller]
2278
2279 *) Load error codes if they are not already present instead of using a
2280 static variable. This allows them to be cleanly unloaded and reloaded.
2281 [Steve Henson]
2282
2283 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
2284
2285 *) Introduce limits to prevent malicious keys being able to
2286 cause a denial of service. (CVE-2006-2940)
2287 [Steve Henson, Bodo Moeller]
2288
2289 *) Fix ASN.1 parsing of certain invalid structures that can result
2290 in a denial of service. (CVE-2006-2937) [Steve Henson]
2291
2292 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2293 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2294
2295 *) Fix SSL client code which could crash if connecting to a
2296 malicious SSLv2 server. (CVE-2006-4343)
2297 [Tavis Ormandy and Will Drewry, Google Security Team]
2298
2299 *) Change ciphersuite string processing so that an explicit
2300 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
2301 will no longer include "AES128-SHA"), and any other similar
2302 ciphersuite (same bitmap) from *other* protocol versions (so that
2303 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
2304 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
2305 changes from 0.9.8b and 0.9.8d.
2306 [Bodo Moeller]
2307
2308 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
2309
2310 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2311 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2312
2313 *) Change the Unix randomness entropy gathering to use poll() when
2314 possible instead of select(), since the latter has some
2315 undesirable limitations.
2316 [Darryl Miles via Richard Levitte and Bodo Moeller]
2317
2318 *) Disable rogue ciphersuites:
2319
2320 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2321 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2322 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2323
2324 The latter two were purportedly from
2325 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2326 appear there.
2327
2328 Also deactive the remaining ciphersuites from
2329 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2330 unofficial, and the ID has long expired.
2331 [Bodo Moeller]
2332
2333 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2334 dual-core machines) and other potential thread-safety issues.
2335 [Bodo Moeller]
2336
2337 Changes between 0.9.7i and 0.9.7j [04 May 2006]
2338
2339 *) Adapt fipsld and the build system to link against the validated FIPS
2340 module in FIPS mode.
2341 [Steve Henson]
2342
2343 *) Fixes for VC++ 2005 build under Windows.
2344 [Steve Henson]
2345
2346 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
2347 from a Windows bash shell such as MSYS. It is autodetected from the
2348 "config" script when run from a VC++ environment. Modify standard VC++
2349 build to use fipscanister.o from the GNU make build.
2350 [Steve Henson]
2351
2352 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
2353
2354 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
2355 The value now differs depending on if you build for FIPS or not.
2356 BEWARE! A program linked with a shared FIPSed libcrypto can't be
2357 safely run with a non-FIPSed libcrypto, as it may crash because of
2358 the difference induced by this change.
2359 [Andy Polyakov]
2360
2361 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
2362
2363 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2364 (part of SSL_OP_ALL). This option used to disable the
2365 countermeasure against man-in-the-middle protocol-version
2366 rollback in the SSL 2.0 server implementation, which is a bad
2367 idea. (CVE-2005-2969)
2368
2369 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2370 for Information Security, National Institute of Advanced Industrial
2371 Science and Technology [AIST], Japan)]
2372
2373 *) Minimal support for X9.31 signatures and PSS padding modes. This is
2374 mainly for FIPS compliance and not fully integrated at this stage.
2375 [Steve Henson]
2376
2377 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
2378 the exponentiation using a fixed-length exponent. (Otherwise,
2379 the information leaked through timing could expose the secret key
2380 after many signatures; cf. Bleichenbacher's attack on DSA with
2381 biased k.)
2382 [Bodo Moeller]
2383
2384 *) Make a new fixed-window mod_exp implementation the default for
2385 RSA, DSA, and DH private-key operations so that the sequence of
2386 squares and multiplies and the memory access pattern are
2387 independent of the particular secret key. This will mitigate
2388 cache-timing and potential related attacks.
2389
2390 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
2391 and this is automatically used by BN_mod_exp_mont() if the new flag
2392 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
2393 will use this BN flag for private exponents unless the flag
2394 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
2395 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
2396
2397 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
2398
2399 *) Change the client implementation for SSLv23_method() and
2400 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
2401 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
2402 (Previously, the SSL 2.0 backwards compatible Client Hello
2403 message format would be used even with SSL_OP_NO_SSLv2.)
2404 [Bodo Moeller]
2405
2406 *) Add support for smime-type MIME parameter in S/MIME messages which some
2407 clients need.
2408 [Steve Henson]
2409
2410 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
2411 a threadsafe manner. Modify rsa code to use new function and add calls
2412 to dsa and dh code (which had race conditions before).
2413 [Steve Henson]
2414
2415 *) Include the fixed error library code in the C error file definitions
2416 instead of fixing them up at runtime. This keeps the error code
2417 structures constant.
2418 [Steve Henson]
2419
2420 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
2421
2422 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
2423 OpenSSL 0.9.8.]
2424
2425 *) Fixes for newer kerberos headers. NB: the casts are needed because
2426 the 'length' field is signed on one version and unsigned on another
2427 with no (?) obvious way to tell the difference, without these VC++
2428 complains. Also the "definition" of FAR (blank) is no longer included
2429 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
2430 some needed definitions.
2431 [Steve Henson]
2432
2433 *) Undo Cygwin change.
2434 [Ulf Möller]
2435
2436 *) Added support for proxy certificates according to RFC 3820.
2437 Because they may be a security thread to unaware applications,
2438 they must be explicitely allowed in run-time. See
2439 docs/HOWTO/proxy_certificates.txt for further information.
2440 [Richard Levitte]
2441
2442 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
2443
2444 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
2445 server and client random values. Previously
2446 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
2447 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
2448
2449 This change has negligible security impact because:
2450
2451 1. Server and client random values still have 24 bytes of pseudo random
2452 data.
2453
2454 2. Server and client random values are sent in the clear in the initial
2455 handshake.
2456
2457 3. The master secret is derived using the premaster secret (48 bytes in
2458 size for static RSA ciphersuites) as well as client server and random
2459 values.
2460
2461 The OpenSSL team would like to thank the UK NISCC for bringing this issue
2462 to our attention.
2463
2464 [Stephen Henson, reported by UK NISCC]
2465
2466 *) Use Windows randomness collection on Cygwin.
2467 [Ulf Möller]
2468
2469 *) Fix hang in EGD/PRNGD query when communication socket is closed
2470 prematurely by EGD/PRNGD.
2471 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
2472
2473 *) Prompt for pass phrases when appropriate for PKCS12 input format.
2474 [Steve Henson]
2475
2476 *) Back-port of selected performance improvements from development
2477 branch, as well as improved support for PowerPC platforms.
2478 [Andy Polyakov]
2479
2480 *) Add lots of checks for memory allocation failure, error codes to indicate
2481 failure and freeing up memory if a failure occurs.
2482 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
2483
2484 *) Add new -passin argument to dgst.
2485 [Steve Henson]
2486
2487 *) Perform some character comparisons of different types in X509_NAME_cmp:
2488 this is needed for some certificates that reencode DNs into UTF8Strings
2489 (in violation of RFC3280) and can't or wont issue name rollover
2490 certificates.
2491 [Steve Henson]
2492
2493 *) Make an explicit check during certificate validation to see that
2494 the CA setting in each certificate on the chain is correct. As a
2495 side effect always do the following basic checks on extensions,
2496 not just when there's an associated purpose to the check:
2497
2498 - if there is an unhandled critical extension (unless the user
2499 has chosen to ignore this fault)
2500 - if the path length has been exceeded (if one is set at all)
2501 - that certain extensions fit the associated purpose (if one has
2502 been given)
2503 [Richard Levitte]
2504
2505 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
2506
2507 *) Avoid a race condition when CRLs are checked in a multi threaded
2508 environment. This would happen due to the reordering of the revoked
2509 entries during signature checking and serial number lookup. Now the
2510 encoding is cached and the serial number sort performed under a lock.
2511 Add new STACK function sk_is_sorted().
2512 [Steve Henson]
2513
2514 *) Add Delta CRL to the extension code.
2515 [Steve Henson]
2516
2517 *) Various fixes to s3_pkt.c so alerts are sent properly.
2518 [David Holmes <d.holmes@f5.com>]
2519
2520 *) Reduce the chances of duplicate issuer name and serial numbers (in
2521 violation of RFC3280) using the OpenSSL certificate creation utilities.
2522 This is done by creating a random 64 bit value for the initial serial
2523 number when a serial number file is created or when a self signed
2524 certificate is created using 'openssl req -x509'. The initial serial
2525 number file is created using 'openssl x509 -next_serial' in CA.pl
2526 rather than being initialized to 1.
2527 [Steve Henson]
2528
2529 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
2530
2531 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2532 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
2533 [Joe Orton, Steve Henson]
2534
2535 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
2536 (CVE-2004-0112)
2537 [Joe Orton, Steve Henson]
2538
2539 *) Make it possible to have multiple active certificates with the same
2540 subject in the CA index file. This is done only if the keyword
2541 'unique_subject' is set to 'no' in the main CA section (default
2542 if 'CA_default') of the configuration file. The value is saved
2543 with the database itself in a separate index attribute file,
2544 named like the index file with '.attr' appended to the name.
2545 [Richard Levitte]
2546
2547 *) X509 verify fixes. Disable broken certificate workarounds when
2548 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
2549 keyUsage extension present. Don't accept CRLs with unhandled critical
2550 extensions: since verify currently doesn't process CRL extensions this
2551 rejects a CRL with *any* critical extensions. Add new verify error codes
2552 for these cases.
2553 [Steve Henson]
2554
2555 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
2556 A clarification of RFC2560 will require the use of OCTET STRINGs and
2557 some implementations cannot handle the current raw format. Since OpenSSL
2558 copies and compares OCSP nonces as opaque blobs without any attempt at
2559 parsing them this should not create any compatibility issues.
2560 [Steve Henson]
2561
2562 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
2563 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
2564 this HMAC (and other) operations are several times slower than OpenSSL
2565 < 0.9.7.
2566 [Steve Henson]
2567
2568 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
2569 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2570
2571 *) Use the correct content when signing type "other".
2572 [Steve Henson]
2573
2574 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
2575
2576 *) Fix various bugs revealed by running the NISCC test suite:
2577
2578 Stop out of bounds reads in the ASN1 code when presented with
2579 invalid tags (CVE-2003-0543 and CVE-2003-0544).
2580
2581 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
2582
2583 If verify callback ignores invalid public key errors don't try to check
2584 certificate signature with the NULL public key.
2585
2586 [Steve Henson]
2587
2588 *) New -ignore_err option in ocsp application to stop the server
2589 exiting on the first error in a request.
2590 [Steve Henson]
2591
2592 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2593 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2594 specifications.
2595 [Steve Henson]
2596
2597 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2598 extra data after the compression methods not only for TLS 1.0
2599 but also for SSL 3.0 (as required by the specification).
2600 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2601
2602 *) Change X509_certificate_type() to mark the key as exported/exportable
2603 when it's 512 *bits* long, not 512 bytes.
2604 [Richard Levitte]
2605
2606 *) Change AES_cbc_encrypt() so it outputs exact multiple of
2607 blocks during encryption.
2608 [Richard Levitte]
2609
2610 *) Various fixes to base64 BIO and non blocking I/O. On write
2611 flushes were not handled properly if the BIO retried. On read
2612 data was not being buffered properly and had various logic bugs.
2613 This also affects blocking I/O when the data being decoded is a
2614 certain size.
2615 [Steve Henson]
2616
2617 *) Various S/MIME bugfixes and compatibility changes:
2618 output correct application/pkcs7 MIME type if
2619 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2620 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2621 of files as .eml work). Correctly handle very long lines in MIME
2622 parser.
2623 [Steve Henson]
2624
2625 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2626
2627 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2628 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2629 a protocol version number mismatch like a decryption error
2630 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2631 [Bodo Moeller]
2632
2633 *) Turn on RSA blinding by default in the default implementation
2634 to avoid a timing attack. Applications that don't want it can call
2635 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2636 They would be ill-advised to do so in most cases.
2637 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2638
2639 *) Change RSA blinding code so that it works when the PRNG is not
2640 seeded (in this case, the secret RSA exponent is abused as
2641 an unpredictable seed -- if it is not unpredictable, there
2642 is no point in blinding anyway). Make RSA blinding thread-safe
2643 by remembering the creator's thread ID in rsa->blinding and
2644 having all other threads use local one-time blinding factors
2645 (this requires more computation than sharing rsa->blinding, but
2646 avoids excessive locking; and if an RSA object is not shared
2647 between threads, blinding will still be very fast).
2648 [Bodo Moeller]
2649
2650 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2651 ENGINE as defaults for all supported algorithms irrespective of
2652 the 'flags' parameter. 'flags' is now honoured, so applications
2653 should make sure they are passing it correctly.
2654 [Geoff Thorpe]
2655
2656 *) Target "mingw" now allows native Windows code to be generated in
2657 the Cygwin environment as well as with the MinGW compiler.
2658 [Ulf Moeller]
2659
2660 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2661
2662 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2663 via timing by performing a MAC computation even if incorrrect
2664 block cipher padding has been found. This is a countermeasure
2665 against active attacks where the attacker has to distinguish
2666 between bad padding and a MAC verification error. (CVE-2003-0078)
2667
2668 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2669 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2670 Martin Vuagnoux (EPFL, Ilion)]
2671
2672 *) Make the no-err option work as intended. The intention with no-err
2673 is not to have the whole error stack handling routines removed from
2674 libcrypto, it's only intended to remove all the function name and
2675 reason texts, thereby removing some of the footprint that may not
2676 be interesting if those errors aren't displayed anyway.
2677
2678 NOTE: it's still possible for any application or module to have it's
2679 own set of error texts inserted. The routines are there, just not
2680 used by default when no-err is given.
2681 [Richard Levitte]
2682
2683 *) Add support for FreeBSD on IA64.
2684 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2685
2686 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2687 Kerberos function mit_des_cbc_cksum(). Before this change,
2688 the value returned by DES_cbc_cksum() was like the one from
2689 mit_des_cbc_cksum(), except the bytes were swapped.
2690 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2691
2692 *) Allow an application to disable the automatic SSL chain building.
2693 Before this a rather primitive chain build was always performed in
2694 ssl3_output_cert_chain(): an application had no way to send the
2695 correct chain if the automatic operation produced an incorrect result.
2696
2697 Now the chain builder is disabled if either:
2698
2699 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2700
2701 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2702
2703 The reasoning behind this is that an application would not want the
2704 auto chain building to take place if extra chain certificates are
2705 present and it might also want a means of sending no additional
2706 certificates (for example the chain has two certificates and the
2707 root is omitted).
2708 [Steve Henson]
2709
2710 *) Add the possibility to build without the ENGINE framework.
2711 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2712
2713 *) Under Win32 gmtime() can return NULL: check return value in
2714 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2715 [Steve Henson]
2716
2717 *) DSA routines: under certain error conditions uninitialized BN objects
2718 could be freed. Solution: make sure initialization is performed early
2719 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2720 Nils Larsch <nla@trustcenter.de> via PR#459)
2721 [Lutz Jaenicke]
2722
2723 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2724 checked on reconnect on the client side, therefore session resumption
2725 could still fail with a "ssl session id is different" error. This
2726 behaviour is masked when SSL_OP_ALL is used due to
2727 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2728 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2729 followup to PR #377.
2730 [Lutz Jaenicke]
2731
2732 *) IA-32 assembler support enhancements: unified ELF targets, support
2733 for SCO/Caldera platforms, fix for Cygwin shared build.
2734 [Andy Polyakov]
2735
2736 *) Add support for FreeBSD on sparc64. As a consequence, support for
2737 FreeBSD on non-x86 processors is separate from x86 processors on
2738 the config script, much like the NetBSD support.
2739 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2740
2741 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2742
2743 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2744 OpenSSL 0.9.7.]
2745
2746 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2747 code (06) was taken as the first octet of the session ID and the last
2748 octet was ignored consequently. As a result SSLv2 client side session
2749 caching could not have worked due to the session ID mismatch between
2750 client and server.
2751 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2752 PR #377.
2753 [Lutz Jaenicke]
2754
2755 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2756 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2757 removed entirely.
2758 [Richard Levitte]
2759
2760 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2761 seems that in spite of existing for more than a year, many application
2762 author have done nothing to provide the necessary callbacks, which
2763 means that this particular engine will not work properly anywhere.
2764 This is a very unfortunate situation which forces us, in the name
2765 of usability, to give the hw_ncipher.c a static lock, which is part
2766 of libcrypto.
2767 NOTE: This is for the 0.9.7 series ONLY. This hack will never
2768 appear in 0.9.8 or later. We EXPECT application authors to have
2769 dealt properly with this when 0.9.8 is released (unless we actually
2770 make such changes in the libcrypto locking code that changes will
2771 have to be made anyway).
2772 [Richard Levitte]
2773
2774 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
2775 octets have been read, EOF or an error occurs. Without this change
2776 some truncated ASN1 structures will not produce an error.
2777 [Steve Henson]
2778
2779 *) Disable Heimdal support, since it hasn't been fully implemented.
2780 Still give the possibility to force the use of Heimdal, but with
2781 warnings and a request that patches get sent to openssl-dev.
2782 [Richard Levitte]
2783
2784 *) Add the VC-CE target, introduce the WINCE sysname, and add
2785 INSTALL.WCE and appropriate conditionals to make it build.
2786 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2787
2788 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
2789 cygssl-x.y.z.dll, where x, y and z are the major, minor and
2790 edit numbers of the version.
2791 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2792
2793 *) Introduce safe string copy and catenation functions
2794 (BUF_strlcpy() and BUF_strlcat()).
2795 [Ben Laurie (CHATS) and Richard Levitte]
2796
2797 *) Avoid using fixed-size buffers for one-line DNs.
2798 [Ben Laurie (CHATS)]
2799
2800 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2801 resizing buffers containing secrets, and use where appropriate.
2802 [Ben Laurie (CHATS)]
2803
2804 *) Avoid using fixed size buffers for configuration file location.
2805 [Ben Laurie (CHATS)]
2806
2807 *) Avoid filename truncation for various CA files.
2808 [Ben Laurie (CHATS)]
2809
2810 *) Use sizeof in preference to magic numbers.
2811 [Ben Laurie (CHATS)]
2812
2813 *) Avoid filename truncation in cert requests.
2814 [Ben Laurie (CHATS)]
2815
2816 *) Add assertions to check for (supposedly impossible) buffer
2817 overflows.
2818 [Ben Laurie (CHATS)]
2819
2820 *) Don't cache truncated DNS entries in the local cache (this could
2821 potentially lead to a spoofing attack).
2822 [Ben Laurie (CHATS)]
2823
2824 *) Fix various buffers to be large enough for hex/decimal
2825 representations in a platform independent manner.
2826 [Ben Laurie (CHATS)]
2827
2828 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2829 resizing buffers containing secrets, and use where appropriate.
2830 [Ben Laurie (CHATS)]
2831
2832 *) Add BIO_indent() to avoid much slightly worrying code to do
2833 indents.
2834 [Ben Laurie (CHATS)]
2835
2836 *) Convert sprintf()/BIO_puts() to BIO_printf().
2837 [Ben Laurie (CHATS)]
2838
2839 *) buffer_gets() could terminate with the buffer only half
2840 full. Fixed.
2841 [Ben Laurie (CHATS)]
2842
2843 *) Add assertions to prevent user-supplied crypto functions from
2844 overflowing internal buffers by having large block sizes, etc.
2845 [Ben Laurie (CHATS)]
2846
2847 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2848 unconditionally).
2849 [Ben Laurie (CHATS)]
2850
2851 *) Eliminate unused copy of key in RC4.
2852 [Ben Laurie (CHATS)]
2853
2854 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2855 [Ben Laurie (CHATS)]
2856
2857 *) Fix off-by-one error in EGD path.
2858 [Ben Laurie (CHATS)]
2859
2860 *) If RANDFILE path is too long, ignore instead of truncating.
2861 [Ben Laurie (CHATS)]
2862
2863 *) Eliminate unused and incorrectly sized X.509 structure
2864 CBCParameter.
2865 [Ben Laurie (CHATS)]
2866
2867 *) Eliminate unused and dangerous function knumber().
2868 [Ben Laurie (CHATS)]
2869
2870 *) Eliminate unused and dangerous structure, KSSL_ERR.
2871 [Ben Laurie (CHATS)]
2872
2873 *) Protect against overlong session ID context length in an encoded
2874 session object. Since these are local, this does not appear to be
2875 exploitable.
2876 [Ben Laurie (CHATS)]
2877
2878 *) Change from security patch (see 0.9.6e below) that did not affect
2879 the 0.9.6 release series:
2880
2881 Remote buffer overflow in SSL3 protocol - an attacker could
2882 supply an oversized master key in Kerberos-enabled versions.
2883 (CVE-2002-0657)
2884 [Ben Laurie (CHATS)]
2885
2886 *) Change the SSL kerb5 codes to match RFC 2712.
2887 [Richard Levitte]
2888
2889 *) Make -nameopt work fully for req and add -reqopt switch.
2890 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2891
2892 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2893 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2894
2895 *) Make sure tests can be performed even if the corresponding algorithms
2896 have been removed entirely. This was also the last step to make
2897 OpenSSL compilable with DJGPP under all reasonable conditions.
2898 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2899
2900 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2901 to allow version independent disabling of normally unselected ciphers,
2902 which may be activated as a side-effect of selecting a single cipher.
2903
2904 (E.g., cipher list string "RSA" enables ciphersuites that are left
2905 out of "ALL" because they do not provide symmetric encryption.
2906 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2907 [Lutz Jaenicke, Bodo Moeller]
2908
2909 *) Add appropriate support for separate platform-dependent build
2910 directories. The recommended way to make a platform-dependent
2911 build directory is the following (tested on Linux), maybe with
2912 some local tweaks:
2913
2914 # Place yourself outside of the OpenSSL source tree. In
2915 # this example, the environment variable OPENSSL_SOURCE
2916 # is assumed to contain the absolute OpenSSL source directory.
2917 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2918 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2919 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2920 mkdir -p `dirname $F`
2921 ln -s $OPENSSL_SOURCE/$F $F
2922 done
2923
2924 To be absolutely sure not to disturb the source tree, a "make clean"
2925 is a good thing. If it isn't successfull, don't worry about it,
2926 it probably means the source directory is very clean.
2927 [Richard Levitte]
2928
2929 *) Make sure any ENGINE control commands make local copies of string
2930 pointers passed to them whenever necessary. Otherwise it is possible
2931 the caller may have overwritten (or deallocated) the original string
2932 data when a later ENGINE operation tries to use the stored values.
2933 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2934
2935 *) Improve diagnostics in file reading and command-line digests.
2936 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2937
2938 *) Add AES modes CFB and OFB to the object database. Correct an
2939 error in AES-CFB decryption.
2940 [Richard Levitte]
2941
2942 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2943 allows existing EVP_CIPHER_CTX structures to be reused after
2944 calling EVP_*Final(). This behaviour is used by encryption
2945 BIOs and some applications. This has the side effect that
2946 applications must explicitly clean up cipher contexts with
2947 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2948 [Steve Henson]
2949
2950 *) Check the values of dna and dnb in bn_mul_recursive before calling
2951 bn_mul_comba (a non zero value means the a or b arrays do not contain
2952 n2 elements) and fallback to bn_mul_normal if either is not zero.
2953 [Steve Henson]
2954
2955 *) Fix escaping of non-ASCII characters when using the -subj option
2956 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2957 [Lutz Jaenicke]
2958
2959 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2960 form for "surname", serialNumber has no short form.
2961 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2962 therefore remove "mail" short name for "internet 7".
2963 The OID for unique identifiers in X509 certificates is
2964 x500UniqueIdentifier, not uniqueIdentifier.
2965 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2966 [Lutz Jaenicke]
2967
2968 *) Add an "init" command to the ENGINE config module and auto initialize
2969 ENGINEs. Without any "init" command the ENGINE will be initialized
2970 after all ctrl commands have been executed on it. If init=1 the
2971 ENGINE is initailized at that point (ctrls before that point are run
2972 on the uninitialized ENGINE and after on the initialized one). If
2973 init=0 then the ENGINE will not be iniatialized at all.
2974 [Steve Henson]
2975
2976 *) Fix the 'app_verify_callback' interface so that the user-defined
2977 argument is actually passed to the callback: In the
2978 SSL_CTX_set_cert_verify_callback() prototype, the callback
2979 declaration has been changed from
2980 int (*cb)()
2981 into
2982 int (*cb)(X509_STORE_CTX *,void *);
2983 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2984 i=s->ctx->app_verify_callback(&ctx)
2985 has been changed into
2986 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2987
2988 To update applications using SSL_CTX_set_cert_verify_callback(),
2989 a dummy argument can be added to their callback functions.
2990 [D. K. Smetters <smetters@parc.xerox.com>]
2991
2992 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2993 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2994
2995 *) Add and OPENSSL_LOAD_CONF define which will cause
2996 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2997 This allows older applications to transparently support certain
2998 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2999 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3000 load the config file and OPENSSL_add_all_algorithms_conf() which will
3001 always load it have also been added.
3002 [Steve Henson]
3003
3004 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3005 Adjust NIDs and EVP layer.
3006 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3007
3008 *) Config modules support in openssl utility.
3009
3010 Most commands now load modules from the config file,
3011 though in a few (such as version) this isn't done
3012 because it couldn't be used for anything.
3013
3014 In the case of ca and req the config file used is
3015 the same as the utility itself: that is the -config
3016 command line option can be used to specify an
3017 alternative file.
3018 [Steve Henson]
3019
3020 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3021 use "openssl_conf" if filename is NULL use default openssl config file.
3022 [Steve Henson]
3023
3024 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3025 config section name. Add a new flag to tolerate a missing config file
3026 and move code to CONF_modules_load_file().
3027 [Steve Henson]
3028
3029 *) Support for crypto accelerator cards from Accelerated Encryption
3030 Processing, www.aep.ie. (Use engine 'aep')
3031 The support was copied from 0.9.6c [engine] and adapted/corrected
3032 to work with the new engine framework.
3033 [AEP Inc. and Richard Levitte]
3034
3035 *) Support for SureWare crypto accelerator cards from Baltimore
3036 Technologies. (Use engine 'sureware')
3037 The support was copied from 0.9.6c [engine] and adapted
3038 to work with the new engine framework.
3039 [Richard Levitte]
3040
3041 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3042 make the newer ENGINE framework commands for the CHIL engine work.
3043 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3044
3045 *) Make it possible to produce shared libraries on ReliantUNIX.
3046 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3047
3048 *) Add the configuration target debug-linux-ppro.
3049 Make 'openssl rsa' use the general key loading routines
3050 implemented in apps.c, and make those routines able to
3051 handle the key format FORMAT_NETSCAPE and the variant
3052 FORMAT_IISSGC.
3053 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3054
3055 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3056 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3057
3058 *) Add -keyform to rsautl, and document -engine.
3059 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3060
3061 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3062 BIO_R_NO_SUCH_FILE error code rather than the generic
3063 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3064 [Ben Laurie]
3065
3066 *) Add new functions
3067 ERR_peek_last_error
3068 ERR_peek_last_error_line
3069 ERR_peek_last_error_line_data.
3070 These are similar to
3071 ERR_peek_error
3072 ERR_peek_error_line
3073 ERR_peek_error_line_data,
3074 but report on the latest error recorded rather than the first one
3075 still in the error queue.
3076 [Ben Laurie, Bodo Moeller]
3077
3078 *) default_algorithms option in ENGINE config module. This allows things
3079 like:
3080 default_algorithms = ALL
3081 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3082 [Steve Henson]
3083
3084 *) Prelminary ENGINE config module.
3085 [Steve Henson]
3086
3087 *) New experimental application configuration code.
3088 [Steve Henson]
3089
3090 *) Change the AES code to follow the same name structure as all other
3091 symmetric ciphers, and behave the same way. Move everything to
3092 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3093 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3094
3095 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3096 [Ben Laurie and Theo de Raadt]
3097
3098 *) Add option to output public keys in req command.
3099 [Massimiliano Pala madwolf@openca.org]
3100
3101 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3102 (up to about 10% better than before for P-192 and P-224).
3103 [Bodo Moeller]
3104
3105 *) New functions/macros
3106
3107 SSL_CTX_set_msg_callback(ctx, cb)
3108 SSL_CTX_set_msg_callback_arg(ctx, arg)
3109 SSL_set_msg_callback(ssl, cb)
3110 SSL_set_msg_callback_arg(ssl, arg)
3111
3112 to request calling a callback function
3113
3114 void cb(int write_p, int version, int content_type,
3115 const void *buf, size_t len, SSL *ssl, void *arg)
3116
3117 whenever a protocol message has been completely received
3118 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3119 protocol version according to which the SSL library interprets
3120 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3121 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3122 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3123 specification (change_cipher_spec(20), alert(21), handshake(22)).
3124 'buf' and 'len' point to the actual message, 'ssl' to the
3125 SSL object, and 'arg' is the application-defined value set by
3126 SSL[_CTX]_set_msg_callback_arg().
3127
3128 'openssl s_client' and 'openssl s_server' have new '-msg' options
3129 to enable a callback that displays all protocol messages.
3130 [Bodo Moeller]
3131
3132 *) Change the shared library support so shared libraries are built as
3133 soon as the corresponding static library is finished, and thereby get
3134 openssl and the test programs linked against the shared library.
3135 This still only happens when the keyword "shard" has been given to
3136 the configuration scripts.
3137
3138 NOTE: shared library support is still an experimental thing, and
3139 backward binary compatibility is still not guaranteed.
3140 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3141
3142 *) Add support for Subject Information Access extension.
3143 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3144
3145 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3146 additional bytes when new memory had to be allocated, not just
3147 when reusing an existing buffer.
3148 [Bodo Moeller]
3149
3150 *) New command line and configuration option 'utf8' for the req command.
3151 This allows field values to be specified as UTF8 strings.
3152 [Steve Henson]
3153
3154 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
3155 runs for the former and machine-readable output for the latter.
3156 [Ben Laurie]
3157
3158 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
3159 of the e-mail address in the DN (i.e., it will go into a certificate
3160 extension only). The new configuration file option 'email_in_dn = no'
3161 has the same effect.
3162 [Massimiliano Pala madwolf@openca.org]
3163
3164 *) Change all functions with names starting with des_ to be starting
3165 with DES_ instead. Add wrappers that are compatible with libdes,
3166 but are named _ossl_old_des_*. Finally, add macros that map the
3167 des_* symbols to the corresponding _ossl_old_des_* if libdes
3168 compatibility is desired. If OpenSSL 0.9.6c compatibility is
3169 desired, the des_* symbols will be mapped to DES_*, with one
3170 exception.
3171
3172 Since we provide two compatibility mappings, the user needs to
3173 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
3174 compatibility is desired. The default (i.e., when that macro
3175 isn't defined) is OpenSSL 0.9.6c compatibility.
3176
3177 There are also macros that enable and disable the support of old
3178 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
3179 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
3180 are defined, the default will apply: to support the old des routines.
3181
3182 In either case, one must include openssl/des.h to get the correct
3183 definitions. Do not try to just include openssl/des_old.h, that
3184 won't work.
3185
3186 NOTE: This is a major break of an old API into a new one. Software
3187 authors are encouraged to switch to the DES_ style functions. Some
3188 time in the future, des_old.h and the libdes compatibility functions
3189 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
3190 default), and then completely removed.
3191 [Richard Levitte]
3192
3193 *) Test for certificates which contain unsupported critical extensions.
3194 If such a certificate is found during a verify operation it is
3195 rejected by default: this behaviour can be overridden by either
3196 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
3197 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
3198 X509_supported_extension() has also been added which returns 1 if a
3199 particular extension is supported.
3200 [Steve Henson]
3201
3202 *) Modify the behaviour of EVP cipher functions in similar way to digests
3203 to retain compatibility with existing code.
3204 [Steve Henson]
3205
3206 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
3207 compatibility with existing code. In particular the 'ctx' parameter does
3208 not have to be to be initialized before the call to EVP_DigestInit() and
3209 it is tidied up after a call to EVP_DigestFinal(). New function
3210 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
3211 EVP_MD_CTX_copy() changed to not require the destination to be
3212 initialized valid and new function EVP_MD_CTX_copy_ex() added which
3213 requires the destination to be valid.
3214
3215 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
3216 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
3217 [Steve Henson]
3218
3219 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
3220 so that complete 'Handshake' protocol structures are kept in memory
3221 instead of overwriting 'msg_type' and 'length' with 'body' data.
3222 [Bodo Moeller]
3223
3224 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
3225 [Massimo Santin via Richard Levitte]
3226
3227 *) Major restructuring to the underlying ENGINE code. This includes
3228 reduction of linker bloat, separation of pure "ENGINE" manipulation
3229 (initialisation, etc) from functionality dealing with implementations
3230 of specific crypto iterfaces. This change also introduces integrated
3231 support for symmetric ciphers and digest implementations - so ENGINEs
3232 can now accelerate these by providing EVP_CIPHER and EVP_MD
3233 implementations of their own. This is detailed in crypto/engine/README
3234 as it couldn't be adequately described here. However, there are a few
3235 API changes worth noting - some RSA, DSA, DH, and RAND functions that
3236 were changed in the original introduction of ENGINE code have now
3237 reverted back - the hooking from this code to ENGINE is now a good
3238 deal more passive and at run-time, operations deal directly with
3239 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
3240 dereferencing through an ENGINE pointer any more. Also, the ENGINE
3241 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
3242 they were not being used by the framework as there is no concept of a
3243 BIGNUM_METHOD and they could not be generalised to the new
3244 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
3245 ENGINE_cpy() has been removed as it cannot be consistently defined in
3246 the new code.
3247 [Geoff Thorpe]
3248
3249 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
3250 [Steve Henson]
3251
3252 *) Change mkdef.pl to sort symbols that get the same entry number,
3253 and make sure the automatically generated functions ERR_load_*
3254 become part of libeay.num as well.
3255 [Richard Levitte]
3256
3257 *) New function SSL_renegotiate_pending(). This returns true once
3258 renegotiation has been requested (either SSL_renegotiate() call
3259 or HelloRequest/ClientHello receveived from the peer) and becomes
3260 false once a handshake has been completed.
3261 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
3262 sends a HelloRequest, but does not ensure that a handshake takes
3263 place. SSL_renegotiate_pending() is useful for checking if the
3264 client has followed the request.)
3265 [Bodo Moeller]
3266
3267 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
3268 By default, clients may request session resumption even during
3269 renegotiation (if session ID contexts permit); with this option,
3270 session resumption is possible only in the first handshake.
3271
3272 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
3273 more bits available for options that should not be part of
3274 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
3275 [Bodo Moeller]
3276
3277 *) Add some demos for certificate and certificate request creation.
3278 [Steve Henson]
3279
3280 *) Make maximum certificate chain size accepted from the peer application
3281 settable (SSL*_get/set_max_cert_list()), as proposed by
3282 "Douglas E. Engert" <deengert@anl.gov>.
3283 [Lutz Jaenicke]
3284
3285 *) Add support for shared libraries for Unixware-7
3286 (Boyd Lynn Gerber <gerberb@zenez.com>).
3287 [Lutz Jaenicke]
3288
3289 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
3290 be done prior to destruction. Use this to unload error strings from
3291 ENGINEs that load their own error strings. NB: This adds two new API
3292 functions to "get" and "set" this destroy handler in an ENGINE.
3293 [Geoff Thorpe]
3294
3295 *) Alter all existing ENGINE implementations (except "openssl" and
3296 "openbsd") to dynamically instantiate their own error strings. This
3297 makes them more flexible to be built both as statically-linked ENGINEs
3298 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
3299 Also, add stub code to each that makes building them as self-contained
3300 shared-libraries easier (see README.ENGINE).
3301 [Geoff Thorpe]
3302
3303 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
3304 implementations into applications that are completely implemented in
3305 self-contained shared-libraries. The "dynamic" ENGINE exposes control
3306 commands that can be used to configure what shared-library to load and
3307 to control aspects of the way it is handled. Also, made an update to
3308 the README.ENGINE file that brings its information up-to-date and
3309 provides some information and instructions on the "dynamic" ENGINE
3310 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
3311 [Geoff Thorpe]
3312
3313 *) Make it possible to unload ranges of ERR strings with a new
3314 "ERR_unload_strings" function.
3315 [Geoff Thorpe]
3316
3317 *) Add a copy() function to EVP_MD.
3318 [Ben Laurie]
3319
3320 *) Make EVP_MD routines take a context pointer instead of just the
3321 md_data void pointer.
3322 [Ben Laurie]
3323
3324 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
3325 that the digest can only process a single chunk of data
3326 (typically because it is provided by a piece of
3327 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
3328 is only going to provide a single chunk of data, and hence the
3329 framework needn't accumulate the data for oneshot drivers.
3330 [Ben Laurie]
3331
3332 *) As with "ERR", make it possible to replace the underlying "ex_data"
3333 functions. This change also alters the storage and management of global
3334 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
3335 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
3336 index counters. The API functions that use this state have been changed
3337 to take a "class_index" rather than pointers to the class's local STACK
3338 and counter, and there is now an API function to dynamically create new
3339 classes. This centralisation allows us to (a) plug a lot of the
3340 thread-safety problems that existed, and (b) makes it possible to clean
3341 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
3342 such data would previously have always leaked in application code and
3343 workarounds were in place to make the memory debugging turn a blind eye
3344 to it. Application code that doesn't use this new function will still
3345 leak as before, but their memory debugging output will announce it now
3346 rather than letting it slide.
3347
3348 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
3349 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
3350 has a return value to indicate success or failure.
3351 [Geoff Thorpe]
3352
3353 *) Make it possible to replace the underlying "ERR" functions such that the
3354 global state (2 LHASH tables and 2 locks) is only used by the "default"
3355 implementation. This change also adds two functions to "get" and "set"
3356 the implementation prior to it being automatically set the first time
3357 any other ERR function takes place. Ie. an application can call "get",
3358 pass the return value to a module it has just loaded, and that module
3359 can call its own "set" function using that value. This means the
3360 module's "ERR" operations will use (and modify) the error state in the
3361 application and not in its own statically linked copy of OpenSSL code.
3362 [Geoff Thorpe]
3363
3364 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
3365 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
3366 the operation, and provides a more encapsulated way for external code
3367 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
3368 to use these functions rather than manually incrementing the counts.
3369
3370 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
3371 [Geoff Thorpe]
3372
3373 *) Add EVP test program.
3374 [Ben Laurie]
3375
3376 *) Add symmetric cipher support to ENGINE. Expect the API to change!
3377 [Ben Laurie]
3378
3379 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
3380 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
3381 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
3382 These allow a CRL to be built without having to access X509_CRL fields
3383 directly. Modify 'ca' application to use new functions.
3384 [Steve Henson]
3385
3386 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
3387 bug workarounds. Rollback attack detection is a security feature.
3388 The problem will only arise on OpenSSL servers when TLSv1 is not
3389 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
3390 Software authors not wanting to support TLSv1 will have special reasons
3391 for their choice and can explicitly enable this option.
3392 [Bodo Moeller, Lutz Jaenicke]
3393
3394 *) Rationalise EVP so it can be extended: don't include a union of
3395 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
3396 (similar to those existing for EVP_CIPHER_CTX).
3397 Usage example:
3398
3399 EVP_MD_CTX md;
3400
3401 EVP_MD_CTX_init(&md); /* new function call */
3402 EVP_DigestInit(&md, EVP_sha1());
3403 EVP_DigestUpdate(&md, in, len);
3404 EVP_DigestFinal(&md, out, NULL);
3405 EVP_MD_CTX_cleanup(&md); /* new function call */
3406
3407 [Ben Laurie]
3408
3409 *) Make DES key schedule conform to the usual scheme, as well as
3410 correcting its structure. This means that calls to DES functions
3411 now have to pass a pointer to a des_key_schedule instead of a
3412 plain des_key_schedule (which was actually always a pointer
3413 anyway): E.g.,
3414
3415 des_key_schedule ks;
3416
3417 des_set_key_checked(..., &ks);
3418 des_ncbc_encrypt(..., &ks, ...);
3419
3420 (Note that a later change renames 'des_...' into 'DES_...'.)
3421 [Ben Laurie]
3422
3423 *) Initial reduction of linker bloat: the use of some functions, such as
3424 PEM causes large amounts of unused functions to be linked in due to
3425 poor organisation. For example pem_all.c contains every PEM function
3426 which has a knock on effect of linking in large amounts of (unused)
3427 ASN1 code. Grouping together similar functions and splitting unrelated
3428 functions prevents this.
3429 [Steve Henson]
3430
3431 *) Cleanup of EVP macros.
3432 [Ben Laurie]
3433
3434 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
3435 correct _ecb suffix.
3436 [Ben Laurie]
3437
3438 *) Add initial OCSP responder support to ocsp application. The
3439 revocation information is handled using the text based index
3440 use by the ca application. The responder can either handle
3441 requests generated internally, supplied in files (for example
3442 via a CGI script) or using an internal minimal server.
3443 [Steve Henson]
3444
3445 *) Add configuration choices to get zlib compression for TLS.
3446 [Richard Levitte]
3447
3448 *) Changes to Kerberos SSL for RFC 2712 compliance:
3449 1. Implemented real KerberosWrapper, instead of just using
3450 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
3451 2. Implemented optional authenticator field of KerberosWrapper.
3452
3453 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
3454 and authenticator structs; see crypto/krb5/.
3455
3456 Generalized Kerberos calls to support multiple Kerberos libraries.
3457 [Vern Staats <staatsvr@asc.hpc.mil>,
3458 Jeffrey Altman <jaltman@columbia.edu>
3459 via Richard Levitte]
3460
3461 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
3462 already does with RSA. testdsa.h now has 'priv_key/pub_key'
3463 values for each of the key sizes rather than having just
3464 parameters (and 'speed' generating keys each time).
3465 [Geoff Thorpe]
3466
3467 *) Speed up EVP routines.
3468 Before:
3469 encrypt
3470 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
3471 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
3472 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
3473 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
3474 decrypt
3475 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
3476 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
3477 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
3478 After:
3479 encrypt
3480 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
3481 decrypt
3482 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
3483 [Ben Laurie]
3484
3485 *) Added the OS2-EMX target.
3486 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
3487
3488 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
3489 to support NCONF routines in extension code. New function CONF_set_nconf()
3490 to allow functions which take an NCONF to also handle the old LHASH
3491 structure: this means that the old CONF compatible routines can be
3492 retained (in particular wrt extensions) without having to duplicate the
3493 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
3494 [Steve Henson]
3495
3496 *) Enhance the general user interface with mechanisms for inner control
3497 and with possibilities to have yes/no kind of prompts.
3498 [Richard Levitte]
3499
3500 *) Change all calls to low level digest routines in the library and
3501 applications to use EVP. Add missing calls to HMAC_cleanup() and
3502 don't assume HMAC_CTX can be copied using memcpy().
3503 [Verdon Walker <VWalker@novell.com>, Steve Henson]
3504
3505 *) Add the possibility to control engines through control names but with
3506 arbitrary arguments instead of just a string.
3507 Change the key loaders to take a UI_METHOD instead of a callback
3508 function pointer. NOTE: this breaks binary compatibility with earlier
3509 versions of OpenSSL [engine].
3510 Adapt the nCipher code for these new conditions and add a card insertion
3511 callback.
3512 [Richard Levitte]
3513
3514 *) Enhance the general user interface with mechanisms to better support
3515 dialog box interfaces, application-defined prompts, the possibility
3516 to use defaults (for example default passwords from somewhere else)
3517 and interrupts/cancellations.
3518 [Richard Levitte]
3519
3520 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
3521 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
3522 [Steve Henson]
3523
3524 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
3525 tidy up some unnecessarily weird code in 'sk_new()').
3526 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
3527
3528 *) Change the key loading routines for ENGINEs to use the same kind
3529 callback (pem_password_cb) as all other routines that need this
3530 kind of callback.
3531 [Richard Levitte]
3532
3533 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
3534 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
3535 than this minimum value is recommended.
3536 [Lutz Jaenicke]
3537
3538 *) New random seeder for OpenVMS, using the system process statistics
3539 that are easily reachable.
3540 [Richard Levitte]
3541
3542 *) Windows apparently can't transparently handle global
3543 variables defined in DLLs. Initialisations such as:
3544
3545 const ASN1_ITEM *it = &ASN1_INTEGER_it;
3546
3547 wont compile. This is used by the any applications that need to
3548 declare their own ASN1 modules. This was fixed by adding the option
3549 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
3550 needed for static libraries under Win32.
3551 [Steve Henson]
3552
3553 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
3554 setting of purpose and trust fields. New X509_STORE trust and
3555 purpose functions and tidy up setting in other SSL functions.
3556 [Steve Henson]
3557
3558 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
3559 structure. These are inherited by X509_STORE_CTX when it is
3560 initialised. This allows various defaults to be set in the
3561 X509_STORE structure (such as flags for CRL checking and custom
3562 purpose or trust settings) for functions which only use X509_STORE_CTX
3563 internally such as S/MIME.
3564
3565 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
3566 trust settings if they are not set in X509_STORE. This allows X509_STORE
3567 purposes and trust (in S/MIME for example) to override any set by default.
3568
3569 Add command line options for CRL checking to smime, s_client and s_server
3570 applications.
3571 [Steve Henson]
3572
3573 *) Initial CRL based revocation checking. If the CRL checking flag(s)
3574 are set then the CRL is looked up in the X509_STORE structure and
3575 its validity and signature checked, then if the certificate is found
3576 in the CRL the verify fails with a revoked error.
3577
3578 Various new CRL related callbacks added to X509_STORE_CTX structure.
3579
3580 Command line options added to 'verify' application to support this.
3581
3582 This needs some additional work, such as being able to handle multiple
3583 CRLs with different times, extension based lookup (rather than just
3584 by subject name) and ultimately more complete V2 CRL extension
3585 handling.
3586 [Steve Henson]
3587
3588 *) Add a general user interface API (crypto/ui/). This is designed
3589 to replace things like des_read_password and friends (backward
3590 compatibility functions using this new API are provided).
3591 The purpose is to remove prompting functions from the DES code
3592 section as well as provide for prompting through dialog boxes in
3593 a window system and the like.
3594 [Richard Levitte]
3595
3596 *) Add "ex_data" support to ENGINE so implementations can add state at a
3597 per-structure level rather than having to store it globally.
3598 [Geoff]
3599
3600 *) Make it possible for ENGINE structures to be copied when retrieved by
3601 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
3602 This causes the "original" ENGINE structure to act like a template,
3603 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
3604 operational state can be localised to each ENGINE structure, despite the
3605 fact they all share the same "methods". New ENGINE structures returned in
3606 this case have no functional references and the return value is the single
3607 structural reference. This matches the single structural reference returned
3608 by ENGINE_by_id() normally, when it is incremented on the pre-existing
3609 ENGINE structure.
3610 [Geoff]
3611
3612 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3613 needs to match any other type at all we need to manually clear the
3614 tag cache.
3615 [Steve Henson]
3616
3617 *) Changes to the "openssl engine" utility to include;
3618 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3619 about an ENGINE's available control commands.
3620 - executing control commands from command line arguments using the
3621 '-pre' and '-post' switches. '-post' is only used if '-t' is
3622 specified and the ENGINE is successfully initialised. The syntax for
3623 the individual commands are colon-separated, for example;
3624 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3625 [Geoff]
3626
3627 *) New dynamic control command support for ENGINEs. ENGINEs can now
3628 declare their own commands (numbers), names (strings), descriptions,
3629 and input types for run-time discovery by calling applications. A
3630 subset of these commands are implicitly classed as "executable"
3631 depending on their input type, and only these can be invoked through
3632 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3633 can be based on user input, config files, etc). The distinction is
3634 that "executable" commands cannot return anything other than a boolean
3635 result and can only support numeric or string input, whereas some
3636 discoverable commands may only be for direct use through
3637 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3638 pointers, or other custom uses. The "executable" commands are to
3639 support parameterisations of ENGINE behaviour that can be
3640 unambiguously defined by ENGINEs and used consistently across any
3641 OpenSSL-based application. Commands have been added to all the
3642 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3643 control over shared-library paths without source code alterations.
3644 [Geoff]
3645
3646 *) Changed all ENGINE implementations to dynamically allocate their
3647 ENGINEs rather than declaring them statically. Apart from this being
3648 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3649 this also allows the implementations to compile without using the
3650 internal engine_int.h header.
3651 [Geoff]
3652
3653 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3654 'const' value. Any code that should be able to modify a RAND_METHOD
3655 should already have non-const pointers to it (ie. they should only
3656 modify their own ones).
3657 [Geoff]
3658
3659 *) Made a variety of little tweaks to the ENGINE code.
3660 - "atalla" and "ubsec" string definitions were moved from header files
3661 to C code. "nuron" string definitions were placed in variables
3662 rather than hard-coded - allowing parameterisation of these values
3663 later on via ctrl() commands.
3664 - Removed unused "#if 0"'d code.
3665 - Fixed engine list iteration code so it uses ENGINE_free() to release
3666 structural references.
3667 - Constified the RAND_METHOD element of ENGINE structures.
3668 - Constified various get/set functions as appropriate and added
3669 missing functions (including a catch-all ENGINE_cpy that duplicates
3670 all ENGINE values onto a new ENGINE except reference counts/state).
3671 - Removed NULL parameter checks in get/set functions. Setting a method
3672 or function to NULL is a way of cancelling out a previously set
3673 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3674 and doesn't justify the extra error symbols and code.
3675 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3676 flags from engine_int.h to engine.h.
3677 - Changed prototypes for ENGINE handler functions (init(), finish(),
3678 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3679 [Geoff]
3680
3681 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3682 to the algorithm using long division. The binary algorithm can be
3683 used only if the modulus is odd. On 32-bit systems, it is faster
3684 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3685 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3686 up to 450 bits. In 64-bit environments, the binary algorithm
3687 appears to be advantageous for much longer moduli; here we use it
3688 for moduli up to 2048 bits.
3689 [Bodo Moeller]
3690
3691 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3692 could not support the combine flag in choice fields.
3693 [Steve Henson]
3694
3695 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3696 extensions from a certificate request to the certificate.
3697 [Steve Henson]
3698
3699 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3700 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3701 file: this allows the display of the certificate about to be
3702 signed to be customised, to allow certain fields to be included
3703 or excluded and extension details. The old system didn't display
3704 multicharacter strings properly, omitted fields not in the policy
3705 and couldn't display additional details such as extensions.
3706 [Steve Henson]
3707
3708 *) Function EC_POINTs_mul for multiple scalar multiplication
3709 of an arbitrary number of elliptic curve points
3710 \sum scalars[i]*points[i],
3711 optionally including the generator defined for the EC_GROUP:
3712 scalar*generator + \sum scalars[i]*points[i].
3713
3714 EC_POINT_mul is a simple wrapper function for the typical case
3715 that the point list has just one item (besides the optional
3716 generator).
3717 [Bodo Moeller]
3718
3719 *) First EC_METHODs for curves over GF(p):
3720
3721 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3722 operations and provides various method functions that can also
3723 operate with faster implementations of modular arithmetic.
3724
3725 EC_GFp_mont_method() reuses most functions that are part of
3726 EC_GFp_simple_method, but uses Montgomery arithmetic.
3727
3728 [Bodo Moeller; point addition and point doubling
3729 implementation directly derived from source code provided by
3730 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3731
3732 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3733 crypto/ec/ec_lib.c):
3734
3735 Curves are EC_GROUP objects (with an optional group generator)
3736 based on EC_METHODs that are built into the library.
3737
3738 Points are EC_POINT objects based on EC_GROUP objects.
3739
3740 Most of the framework would be able to handle curves over arbitrary
3741 finite fields, but as there are no obvious types for fields other
3742 than GF(p), some functions are limited to that for now.
3743 [Bodo Moeller]
3744
3745 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3746 that the file contains a complete HTTP response.
3747 [Richard Levitte]
3748
3749 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3750 change the def and num file printf format specifier from "%-40sXXX"
3751 to "%-39s XXX". The latter will always guarantee a space after the
3752 field while the former will cause them to run together if the field
3753 is 40 of more characters long.
3754 [Steve Henson]
3755
3756 *) Constify the cipher and digest 'method' functions and structures
3757 and modify related functions to take constant EVP_MD and EVP_CIPHER
3758 pointers.
3759 [Steve Henson]
3760
3761 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
3762 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
3763 [Bodo Moeller]
3764
3765 *) Modify EVP_Digest*() routines so they now return values. Although the
3766 internal software routines can never fail additional hardware versions
3767 might.
3768 [Steve Henson]
3769
3770 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
3771
3772 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
3773 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
3774
3775 ASN1 error codes
3776 ERR_R_NESTED_ASN1_ERROR
3777 ...
3778 ERR_R_MISSING_ASN1_EOS
3779 were 4 .. 9, conflicting with
3780 ERR_LIB_RSA (= ERR_R_RSA_LIB)
3781 ...
3782 ERR_LIB_PEM (= ERR_R_PEM_LIB).
3783 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
3784
3785 Add new error code 'ERR_R_INTERNAL_ERROR'.
3786 [Bodo Moeller]
3787
3788 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3789 suffices.
3790 [Bodo Moeller]
3791
3792 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3793 sets the subject name for a new request or supersedes the
3794 subject name in a given request. Formats that can be parsed are
3795 'CN=Some Name, OU=myOU, C=IT'
3796 and
3797 'CN=Some Name/OU=myOU/C=IT'.
3798
3799 Add options '-batch' and '-verbose' to 'openssl req'.
3800 [Massimiliano Pala <madwolf@hackmasters.net>]
3801
3802 *) Introduce the possibility to access global variables through
3803 functions on platform were that's the best way to handle exporting
3804 global variables in shared libraries. To enable this functionality,
3805 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3806 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3807 is normally done by Configure or something similar).
3808
3809 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3810 in the source file (foo.c) like this:
3811
3812 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3813 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3814
3815 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3816 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3817
3818 OPENSSL_DECLARE_GLOBAL(int,foo);
3819 #define foo OPENSSL_GLOBAL_REF(foo)
3820 OPENSSL_DECLARE_GLOBAL(double,bar);
3821 #define bar OPENSSL_GLOBAL_REF(bar)
3822
3823 The #defines are very important, and therefore so is including the
3824 header file everywhere where the defined globals are used.
3825
3826 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3827 of ASN.1 items, but that structure is a bit different.
3828
3829 The largest change is in util/mkdef.pl which has been enhanced with
3830 better and easier to understand logic to choose which symbols should
3831 go into the Windows .def files as well as a number of fixes and code
3832 cleanup (among others, algorithm keywords are now sorted
3833 lexicographically to avoid constant rewrites).
3834 [Richard Levitte]
3835
3836 *) In BN_div() keep a copy of the sign of 'num' before writing the
3837 result to 'rm' because if rm==num the value will be overwritten
3838 and produce the wrong result if 'num' is negative: this caused
3839 problems with BN_mod() and BN_nnmod().
3840 [Steve Henson]
3841
3842 *) Function OCSP_request_verify(). This checks the signature on an
3843 OCSP request and verifies the signer certificate. The signer
3844 certificate is just checked for a generic purpose and OCSP request
3845 trust settings.
3846 [Steve Henson]
3847
3848 *) Add OCSP_check_validity() function to check the validity of OCSP
3849 responses. OCSP responses are prepared in real time and may only
3850 be a few seconds old. Simply checking that the current time lies
3851 between thisUpdate and nextUpdate max reject otherwise valid responses
3852 caused by either OCSP responder or client clock inaccuracy. Instead
3853 we allow thisUpdate and nextUpdate to fall within a certain period of
3854 the current time. The age of the response can also optionally be
3855 checked. Two new options -validity_period and -status_age added to
3856 ocsp utility.
3857 [Steve Henson]
3858
3859 *) If signature or public key algorithm is unrecognized print out its
3860 OID rather that just UNKNOWN.
3861 [Steve Henson]
3862
3863 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3864 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3865 ID to be generated from the issuer certificate alone which can then be
3866 passed to OCSP_id_issuer_cmp().
3867 [Steve Henson]
3868
3869 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3870 ASN1 modules to export functions returning ASN1_ITEM pointers
3871 instead of the ASN1_ITEM structures themselves. This adds several
3872 new macros which allow the underlying ASN1 function/structure to
3873 be accessed transparently. As a result code should not use ASN1_ITEM
3874 references directly (such as &X509_it) but instead use the relevant
3875 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3876 use of the new ASN1 code on platforms where exporting structures
3877 is problematical (for example in shared libraries) but exporting
3878 functions returning pointers to structures is not.
3879 [Steve Henson]
3880
3881 *) Add support for overriding the generation of SSL/TLS session IDs.
3882 These callbacks can be registered either in an SSL_CTX or per SSL.
3883 The purpose of this is to allow applications to control, if they wish,
3884 the arbitrary values chosen for use as session IDs, particularly as it
3885 can be useful for session caching in multiple-server environments. A
3886 command-line switch for testing this (and any client code that wishes
3887 to use such a feature) has been added to "s_server".
3888 [Geoff Thorpe, Lutz Jaenicke]
3889
3890 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3891 of the form '#if defined(...) || defined(...) || ...' and
3892 '#if !defined(...) && !defined(...) && ...'. This also avoids
3893 the growing number of special cases it was previously handling.
3894 [Richard Levitte]
3895
3896 *) Make all configuration macros available for application by making
3897 sure they are available in opensslconf.h, by giving them names starting
3898 with "OPENSSL_" to avoid conflicts with other packages and by making
3899 sure e_os2.h will cover all platform-specific cases together with
3900 opensslconf.h.
3901 Additionally, it is now possible to define configuration/platform-
3902 specific names (called "system identities"). In the C code, these
3903 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3904 macro with the name beginning with "OPENSSL_SYS_", which is determined
3905 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3906 what is available.
3907 [Richard Levitte]
3908
3909 *) New option -set_serial to 'req' and 'x509' this allows the serial
3910 number to use to be specified on the command line. Previously self
3911 signed certificates were hard coded with serial number 0 and the
3912 CA options of 'x509' had to use a serial number in a file which was
3913 auto incremented.
3914 [Steve Henson]
3915
3916 *) New options to 'ca' utility to support V2 CRL entry extensions.
3917 Currently CRL reason, invalidity date and hold instruction are
3918 supported. Add new CRL extensions to V3 code and some new objects.
3919 [Steve Henson]
3920
3921 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3922 disable standard block padding (aka PKCS#5 padding) in the EVP
3923 API, which was previously mandatory. This means that the data is
3924 not padded in any way and so the total length much be a multiple
3925 of the block size, otherwise an error occurs.
3926 [Steve Henson]
3927
3928 *) Initial (incomplete) OCSP SSL support.
3929 [Steve Henson]
3930
3931 *) New function OCSP_parse_url(). This splits up a URL into its host,
3932 port and path components: primarily to parse OCSP URLs. New -url
3933 option to ocsp utility.
3934 [Steve Henson]
3935
3936 *) New nonce behavior. The return value of OCSP_check_nonce() now
3937 reflects the various checks performed. Applications can decide
3938 whether to tolerate certain situations such as an absent nonce
3939 in a response when one was present in a request: the ocsp application
3940 just prints out a warning. New function OCSP_add1_basic_nonce()
3941 this is to allow responders to include a nonce in a response even if
3942 the request is nonce-less.
3943 [Steve Henson]
3944
3945 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3946 skipped when using openssl x509 multiple times on a single input file,
3947 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3948 [Bodo Moeller]
3949
3950 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3951 set string type: to handle setting ASN1_TIME structures. Fix ca
3952 utility to correctly initialize revocation date of CRLs.
3953 [Steve Henson]
3954
3955 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3956 the clients preferred ciphersuites and rather use its own preferences.
3957 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3958 Internet Explorer by ensuring unchanged hash method during stepup.
3959 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3960 [Lutz Jaenicke]
3961
3962 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3963 to aes and add a new 'exist' option to print out symbols that don't
3964 appear to exist.
3965 [Steve Henson]
3966
3967 *) Additional options to ocsp utility to allow flags to be set and
3968 additional certificates supplied.
3969 [Steve Henson]
3970
3971 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3972 OCSP client a number of certificate to only verify the response
3973 signature against.
3974 [Richard Levitte]
3975
3976 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3977 handle the new API. Currently only ECB, CBC modes supported. Add new
3978 AES OIDs.
3979
3980 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3981 Encryption Standard (AES) Ciphersuites for Transport Layer
3982 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3983 not enabled by default and were not part of the "ALL" ciphersuite
3984 alias because they were not yet official; they could be
3985 explicitly requested by specifying the "AESdraft" ciphersuite
3986 group alias. In the final release of OpenSSL 0.9.7, the group
3987 alias is called "AES" and is part of "ALL".)
3988 [Ben Laurie, Steve Henson, Bodo Moeller]
3989
3990 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3991 request to response.
3992 [Steve Henson]
3993
3994 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3995 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3996 extract information from a certificate request. OCSP_response_create()
3997 creates a response and optionally adds a basic response structure.
3998 OCSP_basic_add1_status() adds a complete single response to a basic
3999 response and returns the OCSP_SINGLERESP structure just added (to allow
4000 extensions to be included for example). OCSP_basic_add1_cert() adds a
4001 certificate to a basic response and OCSP_basic_sign() signs a basic
4002 response with various flags. New helper functions ASN1_TIME_check()
4003 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4004 (converts ASN1_TIME to GeneralizedTime).
4005 [Steve Henson]
4006
4007 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4008 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4009 structure from a certificate. X509_pubkey_digest() digests the public_key
4010 contents: this is used in various key identifiers.
4011 [Steve Henson]
4012
4013 *) Make sk_sort() tolerate a NULL argument.
4014 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4015
4016 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4017 passed by the function are trusted implicitly. If any of them signed the
4018 response then it is assumed to be valid and is not verified.
4019 [Steve Henson]
4020
4021 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4022 to data. This was previously part of the PKCS7 ASN1 code. This
4023 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4024 [Steve Henson, reported by Kenneth R. Robinette
4025 <support@securenetterm.com>]
4026
4027 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4028 routines: without these tracing memory leaks is very painful.
4029 Fix leaks in PKCS12 and PKCS7 routines.
4030 [Steve Henson]
4031
4032 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4033 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4034 effectively meant GeneralizedTime would never be used. Now it
4035 is initialised to -1 but X509_time_adj() now has to check the value
4036 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4037 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4038 [Steve Henson, reported by Kenneth R. Robinette
4039 <support@securenetterm.com>]
4040
4041 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4042 result in a zero length in the ASN1_INTEGER structure which was
4043 not consistent with the structure when d2i_ASN1_INTEGER() was used
4044 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4045 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4046 where it did not print out a minus for negative ASN1_INTEGER.
4047 [Steve Henson]
4048
4049 *) Add summary printout to ocsp utility. The various functions which
4050 convert status values to strings have been renamed to:
4051 OCSP_response_status_str(), OCSP_cert_status_str() and
4052 OCSP_crl_reason_str() and are no longer static. New options
4053 to verify nonce values and to disable verification. OCSP response
4054 printout format cleaned up.
4055 [Steve Henson]
4056
4057 *) Add additional OCSP certificate checks. These are those specified
4058 in RFC2560. This consists of two separate checks: the CA of the
4059 certificate being checked must either be the OCSP signer certificate
4060 or the issuer of the OCSP signer certificate. In the latter case the
4061 OCSP signer certificate must contain the OCSP signing extended key
4062 usage. This check is performed by attempting to match the OCSP
4063 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4064 in the OCSP_CERTID structures of the response.
4065 [Steve Henson]
4066
4067 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4068 and related routines. This uses the standard OpenSSL certificate
4069 verify routines to perform initial checks (just CA validity) and
4070 to obtain the certificate chain. Then additional checks will be
4071 performed on the chain. Currently the root CA is checked to see
4072 if it is explicitly trusted for OCSP signing. This is used to set
4073 a root CA as a global signing root: that is any certificate that
4074 chains to that CA is an acceptable OCSP signing certificate.
4075 [Steve Henson]
4076
4077 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4078 extensions from a separate configuration file.
4079 As when reading extensions from the main configuration file,
4080 the '-extensions ...' option may be used for specifying the
4081 section to use.
4082 [Massimiliano Pala <madwolf@comune.modena.it>]
4083
4084 *) New OCSP utility. Allows OCSP requests to be generated or
4085 read. The request can be sent to a responder and the output
4086 parsed, outputed or printed in text form. Not complete yet:
4087 still needs to check the OCSP response validity.
4088 [Steve Henson]
4089
4090 *) New subcommands for 'openssl ca':
4091 'openssl ca -status <serial>' prints the status of the cert with
4092 the given serial number (according to the index file).
4093 'openssl ca -updatedb' updates the expiry status of certificates
4094 in the index file.
4095 [Massimiliano Pala <madwolf@comune.modena.it>]
4096
4097 *) New '-newreq-nodes' command option to CA.pl. This is like
4098 '-newreq', but calls 'openssl req' with the '-nodes' option
4099 so that the resulting key is not encrypted.
4100 [Damien Miller <djm@mindrot.org>]
4101
4102 *) New configuration for the GNU Hurd.
4103 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4104
4105 *) Initial code to implement OCSP basic response verify. This
4106 is currently incomplete. Currently just finds the signer's
4107 certificate and verifies the signature on the response.
4108 [Steve Henson]
4109
4110 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4111 value of OPENSSLDIR. This is available via the new '-d' option
4112 to 'openssl version', and is also included in 'openssl version -a'.
4113 [Bodo Moeller]
4114
4115 *) Allowing defining memory allocation callbacks that will be given
4116 file name and line number information in additional arguments
4117 (a const char* and an int). The basic functionality remains, as
4118 well as the original possibility to just replace malloc(),
4119 realloc() and free() by functions that do not know about these
4120 additional arguments. To register and find out the current
4121 settings for extended allocation functions, the following
4122 functions are provided:
4123
4124 CRYPTO_set_mem_ex_functions
4125 CRYPTO_set_locked_mem_ex_functions
4126 CRYPTO_get_mem_ex_functions
4127 CRYPTO_get_locked_mem_ex_functions
4128
4129 These work the same way as CRYPTO_set_mem_functions and friends.
4130 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4131 extended allocation function is enabled.
4132 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4133 a conventional allocation function is enabled.
4134 [Richard Levitte, Bodo Moeller]
4135
4136 *) Finish off removing the remaining LHASH function pointer casts.
4137 There should no longer be any prototype-casting required when using
4138 the LHASH abstraction, and any casts that remain are "bugs". See
4139 the callback types and macros at the head of lhash.h for details
4140 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4141 [Geoff Thorpe]
4142
4143 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4144 If /dev/[u]random devices are not available or do not return enough
4145 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4146 be queried.
4147 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4148 /etc/entropy will be queried once each in this sequence, quering stops
4149 when enough entropy was collected without querying more sockets.
4150 [Lutz Jaenicke]
4151
4152 *) Change the Unix RAND_poll() variant to be able to poll several
4153 random devices, as specified by DEVRANDOM, until a sufficient amount
4154 of data has been collected. We spend at most 10 ms on each file
4155 (select timeout) and read in non-blocking mode. DEVRANDOM now
4156 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
4157 (previously it was just the string "/dev/urandom"), so on typical
4158 platforms the 10 ms delay will never occur.
4159 Also separate out the Unix variant to its own file, rand_unix.c.
4160 For VMS, there's a currently-empty rand_vms.c.
4161 [Richard Levitte]
4162
4163 *) Move OCSP client related routines to ocsp_cl.c. These
4164 provide utility functions which an application needing
4165 to issue a request to an OCSP responder and analyse the
4166 response will typically need: as opposed to those which an
4167 OCSP responder itself would need which will be added later.
4168
4169 OCSP_request_sign() signs an OCSP request with an API similar
4170 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
4171 response. OCSP_response_get1_basic() extracts basic response
4172 from response. OCSP_resp_find_status(): finds and extracts status
4173 information from an OCSP_CERTID structure (which will be created
4174 when the request structure is built). These are built from lower
4175 level functions which work on OCSP_SINGLERESP structures but
4176 wont normally be used unless the application wishes to examine
4177 extensions in the OCSP response for example.
4178
4179 Replace nonce routines with a pair of functions.
4180 OCSP_request_add1_nonce() adds a nonce value and optionally
4181 generates a random value. OCSP_check_nonce() checks the
4182 validity of the nonce in an OCSP response.
4183 [Steve Henson]
4184
4185 *) Change function OCSP_request_add() to OCSP_request_add0_id().
4186 This doesn't copy the supplied OCSP_CERTID and avoids the
4187 need to free up the newly created id. Change return type
4188 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
4189 This can then be used to add extensions to the request.
4190 Deleted OCSP_request_new(), since most of its functionality
4191 is now in OCSP_REQUEST_new() (and the case insensitive name
4192 clash) apart from the ability to set the request name which
4193 will be added elsewhere.
4194 [Steve Henson]
4195
4196 *) Update OCSP API. Remove obsolete extensions argument from
4197 various functions. Extensions are now handled using the new
4198 OCSP extension code. New simple OCSP HTTP function which
4199 can be used to send requests and parse the response.
4200 [Steve Henson]
4201
4202 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
4203 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
4204 uses the special reorder version of SET OF to sort the attributes
4205 and reorder them to match the encoded order. This resolves a long
4206 standing problem: a verify on a PKCS7 structure just after signing
4207 it used to fail because the attribute order did not match the
4208 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
4209 it uses the received order. This is necessary to tolerate some broken
4210 software that does not order SET OF. This is handled by encoding
4211 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
4212 to produce the required SET OF.
4213 [Steve Henson]
4214
4215 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
4216 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
4217 files to get correct declarations of the ASN.1 item variables.
4218 [Richard Levitte]
4219
4220 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
4221 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
4222 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
4223 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
4224 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
4225 ASN1_ITEM and no wrapper functions.
4226 [Steve Henson]
4227
4228 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
4229 replace the old function pointer based I/O routines. Change most of
4230 the *_d2i_bio() and *_d2i_fp() functions to use these.
4231 [Steve Henson]
4232
4233 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
4234 lines, recognice more "algorithms" that can be deselected, and make
4235 it complain about algorithm deselection that isn't recognised.
4236 [Richard Levitte]
4237
4238 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
4239 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
4240 to use new functions. Add NO_ASN1_OLD which can be set to remove
4241 some old style ASN1 functions: this can be used to determine if old
4242 code will still work when these eventually go away.
4243 [Steve Henson]
4244
4245 *) New extension functions for OCSP structures, these follow the
4246 same conventions as certificates and CRLs.
4247 [Steve Henson]
4248
4249 *) New function X509V3_add1_i2d(). This automatically encodes and
4250 adds an extension. Its behaviour can be customised with various
4251 flags to append, replace or delete. Various wrappers added for
4252 certifcates and CRLs.
4253 [Steve Henson]
4254
4255 *) Fix to avoid calling the underlying ASN1 print routine when
4256 an extension cannot be parsed. Correct a typo in the
4257 OCSP_SERVICELOC extension. Tidy up print OCSP format.
4258 [Steve Henson]
4259
4260 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
4261 entries for variables.
4262 [Steve Henson]
4263
4264 *) Add functionality to apps/openssl.c for detecting locking
4265 problems: As the program is single-threaded, all we have
4266 to do is register a locking callback using an array for
4267 storing which locks are currently held by the program.
4268 [Bodo Moeller]
4269
4270 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
4271 SSL_get_ex_data_X509_STORE_idx(), which is used in
4272 ssl_verify_cert_chain() and thus can be called at any time
4273 during TLS/SSL handshakes so that thread-safety is essential.
4274 Unfortunately, the ex_data design is not at all suited
4275 for multi-threaded use, so it probably should be abolished.
4276 [Bodo Moeller]
4277
4278 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
4279 [Broadcom, tweaked and integrated by Geoff Thorpe]
4280
4281 *) Move common extension printing code to new function
4282 X509V3_print_extensions(). Reorganise OCSP print routines and
4283 implement some needed OCSP ASN1 functions. Add OCSP extensions.
4284 [Steve Henson]
4285
4286 *) New function X509_signature_print() to remove duplication in some
4287 print routines.
4288 [Steve Henson]
4289
4290 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
4291 set (this was treated exactly the same as SET OF previously). This
4292 is used to reorder the STACK representing the structure to match the
4293 encoding. This will be used to get round a problem where a PKCS7
4294 structure which was signed could not be verified because the STACK
4295 order did not reflect the encoded order.
4296 [Steve Henson]
4297
4298 *) Reimplement the OCSP ASN1 module using the new code.
4299 [Steve Henson]
4300
4301 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
4302 for its ASN1 operations. The old style function pointers still exist
4303 for now but they will eventually go away.
4304 [Steve Henson]
4305
4306 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
4307 completely replaces the old ASN1 functionality with a table driven
4308 encoder and decoder which interprets an ASN1_ITEM structure describing
4309 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
4310 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
4311 has also been converted to the new form.
4312 [Steve Henson]
4313
4314 *) Change BN_mod_exp_recp so that negative moduli are tolerated
4315 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
4316 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
4317 for negative moduli.
4318 [Bodo Moeller]
4319
4320 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
4321 of not touching the result's sign bit.
4322 [Bodo Moeller]
4323
4324 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
4325 set.
4326 [Bodo Moeller]
4327
4328 *) Changed the LHASH code to use prototypes for callbacks, and created
4329 macros to declare and implement thin (optionally static) functions
4330 that provide type-safety and avoid function pointer casting for the
4331 type-specific callbacks.
4332 [Geoff Thorpe]
4333
4334 *) Added Kerberos Cipher Suites to be used with TLS, as written in
4335 RFC 2712.
4336 [Veers Staats <staatsvr@asc.hpc.mil>,
4337 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
4338
4339 *) Reformat the FAQ so the different questions and answers can be divided
4340 in sections depending on the subject.
4341 [Richard Levitte]
4342
4343 *) Have the zlib compression code load ZLIB.DLL dynamically under
4344 Windows.
4345 [Richard Levitte]
4346
4347 *) New function BN_mod_sqrt for computing square roots modulo a prime
4348 (using the probabilistic Tonelli-Shanks algorithm unless
4349 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
4350 be handled deterministically).
4351 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4352
4353 *) Make BN_mod_inverse faster by explicitly handling small quotients
4354 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
4355 512 bits], about 30% for larger ones [1024 or 2048 bits].)
4356 [Bodo Moeller]
4357
4358 *) New function BN_kronecker.
4359 [Bodo Moeller]
4360
4361 *) Fix BN_gcd so that it works on negative inputs; the result is
4362 positive unless both parameters are zero.
4363 Previously something reasonably close to an infinite loop was
4364 possible because numbers could be growing instead of shrinking
4365 in the implementation of Euclid's algorithm.
4366 [Bodo Moeller]
4367
4368 *) Fix BN_is_word() and BN_is_one() macros to take into account the
4369 sign of the number in question.
4370
4371 Fix BN_is_word(a,w) to work correctly for w == 0.
4372
4373 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
4374 because its test if the absolute value of 'a' equals 'w'.
4375 Note that BN_abs_is_word does *not* handle w == 0 reliably;
4376 it exists mostly for use in the implementations of BN_is_zero(),
4377 BN_is_one(), and BN_is_word().
4378 [Bodo Moeller]
4379
4380 *) New function BN_swap.
4381 [Bodo Moeller]
4382
4383 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
4384 the exponentiation functions are more likely to produce reasonable
4385 results on negative inputs.
4386 [Bodo Moeller]
4387
4388 *) Change BN_mod_mul so that the result is always non-negative.
4389 Previously, it could be negative if one of the factors was negative;
4390 I don't think anyone really wanted that behaviour.
4391 [Bodo Moeller]
4392
4393 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
4394 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
4395 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
4396 and add new functions:
4397
4398 BN_nnmod
4399 BN_mod_sqr
4400 BN_mod_add
4401 BN_mod_add_quick
4402 BN_mod_sub
4403 BN_mod_sub_quick
4404 BN_mod_lshift1
4405 BN_mod_lshift1_quick
4406 BN_mod_lshift
4407 BN_mod_lshift_quick
4408
4409 These functions always generate non-negative results.
4410
4411 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
4412 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
4413
4414 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
4415 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
4416 be reduced modulo m.
4417 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4418
4419 #if 0
4420 The following entry accidentily appeared in the CHANGES file
4421 distributed with OpenSSL 0.9.7. The modifications described in
4422 it do *not* apply to OpenSSL 0.9.7.
4423
4424 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4425 was actually never needed) and in BN_mul(). The removal in BN_mul()
4426 required a small change in bn_mul_part_recursive() and the addition
4427 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4428 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4429 bn_sub_words() and bn_add_words() except they take arrays with
4430 differing sizes.
4431 [Richard Levitte]
4432 #endif
4433
4434 *) In 'openssl passwd', verify passwords read from the terminal
4435 unless the '-salt' option is used (which usually means that
4436 verification would just waste user's time since the resulting
4437 hash is going to be compared with some given password hash)
4438 or the new '-noverify' option is used.
4439
4440 This is an incompatible change, but it does not affect
4441 non-interactive use of 'openssl passwd' (passwords on the command
4442 line, '-stdin' option, '-in ...' option) and thus should not
4443 cause any problems.
4444 [Bodo Moeller]
4445
4446 *) Remove all references to RSAref, since there's no more need for it.
4447 [Richard Levitte]
4448
4449 *) Make DSO load along a path given through an environment variable
4450 (SHLIB_PATH) with shl_load().
4451 [Richard Levitte]
4452
4453 *) Constify the ENGINE code as a result of BIGNUM constification.
4454 Also constify the RSA code and most things related to it. In a
4455 few places, most notable in the depth of the ASN.1 code, ugly
4456 casts back to non-const were required (to be solved at a later
4457 time)
4458 [Richard Levitte]
4459
4460 *) Make it so the openssl application has all engines loaded by default.
4461 [Richard Levitte]
4462
4463 *) Constify the BIGNUM routines a little more.
4464 [Richard Levitte]
4465
4466 *) Add the following functions:
4467
4468 ENGINE_load_cswift()
4469 ENGINE_load_chil()
4470 ENGINE_load_atalla()
4471 ENGINE_load_nuron()
4472 ENGINE_load_builtin_engines()
4473
4474 That way, an application can itself choose if external engines that
4475 are built-in in OpenSSL shall ever be used or not. The benefit is
4476 that applications won't have to be linked with libdl or other dso
4477 libraries unless it's really needed.
4478
4479 Changed 'openssl engine' to load all engines on demand.
4480 Changed the engine header files to avoid the duplication of some
4481 declarations (they differed!).
4482 [Richard Levitte]
4483
4484 *) 'openssl engine' can now list capabilities.
4485 [Richard Levitte]
4486
4487 *) Better error reporting in 'openssl engine'.
4488 [Richard Levitte]
4489
4490 *) Never call load_dh_param(NULL) in s_server.
4491 [Bodo Moeller]
4492
4493 *) Add engine application. It can currently list engines by name and
4494 identity, and test if they are actually available.
4495 [Richard Levitte]
4496
4497 *) Improve RPM specification file by forcing symbolic linking and making
4498 sure the installed documentation is also owned by root.root.
4499 [Damien Miller <djm@mindrot.org>]
4500
4501 *) Give the OpenSSL applications more possibilities to make use of
4502 keys (public as well as private) handled by engines.
4503 [Richard Levitte]
4504
4505 *) Add OCSP code that comes from CertCo.
4506 [Richard Levitte]
4507
4508 *) Add VMS support for the Rijndael code.
4509 [Richard Levitte]
4510
4511 *) Added untested support for Nuron crypto accelerator.
4512 [Ben Laurie]
4513
4514 *) Add support for external cryptographic devices. This code was
4515 previously distributed separately as the "engine" branch.
4516 [Geoff Thorpe, Richard Levitte]
4517
4518 *) Rework the filename-translation in the DSO code. It is now possible to
4519 have far greater control over how a "name" is turned into a filename
4520 depending on the operating environment and any oddities about the
4521 different shared library filenames on each system.
4522 [Geoff Thorpe]
4523
4524 *) Support threads on FreeBSD-elf in Configure.
4525 [Richard Levitte]
4526
4527 *) Fix for SHA1 assembly problem with MASM: it produces
4528 warnings about corrupt line number information when assembling
4529 with debugging information. This is caused by the overlapping
4530 of two sections.
4531 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
4532
4533 *) NCONF changes.
4534 NCONF_get_number() has no error checking at all. As a replacement,
4535 NCONF_get_number_e() is defined (_e for "error checking") and is
4536 promoted strongly. The old NCONF_get_number is kept around for
4537 binary backward compatibility.
4538 Make it possible for methods to load from something other than a BIO,
4539 by providing a function pointer that is given a name instead of a BIO.
4540 For example, this could be used to load configuration data from an
4541 LDAP server.
4542 [Richard Levitte]
4543
4544 *) Fix for non blocking accept BIOs. Added new I/O special reason
4545 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
4546 with non blocking I/O was not possible because no retry code was
4547 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
4548 this case.
4549 [Steve Henson]
4550
4551 *) Added the beginnings of Rijndael support.
4552 [Ben Laurie]
4553
4554 *) Fix for bug in DirectoryString mask setting. Add support for
4555 X509_NAME_print_ex() in 'req' and X509_print_ex() function
4556 to allow certificate printing to more controllable, additional
4557 'certopt' option to 'x509' to allow new printing options to be
4558 set.
4559 [Steve Henson]
4560
4561 *) Clean old EAY MD5 hack from e_os.h.
4562 [Richard Levitte]
4563
4564 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
4565
4566 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4567 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4568 [Joe Orton, Steve Henson]
4569
4570 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
4571
4572 *) Fix additional bug revealed by the NISCC test suite:
4573
4574 Stop bug triggering large recursion when presented with
4575 certain ASN.1 tags (CVE-2003-0851)
4576 [Steve Henson]
4577
4578 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
4579
4580 *) Fix various bugs revealed by running the NISCC test suite:
4581
4582 Stop out of bounds reads in the ASN1 code when presented with
4583 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4584
4585 If verify callback ignores invalid public key errors don't try to check
4586 certificate signature with the NULL public key.
4587
4588 [Steve Henson]
4589
4590 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4591 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4592 specifications.
4593 [Steve Henson]
4594
4595 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4596 extra data after the compression methods not only for TLS 1.0
4597 but also for SSL 3.0 (as required by the specification).
4598 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4599
4600 *) Change X509_certificate_type() to mark the key as exported/exportable
4601 when it's 512 *bits* long, not 512 bytes.
4602 [Richard Levitte]
4603
4604 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
4605
4606 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4607 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4608 a protocol version number mismatch like a decryption error
4609 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4610 [Bodo Moeller]
4611
4612 *) Turn on RSA blinding by default in the default implementation
4613 to avoid a timing attack. Applications that don't want it can call
4614 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4615 They would be ill-advised to do so in most cases.
4616 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4617
4618 *) Change RSA blinding code so that it works when the PRNG is not
4619 seeded (in this case, the secret RSA exponent is abused as
4620 an unpredictable seed -- if it is not unpredictable, there
4621 is no point in blinding anyway). Make RSA blinding thread-safe
4622 by remembering the creator's thread ID in rsa->blinding and
4623 having all other threads use local one-time blinding factors
4624 (this requires more computation than sharing rsa->blinding, but
4625 avoids excessive locking; and if an RSA object is not shared
4626 between threads, blinding will still be very fast).
4627 [Bodo Moeller]
4628
4629 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4630
4631 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4632 via timing by performing a MAC computation even if incorrrect
4633 block cipher padding has been found. This is a countermeasure
4634 against active attacks where the attacker has to distinguish
4635 between bad padding and a MAC verification error. (CVE-2003-0078)
4636
4637 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4638 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4639 Martin Vuagnoux (EPFL, Ilion)]
4640
4641 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4642
4643 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4644 memory from it's contents. This is done with a counter that will
4645 place alternating values in each byte. This can be used to solve
4646 two issues: 1) the removal of calls to memset() by highly optimizing
4647 compilers, and 2) cleansing with other values than 0, since those can
4648 be read through on certain media, for example a swap space on disk.
4649 [Geoff Thorpe]
4650
4651 *) Bugfix: client side session caching did not work with external caching,
4652 because the session->cipher setting was not restored when reloading
4653 from the external cache. This problem was masked, when
4654 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4655 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4656 [Lutz Jaenicke]
4657
4658 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4659 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4660 [Zeev Lieber <zeev-l@yahoo.com>]
4661
4662 *) Undo an undocumented change introduced in 0.9.6e which caused
4663 repeated calls to OpenSSL_add_all_ciphers() and
4664 OpenSSL_add_all_digests() to be ignored, even after calling
4665 EVP_cleanup().
4666 [Richard Levitte]
4667
4668 *) Change the default configuration reader to deal with last line not
4669 being properly terminated.
4670 [Richard Levitte]
4671
4672 *) Change X509_NAME_cmp() so it applies the special rules on handling
4673 DN values that are of type PrintableString, as well as RDNs of type
4674 emailAddress where the value has the type ia5String.
4675 [stefank@valicert.com via Richard Levitte]
4676
4677 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4678 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4679 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4680 the bitwise-OR of the two for use by the majority of applications
4681 wanting this behaviour, and update the docs. The documented
4682 behaviour and actual behaviour were inconsistent and had been
4683 changing anyway, so this is more a bug-fix than a behavioural
4684 change.
4685 [Geoff Thorpe, diagnosed by Nadav Har'El]
4686
4687 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4688 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4689 [Bodo Moeller]
4690
4691 *) Fix initialization code race conditions in
4692 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4693 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4694 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4695 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4696 ssl2_get_cipher_by_char(),
4697 ssl3_get_cipher_by_char().
4698 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4699
4700 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4701 the cached sessions are flushed, as the remove_cb() might use ex_data
4702 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4703 (see [openssl.org #212]).
4704 [Geoff Thorpe, Lutz Jaenicke]
4705
4706 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4707 length, instead of the encoding length to d2i_ASN1_OBJECT.
4708 [Steve Henson]
4709
4710 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4711
4712 *) [In 0.9.6g-engine release:]
4713 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4714 [Lynn Gazis <lgazis@rainbow.com>]
4715
4716 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4717
4718 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4719 and get fix the header length calculation.
4720 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4721 Alon Kantor <alonk@checkpoint.com> (and others),
4722 Steve Henson]
4723
4724 *) Use proper error handling instead of 'assertions' in buffer
4725 overflow checks added in 0.9.6e. This prevents DoS (the
4726 assertions could call abort()).
4727 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4728
4729 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4730
4731 *) Add various sanity checks to asn1_get_length() to reject
4732 the ASN1 length bytes if they exceed sizeof(long), will appear
4733 negative or the content length exceeds the length of the
4734 supplied buffer.
4735 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4736
4737 *) Fix cipher selection routines: ciphers without encryption had no flags
4738 for the cipher strength set and where therefore not handled correctly
4739 by the selection routines (PR #130).
4740 [Lutz Jaenicke]
4741
4742 *) Fix EVP_dsa_sha macro.
4743 [Nils Larsch]
4744
4745 *) New option
4746 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4747 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4748 that was added in OpenSSL 0.9.6d.
4749
4750 As the countermeasure turned out to be incompatible with some
4751 broken SSL implementations, the new option is part of SSL_OP_ALL.
4752 SSL_OP_ALL is usually employed when compatibility with weird SSL
4753 implementations is desired (e.g. '-bugs' option to 's_client' and
4754 's_server'), so the new option is automatically set in many
4755 applications.
4756 [Bodo Moeller]
4757
4758 *) Changes in security patch:
4759
4760 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4761 Research Projects Agency (DARPA) and Air Force Research Laboratory,
4762 Air Force Materiel Command, USAF, under agreement number
4763 F30602-01-2-0537.
4764
4765 *) Add various sanity checks to asn1_get_length() to reject
4766 the ASN1 length bytes if they exceed sizeof(long), will appear
4767 negative or the content length exceeds the length of the
4768 supplied buffer. (CVE-2002-0659)
4769 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4770
4771 *) Assertions for various potential buffer overflows, not known to
4772 happen in practice.
4773 [Ben Laurie (CHATS)]
4774
4775 *) Various temporary buffers to hold ASCII versions of integers were
4776 too small for 64 bit platforms. (CVE-2002-0655)
4777 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
4778
4779 *) Remote buffer overflow in SSL3 protocol - an attacker could
4780 supply an oversized session ID to a client. (CVE-2002-0656)
4781 [Ben Laurie (CHATS)]
4782
4783 *) Remote buffer overflow in SSL2 protocol - an attacker could
4784 supply an oversized client master key. (CVE-2002-0656)
4785 [Ben Laurie (CHATS)]
4786
4787 Changes between 0.9.6c and 0.9.6d [9 May 2002]
4788
4789 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
4790 encoded as NULL) with id-dsa-with-sha1.
4791 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4792
4793 *) Check various X509_...() return values in apps/req.c.
4794 [Nils Larsch <nla@trustcenter.de>]
4795
4796 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4797 an end-of-file condition would erronously be flagged, when the CRLF
4798 was just at the end of a processed block. The bug was discovered when
4799 processing data through a buffering memory BIO handing the data to a
4800 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4801 <ptsekov@syntrex.com> and Nedelcho Stanev.
4802 [Lutz Jaenicke]
4803
4804 *) Implement a countermeasure against a vulnerability recently found
4805 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4806 before application data chunks to avoid the use of known IVs
4807 with data potentially chosen by the attacker.
4808 [Bodo Moeller]
4809
4810 *) Fix length checks in ssl3_get_client_hello().
4811 [Bodo Moeller]
4812
4813 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4814 to prevent ssl3_read_internal() from incorrectly assuming that
4815 ssl3_read_bytes() found application data while handshake
4816 processing was enabled when in fact s->s3->in_read_app_data was
4817 merely automatically cleared during the initial handshake.
4818 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4819
4820 *) Fix object definitions for Private and Enterprise: they were not
4821 recognized in their shortname (=lowercase) representation. Extend
4822 obj_dat.pl to issue an error when using undefined keywords instead
4823 of silently ignoring the problem (Svenning Sorensen
4824 <sss@sss.dnsalias.net>).
4825 [Lutz Jaenicke]
4826
4827 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4828 generators, i.e. generators other than 2 and 5. (Previously, the
4829 code did not properly initialise the 'add' and 'rem' values to
4830 BN_generate_prime().)
4831
4832 In the new general case, we do not insist that 'generator' is
4833 actually a primitive root: This requirement is rather pointless;
4834 a generator of the order-q subgroup is just as good, if not
4835 better.
4836 [Bodo Moeller]
4837
4838 *) Map new X509 verification errors to alerts. Discovered and submitted by
4839 Tom Wu <tom@arcot.com>.
4840 [Lutz Jaenicke]
4841
4842 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4843 returning non-zero before the data has been completely received
4844 when using non-blocking I/O.
4845 [Bodo Moeller; problem pointed out by John Hughes]
4846
4847 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4848 [Ben Laurie, Lutz Jaenicke]
4849
4850 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4851 Yoram Zahavi <YoramZ@gilian.com>).
4852 [Lutz Jaenicke]
4853
4854 *) Add information about CygWin 1.3 and on, and preserve proper
4855 configuration for the versions before that.
4856 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4857
4858 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4859 check whether we deal with a copy of a session and do not delete from
4860 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4861 <izhar@checkpoint.com>.
4862 [Lutz Jaenicke]
4863
4864 *) Do not store session data into the internal session cache, if it
4865 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4866 flag is set). Proposed by Aslam <aslam@funk.com>.
4867 [Lutz Jaenicke]
4868
4869 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4870 value is 0.
4871 [Richard Levitte]
4872
4873 *) [In 0.9.6d-engine release:]
4874 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4875 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4876
4877 *) Add the configuration target linux-s390x.
4878 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4879
4880 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4881 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4882 variable as an indication that a ClientHello message has been
4883 received. As the flag value will be lost between multiple
4884 invocations of ssl3_accept when using non-blocking I/O, the
4885 function may not be aware that a handshake has actually taken
4886 place, thus preventing a new session from being added to the
4887 session cache.
4888
4889 To avoid this problem, we now set s->new_session to 2 instead of
4890 using a local variable.
4891 [Lutz Jaenicke, Bodo Moeller]
4892
4893 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4894 if the SSL_R_LENGTH_MISMATCH error is detected.
4895 [Geoff Thorpe, Bodo Moeller]
4896
4897 *) New 'shared_ldflag' column in Configure platform table.
4898 [Richard Levitte]
4899
4900 *) Fix EVP_CIPHER_mode macro.
4901 ["Dan S. Camper" <dan@bti.net>]
4902
4903 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4904 type, we must throw them away by setting rr->length to 0.
4905 [D P Chang <dpc@qualys.com>]
4906
4907 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4908
4909 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4910 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4911 worked incorrectly for those cases where range = 10..._2 and
4912 3*range is two bits longer than range.)
4913 [Bodo Moeller]
4914
4915 *) Only add signing time to PKCS7 structures if it is not already
4916 present.
4917 [Steve Henson]
4918
4919 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4920 OBJ_ld_ce should be OBJ_id_ce.
4921 Also some ip-pda OIDs in crypto/objects/objects.txt were
4922 incorrect (cf. RFC 3039).
4923 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4924
4925 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4926 returns early because it has nothing to do.
4927 [Andy Schneider <andy.schneider@bjss.co.uk>]
4928
4929 *) [In 0.9.6c-engine release:]
4930 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4931 [Andy Schneider <andy.schneider@bjss.co.uk>]
4932
4933 *) [In 0.9.6c-engine release:]
4934 Add support for Cryptographic Appliance's keyserver technology.
4935 (Use engine 'keyclient')
4936 [Cryptographic Appliances and Geoff Thorpe]
4937
4938 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4939 is called via tools/c89.sh because arguments have to be
4940 rearranged (all '-L' options must appear before the first object
4941 modules).
4942 [Richard Shapiro <rshapiro@abinitio.com>]
4943
4944 *) [In 0.9.6c-engine release:]
4945 Add support for Broadcom crypto accelerator cards, backported
4946 from 0.9.7.
4947 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4948
4949 *) [In 0.9.6c-engine release:]
4950 Add support for SureWare crypto accelerator cards from
4951 Baltimore Technologies. (Use engine 'sureware')
4952 [Baltimore Technologies and Mark Cox]
4953
4954 *) [In 0.9.6c-engine release:]
4955 Add support for crypto accelerator cards from Accelerated
4956 Encryption Processing, www.aep.ie. (Use engine 'aep')
4957 [AEP Inc. and Mark Cox]
4958
4959 *) Add a configuration entry for gcc on UnixWare.
4960 [Gary Benson <gbenson@redhat.com>]
4961
4962 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4963 messages are stored in a single piece (fixed-length part and
4964 variable-length part combined) and fix various bugs found on the way.
4965 [Bodo Moeller]
4966
4967 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4968 instead. BIO_gethostbyname() does not know what timeouts are
4969 appropriate, so entries would stay in cache even when they have
4970 become invalid.
4971 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4972
4973 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4974 faced with a pathologically small ClientHello fragment that does
4975 not contain client_version: Instead of aborting with an error,
4976 simply choose the highest available protocol version (i.e.,
4977 TLS 1.0 unless it is disabled). In practice, ClientHello
4978 messages are never sent like this, but this change gives us
4979 strictly correct behaviour at least for TLS.
4980 [Bodo Moeller]
4981
4982 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4983 never resets s->method to s->ctx->method when called from within
4984 one of the SSL handshake functions.
4985 [Bodo Moeller; problem pointed out by Niko Baric]
4986
4987 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4988 (sent using the client's version number) if client_version is
4989 smaller than the protocol version in use. Also change
4990 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4991 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4992 the client will at least see that alert.
4993 [Bodo Moeller]
4994
4995 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4996 correctly.
4997 [Bodo Moeller]
4998
4999 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5000 client receives HelloRequest while in a handshake.
5001 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5002
5003 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5004 should end in 'break', not 'goto end' which circuments various
5005 cleanups done in state SSL_ST_OK. But session related stuff
5006 must be disabled for SSL_ST_OK in the case that we just sent a
5007 HelloRequest.
5008
5009 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5010 before just sending a HelloRequest.
5011 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5012
5013 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5014 reveal whether illegal block cipher padding was found or a MAC
5015 verification error occured. (Neither SSLerr() codes nor alerts
5016 are directly visible to potential attackers, but the information
5017 may leak via logfiles.)
5018
5019 Similar changes are not required for the SSL 2.0 implementation
5020 because the number of padding bytes is sent in clear for SSL 2.0,
5021 and the extra bytes are just ignored. However ssl/s2_pkt.c
5022 failed to verify that the purported number of padding bytes is in
5023 the legal range.
5024 [Bodo Moeller]
5025
5026 *) Add OpenUNIX-8 support including shared libraries
5027 (Boyd Lynn Gerber <gerberb@zenez.com>).
5028 [Lutz Jaenicke]
5029
5030 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5031 'wristwatch attack' using huge encoding parameters (cf.
5032 James H. Manger's CRYPTO 2001 paper). Note that the
5033 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5034 encoding parameters and hence was not vulnerable.
5035 [Bodo Moeller]
5036
5037 *) BN_sqr() bug fix.
5038 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5039
5040 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5041 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5042 followed by modular reduction.
5043 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5044
5045 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5046 equivalent based on BN_pseudo_rand() instead of BN_rand().
5047 [Bodo Moeller]
5048
5049 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5050 This function was broken, as the check for a new client hello message
5051 to handle SGC did not allow these large messages.
5052 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5053 [Lutz Jaenicke]
5054
5055 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5056 [Lutz Jaenicke]
5057
5058 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5059 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5060 [Lutz Jaenicke]
5061
5062 *) Rework the configuration and shared library support for Tru64 Unix.
5063 The configuration part makes use of modern compiler features and
5064 still retains old compiler behavior for those that run older versions
5065 of the OS. The shared library support part includes a variant that
5066 uses the RPATH feature, and is available through the special
5067 configuration target "alpha-cc-rpath", which will never be selected
5068 automatically.
5069 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5070
5071 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5072 with the same message size as in ssl3_get_certificate_request().
5073 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5074 messages might inadvertently be reject as too long.
5075 [Petr Lampa <lampa@fee.vutbr.cz>]
5076
5077 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5078 [Andy Polyakov]
5079
5080 *) Modified SSL library such that the verify_callback that has been set
5081 specificly for an SSL object with SSL_set_verify() is actually being
5082 used. Before the change, a verify_callback set with this function was
5083 ignored and the verify_callback() set in the SSL_CTX at the time of
5084 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5085 to allow the necessary settings.
5086 [Lutz Jaenicke]
5087
5088 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5089 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5090 done automatically (in contradiction to the requirements of the C
5091 standard). This made problems when used from OpenSSH.
5092 [Lutz Jaenicke]
5093
5094 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5095 dh->length and always used
5096
5097 BN_rand_range(priv_key, dh->p).
5098
5099 BN_rand_range() is not necessary for Diffie-Hellman, and this
5100 specific range makes Diffie-Hellman unnecessarily inefficient if
5101 dh->length (recommended exponent length) is much smaller than the
5102 length of dh->p. We could use BN_rand_range() if the order of
5103 the subgroup was stored in the DH structure, but we only have
5104 dh->length.
5105
5106 So switch back to
5107
5108 BN_rand(priv_key, l, ...)
5109
5110 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5111 otherwise.
5112 [Bodo Moeller]
5113
5114 *) In
5115
5116 RSA_eay_public_encrypt
5117 RSA_eay_private_decrypt
5118 RSA_eay_private_encrypt (signing)
5119 RSA_eay_public_decrypt (signature verification)
5120
5121 (default implementations for RSA_public_encrypt,
5122 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5123 always reject numbers >= n.
5124 [Bodo Moeller]
5125
5126 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5127 to synchronize access to 'locking_thread'. This is necessary on
5128 systems where access to 'locking_thread' (an 'unsigned long'
5129 variable) is not atomic.
5130 [Bodo Moeller]
5131
5132 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5133 *before* setting the 'crypto_lock_rand' flag. The previous code had
5134 a race condition if 0 is a valid thread ID.
5135 [Travis Vitek <vitek@roguewave.com>]
5136
5137 *) Add support for shared libraries under Irix.
5138 [Albert Chin-A-Young <china@thewrittenword.com>]
5139
5140 *) Add configuration option to build on Linux on both big-endian and
5141 little-endian MIPS.
5142 [Ralf Baechle <ralf@uni-koblenz.de>]
5143
5144 *) Add the possibility to create shared libraries on HP-UX.
5145 [Richard Levitte]
5146
5147 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5148
5149 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
5150 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
5151 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
5152 PRNG state recovery was possible based on the output of
5153 one PRNG request appropriately sized to gain knowledge on
5154 'md' followed by enough consecutive 1-byte PRNG requests
5155 to traverse all of 'state'.
5156
5157 1. When updating 'md_local' (the current thread's copy of 'md')
5158 during PRNG output generation, hash all of the previous
5159 'md_local' value, not just the half used for PRNG output.
5160
5161 2. Make the number of bytes from 'state' included into the hash
5162 independent from the number of PRNG bytes requested.
5163
5164 The first measure alone would be sufficient to avoid
5165 Markku-Juhani's attack. (Actually it had never occurred
5166 to me that the half of 'md_local' used for chaining was the
5167 half from which PRNG output bytes were taken -- I had always
5168 assumed that the secret half would be used.) The second
5169 measure makes sure that additional data from 'state' is never
5170 mixed into 'md_local' in small portions; this heuristically
5171 further strengthens the PRNG.
5172 [Bodo Moeller]
5173
5174 *) Fix crypto/bn/asm/mips3.s.
5175 [Andy Polyakov]
5176
5177 *) When only the key is given to "enc", the IV is undefined. Print out
5178 an error message in this case.
5179 [Lutz Jaenicke]
5180
5181 *) Handle special case when X509_NAME is empty in X509 printing routines.
5182 [Steve Henson]
5183
5184 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
5185 positive and less than q.
5186 [Bodo Moeller]
5187
5188 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
5189 used: it isn't thread safe and the add_lock_callback should handle
5190 that itself.
5191 [Paul Rose <Paul.Rose@bridge.com>]
5192
5193 *) Verify that incoming data obeys the block size in
5194 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
5195 [Bodo Moeller]
5196
5197 *) Fix OAEP check.
5198 [Ulf Möller, Bodo Möller]
5199
5200 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
5201 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
5202 when fixing the server behaviour for backwards-compatible 'client
5203 hello' messages. (Note that the attack is impractical against
5204 SSL 3.0 and TLS 1.0 anyway because length and version checking
5205 means that the probability of guessing a valid ciphertext is
5206 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
5207 paper.)
5208
5209 Before 0.9.5, the countermeasure (hide the error by generating a
5210 random 'decryption result') did not work properly because
5211 ERR_clear_error() was missing, meaning that SSL_get_error() would
5212 detect the supposedly ignored error.
5213
5214 Both problems are now fixed.
5215 [Bodo Moeller]
5216
5217 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
5218 (previously it was 1024).
5219 [Bodo Moeller]
5220
5221 *) Fix for compatibility mode trust settings: ignore trust settings
5222 unless some valid trust or reject settings are present.
5223 [Steve Henson]
5224
5225 *) Fix for blowfish EVP: its a variable length cipher.
5226 [Steve Henson]
5227
5228 *) Fix various bugs related to DSA S/MIME verification. Handle missing
5229 parameters in DSA public key structures and return an error in the
5230 DSA routines if parameters are absent.
5231 [Steve Henson]
5232
5233 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
5234 in the current directory if neither $RANDFILE nor $HOME was set.
5235 RAND_file_name() in 0.9.6a returned NULL in this case. This has
5236 caused some confusion to Windows users who haven't defined $HOME.
5237 Thus RAND_file_name() is changed again: e_os.h can define a
5238 DEFAULT_HOME, which will be used if $HOME is not set.
5239 For Windows, we use "C:"; on other platforms, we still require
5240 environment variables.
5241
5242 *) Move 'if (!initialized) RAND_poll()' into regions protected by
5243 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
5244 having multiple threads call RAND_poll() concurrently.
5245 [Bodo Moeller]
5246
5247 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
5248 combination of a flag and a thread ID variable.
5249 Otherwise while one thread is in ssleay_rand_bytes (which sets the
5250 flag), *other* threads can enter ssleay_add_bytes without obeying
5251 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
5252 that they do not hold after the first thread unsets add_do_not_lock).
5253 [Bodo Moeller]
5254
5255 *) Change bctest again: '-x' expressions are not available in all
5256 versions of 'test'.
5257 [Bodo Moeller]
5258
5259 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
5260
5261 *) Fix a couple of memory leaks in PKCS7_dataDecode()
5262 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
5263
5264 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
5265 the default extension for executables, if any. Also, make the perl
5266 scripts that use symlink() to test if it really exists and use "cp"
5267 if it doesn't. All this made OpenSSL compilable and installable in
5268 CygWin.
5269 [Richard Levitte]
5270
5271 *) Fix for asn1_GetSequence() for indefinite length constructed data.
5272 If SEQUENCE is length is indefinite just set c->slen to the total
5273 amount of data available.
5274 [Steve Henson, reported by shige@FreeBSD.org]
5275 [This change does not apply to 0.9.7.]
5276
5277 *) Change bctest to avoid here-documents inside command substitution
5278 (workaround for FreeBSD /bin/sh bug).
5279 For compatibility with Ultrix, avoid shell functions (introduced
5280 in the bctest version that searches along $PATH).
5281 [Bodo Moeller]
5282
5283 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
5284 with des_encrypt() defined on some operating systems, like Solaris
5285 and UnixWare.
5286 [Richard Levitte]
5287
5288 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
5289 On the Importance of Eliminating Errors in Cryptographic
5290 Computations, J. Cryptology 14 (2001) 2, 101-119,
5291 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
5292 [Ulf Moeller]
5293
5294 *) MIPS assembler BIGNUM division bug fix.
5295 [Andy Polyakov]
5296
5297 *) Disabled incorrect Alpha assembler code.
5298 [Richard Levitte]
5299
5300 *) Fix PKCS#7 decode routines so they correctly update the length
5301 after reading an EOC for the EXPLICIT tag.
5302 [Steve Henson]
5303 [This change does not apply to 0.9.7.]
5304
5305 *) Fix bug in PKCS#12 key generation routines. This was triggered
5306 if a 3DES key was generated with a 0 initial byte. Include
5307 PKCS12_BROKEN_KEYGEN compilation option to retain the old
5308 (but broken) behaviour.
5309 [Steve Henson]
5310
5311 *) Enhance bctest to search for a working bc along $PATH and print
5312 it when found.
5313 [Tim Rice <tim@multitalents.net> via Richard Levitte]
5314
5315 *) Fix memory leaks in err.c: free err_data string if necessary;
5316 don't write to the wrong index in ERR_set_error_data.
5317 [Bodo Moeller]
5318
5319 *) Implement ssl23_peek (analogous to ssl23_read), which previously
5320 did not exist.
5321 [Bodo Moeller]
5322
5323 *) Replace rdtsc with _emit statements for VC++ version 5.
5324 [Jeremy Cooper <jeremy@baymoo.org>]
5325
5326 *) Make it possible to reuse SSLv2 sessions.
5327 [Richard Levitte]
5328
5329 *) In copy_email() check for >= 0 as a return value for
5330 X509_NAME_get_index_by_NID() since 0 is a valid index.
5331 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
5332
5333 *) Avoid coredump with unsupported or invalid public keys by checking if
5334 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
5335 PKCS7_verify() fails with non detached data.
5336 [Steve Henson]
5337
5338 *) Don't use getenv in library functions when run as setuid/setgid.
5339 New function OPENSSL_issetugid().
5340 [Ulf Moeller]
5341
5342 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
5343 due to incorrect handling of multi-threading:
5344
5345 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
5346
5347 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
5348
5349 3. Count how many times MemCheck_off() has been called so that
5350 nested use can be treated correctly. This also avoids
5351 inband-signalling in the previous code (which relied on the
5352 assumption that thread ID 0 is impossible).
5353 [Bodo Moeller]
5354
5355 *) Add "-rand" option also to s_client and s_server.
5356 [Lutz Jaenicke]
5357
5358 *) Fix CPU detection on Irix 6.x.
5359 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
5360 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5361
5362 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
5363 was empty.
5364 [Steve Henson]
5365 [This change does not apply to 0.9.7.]
5366
5367 *) Use the cached encoding of an X509_NAME structure rather than
5368 copying it. This is apparently the reason for the libsafe "errors"
5369 but the code is actually correct.
5370 [Steve Henson]
5371
5372 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
5373 Bleichenbacher's DSA attack.
5374 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
5375 to be set and top=0 forces the highest bit to be set; top=-1 is new
5376 and leaves the highest bit random.
5377 [Ulf Moeller, Bodo Moeller]
5378
5379 *) In the NCONF_...-based implementations for CONF_... queries
5380 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
5381 a temporary CONF structure with the data component set to NULL
5382 (which gives segmentation faults in lh_retrieve).
5383 Instead, use NULL for the CONF pointer in CONF_get_string and
5384 CONF_get_number (which may use environment variables) and directly
5385 return NULL from CONF_get_section.
5386 [Bodo Moeller]
5387
5388 *) Fix potential buffer overrun for EBCDIC.
5389 [Ulf Moeller]
5390
5391 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
5392 keyUsage if basicConstraints absent for a CA.
5393 [Steve Henson]
5394
5395 *) Make SMIME_write_PKCS7() write mail header values with a format that
5396 is more generally accepted (no spaces before the semicolon), since
5397 some programs can't parse those values properly otherwise. Also make
5398 sure BIO's that break lines after each write do not create invalid
5399 headers.
5400 [Richard Levitte]
5401
5402 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
5403 macros previously used would not encode an empty SEQUENCE OF
5404 and break the signature.
5405 [Steve Henson]
5406 [This change does not apply to 0.9.7.]
5407
5408 *) Zero the premaster secret after deriving the master secret in
5409 DH ciphersuites.
5410 [Steve Henson]
5411
5412 *) Add some EVP_add_digest_alias registrations (as found in
5413 OpenSSL_add_all_digests()) to SSL_library_init()
5414 aka OpenSSL_add_ssl_algorithms(). This provides improved
5415 compatibility with peers using X.509 certificates
5416 with unconventional AlgorithmIdentifier OIDs.
5417 [Bodo Moeller]
5418
5419 *) Fix for Irix with NO_ASM.
5420 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5421
5422 *) ./config script fixes.
5423 [Ulf Moeller, Richard Levitte]
5424
5425 *) Fix 'openssl passwd -1'.
5426 [Bodo Moeller]
5427
5428 *) Change PKCS12_key_gen_asc() so it can cope with non null
5429 terminated strings whose length is passed in the passlen
5430 parameter, for example from PEM callbacks. This was done
5431 by adding an extra length parameter to asc2uni().
5432 [Steve Henson, reported by <oddissey@samsung.co.kr>]
5433
5434 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
5435 call failed, free the DSA structure.
5436 [Bodo Moeller]
5437
5438 *) Fix to uni2asc() to cope with zero length Unicode strings.
5439 These are present in some PKCS#12 files.
5440 [Steve Henson]
5441
5442 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
5443 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
5444 when writing a 32767 byte record.
5445 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
5446
5447 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
5448 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
5449
5450 (RSA objects have a reference count access to which is protected
5451 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
5452 so they are meant to be shared between threads.)
5453 [Bodo Moeller, Geoff Thorpe; original patch submitted by
5454 "Reddie, Steven" <Steven.Reddie@ca.com>]
5455
5456 *) Fix a deadlock in CRYPTO_mem_leaks().
5457 [Bodo Moeller]
5458
5459 *) Use better test patterns in bntest.
5460 [Ulf Möller]
5461
5462 *) rand_win.c fix for Borland C.
5463 [Ulf Möller]
5464
5465 *) BN_rshift bugfix for n == 0.
5466 [Bodo Moeller]
5467
5468 *) Add a 'bctest' script that checks for some known 'bc' bugs
5469 so that 'make test' does not abort just because 'bc' is broken.
5470 [Bodo Moeller]
5471
5472 *) Store verify_result within SSL_SESSION also for client side to
5473 avoid potential security hole. (Re-used sessions on the client side
5474 always resulted in verify_result==X509_V_OK, not using the original
5475 result of the server certificate verification.)
5476 [Lutz Jaenicke]
5477
5478 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
5479 SSL3_RT_APPLICATION_DATA, return 0.
5480 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
5481 [Bodo Moeller]
5482
5483 *) Fix SSL_peek:
5484 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
5485 releases, have been re-implemented by renaming the previous
5486 implementations of ssl2_read and ssl3_read to ssl2_read_internal
5487 and ssl3_read_internal, respectively, and adding 'peek' parameters
5488 to them. The new ssl[23]_{read,peek} functions are calls to
5489 ssl[23]_read_internal with the 'peek' flag set appropriately.
5490 A 'peek' parameter has also been added to ssl3_read_bytes, which
5491 does the actual work for ssl3_read_internal.
5492 [Bodo Moeller]
5493
5494 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
5495 the method-specific "init()" handler. Also clean up ex_data after
5496 calling the method-specific "finish()" handler. Previously, this was
5497 happening the other way round.
5498 [Geoff Thorpe]
5499
5500 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
5501 The previous value, 12, was not always sufficient for BN_mod_exp().
5502 [Bodo Moeller]
5503
5504 *) Make sure that shared libraries get the internal name engine with
5505 the full version number and not just 0. This should mark the
5506 shared libraries as not backward compatible. Of course, this should
5507 be changed again when we can guarantee backward binary compatibility.
5508 [Richard Levitte]
5509
5510 *) Fix typo in get_cert_by_subject() in by_dir.c
5511 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
5512
5513 *) Rework the system to generate shared libraries:
5514
5515 - Make note of the expected extension for the shared libraries and
5516 if there is a need for symbolic links from for example libcrypto.so.0
5517 to libcrypto.so.0.9.7. There is extended info in Configure for
5518 that.
5519
5520 - Make as few rebuilds of the shared libraries as possible.
5521
5522 - Still avoid linking the OpenSSL programs with the shared libraries.
5523
5524 - When installing, install the shared libraries separately from the
5525 static ones.
5526 [Richard Levitte]
5527
5528 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
5529
5530 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
5531 and not in SSL_clear because the latter is also used by the
5532 accept/connect functions; previously, the settings made by
5533 SSL_set_read_ahead would be lost during the handshake.
5534 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
5535
5536 *) Correct util/mkdef.pl to be selective about disabled algorithms.
5537 Previously, it would create entries for disableed algorithms no
5538 matter what.
5539 [Richard Levitte]
5540
5541 *) Added several new manual pages for SSL_* function.
5542 [Lutz Jaenicke]
5543
5544 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
5545
5546 *) In ssl23_get_client_hello, generate an error message when faced
5547 with an initial SSL 3.0/TLS record that is too small to contain the
5548 first two bytes of the ClientHello message, i.e. client_version.
5549 (Note that this is a pathologic case that probably has never happened
5550 in real life.) The previous approach was to use the version number
5551 from the record header as a substitute; but our protocol choice
5552 should not depend on that one because it is not authenticated
5553 by the Finished messages.
5554 [Bodo Moeller]
5555
5556 *) More robust randomness gathering functions for Windows.
5557 [Jeffrey Altman <jaltman@columbia.edu>]
5558
5559 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
5560 not set then we don't setup the error code for issuer check errors
5561 to avoid possibly overwriting other errors which the callback does
5562 handle. If an application does set the flag then we assume it knows
5563 what it is doing and can handle the new informational codes
5564 appropriately.
5565 [Steve Henson]
5566
5567 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
5568 a general "ANY" type, as such it should be able to decode anything
5569 including tagged types. However it didn't check the class so it would
5570 wrongly interpret tagged types in the same way as their universal
5571 counterpart and unknown types were just rejected. Changed so that the
5572 tagged and unknown types are handled in the same way as a SEQUENCE:
5573 that is the encoding is stored intact. There is also a new type
5574 "V_ASN1_OTHER" which is used when the class is not universal, in this
5575 case we have no idea what the actual type is so we just lump them all
5576 together.
5577 [Steve Henson]
5578
5579 *) On VMS, stdout may very well lead to a file that is written to
5580 in a record-oriented fashion. That means that every write() will
5581 write a separate record, which will be read separately by the
5582 programs trying to read from it. This can be very confusing.
5583
5584 The solution is to put a BIO filter in the way that will buffer
5585 text until a linefeed is reached, and then write everything a
5586 line at a time, so every record written will be an actual line,
5587 not chunks of lines and not (usually doesn't happen, but I've
5588 seen it once) several lines in one record. BIO_f_linebuffer() is
5589 the answer.
5590
5591 Currently, it's a VMS-only method, because that's where it has
5592 been tested well enough.
5593 [Richard Levitte]
5594
5595 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
5596 it can return incorrect results.
5597 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
5598 but it was in 0.9.6-beta[12].)
5599 [Bodo Moeller]
5600
5601 *) Disable the check for content being present when verifying detached
5602 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
5603 include zero length content when signing messages.
5604 [Steve Henson]
5605
5606 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
5607 BIO_ctrl (for BIO pairs).
5608 [Bodo Möller]
5609
5610 *) Add DSO method for VMS.
5611 [Richard Levitte]
5612
5613 *) Bug fix: Montgomery multiplication could produce results with the
5614 wrong sign.
5615 [Ulf Möller]
5616
5617 *) Add RPM specification openssl.spec and modify it to build three
5618 packages. The default package contains applications, application
5619 documentation and run-time libraries. The devel package contains
5620 include files, static libraries and function documentation. The
5621 doc package contains the contents of the doc directory. The original
5622 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5623 [Richard Levitte]
5624
5625 *) Add a large number of documentation files for many SSL routines.
5626 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5627
5628 *) Add a configuration entry for Sony News 4.
5629 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5630
5631 *) Don't set the two most significant bits to one when generating a
5632 random number < q in the DSA library.
5633 [Ulf Möller]
5634
5635 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5636 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5637 the underlying transport is blocking) if a handshake took place.
5638 (The default behaviour is needed by applications such as s_client
5639 and s_server that use select() to determine when to use SSL_read;
5640 but for applications that know in advance when to expect data, it
5641 just makes things more complicated.)
5642 [Bodo Moeller]
5643
5644 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5645 from EGD.
5646 [Ben Laurie]
5647
5648 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5649 work better on such systems.
5650 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5651
5652 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5653 Update PKCS12_parse() so it copies the friendlyName and the
5654 keyid to the certificates aux info.
5655 [Steve Henson]
5656
5657 *) Fix bug in PKCS7_verify() which caused an infinite loop
5658 if there was more than one signature.
5659 [Sven Uszpelkat <su@celocom.de>]
5660
5661 *) Major change in util/mkdef.pl to include extra information
5662 about each symbol, as well as presentig variables as well
5663 as functions. This change means that there's n more need
5664 to rebuild the .num files when some algorithms are excluded.
5665 [Richard Levitte]
5666
5667 *) Allow the verify time to be set by an application,
5668 rather than always using the current time.
5669 [Steve Henson]
5670
5671 *) Phase 2 verify code reorganisation. The certificate
5672 verify code now looks up an issuer certificate by a
5673 number of criteria: subject name, authority key id
5674 and key usage. It also verifies self signed certificates
5675 by the same criteria. The main comparison function is
5676 X509_check_issued() which performs these checks.
5677
5678 Lot of changes were necessary in order to support this
5679 without completely rewriting the lookup code.
5680
5681 Authority and subject key identifier are now cached.
5682
5683 The LHASH 'certs' is X509_STORE has now been replaced
5684 by a STACK_OF(X509_OBJECT). This is mainly because an
5685 LHASH can't store or retrieve multiple objects with
5686 the same hash value.
5687
5688 As a result various functions (which were all internal
5689 use only) have changed to handle the new X509_STORE
5690 structure. This will break anything that messed round
5691 with X509_STORE internally.
5692
5693 The functions X509_STORE_add_cert() now checks for an
5694 exact match, rather than just subject name.
5695
5696 The X509_STORE API doesn't directly support the retrieval
5697 of multiple certificates matching a given criteria, however
5698 this can be worked round by performing a lookup first
5699 (which will fill the cache with candidate certificates)
5700 and then examining the cache for matches. This is probably
5701 the best we can do without throwing out X509_LOOKUP
5702 entirely (maybe later...).
5703
5704 The X509_VERIFY_CTX structure has been enhanced considerably.
5705
5706 All certificate lookup operations now go via a get_issuer()
5707 callback. Although this currently uses an X509_STORE it
5708 can be replaced by custom lookups. This is a simple way
5709 to bypass the X509_STORE hackery necessary to make this
5710 work and makes it possible to use more efficient techniques
5711 in future. A very simple version which uses a simple
5712 STACK for its trusted certificate store is also provided
5713 using X509_STORE_CTX_trusted_stack().
5714
5715 The verify_cb() and verify() callbacks now have equivalents
5716 in the X509_STORE_CTX structure.
5717
5718 X509_STORE_CTX also has a 'flags' field which can be used
5719 to customise the verify behaviour.
5720 [Steve Henson]
5721
5722 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5723 excludes S/MIME capabilities.
5724 [Steve Henson]
5725
5726 *) When a certificate request is read in keep a copy of the
5727 original encoding of the signed data and use it when outputing
5728 again. Signatures then use the original encoding rather than
5729 a decoded, encoded version which may cause problems if the
5730 request is improperly encoded.
5731 [Steve Henson]
5732
5733 *) For consistency with other BIO_puts implementations, call
5734 buffer_write(b, ...) directly in buffer_puts instead of calling
5735 BIO_write(b, ...).
5736
5737 In BIO_puts, increment b->num_write as in BIO_write.
5738 [Peter.Sylvester@EdelWeb.fr]
5739
5740 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5741 BN_zero, we may not return a BIGNUM with an array consisting of
5742 words set to zero.)
5743 [Bodo Moeller]
5744
5745 *) Avoid calling abort() from within the library when problems are
5746 detected, except if preprocessor symbols have been defined
5747 (such as REF_CHECK, BN_DEBUG etc.).
5748 [Bodo Moeller]
5749
5750 *) New openssl application 'rsautl'. This utility can be
5751 used for low level RSA operations. DER public key
5752 BIO/fp routines also added.
5753 [Steve Henson]
5754
5755 *) New Configure entry and patches for compiling on QNX 4.
5756 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5757
5758 *) A demo state-machine implementation was sponsored by
5759 Nuron (http://www.nuron.com/) and is now available in
5760 demos/state_machine.
5761 [Ben Laurie]
5762
5763 *) New options added to the 'dgst' utility for signature
5764 generation and verification.
5765 [Steve Henson]
5766
5767 *) Unrecognized PKCS#7 content types are now handled via a
5768 catch all ASN1_TYPE structure. This allows unsupported
5769 types to be stored as a "blob" and an application can
5770 encode and decode it manually.
5771 [Steve Henson]
5772
5773 *) Fix various signed/unsigned issues to make a_strex.c
5774 compile under VC++.
5775 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
5776
5777 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
5778 length if passed a buffer. ASN1_INTEGER_to_BN failed
5779 if passed a NULL BN and its argument was negative.
5780 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
5781
5782 *) Modification to PKCS#7 encoding routines to output definite
5783 length encoding. Since currently the whole structures are in
5784 memory there's not real point in using indefinite length
5785 constructed encoding. However if OpenSSL is compiled with
5786 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
5787 [Steve Henson]
5788
5789 *) Added BIO_vprintf() and BIO_vsnprintf().
5790 [Richard Levitte]
5791
5792 *) Added more prefixes to parse for in the the strings written
5793 through a logging bio, to cover all the levels that are available
5794 through syslog. The prefixes are now:
5795
5796 PANIC, EMERG, EMR => LOG_EMERG
5797 ALERT, ALR => LOG_ALERT
5798 CRIT, CRI => LOG_CRIT
5799 ERROR, ERR => LOG_ERR
5800 WARNING, WARN, WAR => LOG_WARNING
5801 NOTICE, NOTE, NOT => LOG_NOTICE
5802 INFO, INF => LOG_INFO
5803 DEBUG, DBG => LOG_DEBUG
5804
5805 and as before, if none of those prefixes are present at the
5806 beginning of the string, LOG_ERR is chosen.
5807
5808 On Win32, the LOG_* levels are mapped according to this:
5809
5810 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5811 LOG_WARNING => EVENTLOG_WARNING_TYPE
5812 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5813
5814 [Richard Levitte]
5815
5816 *) Made it possible to reconfigure with just the configuration
5817 argument "reconf" or "reconfigure". The command line arguments
5818 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5819 and are retrieved from there when reconfiguring.
5820 [Richard Levitte]
5821
5822 *) MD4 implemented.
5823 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5824
5825 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5826 [Richard Levitte]
5827
5828 *) The obj_dat.pl script was messing up the sorting of object
5829 names. The reason was that it compared the quoted version
5830 of strings as a result "OCSP" > "OCSP Signing" because
5831 " > SPACE. Changed script to store unquoted versions of
5832 names and add quotes on output. It was also omitting some
5833 names from the lookup table if they were given a default
5834 value (that is if SN is missing it is given the same
5835 value as LN and vice versa), these are now added on the
5836 grounds that if an object has a name we should be able to
5837 look it up. Finally added warning output when duplicate
5838 short or long names are found.
5839 [Steve Henson]
5840
5841 *) Changes needed for Tandem NSK.
5842 [Scott Uroff <scott@xypro.com>]
5843
5844 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5845 RSA_padding_check_SSLv23(), special padding was never detected
5846 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5847 version rollback attacks was not effective.
5848
5849 In s23_clnt.c, don't use special rollback-attack detection padding
5850 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5851 client; similarly, in s23_srvr.c, don't do the rollback check if
5852 SSL 2.0 is the only protocol enabled in the server.
5853 [Bodo Moeller]
5854
5855 *) Make it possible to get hexdumps of unprintable data with 'openssl
5856 asn1parse'. By implication, the functions ASN1_parse_dump() and
5857 BIO_dump_indent() are added.
5858 [Richard Levitte]
5859
5860 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5861 these print out strings and name structures based on various
5862 flags including RFC2253 support and proper handling of
5863 multibyte characters. Added options to the 'x509' utility
5864 to allow the various flags to be set.
5865 [Steve Henson]
5866
5867 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5868 Also change the functions X509_cmp_current_time() and
5869 X509_gmtime_adj() work with an ASN1_TIME structure,
5870 this will enable certificates using GeneralizedTime in validity
5871 dates to be checked.
5872 [Steve Henson]
5873
5874 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5875 negative public key encodings) on by default,
5876 NO_NEG_PUBKEY_BUG can be set to disable it.
5877 [Steve Henson]
5878
5879 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5880 content octets. An i2c_ASN1_OBJECT is unnecessary because
5881 the encoding can be trivially obtained from the structure.
5882 [Steve Henson]
5883
5884 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5885 not read locks (CRYPTO_r_[un]lock).
5886 [Bodo Moeller]
5887
5888 *) A first attempt at creating official support for shared
5889 libraries through configuration. I've kept it so the
5890 default is static libraries only, and the OpenSSL programs
5891 are always statically linked for now, but there are
5892 preparations for dynamic linking in place.
5893 This has been tested on Linux and Tru64.
5894 [Richard Levitte]
5895
5896 *) Randomness polling function for Win9x, as described in:
5897 Peter Gutmann, Software Generation of Practically Strong
5898 Random Numbers.
5899 [Ulf Möller]
5900
5901 *) Fix so PRNG is seeded in req if using an already existing
5902 DSA key.
5903 [Steve Henson]
5904
5905 *) New options to smime application. -inform and -outform
5906 allow alternative formats for the S/MIME message including
5907 PEM and DER. The -content option allows the content to be
5908 specified separately. This should allow things like Netscape
5909 form signing output easier to verify.
5910 [Steve Henson]
5911
5912 *) Fix the ASN1 encoding of tags using the 'long form'.
5913 [Steve Henson]
5914
5915 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5916 STRING types. These convert content octets to and from the
5917 underlying type. The actual tag and length octets are
5918 already assumed to have been read in and checked. These
5919 are needed because all other string types have virtually
5920 identical handling apart from the tag. By having versions
5921 of the ASN1 functions that just operate on content octets
5922 IMPLICIT tagging can be handled properly. It also allows
5923 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5924 and ASN1_INTEGER are identical apart from the tag.
5925 [Steve Henson]
5926
5927 *) Change the handling of OID objects as follows:
5928
5929 - New object identifiers are inserted in objects.txt, following
5930 the syntax given in objects.README.
5931 - objects.pl is used to process obj_mac.num and create a new
5932 obj_mac.h.
5933 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5934 obj_mac.h.
5935
5936 This is currently kind of a hack, and the perl code in objects.pl
5937 isn't very elegant, but it works as I intended. The simplest way
5938 to check that it worked correctly is to look in obj_dat.h and
5939 check the array nid_objs and make sure the objects haven't moved
5940 around (this is important!). Additions are OK, as well as
5941 consistent name changes.
5942 [Richard Levitte]
5943
5944 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5945 [Bodo Moeller]
5946
5947 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5948 The given file adds to whatever has already been seeded into the
5949 random pool through the RANDFILE configuration file option or
5950 environment variable, or the default random state file.
5951 [Richard Levitte]
5952
5953 *) mkstack.pl now sorts each macro group into lexical order.
5954 Previously the output order depended on the order the files
5955 appeared in the directory, resulting in needless rewriting
5956 of safestack.h .
5957 [Steve Henson]
5958
5959 *) Patches to make OpenSSL compile under Win32 again. Mostly
5960 work arounds for the VC++ problem that it treats func() as
5961 func(void). Also stripped out the parts of mkdef.pl that
5962 added extra typesafe functions: these no longer exist.
5963 [Steve Henson]
5964
5965 *) Reorganisation of the stack code. The macros are now all
5966 collected in safestack.h . Each macro is defined in terms of
5967 a "stack macro" of the form SKM_<name>(type, a, b). The
5968 DEBUG_SAFESTACK is now handled in terms of function casts,
5969 this has the advantage of retaining type safety without the
5970 use of additional functions. If DEBUG_SAFESTACK is not defined
5971 then the non typesafe macros are used instead. Also modified the
5972 mkstack.pl script to handle the new form. Needs testing to see
5973 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5974 the default if no major problems. Similar behaviour for ASN1_SET_OF
5975 and PKCS12_STACK_OF.
5976 [Steve Henson]
5977
5978 *) When some versions of IIS use the 'NET' form of private key the
5979 key derivation algorithm is different. Normally MD5(password) is
5980 used as a 128 bit RC4 key. In the modified case
5981 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5982 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5983 as the old Netscape_RSA functions except they have an additional
5984 'sgckey' parameter which uses the modified algorithm. Also added
5985 an -sgckey command line option to the rsa utility. Thanks to
5986 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5987 algorithm to openssl-dev.
5988 [Steve Henson]
5989
5990 *) The evp_local.h macros were using 'c.##kname' which resulted in
5991 invalid expansion on some systems (SCO 5.0.5 for example).
5992 Corrected to 'c.kname'.
5993 [Phillip Porch <root@theporch.com>]
5994
5995 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5996 a STACK of email addresses from a certificate or request, these look
5997 in the subject name and the subject alternative name extensions and
5998 omit any duplicate addresses.
5999 [Steve Henson]
6000
6001 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6002 This makes DSA verification about 2 % faster.
6003 [Bodo Moeller]
6004
6005 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6006 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6007 plus overhead for 1024 bit moduli).
6008 This makes exponentiations about 0.5 % faster for 1024 bit
6009 exponents (as measured by "openssl speed rsa2048").
6010 [Bodo Moeller]
6011
6012 *) Rename memory handling macros to avoid conflicts with other
6013 software:
6014 Malloc => OPENSSL_malloc
6015 Malloc_locked => OPENSSL_malloc_locked
6016 Realloc => OPENSSL_realloc
6017 Free => OPENSSL_free
6018 [Richard Levitte]
6019
6020 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6021 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6022 [Bodo Moeller]
6023
6024 *) CygWin32 support.
6025 [John Jarvie <jjarvie@newsguy.com>]
6026
6027 *) The type-safe stack code has been rejigged. It is now only compiled
6028 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6029 by default all type-specific stack functions are "#define"d back to
6030 standard stack functions. This results in more streamlined output
6031 but retains the type-safety checking possibilities of the original
6032 approach.
6033 [Geoff Thorpe]
6034
6035 *) The STACK code has been cleaned up, and certain type declarations
6036 that didn't make a lot of sense have been brought in line. This has
6037 also involved a cleanup of sorts in safestack.h to more correctly
6038 map type-safe stack functions onto their plain stack counterparts.
6039 This work has also resulted in a variety of "const"ifications of
6040 lots of the code, especially "_cmp" operations which should normally
6041 be prototyped with "const" parameters anyway.
6042 [Geoff Thorpe]
6043
6044 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6045 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6046 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6047 where all of 'md' is used each time the PRNG is used, but 'state'
6048 is used only indexed by a cyclic counter. As entropy may not be
6049 well distributed from the beginning, 'md' is important as a
6050 chaining variable. However, the output function chains only half
6051 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6052 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6053 in all of 'state' being rewritten, with the new values depending
6054 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6055 [Bodo Moeller]
6056
6057 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6058 the handshake is continued after ssl_verify_cert_chain();
6059 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6060 can lead to 'unexplainable' connection aborts later.
6061 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6062
6063 *) Major EVP API cipher revision.
6064 Add hooks for extra EVP features. This allows various cipher
6065 parameters to be set in the EVP interface. Support added for variable
6066 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6067 setting of RC2 and RC5 parameters.
6068
6069 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6070 ciphers.
6071
6072 Remove lots of duplicated code from the EVP library. For example *every*
6073 cipher init() function handles the 'iv' in the same way according to the
6074 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6075 for CFB and OFB modes they zero ctx->num.
6076
6077 New functionality allows removal of S/MIME code RC2 hack.
6078
6079 Most of the routines have the same form and so can be declared in terms
6080 of macros.
6081
6082 By shifting this to the top level EVP_CipherInit() it can be removed from
6083 all individual ciphers. If the cipher wants to handle IVs or keys
6084 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6085 flags.
6086
6087 Change lots of functions like EVP_EncryptUpdate() to now return a
6088 value: although software versions of the algorithms cannot fail
6089 any installed hardware versions can.
6090 [Steve Henson]
6091
6092 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6093 this option is set, tolerate broken clients that send the negotiated
6094 protocol version number instead of the requested protocol version
6095 number.
6096 [Bodo Moeller]
6097
6098 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6099 i.e. non-zero for export ciphersuites, zero otherwise.
6100 Previous versions had this flag inverted, inconsistent with
6101 rsa_tmp_cb (..._TMP_RSA_CB).
6102 [Bodo Moeller; problem reported by Amit Chopra]
6103
6104 *) Add missing DSA library text string. Work around for some IIS
6105 key files with invalid SEQUENCE encoding.
6106 [Steve Henson]
6107
6108 *) Add a document (doc/standards.txt) that list all kinds of standards
6109 and so on that are implemented in OpenSSL.
6110 [Richard Levitte]
6111
6112 *) Enhance c_rehash script. Old version would mishandle certificates
6113 with the same subject name hash and wouldn't handle CRLs at all.
6114 Added -fingerprint option to crl utility, to support new c_rehash
6115 features.
6116 [Steve Henson]
6117
6118 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6119 [Ulf Möller]
6120
6121 *) Fix for SSL server purpose checking. Server checking was
6122 rejecting certificates which had extended key usage present
6123 but no ssl client purpose.
6124 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6125
6126 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6127 is a little unclear about how a blank password is handled.
6128 Since the password in encoded as a BMPString with terminating
6129 double NULL a zero length password would end up as just the
6130 double NULL. However no password at all is different and is
6131 handled differently in the PKCS#12 key generation code. NS
6132 treats a blank password as zero length. MSIE treats it as no
6133 password on export: but it will try both on import. We now do
6134 the same: PKCS12_parse() tries zero length and no password if
6135 the password is set to "" or NULL (NULL is now a valid password:
6136 it wasn't before) as does the pkcs12 application.
6137 [Steve Henson]
6138
6139 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6140 perror when PEM_read_bio_X509_REQ fails, the error message must
6141 be obtained from the error queue.
6142 [Bodo Moeller]
6143
6144 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6145 it in ERR_remove_state if appropriate, and change ERR_get_state
6146 accordingly to avoid race conditions (this is necessary because
6147 thread_hash is no longer constant once set).
6148 [Bodo Moeller]
6149
6150 *) Bugfix for linux-elf makefile.one.
6151 [Ulf Möller]
6152
6153 *) RSA_get_default_method() will now cause a default
6154 RSA_METHOD to be chosen if one doesn't exist already.
6155 Previously this was only set during a call to RSA_new()
6156 or RSA_new_method(NULL) meaning it was possible for
6157 RSA_get_default_method() to return NULL.
6158 [Geoff Thorpe]
6159
6160 *) Added native name translation to the existing DSO code
6161 that will convert (if the flag to do so is set) filenames
6162 that are sufficiently small and have no path information
6163 into a canonical native form. Eg. "blah" converted to
6164 "libblah.so" or "blah.dll" etc.
6165 [Geoff Thorpe]
6166
6167 *) New function ERR_error_string_n(e, buf, len) which is like
6168 ERR_error_string(e, buf), but writes at most 'len' bytes
6169 including the 0 terminator. For ERR_error_string_n, 'buf'
6170 may not be NULL.
6171 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
6172
6173 *) CONF library reworked to become more general. A new CONF
6174 configuration file reader "class" is implemented as well as a
6175 new functions (NCONF_*, for "New CONF") to handle it. The now
6176 old CONF_* functions are still there, but are reimplemented to
6177 work in terms of the new functions. Also, a set of functions
6178 to handle the internal storage of the configuration data is
6179 provided to make it easier to write new configuration file
6180 reader "classes" (I can definitely see something reading a
6181 configuration file in XML format, for example), called _CONF_*,
6182 or "the configuration storage API"...
6183
6184 The new configuration file reading functions are:
6185
6186 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
6187 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
6188
6189 NCONF_default, NCONF_WIN32
6190
6191 NCONF_dump_fp, NCONF_dump_bio
6192
6193 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
6194 NCONF_new creates a new CONF object. This works in the same way
6195 as other interfaces in OpenSSL, like the BIO interface.
6196 NCONF_dump_* dump the internal storage of the configuration file,
6197 which is useful for debugging. All other functions take the same
6198 arguments as the old CONF_* functions wth the exception of the
6199 first that must be a `CONF *' instead of a `LHASH *'.
6200
6201 To make it easer to use the new classes with the old CONF_* functions,
6202 the function CONF_set_default_method is provided.
6203 [Richard Levitte]
6204
6205 *) Add '-tls1' option to 'openssl ciphers', which was already
6206 mentioned in the documentation but had not been implemented.
6207 (This option is not yet really useful because even the additional
6208 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
6209 [Bodo Moeller]
6210
6211 *) Initial DSO code added into libcrypto for letting OpenSSL (and
6212 OpenSSL-based applications) load shared libraries and bind to
6213 them in a portable way.
6214 [Geoff Thorpe, with contributions from Richard Levitte]
6215
6216 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
6217
6218 *) Make sure _lrotl and _lrotr are only used with MSVC.
6219
6220 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
6221 (the default implementation of RAND_status).
6222
6223 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
6224 to '-clrext' (= clear extensions), as intended and documented.
6225 [Bodo Moeller; inconsistency pointed out by Michael Attili
6226 <attili@amaxo.com>]
6227
6228 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
6229 was larger than the MD block size.
6230 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
6231
6232 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
6233 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
6234 using the passed key: if the passed key was a private key the result
6235 of X509_print(), for example, would be to print out all the private key
6236 components.
6237 [Steve Henson]
6238
6239 *) des_quad_cksum() byte order bug fix.
6240 [Ulf Möller, using the problem description in krb4-0.9.7, where
6241 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
6242
6243 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
6244 discouraged.
6245 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
6246
6247 *) For easily testing in shell scripts whether some command
6248 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
6249 returns with exit code 0 iff no command of the given name is available.
6250 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
6251 the output goes to stdout and nothing is printed to stderr.
6252 Additional arguments are always ignored.
6253
6254 Since for each cipher there is a command of the same name,
6255 the 'no-cipher' compilation switches can be tested this way.
6256
6257 ('openssl no-XXX' is not able to detect pseudo-commands such
6258 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
6259 [Bodo Moeller]
6260
6261 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
6262 [Bodo Moeller]
6263
6264 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
6265 is set; it will be thrown away anyway because each handshake creates
6266 its own key.
6267 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
6268 to parameters -- in previous versions (since OpenSSL 0.9.3) the
6269 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
6270 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
6271 [Bodo Moeller]
6272
6273 *) New s_client option -ign_eof: EOF at stdin is ignored, and
6274 'Q' and 'R' lose their special meanings (quit/renegotiate).
6275 This is part of what -quiet does; unlike -quiet, -ign_eof
6276 does not suppress any output.
6277 [Richard Levitte]
6278
6279 *) Add compatibility options to the purpose and trust code. The
6280 purpose X509_PURPOSE_ANY is "any purpose" which automatically
6281 accepts a certificate or CA, this was the previous behaviour,
6282 with all the associated security issues.
6283
6284 X509_TRUST_COMPAT is the old trust behaviour: only and
6285 automatically trust self signed roots in certificate store. A
6286 new trust setting X509_TRUST_DEFAULT is used to specify that
6287 a purpose has no associated trust setting and it should instead
6288 use the value in the default purpose.
6289 [Steve Henson]
6290
6291 *) Fix the PKCS#8 DSA private key code so it decodes keys again
6292 and fix a memory leak.
6293 [Steve Henson]
6294
6295 *) In util/mkerr.pl (which implements 'make errors'), preserve
6296 reason strings from the previous version of the .c file, as
6297 the default to have only downcase letters (and digits) in
6298 automatically generated reasons codes is not always appropriate.
6299 [Bodo Moeller]
6300
6301 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
6302 using strerror. Previously, ERR_reason_error_string() returned
6303 library names as reason strings for SYSerr; but SYSerr is a special
6304 case where small numbers are errno values, not library numbers.
6305 [Bodo Moeller]
6306
6307 *) Add '-dsaparam' option to 'openssl dhparam' application. This
6308 converts DSA parameters into DH parameters. (When creating parameters,
6309 DSA_generate_parameters is used.)
6310 [Bodo Moeller]
6311
6312 *) Include 'length' (recommended exponent length) in C code generated
6313 by 'openssl dhparam -C'.
6314 [Bodo Moeller]
6315
6316 *) The second argument to set_label in perlasm was already being used
6317 so couldn't be used as a "file scope" flag. Moved to third argument
6318 which was free.
6319 [Steve Henson]
6320
6321 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
6322 instead of RAND_bytes for encryption IVs and salts.
6323 [Bodo Moeller]
6324
6325 *) Include RAND_status() into RAND_METHOD instead of implementing
6326 it only for md_rand.c Otherwise replacing the PRNG by calling
6327 RAND_set_rand_method would be impossible.
6328 [Bodo Moeller]
6329
6330 *) Don't let DSA_generate_key() enter an infinite loop if the random
6331 number generation fails.
6332 [Bodo Moeller]
6333
6334 *) New 'rand' application for creating pseudo-random output.
6335 [Bodo Moeller]
6336
6337 *) Added configuration support for Linux/IA64
6338 [Rolf Haberrecker <rolf@suse.de>]
6339
6340 *) Assembler module support for Mingw32.
6341 [Ulf Möller]
6342
6343 *) Shared library support for HPUX (in shlib/).
6344 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
6345
6346 *) Shared library support for Solaris gcc.
6347 [Lutz Behnke <behnke@trustcenter.de>]
6348
6349 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
6350
6351 *) PKCS7_encrypt() was adding text MIME headers twice because they
6352 were added manually and by SMIME_crlf_copy().
6353 [Steve Henson]
6354
6355 *) In bntest.c don't call BN_rand with zero bits argument.
6356 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
6357
6358 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
6359 case was implemented. This caused BN_div_recp() to fail occasionally.
6360 [Ulf Möller]
6361
6362 *) Add an optional second argument to the set_label() in the perl
6363 assembly language builder. If this argument exists and is set
6364 to 1 it signals that the assembler should use a symbol whose
6365 scope is the entire file, not just the current function. This
6366 is needed with MASM which uses the format label:: for this scope.
6367 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
6368
6369 *) Change the ASN1 types so they are typedefs by default. Before
6370 almost all types were #define'd to ASN1_STRING which was causing
6371 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
6372 for example.
6373 [Steve Henson]
6374
6375 *) Change names of new functions to the new get1/get0 naming
6376 convention: After 'get1', the caller owns a reference count
6377 and has to call ..._free; 'get0' returns a pointer to some
6378 data structure without incrementing reference counters.
6379 (Some of the existing 'get' functions increment a reference
6380 counter, some don't.)
6381 Similarly, 'set1' and 'add1' functions increase reference
6382 counters or duplicate objects.
6383 [Steve Henson]
6384
6385 *) Allow for the possibility of temp RSA key generation failure:
6386 the code used to assume it always worked and crashed on failure.
6387 [Steve Henson]
6388
6389 *) Fix potential buffer overrun problem in BIO_printf().
6390 [Ulf Möller, using public domain code by Patrick Powell; problem
6391 pointed out by David Sacerdote <das33@cornell.edu>]
6392
6393 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
6394 RAND_egd() and RAND_status(). In the command line application,
6395 the EGD socket can be specified like a seed file using RANDFILE
6396 or -rand.
6397 [Ulf Möller]
6398
6399 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
6400 Some CAs (e.g. Verisign) distribute certificates in this form.
6401 [Steve Henson]
6402
6403 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
6404 list to exclude them. This means that no special compilation option
6405 is needed to use anonymous DH: it just needs to be included in the
6406 cipher list.
6407 [Steve Henson]
6408
6409 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
6410 EVP_MD_type. The old functionality is available in a new macro called
6411 EVP_MD_md(). Change code that uses it and update docs.
6412 [Steve Henson]
6413
6414 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
6415 where the 'void *' argument is replaced by a function pointer argument.
6416 Previously 'void *' was abused to point to functions, which works on
6417 many platforms, but is not correct. As these functions are usually
6418 called by macros defined in OpenSSL header files, most source code
6419 should work without changes.
6420 [Richard Levitte]
6421
6422 *) <openssl/opensslconf.h> (which is created by Configure) now contains
6423 sections with information on -D... compiler switches used for
6424 compiling the library so that applications can see them. To enable
6425 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
6426 must be defined. E.g.,
6427 #define OPENSSL_ALGORITHM_DEFINES
6428 #include <openssl/opensslconf.h>
6429 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
6430 [Richard Levitte, Ulf and Bodo Möller]
6431
6432 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
6433 record layer.
6434 [Bodo Moeller]
6435
6436 *) Change the 'other' type in certificate aux info to a STACK_OF
6437 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
6438 the required ASN1 format: arbitrary types determined by an OID.
6439 [Steve Henson]
6440
6441 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
6442 argument to 'req'. This is not because the function is newer or
6443 better than others it just uses the work 'NEW' in the certificate
6444 request header lines. Some software needs this.
6445 [Steve Henson]
6446
6447 *) Reorganise password command line arguments: now passwords can be
6448 obtained from various sources. Delete the PEM_cb function and make
6449 it the default behaviour: i.e. if the callback is NULL and the
6450 usrdata argument is not NULL interpret it as a null terminated pass
6451 phrase. If usrdata and the callback are NULL then the pass phrase
6452 is prompted for as usual.
6453 [Steve Henson]
6454
6455 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
6456 the support is automatically enabled. The resulting binaries will
6457 autodetect the card and use it if present.
6458 [Ben Laurie and Compaq Inc.]
6459
6460 *) Work around for Netscape hang bug. This sends certificate request
6461 and server done in one record. Since this is perfectly legal in the
6462 SSL/TLS protocol it isn't a "bug" option and is on by default. See
6463 the bugs/SSLv3 entry for more info.
6464 [Steve Henson]
6465
6466 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
6467 [Andy Polyakov]
6468
6469 *) Add -rand argument to smime and pkcs12 applications and read/write
6470 of seed file.
6471 [Steve Henson]
6472
6473 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
6474 [Bodo Moeller]
6475
6476 *) Add command line password options to the remaining applications.
6477 [Steve Henson]
6478
6479 *) Bug fix for BN_div_recp() for numerators with an even number of
6480 bits.
6481 [Ulf Möller]
6482
6483 *) More tests in bntest.c, and changed test_bn output.
6484 [Ulf Möller]
6485
6486 *) ./config recognizes MacOS X now.
6487 [Andy Polyakov]
6488
6489 *) Bug fix for BN_div() when the first words of num and divsor are
6490 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
6491 [Ulf Möller]
6492
6493 *) Add support for various broken PKCS#8 formats, and command line
6494 options to produce them.
6495 [Steve Henson]
6496
6497 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
6498 get temporary BIGNUMs from a BN_CTX.
6499 [Ulf Möller]
6500
6501 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
6502 for p == 0.
6503 [Ulf Möller]
6504
6505 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
6506 include a #define from the old name to the new. The original intent
6507 was that statically linked binaries could for example just call
6508 SSLeay_add_all_ciphers() to just add ciphers to the table and not
6509 link with digests. This never worked becayse SSLeay_add_all_digests()
6510 and SSLeay_add_all_ciphers() were in the same source file so calling
6511 one would link with the other. They are now in separate source files.
6512 [Steve Henson]
6513
6514 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
6515 [Steve Henson]
6516
6517 *) Use a less unusual form of the Miller-Rabin primality test (it used
6518 a binary algorithm for exponentiation integrated into the Miller-Rabin
6519 loop, our standard modexp algorithms are faster).
6520 [Bodo Moeller]
6521
6522 *) Support for the EBCDIC character set completed.
6523 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
6524
6525 *) Source code cleanups: use const where appropriate, eliminate casts,
6526 use void * instead of char * in lhash.
6527 [Ulf Möller]
6528
6529 *) Bugfix: ssl3_send_server_key_exchange was not restartable
6530 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
6531 this the server could overwrite ephemeral keys that the client
6532 has already seen).
6533 [Bodo Moeller]
6534
6535 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
6536 using 50 iterations of the Rabin-Miller test.
6537
6538 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
6539 iterations of the Rabin-Miller test as required by the appendix
6540 to FIPS PUB 186[-1]) instead of DSA_is_prime.
6541 As BN_is_prime_fasttest includes trial division, DSA parameter
6542 generation becomes much faster.
6543
6544 This implies a change for the callback functions in DSA_is_prime
6545 and DSA_generate_parameters: The callback function is called once
6546 for each positive witness in the Rabin-Miller test, not just
6547 occasionally in the inner loop; and the parameters to the
6548 callback function now provide an iteration count for the outer
6549 loop rather than for the current invocation of the inner loop.
6550 DSA_generate_parameters additionally can call the callback
6551 function with an 'iteration count' of -1, meaning that a
6552 candidate has passed the trial division test (when q is generated
6553 from an application-provided seed, trial division is skipped).
6554 [Bodo Moeller]
6555
6556 *) New function BN_is_prime_fasttest that optionally does trial
6557 division before starting the Rabin-Miller test and has
6558 an additional BN_CTX * argument (whereas BN_is_prime always
6559 has to allocate at least one BN_CTX).
6560 'callback(1, -1, cb_arg)' is called when a number has passed the
6561 trial division stage.
6562 [Bodo Moeller]
6563
6564 *) Fix for bug in CRL encoding. The validity dates weren't being handled
6565 as ASN1_TIME.
6566 [Steve Henson]
6567
6568 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
6569 [Steve Henson]
6570
6571 *) New function BN_pseudo_rand().
6572 [Ulf Möller]
6573
6574 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
6575 bignum version of BN_from_montgomery() with the working code from
6576 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
6577 the comments.
6578 [Ulf Möller]
6579
6580 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
6581 made it impossible to use the same SSL_SESSION data structure in
6582 SSL2 clients in multiple threads.
6583 [Bodo Moeller]
6584
6585 *) The return value of RAND_load_file() no longer counts bytes obtained
6586 by stat(). RAND_load_file(..., -1) is new and uses the complete file
6587 to seed the PRNG (previously an explicit byte count was required).
6588 [Ulf Möller, Bodo Möller]
6589
6590 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
6591 used (char *) instead of (void *) and had casts all over the place.
6592 [Steve Henson]
6593
6594 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
6595 [Ulf Möller]
6596
6597 *) Retain source code compatibility for BN_prime_checks macro:
6598 BN_is_prime(..., BN_prime_checks, ...) now uses
6599 BN_prime_checks_for_size to determine the appropriate number of
6600 Rabin-Miller iterations.
6601 [Ulf Möller]
6602
6603 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
6604 DH_CHECK_P_NOT_SAFE_PRIME.
6605 (Check if this is true? OpenPGP calls them "strong".)
6606 [Ulf Möller]
6607
6608 *) Merge the functionality of "dh" and "gendh" programs into a new program
6609 "dhparam". The old programs are retained for now but will handle DH keys
6610 (instead of parameters) in future.
6611 [Steve Henson]
6612
6613 *) Make the ciphers, s_server and s_client programs check the return values
6614 when a new cipher list is set.
6615 [Steve Henson]
6616
6617 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6618 ciphers. Before when the 56bit ciphers were enabled the sorting was
6619 wrong.
6620
6621 The syntax for the cipher sorting has been extended to support sorting by
6622 cipher-strength (using the strength_bits hard coded in the tables).
6623 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6624
6625 Fix a bug in the cipher-command parser: when supplying a cipher command
6626 string with an "undefined" symbol (neither command nor alphanumeric
6627 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6628 an error is flagged.
6629
6630 Due to the strength-sorting extension, the code of the
6631 ssl_create_cipher_list() function was completely rearranged. I hope that
6632 the readability was also increased :-)
6633 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6634
6635 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6636 for the first serial number and places 2 in the serial number file. This
6637 avoids problems when the root CA is created with serial number zero and
6638 the first user certificate has the same issuer name and serial number
6639 as the root CA.
6640 [Steve Henson]
6641
6642 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6643 the new code. Add documentation for this stuff.
6644 [Steve Henson]
6645
6646 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6647 X509_*() to X509at_*() on the grounds that they don't handle X509
6648 structures and behave in an analagous way to the X509v3 functions:
6649 they shouldn't be called directly but wrapper functions should be used
6650 instead.
6651
6652 So we also now have some wrapper functions that call the X509at functions
6653 when passed certificate requests. (TO DO: similar things can be done with
6654 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6655 things. Some of these need some d2i or i2d and print functionality
6656 because they handle more complex structures.)
6657 [Steve Henson]
6658
6659 *) Add missing #ifndefs that caused missing symbols when building libssl
6660 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6661 NO_RSA in ssl/s2*.c.
6662 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6663
6664 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6665 has a return value which indicates the quality of the random data
6666 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6667 error queue. New function RAND_pseudo_bytes() generates output that is
6668 guaranteed to be unique but not unpredictable. RAND_add is like
6669 RAND_seed, but takes an extra argument for an entropy estimate
6670 (RAND_seed always assumes full entropy).
6671 [Ulf Möller]
6672
6673 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6674 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6675 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6676 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6677 false-positive rate of at most 2^-80 for random input.
6678 [Bodo Moeller]
6679
6680 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6681 [Bodo Moeller]
6682
6683 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6684 in the 0.9.5 release), this returns the chain
6685 from an X509_CTX structure with a dup of the stack and all
6686 the X509 reference counts upped: so the stack will exist
6687 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6688 to use this.
6689
6690 Also make SSL_SESSION_print() print out the verify return
6691 code.
6692 [Steve Henson]
6693
6694 *) Add manpage for the pkcs12 command. Also change the default
6695 behaviour so MAC iteration counts are used unless the new
6696 -nomaciter option is used. This improves file security and
6697 only older versions of MSIE (4.0 for example) need it.
6698 [Steve Henson]
6699
6700 *) Honor the no-xxx Configure options when creating .DEF files.
6701 [Ulf Möller]
6702
6703 *) Add PKCS#10 attributes to field table: challengePassword,
6704 unstructuredName and unstructuredAddress. These are taken from
6705 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6706 international characters are used.
6707
6708 More changes to X509_ATTRIBUTE code: allow the setting of types
6709 based on strings. Remove the 'loc' parameter when adding
6710 attributes because these will be a SET OF encoding which is sorted
6711 in ASN1 order.
6712 [Steve Henson]
6713
6714 *) Initial changes to the 'req' utility to allow request generation
6715 automation. This will allow an application to just generate a template
6716 file containing all the field values and have req construct the
6717 request.
6718
6719 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6720 used all over the place including certificate requests and PKCS#7
6721 structures. They are currently handled manually where necessary with
6722 some primitive wrappers for PKCS#7. The new functions behave in a
6723 manner analogous to the X509 extension functions: they allow
6724 attributes to be looked up by NID and added.
6725
6726 Later something similar to the X509V3 code would be desirable to
6727 automatically handle the encoding, decoding and printing of the
6728 more complex types. The string types like challengePassword can
6729 be handled by the string table functions.
6730
6731 Also modified the multi byte string table handling. Now there is
6732 a 'global mask' which masks out certain types. The table itself
6733 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6734 is useful when for example there is only one permissible type
6735 (as in countryName) and using the mask might result in no valid
6736 types at all.
6737 [Steve Henson]
6738
6739 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6740 SSL_get_peer_finished to allow applications to obtain the latest
6741 Finished messages sent to the peer or expected from the peer,
6742 respectively. (SSL_get_peer_finished is usually the Finished message
6743 actually received from the peer, otherwise the protocol will be aborted.)
6744
6745 As the Finished message are message digests of the complete handshake
6746 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6747 be used for external authentication procedures when the authentication
6748 provided by SSL/TLS is not desired or is not enough.
6749 [Bodo Moeller]
6750
6751 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6752 the host supports BWX extension and if Compaq C is present on the
6753 $PATH. Just exploiting of the BWX extension results in 20-30%
6754 performance kick for some algorithms, e.g. DES and RC4 to mention
6755 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6756 SHA1.
6757 [Andy Polyakov]
6758
6759 *) Add support for MS "fast SGC". This is arguably a violation of the
6760 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6761 weak crypto and after checking the certificate is SGC a second one
6762 with strong crypto. MS SGC stops the first handshake after receiving
6763 the server certificate message and sends a second client hello. Since
6764 a server will typically do all the time consuming operations before
6765 expecting any further messages from the client (server key exchange
6766 is the most expensive) there is little difference between the two.
6767
6768 To get OpenSSL to support MS SGC we have to permit a second client
6769 hello message after we have sent server done. In addition we have to
6770 reset the MAC if we do get this second client hello.
6771 [Steve Henson]
6772
6773 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
6774 if a DER encoded private key is RSA or DSA traditional format. Changed
6775 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
6776 format DER encoded private key. Newer code should use PKCS#8 format which
6777 has the key type encoded in the ASN1 structure. Added DER private key
6778 support to pkcs8 application.
6779 [Steve Henson]
6780
6781 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
6782 ciphersuites has been selected (as required by the SSL 3/TLS 1
6783 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
6784 is set, we interpret this as a request to violate the specification
6785 (the worst that can happen is a handshake failure, and 'correct'
6786 behaviour would result in a handshake failure anyway).
6787 [Bodo Moeller]
6788
6789 *) In SSL_CTX_add_session, take into account that there might be multiple
6790 SSL_SESSION structures with the same session ID (e.g. when two threads
6791 concurrently obtain them from an external cache).
6792 The internal cache can handle only one SSL_SESSION with a given ID,
6793 so if there's a conflict, we now throw out the old one to achieve
6794 consistency.
6795 [Bodo Moeller]
6796
6797 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6798 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6799 some routines that use cipher OIDs: some ciphers do not have OIDs
6800 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6801 example.
6802 [Steve Henson]
6803
6804 *) Simplify the trust setting structure and code. Now we just have
6805 two sequences of OIDs for trusted and rejected settings. These will
6806 typically have values the same as the extended key usage extension
6807 and any application specific purposes.
6808
6809 The trust checking code now has a default behaviour: it will just
6810 check for an object with the same NID as the passed id. Functions can
6811 be provided to override either the default behaviour or the behaviour
6812 for a given id. SSL client, server and email already have functions
6813 in place for compatibility: they check the NID and also return "trusted"
6814 if the certificate is self signed.
6815 [Steve Henson]
6816
6817 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6818 traditional format into an EVP_PKEY structure.
6819 [Steve Henson]
6820
6821 *) Add a password callback function PEM_cb() which either prompts for
6822 a password if usr_data is NULL or otherwise assumes it is a null
6823 terminated password. Allow passwords to be passed on command line
6824 environment or config files in a few more utilities.
6825 [Steve Henson]
6826
6827 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6828 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6829 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6830 Update documentation.
6831 [Steve Henson]
6832
6833 *) Support for ASN1 "NULL" type. This could be handled before by using
6834 ASN1_TYPE but there wasn't any function that would try to read a NULL
6835 and produce an error if it couldn't. For compatibility we also have
6836 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6837 don't allocate anything because they don't need to.
6838 [Steve Henson]
6839
6840 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6841 for details.
6842 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6843
6844 *) Rebuild of the memory allocation routines used by OpenSSL code and
6845 possibly others as well. The purpose is to make an interface that
6846 provide hooks so anyone can build a separate set of allocation and
6847 deallocation routines to be used by OpenSSL, for example memory
6848 pool implementations, or something else, which was previously hard
6849 since Malloc(), Realloc() and Free() were defined as macros having
6850 the values malloc, realloc and free, respectively (except for Win32
6851 compilations). The same is provided for memory debugging code.
6852 OpenSSL already comes with functionality to find memory leaks, but
6853 this gives people a chance to debug other memory problems.
6854
6855 With these changes, a new set of functions and macros have appeared:
6856
6857 CRYPTO_set_mem_debug_functions() [F]
6858 CRYPTO_get_mem_debug_functions() [F]
6859 CRYPTO_dbg_set_options() [F]
6860 CRYPTO_dbg_get_options() [F]
6861 CRYPTO_malloc_debug_init() [M]
6862
6863 The memory debug functions are NULL by default, unless the library
6864 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6865 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6866 gives the standard debugging functions that come with OpenSSL) or
6867 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6868 provided by the library user) must be used. When the standard
6869 debugging functions are used, CRYPTO_dbg_set_options can be used to
6870 request additional information:
6871 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6872 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6873
6874 Also, things like CRYPTO_set_mem_functions will always give the
6875 expected result (the new set of functions is used for allocation
6876 and deallocation) at all times, regardless of platform and compiler
6877 options.
6878
6879 To finish it up, some functions that were never use in any other
6880 way than through macros have a new API and new semantic:
6881
6882 CRYPTO_dbg_malloc()
6883 CRYPTO_dbg_realloc()
6884 CRYPTO_dbg_free()
6885
6886 All macros of value have retained their old syntax.
6887 [Richard Levitte and Bodo Moeller]
6888
6889 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6890 ordering of SMIMECapabilities wasn't in "strength order" and there
6891 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6892 algorithm.
6893 [Steve Henson]
6894
6895 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6896 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6897 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6898
6899 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6900 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6901 functionality to handle multipart/signed properly) and a utility
6902 called 'smime' to call all this stuff. This is based on code I
6903 originally wrote for Celo who have kindly allowed it to be
6904 included in OpenSSL.
6905 [Steve Henson]
6906
6907 *) Add variants des_set_key_checked and des_set_key_unchecked of
6908 des_set_key (aka des_key_sched). Global variable des_check_key
6909 decides which of these is called by des_set_key; this way
6910 des_check_key behaves as it always did, but applications and
6911 the library itself, which was buggy for des_check_key == 1,
6912 have a cleaner way to pick the version they need.
6913 [Bodo Moeller]
6914
6915 *) New function PKCS12_newpass() which changes the password of a
6916 PKCS12 structure.
6917 [Steve Henson]
6918
6919 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6920 dynamic mix. In both cases the ids can be used as an index into the
6921 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6922 functions so they accept a list of the field values and the
6923 application doesn't need to directly manipulate the X509_TRUST
6924 structure.
6925 [Steve Henson]
6926
6927 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6928 need initialising.
6929 [Steve Henson]
6930
6931 *) Modify the way the V3 extension code looks up extensions. This now
6932 works in a similar way to the object code: we have some "standard"
6933 extensions in a static table which is searched with OBJ_bsearch()
6934 and the application can add dynamic ones if needed. The file
6935 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6936 updated whenever a new extension is added to the core code and kept
6937 in ext_nid order. There is a simple program 'tabtest.c' which checks
6938 this. New extensions are not added too often so this file can readily
6939 be maintained manually.
6940
6941 There are two big advantages in doing things this way. The extensions
6942 can be looked up immediately and no longer need to be "added" using
6943 X509V3_add_standard_extensions(): this function now does nothing.
6944 [Side note: I get *lots* of email saying the extension code doesn't
6945 work because people forget to call this function]
6946 Also no dynamic allocation is done unless new extensions are added:
6947 so if we don't add custom extensions there is no need to call
6948 X509V3_EXT_cleanup().
6949 [Steve Henson]
6950
6951 *) Modify enc utility's salting as follows: make salting the default. Add a
6952 magic header, so unsalted files fail gracefully instead of just decrypting
6953 to garbage. This is because not salting is a big security hole, so people
6954 should be discouraged from doing it.
6955 [Ben Laurie]
6956
6957 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6958 digest to be passed on the command line but it only used this
6959 parameter when signing a certificate. Modified so all relevant
6960 operations are affected by the digest parameter including the
6961 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6962 DSA key was used because it didn't fix the digest.
6963 [Steve Henson]
6964
6965 *) Initial certificate chain verify code. Currently tests the untrusted
6966 certificates for consistency with the verify purpose (which is set
6967 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6968
6969 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6970 this is because it will reject chains with invalid extensions whereas
6971 every previous version of OpenSSL and SSLeay made no checks at all.
6972
6973 Trust code: checks the root CA for the relevant trust settings. Trust
6974 settings have an initial value consistent with the verify purpose: e.g.
6975 if the verify purpose is for SSL client use it expects the CA to be
6976 trusted for SSL client use. However the default value can be changed to
6977 permit custom trust settings: one example of this would be to only trust
6978 certificates from a specific "secure" set of CAs.
6979
6980 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6981 which should be used for version portability: especially since the
6982 verify structure is likely to change more often now.
6983
6984 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6985 to set them. If not set then assume SSL clients will verify SSL servers
6986 and vice versa.
6987
6988 Two new options to the verify program: -untrusted allows a set of
6989 untrusted certificates to be passed in and -purpose which sets the
6990 intended purpose of the certificate. If a purpose is set then the
6991 new chain verify code is used to check extension consistency.
6992 [Steve Henson]
6993
6994 *) Support for the authority information access extension.
6995 [Steve Henson]
6996
6997 *) Modify RSA and DSA PEM read routines to transparently handle
6998 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6999 public keys in a format compatible with certificate
7000 SubjectPublicKeyInfo structures. Unfortunately there were already
7001 functions called *_PublicKey_* which used various odd formats so
7002 these are retained for compatibility: however the DSA variants were
7003 never in a public release so they have been deleted. Changed dsa/rsa
7004 utilities to handle the new format: note no releases ever handled public
7005 keys so we should be OK.
7006
7007 The primary motivation for this change is to avoid the same fiasco
7008 that dogs private keys: there are several incompatible private key
7009 formats some of which are standard and some OpenSSL specific and
7010 require various evil hacks to allow partial transparent handling and
7011 even then it doesn't work with DER formats. Given the option anything
7012 other than PKCS#8 should be dumped: but the other formats have to
7013 stay in the name of compatibility.
7014
7015 With public keys and the benefit of hindsight one standard format
7016 is used which works with EVP_PKEY, RSA or DSA structures: though
7017 it clearly returns an error if you try to read the wrong kind of key.
7018
7019 Added a -pubkey option to the 'x509' utility to output the public key.
7020 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7021 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7022 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7023 that do the same as the EVP_PKEY_assign_*() except they up the
7024 reference count of the added key (they don't "swallow" the
7025 supplied key).
7026 [Steve Henson]
7027
7028 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7029 CRLs would fail if the file contained no certificates or no CRLs:
7030 added a new function to read in both types and return the number
7031 read: this means that if none are read it will be an error. The
7032 DER versions of the certificate and CRL reader would always fail
7033 because it isn't possible to mix certificates and CRLs in DER format
7034 without choking one or the other routine. Changed this to just read
7035 a certificate: this is the best we can do. Also modified the code
7036 in apps/verify.c to take notice of return codes: it was previously
7037 attempting to read in certificates from NULL pointers and ignoring
7038 any errors: this is one reason why the cert and CRL reader seemed
7039 to work. It doesn't check return codes from the default certificate
7040 routines: these may well fail if the certificates aren't installed.
7041 [Steve Henson]
7042
7043 *) Code to support otherName option in GeneralName.
7044 [Steve Henson]
7045
7046 *) First update to verify code. Change the verify utility
7047 so it warns if it is passed a self signed certificate:
7048 for consistency with the normal behaviour. X509_verify
7049 has been modified to it will now verify a self signed
7050 certificate if *exactly* the same certificate appears
7051 in the store: it was previously impossible to trust a
7052 single self signed certificate. This means that:
7053 openssl verify ss.pem
7054 now gives a warning about a self signed certificate but
7055 openssl verify -CAfile ss.pem ss.pem
7056 is OK.
7057 [Steve Henson]
7058
7059 *) For servers, store verify_result in SSL_SESSION data structure
7060 (and add it to external session representation).
7061 This is needed when client certificate verifications fails,
7062 but an application-provided verification callback (set by
7063 SSL_CTX_set_cert_verify_callback) allows accepting the session
7064 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7065 but returns 1): When the session is reused, we have to set
7066 ssl->verify_result to the appropriate error code to avoid
7067 security holes.
7068 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7069
7070 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7071 case in PKCS7_dataInit() where the signed PKCS7 structure
7072 didn't contain any existing data because it was being created.
7073 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7074
7075 *) Add a salt to the key derivation routines in enc.c. This
7076 forms the first 8 bytes of the encrypted file. Also add a
7077 -S option to allow a salt to be input on the command line.
7078 [Steve Henson]
7079
7080 *) New function X509_cmp(). Oddly enough there wasn't a function
7081 to compare two certificates. We do this by working out the SHA1
7082 hash and comparing that. X509_cmp() will be needed by the trust
7083 code.
7084 [Steve Henson]
7085
7086 *) SSL_get1_session() is like SSL_get_session(), but increments
7087 the reference count in the SSL_SESSION returned.
7088 [Geoff Thorpe <geoff@eu.c2.net>]
7089
7090 *) Fix for 'req': it was adding a null to request attributes.
7091 Also change the X509_LOOKUP and X509_INFO code to handle
7092 certificate auxiliary information.
7093 [Steve Henson]
7094
7095 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7096 the 'enc' command.
7097 [Steve Henson]
7098
7099 *) Add the possibility to add extra information to the memory leak
7100 detecting output, to form tracebacks, showing from where each
7101 allocation was originated: CRYPTO_push_info("constant string") adds
7102 the string plus current file name and line number to a per-thread
7103 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7104 is like calling CYRPTO_pop_info() until the stack is empty.
7105 Also updated memory leak detection code to be multi-thread-safe.
7106 [Richard Levitte]
7107
7108 *) Add options -text and -noout to pkcs7 utility and delete the
7109 encryption options which never did anything. Update docs.
7110 [Steve Henson]
7111
7112 *) Add options to some of the utilities to allow the pass phrase
7113 to be included on either the command line (not recommended on
7114 OSes like Unix) or read from the environment. Update the
7115 manpages and fix a few bugs.
7116 [Steve Henson]
7117
7118 *) Add a few manpages for some of the openssl commands.
7119 [Steve Henson]
7120
7121 *) Fix the -revoke option in ca. It was freeing up memory twice,
7122 leaking and not finding already revoked certificates.
7123 [Steve Henson]
7124
7125 *) Extensive changes to support certificate auxiliary information.
7126 This involves the use of X509_CERT_AUX structure and X509_AUX
7127 functions. An X509_AUX function such as PEM_read_X509_AUX()
7128 can still read in a certificate file in the usual way but it
7129 will also read in any additional "auxiliary information". By
7130 doing things this way a fair degree of compatibility can be
7131 retained: existing certificates can have this information added
7132 using the new 'x509' options.
7133
7134 Current auxiliary information includes an "alias" and some trust
7135 settings. The trust settings will ultimately be used in enhanced
7136 certificate chain verification routines: currently a certificate
7137 can only be trusted if it is self signed and then it is trusted
7138 for all purposes.
7139 [Steve Henson]
7140
7141 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7142 The problem was that one of the replacement routines had not been working
7143 since SSLeay releases. For now the offending routine has been replaced
7144 with non-optimised assembler. Even so, this now gives around 95%
7145 performance improvement for 1024 bit RSA signs.
7146 [Mark Cox]
7147
7148 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7149 handling. Most clients have the effective key size in bits equal to
7150 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
7151 A few however don't do this and instead use the size of the decrypted key
7152 to determine the RC2 key length and the AlgorithmIdentifier to determine
7153 the effective key length. In this case the effective key length can still
7154 be 40 bits but the key length can be 168 bits for example. This is fixed
7155 by manually forcing an RC2 key into the EVP_PKEY structure because the
7156 EVP code can't currently handle unusual RC2 key sizes: it always assumes
7157 the key length and effective key length are equal.
7158 [Steve Henson]
7159
7160 *) Add a bunch of functions that should simplify the creation of
7161 X509_NAME structures. Now you should be able to do:
7162 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
7163 and have it automatically work out the correct field type and fill in
7164 the structures. The more adventurous can try:
7165 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
7166 and it will (hopefully) work out the correct multibyte encoding.
7167 [Steve Henson]
7168
7169 *) Change the 'req' utility to use the new field handling and multibyte
7170 copy routines. Before the DN field creation was handled in an ad hoc
7171 way in req, ca, and x509 which was rather broken and didn't support
7172 BMPStrings or UTF8Strings. Since some software doesn't implement
7173 BMPStrings or UTF8Strings yet, they can be enabled using the config file
7174 using the dirstring_type option. See the new comment in the default
7175 openssl.cnf for more info.
7176 [Steve Henson]
7177
7178 *) Make crypto/rand/md_rand.c more robust:
7179 - Assure unique random numbers after fork().
7180 - Make sure that concurrent threads access the global counter and
7181 md serializably so that we never lose entropy in them
7182 or use exactly the same state in multiple threads.
7183 Access to the large state is not always serializable because
7184 the additional locking could be a performance killer, and
7185 md should be large enough anyway.
7186 [Bodo Moeller]
7187
7188 *) New file apps/app_rand.c with commonly needed functionality
7189 for handling the random seed file.
7190
7191 Use the random seed file in some applications that previously did not:
7192 ca,
7193 dsaparam -genkey (which also ignored its '-rand' option),
7194 s_client,
7195 s_server,
7196 x509 (when signing).
7197 Except on systems with /dev/urandom, it is crucial to have a random
7198 seed file at least for key creation, DSA signing, and for DH exchanges;
7199 for RSA signatures we could do without one.
7200
7201 gendh and gendsa (unlike genrsa) used to read only the first byte
7202 of each file listed in the '-rand' option. The function as previously
7203 found in genrsa is now in app_rand.c and is used by all programs
7204 that support '-rand'.
7205 [Bodo Moeller]
7206
7207 *) In RAND_write_file, use mode 0600 for creating files;
7208 don't just chmod when it may be too late.
7209 [Bodo Moeller]
7210
7211 *) Report an error from X509_STORE_load_locations
7212 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
7213 [Bill Perry]
7214
7215 *) New function ASN1_mbstring_copy() this copies a string in either
7216 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
7217 into an ASN1_STRING type. A mask of permissible types is passed
7218 and it chooses the "minimal" type to use or an error if not type
7219 is suitable.
7220 [Steve Henson]
7221
7222 *) Add function equivalents to the various macros in asn1.h. The old
7223 macros are retained with an M_ prefix. Code inside the library can
7224 use the M_ macros. External code (including the openssl utility)
7225 should *NOT* in order to be "shared library friendly".
7226 [Steve Henson]
7227
7228 *) Add various functions that can check a certificate's extensions
7229 to see if it usable for various purposes such as SSL client,
7230 server or S/MIME and CAs of these types. This is currently
7231 VERY EXPERIMENTAL but will ultimately be used for certificate chain
7232 verification. Also added a -purpose flag to x509 utility to
7233 print out all the purposes.
7234 [Steve Henson]
7235
7236 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
7237 functions.
7238 [Steve Henson]
7239
7240 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
7241 for, obtain and decode and extension and obtain its critical flag.
7242 This allows all the necessary extension code to be handled in a
7243 single function call.
7244 [Steve Henson]
7245
7246 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
7247 platforms. See crypto/rc4/rc4_enc.c for further details.
7248 [Andy Polyakov]
7249
7250 *) New -noout option to asn1parse. This causes no output to be produced
7251 its main use is when combined with -strparse and -out to extract data
7252 from a file (which may not be in ASN.1 format).
7253 [Steve Henson]
7254
7255 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
7256 when producing the local key id.
7257 [Richard Levitte <levitte@stacken.kth.se>]
7258
7259 *) New option -dhparam in s_server. This allows a DH parameter file to be
7260 stated explicitly. If it is not stated then it tries the first server
7261 certificate file. The previous behaviour hard coded the filename
7262 "server.pem".
7263 [Steve Henson]
7264
7265 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
7266 a public key to be input or output. For example:
7267 openssl rsa -in key.pem -pubout -out pubkey.pem
7268 Also added necessary DSA public key functions to handle this.
7269 [Steve Henson]
7270
7271 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
7272 in the message. This was handled by allowing
7273 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
7274 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
7275
7276 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
7277 to the end of the strings whereas this didn't. This would cause problems
7278 if strings read with d2i_ASN1_bytes() were later modified.
7279 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
7280
7281 *) Fix for base64 decode bug. When a base64 bio reads only one line of
7282 data and it contains EOF it will end up returning an error. This is
7283 caused by input 46 bytes long. The cause is due to the way base64
7284 BIOs find the start of base64 encoded data. They do this by trying a
7285 trial decode on each line until they find one that works. When they
7286 do a flag is set and it starts again knowing it can pass all the
7287 data directly through the decoder. Unfortunately it doesn't reset
7288 the context it uses. This means that if EOF is reached an attempt
7289 is made to pass two EOFs through the context and this causes the
7290 resulting error. This can also cause other problems as well. As is
7291 usual with these problems it takes *ages* to find and the fix is
7292 trivial: move one line.
7293 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
7294
7295 *) Ugly workaround to get s_client and s_server working under Windows. The
7296 old code wouldn't work because it needed to select() on sockets and the
7297 tty (for keypresses and to see if data could be written). Win32 only
7298 supports select() on sockets so we select() with a 1s timeout on the
7299 sockets and then see if any characters are waiting to be read, if none
7300 are present then we retry, we also assume we can always write data to
7301 the tty. This isn't nice because the code then blocks until we've
7302 received a complete line of data and it is effectively polling the
7303 keyboard at 1s intervals: however it's quite a bit better than not
7304 working at all :-) A dedicated Windows application might handle this
7305 with an event loop for example.
7306 [Steve Henson]
7307
7308 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
7309 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
7310 will be called when RSA_sign() and RSA_verify() are used. This is useful
7311 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
7312 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
7313 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
7314 This necessitated the support of an extra signature type NID_md5_sha1
7315 for SSL signatures and modifications to the SSL library to use it instead
7316 of calling RSA_public_decrypt() and RSA_private_encrypt().
7317 [Steve Henson]
7318
7319 *) Add new -verify -CAfile and -CApath options to the crl program, these
7320 will lookup a CRL issuers certificate and verify the signature in a
7321 similar way to the verify program. Tidy up the crl program so it
7322 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
7323 less strict. It will now permit CRL extensions even if it is not
7324 a V2 CRL: this will allow it to tolerate some broken CRLs.
7325 [Steve Henson]
7326
7327 *) Initialize all non-automatic variables each time one of the openssl
7328 sub-programs is started (this is necessary as they may be started
7329 multiple times from the "OpenSSL>" prompt).
7330 [Lennart Bang, Bodo Moeller]
7331
7332 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
7333 removing all other RSA functionality (this is what NO_RSA does). This
7334 is so (for example) those in the US can disable those operations covered
7335 by the RSA patent while allowing storage and parsing of RSA keys and RSA
7336 key generation.
7337 [Steve Henson]
7338
7339 *) Non-copying interface to BIO pairs.
7340 (still largely untested)
7341 [Bodo Moeller]
7342
7343 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
7344 ASCII string. This was handled independently in various places before.
7345 [Steve Henson]
7346
7347 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
7348 UTF8 strings a character at a time.
7349 [Steve Henson]
7350
7351 *) Use client_version from client hello to select the protocol
7352 (s23_srvr.c) and for RSA client key exchange verification
7353 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
7354 [Bodo Moeller]
7355
7356 *) Add various utility functions to handle SPKACs, these were previously
7357 handled by poking round in the structure internals. Added new function
7358 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
7359 print, verify and generate SPKACs. Based on an original idea from
7360 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
7361 [Steve Henson]
7362
7363 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
7364 [Andy Polyakov]
7365
7366 *) Allow the config file extension section to be overwritten on the
7367 command line. Based on an original idea from Massimiliano Pala
7368 <madwolf@comune.modena.it>. The new option is called -extensions
7369 and can be applied to ca, req and x509. Also -reqexts to override
7370 the request extensions in req and -crlexts to override the crl extensions
7371 in ca.
7372 [Steve Henson]
7373
7374 *) Add new feature to the SPKAC handling in ca. Now you can include
7375 the same field multiple times by preceding it by "XXXX." for example:
7376 1.OU="Unit name 1"
7377 2.OU="Unit name 2"
7378 this is the same syntax as used in the req config file.
7379 [Steve Henson]
7380
7381 *) Allow certificate extensions to be added to certificate requests. These
7382 are specified in a 'req_extensions' option of the req section of the
7383 config file. They can be printed out with the -text option to req but
7384 are otherwise ignored at present.
7385 [Steve Henson]
7386
7387 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
7388 data read consists of only the final block it would not decrypted because
7389 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
7390 A misplaced 'break' also meant the decrypted final block might not be
7391 copied until the next read.
7392 [Steve Henson]
7393
7394 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
7395 a few extra parameters to the DH structure: these will be useful if
7396 for example we want the value of 'q' or implement X9.42 DH.
7397 [Steve Henson]
7398
7399 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
7400 provides hooks that allow the default DSA functions or functions on a
7401 "per key" basis to be replaced. This allows hardware acceleration and
7402 hardware key storage to be handled without major modification to the
7403 library. Also added low level modexp hooks and CRYPTO_EX structure and
7404 associated functions.
7405 [Steve Henson]
7406
7407 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
7408 as "read only": it can't be written to and the buffer it points to will
7409 not be freed. Reading from a read only BIO is much more efficient than
7410 a normal memory BIO. This was added because there are several times when
7411 an area of memory needs to be read from a BIO. The previous method was
7412 to create a memory BIO and write the data to it, this results in two
7413 copies of the data and an O(n^2) reading algorithm. There is a new
7414 function BIO_new_mem_buf() which creates a read only memory BIO from
7415 an area of memory. Also modified the PKCS#7 routines to use read only
7416 memory BIOs.
7417 [Steve Henson]
7418
7419 *) Bugfix: ssl23_get_client_hello did not work properly when called in
7420 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
7421 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
7422 but a retry condition occured while trying to read the rest.
7423 [Bodo Moeller]
7424
7425 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
7426 NID_pkcs7_encrypted by default: this was wrong since this should almost
7427 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
7428 the encrypted data type: this is a more sensible place to put it and it
7429 allows the PKCS#12 code to be tidied up that duplicated this
7430 functionality.
7431 [Steve Henson]
7432
7433 *) Changed obj_dat.pl script so it takes its input and output files on
7434 the command line. This should avoid shell escape redirection problems
7435 under Win32.
7436 [Steve Henson]
7437
7438 *) Initial support for certificate extension requests, these are included
7439 in things like Xenroll certificate requests. Included functions to allow
7440 extensions to be obtained and added.
7441 [Steve Henson]
7442
7443 *) -crlf option to s_client and s_server for sending newlines as
7444 CRLF (as required by many protocols).
7445 [Bodo Moeller]
7446
7447 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
7448
7449 *) Install libRSAglue.a when OpenSSL is built with RSAref.
7450 [Ralf S. Engelschall]
7451
7452 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
7453 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
7454
7455 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
7456 program.
7457 [Steve Henson]
7458
7459 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
7460 DH parameters/keys (q is lost during that conversion, but the resulting
7461 DH parameters contain its length).
7462
7463 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
7464 much faster than DH_generate_parameters (which creates parameters
7465 where p = 2*q + 1), and also the smaller q makes DH computations
7466 much more efficient (160-bit exponentiation instead of 1024-bit
7467 exponentiation); so this provides a convenient way to support DHE
7468 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
7469 utter importance to use
7470 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7471 or
7472 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7473 when such DH parameters are used, because otherwise small subgroup
7474 attacks may become possible!
7475 [Bodo Moeller]
7476
7477 *) Avoid memory leak in i2d_DHparams.
7478 [Bodo Moeller]
7479
7480 *) Allow the -k option to be used more than once in the enc program:
7481 this allows the same encrypted message to be read by multiple recipients.
7482 [Steve Henson]
7483
7484 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
7485 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
7486 it will always use the numerical form of the OID, even if it has a short
7487 or long name.
7488 [Steve Henson]
7489
7490 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
7491 method only got called if p,q,dmp1,dmq1,iqmp components were present,
7492 otherwise bn_mod_exp was called. In the case of hardware keys for example
7493 no private key components need be present and it might store extra data
7494 in the RSA structure, which cannot be accessed from bn_mod_exp.
7495 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
7496 private key operations.
7497 [Steve Henson]
7498
7499 *) Added support for SPARC Linux.
7500 [Andy Polyakov]
7501
7502 *) pem_password_cb function type incompatibly changed from
7503 typedef int pem_password_cb(char *buf, int size, int rwflag);
7504 to
7505 ....(char *buf, int size, int rwflag, void *userdata);
7506 so that applications can pass data to their callbacks:
7507 The PEM[_ASN1]_{read,write}... functions and macros now take an
7508 additional void * argument, which is just handed through whenever
7509 the password callback is called.
7510 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
7511
7512 New function SSL_CTX_set_default_passwd_cb_userdata.
7513
7514 Compatibility note: As many C implementations push function arguments
7515 onto the stack in reverse order, the new library version is likely to
7516 interoperate with programs that have been compiled with the old
7517 pem_password_cb definition (PEM_whatever takes some data that
7518 happens to be on the stack as its last argument, and the callback
7519 just ignores this garbage); but there is no guarantee whatsoever that
7520 this will work.
7521
7522 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
7523 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
7524 problems not only on Windows, but also on some Unix platforms.
7525 To avoid problematic command lines, these definitions are now in an
7526 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
7527 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
7528 [Bodo Moeller]
7529
7530 *) MIPS III/IV assembler module is reimplemented.
7531 [Andy Polyakov]
7532
7533 *) More DES library cleanups: remove references to srand/rand and
7534 delete an unused file.
7535 [Ulf Möller]
7536
7537 *) Add support for the the free Netwide assembler (NASM) under Win32,
7538 since not many people have MASM (ml) and it can be hard to obtain.
7539 This is currently experimental but it seems to work OK and pass all
7540 the tests. Check out INSTALL.W32 for info.
7541 [Steve Henson]
7542
7543 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
7544 without temporary keys kept an extra copy of the server key,
7545 and connections with temporary keys did not free everything in case
7546 of an error.
7547 [Bodo Moeller]
7548
7549 *) New function RSA_check_key and new openssl rsa option -check
7550 for verifying the consistency of RSA keys.
7551 [Ulf Moeller, Bodo Moeller]
7552
7553 *) Various changes to make Win32 compile work:
7554 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
7555 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
7556 comparison" warnings.
7557 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
7558 [Steve Henson]
7559
7560 *) Add a debugging option to PKCS#5 v2 key generation function: when
7561 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
7562 derived keys are printed to stderr.
7563 [Steve Henson]
7564
7565 *) Copy the flags in ASN1_STRING_dup().
7566 [Roman E. Pavlov <pre@mo.msk.ru>]
7567
7568 *) The x509 application mishandled signing requests containing DSA
7569 keys when the signing key was also DSA and the parameters didn't match.
7570
7571 It was supposed to omit the parameters when they matched the signing key:
7572 the verifying software was then supposed to automatically use the CA's
7573 parameters if they were absent from the end user certificate.
7574
7575 Omitting parameters is no longer recommended. The test was also
7576 the wrong way round! This was probably due to unusual behaviour in
7577 EVP_cmp_parameters() which returns 1 if the parameters match.
7578 This meant that parameters were omitted when they *didn't* match and
7579 the certificate was useless. Certificates signed with 'ca' didn't have
7580 this bug.
7581 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
7582
7583 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
7584 The interface is as follows:
7585 Applications can use
7586 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
7587 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
7588 "off" is now the default.
7589 The library internally uses
7590 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
7591 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
7592 to disable memory-checking temporarily.
7593
7594 Some inconsistent states that previously were possible (and were
7595 even the default) are now avoided.
7596
7597 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
7598 with each memory chunk allocated; this is occasionally more helpful
7599 than just having a counter.
7600
7601 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
7602
7603 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
7604 extensions.
7605 [Bodo Moeller]
7606
7607 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
7608 which largely parallels "options", but is for changing API behaviour,
7609 whereas "options" are about protocol behaviour.
7610 Initial "mode" flags are:
7611
7612 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7613 a single record has been written.
7614 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7615 retries use the same buffer location.
7616 (But all of the contents must be
7617 copied!)
7618 [Bodo Moeller]
7619
7620 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7621 worked.
7622
7623 *) Fix problems with no-hmac etc.
7624 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7625
7626 *) New functions RSA_get_default_method(), RSA_set_method() and
7627 RSA_get_method(). These allows replacement of RSA_METHODs without having
7628 to mess around with the internals of an RSA structure.
7629 [Steve Henson]
7630
7631 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7632 Also really enable memory leak checks in openssl.c and in some
7633 test programs.
7634 [Chad C. Mulligan, Bodo Moeller]
7635
7636 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7637 up the length of negative integers. This has now been simplified to just
7638 store the length when it is first determined and use it later, rather
7639 than trying to keep track of where data is copied and updating it to
7640 point to the end.
7641 [Steve Henson, reported by Brien Wheeler
7642 <bwheeler@authentica-security.com>]
7643
7644 *) Add a new function PKCS7_signatureVerify. This allows the verification
7645 of a PKCS#7 signature but with the signing certificate passed to the
7646 function itself. This contrasts with PKCS7_dataVerify which assumes the
7647 certificate is present in the PKCS#7 structure. This isn't always the
7648 case: certificates can be omitted from a PKCS#7 structure and be
7649 distributed by "out of band" means (such as a certificate database).
7650 [Steve Henson]
7651
7652 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7653 function prototypes in pem.h, also change util/mkdef.pl to add the
7654 necessary function names.
7655 [Steve Henson]
7656
7657 *) mk1mf.pl (used by Windows builds) did not properly read the
7658 options set by Configure in the top level Makefile, and Configure
7659 was not even able to write more than one option correctly.
7660 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7661 [Bodo Moeller]
7662
7663 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7664 file to be loaded from a BIO or FILE pointer. The BIO version will
7665 for example allow memory BIOs to contain config info.
7666 [Steve Henson]
7667
7668 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7669 Whoever hopes to achieve shared-library compatibility across versions
7670 must use this, not the compile-time macro.
7671 (Exercise 0.9.4: Which is the minimum library version required by
7672 such programs?)
7673 Note: All this applies only to multi-threaded programs, others don't
7674 need locks.
7675 [Bodo Moeller]
7676
7677 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7678 through a BIO pair triggered the default case, i.e.
7679 SSLerr(...,SSL_R_UNKNOWN_STATE).
7680 [Bodo Moeller]
7681
7682 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7683 can use the SSL library even if none of the specific BIOs is
7684 appropriate.
7685 [Bodo Moeller]
7686
7687 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7688 for the encoded length.
7689 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7690
7691 *) Add initial documentation of the X509V3 functions.
7692 [Steve Henson]
7693
7694 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7695 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7696 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7697 secure PKCS#8 private key format with a high iteration count.
7698 [Steve Henson]
7699
7700 *) Fix determination of Perl interpreter: A perl or perl5
7701 _directory_ in $PATH was also accepted as the interpreter.
7702 [Ralf S. Engelschall]
7703
7704 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7705 wrong with it but it was very old and did things like calling
7706 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7707 unusual formatting.
7708 [Steve Henson]
7709
7710 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7711 to use the new extension code.
7712 [Steve Henson]
7713
7714 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7715 with macros. This should make it easier to change their form, add extra
7716 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7717 constant.
7718 [Steve Henson]
7719
7720 *) Add to configuration table a new entry that can specify an alternative
7721 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7722 according to Mark Crispin <MRC@Panda.COM>.
7723 [Bodo Moeller]
7724
7725 #if 0
7726 *) DES CBC did not update the IV. Weird.
7727 [Ben Laurie]
7728 #else
7729 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7730 Changing the behaviour of the former might break existing programs --
7731 where IV updating is needed, des_ncbc_encrypt can be used.
7732 #endif
7733
7734 *) When bntest is run from "make test" it drives bc to check its
7735 calculations, as well as internally checking them. If an internal check
7736 fails, it needs to cause bc to give a non-zero result or make test carries
7737 on without noticing the failure. Fixed.
7738 [Ben Laurie]
7739
7740 *) DES library cleanups.
7741 [Ulf Möller]
7742
7743 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7744 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7745 ciphers. NOTE: although the key derivation function has been verified
7746 against some published test vectors it has not been extensively tested
7747 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7748 of v2.0.
7749 [Steve Henson]
7750
7751 *) Instead of "mkdir -p", which is not fully portable, use new
7752 Perl script "util/mkdir-p.pl".
7753 [Bodo Moeller]
7754
7755 *) Rewrite the way password based encryption (PBE) is handled. It used to
7756 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7757 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7758 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7759 the 'parameter' field of the AlgorithmIdentifier is passed to the
7760 underlying key generation function so it must do its own ASN1 parsing.
7761 This has also changed the EVP_PBE_CipherInit() function which now has a
7762 'parameter' argument instead of literal salt and iteration count values
7763 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
7764 [Steve Henson]
7765
7766 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
7767 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
7768 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
7769 KEY" because this clashed with PKCS#8 unencrypted string. Since this
7770 value was just used as a "magic string" and not used directly its
7771 value doesn't matter.
7772 [Steve Henson]
7773
7774 *) Introduce some semblance of const correctness to BN. Shame C doesn't
7775 support mutable.
7776 [Ben Laurie]
7777
7778 *) "linux-sparc64" configuration (ultrapenguin).
7779 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
7780 "linux-sparc" configuration.
7781 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
7782
7783 *) config now generates no-xxx options for missing ciphers.
7784 [Ulf Möller]
7785
7786 *) Support the EBCDIC character set (work in progress).
7787 File ebcdic.c not yet included because it has a different license.
7788 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7789
7790 *) Support BS2000/OSD-POSIX.
7791 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7792
7793 *) Make callbacks for key generation use void * instead of char *.
7794 [Ben Laurie]
7795
7796 *) Make S/MIME samples compile (not yet tested).
7797 [Ben Laurie]
7798
7799 *) Additional typesafe stacks.
7800 [Ben Laurie]
7801
7802 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7803 [Bodo Moeller]
7804
7805
7806 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7807
7808 *) New configuration variant "sco5-gcc".
7809
7810 *) Updated some demos.
7811 [Sean O Riordain, Wade Scholine]
7812
7813 *) Add missing BIO_free at exit of pkcs12 application.
7814 [Wu Zhigang]
7815
7816 *) Fix memory leak in conf.c.
7817 [Steve Henson]
7818
7819 *) Updates for Win32 to assembler version of MD5.
7820 [Steve Henson]
7821
7822 *) Set #! path to perl in apps/der_chop to where we found it
7823 instead of using a fixed path.
7824 [Bodo Moeller]
7825
7826 *) SHA library changes for irix64-mips4-cc.
7827 [Andy Polyakov]
7828
7829 *) Improvements for VMS support.
7830 [Richard Levitte]
7831
7832
7833 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7834
7835 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7836 This also avoids the problems with SC4.2 and unpatched SC5.
7837 [Andy Polyakov <appro@fy.chalmers.se>]
7838
7839 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7840 These are required because of the typesafe stack would otherwise break
7841 existing code. If old code used a structure member which used to be STACK
7842 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7843 sk_num or sk_value it would produce an error because the num, data members
7844 are not present in STACK_OF. Now it just produces a warning. sk_set
7845 replaces the old method of assigning a value to sk_value
7846 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7847 that does this will no longer work (and should use sk_set instead) but
7848 this could be regarded as a "questionable" behaviour anyway.
7849 [Steve Henson]
7850
7851 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7852 correctly handle encrypted S/MIME data.
7853 [Steve Henson]
7854
7855 *) Change type of various DES function arguments from des_cblock
7856 (which means, in function argument declarations, pointer to char)
7857 to des_cblock * (meaning pointer to array with 8 char elements),
7858 which allows the compiler to do more typechecking; it was like
7859 that back in SSLeay, but with lots of ugly casts.
7860
7861 Introduce new type const_des_cblock.
7862 [Bodo Moeller]
7863
7864 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7865 problems: find RecipientInfo structure that matches recipient certificate
7866 and initialise the ASN1 structures properly based on passed cipher.
7867 [Steve Henson]
7868
7869 *) Belatedly make the BN tests actually check the results.
7870 [Ben Laurie]
7871
7872 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7873 to and from BNs: it was completely broken. New compilation option
7874 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7875 key elements as negative integers.
7876 [Steve Henson]
7877
7878 *) Reorganize and speed up MD5.
7879 [Andy Polyakov <appro@fy.chalmers.se>]
7880
7881 *) VMS support.
7882 [Richard Levitte <richard@levitte.org>]
7883
7884 *) New option -out to asn1parse to allow the parsed structure to be
7885 output to a file. This is most useful when combined with the -strparse
7886 option to examine the output of things like OCTET STRINGS.
7887 [Steve Henson]
7888
7889 *) Make SSL library a little more fool-proof by not requiring any longer
7890 that SSL_set_{accept,connect}_state be called before
7891 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7892 in many applications because usually everything *appeared* to work as
7893 intended anyway -- now it really works as intended).
7894 [Bodo Moeller]
7895
7896 *) Move openssl.cnf out of lib/.
7897 [Ulf Möller]
7898
7899 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7900 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7901 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7902 [Ralf S. Engelschall]
7903
7904 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7905 handle PKCS#7 enveloped data properly.
7906 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7907
7908 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7909 copying pointers. The cert_st handling is changed by this in
7910 various ways (and thus what used to be known as ctx->default_cert
7911 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7912 any longer when s->cert does not give us what we need).
7913 ssl_cert_instantiate becomes obsolete by this change.
7914 As soon as we've got the new code right (possibly it already is?),
7915 we have solved a couple of bugs of the earlier code where s->cert
7916 was used as if it could not have been shared with other SSL structures.
7917
7918 Note that using the SSL API in certain dirty ways now will result
7919 in different behaviour than observed with earlier library versions:
7920 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7921 does not influence s as it used to.
7922
7923 In order to clean up things more thoroughly, inside SSL_SESSION
7924 we don't use CERT any longer, but a new structure SESS_CERT
7925 that holds per-session data (if available); currently, this is
7926 the peer's certificate chain and, for clients, the server's certificate
7927 and temporary key. CERT holds only those values that can have
7928 meaningful defaults in an SSL_CTX.
7929 [Bodo Moeller]
7930
7931 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7932 from the internal representation. Various PKCS#7 fixes: remove some
7933 evil casts and set the enc_dig_alg field properly based on the signing
7934 key type.
7935 [Steve Henson]
7936
7937 *) Allow PKCS#12 password to be set from the command line or the
7938 environment. Let 'ca' get its config file name from the environment
7939 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7940 and 'x509').
7941 [Steve Henson]
7942
7943 *) Allow certificate policies extension to use an IA5STRING for the
7944 organization field. This is contrary to the PKIX definition but
7945 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7946 extension option.
7947 [Steve Henson]
7948
7949 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7950 without disallowing inline assembler and the like for non-pedantic builds.
7951 [Ben Laurie]
7952
7953 *) Support Borland C++ builder.
7954 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7955
7956 *) Support Mingw32.
7957 [Ulf Möller]
7958
7959 *) SHA-1 cleanups and performance enhancements.
7960 [Andy Polyakov <appro@fy.chalmers.se>]
7961
7962 *) Sparc v8plus assembler for the bignum library.
7963 [Andy Polyakov <appro@fy.chalmers.se>]
7964
7965 *) Accept any -xxx and +xxx compiler options in Configure.
7966 [Ulf Möller]
7967
7968 *) Update HPUX configuration.
7969 [Anonymous]
7970
7971 *) Add missing sk_<type>_unshift() function to safestack.h
7972 [Ralf S. Engelschall]
7973
7974 *) New function SSL_CTX_use_certificate_chain_file that sets the
7975 "extra_cert"s in addition to the certificate. (This makes sense
7976 only for "PEM" format files, as chains as a whole are not
7977 DER-encoded.)
7978 [Bodo Moeller]
7979
7980 *) Support verify_depth from the SSL API.
7981 x509_vfy.c had what can be considered an off-by-one-error:
7982 Its depth (which was not part of the external interface)
7983 was actually counting the number of certificates in a chain;
7984 now it really counts the depth.
7985 [Bodo Moeller]
7986
7987 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7988 instead of X509err, which often resulted in confusing error
7989 messages since the error codes are not globally unique
7990 (e.g. an alleged error in ssl3_accept when a certificate
7991 didn't match the private key).
7992
7993 *) New function SSL_CTX_set_session_id_context that allows to set a default
7994 value (so that you don't need SSL_set_session_id_context for each
7995 connection using the SSL_CTX).
7996 [Bodo Moeller]
7997
7998 *) OAEP decoding bug fix.
7999 [Ulf Möller]
8000
8001 *) Support INSTALL_PREFIX for package builders, as proposed by
8002 David Harris.
8003 [Bodo Moeller]
8004
8005 *) New Configure options "threads" and "no-threads". For systems
8006 where the proper compiler options are known (currently Solaris
8007 and Linux), "threads" is the default.
8008 [Bodo Moeller]
8009
8010 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8011 [Bodo Moeller]
8012
8013 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8014 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8015 such as /usr/local/bin.
8016 [Bodo Moeller]
8017
8018 *) "make linux-shared" to build shared libraries.
8019 [Niels Poppe <niels@netbox.org>]
8020
8021 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8022 [Ulf Möller]
8023
8024 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8025 extension adding in x509 utility.
8026 [Steve Henson]
8027
8028 *) Remove NOPROTO sections and error code comments.
8029 [Ulf Möller]
8030
8031 *) Partial rewrite of the DEF file generator to now parse the ANSI
8032 prototypes.
8033 [Steve Henson]
8034
8035 *) New Configure options --prefix=DIR and --openssldir=DIR.
8036 [Ulf Möller]
8037
8038 *) Complete rewrite of the error code script(s). It is all now handled
8039 by one script at the top level which handles error code gathering,
8040 header rewriting and C source file generation. It should be much better
8041 than the old method: it now uses a modified version of Ulf's parser to
8042 read the ANSI prototypes in all header files (thus the old K&R definitions
8043 aren't needed for error creation any more) and do a better job of
8044 translating function codes into names. The old 'ASN1 error code imbedded
8045 in a comment' is no longer necessary and it doesn't use .err files which
8046 have now been deleted. Also the error code call doesn't have to appear all
8047 on one line (which resulted in some large lines...).
8048 [Steve Henson]
8049
8050 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8051 [Bodo Moeller]
8052
8053 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8054 0 (which usually indicates a closed connection), but continue reading.
8055 [Bodo Moeller]
8056
8057 *) Fix some race conditions.
8058 [Bodo Moeller]
8059
8060 *) Add support for CRL distribution points extension. Add Certificate
8061 Policies and CRL distribution points documentation.
8062 [Steve Henson]
8063
8064 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8065 [Ulf Möller]
8066
8067 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8068 8 of keying material. Merlin has also confirmed interop with this fix
8069 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8070 [Merlin Hughes <merlin@baltimore.ie>]
8071
8072 *) Fix lots of warnings.
8073 [Richard Levitte <levitte@stacken.kth.se>]
8074
8075 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8076 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8077 [Richard Levitte <levitte@stacken.kth.se>]
8078
8079 *) Fix problems with sizeof(long) == 8.
8080 [Andy Polyakov <appro@fy.chalmers.se>]
8081
8082 *) Change functions to ANSI C.
8083 [Ulf Möller]
8084
8085 *) Fix typos in error codes.
8086 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8087
8088 *) Remove defunct assembler files from Configure.
8089 [Ulf Möller]
8090
8091 *) SPARC v8 assembler BIGNUM implementation.
8092 [Andy Polyakov <appro@fy.chalmers.se>]
8093
8094 *) Support for Certificate Policies extension: both print and set.
8095 Various additions to support the r2i method this uses.
8096 [Steve Henson]
8097
8098 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8099 return a const string when you are expecting an allocated buffer.
8100 [Ben Laurie]
8101
8102 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8103 types DirectoryString and DisplayText.
8104 [Steve Henson]
8105
8106 *) Add code to allow r2i extensions to access the configuration database,
8107 add an LHASH database driver and add several ctx helper functions.
8108 [Steve Henson]
8109
8110 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8111 fail when they extended the size of a BIGNUM.
8112 [Steve Henson]
8113
8114 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8115 support typesafe stack.
8116 [Steve Henson]
8117
8118 *) Fix typo in SSL_[gs]et_options().
8119 [Nils Frostberg <nils@medcom.se>]
8120
8121 *) Delete various functions and files that belonged to the (now obsolete)
8122 old X509V3 handling code.
8123 [Steve Henson]
8124
8125 *) New Configure option "rsaref".
8126 [Ulf Möller]
8127
8128 *) Don't auto-generate pem.h.
8129 [Bodo Moeller]
8130
8131 *) Introduce type-safe ASN.1 SETs.
8132 [Ben Laurie]
8133
8134 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8135 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8136
8137 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8138 that links with OpenSSL (well at least cause lots of warnings), but fear
8139 not: the conversion is trivial, and it eliminates loads of evil casts. A
8140 few STACKed things have been converted already. Feel free to convert more.
8141 In the fullness of time, I'll do away with the STACK type altogether.
8142 [Ben Laurie]
8143
8144 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8145 specified in <certfile> by updating the entry in the index.txt file.
8146 This way one no longer has to edit the index.txt file manually for
8147 revoking a certificate. The -revoke option does the gory details now.
8148 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
8149
8150 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
8151 `-text' option at all and this way the `-noout -text' combination was
8152 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
8153 [Ralf S. Engelschall]
8154
8155 *) Make sure a corresponding plain text error message exists for the
8156 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
8157 verify callback function determined that a certificate was revoked.
8158 [Ralf S. Engelschall]
8159
8160 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
8161 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
8162 all available cipers including rc5, which was forgotten until now.
8163 In order to let the testing shell script know which algorithms
8164 are available, a new (up to now undocumented) command
8165 "openssl list-cipher-commands" is used.
8166 [Bodo Moeller]
8167
8168 *) Bugfix: s_client occasionally would sleep in select() when
8169 it should have checked SSL_pending() first.
8170 [Bodo Moeller]
8171
8172 *) New functions DSA_do_sign and DSA_do_verify to provide access to
8173 the raw DSA values prior to ASN.1 encoding.
8174 [Ulf Möller]
8175
8176 *) Tweaks to Configure
8177 [Niels Poppe <niels@netbox.org>]
8178
8179 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
8180 yet...
8181 [Steve Henson]
8182
8183 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
8184 [Ulf Möller]
8185
8186 *) New config option to avoid instructions that are illegal on the 80386.
8187 The default code is faster, but requires at least a 486.
8188 [Ulf Möller]
8189
8190 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
8191 SSL2_SERVER_VERSION (not used at all) macros, which are now the
8192 same as SSL2_VERSION anyway.
8193 [Bodo Moeller]
8194
8195 *) New "-showcerts" option for s_client.
8196 [Bodo Moeller]
8197
8198 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
8199 application. Various cleanups and fixes.
8200 [Steve Henson]
8201
8202 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
8203 modify error routines to work internally. Add error codes and PBE init
8204 to library startup routines.
8205 [Steve Henson]
8206
8207 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
8208 packing functions to asn1 and evp. Changed function names and error
8209 codes along the way.
8210 [Steve Henson]
8211
8212 *) PKCS12 integration: and so it begins... First of several patches to
8213 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
8214 objects to objects.h
8215 [Steve Henson]
8216
8217 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
8218 and display support for Thawte strong extranet extension.
8219 [Steve Henson]
8220
8221 *) Add LinuxPPC support.
8222 [Jeff Dubrule <igor@pobox.org>]
8223
8224 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
8225 bn_div_words in alpha.s.
8226 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
8227
8228 *) Make sure the RSA OAEP test is skipped under -DRSAref because
8229 OAEP isn't supported when OpenSSL is built with RSAref.
8230 [Ulf Moeller <ulf@fitug.de>]
8231
8232 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
8233 so they no longer are missing under -DNOPROTO.
8234 [Soren S. Jorvang <soren@t.dk>]
8235
8236
8237 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
8238
8239 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
8240 doesn't work when the session is reused. Coming soon!
8241 [Ben Laurie]
8242
8243 *) Fix a security hole, that allows sessions to be reused in the wrong
8244 context thus bypassing client cert protection! All software that uses
8245 client certs and session caches in multiple contexts NEEDS PATCHING to
8246 allow session reuse! A fuller solution is in the works.
8247 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
8248
8249 *) Some more source tree cleanups (removed obsolete files
8250 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
8251 permission on "config" script to be executable) and a fix for the INSTALL
8252 document.
8253 [Ulf Moeller <ulf@fitug.de>]
8254
8255 *) Remove some legacy and erroneous uses of malloc, free instead of
8256 Malloc, Free.
8257 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
8258
8259 *) Make rsa_oaep_test return non-zero on error.
8260 [Ulf Moeller <ulf@fitug.de>]
8261
8262 *) Add support for native Solaris shared libraries. Configure
8263 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
8264 if someone would make that last step automatic.
8265 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
8266
8267 *) ctx_size was not built with the right compiler during "make links". Fixed.
8268 [Ben Laurie]
8269
8270 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
8271 except NULL ciphers". This means the default cipher list will no longer
8272 enable NULL ciphers. They need to be specifically enabled e.g. with
8273 the string "DEFAULT:eNULL".
8274 [Steve Henson]
8275
8276 *) Fix to RSA private encryption routines: if p < q then it would
8277 occasionally produce an invalid result. This will only happen with
8278 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
8279 [Steve Henson]
8280
8281 *) Be less restrictive and allow also `perl util/perlpath.pl
8282 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
8283 because this way one can also use an interpreter named `perl5' (which is
8284 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
8285 installed as `perl').
8286 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8287
8288 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
8289 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8290
8291 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
8292 advapi32.lib to Win32 build and change the pem test comparision
8293 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
8294 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
8295 and crypto/des/ede_cbcm_enc.c.
8296 [Steve Henson]
8297
8298 *) DES quad checksum was broken on big-endian architectures. Fixed.
8299 [Ben Laurie]
8300
8301 *) Comment out two functions in bio.h that aren't implemented. Fix up the
8302 Win32 test batch file so it (might) work again. The Win32 test batch file
8303 is horrible: I feel ill....
8304 [Steve Henson]
8305
8306 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
8307 in e_os.h. Audit of header files to check ANSI and non ANSI
8308 sections: 10 functions were absent from non ANSI section and not exported
8309 from Windows DLLs. Fixed up libeay.num for new functions.
8310 [Steve Henson]
8311
8312 *) Make `openssl version' output lines consistent.
8313 [Ralf S. Engelschall]
8314
8315 *) Fix Win32 symbol export lists for BIO functions: Added
8316 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
8317 to ms/libeay{16,32}.def.
8318 [Ralf S. Engelschall]
8319
8320 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
8321 fine under Unix and passes some trivial tests I've now added. But the
8322 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
8323 added to make sure no one expects that this stuff really works in the
8324 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
8325 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
8326 openssl_bio.xs.
8327 [Ralf S. Engelschall]
8328
8329 *) Fix the generation of two part addresses in perl.
8330 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
8331
8332 *) Add config entry for Linux on MIPS.
8333 [John Tobey <jtobey@channel1.com>]
8334
8335 *) Make links whenever Configure is run, unless we are on Windoze.
8336 [Ben Laurie]
8337
8338 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
8339 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
8340 in CRLs.
8341 [Steve Henson]
8342
8343 *) Add a useful kludge to allow package maintainers to specify compiler and
8344 other platforms details on the command line without having to patch the
8345 Configure script everytime: One now can use ``perl Configure
8346 <id>:<details>'', i.e. platform ids are allowed to have details appended
8347 to them (seperated by colons). This is treated as there would be a static
8348 pre-configured entry in Configure's %table under key <id> with value
8349 <details> and ``perl Configure <id>'' is called. So, when you want to
8350 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
8351 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
8352 now, which overrides the FreeBSD-elf entry on-the-fly.
8353 [Ralf S. Engelschall]
8354
8355 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
8356 [Ben Laurie]
8357
8358 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
8359 on the `perl Configure ...' command line. This way one can compile
8360 OpenSSL libraries with Position Independent Code (PIC) which is needed
8361 for linking it into DSOs.
8362 [Ralf S. Engelschall]
8363
8364 *) Remarkably, export ciphers were totally broken and no-one had noticed!
8365 Fixed.
8366 [Ben Laurie]
8367
8368 *) Cleaned up the LICENSE document: The official contact for any license
8369 questions now is the OpenSSL core team under openssl-core@openssl.org.
8370 And add a paragraph about the dual-license situation to make sure people
8371 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
8372 to the OpenSSL toolkit.
8373 [Ralf S. Engelschall]
8374
8375 *) General source tree makefile cleanups: Made `making xxx in yyy...'
8376 display consistent in the source tree and replaced `/bin/rm' by `rm'.
8377 Additonally cleaned up the `make links' target: Remove unnecessary
8378 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
8379 to speed processing and no longer clutter the display with confusing
8380 stuff. Instead only the actually done links are displayed.
8381 [Ralf S. Engelschall]
8382
8383 *) Permit null encryption ciphersuites, used for authentication only. It used
8384 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
8385 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
8386 encryption.
8387 [Ben Laurie]
8388
8389 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
8390 signed attributes when verifying signatures (this would break them),
8391 the detached data encoding was wrong and public keys obtained using
8392 X509_get_pubkey() weren't freed.
8393 [Steve Henson]
8394
8395 *) Add text documentation for the BUFFER functions. Also added a work around
8396 to a Win95 console bug. This was triggered by the password read stuff: the
8397 last character typed gets carried over to the next fread(). If you were
8398 generating a new cert request using 'req' for example then the last
8399 character of the passphrase would be CR which would then enter the first
8400 field as blank.
8401 [Steve Henson]
8402
8403 *) Added the new `Includes OpenSSL Cryptography Software' button as
8404 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
8405 button and can be used by applications based on OpenSSL to show the
8406 relationship to the OpenSSL project.
8407 [Ralf S. Engelschall]
8408
8409 *) Remove confusing variables in function signatures in files
8410 ssl/ssl_lib.c and ssl/ssl.h.
8411 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8412
8413 *) Don't install bss_file.c under PREFIX/include/
8414 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8415
8416 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
8417 functions that return function pointers and has support for NT specific
8418 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
8419 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
8420 unsigned to signed types: this was killing the Win32 compile.
8421 [Steve Henson]
8422
8423 *) Add new certificate file to stack functions,
8424 SSL_add_dir_cert_subjects_to_stack() and
8425 SSL_add_file_cert_subjects_to_stack(). These largely supplant
8426 SSL_load_client_CA_file(), and can be used to add multiple certs easily
8427 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
8428 This means that Apache-SSL and similar packages don't have to mess around
8429 to add as many CAs as they want to the preferred list.
8430 [Ben Laurie]
8431
8432 *) Experiment with doxygen documentation. Currently only partially applied to
8433 ssl/ssl_lib.c.
8434 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
8435 openssl.doxy as the configuration file.
8436 [Ben Laurie]
8437
8438 *) Get rid of remaining C++-style comments which strict C compilers hate.
8439 [Ralf S. Engelschall, pointed out by Carlos Amengual]
8440
8441 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
8442 compiled in by default: it has problems with large keys.
8443 [Steve Henson]
8444
8445 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
8446 DH private keys and/or callback functions which directly correspond to
8447 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
8448 is needed for applications which have to configure certificates on a
8449 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
8450 (e.g. s_server).
8451 For the RSA certificate situation is makes no difference, but
8452 for the DSA certificate situation this fixes the "no shared cipher"
8453 problem where the OpenSSL cipher selection procedure failed because the
8454 temporary keys were not overtaken from the context and the API provided
8455 no way to reconfigure them.
8456 The new functions now let applications reconfigure the stuff and they
8457 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
8458 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
8459 non-public-API function ssl_cert_instantiate() is used as a helper
8460 function and also to reduce code redundancy inside ssl_rsa.c.
8461 [Ralf S. Engelschall]
8462
8463 *) Move s_server -dcert and -dkey options out of the undocumented feature
8464 area because they are useful for the DSA situation and should be
8465 recognized by the users.
8466 [Ralf S. Engelschall]
8467
8468 *) Fix the cipher decision scheme for export ciphers: the export bits are
8469 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
8470 SSL_EXP_MASK. So, the original variable has to be used instead of the
8471 already masked variable.
8472 [Richard Levitte <levitte@stacken.kth.se>]
8473
8474 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
8475 [Richard Levitte <levitte@stacken.kth.se>]
8476
8477 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
8478 from `int' to `unsigned int' because it's a length and initialized by
8479 EVP_DigestFinal() which expects an `unsigned int *'.
8480 [Richard Levitte <levitte@stacken.kth.se>]
8481
8482 *) Don't hard-code path to Perl interpreter on shebang line of Configure
8483 script. Instead use the usual Shell->Perl transition trick.
8484 [Ralf S. Engelschall]
8485
8486 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
8487 (in addition to RSA certificates) to match the behaviour of `openssl dsa
8488 -noout -modulus' as it's already the case for `openssl rsa -noout
8489 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
8490 currently the public key is printed (a decision which was already done by
8491 `openssl dsa -modulus' in the past) which serves a similar purpose.
8492 Additionally the NO_RSA no longer completely removes the whole -modulus
8493 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
8494 now, too.
8495 [Ralf S. Engelschall]
8496
8497 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
8498 BIO. See the source (crypto/evp/bio_ok.c) for more info.
8499 [Arne Ansper <arne@ats.cyber.ee>]
8500
8501 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
8502 to be added. Now both 'req' and 'ca' can use new objects defined in the
8503 config file.
8504 [Steve Henson]
8505
8506 *) Add cool BIO that does syslog (or event log on NT).
8507 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
8508
8509 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
8510 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
8511 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
8512 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
8513 [Ben Laurie]
8514
8515 *) Add preliminary config info for new extension code.
8516 [Steve Henson]
8517
8518 *) Make RSA_NO_PADDING really use no padding.
8519 [Ulf Moeller <ulf@fitug.de>]
8520
8521 *) Generate errors when private/public key check is done.
8522 [Ben Laurie]
8523
8524 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
8525 for some CRL extensions and new objects added.
8526 [Steve Henson]
8527
8528 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
8529 key usage extension and fuller support for authority key id.
8530 [Steve Henson]
8531
8532 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
8533 padding method for RSA, which is recommended for new applications in PKCS
8534 #1 v2.0 (RFC 2437, October 1998).
8535 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
8536 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
8537 against Bleichbacher's attack on RSA.
8538 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
8539 Ben Laurie]
8540
8541 *) Updates to the new SSL compression code
8542 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8543
8544 *) Fix so that the version number in the master secret, when passed
8545 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
8546 (because the server will not accept higher), that the version number
8547 is 0x03,0x01, not 0x03,0x00
8548 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8549
8550 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
8551 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
8552 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
8553 [Steve Henson]
8554
8555 *) Support for RAW extensions where an arbitrary extension can be
8556 created by including its DER encoding. See apps/openssl.cnf for
8557 an example.
8558 [Steve Henson]
8559
8560 *) Make sure latest Perl versions don't interpret some generated C array
8561 code as Perl array code in the crypto/err/err_genc.pl script.
8562 [Lars Weber <3weber@informatik.uni-hamburg.de>]
8563
8564 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
8565 not many people have the assembler. Various Win32 compilation fixes and
8566 update to the INSTALL.W32 file with (hopefully) more accurate Win32
8567 build instructions.
8568 [Steve Henson]
8569
8570 *) Modify configure script 'Configure' to automatically create crypto/date.h
8571 file under Win32 and also build pem.h from pem.org. New script
8572 util/mkfiles.pl to create the MINFO file on environments that can't do a
8573 'make files': perl util/mkfiles.pl >MINFO should work.
8574 [Steve Henson]
8575
8576 *) Major rework of DES function declarations, in the pursuit of correctness
8577 and purity. As a result, many evil casts evaporated, and some weirdness,
8578 too. You may find this causes warnings in your code. Zapping your evil
8579 casts will probably fix them. Mostly.
8580 [Ben Laurie]
8581
8582 *) Fix for a typo in asn1.h. Bug fix to object creation script
8583 obj_dat.pl. It considered a zero in an object definition to mean
8584 "end of object": none of the objects in objects.h have any zeros
8585 so it wasn't spotted.
8586 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
8587
8588 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
8589 Masking (CBCM). In the absence of test vectors, the best I have been able
8590 to do is check that the decrypt undoes the encrypt, so far. Send me test
8591 vectors if you have them.
8592 [Ben Laurie]
8593
8594 *) Correct calculation of key length for export ciphers (too much space was
8595 allocated for null ciphers). This has not been tested!
8596 [Ben Laurie]
8597
8598 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
8599 message is now correct (it understands "crypto" and "ssl" on its
8600 command line). There is also now an "update" option. This will update
8601 the util/ssleay.num and util/libeay.num files with any new functions.
8602 If you do a:
8603 perl util/mkdef.pl crypto ssl update
8604 it will update them.
8605 [Steve Henson]
8606
8607 *) Overhauled the Perl interface (perl/*):
8608 - ported BN stuff to OpenSSL's different BN library
8609 - made the perl/ source tree CVS-aware
8610 - renamed the package from SSLeay to OpenSSL (the files still contain
8611 their history because I've copied them in the repository)
8612 - removed obsolete files (the test scripts will be replaced
8613 by better Test::Harness variants in the future)
8614 [Ralf S. Engelschall]
8615
8616 *) First cut for a very conservative source tree cleanup:
8617 1. merge various obsolete readme texts into doc/ssleay.txt
8618 where we collect the old documents and readme texts.
8619 2. remove the first part of files where I'm already sure that we no
8620 longer need them because of three reasons: either they are just temporary
8621 files which were left by Eric or they are preserved original files where
8622 I've verified that the diff is also available in the CVS via "cvs diff
8623 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8624 the crypto/md/ stuff).
8625 [Ralf S. Engelschall]
8626
8627 *) More extension code. Incomplete support for subject and issuer alt
8628 name, issuer and authority key id. Change the i2v function parameters
8629 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8630 what that's for :-) Fix to ASN1 macro which messed up
8631 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8632 [Steve Henson]
8633
8634 *) Preliminary support for ENUMERATED type. This is largely copied from the
8635 INTEGER code.
8636 [Steve Henson]
8637
8638 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8639 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8640
8641 *) Make sure `make rehash' target really finds the `openssl' program.
8642 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8643
8644 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8645 like to hear about it if this slows down other processors.
8646 [Ben Laurie]
8647
8648 *) Add CygWin32 platform information to Configure script.
8649 [Alan Batie <batie@aahz.jf.intel.com>]
8650
8651 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8652 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8653
8654 *) New program nseq to manipulate netscape certificate sequences
8655 [Steve Henson]
8656
8657 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8658 few typos.
8659 [Steve Henson]
8660
8661 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8662 but the BN code had some problems that would cause failures when
8663 doing certificate verification and some other functions.
8664 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8665
8666 *) Add ASN1 and PEM code to support netscape certificate sequences.
8667 [Steve Henson]
8668
8669 *) Add ASN1 and PEM code to support netscape certificate sequences.
8670 [Steve Henson]
8671
8672 *) Add several PKIX and private extended key usage OIDs.
8673 [Steve Henson]
8674
8675 *) Modify the 'ca' program to handle the new extension code. Modify
8676 openssl.cnf for new extension format, add comments.
8677 [Steve Henson]
8678
8679 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8680 and add a sample to openssl.cnf so req -x509 now adds appropriate
8681 CA extensions.
8682 [Steve Henson]
8683
8684 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8685 error code, add initial support to X509_print() and x509 application.
8686 [Steve Henson]
8687
8688 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8689 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8690 stuff is currently isolated and isn't even compiled yet.
8691 [Steve Henson]
8692
8693 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8694 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8695 Removed the versions check from X509 routines when loading extensions:
8696 this allows certain broken certificates that don't set the version
8697 properly to be processed.
8698 [Steve Henson]
8699
8700 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8701 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8702 can still be regenerated with "make depend".
8703 [Ben Laurie]
8704
8705 *) Spelling mistake in C version of CAST-128.
8706 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8707
8708 *) Changes to the error generation code. The perl script err-code.pl
8709 now reads in the old error codes and retains the old numbers, only
8710 adding new ones if necessary. It also only changes the .err files if new
8711 codes are added. The makefiles have been modified to only insert errors
8712 when needed (to avoid needlessly modifying header files). This is done
8713 by only inserting errors if the .err file is newer than the auto generated
8714 C file. To rebuild all the error codes from scratch (the old behaviour)
8715 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8716 or delete all the .err files.
8717 [Steve Henson]
8718
8719 *) CAST-128 was incorrectly implemented for short keys. The C version has
8720 been fixed, but is untested. The assembler versions are also fixed, but
8721 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8722 to regenerate it if needed.
8723 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8724 Hagino <itojun@kame.net>]
8725
8726 *) File was opened incorrectly in randfile.c.
8727 [Ulf Möller <ulf@fitug.de>]
8728
8729 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8730 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8731 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8732 al: it's just almost always a UTCTime. Note this patch adds new error
8733 codes so do a "make errors" if there are problems.
8734 [Steve Henson]
8735
8736 *) Correct Linux 1 recognition in config.
8737 [Ulf Möller <ulf@fitug.de>]
8738
8739 *) Remove pointless MD5 hash when using DSA keys in ca.
8740 [Anonymous <nobody@replay.com>]
8741
8742 *) Generate an error if given an empty string as a cert directory. Also
8743 generate an error if handed NULL (previously returned 0 to indicate an
8744 error, but didn't set one).
8745 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8746
8747 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8748 [Ben Laurie]
8749
8750 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8751 parameters. This was causing a warning which killed off the Win32 compile.
8752 [Steve Henson]
8753
8754 *) Remove C++ style comments from crypto/bn/bn_local.h.
8755 [Neil Costigan <neil.costigan@celocom.com>]
8756
8757 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8758 based on a text string, looking up short and long names and finally
8759 "dot" format. The "dot" format stuff didn't work. Added new function
8760 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8761 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
8762 OID is not part of the table.
8763 [Steve Henson]
8764
8765 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
8766 X509_LOOKUP_by_alias().
8767 [Ben Laurie]
8768
8769 *) Sort openssl functions by name.
8770 [Ben Laurie]
8771
8772 *) Get the gendsa program working (hopefully) and add it to app list. Remove
8773 encryption from sample DSA keys (in case anyone is interested the password
8774 was "1234").
8775 [Steve Henson]
8776
8777 *) Make _all_ *_free functions accept a NULL pointer.
8778 [Frans Heymans <fheymans@isaserver.be>]
8779
8780 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
8781 NULL pointers.
8782 [Anonymous <nobody@replay.com>]
8783
8784 *) s_server should send the CAfile as acceptable CAs, not its own cert.
8785 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8786
8787 *) Don't blow it for numeric -newkey arguments to apps/req.
8788 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8789
8790 *) Temp key "for export" tests were wrong in s3_srvr.c.
8791 [Anonymous <nobody@replay.com>]
8792
8793 *) Add prototype for temp key callback functions
8794 SSL_CTX_set_tmp_{rsa,dh}_callback().
8795 [Ben Laurie]
8796
8797 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8798 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8799 [Steve Henson]
8800
8801 *) X509_name_add_entry() freed the wrong thing after an error.
8802 [Arne Ansper <arne@ats.cyber.ee>]
8803
8804 *) rsa_eay.c would attempt to free a NULL context.
8805 [Arne Ansper <arne@ats.cyber.ee>]
8806
8807 *) BIO_s_socket() had a broken should_retry() on Windoze.
8808 [Arne Ansper <arne@ats.cyber.ee>]
8809
8810 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8811 [Arne Ansper <arne@ats.cyber.ee>]
8812
8813 *) Make sure the already existing X509_STORE->depth variable is initialized
8814 in X509_STORE_new(), but document the fact that this variable is still
8815 unused in the certificate verification process.
8816 [Ralf S. Engelschall]
8817
8818 *) Fix the various library and apps files to free up pkeys obtained from
8819 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8820 [Steve Henson]
8821
8822 *) Fix reference counting in X509_PUBKEY_get(). This makes
8823 demos/maurice/example2.c work, amongst others, probably.
8824 [Steve Henson and Ben Laurie]
8825
8826 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8827 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8828 are no longer created. This way we have a single and consistent command
8829 line interface `openssl <command>', similar to `cvs <command>'.
8830 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8831
8832 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8833 BIT STRING wrapper always have zero unused bits.
8834 [Steve Henson]
8835
8836 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8837 [Steve Henson]
8838
8839 *) Make the top-level INSTALL documentation easier to understand.
8840 [Paul Sutton]
8841
8842 *) Makefiles updated to exit if an error occurs in a sub-directory
8843 make (including if user presses ^C) [Paul Sutton]
8844
8845 *) Make Montgomery context stuff explicit in RSA data structure.
8846 [Ben Laurie]
8847
8848 *) Fix build order of pem and err to allow for generated pem.h.
8849 [Ben Laurie]
8850
8851 *) Fix renumbering bug in X509_NAME_delete_entry().
8852 [Ben Laurie]
8853
8854 *) Enhanced the err-ins.pl script so it makes the error library number
8855 global and can add a library name. This is needed for external ASN1 and
8856 other error libraries.
8857 [Steve Henson]
8858
8859 *) Fixed sk_insert which never worked properly.
8860 [Steve Henson]
8861
8862 *) Fix ASN1 macros so they can handle indefinite length construted
8863 EXPLICIT tags. Some non standard certificates use these: they can now
8864 be read in.
8865 [Steve Henson]
8866
8867 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8868 into a single doc/ssleay.txt bundle. This way the information is still
8869 preserved but no longer messes up this directory. Now it's new room for
8870 the new set of documenation files.
8871 [Ralf S. Engelschall]
8872
8873 *) SETs were incorrectly DER encoded. This was a major pain, because they
8874 shared code with SEQUENCEs, which aren't coded the same. This means that
8875 almost everything to do with SETs or SEQUENCEs has either changed name or
8876 number of arguments.
8877 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8878
8879 *) Fix test data to work with the above.
8880 [Ben Laurie]
8881
8882 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8883 was already fixed by Eric for 0.9.1 it seems.
8884 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8885
8886 *) Autodetect FreeBSD3.
8887 [Ben Laurie]
8888
8889 *) Fix various bugs in Configure. This affects the following platforms:
8890 nextstep
8891 ncr-scde
8892 unixware-2.0
8893 unixware-2.0-pentium
8894 sco5-cc.
8895 [Ben Laurie]
8896
8897 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8898 before they are needed.
8899 [Ben Laurie]
8900
8901 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8902 [Ben Laurie]
8903
8904
8905 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8906
8907 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8908 changed SSLeay to OpenSSL in version strings.
8909 [Ralf S. Engelschall]
8910
8911 *) Some fixups to the top-level documents.
8912 [Paul Sutton]
8913
8914 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8915 because the symlink to include/ was missing.
8916 [Ralf S. Engelschall]
8917
8918 *) Incorporated the popular no-RSA/DSA-only patches
8919 which allow to compile a RSA-free SSLeay.
8920 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8921
8922 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8923 when "ssleay" is still not found.
8924 [Ralf S. Engelschall]
8925
8926 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8927 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8928
8929 *) Updated the README file.
8930 [Ralf S. Engelschall]
8931
8932 *) Added various .cvsignore files in the CVS repository subdirs
8933 to make a "cvs update" really silent.
8934 [Ralf S. Engelschall]
8935
8936 *) Recompiled the error-definition header files and added
8937 missing symbols to the Win32 linker tables.
8938 [Ralf S. Engelschall]
8939
8940 *) Cleaned up the top-level documents;
8941 o new files: CHANGES and LICENSE
8942 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8943 o merged COPYRIGHT into LICENSE
8944 o removed obsolete TODO file
8945 o renamed MICROSOFT to INSTALL.W32
8946 [Ralf S. Engelschall]
8947
8948 *) Removed dummy files from the 0.9.1b source tree:
8949 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8950 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8951 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8952 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8953 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8954 [Ralf S. Engelschall]
8955
8956 *) Added various platform portability fixes.
8957 [Mark J. Cox]
8958
8959 *) The Genesis of the OpenSSL rpject:
8960 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8961 Young and Tim J. Hudson created while they were working for C2Net until
8962 summer 1998.
8963 [The OpenSSL Project]
8964
8965
8966 Changes between 0.9.0b and 0.9.1b [not released]
8967
8968 *) Updated a few CA certificates under certs/
8969 [Eric A. Young]
8970
8971 *) Changed some BIGNUM api stuff.
8972 [Eric A. Young]
8973
8974 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8975 DGUX x86, Linux Alpha, etc.
8976 [Eric A. Young]
8977
8978 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8979 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8980 available).
8981 [Eric A. Young]
8982
8983 *) Add -strparse option to asn1pars program which parses nested
8984 binary structures
8985 [Dr Stephen Henson <shenson@bigfoot.com>]
8986
8987 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8988 [Eric A. Young]
8989
8990 *) DSA fix for "ca" program.
8991 [Eric A. Young]
8992
8993 *) Added "-genkey" option to "dsaparam" program.
8994 [Eric A. Young]
8995
8996 *) Added RIPE MD160 (rmd160) message digest.
8997 [Eric A. Young]
8998
8999 *) Added -a (all) option to "ssleay version" command.
9000 [Eric A. Young]
9001
9002 *) Added PLATFORM define which is the id given to Configure.
9003 [Eric A. Young]
9004
9005 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9006 [Eric A. Young]
9007
9008 *) Extended the ASN.1 parser routines.
9009 [Eric A. Young]
9010
9011 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9012 [Eric A. Young]
9013
9014 *) Added a BN_CTX to the BN library.
9015 [Eric A. Young]
9016
9017 *) Fixed the weak key values in DES library
9018 [Eric A. Young]
9019
9020 *) Changed API in EVP library for cipher aliases.
9021 [Eric A. Young]
9022
9023 *) Added support for RC2/64bit cipher.
9024 [Eric A. Young]
9025
9026 *) Converted the lhash library to the crypto/mem.c functions.
9027 [Eric A. Young]
9028
9029 *) Added more recognized ASN.1 object ids.
9030 [Eric A. Young]
9031
9032 *) Added more RSA padding checks for SSL/TLS.
9033 [Eric A. Young]
9034
9035 *) Added BIO proxy/filter functionality.
9036 [Eric A. Young]
9037
9038 *) Added extra_certs to SSL_CTX which can be used
9039 send extra CA certificates to the client in the CA cert chain sending
9040 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9041 [Eric A. Young]
9042
9043 *) Now Fortezza is denied in the authentication phase because
9044 this is key exchange mechanism is not supported by SSLeay at all.
9045 [Eric A. Young]
9046
9047 *) Additional PKCS1 checks.
9048 [Eric A. Young]
9049
9050 *) Support the string "TLSv1" for all TLS v1 ciphers.
9051 [Eric A. Young]
9052
9053 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9054 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9055 [Eric A. Young]
9056
9057 *) Fixed a few memory leaks.
9058 [Eric A. Young]
9059
9060 *) Fixed various code and comment typos.
9061 [Eric A. Young]
9062
9063 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9064 bytes sent in the client random.
9065 [Edward Bishop <ebishop@spyglass.com>]
9066