]> git.ipfire.org Git - thirdparty/openssl.git/blob - CHANGES
Typo
[thirdparty/openssl.git] / CHANGES
1
2 OpenSSL CHANGES
3 _______________
4
5 Changes between 1.0.0 and 1.1.0 [xx XXX xxxx]
6
7 *) Add an "external" session cache for debugging purposes to s_server. This
8 should help trace issues which normally are only apparent in deployed
9 multi-process servers.
10 [Steve Henson]
11
12 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
13 a few changes are required:
14
15 Add SSL_OP_NO_TLSv1_1 flag.
16 Add TLSv1_1 methods.
17 Update version checking logic to handle version 1.1.
18 Add explicit IV handling (ported from DTLS code).
19 Add command line options to s_client/s_server.
20 [Steve Henson]
21
22 *) Experiemental password based recipient info support for CMS library:
23 implementing RFC3211.
24 [Steve Henson]
25
26 *) Split password based encryption into PBES2 and PBKDF2 functions. This
27 neatly separates the code into cipher and PBE sections and is required
28 for some algorithms that split PBES2 into separate pieces (such as
29 password based CMS).
30 [Steve Henson]
31
32 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
33 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
34 BIO_set_cipher() and some obscure PEM functions were changed so they
35 can now return an error. The RAND changes required a change to the
36 RAND_METHOD structure.
37 [Steve Henson]
38
39 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
40 a gcc attribute to warn if the result of a function is ignored. This
41 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
42 whose return value is often ignored.
43 [Steve Henson]
44
45 Changes between 0.9.8m (?) and 1.0.0 [xx XXX xxxx]
46
47 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
48 [Steve Henson]
49
50 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
51 output hashes compatible with older versions of OpenSSL.
52 [Willy Weisz <weisz@vcpc.univie.ac.at>]
53
54 *) Fix compression algorithm handling: if resuming a session use the
55 compression algorithm of the resumed session instead of determining
56 it from client hello again. Don't allow server to change algorithm.
57 [Steve Henson]
58
59 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
60 change when encrypting or decrypting.
61 [Bodo Moeller]
62
63 *) Add load_crls() function to apps tidying load_certs() too. Add option
64 to verify utility to allow additional CRLs to be included.
65 [Steve Henson]
66
67 *) Update OCSP request code to permit adding custom headers to the request:
68 some responders need this.
69 [Steve Henson]
70
71 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
72 correctly.
73 [Julia Lawall <julia@diku.dk>]
74
75 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
76 needlessly dereferenced structures, used obsolete functions and
77 didn't handle all updated verify codes correctly.
78 [Steve Henson]
79
80 *) Disable MD2 in the default configuration.
81 [Steve Henson]
82
83 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
84 indicate the initial BIO being pushed or popped. This makes it possible
85 to determine whether the BIO is the one explicitly called or as a result
86 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
87 it handles reference counts correctly and doesn't zero out the I/O bio
88 when it is not being explicitly popped. WARNING: applications which
89 included workarounds for the old buggy behaviour will need to be modified
90 or they could free up already freed BIOs.
91 [Steve Henson]
92
93 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
94 renaming to all platforms (within the 0.9.8 branch, this was
95 done conditionally on Netware platforms to avoid a name clash).
96 [Guenter <lists@gknw.net>]
97
98 *) Add ECDHE and PSK support to DTLS.
99 [Michael Tuexen <tuexen@fh-muenster.de>]
100
101 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
102 be used on C++.
103 [Steve Henson]
104
105 *) Add "missing" function EVP_MD_flags() (without this the only way to
106 retrieve a digest flags is by accessing the structure directly. Update
107 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
108 or cipher is registered as in the "from" argument. Print out all
109 registered digests in the dgst usage message instead of manually
110 attempting to work them out.
111 [Steve Henson]
112
113 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
114 this allows the use of compression and extensions. Change default cipher
115 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
116 by default unless an application cipher string requests it.
117 [Steve Henson]
118
119 *) Alter match criteria in PKCS12_parse(). It used to try to use local
120 key ids to find matching certificates and keys but some PKCS#12 files
121 don't follow the (somewhat unwritten) rules and this strategy fails.
122 Now just gather all certificates together and the first private key
123 then look for the first certificate that matches the key.
124 [Steve Henson]
125
126 *) Support use of registered digest and cipher names for dgst and cipher
127 commands instead of having to add each one as a special case. So now
128 you can do:
129
130 openssl sha256 foo
131
132 as well as:
133
134 openssl dgst -sha256 foo
135
136 and this works for ENGINE based algorithms too.
137
138 [Steve Henson]
139
140 *) Update Gost ENGINE to support parameter files.
141 [Victor B. Wagner <vitus@cryptocom.ru>]
142
143 *) Support GeneralizedTime in ca utility.
144 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
145
146 *) Enhance the hash format used for certificate directory links. The new
147 form uses the canonical encoding (meaning equivalent names will work
148 even if they aren't identical) and uses SHA1 instead of MD5. This form
149 is incompatible with the older format and as a result c_rehash should
150 be used to rebuild symbolic links.
151 [Steve Henson]
152
153 *) Make PKCS#8 the default write format for private keys, replacing the
154 traditional format. This form is standardised, more secure and doesn't
155 include an implicit MD5 dependency.
156 [Steve Henson]
157
158 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
159 committed to OpenSSL should pass this lot as a minimum.
160 [Steve Henson]
161
162 *) Add session ticket override functionality for use by EAP-FAST.
163 [Jouni Malinen <j@w1.fi>]
164
165 *) Modify HMAC functions to return a value. Since these can be implemented
166 in an ENGINE errors can occur.
167 [Steve Henson]
168
169 *) Type-checked OBJ_bsearch_ex.
170 [Ben Laurie]
171
172 *) Type-checked OBJ_bsearch. Also some constification necessitated
173 by type-checking. Still to come: TXT_DB, bsearch(?),
174 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
175 CONF_VALUE.
176 [Ben Laurie]
177
178 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
179 seconds to a tm structure directly, instead of going through OS
180 specific date routines. This avoids any issues with OS routines such
181 as the year 2038 bug. New *_adj() functions for ASN1 time structures
182 and X509_time_adj_ex() to cover the extended range. The existing
183 X509_time_adj() is still usable and will no longer have any date issues.
184 [Steve Henson]
185
186 *) Delta CRL support. New use deltas option which will attempt to locate
187 and search any appropriate delta CRLs available.
188
189 This work was sponsored by Google.
190 [Steve Henson]
191
192 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
193 code and add additional score elements. Validate alternate CRL paths
194 as part of the CRL checking and indicate a new error "CRL path validation
195 error" in this case. Applications wanting additional details can use
196 the verify callback and check the new "parent" field. If this is not
197 NULL CRL path validation is taking place. Existing applications wont
198 see this because it requires extended CRL support which is off by
199 default.
200
201 This work was sponsored by Google.
202 [Steve Henson]
203
204 *) Support for freshest CRL extension.
205
206 This work was sponsored by Google.
207 [Steve Henson]
208
209 *) Initial indirect CRL support. Currently only supported in the CRLs
210 passed directly and not via lookup. Process certificate issuer
211 CRL entry extension and lookup CRL entries by bother issuer name
212 and serial number. Check and process CRL issuer entry in IDP extension.
213
214 This work was sponsored by Google.
215 [Steve Henson]
216
217 *) Add support for distinct certificate and CRL paths. The CRL issuer
218 certificate is validated separately in this case. Only enabled if
219 an extended CRL support flag is set: this flag will enable additional
220 CRL functionality in future.
221
222 This work was sponsored by Google.
223 [Steve Henson]
224
225 *) Add support for policy mappings extension.
226
227 This work was sponsored by Google.
228 [Steve Henson]
229
230 *) Fixes to pathlength constraint, self issued certificate handling,
231 policy processing to align with RFC3280 and PKITS tests.
232
233 This work was sponsored by Google.
234 [Steve Henson]
235
236 *) Support for name constraints certificate extension. DN, email, DNS
237 and URI types are currently supported.
238
239 This work was sponsored by Google.
240 [Steve Henson]
241
242 *) To cater for systems that provide a pointer-based thread ID rather
243 than numeric, deprecate the current numeric thread ID mechanism and
244 replace it with a structure and associated callback type. This
245 mechanism allows a numeric "hash" to be extracted from a thread ID in
246 either case, and on platforms where pointers are larger than 'long',
247 mixing is done to help ensure the numeric 'hash' is usable even if it
248 can't be guaranteed unique. The default mechanism is to use "&errno"
249 as a pointer-based thread ID to distinguish between threads.
250
251 Applications that want to provide their own thread IDs should now use
252 CRYPTO_THREADID_set_callback() to register a callback that will call
253 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
254
255 Note that ERR_remove_state() is now deprecated, because it is tied
256 to the assumption that thread IDs are numeric. ERR_remove_state(0)
257 to free the current thread's error state should be replaced by
258 ERR_remove_thread_state(NULL).
259
260 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
261 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
262 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
263 application was previously providing a numeric thread callback that
264 was inappropriate for distinguishing threads, then uniqueness might
265 have been obtained with &errno that happened immediately in the
266 intermediate development versions of OpenSSL; this is no longer the
267 case, the numeric thread callback will now override the automatic use
268 of &errno.)
269 [Geoff Thorpe, with help from Bodo Moeller]
270
271 *) Initial support for different CRL issuing certificates. This covers a
272 simple case where the self issued certificates in the chain exist and
273 the real CRL issuer is higher in the existing chain.
274
275 This work was sponsored by Google.
276 [Steve Henson]
277
278 *) Removed effectively defunct crypto/store from the build.
279 [Ben Laurie]
280
281 *) Revamp of STACK to provide stronger type-checking. Still to come:
282 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
283 ASN1_STRING, CONF_VALUE.
284 [Ben Laurie]
285
286 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
287 RAM on SSL connections. This option can save about 34k per idle SSL.
288 [Nick Mathewson]
289
290 *) Revamp of LHASH to provide stronger type-checking. Still to come:
291 STACK, TXT_DB, bsearch, qsort.
292 [Ben Laurie]
293
294 *) Initial support for Cryptographic Message Syntax (aka CMS) based
295 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
296 support for data, signedData, compressedData, digestedData and
297 encryptedData, envelopedData types included. Scripts to check against
298 RFC4134 examples draft and interop and consistency checks of many
299 content types and variants.
300 [Steve Henson]
301
302 *) Add options to enc utility to support use of zlib compression BIO.
303 [Steve Henson]
304
305 *) Extend mk1mf to support importing of options and assembly language
306 files from Configure script, currently only included in VC-WIN32.
307 The assembly language rules can now optionally generate the source
308 files from the associated perl scripts.
309 [Steve Henson]
310
311 *) Implement remaining functionality needed to support GOST ciphersuites.
312 Interop testing has been performed using CryptoPro implementations.
313 [Victor B. Wagner <vitus@cryptocom.ru>]
314
315 *) s390x assembler pack.
316 [Andy Polyakov]
317
318 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
319 "family."
320 [Andy Polyakov]
321
322 *) Implement Opaque PRF Input TLS extension as specified in
323 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
324 official specification yet and no extension type assignment by
325 IANA exists, this extension (for now) will have to be explicitly
326 enabled when building OpenSSL by providing the extension number
327 to use. For example, specify an option
328
329 -DTLSEXT_TYPE_opaque_prf_input=0x9527
330
331 to the "config" or "Configure" script to enable the extension,
332 assuming extension number 0x9527 (which is a completely arbitrary
333 and unofficial assignment based on the MD5 hash of the Internet
334 Draft). Note that by doing so, you potentially lose
335 interoperability with other TLS implementations since these might
336 be using the same extension number for other purposes.
337
338 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
339 opaque PRF input value to use in the handshake. This will create
340 an interal copy of the length-'len' string at 'src', and will
341 return non-zero for success.
342
343 To get more control and flexibility, provide a callback function
344 by using
345
346 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
347 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
348
349 where
350
351 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
352 void *arg;
353
354 Callback function 'cb' will be called in handshakes, and is
355 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
356 Argument 'arg' is for application purposes (the value as given to
357 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
358 be provided to the callback function). The callback function
359 has to return non-zero to report success: usually 1 to use opaque
360 PRF input just if possible, or 2 to enforce use of the opaque PRF
361 input. In the latter case, the library will abort the handshake
362 if opaque PRF input is not successfully negotiated.
363
364 Arguments 'peerinput' and 'len' given to the callback function
365 will always be NULL and 0 in the case of a client. A server will
366 see the client's opaque PRF input through these variables if
367 available (NULL and 0 otherwise). Note that if the server
368 provides an opaque PRF input, the length must be the same as the
369 length of the client's opaque PRF input.
370
371 Note that the callback function will only be called when creating
372 a new session (session resumption can resume whatever was
373 previously negotiated), and will not be called in SSL 2.0
374 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
375 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
376 for applications that need to enforce opaque PRF input.
377
378 [Bodo Moeller]
379
380 *) Update ssl code to support digests other than SHA1+MD5 for handshake
381 MAC.
382
383 [Victor B. Wagner <vitus@cryptocom.ru>]
384
385 *) Add RFC4507 support to OpenSSL. This includes the corrections in
386 RFC4507bis. The encrypted ticket format is an encrypted encoded
387 SSL_SESSION structure, that way new session features are automatically
388 supported.
389
390 If a client application caches session in an SSL_SESSION structure
391 support is transparent because tickets are now stored in the encoded
392 SSL_SESSION.
393
394 The SSL_CTX structure automatically generates keys for ticket
395 protection in servers so again support should be possible
396 with no application modification.
397
398 If a client or server wishes to disable RFC4507 support then the option
399 SSL_OP_NO_TICKET can be set.
400
401 Add a TLS extension debugging callback to allow the contents of any client
402 or server extensions to be examined.
403
404 This work was sponsored by Google.
405 [Steve Henson]
406
407 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
408 OpenSSL should now compile cleanly on gcc 4.2
409 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
410
411 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
412 support including streaming MAC support: this is required for GOST
413 ciphersuite support.
414 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
415
416 *) Add option -stream to use PKCS#7 streaming in smime utility. New
417 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
418 to output in BER and PEM format.
419 [Steve Henson]
420
421 *) Experimental support for use of HMAC via EVP_PKEY interface. This
422 allows HMAC to be handled via the EVP_DigestSign*() interface. The
423 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
424 ENGINE support for HMAC keys which are unextractable. New -mac and
425 -macopt options to dgst utility.
426 [Steve Henson]
427
428 *) New option -sigopt to dgst utility. Update dgst to use
429 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
430 alternative signing paramaters such as X9.31 or PSS in the dgst
431 utility.
432 [Steve Henson]
433
434 *) Change ssl_cipher_apply_rule(), the internal function that does
435 the work each time a ciphersuite string requests enabling
436 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
437 removing ("!foo+bar") a class of ciphersuites: Now it maintains
438 the order of disabled ciphersuites such that those ciphersuites
439 that most recently went from enabled to disabled not only stay
440 in order with respect to each other, but also have higher priority
441 than other disabled ciphersuites the next time ciphersuites are
442 enabled again.
443
444 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
445 the same ciphersuites as with "HIGH" alone, but in a specific
446 order where the PSK ciphersuites come first (since they are the
447 most recently disabled ciphersuites when "HIGH" is parsed).
448
449 Also, change ssl_create_cipher_list() (using this new
450 funcionality) such that between otherwise identical
451 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
452 the default order.
453 [Bodo Moeller]
454
455 *) Change ssl_create_cipher_list() so that it automatically
456 arranges the ciphersuites in reasonable order before starting
457 to process the rule string. Thus, the definition for "DEFAULT"
458 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
459 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
460 This makes it much easier to arrive at a reasonable default order
461 in applications for which anonymous ciphers are OK (meaning
462 that you can't actually use DEFAULT).
463 [Bodo Moeller; suggested by Victor Duchovni]
464
465 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
466 processing) into multiple integers instead of setting
467 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
468 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
469 (These masks as well as the individual bit definitions are hidden
470 away into the non-exported interface ssl/ssl_locl.h, so this
471 change to the definition of the SSL_CIPHER structure shouldn't
472 affect applications.) This give us more bits for each of these
473 categories, so there is no longer a need to coagulate AES128 and
474 AES256 into a single algorithm bit, and to coagulate Camellia128
475 and Camellia256 into a single algorithm bit, which has led to all
476 kinds of kludges.
477
478 Thus, among other things, the kludge introduced in 0.9.7m and
479 0.9.8e for masking out AES256 independently of AES128 or masking
480 out Camellia256 independently of AES256 is not needed here in 0.9.9.
481
482 With the change, we also introduce new ciphersuite aliases that
483 so far were missing: "AES128", "AES256", "CAMELLIA128", and
484 "CAMELLIA256".
485 [Bodo Moeller]
486
487 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
488 Use the leftmost N bytes of the signature input if the input is
489 larger than the prime q (with N being the size in bytes of q).
490 [Nils Larsch]
491
492 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
493 it yet and it is largely untested.
494 [Steve Henson]
495
496 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
497 [Nils Larsch]
498
499 *) Initial incomplete changes to avoid need for function casts in OpenSSL
500 some compilers (gcc 4.2 and later) reject their use. Safestack is
501 reimplemented. Update ASN1 to avoid use of legacy functions.
502 [Steve Henson]
503
504 *) Win32/64 targets are linked with Winsock2.
505 [Andy Polyakov]
506
507 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
508 to external functions. This can be used to increase CRL handling
509 efficiency especially when CRLs are very large by (for example) storing
510 the CRL revoked certificates in a database.
511 [Steve Henson]
512
513 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
514 new CRLs added to a directory can be used. New command line option
515 -verify_return_error to s_client and s_server. This causes real errors
516 to be returned by the verify callback instead of carrying on no matter
517 what. This reflects the way a "real world" verify callback would behave.
518 [Steve Henson]
519
520 *) GOST engine, supporting several GOST algorithms and public key formats.
521 Kindly donated by Cryptocom.
522 [Cryptocom]
523
524 *) Partial support for Issuing Distribution Point CRL extension. CRLs
525 partitioned by DP are handled but no indirect CRL or reason partitioning
526 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
527 selected via a scoring technique which handles IDP and AKID in CRLs.
528 [Steve Henson]
529
530 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
531 will ultimately be used for all verify operations: this will remove the
532 X509_STORE dependency on certificate verification and allow alternative
533 lookup methods. X509_STORE based implementations of these two callbacks.
534 [Steve Henson]
535
536 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
537 Modify get_crl() to find a valid (unexpired) CRL if possible.
538 [Steve Henson]
539
540 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
541 this would be called X509_CRL_cmp() but that name is already used by
542 a function that just compares CRL issuer names. Cache several CRL
543 extensions in X509_CRL structure and cache CRLDP in X509.
544 [Steve Henson]
545
546 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
547 this maps equivalent X509_NAME structures into a consistent structure.
548 Name comparison can then be performed rapidly using memcmp().
549 [Steve Henson]
550
551 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
552 utility.
553 [Steve Henson]
554
555 *) Allow digests to supply their own micalg string for S/MIME type using
556 the ctrl EVP_MD_CTRL_MICALG.
557 [Steve Henson]
558
559 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
560 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
561 ctrl. It can then customise the structure before and/or after signing
562 if necessary.
563 [Steve Henson]
564
565 *) New function OBJ_add_sigid() to allow application defined signature OIDs
566 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
567 to free up any added signature OIDs.
568 [Steve Henson]
569
570 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
571 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
572 digest and cipher tables. New options added to openssl utility:
573 list-message-digest-algorithms and list-cipher-algorithms.
574 [Steve Henson]
575
576 *) Change the array representation of binary polynomials: the list
577 of degrees of non-zero coefficients is now terminated with -1.
578 Previously it was terminated with 0, which was also part of the
579 value; thus, the array representation was not applicable to
580 polynomials where t^0 has coefficient zero. This change makes
581 the array representation useful in a more general context.
582 [Douglas Stebila]
583
584 *) Various modifications and fixes to SSL/TLS cipher string
585 handling. For ECC, the code now distinguishes between fixed ECDH
586 with RSA certificates on the one hand and with ECDSA certificates
587 on the other hand, since these are separate ciphersuites. The
588 unused code for Fortezza ciphersuites has been removed.
589
590 For consistency with EDH, ephemeral ECDH is now called "EECDH"
591 (not "ECDHE"). For consistency with the code for DH
592 certificates, use of ECDH certificates is now considered ECDH
593 authentication, not RSA or ECDSA authentication (the latter is
594 merely the CA's signing algorithm and not actively used in the
595 protocol).
596
597 The temporary ciphersuite alias "ECCdraft" is no longer
598 available, and ECC ciphersuites are no longer excluded from "ALL"
599 and "DEFAULT". The following aliases now exist for RFC 4492
600 ciphersuites, most of these by analogy with the DH case:
601
602 kECDHr - ECDH cert, signed with RSA
603 kECDHe - ECDH cert, signed with ECDSA
604 kECDH - ECDH cert (signed with either RSA or ECDSA)
605 kEECDH - ephemeral ECDH
606 ECDH - ECDH cert or ephemeral ECDH
607
608 aECDH - ECDH cert
609 aECDSA - ECDSA cert
610 ECDSA - ECDSA cert
611
612 AECDH - anonymous ECDH
613 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
614
615 [Bodo Moeller]
616
617 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
618 Use correct micalg parameters depending on digest(s) in signed message.
619 [Steve Henson]
620
621 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
622 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
623 [Steve Henson]
624
625 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
626 an engine to register a method. Add ENGINE lookups for methods and
627 functional reference processing.
628 [Steve Henson]
629
630 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
631 EVP_{Sign,Verify}* which allow an application to customise the signature
632 process.
633 [Steve Henson]
634
635 *) New -resign option to smime utility. This adds one or more signers
636 to an existing PKCS#7 signedData structure. Also -md option to use an
637 alternative message digest algorithm for signing.
638 [Steve Henson]
639
640 *) Tidy up PKCS#7 routines and add new functions to make it easier to
641 create PKCS7 structures containing multiple signers. Update smime
642 application to support multiple signers.
643 [Steve Henson]
644
645 *) New -macalg option to pkcs12 utility to allow setting of an alternative
646 digest MAC.
647 [Steve Henson]
648
649 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
650 Reorganize PBE internals to lookup from a static table using NIDs,
651 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
652 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
653 PRF which will be automatically used with PBES2.
654 [Steve Henson]
655
656 *) Replace the algorithm specific calls to generate keys in "req" with the
657 new API.
658 [Steve Henson]
659
660 *) Update PKCS#7 enveloped data routines to use new API. This is now
661 supported by any public key method supporting the encrypt operation. A
662 ctrl is added to allow the public key algorithm to examine or modify
663 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
664 a no op.
665 [Steve Henson]
666
667 *) Add a ctrl to asn1 method to allow a public key algorithm to express
668 a default digest type to use. In most cases this will be SHA1 but some
669 algorithms (such as GOST) need to specify an alternative digest. The
670 return value indicates how strong the prefernce is 1 means optional and
671 2 is mandatory (that is it is the only supported type). Modify
672 ASN1_item_sign() to accept a NULL digest argument to indicate it should
673 use the default md. Update openssl utilities to use the default digest
674 type for signing if it is not explicitly indicated.
675 [Steve Henson]
676
677 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
678 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
679 signing method from the key type. This effectively removes the link
680 between digests and public key types.
681 [Steve Henson]
682
683 *) Add an OID cross reference table and utility functions. Its purpose is to
684 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
685 rsaEncryption. This will allow some of the algorithm specific hackery
686 needed to use the correct OID to be removed.
687 [Steve Henson]
688
689 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
690 structures for PKCS7_sign(). They are now set up by the relevant public
691 key ASN1 method.
692 [Steve Henson]
693
694 *) Add provisional EC pkey method with support for ECDSA and ECDH.
695 [Steve Henson]
696
697 *) Add support for key derivation (agreement) in the API, DH method and
698 pkeyutl.
699 [Steve Henson]
700
701 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
702 public and private key formats. As a side effect these add additional
703 command line functionality not previously available: DSA signatures can be
704 generated and verified using pkeyutl and DH key support and generation in
705 pkey, genpkey.
706 [Steve Henson]
707
708 *) BeOS support.
709 [Oliver Tappe <zooey@hirschkaefer.de>]
710
711 *) New make target "install_html_docs" installs HTML renditions of the
712 manual pages.
713 [Oliver Tappe <zooey@hirschkaefer.de>]
714
715 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
716 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
717 support key and parameter generation and add initial key generation
718 functionality for RSA.
719 [Steve Henson]
720
721 *) Add functions for main EVP_PKEY_method operations. The undocumented
722 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
723 EVP_PKEY_{encrypt,decrypt}_old.
724 [Steve Henson]
725
726 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
727 key API, doesn't do much yet.
728 [Steve Henson]
729
730 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
731 public key algorithms. New option to openssl utility:
732 "list-public-key-algorithms" to print out info.
733 [Steve Henson]
734
735 *) Implement the Supported Elliptic Curves Extension for
736 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
737 [Douglas Stebila]
738
739 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
740 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
741 [Steve Henson]
742
743 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
744 utilities such as rsa, dsa, dsaparam etc except they process any key
745 type.
746 [Steve Henson]
747
748 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
749 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
750 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
751 structure.
752 [Steve Henson]
753
754 *) Initial support for pluggable public key ASN1.
755 De-spaghettify the public key ASN1 handling. Move public and private
756 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
757 algorithm specific handling to a single module within the relevant
758 algorithm directory. Add functions to allow (near) opaque processing
759 of public and private key structures.
760 [Steve Henson]
761
762 *) Implement the Supported Point Formats Extension for
763 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
764 [Douglas Stebila]
765
766 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
767 for the psk identity [hint] and the psk callback functions to the
768 SSL_SESSION, SSL and SSL_CTX structure.
769
770 New ciphersuites:
771 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
772 PSK-AES256-CBC-SHA
773
774 New functions:
775 SSL_CTX_use_psk_identity_hint
776 SSL_get_psk_identity_hint
777 SSL_get_psk_identity
778 SSL_use_psk_identity_hint
779
780 [Mika Kousa and Pasi Eronen of Nokia Corporation]
781
782 *) Add RFC 3161 compliant time stamp request creation, response generation
783 and response verification functionality.
784 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
785
786 *) Add initial support for TLS extensions, specifically for the server_name
787 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
788 have new members for a host name. The SSL data structure has an
789 additional member SSL_CTX *initial_ctx so that new sessions can be
790 stored in that context to allow for session resumption, even after the
791 SSL has been switched to a new SSL_CTX in reaction to a client's
792 server_name extension.
793
794 New functions (subject to change):
795
796 SSL_get_servername()
797 SSL_get_servername_type()
798 SSL_set_SSL_CTX()
799
800 New CTRL codes and macros (subject to change):
801
802 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
803 - SSL_CTX_set_tlsext_servername_callback()
804 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
805 - SSL_CTX_set_tlsext_servername_arg()
806 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
807
808 openssl s_client has a new '-servername ...' option.
809
810 openssl s_server has new options '-servername_host ...', '-cert2 ...',
811 '-key2 ...', '-servername_fatal' (subject to change). This allows
812 testing the HostName extension for a specific single host name ('-cert'
813 and '-key' remain fallbacks for handshakes without HostName
814 negotiation). If the unrecogninzed_name alert has to be sent, this by
815 default is a warning; it becomes fatal with the '-servername_fatal'
816 option.
817
818 [Peter Sylvester, Remy Allais, Christophe Renou]
819
820 *) Whirlpool hash implementation is added.
821 [Andy Polyakov]
822
823 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
824 bn(64,32). Because of instruction set limitations it doesn't have
825 any negative impact on performance. This was done mostly in order
826 to make it possible to share assembler modules, such as bn_mul_mont
827 implementations, between 32- and 64-bit builds without hassle.
828 [Andy Polyakov]
829
830 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
831 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
832 macro.
833 [Bodo Moeller]
834
835 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
836 dedicated Montgomery multiplication procedure, is introduced.
837 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
838 "64-bit" performance on certain 32-bit targets.
839 [Andy Polyakov]
840
841 *) New option SSL_OP_NO_COMP to disable use of compression selectively
842 in SSL structures. New SSL ctrl to set maximum send fragment size.
843 Save memory by seeting the I/O buffer sizes dynamically instead of
844 using the maximum available value.
845 [Steve Henson]
846
847 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
848 in addition to the text details.
849 [Bodo Moeller]
850
851 *) Very, very preliminary EXPERIMENTAL support for printing of general
852 ASN1 structures. This currently produces rather ugly output and doesn't
853 handle several customised structures at all.
854 [Steve Henson]
855
856 *) Integrated support for PVK file format and some related formats such
857 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
858 these in the 'rsa' and 'dsa' utilities.
859 [Steve Henson]
860
861 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
862 [Steve Henson]
863
864 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
865 place for the (very old) "NETSCAPE" format certificates which are now
866 handled using new ASN1 code equivalents.
867 [Steve Henson]
868
869 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
870 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
871 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
872 [Nils Larsch]
873
874 *) Modify CRL distribution points extension code to print out previously
875 unsupported fields. Enhance extension setting code to allow setting of
876 all fields.
877 [Steve Henson]
878
879 *) Add print and set support for Issuing Distribution Point CRL extension.
880 [Steve Henson]
881
882 *) Change 'Configure' script to enable Camellia by default.
883 [NTT]
884
885 Changes between 0.9.8l (?) and 0.9.8m (?) [xx XXX xxxx]
886
887 *) Handle TLS versions 2.0 and later properly and correctly use the
888 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
889 off ancient servers have a habit of sticking around for a while...
890 [Steve Henson]
891
892 *) Modify compression code so it frees up structures without using the
893 ex_data callbacks. This works around a problem where some applications
894 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
895 restarting) then use compression (e.g. SSL with compression) later.
896 This results in significant per-connection memory leaks and
897 has caused some security issues including CVE-2008-1678 and
898 CVE-2009-4355.
899 [Steve Henson]
900
901 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
902 connect (but not renegotiate) with servers which do not support RI.
903 Until RI is more widely deployed this option is enabled by default.
904 [Steve Henson]
905
906 *) Add "missing" ssl ctrls to clear options and mode.
907 [Steve Henson]
908
909 *) If client attempts to renegotiate and doesn't support RI respond with
910 a no_renegotiation alert as required by draft-ietf-tls-renegotiation.
911 Some renegotiating TLS clients will continue a connection gracefully
912 when they receive the alert. Unfortunately OpenSSL mishandled
913 this alert and would hang waiting for a server hello which it will never
914 receive. Now we treat a received no_renegotiation alert as a fatal
915 error. This is because applications requesting a renegotiation might well
916 expect it to succeed and would have no code in place to handle the server
917 denying it so the only safe thing to do is to terminate the connection.
918 [Steve Henson]
919
920 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
921 peer supports secure renegotiation and 0 otherwise. Print out peer
922 renegotiation support in s_client/s_server.
923 [Steve Henson]
924
925 *) Replace the highly broken and deprecated SPKAC certification method with
926 the updated NID creation version. This should correctly handle UTF8.
927 [Steve Henson]
928
929 *) Implement draft-ietf-tls-renegotiation-03. Re-enable
930 renegotiation but require the extension as needed. Unfortunately,
931 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION turns out to be a
932 bad idea. It has been replaced by
933 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
934 SSL_CTX_set_options(). This is really not recommended unless you
935 know what you are doing.
936 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
937
938 *) Fixes to stateless session resumption handling. Use initial_ctx when
939 issuing and attempting to decrypt tickets in case it has changed during
940 servername handling. Use a non-zero length session ID when attempting
941 stateless session resumption: this makes it possible to determine if
942 a resumption has occurred immediately after receiving server hello
943 (several places in OpenSSL subtly assume this) instead of later in
944 the handshake.
945 [Steve Henson]
946
947 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
948 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
949 fixes for a few places where the return code is not checked
950 correctly.
951 [Julia Lawall <julia@diku.dk>]
952
953 *) Add --strict-warnings option to Configure script to include devteam
954 warnings in other configurations.
955 [Steve Henson]
956
957 *) Add support for --libdir option and LIBDIR variable in makefiles. This
958 makes it possible to install openssl libraries in locations which
959 have names other than "lib", for example "/usr/lib64" which some
960 systems need.
961 [Steve Henson, based on patch from Jeremy Utley]
962
963 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
964 X690 8.9.12 and can produce some misleading textual output of OIDs.
965 [Steve Henson, reported by Dan Kaminsky]
966
967 *) Delete MD2 from algorithm tables. This follows the recommendation in
968 several standards that it is not used in new applications due to
969 several cryptographic weaknesses. For binary compatibility reasons
970 the MD2 API is still compiled in by default.
971 [Steve Henson]
972
973 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
974 and restored.
975 [Steve Henson]
976
977 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
978 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
979 clash.
980 [Guenter <lists@gknw.net>]
981
982 *) Fix the server certificate chain building code to use X509_verify_cert(),
983 it used to have an ad-hoc builder which was unable to cope with anything
984 other than a simple chain.
985 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
986
987 *) Don't check self signed certificate signatures in X509_verify_cert()
988 by default (a flag can override this): it just wastes time without
989 adding any security. As a useful side effect self signed root CAs
990 with non-FIPS digests are now usable in FIPS mode.
991 [Steve Henson]
992
993 *) In dtls1_process_out_of_seq_message() the check if the current message
994 is already buffered was missing. For every new message was memory
995 allocated, allowing an attacker to perform an denial of service attack
996 with sending out of seq handshake messages until there is no memory
997 left. Additionally every future messege was buffered, even if the
998 sequence number made no sense and would be part of another handshake.
999 So only messages with sequence numbers less than 10 in advance will be
1000 buffered. (CVE-2009-1378)
1001 [Robin Seggelmann, discovered by Daniel Mentz]
1002
1003 *) Records are buffered if they arrive with a future epoch to be
1004 processed after finishing the corresponding handshake. There is
1005 currently no limitation to this buffer allowing an attacker to perform
1006 a DOS attack with sending records with future epochs until there is no
1007 memory left. This patch adds the pqueue_size() function to detemine
1008 the size of a buffer and limits the record buffer to 100 entries.
1009 (CVE-2009-1377)
1010 [Robin Seggelmann, discovered by Daniel Mentz]
1011
1012 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1013 parent structure is freed. (CVE-2009-1379)
1014 [Daniel Mentz]
1015
1016 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1017 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1018
1019 *) Add 2.5.4.* OIDs
1020 [Ilya O. <vrghost@gmail.com>]
1021
1022 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1023
1024 *) Disable renegotiation completely - this fixes a severe security
1025 problem (CVE-2009-3555) at the cost of breaking all
1026 renegotiation. Renegotiation can be re-enabled by setting
1027 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1028 run-time. This is really not recommended unless you know what
1029 you're doing.
1030 [Ben Laurie]
1031
1032 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1033
1034 *) Don't set val to NULL when freeing up structures, it is freed up by
1035 underlying code. If sizeof(void *) > sizeof(long) this can result in
1036 zeroing past the valid field. (CVE-2009-0789)
1037 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1038
1039 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1040 checked correctly. This would allow some invalid signed attributes to
1041 appear to verify correctly. (CVE-2009-0591)
1042 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1043
1044 *) Reject UniversalString and BMPString types with invalid lengths. This
1045 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1046 a legal length. (CVE-2009-0590)
1047 [Steve Henson]
1048
1049 *) Set S/MIME signing as the default purpose rather than setting it
1050 unconditionally. This allows applications to override it at the store
1051 level.
1052 [Steve Henson]
1053
1054 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1055 to handle some structures.
1056 [Steve Henson]
1057
1058 *) Improve efficiency of mem_gets: don't search whole buffer each time
1059 for a '\n'
1060 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1061
1062 *) New -hex option for openssl rand.
1063 [Matthieu Herrb]
1064
1065 *) Print out UTF8String and NumericString when parsing ASN1.
1066 [Steve Henson]
1067
1068 *) Support NumericString type for name components.
1069 [Steve Henson]
1070
1071 *) Allow CC in the environment to override the automatically chosen
1072 compiler. Note that nothing is done to ensure flags work with the
1073 chosen compiler.
1074 [Ben Laurie]
1075
1076 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1077
1078 *) Properly check EVP_VerifyFinal() and similar return values
1079 (CVE-2008-5077).
1080 [Ben Laurie, Bodo Moeller, Google Security Team]
1081
1082 *) Enable TLS extensions by default.
1083 [Ben Laurie]
1084
1085 *) Allow the CHIL engine to be loaded, whether the application is
1086 multithreaded or not. (This does not release the developer from the
1087 obligation to set up the dynamic locking callbacks.)
1088 [Sander Temme <sander@temme.net>]
1089
1090 *) Use correct exit code if there is an error in dgst command.
1091 [Steve Henson; problem pointed out by Roland Dirlewanger]
1092
1093 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1094 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1095 [Bodo Moeller]
1096
1097 *) Add experimental JPAKE support, including demo authentication in
1098 s_client and s_server.
1099 [Ben Laurie]
1100
1101 *) Set the comparison function in v3_addr_canonize().
1102 [Rob Austein <sra@hactrn.net>]
1103
1104 *) Add support for XMPP STARTTLS in s_client.
1105 [Philip Paeps <philip@freebsd.org>]
1106
1107 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1108 to ensure that even with this option, only ciphersuites in the
1109 server's preference list will be accepted. (Note that the option
1110 applies only when resuming a session, so the earlier behavior was
1111 just about the algorithm choice for symmetric cryptography.)
1112 [Bodo Moeller]
1113
1114 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1115
1116 *) Fix NULL pointer dereference if a DTLS server received
1117 ChangeCipherSpec as first record (CVE-2009-1386).
1118 [PR #1679]
1119
1120 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1121 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1122 [Nagendra Modadugu]
1123
1124 *) The fix in 0.9.8c that supposedly got rid of unsafe
1125 double-checked locking was incomplete for RSA blinding,
1126 addressing just one layer of what turns out to have been
1127 doubly unsafe triple-checked locking.
1128
1129 So now fix this for real by retiring the MONT_HELPER macro
1130 in crypto/rsa/rsa_eay.c.
1131
1132 [Bodo Moeller; problem pointed out by Marius Schilder]
1133
1134 *) Various precautionary measures:
1135
1136 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1137
1138 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1139 (NB: This would require knowledge of the secret session ticket key
1140 to exploit, in which case you'd be SOL either way.)
1141
1142 - Change bn_nist.c so that it will properly handle input BIGNUMs
1143 outside the expected range.
1144
1145 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1146 builds.
1147
1148 [Neel Mehta, Bodo Moeller]
1149
1150 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1151 the load fails. Useful for distros.
1152 [Ben Laurie and the FreeBSD team]
1153
1154 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1155 [Steve Henson]
1156
1157 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1158 [Huang Ying]
1159
1160 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1161
1162 This work was sponsored by Logica.
1163 [Steve Henson]
1164
1165 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1166 keystores. Support for SSL/TLS client authentication too.
1167 Not compiled unless enable-capieng specified to Configure.
1168
1169 This work was sponsored by Logica.
1170 [Steve Henson]
1171
1172 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1173 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1174 attribute creation routines such as certifcate requests and PKCS#12
1175 files.
1176 [Steve Henson]
1177
1178 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1179
1180 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1181 handshake which could lead to a cilent crash as found using the
1182 Codenomicon TLS test suite (CVE-2008-1672)
1183 [Steve Henson, Mark Cox]
1184
1185 *) Fix double free in TLS server name extensions which could lead to
1186 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1187 [Joe Orton]
1188
1189 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1190
1191 Clear the error queue to ensure that error entries left from
1192 older function calls do not interfere with the correct operation.
1193 [Lutz Jaenicke, Erik de Castro Lopo]
1194
1195 *) Remove root CA certificates of commercial CAs:
1196
1197 The OpenSSL project does not recommend any specific CA and does not
1198 have any policy with respect to including or excluding any CA.
1199 Therefore it does not make any sense to ship an arbitrary selection
1200 of root CA certificates with the OpenSSL software.
1201 [Lutz Jaenicke]
1202
1203 *) RSA OAEP patches to fix two separate invalid memory reads.
1204 The first one involves inputs when 'lzero' is greater than
1205 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1206 before the beginning of from). The second one involves inputs where
1207 the 'db' section contains nothing but zeroes (there is a one-byte
1208 invalid read after the end of 'db').
1209 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1210
1211 *) Partial backport from 0.9.9-dev:
1212
1213 Introduce bn_mul_mont (dedicated Montgomery multiplication
1214 procedure) as a candidate for BIGNUM assembler implementation.
1215 While 0.9.9-dev uses assembler for various architectures, only
1216 x86_64 is available by default here in the 0.9.8 branch, and
1217 32-bit x86 is available through a compile-time setting.
1218
1219 To try the 32-bit x86 assembler implementation, use Configure
1220 option "enable-montasm" (which exists only for this backport).
1221
1222 As "enable-montasm" for 32-bit x86 disclaims code stability
1223 anyway, in this constellation we activate additional code
1224 backported from 0.9.9-dev for further performance improvements,
1225 namely BN_from_montgomery_word. (To enable this otherwise,
1226 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1227
1228 [Andy Polyakov (backport partially by Bodo Moeller)]
1229
1230 *) Add TLS session ticket callback. This allows an application to set
1231 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1232 values. This is useful for key rollover for example where several key
1233 sets may exist with different names.
1234 [Steve Henson]
1235
1236 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1237 This was broken until now in 0.9.8 releases, such that the only way
1238 a registered ENGINE could be used (assuming it initialises
1239 successfully on the host) was to explicitly set it as the default
1240 for the relevant algorithms. This is in contradiction with 0.9.7
1241 behaviour and the documentation. With this fix, when an ENGINE is
1242 registered into a given algorithm's table of implementations, the
1243 'uptodate' flag is reset so that auto-discovery will be used next
1244 time a new context for that algorithm attempts to select an
1245 implementation.
1246 [Ian Lister (tweaked by Geoff Thorpe)]
1247
1248 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1249 implemention in the following ways:
1250
1251 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1252 hard coded.
1253
1254 Lack of BER streaming support means one pass streaming processing is
1255 only supported if data is detached: setting the streaming flag is
1256 ignored for embedded content.
1257
1258 CMS support is disabled by default and must be explicitly enabled
1259 with the enable-cms configuration option.
1260 [Steve Henson]
1261
1262 *) Update the GMP engine glue to do direct copies between BIGNUM and
1263 mpz_t when openssl and GMP use the same limb size. Otherwise the
1264 existing "conversion via a text string export" trick is still used.
1265 [Paul Sheer <paulsheer@gmail.com>]
1266
1267 *) Zlib compression BIO. This is a filter BIO which compressed and
1268 uncompresses any data passed through it.
1269 [Steve Henson]
1270
1271 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1272 RFC3394 compatible AES key wrapping.
1273 [Steve Henson]
1274
1275 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1276 sets string data without copying. X509_ALGOR_set0() and
1277 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1278 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1279 from an X509_ATTRIBUTE structure optionally checking it occurs only
1280 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1281 data.
1282 [Steve Henson]
1283
1284 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1285 to get the expected BN_FLG_CONSTTIME behavior.
1286 [Bodo Moeller (Google)]
1287
1288 *) Netware support:
1289
1290 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1291 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1292 - added some more tests to do_tests.pl
1293 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1294 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1295 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1296 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1297 - various changes to netware.pl to enable gcc-cross builds on Win32
1298 platform
1299 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1300 - various changes to fix missing prototype warnings
1301 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1302 - added AES, WHIRLPOOL and CPUID assembler code to build files
1303 - added missing AES assembler make rules to mk1mf.pl
1304 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1305 [Guenter Knauf <eflash@gmx.net>]
1306
1307 *) Implement certificate status request TLS extension defined in RFC3546.
1308 A client can set the appropriate parameters and receive the encoded
1309 OCSP response via a callback. A server can query the supplied parameters
1310 and set the encoded OCSP response in the callback. Add simplified examples
1311 to s_client and s_server.
1312 [Steve Henson]
1313
1314 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1315
1316 *) Fix various bugs:
1317 + Binary incompatibility of ssl_ctx_st structure
1318 + DTLS interoperation with non-compliant servers
1319 + Don't call get_session_cb() without proposed session
1320 + Fix ia64 assembler code
1321 [Andy Polyakov, Steve Henson]
1322
1323 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1324
1325 *) DTLS Handshake overhaul. There were longstanding issues with
1326 OpenSSL DTLS implementation, which were making it impossible for
1327 RFC 4347 compliant client to communicate with OpenSSL server.
1328 Unfortunately just fixing these incompatibilities would "cut off"
1329 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1330 server keeps tolerating non RFC compliant syntax. The opposite is
1331 not true, 0.9.8f client can not communicate with earlier server.
1332 This update even addresses CVE-2007-4995.
1333 [Andy Polyakov]
1334
1335 *) Changes to avoid need for function casts in OpenSSL: some compilers
1336 (gcc 4.2 and later) reject their use.
1337 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1338 Steve Henson]
1339
1340 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1341 RFC4507bis. The encrypted ticket format is an encrypted encoded
1342 SSL_SESSION structure, that way new session features are automatically
1343 supported.
1344
1345 If a client application caches session in an SSL_SESSION structure
1346 support is transparent because tickets are now stored in the encoded
1347 SSL_SESSION.
1348
1349 The SSL_CTX structure automatically generates keys for ticket
1350 protection in servers so again support should be possible
1351 with no application modification.
1352
1353 If a client or server wishes to disable RFC4507 support then the option
1354 SSL_OP_NO_TICKET can be set.
1355
1356 Add a TLS extension debugging callback to allow the contents of any client
1357 or server extensions to be examined.
1358
1359 This work was sponsored by Google.
1360 [Steve Henson]
1361
1362 *) Add initial support for TLS extensions, specifically for the server_name
1363 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1364 have new members for a host name. The SSL data structure has an
1365 additional member SSL_CTX *initial_ctx so that new sessions can be
1366 stored in that context to allow for session resumption, even after the
1367 SSL has been switched to a new SSL_CTX in reaction to a client's
1368 server_name extension.
1369
1370 New functions (subject to change):
1371
1372 SSL_get_servername()
1373 SSL_get_servername_type()
1374 SSL_set_SSL_CTX()
1375
1376 New CTRL codes and macros (subject to change):
1377
1378 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1379 - SSL_CTX_set_tlsext_servername_callback()
1380 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1381 - SSL_CTX_set_tlsext_servername_arg()
1382 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1383
1384 openssl s_client has a new '-servername ...' option.
1385
1386 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1387 '-key2 ...', '-servername_fatal' (subject to change). This allows
1388 testing the HostName extension for a specific single host name ('-cert'
1389 and '-key' remain fallbacks for handshakes without HostName
1390 negotiation). If the unrecogninzed_name alert has to be sent, this by
1391 default is a warning; it becomes fatal with the '-servername_fatal'
1392 option.
1393
1394 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1395
1396 *) Add AES and SSE2 assembly language support to VC++ build.
1397 [Steve Henson]
1398
1399 *) Mitigate attack on final subtraction in Montgomery reduction.
1400 [Andy Polyakov]
1401
1402 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1403 (which previously caused an internal error).
1404 [Bodo Moeller]
1405
1406 *) Squeeze another 10% out of IGE mode when in != out.
1407 [Ben Laurie]
1408
1409 *) AES IGE mode speedup.
1410 [Dean Gaudet (Google)]
1411
1412 *) Add the Korean symmetric 128-bit cipher SEED (see
1413 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1414 add SEED ciphersuites from RFC 4162:
1415
1416 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1417 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1418 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1419 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1420
1421 To minimize changes between patchlevels in the OpenSSL 0.9.8
1422 series, SEED remains excluded from compilation unless OpenSSL
1423 is configured with 'enable-seed'.
1424 [KISA, Bodo Moeller]
1425
1426 *) Mitigate branch prediction attacks, which can be practical if a
1427 single processor is shared, allowing a spy process to extract
1428 information. For detailed background information, see
1429 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1430 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
1431 and Necessary Software Countermeasures"). The core of the change
1432 are new versions BN_div_no_branch() and
1433 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
1434 respectively, which are slower, but avoid the security-relevant
1435 conditional branches. These are automatically called by BN_div()
1436 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
1437 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
1438 remove a conditional branch.
1439
1440 BN_FLG_CONSTTIME is the new name for the previous
1441 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
1442 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
1443 in the exponent causes BN_mod_exp_mont() to use the alternative
1444 implementation in BN_mod_exp_mont_consttime().) The old name
1445 remains as a deprecated alias.
1446
1447 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
1448 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
1449 constant-time implementations for more than just exponentiation.
1450 Here too the old name is kept as a deprecated alias.
1451
1452 BN_BLINDING_new() will now use BN_dup() for the modulus so that
1453 the BN_BLINDING structure gets an independent copy of the
1454 modulus. This means that the previous "BIGNUM *m" argument to
1455 BN_BLINDING_new() and to BN_BLINDING_create_param() now
1456 essentially becomes "const BIGNUM *m", although we can't actually
1457 change this in the header file before 0.9.9. It allows
1458 RSA_setup_blinding() to use BN_with_flags() on the modulus to
1459 enable BN_FLG_CONSTTIME.
1460
1461 [Matthew D Wood (Intel Corp)]
1462
1463 *) In the SSL/TLS server implementation, be strict about session ID
1464 context matching (which matters if an application uses a single
1465 external cache for different purposes). Previously,
1466 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1467 set. This did ensure strict client verification, but meant that,
1468 with applications using a single external cache for quite
1469 different requirements, clients could circumvent ciphersuite
1470 restrictions for a given session ID context by starting a session
1471 in a different context.
1472 [Bodo Moeller]
1473
1474 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1475 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1476 authentication-only ciphersuites.
1477 [Bodo Moeller]
1478
1479 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
1480 not complete and could lead to a possible single byte overflow
1481 (CVE-2007-5135) [Ben Laurie]
1482
1483 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
1484
1485 *) Since AES128 and AES256 (and similarly Camellia128 and
1486 Camellia256) share a single mask bit in the logic of
1487 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1488 kludge to work properly if AES128 is available and AES256 isn't
1489 (or if Camellia128 is available and Camellia256 isn't).
1490 [Victor Duchovni]
1491
1492 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
1493 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
1494 When a point or a seed is encoded in a BIT STRING, we need to
1495 prevent the removal of trailing zero bits to get the proper DER
1496 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
1497 of a NamedBitList, for which trailing 0 bits need to be removed.)
1498 [Bodo Moeller]
1499
1500 *) Have SSL/TLS server implementation tolerate "mismatched" record
1501 protocol version while receiving ClientHello even if the
1502 ClientHello is fragmented. (The server can't insist on the
1503 particular protocol version it has chosen before the ServerHello
1504 message has informed the client about his choice.)
1505 [Bodo Moeller]
1506
1507 *) Add RFC 3779 support.
1508 [Rob Austein for ARIN, Ben Laurie]
1509
1510 *) Load error codes if they are not already present instead of using a
1511 static variable. This allows them to be cleanly unloaded and reloaded.
1512 Improve header file function name parsing.
1513 [Steve Henson]
1514
1515 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
1516 or CAPABILITY handshake as required by RFCs.
1517 [Goetz Babin-Ebell]
1518
1519 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
1520
1521 *) Introduce limits to prevent malicious keys being able to
1522 cause a denial of service. (CVE-2006-2940)
1523 [Steve Henson, Bodo Moeller]
1524
1525 *) Fix ASN.1 parsing of certain invalid structures that can result
1526 in a denial of service. (CVE-2006-2937) [Steve Henson]
1527
1528 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1529 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1530
1531 *) Fix SSL client code which could crash if connecting to a
1532 malicious SSLv2 server. (CVE-2006-4343)
1533 [Tavis Ormandy and Will Drewry, Google Security Team]
1534
1535 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
1536 match only those. Before that, "AES256-SHA" would be interpreted
1537 as a pattern and match "AES128-SHA" too (since AES128-SHA got
1538 the same strength classification in 0.9.7h) as we currently only
1539 have a single AES bit in the ciphersuite description bitmap.
1540 That change, however, also applied to ciphersuite strings such as
1541 "RC4-MD5" that intentionally matched multiple ciphersuites --
1542 namely, SSL 2.0 ciphersuites in addition to the more common ones
1543 from SSL 3.0/TLS 1.0.
1544
1545 So we change the selection algorithm again: Naming an explicit
1546 ciphersuite selects this one ciphersuite, and any other similar
1547 ciphersuite (same bitmap) from *other* protocol versions.
1548 Thus, "RC4-MD5" again will properly select both the SSL 2.0
1549 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
1550
1551 Since SSL 2.0 does not have any ciphersuites for which the
1552 128/256 bit distinction would be relevant, this works for now.
1553 The proper fix will be to use different bits for AES128 and
1554 AES256, which would have avoided the problems from the beginning;
1555 however, bits are scarce, so we can only do this in a new release
1556 (not just a patchlevel) when we can change the SSL_CIPHER
1557 definition to split the single 'unsigned long mask' bitmap into
1558 multiple values to extend the available space.
1559
1560 [Bodo Moeller]
1561
1562 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
1563
1564 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1565 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1566
1567 *) Add AES IGE and biIGE modes.
1568 [Ben Laurie]
1569
1570 *) Change the Unix randomness entropy gathering to use poll() when
1571 possible instead of select(), since the latter has some
1572 undesirable limitations.
1573 [Darryl Miles via Richard Levitte and Bodo Moeller]
1574
1575 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
1576 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
1577 cannot be implicitly activated as part of, e.g., the "AES" alias.
1578 However, please upgrade to OpenSSL 0.9.9[-dev] for
1579 non-experimental use of the ECC ciphersuites to get TLS extension
1580 support, which is required for curve and point format negotiation
1581 to avoid potential handshake problems.
1582 [Bodo Moeller]
1583
1584 *) Disable rogue ciphersuites:
1585
1586 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1587 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1588 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1589
1590 The latter two were purportedly from
1591 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1592 appear there.
1593
1594 Also deactivate the remaining ciphersuites from
1595 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1596 unofficial, and the ID has long expired.
1597 [Bodo Moeller]
1598
1599 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1600 dual-core machines) and other potential thread-safety issues.
1601 [Bodo Moeller]
1602
1603 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
1604 versions), which is now available for royalty-free use
1605 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
1606 Also, add Camellia TLS ciphersuites from RFC 4132.
1607
1608 To minimize changes between patchlevels in the OpenSSL 0.9.8
1609 series, Camellia remains excluded from compilation unless OpenSSL
1610 is configured with 'enable-camellia'.
1611 [NTT]
1612
1613 *) Disable the padding bug check when compression is in use. The padding
1614 bug check assumes the first packet is of even length, this is not
1615 necessarily true if compresssion is enabled and can result in false
1616 positives causing handshake failure. The actual bug test is ancient
1617 code so it is hoped that implementations will either have fixed it by
1618 now or any which still have the bug do not support compression.
1619 [Steve Henson]
1620
1621 Changes between 0.9.8a and 0.9.8b [04 May 2006]
1622
1623 *) When applying a cipher rule check to see if string match is an explicit
1624 cipher suite and only match that one cipher suite if it is.
1625 [Steve Henson]
1626
1627 *) Link in manifests for VC++ if needed.
1628 [Austin Ziegler <halostatue@gmail.com>]
1629
1630 *) Update support for ECC-based TLS ciphersuites according to
1631 draft-ietf-tls-ecc-12.txt with proposed changes (but without
1632 TLS extensions, which are supported starting with the 0.9.9
1633 branch, not in the OpenSSL 0.9.8 branch).
1634 [Douglas Stebila]
1635
1636 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
1637 opaque EVP_CIPHER_CTX handling.
1638 [Steve Henson]
1639
1640 *) Fixes and enhancements to zlib compression code. We now only use
1641 "zlib1.dll" and use the default __cdecl calling convention on Win32
1642 to conform with the standards mentioned here:
1643 http://www.zlib.net/DLL_FAQ.txt
1644 Static zlib linking now works on Windows and the new --with-zlib-include
1645 --with-zlib-lib options to Configure can be used to supply the location
1646 of the headers and library. Gracefully handle case where zlib library
1647 can't be loaded.
1648 [Steve Henson]
1649
1650 *) Several fixes and enhancements to the OID generation code. The old code
1651 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
1652 handle numbers larger than ULONG_MAX, truncated printing and had a
1653 non standard OBJ_obj2txt() behaviour.
1654 [Steve Henson]
1655
1656 *) Add support for building of engines under engine/ as shared libraries
1657 under VC++ build system.
1658 [Steve Henson]
1659
1660 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
1661 Hopefully, we will not see any false combination of paths any more.
1662 [Richard Levitte]
1663
1664 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
1665
1666 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1667 (part of SSL_OP_ALL). This option used to disable the
1668 countermeasure against man-in-the-middle protocol-version
1669 rollback in the SSL 2.0 server implementation, which is a bad
1670 idea. (CVE-2005-2969)
1671
1672 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1673 for Information Security, National Institute of Advanced Industrial
1674 Science and Technology [AIST], Japan)]
1675
1676 *) Add two function to clear and return the verify parameter flags.
1677 [Steve Henson]
1678
1679 *) Keep cipherlists sorted in the source instead of sorting them at
1680 runtime, thus removing the need for a lock.
1681 [Nils Larsch]
1682
1683 *) Avoid some small subgroup attacks in Diffie-Hellman.
1684 [Nick Mathewson and Ben Laurie]
1685
1686 *) Add functions for well-known primes.
1687 [Nick Mathewson]
1688
1689 *) Extended Windows CE support.
1690 [Satoshi Nakamura and Andy Polyakov]
1691
1692 *) Initialize SSL_METHOD structures at compile time instead of during
1693 runtime, thus removing the need for a lock.
1694 [Steve Henson]
1695
1696 *) Make PKCS7_decrypt() work even if no certificate is supplied by
1697 attempting to decrypt each encrypted key in turn. Add support to
1698 smime utility.
1699 [Steve Henson]
1700
1701 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
1702
1703 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
1704 OpenSSL 0.9.8.]
1705
1706 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
1707 [Richard Levitte]
1708
1709 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
1710 key into the same file any more.
1711 [Richard Levitte]
1712
1713 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
1714 [Andy Polyakov]
1715
1716 *) Add -utf8 command line and config file option to 'ca'.
1717 [Stefan <stf@udoma.org]
1718
1719 *) Removed the macro des_crypt(), as it seems to conflict with some
1720 libraries. Use DES_crypt().
1721 [Richard Levitte]
1722
1723 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
1724 involves renaming the source and generated shared-libs for
1725 both. The engines will accept the corrected or legacy ids
1726 ('ncipher' and '4758_cca' respectively) when binding. NB,
1727 this only applies when building 'shared'.
1728 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
1729
1730 *) Add attribute functions to EVP_PKEY structure. Modify
1731 PKCS12_create() to recognize a CSP name attribute and
1732 use it. Make -CSP option work again in pkcs12 utility.
1733 [Steve Henson]
1734
1735 *) Add new functionality to the bn blinding code:
1736 - automatic re-creation of the BN_BLINDING parameters after
1737 a fixed number of uses (currently 32)
1738 - add new function for parameter creation
1739 - introduce flags to control the update behaviour of the
1740 BN_BLINDING parameters
1741 - hide BN_BLINDING structure
1742 Add a second BN_BLINDING slot to the RSA structure to improve
1743 performance when a single RSA object is shared among several
1744 threads.
1745 [Nils Larsch]
1746
1747 *) Add support for DTLS.
1748 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
1749
1750 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
1751 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
1752 [Walter Goulet]
1753
1754 *) Remove buggy and incompletet DH cert support from
1755 ssl/ssl_rsa.c and ssl/s3_both.c
1756 [Nils Larsch]
1757
1758 *) Use SHA-1 instead of MD5 as the default digest algorithm for
1759 the apps/openssl applications.
1760 [Nils Larsch]
1761
1762 *) Compile clean with "-Wall -Wmissing-prototypes
1763 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
1764 DEBUG_SAFESTACK must also be set.
1765 [Ben Laurie]
1766
1767 *) Change ./Configure so that certain algorithms can be disabled by default.
1768 The new counterpiece to "no-xxx" is "enable-xxx".
1769
1770 The patented RC5 and MDC2 algorithms will now be disabled unless
1771 "enable-rc5" and "enable-mdc2", respectively, are specified.
1772
1773 (IDEA remains enabled despite being patented. This is because IDEA
1774 is frequently required for interoperability, and there is no license
1775 fee for non-commercial use. As before, "no-idea" can be used to
1776 avoid this algorithm.)
1777
1778 [Bodo Moeller]
1779
1780 *) Add processing of proxy certificates (see RFC 3820). This work was
1781 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
1782 EGEE (Enabling Grids for E-science in Europe).
1783 [Richard Levitte]
1784
1785 *) RC4 performance overhaul on modern architectures/implementations, such
1786 as Intel P4, IA-64 and AMD64.
1787 [Andy Polyakov]
1788
1789 *) New utility extract-section.pl. This can be used specify an alternative
1790 section number in a pod file instead of having to treat each file as
1791 a separate case in Makefile. This can be done by adding two lines to the
1792 pod file:
1793
1794 =for comment openssl_section:XXX
1795
1796 The blank line is mandatory.
1797
1798 [Steve Henson]
1799
1800 *) New arguments -certform, -keyform and -pass for s_client and s_server
1801 to allow alternative format key and certificate files and passphrase
1802 sources.
1803 [Steve Henson]
1804
1805 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
1806 update associated structures and add various utility functions.
1807
1808 Add new policy related verify parameters, include policy checking in
1809 standard verify code. Enhance 'smime' application with extra parameters
1810 to support policy checking and print out.
1811 [Steve Henson]
1812
1813 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
1814 Nehemiah processors. These extensions support AES encryption in hardware
1815 as well as RNG (though RNG support is currently disabled).
1816 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
1817
1818 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
1819 [Geoff Thorpe]
1820
1821 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
1822 [Andy Polyakov and a number of other people]
1823
1824 *) Improved PowerPC platform support. Most notably BIGNUM assembler
1825 implementation contributed by IBM.
1826 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
1827
1828 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
1829 exponent rather than 'unsigned long'. There is a corresponding change to
1830 the new 'rsa_keygen' element of the RSA_METHOD structure.
1831 [Jelte Jansen, Geoff Thorpe]
1832
1833 *) Functionality for creating the initial serial number file is now
1834 moved from CA.pl to the 'ca' utility with a new option -create_serial.
1835
1836 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
1837 number file to 1, which is bound to cause problems. To avoid
1838 the problems while respecting compatibility between different 0.9.7
1839 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
1840 CA.pl for serial number initialization. With the new release 0.9.8,
1841 we can fix the problem directly in the 'ca' utility.)
1842 [Steve Henson]
1843
1844 *) Reduced header interdepencies by declaring more opaque objects in
1845 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
1846 give fewer recursive includes, which could break lazy source code - so
1847 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
1848 developers should define this symbol when building and using openssl to
1849 ensure they track the recommended behaviour, interfaces, [etc], but
1850 backwards-compatible behaviour prevails when this isn't defined.
1851 [Geoff Thorpe]
1852
1853 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
1854 [Steve Henson]
1855
1856 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
1857 This will generate a random key of the appropriate length based on the
1858 cipher context. The EVP_CIPHER can provide its own random key generation
1859 routine to support keys of a specific form. This is used in the des and
1860 3des routines to generate a key of the correct parity. Update S/MIME
1861 code to use new functions and hence generate correct parity DES keys.
1862 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
1863 valid (weak or incorrect parity).
1864 [Steve Henson]
1865
1866 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
1867 as looking them up. This is useful when the verified structure may contain
1868 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
1869 present unless the new PKCS7_NO_CRL flag is asserted.
1870 [Steve Henson]
1871
1872 *) Extend ASN1 oid configuration module. It now additionally accepts the
1873 syntax:
1874
1875 shortName = some long name, 1.2.3.4
1876 [Steve Henson]
1877
1878 *) Reimplemented the BN_CTX implementation. There is now no more static
1879 limitation on the number of variables it can handle nor the depth of the
1880 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
1881 information can now expand as required, and rather than having a single
1882 static array of bignums, BN_CTX now uses a linked-list of such arrays
1883 allowing it to expand on demand whilst maintaining the usefulness of
1884 BN_CTX's "bundling".
1885 [Geoff Thorpe]
1886
1887 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
1888 to allow all RSA operations to function using a single BN_CTX.
1889 [Geoff Thorpe]
1890
1891 *) Preliminary support for certificate policy evaluation and checking. This
1892 is initially intended to pass the tests outlined in "Conformance Testing
1893 of Relying Party Client Certificate Path Processing Logic" v1.07.
1894 [Steve Henson]
1895
1896 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1897 remained unused and not that useful. A variety of other little bignum
1898 tweaks and fixes have also been made continuing on from the audit (see
1899 below).
1900 [Geoff Thorpe]
1901
1902 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1903 associated ASN1, EVP and SSL functions and old ASN1 macros.
1904 [Richard Levitte]
1905
1906 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1907 and this should never fail. So the return value from the use of
1908 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1909 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1910 [Geoff Thorpe]
1911
1912 *) BN_CTX_get() should return zero-valued bignums, providing the same
1913 initialised value as BN_new().
1914 [Geoff Thorpe, suggested by Ulf Möller]
1915
1916 *) Support for inhibitAnyPolicy certificate extension.
1917 [Steve Henson]
1918
1919 *) An audit of the BIGNUM code is underway, for which debugging code is
1920 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1921 is considered valid when processing BIGNUMs, and causes execution to
1922 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1923 further steps are taken to deliberately pollute unused data in BIGNUM
1924 structures to try and expose faulty code further on. For now, openssl will
1925 (in its default mode of operation) continue to tolerate the inconsistent
1926 forms that it has tolerated in the past, but authors and packagers should
1927 consider trying openssl and their own applications when compiled with
1928 these debugging symbols defined. It will help highlight potential bugs in
1929 their own code, and will improve the test coverage for OpenSSL itself. At
1930 some point, these tighter rules will become openssl's default to improve
1931 maintainability, though the assert()s and other overheads will remain only
1932 in debugging configurations. See bn.h for more details.
1933 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1934
1935 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1936 that can only be obtained through BN_CTX_new() (which implicitly
1937 initialises it). The presence of this function only made it possible
1938 to overwrite an existing structure (and cause memory leaks).
1939 [Geoff Thorpe]
1940
1941 *) Because of the callback-based approach for implementing LHASH as a
1942 template type, lh_insert() adds opaque objects to hash-tables and
1943 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1944 to clean up those corresponding objects before destroying the hash table
1945 (and losing the object pointers). So some over-zealous constifications in
1946 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1947 objects as "const" and the lh_doall[_arg] callback wrappers are not
1948 prototyped to have "const" restrictions on the object pointers they are
1949 given (and so aren't required to cast them away any more).
1950 [Geoff Thorpe]
1951
1952 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1953 (speed) prefers to use its own implementation. The two implementations
1954 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1955 its object type properly exposed (MS_TM) instead of casting to/from "char
1956 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1957 aren't necessarily the greatest nomenclatures - but this is what was used
1958 internally to the implementation so I've used that for now.
1959 [Geoff Thorpe]
1960
1961 *) Ensure that deprecated functions do not get compiled when
1962 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1963 the self-tests were still using deprecated key-generation functions so
1964 these have been updated also.
1965 [Geoff Thorpe]
1966
1967 *) Reorganise PKCS#7 code to separate the digest location functionality
1968 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1969 New function PKCS7_set_digest() to set the digest type for PKCS#7
1970 digestedData type. Add additional code to correctly generate the
1971 digestedData type and add support for this type in PKCS7 initialization
1972 functions.
1973 [Steve Henson]
1974
1975 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1976 structure of type "other".
1977 [Steve Henson]
1978
1979 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1980 sure the loop does correctly stop and breaking ("division by zero")
1981 modulus operations are not performed. The (pre-generated) prime
1982 table crypto/bn/bn_prime.h was already correct, but it could not be
1983 re-generated on some platforms because of the "division by zero"
1984 situation in the script.
1985 [Ralf S. Engelschall]
1986
1987 *) Update support for ECC-based TLS ciphersuites according to
1988 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1989 SHA-1 now is only used for "small" curves (where the
1990 representation of a field element takes up to 24 bytes); for
1991 larger curves, the field element resulting from ECDH is directly
1992 used as premaster secret.
1993 [Douglas Stebila (Sun Microsystems Laboratories)]
1994
1995 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1996 curve secp160r1 to the tests.
1997 [Douglas Stebila (Sun Microsystems Laboratories)]
1998
1999 *) Add the possibility to load symbols globally with DSO.
2000 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2001
2002 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2003 control of the error stack.
2004 [Richard Levitte]
2005
2006 *) Add support for STORE in ENGINE.
2007 [Richard Levitte]
2008
2009 *) Add the STORE type. The intention is to provide a common interface
2010 to certificate and key stores, be they simple file-based stores, or
2011 HSM-type store, or LDAP stores, or...
2012 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2013 [Richard Levitte]
2014
2015 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2016 pass a list of arguments to any function as well as provide a way
2017 for a function to pass data back to the caller.
2018 [Richard Levitte]
2019
2020 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2021 works like BUF_strdup() but can be used to duplicate a portion of
2022 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2023 a memory area.
2024 [Richard Levitte]
2025
2026 *) Add the function sk_find_ex() which works like sk_find(), but will
2027 return an index to an element even if an exact match couldn't be
2028 found. The index is guaranteed to point at the element where the
2029 searched-for key would be inserted to preserve sorting order.
2030 [Richard Levitte]
2031
2032 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2033 takes an extra flags argument for optional functionality. Currently,
2034 the following flags are defined:
2035
2036 OBJ_BSEARCH_VALUE_ON_NOMATCH
2037 This one gets OBJ_bsearch_ex() to return a pointer to the first
2038 element where the comparing function returns a negative or zero
2039 number.
2040
2041 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2042 This one gets OBJ_bsearch_ex() to return a pointer to the first
2043 element where the comparing function returns zero. This is useful
2044 if there are more than one element where the comparing function
2045 returns zero.
2046 [Richard Levitte]
2047
2048 *) Make it possible to create self-signed certificates with 'openssl ca'
2049 in such a way that the self-signed certificate becomes part of the
2050 CA database and uses the same mechanisms for serial number generation
2051 as all other certificate signing. The new flag '-selfsign' enables
2052 this functionality. Adapt CA.sh and CA.pl.in.
2053 [Richard Levitte]
2054
2055 *) Add functionality to check the public key of a certificate request
2056 against a given private. This is useful to check that a certificate
2057 request can be signed by that key (self-signing).
2058 [Richard Levitte]
2059
2060 *) Make it possible to have multiple active certificates with the same
2061 subject in the CA index file. This is done only if the keyword
2062 'unique_subject' is set to 'no' in the main CA section (default
2063 if 'CA_default') of the configuration file. The value is saved
2064 with the database itself in a separate index attribute file,
2065 named like the index file with '.attr' appended to the name.
2066 [Richard Levitte]
2067
2068 *) Generate muti valued AVAs using '+' notation in config files for
2069 req and dirName.
2070 [Steve Henson]
2071
2072 *) Support for nameConstraints certificate extension.
2073 [Steve Henson]
2074
2075 *) Support for policyConstraints certificate extension.
2076 [Steve Henson]
2077
2078 *) Support for policyMappings certificate extension.
2079 [Steve Henson]
2080
2081 *) Make sure the default DSA_METHOD implementation only uses its
2082 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2083 and change its own handlers to be NULL so as to remove unnecessary
2084 indirection. This lets alternative implementations fallback to the
2085 default implementation more easily.
2086 [Geoff Thorpe]
2087
2088 *) Support for directoryName in GeneralName related extensions
2089 in config files.
2090 [Steve Henson]
2091
2092 *) Make it possible to link applications using Makefile.shared.
2093 Make that possible even when linking against static libraries!
2094 [Richard Levitte]
2095
2096 *) Support for single pass processing for S/MIME signing. This now
2097 means that S/MIME signing can be done from a pipe, in addition
2098 cleartext signing (multipart/signed type) is effectively streaming
2099 and the signed data does not need to be all held in memory.
2100
2101 This is done with a new flag PKCS7_STREAM. When this flag is set
2102 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2103 is done after the data is output (and digests calculated) in
2104 SMIME_write_PKCS7().
2105 [Steve Henson]
2106
2107 *) Add full support for -rpath/-R, both in shared libraries and
2108 applications, at least on the platforms where it's known how
2109 to do it.
2110 [Richard Levitte]
2111
2112 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2113 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2114 will now compute a table of multiples of the generator that
2115 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2116 faster (notably in the case of a single point multiplication,
2117 scalar * generator).
2118 [Nils Larsch, Bodo Moeller]
2119
2120 *) IPv6 support for certificate extensions. The various extensions
2121 which use the IP:a.b.c.d can now take IPv6 addresses using the
2122 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2123 correctly.
2124 [Steve Henson]
2125
2126 *) Added an ENGINE that implements RSA by performing private key
2127 exponentiations with the GMP library. The conversions to and from
2128 GMP's mpz_t format aren't optimised nor are any montgomery forms
2129 cached, and on x86 it appears OpenSSL's own performance has caught up.
2130 However there are likely to be other architectures where GMP could
2131 provide a boost. This ENGINE is not built in by default, but it can be
2132 specified at Configure time and should be accompanied by the necessary
2133 linker additions, eg;
2134 ./config -DOPENSSL_USE_GMP -lgmp
2135 [Geoff Thorpe]
2136
2137 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2138 testing availability of engines with "-t" - the old behaviour is
2139 produced by increasing the feature's verbosity with "-tt".
2140 [Geoff Thorpe]
2141
2142 *) ECDSA routines: under certain error conditions uninitialized BN objects
2143 could be freed. Solution: make sure initialization is performed early
2144 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2145 via PR#459)
2146 [Lutz Jaenicke]
2147
2148 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2149 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2150 software implementations. For DSA and DH, parameter generation can
2151 also be overriden by providing the appropriate method callbacks.
2152 [Geoff Thorpe]
2153
2154 *) Change the "progress" mechanism used in key-generation and
2155 primality testing to functions that take a new BN_GENCB pointer in
2156 place of callback/argument pairs. The new API functions have "_ex"
2157 postfixes and the older functions are reimplemented as wrappers for
2158 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2159 declarations of the old functions to help (graceful) attempts to
2160 migrate to the new functions. Also, the new key-generation API
2161 functions operate on a caller-supplied key-structure and return
2162 success/failure rather than returning a key or NULL - this is to
2163 help make "keygen" another member function of RSA_METHOD etc.
2164
2165 Example for using the new callback interface:
2166
2167 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2168 void *my_arg = ...;
2169 BN_GENCB my_cb;
2170
2171 BN_GENCB_set(&my_cb, my_callback, my_arg);
2172
2173 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2174 /* For the meaning of a, b in calls to my_callback(), see the
2175 * documentation of the function that calls the callback.
2176 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2177 * my_callback should return 1 if it wants BN_is_prime_ex()
2178 * to continue, or 0 to stop.
2179 */
2180
2181 [Geoff Thorpe]
2182
2183 *) Change the ZLIB compression method to be stateful, and make it
2184 available to TLS with the number defined in
2185 draft-ietf-tls-compression-04.txt.
2186 [Richard Levitte]
2187
2188 *) Add the ASN.1 structures and functions for CertificatePair, which
2189 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2190
2191 CertificatePair ::= SEQUENCE {
2192 forward [0] Certificate OPTIONAL,
2193 reverse [1] Certificate OPTIONAL,
2194 -- at least one of the pair shall be present -- }
2195
2196 Also implement the PEM functions to read and write certificate
2197 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2198
2199 This needed to be defined, mostly for the sake of the LDAP
2200 attribute crossCertificatePair, but may prove useful elsewhere as
2201 well.
2202 [Richard Levitte]
2203
2204 *) Make it possible to inhibit symlinking of shared libraries in
2205 Makefile.shared, for Cygwin's sake.
2206 [Richard Levitte]
2207
2208 *) Extend the BIGNUM API by creating a function
2209 void BN_set_negative(BIGNUM *a, int neg);
2210 and a macro that behave like
2211 int BN_is_negative(const BIGNUM *a);
2212
2213 to avoid the need to access 'a->neg' directly in applications.
2214 [Nils Larsch]
2215
2216 *) Implement fast modular reduction for pseudo-Mersenne primes
2217 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2218 EC_GROUP_new_curve_GFp() will now automatically use this
2219 if applicable.
2220 [Nils Larsch <nla@trustcenter.de>]
2221
2222 *) Add new lock type (CRYPTO_LOCK_BN).
2223 [Bodo Moeller]
2224
2225 *) Change the ENGINE framework to automatically load engines
2226 dynamically from specific directories unless they could be
2227 found to already be built in or loaded. Move all the
2228 current engines except for the cryptodev one to a new
2229 directory engines/.
2230 The engines in engines/ are built as shared libraries if
2231 the "shared" options was given to ./Configure or ./config.
2232 Otherwise, they are inserted in libcrypto.a.
2233 /usr/local/ssl/engines is the default directory for dynamic
2234 engines, but that can be overriden at configure time through
2235 the usual use of --prefix and/or --openssldir, and at run
2236 time with the environment variable OPENSSL_ENGINES.
2237 [Geoff Thorpe and Richard Levitte]
2238
2239 *) Add Makefile.shared, a helper makefile to build shared
2240 libraries. Addapt Makefile.org.
2241 [Richard Levitte]
2242
2243 *) Add version info to Win32 DLLs.
2244 [Peter 'Luna' Runestig" <peter@runestig.com>]
2245
2246 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2247 can be added using this API to created arbitrary PKCS#12
2248 files while avoiding the low level API.
2249
2250 New options to PKCS12_create(), key or cert can be NULL and
2251 will then be omitted from the output file. The encryption
2252 algorithm NIDs can be set to -1 for no encryption, the mac
2253 iteration count can be set to 0 to omit the mac.
2254
2255 Enhance pkcs12 utility by making the -nokeys and -nocerts
2256 options work when creating a PKCS#12 file. New option -nomac
2257 to omit the mac, NONE can be set for an encryption algorithm.
2258 New code is modified to use the enhanced PKCS12_create()
2259 instead of the low level API.
2260 [Steve Henson]
2261
2262 *) Extend ASN1 encoder to support indefinite length constructed
2263 encoding. This can output sequences tags and octet strings in
2264 this form. Modify pk7_asn1.c to support indefinite length
2265 encoding. This is experimental and needs additional code to
2266 be useful, such as an ASN1 bio and some enhanced streaming
2267 PKCS#7 code.
2268
2269 Extend template encode functionality so that tagging is passed
2270 down to the template encoder.
2271 [Steve Henson]
2272
2273 *) Let 'openssl req' fail if an argument to '-newkey' is not
2274 recognized instead of using RSA as a default.
2275 [Bodo Moeller]
2276
2277 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2278 As these are not official, they are not included in "ALL";
2279 the "ECCdraft" ciphersuite group alias can be used to select them.
2280 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2281
2282 *) Add ECDH engine support.
2283 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2284
2285 *) Add ECDH in new directory crypto/ecdh/.
2286 [Douglas Stebila (Sun Microsystems Laboratories)]
2287
2288 *) Let BN_rand_range() abort with an error after 100 iterations
2289 without success (which indicates a broken PRNG).
2290 [Bodo Moeller]
2291
2292 *) Change BN_mod_sqrt() so that it verifies that the input value
2293 is really the square of the return value. (Previously,
2294 BN_mod_sqrt would show GIGO behaviour.)
2295 [Bodo Moeller]
2296
2297 *) Add named elliptic curves over binary fields from X9.62, SECG,
2298 and WAP/WTLS; add OIDs that were still missing.
2299
2300 [Sheueling Chang Shantz and Douglas Stebila
2301 (Sun Microsystems Laboratories)]
2302
2303 *) Extend the EC library for elliptic curves over binary fields
2304 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2305 New EC_METHOD:
2306
2307 EC_GF2m_simple_method
2308
2309 New API functions:
2310
2311 EC_GROUP_new_curve_GF2m
2312 EC_GROUP_set_curve_GF2m
2313 EC_GROUP_get_curve_GF2m
2314 EC_POINT_set_affine_coordinates_GF2m
2315 EC_POINT_get_affine_coordinates_GF2m
2316 EC_POINT_set_compressed_coordinates_GF2m
2317
2318 Point compression for binary fields is disabled by default for
2319 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2320 enable it).
2321
2322 As binary polynomials are represented as BIGNUMs, various members
2323 of the EC_GROUP and EC_POINT data structures can be shared
2324 between the implementations for prime fields and binary fields;
2325 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2326 are essentially identical to their ..._GFp counterparts.
2327 (For simplicity, the '..._GFp' prefix has been dropped from
2328 various internal method names.)
2329
2330 An internal 'field_div' method (similar to 'field_mul' and
2331 'field_sqr') has been added; this is used only for binary fields.
2332
2333 [Sheueling Chang Shantz and Douglas Stebila
2334 (Sun Microsystems Laboratories)]
2335
2336 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2337 through methods ('mul', 'precompute_mult').
2338
2339 The generic implementations (now internally called 'ec_wNAF_mul'
2340 and 'ec_wNAF_precomputed_mult') remain the default if these
2341 methods are undefined.
2342
2343 [Sheueling Chang Shantz and Douglas Stebila
2344 (Sun Microsystems Laboratories)]
2345
2346 *) New function EC_GROUP_get_degree, which is defined through
2347 EC_METHOD. For curves over prime fields, this returns the bit
2348 length of the modulus.
2349
2350 [Sheueling Chang Shantz and Douglas Stebila
2351 (Sun Microsystems Laboratories)]
2352
2353 *) New functions EC_GROUP_dup, EC_POINT_dup.
2354 (These simply call ..._new and ..._copy).
2355
2356 [Sheueling Chang Shantz and Douglas Stebila
2357 (Sun Microsystems Laboratories)]
2358
2359 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2360 Polynomials are represented as BIGNUMs (where the sign bit is not
2361 used) in the following functions [macros]:
2362
2363 BN_GF2m_add
2364 BN_GF2m_sub [= BN_GF2m_add]
2365 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2366 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2367 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2368 BN_GF2m_mod_inv
2369 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2370 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2371 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2372 BN_GF2m_cmp [= BN_ucmp]
2373
2374 (Note that only the 'mod' functions are actually for fields GF(2^m).
2375 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2376
2377 For some functions, an the irreducible polynomial defining a
2378 field can be given as an 'unsigned int[]' with strictly
2379 decreasing elements giving the indices of those bits that are set;
2380 i.e., p[] represents the polynomial
2381 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2382 where
2383 p[0] > p[1] > ... > p[k] = 0.
2384 This applies to the following functions:
2385
2386 BN_GF2m_mod_arr
2387 BN_GF2m_mod_mul_arr
2388 BN_GF2m_mod_sqr_arr
2389 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2390 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2391 BN_GF2m_mod_exp_arr
2392 BN_GF2m_mod_sqrt_arr
2393 BN_GF2m_mod_solve_quad_arr
2394 BN_GF2m_poly2arr
2395 BN_GF2m_arr2poly
2396
2397 Conversion can be performed by the following functions:
2398
2399 BN_GF2m_poly2arr
2400 BN_GF2m_arr2poly
2401
2402 bntest.c has additional tests for binary polynomial arithmetic.
2403
2404 Two implementations for BN_GF2m_mod_div() are available.
2405 The default algorithm simply uses BN_GF2m_mod_inv() and
2406 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2407 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2408 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2409
2410 [Sheueling Chang Shantz and Douglas Stebila
2411 (Sun Microsystems Laboratories)]
2412
2413 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2414 functionality is disabled at compile-time.
2415 [Douglas Stebila <douglas.stebila@sun.com>]
2416
2417 *) Change default behaviour of 'openssl asn1parse' so that more
2418 information is visible when viewing, e.g., a certificate:
2419
2420 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2421 mode the content of non-printable OCTET STRINGs is output in a
2422 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2423 avoid the appearance of a printable string.
2424 [Nils Larsch <nla@trustcenter.de>]
2425
2426 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2427 functions
2428 EC_GROUP_set_asn1_flag()
2429 EC_GROUP_get_asn1_flag()
2430 EC_GROUP_set_point_conversion_form()
2431 EC_GROUP_get_point_conversion_form()
2432 These control ASN1 encoding details:
2433 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
2434 has been set to OPENSSL_EC_NAMED_CURVE.
2435 - Points are encoded in uncompressed form by default; options for
2436 asn1_for are as for point2oct, namely
2437 POINT_CONVERSION_COMPRESSED
2438 POINT_CONVERSION_UNCOMPRESSED
2439 POINT_CONVERSION_HYBRID
2440
2441 Also add 'seed' and 'seed_len' members to EC_GROUP with access
2442 functions
2443 EC_GROUP_set_seed()
2444 EC_GROUP_get0_seed()
2445 EC_GROUP_get_seed_len()
2446 This is used only for ASN1 purposes (so far).
2447 [Nils Larsch <nla@trustcenter.de>]
2448
2449 *) Add 'field_type' member to EC_METHOD, which holds the NID
2450 of the appropriate field type OID. The new function
2451 EC_METHOD_get_field_type() returns this value.
2452 [Nils Larsch <nla@trustcenter.de>]
2453
2454 *) Add functions
2455 EC_POINT_point2bn()
2456 EC_POINT_bn2point()
2457 EC_POINT_point2hex()
2458 EC_POINT_hex2point()
2459 providing useful interfaces to EC_POINT_point2oct() and
2460 EC_POINT_oct2point().
2461 [Nils Larsch <nla@trustcenter.de>]
2462
2463 *) Change internals of the EC library so that the functions
2464 EC_GROUP_set_generator()
2465 EC_GROUP_get_generator()
2466 EC_GROUP_get_order()
2467 EC_GROUP_get_cofactor()
2468 are implemented directly in crypto/ec/ec_lib.c and not dispatched
2469 to methods, which would lead to unnecessary code duplication when
2470 adding different types of curves.
2471 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
2472
2473 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
2474 arithmetic, and such that modified wNAFs are generated
2475 (which avoid length expansion in many cases).
2476 [Bodo Moeller]
2477
2478 *) Add a function EC_GROUP_check_discriminant() (defined via
2479 EC_METHOD) that verifies that the curve discriminant is non-zero.
2480
2481 Add a function EC_GROUP_check() that makes some sanity tests
2482 on a EC_GROUP, its generator and order. This includes
2483 EC_GROUP_check_discriminant().
2484 [Nils Larsch <nla@trustcenter.de>]
2485
2486 *) Add ECDSA in new directory crypto/ecdsa/.
2487
2488 Add applications 'openssl ecparam' and 'openssl ecdsa'
2489 (these are based on 'openssl dsaparam' and 'openssl dsa').
2490
2491 ECDSA support is also included in various other files across the
2492 library. Most notably,
2493 - 'openssl req' now has a '-newkey ecdsa:file' option;
2494 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
2495 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
2496 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
2497 them suitable for ECDSA where domain parameters must be
2498 extracted before the specific public key;
2499 - ECDSA engine support has been added.
2500 [Nils Larsch <nla@trustcenter.de>]
2501
2502 *) Include some named elliptic curves, and add OIDs from X9.62,
2503 SECG, and WAP/WTLS. Each curve can be obtained from the new
2504 function
2505 EC_GROUP_new_by_curve_name(),
2506 and the list of available named curves can be obtained with
2507 EC_get_builtin_curves().
2508 Also add a 'curve_name' member to EC_GROUP objects, which can be
2509 accessed via
2510 EC_GROUP_set_curve_name()
2511 EC_GROUP_get_curve_name()
2512 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
2513
2514 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2515 was actually never needed) and in BN_mul(). The removal in BN_mul()
2516 required a small change in bn_mul_part_recursive() and the addition
2517 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2518 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2519 bn_sub_words() and bn_add_words() except they take arrays with
2520 differing sizes.
2521 [Richard Levitte]
2522
2523 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
2524
2525 *) Cleanse PEM buffers before freeing them since they may contain
2526 sensitive data.
2527 [Benjamin Bennett <ben@psc.edu>]
2528
2529 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2530 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2531 authentication-only ciphersuites.
2532 [Bodo Moeller]
2533
2534 *) Since AES128 and AES256 share a single mask bit in the logic of
2535 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2536 kludge to work properly if AES128 is available and AES256 isn't.
2537 [Victor Duchovni]
2538
2539 *) Expand security boundary to match 1.1.1 module.
2540 [Steve Henson]
2541
2542 *) Remove redundant features: hash file source, editing of test vectors
2543 modify fipsld to use external fips_premain.c signature.
2544 [Steve Henson]
2545
2546 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
2547 run algorithm test programs.
2548 [Steve Henson]
2549
2550 *) Make algorithm test programs more tolerant of whitespace.
2551 [Steve Henson]
2552
2553 *) Have SSL/TLS server implementation tolerate "mismatched" record
2554 protocol version while receiving ClientHello even if the
2555 ClientHello is fragmented. (The server can't insist on the
2556 particular protocol version it has chosen before the ServerHello
2557 message has informed the client about his choice.)
2558 [Bodo Moeller]
2559
2560 *) Load error codes if they are not already present instead of using a
2561 static variable. This allows them to be cleanly unloaded and reloaded.
2562 [Steve Henson]
2563
2564 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
2565
2566 *) Introduce limits to prevent malicious keys being able to
2567 cause a denial of service. (CVE-2006-2940)
2568 [Steve Henson, Bodo Moeller]
2569
2570 *) Fix ASN.1 parsing of certain invalid structures that can result
2571 in a denial of service. (CVE-2006-2937) [Steve Henson]
2572
2573 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2574 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2575
2576 *) Fix SSL client code which could crash if connecting to a
2577 malicious SSLv2 server. (CVE-2006-4343)
2578 [Tavis Ormandy and Will Drewry, Google Security Team]
2579
2580 *) Change ciphersuite string processing so that an explicit
2581 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
2582 will no longer include "AES128-SHA"), and any other similar
2583 ciphersuite (same bitmap) from *other* protocol versions (so that
2584 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
2585 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
2586 changes from 0.9.8b and 0.9.8d.
2587 [Bodo Moeller]
2588
2589 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
2590
2591 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2592 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2593
2594 *) Change the Unix randomness entropy gathering to use poll() when
2595 possible instead of select(), since the latter has some
2596 undesirable limitations.
2597 [Darryl Miles via Richard Levitte and Bodo Moeller]
2598
2599 *) Disable rogue ciphersuites:
2600
2601 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2602 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2603 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2604
2605 The latter two were purportedly from
2606 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2607 appear there.
2608
2609 Also deactive the remaining ciphersuites from
2610 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2611 unofficial, and the ID has long expired.
2612 [Bodo Moeller]
2613
2614 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2615 dual-core machines) and other potential thread-safety issues.
2616 [Bodo Moeller]
2617
2618 Changes between 0.9.7i and 0.9.7j [04 May 2006]
2619
2620 *) Adapt fipsld and the build system to link against the validated FIPS
2621 module in FIPS mode.
2622 [Steve Henson]
2623
2624 *) Fixes for VC++ 2005 build under Windows.
2625 [Steve Henson]
2626
2627 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
2628 from a Windows bash shell such as MSYS. It is autodetected from the
2629 "config" script when run from a VC++ environment. Modify standard VC++
2630 build to use fipscanister.o from the GNU make build.
2631 [Steve Henson]
2632
2633 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
2634
2635 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
2636 The value now differs depending on if you build for FIPS or not.
2637 BEWARE! A program linked with a shared FIPSed libcrypto can't be
2638 safely run with a non-FIPSed libcrypto, as it may crash because of
2639 the difference induced by this change.
2640 [Andy Polyakov]
2641
2642 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
2643
2644 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2645 (part of SSL_OP_ALL). This option used to disable the
2646 countermeasure against man-in-the-middle protocol-version
2647 rollback in the SSL 2.0 server implementation, which is a bad
2648 idea. (CVE-2005-2969)
2649
2650 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2651 for Information Security, National Institute of Advanced Industrial
2652 Science and Technology [AIST], Japan)]
2653
2654 *) Minimal support for X9.31 signatures and PSS padding modes. This is
2655 mainly for FIPS compliance and not fully integrated at this stage.
2656 [Steve Henson]
2657
2658 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
2659 the exponentiation using a fixed-length exponent. (Otherwise,
2660 the information leaked through timing could expose the secret key
2661 after many signatures; cf. Bleichenbacher's attack on DSA with
2662 biased k.)
2663 [Bodo Moeller]
2664
2665 *) Make a new fixed-window mod_exp implementation the default for
2666 RSA, DSA, and DH private-key operations so that the sequence of
2667 squares and multiplies and the memory access pattern are
2668 independent of the particular secret key. This will mitigate
2669 cache-timing and potential related attacks.
2670
2671 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
2672 and this is automatically used by BN_mod_exp_mont() if the new flag
2673 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
2674 will use this BN flag for private exponents unless the flag
2675 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
2676 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
2677
2678 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
2679
2680 *) Change the client implementation for SSLv23_method() and
2681 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
2682 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
2683 (Previously, the SSL 2.0 backwards compatible Client Hello
2684 message format would be used even with SSL_OP_NO_SSLv2.)
2685 [Bodo Moeller]
2686
2687 *) Add support for smime-type MIME parameter in S/MIME messages which some
2688 clients need.
2689 [Steve Henson]
2690
2691 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
2692 a threadsafe manner. Modify rsa code to use new function and add calls
2693 to dsa and dh code (which had race conditions before).
2694 [Steve Henson]
2695
2696 *) Include the fixed error library code in the C error file definitions
2697 instead of fixing them up at runtime. This keeps the error code
2698 structures constant.
2699 [Steve Henson]
2700
2701 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
2702
2703 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
2704 OpenSSL 0.9.8.]
2705
2706 *) Fixes for newer kerberos headers. NB: the casts are needed because
2707 the 'length' field is signed on one version and unsigned on another
2708 with no (?) obvious way to tell the difference, without these VC++
2709 complains. Also the "definition" of FAR (blank) is no longer included
2710 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
2711 some needed definitions.
2712 [Steve Henson]
2713
2714 *) Undo Cygwin change.
2715 [Ulf Möller]
2716
2717 *) Added support for proxy certificates according to RFC 3820.
2718 Because they may be a security thread to unaware applications,
2719 they must be explicitely allowed in run-time. See
2720 docs/HOWTO/proxy_certificates.txt for further information.
2721 [Richard Levitte]
2722
2723 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
2724
2725 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
2726 server and client random values. Previously
2727 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
2728 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
2729
2730 This change has negligible security impact because:
2731
2732 1. Server and client random values still have 24 bytes of pseudo random
2733 data.
2734
2735 2. Server and client random values are sent in the clear in the initial
2736 handshake.
2737
2738 3. The master secret is derived using the premaster secret (48 bytes in
2739 size for static RSA ciphersuites) as well as client server and random
2740 values.
2741
2742 The OpenSSL team would like to thank the UK NISCC for bringing this issue
2743 to our attention.
2744
2745 [Stephen Henson, reported by UK NISCC]
2746
2747 *) Use Windows randomness collection on Cygwin.
2748 [Ulf Möller]
2749
2750 *) Fix hang in EGD/PRNGD query when communication socket is closed
2751 prematurely by EGD/PRNGD.
2752 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
2753
2754 *) Prompt for pass phrases when appropriate for PKCS12 input format.
2755 [Steve Henson]
2756
2757 *) Back-port of selected performance improvements from development
2758 branch, as well as improved support for PowerPC platforms.
2759 [Andy Polyakov]
2760
2761 *) Add lots of checks for memory allocation failure, error codes to indicate
2762 failure and freeing up memory if a failure occurs.
2763 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
2764
2765 *) Add new -passin argument to dgst.
2766 [Steve Henson]
2767
2768 *) Perform some character comparisons of different types in X509_NAME_cmp:
2769 this is needed for some certificates that reencode DNs into UTF8Strings
2770 (in violation of RFC3280) and can't or wont issue name rollover
2771 certificates.
2772 [Steve Henson]
2773
2774 *) Make an explicit check during certificate validation to see that
2775 the CA setting in each certificate on the chain is correct. As a
2776 side effect always do the following basic checks on extensions,
2777 not just when there's an associated purpose to the check:
2778
2779 - if there is an unhandled critical extension (unless the user
2780 has chosen to ignore this fault)
2781 - if the path length has been exceeded (if one is set at all)
2782 - that certain extensions fit the associated purpose (if one has
2783 been given)
2784 [Richard Levitte]
2785
2786 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
2787
2788 *) Avoid a race condition when CRLs are checked in a multi threaded
2789 environment. This would happen due to the reordering of the revoked
2790 entries during signature checking and serial number lookup. Now the
2791 encoding is cached and the serial number sort performed under a lock.
2792 Add new STACK function sk_is_sorted().
2793 [Steve Henson]
2794
2795 *) Add Delta CRL to the extension code.
2796 [Steve Henson]
2797
2798 *) Various fixes to s3_pkt.c so alerts are sent properly.
2799 [David Holmes <d.holmes@f5.com>]
2800
2801 *) Reduce the chances of duplicate issuer name and serial numbers (in
2802 violation of RFC3280) using the OpenSSL certificate creation utilities.
2803 This is done by creating a random 64 bit value for the initial serial
2804 number when a serial number file is created or when a self signed
2805 certificate is created using 'openssl req -x509'. The initial serial
2806 number file is created using 'openssl x509 -next_serial' in CA.pl
2807 rather than being initialized to 1.
2808 [Steve Henson]
2809
2810 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
2811
2812 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2813 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
2814 [Joe Orton, Steve Henson]
2815
2816 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
2817 (CVE-2004-0112)
2818 [Joe Orton, Steve Henson]
2819
2820 *) Make it possible to have multiple active certificates with the same
2821 subject in the CA index file. This is done only if the keyword
2822 'unique_subject' is set to 'no' in the main CA section (default
2823 if 'CA_default') of the configuration file. The value is saved
2824 with the database itself in a separate index attribute file,
2825 named like the index file with '.attr' appended to the name.
2826 [Richard Levitte]
2827
2828 *) X509 verify fixes. Disable broken certificate workarounds when
2829 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
2830 keyUsage extension present. Don't accept CRLs with unhandled critical
2831 extensions: since verify currently doesn't process CRL extensions this
2832 rejects a CRL with *any* critical extensions. Add new verify error codes
2833 for these cases.
2834 [Steve Henson]
2835
2836 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
2837 A clarification of RFC2560 will require the use of OCTET STRINGs and
2838 some implementations cannot handle the current raw format. Since OpenSSL
2839 copies and compares OCSP nonces as opaque blobs without any attempt at
2840 parsing them this should not create any compatibility issues.
2841 [Steve Henson]
2842
2843 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
2844 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
2845 this HMAC (and other) operations are several times slower than OpenSSL
2846 < 0.9.7.
2847 [Steve Henson]
2848
2849 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
2850 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2851
2852 *) Use the correct content when signing type "other".
2853 [Steve Henson]
2854
2855 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
2856
2857 *) Fix various bugs revealed by running the NISCC test suite:
2858
2859 Stop out of bounds reads in the ASN1 code when presented with
2860 invalid tags (CVE-2003-0543 and CVE-2003-0544).
2861
2862 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
2863
2864 If verify callback ignores invalid public key errors don't try to check
2865 certificate signature with the NULL public key.
2866
2867 [Steve Henson]
2868
2869 *) New -ignore_err option in ocsp application to stop the server
2870 exiting on the first error in a request.
2871 [Steve Henson]
2872
2873 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2874 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2875 specifications.
2876 [Steve Henson]
2877
2878 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2879 extra data after the compression methods not only for TLS 1.0
2880 but also for SSL 3.0 (as required by the specification).
2881 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2882
2883 *) Change X509_certificate_type() to mark the key as exported/exportable
2884 when it's 512 *bits* long, not 512 bytes.
2885 [Richard Levitte]
2886
2887 *) Change AES_cbc_encrypt() so it outputs exact multiple of
2888 blocks during encryption.
2889 [Richard Levitte]
2890
2891 *) Various fixes to base64 BIO and non blocking I/O. On write
2892 flushes were not handled properly if the BIO retried. On read
2893 data was not being buffered properly and had various logic bugs.
2894 This also affects blocking I/O when the data being decoded is a
2895 certain size.
2896 [Steve Henson]
2897
2898 *) Various S/MIME bugfixes and compatibility changes:
2899 output correct application/pkcs7 MIME type if
2900 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2901 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2902 of files as .eml work). Correctly handle very long lines in MIME
2903 parser.
2904 [Steve Henson]
2905
2906 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2907
2908 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2909 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2910 a protocol version number mismatch like a decryption error
2911 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2912 [Bodo Moeller]
2913
2914 *) Turn on RSA blinding by default in the default implementation
2915 to avoid a timing attack. Applications that don't want it can call
2916 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2917 They would be ill-advised to do so in most cases.
2918 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2919
2920 *) Change RSA blinding code so that it works when the PRNG is not
2921 seeded (in this case, the secret RSA exponent is abused as
2922 an unpredictable seed -- if it is not unpredictable, there
2923 is no point in blinding anyway). Make RSA blinding thread-safe
2924 by remembering the creator's thread ID in rsa->blinding and
2925 having all other threads use local one-time blinding factors
2926 (this requires more computation than sharing rsa->blinding, but
2927 avoids excessive locking; and if an RSA object is not shared
2928 between threads, blinding will still be very fast).
2929 [Bodo Moeller]
2930
2931 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2932 ENGINE as defaults for all supported algorithms irrespective of
2933 the 'flags' parameter. 'flags' is now honoured, so applications
2934 should make sure they are passing it correctly.
2935 [Geoff Thorpe]
2936
2937 *) Target "mingw" now allows native Windows code to be generated in
2938 the Cygwin environment as well as with the MinGW compiler.
2939 [Ulf Moeller]
2940
2941 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2942
2943 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2944 via timing by performing a MAC computation even if incorrrect
2945 block cipher padding has been found. This is a countermeasure
2946 against active attacks where the attacker has to distinguish
2947 between bad padding and a MAC verification error. (CVE-2003-0078)
2948
2949 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2950 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2951 Martin Vuagnoux (EPFL, Ilion)]
2952
2953 *) Make the no-err option work as intended. The intention with no-err
2954 is not to have the whole error stack handling routines removed from
2955 libcrypto, it's only intended to remove all the function name and
2956 reason texts, thereby removing some of the footprint that may not
2957 be interesting if those errors aren't displayed anyway.
2958
2959 NOTE: it's still possible for any application or module to have it's
2960 own set of error texts inserted. The routines are there, just not
2961 used by default when no-err is given.
2962 [Richard Levitte]
2963
2964 *) Add support for FreeBSD on IA64.
2965 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2966
2967 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2968 Kerberos function mit_des_cbc_cksum(). Before this change,
2969 the value returned by DES_cbc_cksum() was like the one from
2970 mit_des_cbc_cksum(), except the bytes were swapped.
2971 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2972
2973 *) Allow an application to disable the automatic SSL chain building.
2974 Before this a rather primitive chain build was always performed in
2975 ssl3_output_cert_chain(): an application had no way to send the
2976 correct chain if the automatic operation produced an incorrect result.
2977
2978 Now the chain builder is disabled if either:
2979
2980 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2981
2982 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2983
2984 The reasoning behind this is that an application would not want the
2985 auto chain building to take place if extra chain certificates are
2986 present and it might also want a means of sending no additional
2987 certificates (for example the chain has two certificates and the
2988 root is omitted).
2989 [Steve Henson]
2990
2991 *) Add the possibility to build without the ENGINE framework.
2992 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2993
2994 *) Under Win32 gmtime() can return NULL: check return value in
2995 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2996 [Steve Henson]
2997
2998 *) DSA routines: under certain error conditions uninitialized BN objects
2999 could be freed. Solution: make sure initialization is performed early
3000 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3001 Nils Larsch <nla@trustcenter.de> via PR#459)
3002 [Lutz Jaenicke]
3003
3004 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3005 checked on reconnect on the client side, therefore session resumption
3006 could still fail with a "ssl session id is different" error. This
3007 behaviour is masked when SSL_OP_ALL is used due to
3008 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3009 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3010 followup to PR #377.
3011 [Lutz Jaenicke]
3012
3013 *) IA-32 assembler support enhancements: unified ELF targets, support
3014 for SCO/Caldera platforms, fix for Cygwin shared build.
3015 [Andy Polyakov]
3016
3017 *) Add support for FreeBSD on sparc64. As a consequence, support for
3018 FreeBSD on non-x86 processors is separate from x86 processors on
3019 the config script, much like the NetBSD support.
3020 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3021
3022 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3023
3024 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3025 OpenSSL 0.9.7.]
3026
3027 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3028 code (06) was taken as the first octet of the session ID and the last
3029 octet was ignored consequently. As a result SSLv2 client side session
3030 caching could not have worked due to the session ID mismatch between
3031 client and server.
3032 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3033 PR #377.
3034 [Lutz Jaenicke]
3035
3036 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3037 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3038 removed entirely.
3039 [Richard Levitte]
3040
3041 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3042 seems that in spite of existing for more than a year, many application
3043 author have done nothing to provide the necessary callbacks, which
3044 means that this particular engine will not work properly anywhere.
3045 This is a very unfortunate situation which forces us, in the name
3046 of usability, to give the hw_ncipher.c a static lock, which is part
3047 of libcrypto.
3048 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3049 appear in 0.9.8 or later. We EXPECT application authors to have
3050 dealt properly with this when 0.9.8 is released (unless we actually
3051 make such changes in the libcrypto locking code that changes will
3052 have to be made anyway).
3053 [Richard Levitte]
3054
3055 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3056 octets have been read, EOF or an error occurs. Without this change
3057 some truncated ASN1 structures will not produce an error.
3058 [Steve Henson]
3059
3060 *) Disable Heimdal support, since it hasn't been fully implemented.
3061 Still give the possibility to force the use of Heimdal, but with
3062 warnings and a request that patches get sent to openssl-dev.
3063 [Richard Levitte]
3064
3065 *) Add the VC-CE target, introduce the WINCE sysname, and add
3066 INSTALL.WCE and appropriate conditionals to make it build.
3067 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3068
3069 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3070 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3071 edit numbers of the version.
3072 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3073
3074 *) Introduce safe string copy and catenation functions
3075 (BUF_strlcpy() and BUF_strlcat()).
3076 [Ben Laurie (CHATS) and Richard Levitte]
3077
3078 *) Avoid using fixed-size buffers for one-line DNs.
3079 [Ben Laurie (CHATS)]
3080
3081 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3082 resizing buffers containing secrets, and use where appropriate.
3083 [Ben Laurie (CHATS)]
3084
3085 *) Avoid using fixed size buffers for configuration file location.
3086 [Ben Laurie (CHATS)]
3087
3088 *) Avoid filename truncation for various CA files.
3089 [Ben Laurie (CHATS)]
3090
3091 *) Use sizeof in preference to magic numbers.
3092 [Ben Laurie (CHATS)]
3093
3094 *) Avoid filename truncation in cert requests.
3095 [Ben Laurie (CHATS)]
3096
3097 *) Add assertions to check for (supposedly impossible) buffer
3098 overflows.
3099 [Ben Laurie (CHATS)]
3100
3101 *) Don't cache truncated DNS entries in the local cache (this could
3102 potentially lead to a spoofing attack).
3103 [Ben Laurie (CHATS)]
3104
3105 *) Fix various buffers to be large enough for hex/decimal
3106 representations in a platform independent manner.
3107 [Ben Laurie (CHATS)]
3108
3109 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3110 resizing buffers containing secrets, and use where appropriate.
3111 [Ben Laurie (CHATS)]
3112
3113 *) Add BIO_indent() to avoid much slightly worrying code to do
3114 indents.
3115 [Ben Laurie (CHATS)]
3116
3117 *) Convert sprintf()/BIO_puts() to BIO_printf().
3118 [Ben Laurie (CHATS)]
3119
3120 *) buffer_gets() could terminate with the buffer only half
3121 full. Fixed.
3122 [Ben Laurie (CHATS)]
3123
3124 *) Add assertions to prevent user-supplied crypto functions from
3125 overflowing internal buffers by having large block sizes, etc.
3126 [Ben Laurie (CHATS)]
3127
3128 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3129 unconditionally).
3130 [Ben Laurie (CHATS)]
3131
3132 *) Eliminate unused copy of key in RC4.
3133 [Ben Laurie (CHATS)]
3134
3135 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3136 [Ben Laurie (CHATS)]
3137
3138 *) Fix off-by-one error in EGD path.
3139 [Ben Laurie (CHATS)]
3140
3141 *) If RANDFILE path is too long, ignore instead of truncating.
3142 [Ben Laurie (CHATS)]
3143
3144 *) Eliminate unused and incorrectly sized X.509 structure
3145 CBCParameter.
3146 [Ben Laurie (CHATS)]
3147
3148 *) Eliminate unused and dangerous function knumber().
3149 [Ben Laurie (CHATS)]
3150
3151 *) Eliminate unused and dangerous structure, KSSL_ERR.
3152 [Ben Laurie (CHATS)]
3153
3154 *) Protect against overlong session ID context length in an encoded
3155 session object. Since these are local, this does not appear to be
3156 exploitable.
3157 [Ben Laurie (CHATS)]
3158
3159 *) Change from security patch (see 0.9.6e below) that did not affect
3160 the 0.9.6 release series:
3161
3162 Remote buffer overflow in SSL3 protocol - an attacker could
3163 supply an oversized master key in Kerberos-enabled versions.
3164 (CVE-2002-0657)
3165 [Ben Laurie (CHATS)]
3166
3167 *) Change the SSL kerb5 codes to match RFC 2712.
3168 [Richard Levitte]
3169
3170 *) Make -nameopt work fully for req and add -reqopt switch.
3171 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3172
3173 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3174 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3175
3176 *) Make sure tests can be performed even if the corresponding algorithms
3177 have been removed entirely. This was also the last step to make
3178 OpenSSL compilable with DJGPP under all reasonable conditions.
3179 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3180
3181 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3182 to allow version independent disabling of normally unselected ciphers,
3183 which may be activated as a side-effect of selecting a single cipher.
3184
3185 (E.g., cipher list string "RSA" enables ciphersuites that are left
3186 out of "ALL" because they do not provide symmetric encryption.
3187 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3188 [Lutz Jaenicke, Bodo Moeller]
3189
3190 *) Add appropriate support for separate platform-dependent build
3191 directories. The recommended way to make a platform-dependent
3192 build directory is the following (tested on Linux), maybe with
3193 some local tweaks:
3194
3195 # Place yourself outside of the OpenSSL source tree. In
3196 # this example, the environment variable OPENSSL_SOURCE
3197 # is assumed to contain the absolute OpenSSL source directory.
3198 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3199 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3200 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3201 mkdir -p `dirname $F`
3202 ln -s $OPENSSL_SOURCE/$F $F
3203 done
3204
3205 To be absolutely sure not to disturb the source tree, a "make clean"
3206 is a good thing. If it isn't successfull, don't worry about it,
3207 it probably means the source directory is very clean.
3208 [Richard Levitte]
3209
3210 *) Make sure any ENGINE control commands make local copies of string
3211 pointers passed to them whenever necessary. Otherwise it is possible
3212 the caller may have overwritten (or deallocated) the original string
3213 data when a later ENGINE operation tries to use the stored values.
3214 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3215
3216 *) Improve diagnostics in file reading and command-line digests.
3217 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3218
3219 *) Add AES modes CFB and OFB to the object database. Correct an
3220 error in AES-CFB decryption.
3221 [Richard Levitte]
3222
3223 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3224 allows existing EVP_CIPHER_CTX structures to be reused after
3225 calling EVP_*Final(). This behaviour is used by encryption
3226 BIOs and some applications. This has the side effect that
3227 applications must explicitly clean up cipher contexts with
3228 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3229 [Steve Henson]
3230
3231 *) Check the values of dna and dnb in bn_mul_recursive before calling
3232 bn_mul_comba (a non zero value means the a or b arrays do not contain
3233 n2 elements) and fallback to bn_mul_normal if either is not zero.
3234 [Steve Henson]
3235
3236 *) Fix escaping of non-ASCII characters when using the -subj option
3237 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3238 [Lutz Jaenicke]
3239
3240 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3241 form for "surname", serialNumber has no short form.
3242 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3243 therefore remove "mail" short name for "internet 7".
3244 The OID for unique identifiers in X509 certificates is
3245 x500UniqueIdentifier, not uniqueIdentifier.
3246 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3247 [Lutz Jaenicke]
3248
3249 *) Add an "init" command to the ENGINE config module and auto initialize
3250 ENGINEs. Without any "init" command the ENGINE will be initialized
3251 after all ctrl commands have been executed on it. If init=1 the
3252 ENGINE is initailized at that point (ctrls before that point are run
3253 on the uninitialized ENGINE and after on the initialized one). If
3254 init=0 then the ENGINE will not be iniatialized at all.
3255 [Steve Henson]
3256
3257 *) Fix the 'app_verify_callback' interface so that the user-defined
3258 argument is actually passed to the callback: In the
3259 SSL_CTX_set_cert_verify_callback() prototype, the callback
3260 declaration has been changed from
3261 int (*cb)()
3262 into
3263 int (*cb)(X509_STORE_CTX *,void *);
3264 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3265 i=s->ctx->app_verify_callback(&ctx)
3266 has been changed into
3267 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3268
3269 To update applications using SSL_CTX_set_cert_verify_callback(),
3270 a dummy argument can be added to their callback functions.
3271 [D. K. Smetters <smetters@parc.xerox.com>]
3272
3273 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3274 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3275
3276 *) Add and OPENSSL_LOAD_CONF define which will cause
3277 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3278 This allows older applications to transparently support certain
3279 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3280 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3281 load the config file and OPENSSL_add_all_algorithms_conf() which will
3282 always load it have also been added.
3283 [Steve Henson]
3284
3285 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3286 Adjust NIDs and EVP layer.
3287 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3288
3289 *) Config modules support in openssl utility.
3290
3291 Most commands now load modules from the config file,
3292 though in a few (such as version) this isn't done
3293 because it couldn't be used for anything.
3294
3295 In the case of ca and req the config file used is
3296 the same as the utility itself: that is the -config
3297 command line option can be used to specify an
3298 alternative file.
3299 [Steve Henson]
3300
3301 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3302 use "openssl_conf" if filename is NULL use default openssl config file.
3303 [Steve Henson]
3304
3305 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3306 config section name. Add a new flag to tolerate a missing config file
3307 and move code to CONF_modules_load_file().
3308 [Steve Henson]
3309
3310 *) Support for crypto accelerator cards from Accelerated Encryption
3311 Processing, www.aep.ie. (Use engine 'aep')
3312 The support was copied from 0.9.6c [engine] and adapted/corrected
3313 to work with the new engine framework.
3314 [AEP Inc. and Richard Levitte]
3315
3316 *) Support for SureWare crypto accelerator cards from Baltimore
3317 Technologies. (Use engine 'sureware')
3318 The support was copied from 0.9.6c [engine] and adapted
3319 to work with the new engine framework.
3320 [Richard Levitte]
3321
3322 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3323 make the newer ENGINE framework commands for the CHIL engine work.
3324 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3325
3326 *) Make it possible to produce shared libraries on ReliantUNIX.
3327 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3328
3329 *) Add the configuration target debug-linux-ppro.
3330 Make 'openssl rsa' use the general key loading routines
3331 implemented in apps.c, and make those routines able to
3332 handle the key format FORMAT_NETSCAPE and the variant
3333 FORMAT_IISSGC.
3334 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3335
3336 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3337 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3338
3339 *) Add -keyform to rsautl, and document -engine.
3340 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3341
3342 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3343 BIO_R_NO_SUCH_FILE error code rather than the generic
3344 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3345 [Ben Laurie]
3346
3347 *) Add new functions
3348 ERR_peek_last_error
3349 ERR_peek_last_error_line
3350 ERR_peek_last_error_line_data.
3351 These are similar to
3352 ERR_peek_error
3353 ERR_peek_error_line
3354 ERR_peek_error_line_data,
3355 but report on the latest error recorded rather than the first one
3356 still in the error queue.
3357 [Ben Laurie, Bodo Moeller]
3358
3359 *) default_algorithms option in ENGINE config module. This allows things
3360 like:
3361 default_algorithms = ALL
3362 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3363 [Steve Henson]
3364
3365 *) Prelminary ENGINE config module.
3366 [Steve Henson]
3367
3368 *) New experimental application configuration code.
3369 [Steve Henson]
3370
3371 *) Change the AES code to follow the same name structure as all other
3372 symmetric ciphers, and behave the same way. Move everything to
3373 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3374 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3375
3376 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3377 [Ben Laurie and Theo de Raadt]
3378
3379 *) Add option to output public keys in req command.
3380 [Massimiliano Pala madwolf@openca.org]
3381
3382 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3383 (up to about 10% better than before for P-192 and P-224).
3384 [Bodo Moeller]
3385
3386 *) New functions/macros
3387
3388 SSL_CTX_set_msg_callback(ctx, cb)
3389 SSL_CTX_set_msg_callback_arg(ctx, arg)
3390 SSL_set_msg_callback(ssl, cb)
3391 SSL_set_msg_callback_arg(ssl, arg)
3392
3393 to request calling a callback function
3394
3395 void cb(int write_p, int version, int content_type,
3396 const void *buf, size_t len, SSL *ssl, void *arg)
3397
3398 whenever a protocol message has been completely received
3399 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3400 protocol version according to which the SSL library interprets
3401 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3402 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3403 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3404 specification (change_cipher_spec(20), alert(21), handshake(22)).
3405 'buf' and 'len' point to the actual message, 'ssl' to the
3406 SSL object, and 'arg' is the application-defined value set by
3407 SSL[_CTX]_set_msg_callback_arg().
3408
3409 'openssl s_client' and 'openssl s_server' have new '-msg' options
3410 to enable a callback that displays all protocol messages.
3411 [Bodo Moeller]
3412
3413 *) Change the shared library support so shared libraries are built as
3414 soon as the corresponding static library is finished, and thereby get
3415 openssl and the test programs linked against the shared library.
3416 This still only happens when the keyword "shard" has been given to
3417 the configuration scripts.
3418
3419 NOTE: shared library support is still an experimental thing, and
3420 backward binary compatibility is still not guaranteed.
3421 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3422
3423 *) Add support for Subject Information Access extension.
3424 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3425
3426 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3427 additional bytes when new memory had to be allocated, not just
3428 when reusing an existing buffer.
3429 [Bodo Moeller]
3430
3431 *) New command line and configuration option 'utf8' for the req command.
3432 This allows field values to be specified as UTF8 strings.
3433 [Steve Henson]
3434
3435 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
3436 runs for the former and machine-readable output for the latter.
3437 [Ben Laurie]
3438
3439 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
3440 of the e-mail address in the DN (i.e., it will go into a certificate
3441 extension only). The new configuration file option 'email_in_dn = no'
3442 has the same effect.
3443 [Massimiliano Pala madwolf@openca.org]
3444
3445 *) Change all functions with names starting with des_ to be starting
3446 with DES_ instead. Add wrappers that are compatible with libdes,
3447 but are named _ossl_old_des_*. Finally, add macros that map the
3448 des_* symbols to the corresponding _ossl_old_des_* if libdes
3449 compatibility is desired. If OpenSSL 0.9.6c compatibility is
3450 desired, the des_* symbols will be mapped to DES_*, with one
3451 exception.
3452
3453 Since we provide two compatibility mappings, the user needs to
3454 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
3455 compatibility is desired. The default (i.e., when that macro
3456 isn't defined) is OpenSSL 0.9.6c compatibility.
3457
3458 There are also macros that enable and disable the support of old
3459 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
3460 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
3461 are defined, the default will apply: to support the old des routines.
3462
3463 In either case, one must include openssl/des.h to get the correct
3464 definitions. Do not try to just include openssl/des_old.h, that
3465 won't work.
3466
3467 NOTE: This is a major break of an old API into a new one. Software
3468 authors are encouraged to switch to the DES_ style functions. Some
3469 time in the future, des_old.h and the libdes compatibility functions
3470 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
3471 default), and then completely removed.
3472 [Richard Levitte]
3473
3474 *) Test for certificates which contain unsupported critical extensions.
3475 If such a certificate is found during a verify operation it is
3476 rejected by default: this behaviour can be overridden by either
3477 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
3478 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
3479 X509_supported_extension() has also been added which returns 1 if a
3480 particular extension is supported.
3481 [Steve Henson]
3482
3483 *) Modify the behaviour of EVP cipher functions in similar way to digests
3484 to retain compatibility with existing code.
3485 [Steve Henson]
3486
3487 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
3488 compatibility with existing code. In particular the 'ctx' parameter does
3489 not have to be to be initialized before the call to EVP_DigestInit() and
3490 it is tidied up after a call to EVP_DigestFinal(). New function
3491 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
3492 EVP_MD_CTX_copy() changed to not require the destination to be
3493 initialized valid and new function EVP_MD_CTX_copy_ex() added which
3494 requires the destination to be valid.
3495
3496 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
3497 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
3498 [Steve Henson]
3499
3500 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
3501 so that complete 'Handshake' protocol structures are kept in memory
3502 instead of overwriting 'msg_type' and 'length' with 'body' data.
3503 [Bodo Moeller]
3504
3505 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
3506 [Massimo Santin via Richard Levitte]
3507
3508 *) Major restructuring to the underlying ENGINE code. This includes
3509 reduction of linker bloat, separation of pure "ENGINE" manipulation
3510 (initialisation, etc) from functionality dealing with implementations
3511 of specific crypto iterfaces. This change also introduces integrated
3512 support for symmetric ciphers and digest implementations - so ENGINEs
3513 can now accelerate these by providing EVP_CIPHER and EVP_MD
3514 implementations of their own. This is detailed in crypto/engine/README
3515 as it couldn't be adequately described here. However, there are a few
3516 API changes worth noting - some RSA, DSA, DH, and RAND functions that
3517 were changed in the original introduction of ENGINE code have now
3518 reverted back - the hooking from this code to ENGINE is now a good
3519 deal more passive and at run-time, operations deal directly with
3520 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
3521 dereferencing through an ENGINE pointer any more. Also, the ENGINE
3522 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
3523 they were not being used by the framework as there is no concept of a
3524 BIGNUM_METHOD and they could not be generalised to the new
3525 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
3526 ENGINE_cpy() has been removed as it cannot be consistently defined in
3527 the new code.
3528 [Geoff Thorpe]
3529
3530 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
3531 [Steve Henson]
3532
3533 *) Change mkdef.pl to sort symbols that get the same entry number,
3534 and make sure the automatically generated functions ERR_load_*
3535 become part of libeay.num as well.
3536 [Richard Levitte]
3537
3538 *) New function SSL_renegotiate_pending(). This returns true once
3539 renegotiation has been requested (either SSL_renegotiate() call
3540 or HelloRequest/ClientHello receveived from the peer) and becomes
3541 false once a handshake has been completed.
3542 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
3543 sends a HelloRequest, but does not ensure that a handshake takes
3544 place. SSL_renegotiate_pending() is useful for checking if the
3545 client has followed the request.)
3546 [Bodo Moeller]
3547
3548 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
3549 By default, clients may request session resumption even during
3550 renegotiation (if session ID contexts permit); with this option,
3551 session resumption is possible only in the first handshake.
3552
3553 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
3554 more bits available for options that should not be part of
3555 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
3556 [Bodo Moeller]
3557
3558 *) Add some demos for certificate and certificate request creation.
3559 [Steve Henson]
3560
3561 *) Make maximum certificate chain size accepted from the peer application
3562 settable (SSL*_get/set_max_cert_list()), as proposed by
3563 "Douglas E. Engert" <deengert@anl.gov>.
3564 [Lutz Jaenicke]
3565
3566 *) Add support for shared libraries for Unixware-7
3567 (Boyd Lynn Gerber <gerberb@zenez.com>).
3568 [Lutz Jaenicke]
3569
3570 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
3571 be done prior to destruction. Use this to unload error strings from
3572 ENGINEs that load their own error strings. NB: This adds two new API
3573 functions to "get" and "set" this destroy handler in an ENGINE.
3574 [Geoff Thorpe]
3575
3576 *) Alter all existing ENGINE implementations (except "openssl" and
3577 "openbsd") to dynamically instantiate their own error strings. This
3578 makes them more flexible to be built both as statically-linked ENGINEs
3579 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
3580 Also, add stub code to each that makes building them as self-contained
3581 shared-libraries easier (see README.ENGINE).
3582 [Geoff Thorpe]
3583
3584 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
3585 implementations into applications that are completely implemented in
3586 self-contained shared-libraries. The "dynamic" ENGINE exposes control
3587 commands that can be used to configure what shared-library to load and
3588 to control aspects of the way it is handled. Also, made an update to
3589 the README.ENGINE file that brings its information up-to-date and
3590 provides some information and instructions on the "dynamic" ENGINE
3591 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
3592 [Geoff Thorpe]
3593
3594 *) Make it possible to unload ranges of ERR strings with a new
3595 "ERR_unload_strings" function.
3596 [Geoff Thorpe]
3597
3598 *) Add a copy() function to EVP_MD.
3599 [Ben Laurie]
3600
3601 *) Make EVP_MD routines take a context pointer instead of just the
3602 md_data void pointer.
3603 [Ben Laurie]
3604
3605 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
3606 that the digest can only process a single chunk of data
3607 (typically because it is provided by a piece of
3608 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
3609 is only going to provide a single chunk of data, and hence the
3610 framework needn't accumulate the data for oneshot drivers.
3611 [Ben Laurie]
3612
3613 *) As with "ERR", make it possible to replace the underlying "ex_data"
3614 functions. This change also alters the storage and management of global
3615 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
3616 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
3617 index counters. The API functions that use this state have been changed
3618 to take a "class_index" rather than pointers to the class's local STACK
3619 and counter, and there is now an API function to dynamically create new
3620 classes. This centralisation allows us to (a) plug a lot of the
3621 thread-safety problems that existed, and (b) makes it possible to clean
3622 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
3623 such data would previously have always leaked in application code and
3624 workarounds were in place to make the memory debugging turn a blind eye
3625 to it. Application code that doesn't use this new function will still
3626 leak as before, but their memory debugging output will announce it now
3627 rather than letting it slide.
3628
3629 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
3630 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
3631 has a return value to indicate success or failure.
3632 [Geoff Thorpe]
3633
3634 *) Make it possible to replace the underlying "ERR" functions such that the
3635 global state (2 LHASH tables and 2 locks) is only used by the "default"
3636 implementation. This change also adds two functions to "get" and "set"
3637 the implementation prior to it being automatically set the first time
3638 any other ERR function takes place. Ie. an application can call "get",
3639 pass the return value to a module it has just loaded, and that module
3640 can call its own "set" function using that value. This means the
3641 module's "ERR" operations will use (and modify) the error state in the
3642 application and not in its own statically linked copy of OpenSSL code.
3643 [Geoff Thorpe]
3644
3645 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
3646 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
3647 the operation, and provides a more encapsulated way for external code
3648 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
3649 to use these functions rather than manually incrementing the counts.
3650
3651 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
3652 [Geoff Thorpe]
3653
3654 *) Add EVP test program.
3655 [Ben Laurie]
3656
3657 *) Add symmetric cipher support to ENGINE. Expect the API to change!
3658 [Ben Laurie]
3659
3660 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
3661 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
3662 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
3663 These allow a CRL to be built without having to access X509_CRL fields
3664 directly. Modify 'ca' application to use new functions.
3665 [Steve Henson]
3666
3667 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
3668 bug workarounds. Rollback attack detection is a security feature.
3669 The problem will only arise on OpenSSL servers when TLSv1 is not
3670 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
3671 Software authors not wanting to support TLSv1 will have special reasons
3672 for their choice and can explicitly enable this option.
3673 [Bodo Moeller, Lutz Jaenicke]
3674
3675 *) Rationalise EVP so it can be extended: don't include a union of
3676 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
3677 (similar to those existing for EVP_CIPHER_CTX).
3678 Usage example:
3679
3680 EVP_MD_CTX md;
3681
3682 EVP_MD_CTX_init(&md); /* new function call */
3683 EVP_DigestInit(&md, EVP_sha1());
3684 EVP_DigestUpdate(&md, in, len);
3685 EVP_DigestFinal(&md, out, NULL);
3686 EVP_MD_CTX_cleanup(&md); /* new function call */
3687
3688 [Ben Laurie]
3689
3690 *) Make DES key schedule conform to the usual scheme, as well as
3691 correcting its structure. This means that calls to DES functions
3692 now have to pass a pointer to a des_key_schedule instead of a
3693 plain des_key_schedule (which was actually always a pointer
3694 anyway): E.g.,
3695
3696 des_key_schedule ks;
3697
3698 des_set_key_checked(..., &ks);
3699 des_ncbc_encrypt(..., &ks, ...);
3700
3701 (Note that a later change renames 'des_...' into 'DES_...'.)
3702 [Ben Laurie]
3703
3704 *) Initial reduction of linker bloat: the use of some functions, such as
3705 PEM causes large amounts of unused functions to be linked in due to
3706 poor organisation. For example pem_all.c contains every PEM function
3707 which has a knock on effect of linking in large amounts of (unused)
3708 ASN1 code. Grouping together similar functions and splitting unrelated
3709 functions prevents this.
3710 [Steve Henson]
3711
3712 *) Cleanup of EVP macros.
3713 [Ben Laurie]
3714
3715 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
3716 correct _ecb suffix.
3717 [Ben Laurie]
3718
3719 *) Add initial OCSP responder support to ocsp application. The
3720 revocation information is handled using the text based index
3721 use by the ca application. The responder can either handle
3722 requests generated internally, supplied in files (for example
3723 via a CGI script) or using an internal minimal server.
3724 [Steve Henson]
3725
3726 *) Add configuration choices to get zlib compression for TLS.
3727 [Richard Levitte]
3728
3729 *) Changes to Kerberos SSL for RFC 2712 compliance:
3730 1. Implemented real KerberosWrapper, instead of just using
3731 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
3732 2. Implemented optional authenticator field of KerberosWrapper.
3733
3734 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
3735 and authenticator structs; see crypto/krb5/.
3736
3737 Generalized Kerberos calls to support multiple Kerberos libraries.
3738 [Vern Staats <staatsvr@asc.hpc.mil>,
3739 Jeffrey Altman <jaltman@columbia.edu>
3740 via Richard Levitte]
3741
3742 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
3743 already does with RSA. testdsa.h now has 'priv_key/pub_key'
3744 values for each of the key sizes rather than having just
3745 parameters (and 'speed' generating keys each time).
3746 [Geoff Thorpe]
3747
3748 *) Speed up EVP routines.
3749 Before:
3750 encrypt
3751 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
3752 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
3753 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
3754 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
3755 decrypt
3756 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
3757 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
3758 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
3759 After:
3760 encrypt
3761 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
3762 decrypt
3763 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
3764 [Ben Laurie]
3765
3766 *) Added the OS2-EMX target.
3767 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
3768
3769 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
3770 to support NCONF routines in extension code. New function CONF_set_nconf()
3771 to allow functions which take an NCONF to also handle the old LHASH
3772 structure: this means that the old CONF compatible routines can be
3773 retained (in particular wrt extensions) without having to duplicate the
3774 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
3775 [Steve Henson]
3776
3777 *) Enhance the general user interface with mechanisms for inner control
3778 and with possibilities to have yes/no kind of prompts.
3779 [Richard Levitte]
3780
3781 *) Change all calls to low level digest routines in the library and
3782 applications to use EVP. Add missing calls to HMAC_cleanup() and
3783 don't assume HMAC_CTX can be copied using memcpy().
3784 [Verdon Walker <VWalker@novell.com>, Steve Henson]
3785
3786 *) Add the possibility to control engines through control names but with
3787 arbitrary arguments instead of just a string.
3788 Change the key loaders to take a UI_METHOD instead of a callback
3789 function pointer. NOTE: this breaks binary compatibility with earlier
3790 versions of OpenSSL [engine].
3791 Adapt the nCipher code for these new conditions and add a card insertion
3792 callback.
3793 [Richard Levitte]
3794
3795 *) Enhance the general user interface with mechanisms to better support
3796 dialog box interfaces, application-defined prompts, the possibility
3797 to use defaults (for example default passwords from somewhere else)
3798 and interrupts/cancellations.
3799 [Richard Levitte]
3800
3801 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
3802 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
3803 [Steve Henson]
3804
3805 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
3806 tidy up some unnecessarily weird code in 'sk_new()').
3807 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
3808
3809 *) Change the key loading routines for ENGINEs to use the same kind
3810 callback (pem_password_cb) as all other routines that need this
3811 kind of callback.
3812 [Richard Levitte]
3813
3814 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
3815 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
3816 than this minimum value is recommended.
3817 [Lutz Jaenicke]
3818
3819 *) New random seeder for OpenVMS, using the system process statistics
3820 that are easily reachable.
3821 [Richard Levitte]
3822
3823 *) Windows apparently can't transparently handle global
3824 variables defined in DLLs. Initialisations such as:
3825
3826 const ASN1_ITEM *it = &ASN1_INTEGER_it;
3827
3828 wont compile. This is used by the any applications that need to
3829 declare their own ASN1 modules. This was fixed by adding the option
3830 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
3831 needed for static libraries under Win32.
3832 [Steve Henson]
3833
3834 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
3835 setting of purpose and trust fields. New X509_STORE trust and
3836 purpose functions and tidy up setting in other SSL functions.
3837 [Steve Henson]
3838
3839 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
3840 structure. These are inherited by X509_STORE_CTX when it is
3841 initialised. This allows various defaults to be set in the
3842 X509_STORE structure (such as flags for CRL checking and custom
3843 purpose or trust settings) for functions which only use X509_STORE_CTX
3844 internally such as S/MIME.
3845
3846 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
3847 trust settings if they are not set in X509_STORE. This allows X509_STORE
3848 purposes and trust (in S/MIME for example) to override any set by default.
3849
3850 Add command line options for CRL checking to smime, s_client and s_server
3851 applications.
3852 [Steve Henson]
3853
3854 *) Initial CRL based revocation checking. If the CRL checking flag(s)
3855 are set then the CRL is looked up in the X509_STORE structure and
3856 its validity and signature checked, then if the certificate is found
3857 in the CRL the verify fails with a revoked error.
3858
3859 Various new CRL related callbacks added to X509_STORE_CTX structure.
3860
3861 Command line options added to 'verify' application to support this.
3862
3863 This needs some additional work, such as being able to handle multiple
3864 CRLs with different times, extension based lookup (rather than just
3865 by subject name) and ultimately more complete V2 CRL extension
3866 handling.
3867 [Steve Henson]
3868
3869 *) Add a general user interface API (crypto/ui/). This is designed
3870 to replace things like des_read_password and friends (backward
3871 compatibility functions using this new API are provided).
3872 The purpose is to remove prompting functions from the DES code
3873 section as well as provide for prompting through dialog boxes in
3874 a window system and the like.
3875 [Richard Levitte]
3876
3877 *) Add "ex_data" support to ENGINE so implementations can add state at a
3878 per-structure level rather than having to store it globally.
3879 [Geoff]
3880
3881 *) Make it possible for ENGINE structures to be copied when retrieved by
3882 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
3883 This causes the "original" ENGINE structure to act like a template,
3884 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
3885 operational state can be localised to each ENGINE structure, despite the
3886 fact they all share the same "methods". New ENGINE structures returned in
3887 this case have no functional references and the return value is the single
3888 structural reference. This matches the single structural reference returned
3889 by ENGINE_by_id() normally, when it is incremented on the pre-existing
3890 ENGINE structure.
3891 [Geoff]
3892
3893 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3894 needs to match any other type at all we need to manually clear the
3895 tag cache.
3896 [Steve Henson]
3897
3898 *) Changes to the "openssl engine" utility to include;
3899 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3900 about an ENGINE's available control commands.
3901 - executing control commands from command line arguments using the
3902 '-pre' and '-post' switches. '-post' is only used if '-t' is
3903 specified and the ENGINE is successfully initialised. The syntax for
3904 the individual commands are colon-separated, for example;
3905 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3906 [Geoff]
3907
3908 *) New dynamic control command support for ENGINEs. ENGINEs can now
3909 declare their own commands (numbers), names (strings), descriptions,
3910 and input types for run-time discovery by calling applications. A
3911 subset of these commands are implicitly classed as "executable"
3912 depending on their input type, and only these can be invoked through
3913 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3914 can be based on user input, config files, etc). The distinction is
3915 that "executable" commands cannot return anything other than a boolean
3916 result and can only support numeric or string input, whereas some
3917 discoverable commands may only be for direct use through
3918 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3919 pointers, or other custom uses. The "executable" commands are to
3920 support parameterisations of ENGINE behaviour that can be
3921 unambiguously defined by ENGINEs and used consistently across any
3922 OpenSSL-based application. Commands have been added to all the
3923 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3924 control over shared-library paths without source code alterations.
3925 [Geoff]
3926
3927 *) Changed all ENGINE implementations to dynamically allocate their
3928 ENGINEs rather than declaring them statically. Apart from this being
3929 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3930 this also allows the implementations to compile without using the
3931 internal engine_int.h header.
3932 [Geoff]
3933
3934 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3935 'const' value. Any code that should be able to modify a RAND_METHOD
3936 should already have non-const pointers to it (ie. they should only
3937 modify their own ones).
3938 [Geoff]
3939
3940 *) Made a variety of little tweaks to the ENGINE code.
3941 - "atalla" and "ubsec" string definitions were moved from header files
3942 to C code. "nuron" string definitions were placed in variables
3943 rather than hard-coded - allowing parameterisation of these values
3944 later on via ctrl() commands.
3945 - Removed unused "#if 0"'d code.
3946 - Fixed engine list iteration code so it uses ENGINE_free() to release
3947 structural references.
3948 - Constified the RAND_METHOD element of ENGINE structures.
3949 - Constified various get/set functions as appropriate and added
3950 missing functions (including a catch-all ENGINE_cpy that duplicates
3951 all ENGINE values onto a new ENGINE except reference counts/state).
3952 - Removed NULL parameter checks in get/set functions. Setting a method
3953 or function to NULL is a way of cancelling out a previously set
3954 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3955 and doesn't justify the extra error symbols and code.
3956 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3957 flags from engine_int.h to engine.h.
3958 - Changed prototypes for ENGINE handler functions (init(), finish(),
3959 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3960 [Geoff]
3961
3962 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3963 to the algorithm using long division. The binary algorithm can be
3964 used only if the modulus is odd. On 32-bit systems, it is faster
3965 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3966 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3967 up to 450 bits. In 64-bit environments, the binary algorithm
3968 appears to be advantageous for much longer moduli; here we use it
3969 for moduli up to 2048 bits.
3970 [Bodo Moeller]
3971
3972 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3973 could not support the combine flag in choice fields.
3974 [Steve Henson]
3975
3976 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3977 extensions from a certificate request to the certificate.
3978 [Steve Henson]
3979
3980 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3981 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3982 file: this allows the display of the certificate about to be
3983 signed to be customised, to allow certain fields to be included
3984 or excluded and extension details. The old system didn't display
3985 multicharacter strings properly, omitted fields not in the policy
3986 and couldn't display additional details such as extensions.
3987 [Steve Henson]
3988
3989 *) Function EC_POINTs_mul for multiple scalar multiplication
3990 of an arbitrary number of elliptic curve points
3991 \sum scalars[i]*points[i],
3992 optionally including the generator defined for the EC_GROUP:
3993 scalar*generator + \sum scalars[i]*points[i].
3994
3995 EC_POINT_mul is a simple wrapper function for the typical case
3996 that the point list has just one item (besides the optional
3997 generator).
3998 [Bodo Moeller]
3999
4000 *) First EC_METHODs for curves over GF(p):
4001
4002 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4003 operations and provides various method functions that can also
4004 operate with faster implementations of modular arithmetic.
4005
4006 EC_GFp_mont_method() reuses most functions that are part of
4007 EC_GFp_simple_method, but uses Montgomery arithmetic.
4008
4009 [Bodo Moeller; point addition and point doubling
4010 implementation directly derived from source code provided by
4011 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4012
4013 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4014 crypto/ec/ec_lib.c):
4015
4016 Curves are EC_GROUP objects (with an optional group generator)
4017 based on EC_METHODs that are built into the library.
4018
4019 Points are EC_POINT objects based on EC_GROUP objects.
4020
4021 Most of the framework would be able to handle curves over arbitrary
4022 finite fields, but as there are no obvious types for fields other
4023 than GF(p), some functions are limited to that for now.
4024 [Bodo Moeller]
4025
4026 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4027 that the file contains a complete HTTP response.
4028 [Richard Levitte]
4029
4030 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4031 change the def and num file printf format specifier from "%-40sXXX"
4032 to "%-39s XXX". The latter will always guarantee a space after the
4033 field while the former will cause them to run together if the field
4034 is 40 of more characters long.
4035 [Steve Henson]
4036
4037 *) Constify the cipher and digest 'method' functions and structures
4038 and modify related functions to take constant EVP_MD and EVP_CIPHER
4039 pointers.
4040 [Steve Henson]
4041
4042 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4043 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4044 [Bodo Moeller]
4045
4046 *) Modify EVP_Digest*() routines so they now return values. Although the
4047 internal software routines can never fail additional hardware versions
4048 might.
4049 [Steve Henson]
4050
4051 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4052
4053 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4054 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4055
4056 ASN1 error codes
4057 ERR_R_NESTED_ASN1_ERROR
4058 ...
4059 ERR_R_MISSING_ASN1_EOS
4060 were 4 .. 9, conflicting with
4061 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4062 ...
4063 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4064 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4065
4066 Add new error code 'ERR_R_INTERNAL_ERROR'.
4067 [Bodo Moeller]
4068
4069 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4070 suffices.
4071 [Bodo Moeller]
4072
4073 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4074 sets the subject name for a new request or supersedes the
4075 subject name in a given request. Formats that can be parsed are
4076 'CN=Some Name, OU=myOU, C=IT'
4077 and
4078 'CN=Some Name/OU=myOU/C=IT'.
4079
4080 Add options '-batch' and '-verbose' to 'openssl req'.
4081 [Massimiliano Pala <madwolf@hackmasters.net>]
4082
4083 *) Introduce the possibility to access global variables through
4084 functions on platform were that's the best way to handle exporting
4085 global variables in shared libraries. To enable this functionality,
4086 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4087 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4088 is normally done by Configure or something similar).
4089
4090 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4091 in the source file (foo.c) like this:
4092
4093 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4094 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4095
4096 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4097 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4098
4099 OPENSSL_DECLARE_GLOBAL(int,foo);
4100 #define foo OPENSSL_GLOBAL_REF(foo)
4101 OPENSSL_DECLARE_GLOBAL(double,bar);
4102 #define bar OPENSSL_GLOBAL_REF(bar)
4103
4104 The #defines are very important, and therefore so is including the
4105 header file everywhere where the defined globals are used.
4106
4107 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4108 of ASN.1 items, but that structure is a bit different.
4109
4110 The largest change is in util/mkdef.pl which has been enhanced with
4111 better and easier to understand logic to choose which symbols should
4112 go into the Windows .def files as well as a number of fixes and code
4113 cleanup (among others, algorithm keywords are now sorted
4114 lexicographically to avoid constant rewrites).
4115 [Richard Levitte]
4116
4117 *) In BN_div() keep a copy of the sign of 'num' before writing the
4118 result to 'rm' because if rm==num the value will be overwritten
4119 and produce the wrong result if 'num' is negative: this caused
4120 problems with BN_mod() and BN_nnmod().
4121 [Steve Henson]
4122
4123 *) Function OCSP_request_verify(). This checks the signature on an
4124 OCSP request and verifies the signer certificate. The signer
4125 certificate is just checked for a generic purpose and OCSP request
4126 trust settings.
4127 [Steve Henson]
4128
4129 *) Add OCSP_check_validity() function to check the validity of OCSP
4130 responses. OCSP responses are prepared in real time and may only
4131 be a few seconds old. Simply checking that the current time lies
4132 between thisUpdate and nextUpdate max reject otherwise valid responses
4133 caused by either OCSP responder or client clock inaccuracy. Instead
4134 we allow thisUpdate and nextUpdate to fall within a certain period of
4135 the current time. The age of the response can also optionally be
4136 checked. Two new options -validity_period and -status_age added to
4137 ocsp utility.
4138 [Steve Henson]
4139
4140 *) If signature or public key algorithm is unrecognized print out its
4141 OID rather that just UNKNOWN.
4142 [Steve Henson]
4143
4144 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4145 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4146 ID to be generated from the issuer certificate alone which can then be
4147 passed to OCSP_id_issuer_cmp().
4148 [Steve Henson]
4149
4150 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4151 ASN1 modules to export functions returning ASN1_ITEM pointers
4152 instead of the ASN1_ITEM structures themselves. This adds several
4153 new macros which allow the underlying ASN1 function/structure to
4154 be accessed transparently. As a result code should not use ASN1_ITEM
4155 references directly (such as &X509_it) but instead use the relevant
4156 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4157 use of the new ASN1 code on platforms where exporting structures
4158 is problematical (for example in shared libraries) but exporting
4159 functions returning pointers to structures is not.
4160 [Steve Henson]
4161
4162 *) Add support for overriding the generation of SSL/TLS session IDs.
4163 These callbacks can be registered either in an SSL_CTX or per SSL.
4164 The purpose of this is to allow applications to control, if they wish,
4165 the arbitrary values chosen for use as session IDs, particularly as it
4166 can be useful for session caching in multiple-server environments. A
4167 command-line switch for testing this (and any client code that wishes
4168 to use such a feature) has been added to "s_server".
4169 [Geoff Thorpe, Lutz Jaenicke]
4170
4171 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4172 of the form '#if defined(...) || defined(...) || ...' and
4173 '#if !defined(...) && !defined(...) && ...'. This also avoids
4174 the growing number of special cases it was previously handling.
4175 [Richard Levitte]
4176
4177 *) Make all configuration macros available for application by making
4178 sure they are available in opensslconf.h, by giving them names starting
4179 with "OPENSSL_" to avoid conflicts with other packages and by making
4180 sure e_os2.h will cover all platform-specific cases together with
4181 opensslconf.h.
4182 Additionally, it is now possible to define configuration/platform-
4183 specific names (called "system identities"). In the C code, these
4184 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4185 macro with the name beginning with "OPENSSL_SYS_", which is determined
4186 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4187 what is available.
4188 [Richard Levitte]
4189
4190 *) New option -set_serial to 'req' and 'x509' this allows the serial
4191 number to use to be specified on the command line. Previously self
4192 signed certificates were hard coded with serial number 0 and the
4193 CA options of 'x509' had to use a serial number in a file which was
4194 auto incremented.
4195 [Steve Henson]
4196
4197 *) New options to 'ca' utility to support V2 CRL entry extensions.
4198 Currently CRL reason, invalidity date and hold instruction are
4199 supported. Add new CRL extensions to V3 code and some new objects.
4200 [Steve Henson]
4201
4202 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4203 disable standard block padding (aka PKCS#5 padding) in the EVP
4204 API, which was previously mandatory. This means that the data is
4205 not padded in any way and so the total length much be a multiple
4206 of the block size, otherwise an error occurs.
4207 [Steve Henson]
4208
4209 *) Initial (incomplete) OCSP SSL support.
4210 [Steve Henson]
4211
4212 *) New function OCSP_parse_url(). This splits up a URL into its host,
4213 port and path components: primarily to parse OCSP URLs. New -url
4214 option to ocsp utility.
4215 [Steve Henson]
4216
4217 *) New nonce behavior. The return value of OCSP_check_nonce() now
4218 reflects the various checks performed. Applications can decide
4219 whether to tolerate certain situations such as an absent nonce
4220 in a response when one was present in a request: the ocsp application
4221 just prints out a warning. New function OCSP_add1_basic_nonce()
4222 this is to allow responders to include a nonce in a response even if
4223 the request is nonce-less.
4224 [Steve Henson]
4225
4226 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4227 skipped when using openssl x509 multiple times on a single input file,
4228 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4229 [Bodo Moeller]
4230
4231 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4232 set string type: to handle setting ASN1_TIME structures. Fix ca
4233 utility to correctly initialize revocation date of CRLs.
4234 [Steve Henson]
4235
4236 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4237 the clients preferred ciphersuites and rather use its own preferences.
4238 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4239 Internet Explorer by ensuring unchanged hash method during stepup.
4240 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4241 [Lutz Jaenicke]
4242
4243 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4244 to aes and add a new 'exist' option to print out symbols that don't
4245 appear to exist.
4246 [Steve Henson]
4247
4248 *) Additional options to ocsp utility to allow flags to be set and
4249 additional certificates supplied.
4250 [Steve Henson]
4251
4252 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4253 OCSP client a number of certificate to only verify the response
4254 signature against.
4255 [Richard Levitte]
4256
4257 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4258 handle the new API. Currently only ECB, CBC modes supported. Add new
4259 AES OIDs.
4260
4261 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4262 Encryption Standard (AES) Ciphersuites for Transport Layer
4263 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4264 not enabled by default and were not part of the "ALL" ciphersuite
4265 alias because they were not yet official; they could be
4266 explicitly requested by specifying the "AESdraft" ciphersuite
4267 group alias. In the final release of OpenSSL 0.9.7, the group
4268 alias is called "AES" and is part of "ALL".)
4269 [Ben Laurie, Steve Henson, Bodo Moeller]
4270
4271 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4272 request to response.
4273 [Steve Henson]
4274
4275 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4276 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4277 extract information from a certificate request. OCSP_response_create()
4278 creates a response and optionally adds a basic response structure.
4279 OCSP_basic_add1_status() adds a complete single response to a basic
4280 response and returns the OCSP_SINGLERESP structure just added (to allow
4281 extensions to be included for example). OCSP_basic_add1_cert() adds a
4282 certificate to a basic response and OCSP_basic_sign() signs a basic
4283 response with various flags. New helper functions ASN1_TIME_check()
4284 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4285 (converts ASN1_TIME to GeneralizedTime).
4286 [Steve Henson]
4287
4288 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4289 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4290 structure from a certificate. X509_pubkey_digest() digests the public_key
4291 contents: this is used in various key identifiers.
4292 [Steve Henson]
4293
4294 *) Make sk_sort() tolerate a NULL argument.
4295 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4296
4297 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4298 passed by the function are trusted implicitly. If any of them signed the
4299 response then it is assumed to be valid and is not verified.
4300 [Steve Henson]
4301
4302 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4303 to data. This was previously part of the PKCS7 ASN1 code. This
4304 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4305 [Steve Henson, reported by Kenneth R. Robinette
4306 <support@securenetterm.com>]
4307
4308 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4309 routines: without these tracing memory leaks is very painful.
4310 Fix leaks in PKCS12 and PKCS7 routines.
4311 [Steve Henson]
4312
4313 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4314 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4315 effectively meant GeneralizedTime would never be used. Now it
4316 is initialised to -1 but X509_time_adj() now has to check the value
4317 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4318 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4319 [Steve Henson, reported by Kenneth R. Robinette
4320 <support@securenetterm.com>]
4321
4322 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4323 result in a zero length in the ASN1_INTEGER structure which was
4324 not consistent with the structure when d2i_ASN1_INTEGER() was used
4325 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4326 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4327 where it did not print out a minus for negative ASN1_INTEGER.
4328 [Steve Henson]
4329
4330 *) Add summary printout to ocsp utility. The various functions which
4331 convert status values to strings have been renamed to:
4332 OCSP_response_status_str(), OCSP_cert_status_str() and
4333 OCSP_crl_reason_str() and are no longer static. New options
4334 to verify nonce values and to disable verification. OCSP response
4335 printout format cleaned up.
4336 [Steve Henson]
4337
4338 *) Add additional OCSP certificate checks. These are those specified
4339 in RFC2560. This consists of two separate checks: the CA of the
4340 certificate being checked must either be the OCSP signer certificate
4341 or the issuer of the OCSP signer certificate. In the latter case the
4342 OCSP signer certificate must contain the OCSP signing extended key
4343 usage. This check is performed by attempting to match the OCSP
4344 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4345 in the OCSP_CERTID structures of the response.
4346 [Steve Henson]
4347
4348 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4349 and related routines. This uses the standard OpenSSL certificate
4350 verify routines to perform initial checks (just CA validity) and
4351 to obtain the certificate chain. Then additional checks will be
4352 performed on the chain. Currently the root CA is checked to see
4353 if it is explicitly trusted for OCSP signing. This is used to set
4354 a root CA as a global signing root: that is any certificate that
4355 chains to that CA is an acceptable OCSP signing certificate.
4356 [Steve Henson]
4357
4358 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4359 extensions from a separate configuration file.
4360 As when reading extensions from the main configuration file,
4361 the '-extensions ...' option may be used for specifying the
4362 section to use.
4363 [Massimiliano Pala <madwolf@comune.modena.it>]
4364
4365 *) New OCSP utility. Allows OCSP requests to be generated or
4366 read. The request can be sent to a responder and the output
4367 parsed, outputed or printed in text form. Not complete yet:
4368 still needs to check the OCSP response validity.
4369 [Steve Henson]
4370
4371 *) New subcommands for 'openssl ca':
4372 'openssl ca -status <serial>' prints the status of the cert with
4373 the given serial number (according to the index file).
4374 'openssl ca -updatedb' updates the expiry status of certificates
4375 in the index file.
4376 [Massimiliano Pala <madwolf@comune.modena.it>]
4377
4378 *) New '-newreq-nodes' command option to CA.pl. This is like
4379 '-newreq', but calls 'openssl req' with the '-nodes' option
4380 so that the resulting key is not encrypted.
4381 [Damien Miller <djm@mindrot.org>]
4382
4383 *) New configuration for the GNU Hurd.
4384 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4385
4386 *) Initial code to implement OCSP basic response verify. This
4387 is currently incomplete. Currently just finds the signer's
4388 certificate and verifies the signature on the response.
4389 [Steve Henson]
4390
4391 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4392 value of OPENSSLDIR. This is available via the new '-d' option
4393 to 'openssl version', and is also included in 'openssl version -a'.
4394 [Bodo Moeller]
4395
4396 *) Allowing defining memory allocation callbacks that will be given
4397 file name and line number information in additional arguments
4398 (a const char* and an int). The basic functionality remains, as
4399 well as the original possibility to just replace malloc(),
4400 realloc() and free() by functions that do not know about these
4401 additional arguments. To register and find out the current
4402 settings for extended allocation functions, the following
4403 functions are provided:
4404
4405 CRYPTO_set_mem_ex_functions
4406 CRYPTO_set_locked_mem_ex_functions
4407 CRYPTO_get_mem_ex_functions
4408 CRYPTO_get_locked_mem_ex_functions
4409
4410 These work the same way as CRYPTO_set_mem_functions and friends.
4411 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4412 extended allocation function is enabled.
4413 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4414 a conventional allocation function is enabled.
4415 [Richard Levitte, Bodo Moeller]
4416
4417 *) Finish off removing the remaining LHASH function pointer casts.
4418 There should no longer be any prototype-casting required when using
4419 the LHASH abstraction, and any casts that remain are "bugs". See
4420 the callback types and macros at the head of lhash.h for details
4421 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4422 [Geoff Thorpe]
4423
4424 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4425 If /dev/[u]random devices are not available or do not return enough
4426 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4427 be queried.
4428 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4429 /etc/entropy will be queried once each in this sequence, quering stops
4430 when enough entropy was collected without querying more sockets.
4431 [Lutz Jaenicke]
4432
4433 *) Change the Unix RAND_poll() variant to be able to poll several
4434 random devices, as specified by DEVRANDOM, until a sufficient amount
4435 of data has been collected. We spend at most 10 ms on each file
4436 (select timeout) and read in non-blocking mode. DEVRANDOM now
4437 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
4438 (previously it was just the string "/dev/urandom"), so on typical
4439 platforms the 10 ms delay will never occur.
4440 Also separate out the Unix variant to its own file, rand_unix.c.
4441 For VMS, there's a currently-empty rand_vms.c.
4442 [Richard Levitte]
4443
4444 *) Move OCSP client related routines to ocsp_cl.c. These
4445 provide utility functions which an application needing
4446 to issue a request to an OCSP responder and analyse the
4447 response will typically need: as opposed to those which an
4448 OCSP responder itself would need which will be added later.
4449
4450 OCSP_request_sign() signs an OCSP request with an API similar
4451 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
4452 response. OCSP_response_get1_basic() extracts basic response
4453 from response. OCSP_resp_find_status(): finds and extracts status
4454 information from an OCSP_CERTID structure (which will be created
4455 when the request structure is built). These are built from lower
4456 level functions which work on OCSP_SINGLERESP structures but
4457 wont normally be used unless the application wishes to examine
4458 extensions in the OCSP response for example.
4459
4460 Replace nonce routines with a pair of functions.
4461 OCSP_request_add1_nonce() adds a nonce value and optionally
4462 generates a random value. OCSP_check_nonce() checks the
4463 validity of the nonce in an OCSP response.
4464 [Steve Henson]
4465
4466 *) Change function OCSP_request_add() to OCSP_request_add0_id().
4467 This doesn't copy the supplied OCSP_CERTID and avoids the
4468 need to free up the newly created id. Change return type
4469 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
4470 This can then be used to add extensions to the request.
4471 Deleted OCSP_request_new(), since most of its functionality
4472 is now in OCSP_REQUEST_new() (and the case insensitive name
4473 clash) apart from the ability to set the request name which
4474 will be added elsewhere.
4475 [Steve Henson]
4476
4477 *) Update OCSP API. Remove obsolete extensions argument from
4478 various functions. Extensions are now handled using the new
4479 OCSP extension code. New simple OCSP HTTP function which
4480 can be used to send requests and parse the response.
4481 [Steve Henson]
4482
4483 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
4484 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
4485 uses the special reorder version of SET OF to sort the attributes
4486 and reorder them to match the encoded order. This resolves a long
4487 standing problem: a verify on a PKCS7 structure just after signing
4488 it used to fail because the attribute order did not match the
4489 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
4490 it uses the received order. This is necessary to tolerate some broken
4491 software that does not order SET OF. This is handled by encoding
4492 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
4493 to produce the required SET OF.
4494 [Steve Henson]
4495
4496 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
4497 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
4498 files to get correct declarations of the ASN.1 item variables.
4499 [Richard Levitte]
4500
4501 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
4502 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
4503 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
4504 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
4505 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
4506 ASN1_ITEM and no wrapper functions.
4507 [Steve Henson]
4508
4509 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
4510 replace the old function pointer based I/O routines. Change most of
4511 the *_d2i_bio() and *_d2i_fp() functions to use these.
4512 [Steve Henson]
4513
4514 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
4515 lines, recognice more "algorithms" that can be deselected, and make
4516 it complain about algorithm deselection that isn't recognised.
4517 [Richard Levitte]
4518
4519 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
4520 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
4521 to use new functions. Add NO_ASN1_OLD which can be set to remove
4522 some old style ASN1 functions: this can be used to determine if old
4523 code will still work when these eventually go away.
4524 [Steve Henson]
4525
4526 *) New extension functions for OCSP structures, these follow the
4527 same conventions as certificates and CRLs.
4528 [Steve Henson]
4529
4530 *) New function X509V3_add1_i2d(). This automatically encodes and
4531 adds an extension. Its behaviour can be customised with various
4532 flags to append, replace or delete. Various wrappers added for
4533 certifcates and CRLs.
4534 [Steve Henson]
4535
4536 *) Fix to avoid calling the underlying ASN1 print routine when
4537 an extension cannot be parsed. Correct a typo in the
4538 OCSP_SERVICELOC extension. Tidy up print OCSP format.
4539 [Steve Henson]
4540
4541 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
4542 entries for variables.
4543 [Steve Henson]
4544
4545 *) Add functionality to apps/openssl.c for detecting locking
4546 problems: As the program is single-threaded, all we have
4547 to do is register a locking callback using an array for
4548 storing which locks are currently held by the program.
4549 [Bodo Moeller]
4550
4551 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
4552 SSL_get_ex_data_X509_STORE_idx(), which is used in
4553 ssl_verify_cert_chain() and thus can be called at any time
4554 during TLS/SSL handshakes so that thread-safety is essential.
4555 Unfortunately, the ex_data design is not at all suited
4556 for multi-threaded use, so it probably should be abolished.
4557 [Bodo Moeller]
4558
4559 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
4560 [Broadcom, tweaked and integrated by Geoff Thorpe]
4561
4562 *) Move common extension printing code to new function
4563 X509V3_print_extensions(). Reorganise OCSP print routines and
4564 implement some needed OCSP ASN1 functions. Add OCSP extensions.
4565 [Steve Henson]
4566
4567 *) New function X509_signature_print() to remove duplication in some
4568 print routines.
4569 [Steve Henson]
4570
4571 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
4572 set (this was treated exactly the same as SET OF previously). This
4573 is used to reorder the STACK representing the structure to match the
4574 encoding. This will be used to get round a problem where a PKCS7
4575 structure which was signed could not be verified because the STACK
4576 order did not reflect the encoded order.
4577 [Steve Henson]
4578
4579 *) Reimplement the OCSP ASN1 module using the new code.
4580 [Steve Henson]
4581
4582 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
4583 for its ASN1 operations. The old style function pointers still exist
4584 for now but they will eventually go away.
4585 [Steve Henson]
4586
4587 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
4588 completely replaces the old ASN1 functionality with a table driven
4589 encoder and decoder which interprets an ASN1_ITEM structure describing
4590 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
4591 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
4592 has also been converted to the new form.
4593 [Steve Henson]
4594
4595 *) Change BN_mod_exp_recp so that negative moduli are tolerated
4596 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
4597 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
4598 for negative moduli.
4599 [Bodo Moeller]
4600
4601 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
4602 of not touching the result's sign bit.
4603 [Bodo Moeller]
4604
4605 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
4606 set.
4607 [Bodo Moeller]
4608
4609 *) Changed the LHASH code to use prototypes for callbacks, and created
4610 macros to declare and implement thin (optionally static) functions
4611 that provide type-safety and avoid function pointer casting for the
4612 type-specific callbacks.
4613 [Geoff Thorpe]
4614
4615 *) Added Kerberos Cipher Suites to be used with TLS, as written in
4616 RFC 2712.
4617 [Veers Staats <staatsvr@asc.hpc.mil>,
4618 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
4619
4620 *) Reformat the FAQ so the different questions and answers can be divided
4621 in sections depending on the subject.
4622 [Richard Levitte]
4623
4624 *) Have the zlib compression code load ZLIB.DLL dynamically under
4625 Windows.
4626 [Richard Levitte]
4627
4628 *) New function BN_mod_sqrt for computing square roots modulo a prime
4629 (using the probabilistic Tonelli-Shanks algorithm unless
4630 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
4631 be handled deterministically).
4632 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4633
4634 *) Make BN_mod_inverse faster by explicitly handling small quotients
4635 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
4636 512 bits], about 30% for larger ones [1024 or 2048 bits].)
4637 [Bodo Moeller]
4638
4639 *) New function BN_kronecker.
4640 [Bodo Moeller]
4641
4642 *) Fix BN_gcd so that it works on negative inputs; the result is
4643 positive unless both parameters are zero.
4644 Previously something reasonably close to an infinite loop was
4645 possible because numbers could be growing instead of shrinking
4646 in the implementation of Euclid's algorithm.
4647 [Bodo Moeller]
4648
4649 *) Fix BN_is_word() and BN_is_one() macros to take into account the
4650 sign of the number in question.
4651
4652 Fix BN_is_word(a,w) to work correctly for w == 0.
4653
4654 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
4655 because its test if the absolute value of 'a' equals 'w'.
4656 Note that BN_abs_is_word does *not* handle w == 0 reliably;
4657 it exists mostly for use in the implementations of BN_is_zero(),
4658 BN_is_one(), and BN_is_word().
4659 [Bodo Moeller]
4660
4661 *) New function BN_swap.
4662 [Bodo Moeller]
4663
4664 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
4665 the exponentiation functions are more likely to produce reasonable
4666 results on negative inputs.
4667 [Bodo Moeller]
4668
4669 *) Change BN_mod_mul so that the result is always non-negative.
4670 Previously, it could be negative if one of the factors was negative;
4671 I don't think anyone really wanted that behaviour.
4672 [Bodo Moeller]
4673
4674 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
4675 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
4676 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
4677 and add new functions:
4678
4679 BN_nnmod
4680 BN_mod_sqr
4681 BN_mod_add
4682 BN_mod_add_quick
4683 BN_mod_sub
4684 BN_mod_sub_quick
4685 BN_mod_lshift1
4686 BN_mod_lshift1_quick
4687 BN_mod_lshift
4688 BN_mod_lshift_quick
4689
4690 These functions always generate non-negative results.
4691
4692 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
4693 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
4694
4695 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
4696 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
4697 be reduced modulo m.
4698 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4699
4700 #if 0
4701 The following entry accidentily appeared in the CHANGES file
4702 distributed with OpenSSL 0.9.7. The modifications described in
4703 it do *not* apply to OpenSSL 0.9.7.
4704
4705 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4706 was actually never needed) and in BN_mul(). The removal in BN_mul()
4707 required a small change in bn_mul_part_recursive() and the addition
4708 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4709 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4710 bn_sub_words() and bn_add_words() except they take arrays with
4711 differing sizes.
4712 [Richard Levitte]
4713 #endif
4714
4715 *) In 'openssl passwd', verify passwords read from the terminal
4716 unless the '-salt' option is used (which usually means that
4717 verification would just waste user's time since the resulting
4718 hash is going to be compared with some given password hash)
4719 or the new '-noverify' option is used.
4720
4721 This is an incompatible change, but it does not affect
4722 non-interactive use of 'openssl passwd' (passwords on the command
4723 line, '-stdin' option, '-in ...' option) and thus should not
4724 cause any problems.
4725 [Bodo Moeller]
4726
4727 *) Remove all references to RSAref, since there's no more need for it.
4728 [Richard Levitte]
4729
4730 *) Make DSO load along a path given through an environment variable
4731 (SHLIB_PATH) with shl_load().
4732 [Richard Levitte]
4733
4734 *) Constify the ENGINE code as a result of BIGNUM constification.
4735 Also constify the RSA code and most things related to it. In a
4736 few places, most notable in the depth of the ASN.1 code, ugly
4737 casts back to non-const were required (to be solved at a later
4738 time)
4739 [Richard Levitte]
4740
4741 *) Make it so the openssl application has all engines loaded by default.
4742 [Richard Levitte]
4743
4744 *) Constify the BIGNUM routines a little more.
4745 [Richard Levitte]
4746
4747 *) Add the following functions:
4748
4749 ENGINE_load_cswift()
4750 ENGINE_load_chil()
4751 ENGINE_load_atalla()
4752 ENGINE_load_nuron()
4753 ENGINE_load_builtin_engines()
4754
4755 That way, an application can itself choose if external engines that
4756 are built-in in OpenSSL shall ever be used or not. The benefit is
4757 that applications won't have to be linked with libdl or other dso
4758 libraries unless it's really needed.
4759
4760 Changed 'openssl engine' to load all engines on demand.
4761 Changed the engine header files to avoid the duplication of some
4762 declarations (they differed!).
4763 [Richard Levitte]
4764
4765 *) 'openssl engine' can now list capabilities.
4766 [Richard Levitte]
4767
4768 *) Better error reporting in 'openssl engine'.
4769 [Richard Levitte]
4770
4771 *) Never call load_dh_param(NULL) in s_server.
4772 [Bodo Moeller]
4773
4774 *) Add engine application. It can currently list engines by name and
4775 identity, and test if they are actually available.
4776 [Richard Levitte]
4777
4778 *) Improve RPM specification file by forcing symbolic linking and making
4779 sure the installed documentation is also owned by root.root.
4780 [Damien Miller <djm@mindrot.org>]
4781
4782 *) Give the OpenSSL applications more possibilities to make use of
4783 keys (public as well as private) handled by engines.
4784 [Richard Levitte]
4785
4786 *) Add OCSP code that comes from CertCo.
4787 [Richard Levitte]
4788
4789 *) Add VMS support for the Rijndael code.
4790 [Richard Levitte]
4791
4792 *) Added untested support for Nuron crypto accelerator.
4793 [Ben Laurie]
4794
4795 *) Add support for external cryptographic devices. This code was
4796 previously distributed separately as the "engine" branch.
4797 [Geoff Thorpe, Richard Levitte]
4798
4799 *) Rework the filename-translation in the DSO code. It is now possible to
4800 have far greater control over how a "name" is turned into a filename
4801 depending on the operating environment and any oddities about the
4802 different shared library filenames on each system.
4803 [Geoff Thorpe]
4804
4805 *) Support threads on FreeBSD-elf in Configure.
4806 [Richard Levitte]
4807
4808 *) Fix for SHA1 assembly problem with MASM: it produces
4809 warnings about corrupt line number information when assembling
4810 with debugging information. This is caused by the overlapping
4811 of two sections.
4812 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
4813
4814 *) NCONF changes.
4815 NCONF_get_number() has no error checking at all. As a replacement,
4816 NCONF_get_number_e() is defined (_e for "error checking") and is
4817 promoted strongly. The old NCONF_get_number is kept around for
4818 binary backward compatibility.
4819 Make it possible for methods to load from something other than a BIO,
4820 by providing a function pointer that is given a name instead of a BIO.
4821 For example, this could be used to load configuration data from an
4822 LDAP server.
4823 [Richard Levitte]
4824
4825 *) Fix for non blocking accept BIOs. Added new I/O special reason
4826 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
4827 with non blocking I/O was not possible because no retry code was
4828 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
4829 this case.
4830 [Steve Henson]
4831
4832 *) Added the beginnings of Rijndael support.
4833 [Ben Laurie]
4834
4835 *) Fix for bug in DirectoryString mask setting. Add support for
4836 X509_NAME_print_ex() in 'req' and X509_print_ex() function
4837 to allow certificate printing to more controllable, additional
4838 'certopt' option to 'x509' to allow new printing options to be
4839 set.
4840 [Steve Henson]
4841
4842 *) Clean old EAY MD5 hack from e_os.h.
4843 [Richard Levitte]
4844
4845 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
4846
4847 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4848 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4849 [Joe Orton, Steve Henson]
4850
4851 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
4852
4853 *) Fix additional bug revealed by the NISCC test suite:
4854
4855 Stop bug triggering large recursion when presented with
4856 certain ASN.1 tags (CVE-2003-0851)
4857 [Steve Henson]
4858
4859 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
4860
4861 *) Fix various bugs revealed by running the NISCC test suite:
4862
4863 Stop out of bounds reads in the ASN1 code when presented with
4864 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4865
4866 If verify callback ignores invalid public key errors don't try to check
4867 certificate signature with the NULL public key.
4868
4869 [Steve Henson]
4870
4871 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4872 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4873 specifications.
4874 [Steve Henson]
4875
4876 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4877 extra data after the compression methods not only for TLS 1.0
4878 but also for SSL 3.0 (as required by the specification).
4879 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4880
4881 *) Change X509_certificate_type() to mark the key as exported/exportable
4882 when it's 512 *bits* long, not 512 bytes.
4883 [Richard Levitte]
4884
4885 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
4886
4887 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4888 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4889 a protocol version number mismatch like a decryption error
4890 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4891 [Bodo Moeller]
4892
4893 *) Turn on RSA blinding by default in the default implementation
4894 to avoid a timing attack. Applications that don't want it can call
4895 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4896 They would be ill-advised to do so in most cases.
4897 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4898
4899 *) Change RSA blinding code so that it works when the PRNG is not
4900 seeded (in this case, the secret RSA exponent is abused as
4901 an unpredictable seed -- if it is not unpredictable, there
4902 is no point in blinding anyway). Make RSA blinding thread-safe
4903 by remembering the creator's thread ID in rsa->blinding and
4904 having all other threads use local one-time blinding factors
4905 (this requires more computation than sharing rsa->blinding, but
4906 avoids excessive locking; and if an RSA object is not shared
4907 between threads, blinding will still be very fast).
4908 [Bodo Moeller]
4909
4910 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4911
4912 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4913 via timing by performing a MAC computation even if incorrrect
4914 block cipher padding has been found. This is a countermeasure
4915 against active attacks where the attacker has to distinguish
4916 between bad padding and a MAC verification error. (CVE-2003-0078)
4917
4918 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4919 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4920 Martin Vuagnoux (EPFL, Ilion)]
4921
4922 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4923
4924 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4925 memory from it's contents. This is done with a counter that will
4926 place alternating values in each byte. This can be used to solve
4927 two issues: 1) the removal of calls to memset() by highly optimizing
4928 compilers, and 2) cleansing with other values than 0, since those can
4929 be read through on certain media, for example a swap space on disk.
4930 [Geoff Thorpe]
4931
4932 *) Bugfix: client side session caching did not work with external caching,
4933 because the session->cipher setting was not restored when reloading
4934 from the external cache. This problem was masked, when
4935 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4936 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4937 [Lutz Jaenicke]
4938
4939 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4940 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4941 [Zeev Lieber <zeev-l@yahoo.com>]
4942
4943 *) Undo an undocumented change introduced in 0.9.6e which caused
4944 repeated calls to OpenSSL_add_all_ciphers() and
4945 OpenSSL_add_all_digests() to be ignored, even after calling
4946 EVP_cleanup().
4947 [Richard Levitte]
4948
4949 *) Change the default configuration reader to deal with last line not
4950 being properly terminated.
4951 [Richard Levitte]
4952
4953 *) Change X509_NAME_cmp() so it applies the special rules on handling
4954 DN values that are of type PrintableString, as well as RDNs of type
4955 emailAddress where the value has the type ia5String.
4956 [stefank@valicert.com via Richard Levitte]
4957
4958 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4959 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4960 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4961 the bitwise-OR of the two for use by the majority of applications
4962 wanting this behaviour, and update the docs. The documented
4963 behaviour and actual behaviour were inconsistent and had been
4964 changing anyway, so this is more a bug-fix than a behavioural
4965 change.
4966 [Geoff Thorpe, diagnosed by Nadav Har'El]
4967
4968 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4969 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4970 [Bodo Moeller]
4971
4972 *) Fix initialization code race conditions in
4973 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4974 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4975 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4976 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4977 ssl2_get_cipher_by_char(),
4978 ssl3_get_cipher_by_char().
4979 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4980
4981 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4982 the cached sessions are flushed, as the remove_cb() might use ex_data
4983 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4984 (see [openssl.org #212]).
4985 [Geoff Thorpe, Lutz Jaenicke]
4986
4987 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4988 length, instead of the encoding length to d2i_ASN1_OBJECT.
4989 [Steve Henson]
4990
4991 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4992
4993 *) [In 0.9.6g-engine release:]
4994 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4995 [Lynn Gazis <lgazis@rainbow.com>]
4996
4997 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4998
4999 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5000 and get fix the header length calculation.
5001 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5002 Alon Kantor <alonk@checkpoint.com> (and others),
5003 Steve Henson]
5004
5005 *) Use proper error handling instead of 'assertions' in buffer
5006 overflow checks added in 0.9.6e. This prevents DoS (the
5007 assertions could call abort()).
5008 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5009
5010 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5011
5012 *) Add various sanity checks to asn1_get_length() to reject
5013 the ASN1 length bytes if they exceed sizeof(long), will appear
5014 negative or the content length exceeds the length of the
5015 supplied buffer.
5016 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5017
5018 *) Fix cipher selection routines: ciphers without encryption had no flags
5019 for the cipher strength set and where therefore not handled correctly
5020 by the selection routines (PR #130).
5021 [Lutz Jaenicke]
5022
5023 *) Fix EVP_dsa_sha macro.
5024 [Nils Larsch]
5025
5026 *) New option
5027 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5028 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5029 that was added in OpenSSL 0.9.6d.
5030
5031 As the countermeasure turned out to be incompatible with some
5032 broken SSL implementations, the new option is part of SSL_OP_ALL.
5033 SSL_OP_ALL is usually employed when compatibility with weird SSL
5034 implementations is desired (e.g. '-bugs' option to 's_client' and
5035 's_server'), so the new option is automatically set in many
5036 applications.
5037 [Bodo Moeller]
5038
5039 *) Changes in security patch:
5040
5041 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5042 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5043 Air Force Materiel Command, USAF, under agreement number
5044 F30602-01-2-0537.
5045
5046 *) Add various sanity checks to asn1_get_length() to reject
5047 the ASN1 length bytes if they exceed sizeof(long), will appear
5048 negative or the content length exceeds the length of the
5049 supplied buffer. (CVE-2002-0659)
5050 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5051
5052 *) Assertions for various potential buffer overflows, not known to
5053 happen in practice.
5054 [Ben Laurie (CHATS)]
5055
5056 *) Various temporary buffers to hold ASCII versions of integers were
5057 too small for 64 bit platforms. (CVE-2002-0655)
5058 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5059
5060 *) Remote buffer overflow in SSL3 protocol - an attacker could
5061 supply an oversized session ID to a client. (CVE-2002-0656)
5062 [Ben Laurie (CHATS)]
5063
5064 *) Remote buffer overflow in SSL2 protocol - an attacker could
5065 supply an oversized client master key. (CVE-2002-0656)
5066 [Ben Laurie (CHATS)]
5067
5068 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5069
5070 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5071 encoded as NULL) with id-dsa-with-sha1.
5072 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5073
5074 *) Check various X509_...() return values in apps/req.c.
5075 [Nils Larsch <nla@trustcenter.de>]
5076
5077 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5078 an end-of-file condition would erronously be flagged, when the CRLF
5079 was just at the end of a processed block. The bug was discovered when
5080 processing data through a buffering memory BIO handing the data to a
5081 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5082 <ptsekov@syntrex.com> and Nedelcho Stanev.
5083 [Lutz Jaenicke]
5084
5085 *) Implement a countermeasure against a vulnerability recently found
5086 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5087 before application data chunks to avoid the use of known IVs
5088 with data potentially chosen by the attacker.
5089 [Bodo Moeller]
5090
5091 *) Fix length checks in ssl3_get_client_hello().
5092 [Bodo Moeller]
5093
5094 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5095 to prevent ssl3_read_internal() from incorrectly assuming that
5096 ssl3_read_bytes() found application data while handshake
5097 processing was enabled when in fact s->s3->in_read_app_data was
5098 merely automatically cleared during the initial handshake.
5099 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5100
5101 *) Fix object definitions for Private and Enterprise: they were not
5102 recognized in their shortname (=lowercase) representation. Extend
5103 obj_dat.pl to issue an error when using undefined keywords instead
5104 of silently ignoring the problem (Svenning Sorensen
5105 <sss@sss.dnsalias.net>).
5106 [Lutz Jaenicke]
5107
5108 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5109 generators, i.e. generators other than 2 and 5. (Previously, the
5110 code did not properly initialise the 'add' and 'rem' values to
5111 BN_generate_prime().)
5112
5113 In the new general case, we do not insist that 'generator' is
5114 actually a primitive root: This requirement is rather pointless;
5115 a generator of the order-q subgroup is just as good, if not
5116 better.
5117 [Bodo Moeller]
5118
5119 *) Map new X509 verification errors to alerts. Discovered and submitted by
5120 Tom Wu <tom@arcot.com>.
5121 [Lutz Jaenicke]
5122
5123 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5124 returning non-zero before the data has been completely received
5125 when using non-blocking I/O.
5126 [Bodo Moeller; problem pointed out by John Hughes]
5127
5128 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5129 [Ben Laurie, Lutz Jaenicke]
5130
5131 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5132 Yoram Zahavi <YoramZ@gilian.com>).
5133 [Lutz Jaenicke]
5134
5135 *) Add information about CygWin 1.3 and on, and preserve proper
5136 configuration for the versions before that.
5137 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5138
5139 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5140 check whether we deal with a copy of a session and do not delete from
5141 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5142 <izhar@checkpoint.com>.
5143 [Lutz Jaenicke]
5144
5145 *) Do not store session data into the internal session cache, if it
5146 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5147 flag is set). Proposed by Aslam <aslam@funk.com>.
5148 [Lutz Jaenicke]
5149
5150 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5151 value is 0.
5152 [Richard Levitte]
5153
5154 *) [In 0.9.6d-engine release:]
5155 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5156 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5157
5158 *) Add the configuration target linux-s390x.
5159 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5160
5161 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5162 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5163 variable as an indication that a ClientHello message has been
5164 received. As the flag value will be lost between multiple
5165 invocations of ssl3_accept when using non-blocking I/O, the
5166 function may not be aware that a handshake has actually taken
5167 place, thus preventing a new session from being added to the
5168 session cache.
5169
5170 To avoid this problem, we now set s->new_session to 2 instead of
5171 using a local variable.
5172 [Lutz Jaenicke, Bodo Moeller]
5173
5174 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5175 if the SSL_R_LENGTH_MISMATCH error is detected.
5176 [Geoff Thorpe, Bodo Moeller]
5177
5178 *) New 'shared_ldflag' column in Configure platform table.
5179 [Richard Levitte]
5180
5181 *) Fix EVP_CIPHER_mode macro.
5182 ["Dan S. Camper" <dan@bti.net>]
5183
5184 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5185 type, we must throw them away by setting rr->length to 0.
5186 [D P Chang <dpc@qualys.com>]
5187
5188 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5189
5190 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5191 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5192 worked incorrectly for those cases where range = 10..._2 and
5193 3*range is two bits longer than range.)
5194 [Bodo Moeller]
5195
5196 *) Only add signing time to PKCS7 structures if it is not already
5197 present.
5198 [Steve Henson]
5199
5200 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5201 OBJ_ld_ce should be OBJ_id_ce.
5202 Also some ip-pda OIDs in crypto/objects/objects.txt were
5203 incorrect (cf. RFC 3039).
5204 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5205
5206 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5207 returns early because it has nothing to do.
5208 [Andy Schneider <andy.schneider@bjss.co.uk>]
5209
5210 *) [In 0.9.6c-engine release:]
5211 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5212 [Andy Schneider <andy.schneider@bjss.co.uk>]
5213
5214 *) [In 0.9.6c-engine release:]
5215 Add support for Cryptographic Appliance's keyserver technology.
5216 (Use engine 'keyclient')
5217 [Cryptographic Appliances and Geoff Thorpe]
5218
5219 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5220 is called via tools/c89.sh because arguments have to be
5221 rearranged (all '-L' options must appear before the first object
5222 modules).
5223 [Richard Shapiro <rshapiro@abinitio.com>]
5224
5225 *) [In 0.9.6c-engine release:]
5226 Add support for Broadcom crypto accelerator cards, backported
5227 from 0.9.7.
5228 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5229
5230 *) [In 0.9.6c-engine release:]
5231 Add support for SureWare crypto accelerator cards from
5232 Baltimore Technologies. (Use engine 'sureware')
5233 [Baltimore Technologies and Mark Cox]
5234
5235 *) [In 0.9.6c-engine release:]
5236 Add support for crypto accelerator cards from Accelerated
5237 Encryption Processing, www.aep.ie. (Use engine 'aep')
5238 [AEP Inc. and Mark Cox]
5239
5240 *) Add a configuration entry for gcc on UnixWare.
5241 [Gary Benson <gbenson@redhat.com>]
5242
5243 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5244 messages are stored in a single piece (fixed-length part and
5245 variable-length part combined) and fix various bugs found on the way.
5246 [Bodo Moeller]
5247
5248 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5249 instead. BIO_gethostbyname() does not know what timeouts are
5250 appropriate, so entries would stay in cache even when they have
5251 become invalid.
5252 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5253
5254 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5255 faced with a pathologically small ClientHello fragment that does
5256 not contain client_version: Instead of aborting with an error,
5257 simply choose the highest available protocol version (i.e.,
5258 TLS 1.0 unless it is disabled). In practice, ClientHello
5259 messages are never sent like this, but this change gives us
5260 strictly correct behaviour at least for TLS.
5261 [Bodo Moeller]
5262
5263 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5264 never resets s->method to s->ctx->method when called from within
5265 one of the SSL handshake functions.
5266 [Bodo Moeller; problem pointed out by Niko Baric]
5267
5268 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5269 (sent using the client's version number) if client_version is
5270 smaller than the protocol version in use. Also change
5271 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5272 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5273 the client will at least see that alert.
5274 [Bodo Moeller]
5275
5276 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5277 correctly.
5278 [Bodo Moeller]
5279
5280 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5281 client receives HelloRequest while in a handshake.
5282 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5283
5284 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5285 should end in 'break', not 'goto end' which circuments various
5286 cleanups done in state SSL_ST_OK. But session related stuff
5287 must be disabled for SSL_ST_OK in the case that we just sent a
5288 HelloRequest.
5289
5290 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5291 before just sending a HelloRequest.
5292 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5293
5294 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5295 reveal whether illegal block cipher padding was found or a MAC
5296 verification error occured. (Neither SSLerr() codes nor alerts
5297 are directly visible to potential attackers, but the information
5298 may leak via logfiles.)
5299
5300 Similar changes are not required for the SSL 2.0 implementation
5301 because the number of padding bytes is sent in clear for SSL 2.0,
5302 and the extra bytes are just ignored. However ssl/s2_pkt.c
5303 failed to verify that the purported number of padding bytes is in
5304 the legal range.
5305 [Bodo Moeller]
5306
5307 *) Add OpenUNIX-8 support including shared libraries
5308 (Boyd Lynn Gerber <gerberb@zenez.com>).
5309 [Lutz Jaenicke]
5310
5311 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5312 'wristwatch attack' using huge encoding parameters (cf.
5313 James H. Manger's CRYPTO 2001 paper). Note that the
5314 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5315 encoding parameters and hence was not vulnerable.
5316 [Bodo Moeller]
5317
5318 *) BN_sqr() bug fix.
5319 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5320
5321 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5322 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5323 followed by modular reduction.
5324 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5325
5326 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5327 equivalent based on BN_pseudo_rand() instead of BN_rand().
5328 [Bodo Moeller]
5329
5330 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5331 This function was broken, as the check for a new client hello message
5332 to handle SGC did not allow these large messages.
5333 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5334 [Lutz Jaenicke]
5335
5336 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5337 [Lutz Jaenicke]
5338
5339 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5340 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5341 [Lutz Jaenicke]
5342
5343 *) Rework the configuration and shared library support for Tru64 Unix.
5344 The configuration part makes use of modern compiler features and
5345 still retains old compiler behavior for those that run older versions
5346 of the OS. The shared library support part includes a variant that
5347 uses the RPATH feature, and is available through the special
5348 configuration target "alpha-cc-rpath", which will never be selected
5349 automatically.
5350 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5351
5352 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5353 with the same message size as in ssl3_get_certificate_request().
5354 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5355 messages might inadvertently be reject as too long.
5356 [Petr Lampa <lampa@fee.vutbr.cz>]
5357
5358 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5359 [Andy Polyakov]
5360
5361 *) Modified SSL library such that the verify_callback that has been set
5362 specificly for an SSL object with SSL_set_verify() is actually being
5363 used. Before the change, a verify_callback set with this function was
5364 ignored and the verify_callback() set in the SSL_CTX at the time of
5365 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5366 to allow the necessary settings.
5367 [Lutz Jaenicke]
5368
5369 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5370 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5371 done automatically (in contradiction to the requirements of the C
5372 standard). This made problems when used from OpenSSH.
5373 [Lutz Jaenicke]
5374
5375 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5376 dh->length and always used
5377
5378 BN_rand_range(priv_key, dh->p).
5379
5380 BN_rand_range() is not necessary for Diffie-Hellman, and this
5381 specific range makes Diffie-Hellman unnecessarily inefficient if
5382 dh->length (recommended exponent length) is much smaller than the
5383 length of dh->p. We could use BN_rand_range() if the order of
5384 the subgroup was stored in the DH structure, but we only have
5385 dh->length.
5386
5387 So switch back to
5388
5389 BN_rand(priv_key, l, ...)
5390
5391 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5392 otherwise.
5393 [Bodo Moeller]
5394
5395 *) In
5396
5397 RSA_eay_public_encrypt
5398 RSA_eay_private_decrypt
5399 RSA_eay_private_encrypt (signing)
5400 RSA_eay_public_decrypt (signature verification)
5401
5402 (default implementations for RSA_public_encrypt,
5403 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5404 always reject numbers >= n.
5405 [Bodo Moeller]
5406
5407 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5408 to synchronize access to 'locking_thread'. This is necessary on
5409 systems where access to 'locking_thread' (an 'unsigned long'
5410 variable) is not atomic.
5411 [Bodo Moeller]
5412
5413 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5414 *before* setting the 'crypto_lock_rand' flag. The previous code had
5415 a race condition if 0 is a valid thread ID.
5416 [Travis Vitek <vitek@roguewave.com>]
5417
5418 *) Add support for shared libraries under Irix.
5419 [Albert Chin-A-Young <china@thewrittenword.com>]
5420
5421 *) Add configuration option to build on Linux on both big-endian and
5422 little-endian MIPS.
5423 [Ralf Baechle <ralf@uni-koblenz.de>]
5424
5425 *) Add the possibility to create shared libraries on HP-UX.
5426 [Richard Levitte]
5427
5428 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5429
5430 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
5431 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
5432 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
5433 PRNG state recovery was possible based on the output of
5434 one PRNG request appropriately sized to gain knowledge on
5435 'md' followed by enough consecutive 1-byte PRNG requests
5436 to traverse all of 'state'.
5437
5438 1. When updating 'md_local' (the current thread's copy of 'md')
5439 during PRNG output generation, hash all of the previous
5440 'md_local' value, not just the half used for PRNG output.
5441
5442 2. Make the number of bytes from 'state' included into the hash
5443 independent from the number of PRNG bytes requested.
5444
5445 The first measure alone would be sufficient to avoid
5446 Markku-Juhani's attack. (Actually it had never occurred
5447 to me that the half of 'md_local' used for chaining was the
5448 half from which PRNG output bytes were taken -- I had always
5449 assumed that the secret half would be used.) The second
5450 measure makes sure that additional data from 'state' is never
5451 mixed into 'md_local' in small portions; this heuristically
5452 further strengthens the PRNG.
5453 [Bodo Moeller]
5454
5455 *) Fix crypto/bn/asm/mips3.s.
5456 [Andy Polyakov]
5457
5458 *) When only the key is given to "enc", the IV is undefined. Print out
5459 an error message in this case.
5460 [Lutz Jaenicke]
5461
5462 *) Handle special case when X509_NAME is empty in X509 printing routines.
5463 [Steve Henson]
5464
5465 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
5466 positive and less than q.
5467 [Bodo Moeller]
5468
5469 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
5470 used: it isn't thread safe and the add_lock_callback should handle
5471 that itself.
5472 [Paul Rose <Paul.Rose@bridge.com>]
5473
5474 *) Verify that incoming data obeys the block size in
5475 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
5476 [Bodo Moeller]
5477
5478 *) Fix OAEP check.
5479 [Ulf Möller, Bodo Möller]
5480
5481 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
5482 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
5483 when fixing the server behaviour for backwards-compatible 'client
5484 hello' messages. (Note that the attack is impractical against
5485 SSL 3.0 and TLS 1.0 anyway because length and version checking
5486 means that the probability of guessing a valid ciphertext is
5487 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
5488 paper.)
5489
5490 Before 0.9.5, the countermeasure (hide the error by generating a
5491 random 'decryption result') did not work properly because
5492 ERR_clear_error() was missing, meaning that SSL_get_error() would
5493 detect the supposedly ignored error.
5494
5495 Both problems are now fixed.
5496 [Bodo Moeller]
5497
5498 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
5499 (previously it was 1024).
5500 [Bodo Moeller]
5501
5502 *) Fix for compatibility mode trust settings: ignore trust settings
5503 unless some valid trust or reject settings are present.
5504 [Steve Henson]
5505
5506 *) Fix for blowfish EVP: its a variable length cipher.
5507 [Steve Henson]
5508
5509 *) Fix various bugs related to DSA S/MIME verification. Handle missing
5510 parameters in DSA public key structures and return an error in the
5511 DSA routines if parameters are absent.
5512 [Steve Henson]
5513
5514 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
5515 in the current directory if neither $RANDFILE nor $HOME was set.
5516 RAND_file_name() in 0.9.6a returned NULL in this case. This has
5517 caused some confusion to Windows users who haven't defined $HOME.
5518 Thus RAND_file_name() is changed again: e_os.h can define a
5519 DEFAULT_HOME, which will be used if $HOME is not set.
5520 For Windows, we use "C:"; on other platforms, we still require
5521 environment variables.
5522
5523 *) Move 'if (!initialized) RAND_poll()' into regions protected by
5524 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
5525 having multiple threads call RAND_poll() concurrently.
5526 [Bodo Moeller]
5527
5528 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
5529 combination of a flag and a thread ID variable.
5530 Otherwise while one thread is in ssleay_rand_bytes (which sets the
5531 flag), *other* threads can enter ssleay_add_bytes without obeying
5532 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
5533 that they do not hold after the first thread unsets add_do_not_lock).
5534 [Bodo Moeller]
5535
5536 *) Change bctest again: '-x' expressions are not available in all
5537 versions of 'test'.
5538 [Bodo Moeller]
5539
5540 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
5541
5542 *) Fix a couple of memory leaks in PKCS7_dataDecode()
5543 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
5544
5545 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
5546 the default extension for executables, if any. Also, make the perl
5547 scripts that use symlink() to test if it really exists and use "cp"
5548 if it doesn't. All this made OpenSSL compilable and installable in
5549 CygWin.
5550 [Richard Levitte]
5551
5552 *) Fix for asn1_GetSequence() for indefinite length constructed data.
5553 If SEQUENCE is length is indefinite just set c->slen to the total
5554 amount of data available.
5555 [Steve Henson, reported by shige@FreeBSD.org]
5556 [This change does not apply to 0.9.7.]
5557
5558 *) Change bctest to avoid here-documents inside command substitution
5559 (workaround for FreeBSD /bin/sh bug).
5560 For compatibility with Ultrix, avoid shell functions (introduced
5561 in the bctest version that searches along $PATH).
5562 [Bodo Moeller]
5563
5564 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
5565 with des_encrypt() defined on some operating systems, like Solaris
5566 and UnixWare.
5567 [Richard Levitte]
5568
5569 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
5570 On the Importance of Eliminating Errors in Cryptographic
5571 Computations, J. Cryptology 14 (2001) 2, 101-119,
5572 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
5573 [Ulf Moeller]
5574
5575 *) MIPS assembler BIGNUM division bug fix.
5576 [Andy Polyakov]
5577
5578 *) Disabled incorrect Alpha assembler code.
5579 [Richard Levitte]
5580
5581 *) Fix PKCS#7 decode routines so they correctly update the length
5582 after reading an EOC for the EXPLICIT tag.
5583 [Steve Henson]
5584 [This change does not apply to 0.9.7.]
5585
5586 *) Fix bug in PKCS#12 key generation routines. This was triggered
5587 if a 3DES key was generated with a 0 initial byte. Include
5588 PKCS12_BROKEN_KEYGEN compilation option to retain the old
5589 (but broken) behaviour.
5590 [Steve Henson]
5591
5592 *) Enhance bctest to search for a working bc along $PATH and print
5593 it when found.
5594 [Tim Rice <tim@multitalents.net> via Richard Levitte]
5595
5596 *) Fix memory leaks in err.c: free err_data string if necessary;
5597 don't write to the wrong index in ERR_set_error_data.
5598 [Bodo Moeller]
5599
5600 *) Implement ssl23_peek (analogous to ssl23_read), which previously
5601 did not exist.
5602 [Bodo Moeller]
5603
5604 *) Replace rdtsc with _emit statements for VC++ version 5.
5605 [Jeremy Cooper <jeremy@baymoo.org>]
5606
5607 *) Make it possible to reuse SSLv2 sessions.
5608 [Richard Levitte]
5609
5610 *) In copy_email() check for >= 0 as a return value for
5611 X509_NAME_get_index_by_NID() since 0 is a valid index.
5612 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
5613
5614 *) Avoid coredump with unsupported or invalid public keys by checking if
5615 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
5616 PKCS7_verify() fails with non detached data.
5617 [Steve Henson]
5618
5619 *) Don't use getenv in library functions when run as setuid/setgid.
5620 New function OPENSSL_issetugid().
5621 [Ulf Moeller]
5622
5623 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
5624 due to incorrect handling of multi-threading:
5625
5626 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
5627
5628 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
5629
5630 3. Count how many times MemCheck_off() has been called so that
5631 nested use can be treated correctly. This also avoids
5632 inband-signalling in the previous code (which relied on the
5633 assumption that thread ID 0 is impossible).
5634 [Bodo Moeller]
5635
5636 *) Add "-rand" option also to s_client and s_server.
5637 [Lutz Jaenicke]
5638
5639 *) Fix CPU detection on Irix 6.x.
5640 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
5641 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5642
5643 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
5644 was empty.
5645 [Steve Henson]
5646 [This change does not apply to 0.9.7.]
5647
5648 *) Use the cached encoding of an X509_NAME structure rather than
5649 copying it. This is apparently the reason for the libsafe "errors"
5650 but the code is actually correct.
5651 [Steve Henson]
5652
5653 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
5654 Bleichenbacher's DSA attack.
5655 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
5656 to be set and top=0 forces the highest bit to be set; top=-1 is new
5657 and leaves the highest bit random.
5658 [Ulf Moeller, Bodo Moeller]
5659
5660 *) In the NCONF_...-based implementations for CONF_... queries
5661 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
5662 a temporary CONF structure with the data component set to NULL
5663 (which gives segmentation faults in lh_retrieve).
5664 Instead, use NULL for the CONF pointer in CONF_get_string and
5665 CONF_get_number (which may use environment variables) and directly
5666 return NULL from CONF_get_section.
5667 [Bodo Moeller]
5668
5669 *) Fix potential buffer overrun for EBCDIC.
5670 [Ulf Moeller]
5671
5672 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
5673 keyUsage if basicConstraints absent for a CA.
5674 [Steve Henson]
5675
5676 *) Make SMIME_write_PKCS7() write mail header values with a format that
5677 is more generally accepted (no spaces before the semicolon), since
5678 some programs can't parse those values properly otherwise. Also make
5679 sure BIO's that break lines after each write do not create invalid
5680 headers.
5681 [Richard Levitte]
5682
5683 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
5684 macros previously used would not encode an empty SEQUENCE OF
5685 and break the signature.
5686 [Steve Henson]
5687 [This change does not apply to 0.9.7.]
5688
5689 *) Zero the premaster secret after deriving the master secret in
5690 DH ciphersuites.
5691 [Steve Henson]
5692
5693 *) Add some EVP_add_digest_alias registrations (as found in
5694 OpenSSL_add_all_digests()) to SSL_library_init()
5695 aka OpenSSL_add_ssl_algorithms(). This provides improved
5696 compatibility with peers using X.509 certificates
5697 with unconventional AlgorithmIdentifier OIDs.
5698 [Bodo Moeller]
5699
5700 *) Fix for Irix with NO_ASM.
5701 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5702
5703 *) ./config script fixes.
5704 [Ulf Moeller, Richard Levitte]
5705
5706 *) Fix 'openssl passwd -1'.
5707 [Bodo Moeller]
5708
5709 *) Change PKCS12_key_gen_asc() so it can cope with non null
5710 terminated strings whose length is passed in the passlen
5711 parameter, for example from PEM callbacks. This was done
5712 by adding an extra length parameter to asc2uni().
5713 [Steve Henson, reported by <oddissey@samsung.co.kr>]
5714
5715 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
5716 call failed, free the DSA structure.
5717 [Bodo Moeller]
5718
5719 *) Fix to uni2asc() to cope with zero length Unicode strings.
5720 These are present in some PKCS#12 files.
5721 [Steve Henson]
5722
5723 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
5724 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
5725 when writing a 32767 byte record.
5726 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
5727
5728 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
5729 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
5730
5731 (RSA objects have a reference count access to which is protected
5732 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
5733 so they are meant to be shared between threads.)
5734 [Bodo Moeller, Geoff Thorpe; original patch submitted by
5735 "Reddie, Steven" <Steven.Reddie@ca.com>]
5736
5737 *) Fix a deadlock in CRYPTO_mem_leaks().
5738 [Bodo Moeller]
5739
5740 *) Use better test patterns in bntest.
5741 [Ulf Möller]
5742
5743 *) rand_win.c fix for Borland C.
5744 [Ulf Möller]
5745
5746 *) BN_rshift bugfix for n == 0.
5747 [Bodo Moeller]
5748
5749 *) Add a 'bctest' script that checks for some known 'bc' bugs
5750 so that 'make test' does not abort just because 'bc' is broken.
5751 [Bodo Moeller]
5752
5753 *) Store verify_result within SSL_SESSION also for client side to
5754 avoid potential security hole. (Re-used sessions on the client side
5755 always resulted in verify_result==X509_V_OK, not using the original
5756 result of the server certificate verification.)
5757 [Lutz Jaenicke]
5758
5759 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
5760 SSL3_RT_APPLICATION_DATA, return 0.
5761 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
5762 [Bodo Moeller]
5763
5764 *) Fix SSL_peek:
5765 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
5766 releases, have been re-implemented by renaming the previous
5767 implementations of ssl2_read and ssl3_read to ssl2_read_internal
5768 and ssl3_read_internal, respectively, and adding 'peek' parameters
5769 to them. The new ssl[23]_{read,peek} functions are calls to
5770 ssl[23]_read_internal with the 'peek' flag set appropriately.
5771 A 'peek' parameter has also been added to ssl3_read_bytes, which
5772 does the actual work for ssl3_read_internal.
5773 [Bodo Moeller]
5774
5775 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
5776 the method-specific "init()" handler. Also clean up ex_data after
5777 calling the method-specific "finish()" handler. Previously, this was
5778 happening the other way round.
5779 [Geoff Thorpe]
5780
5781 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
5782 The previous value, 12, was not always sufficient for BN_mod_exp().
5783 [Bodo Moeller]
5784
5785 *) Make sure that shared libraries get the internal name engine with
5786 the full version number and not just 0. This should mark the
5787 shared libraries as not backward compatible. Of course, this should
5788 be changed again when we can guarantee backward binary compatibility.
5789 [Richard Levitte]
5790
5791 *) Fix typo in get_cert_by_subject() in by_dir.c
5792 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
5793
5794 *) Rework the system to generate shared libraries:
5795
5796 - Make note of the expected extension for the shared libraries and
5797 if there is a need for symbolic links from for example libcrypto.so.0
5798 to libcrypto.so.0.9.7. There is extended info in Configure for
5799 that.
5800
5801 - Make as few rebuilds of the shared libraries as possible.
5802
5803 - Still avoid linking the OpenSSL programs with the shared libraries.
5804
5805 - When installing, install the shared libraries separately from the
5806 static ones.
5807 [Richard Levitte]
5808
5809 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
5810
5811 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
5812 and not in SSL_clear because the latter is also used by the
5813 accept/connect functions; previously, the settings made by
5814 SSL_set_read_ahead would be lost during the handshake.
5815 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
5816
5817 *) Correct util/mkdef.pl to be selective about disabled algorithms.
5818 Previously, it would create entries for disableed algorithms no
5819 matter what.
5820 [Richard Levitte]
5821
5822 *) Added several new manual pages for SSL_* function.
5823 [Lutz Jaenicke]
5824
5825 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
5826
5827 *) In ssl23_get_client_hello, generate an error message when faced
5828 with an initial SSL 3.0/TLS record that is too small to contain the
5829 first two bytes of the ClientHello message, i.e. client_version.
5830 (Note that this is a pathologic case that probably has never happened
5831 in real life.) The previous approach was to use the version number
5832 from the record header as a substitute; but our protocol choice
5833 should not depend on that one because it is not authenticated
5834 by the Finished messages.
5835 [Bodo Moeller]
5836
5837 *) More robust randomness gathering functions for Windows.
5838 [Jeffrey Altman <jaltman@columbia.edu>]
5839
5840 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
5841 not set then we don't setup the error code for issuer check errors
5842 to avoid possibly overwriting other errors which the callback does
5843 handle. If an application does set the flag then we assume it knows
5844 what it is doing and can handle the new informational codes
5845 appropriately.
5846 [Steve Henson]
5847
5848 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
5849 a general "ANY" type, as such it should be able to decode anything
5850 including tagged types. However it didn't check the class so it would
5851 wrongly interpret tagged types in the same way as their universal
5852 counterpart and unknown types were just rejected. Changed so that the
5853 tagged and unknown types are handled in the same way as a SEQUENCE:
5854 that is the encoding is stored intact. There is also a new type
5855 "V_ASN1_OTHER" which is used when the class is not universal, in this
5856 case we have no idea what the actual type is so we just lump them all
5857 together.
5858 [Steve Henson]
5859
5860 *) On VMS, stdout may very well lead to a file that is written to
5861 in a record-oriented fashion. That means that every write() will
5862 write a separate record, which will be read separately by the
5863 programs trying to read from it. This can be very confusing.
5864
5865 The solution is to put a BIO filter in the way that will buffer
5866 text until a linefeed is reached, and then write everything a
5867 line at a time, so every record written will be an actual line,
5868 not chunks of lines and not (usually doesn't happen, but I've
5869 seen it once) several lines in one record. BIO_f_linebuffer() is
5870 the answer.
5871
5872 Currently, it's a VMS-only method, because that's where it has
5873 been tested well enough.
5874 [Richard Levitte]
5875
5876 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
5877 it can return incorrect results.
5878 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
5879 but it was in 0.9.6-beta[12].)
5880 [Bodo Moeller]
5881
5882 *) Disable the check for content being present when verifying detached
5883 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
5884 include zero length content when signing messages.
5885 [Steve Henson]
5886
5887 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
5888 BIO_ctrl (for BIO pairs).
5889 [Bodo Möller]
5890
5891 *) Add DSO method for VMS.
5892 [Richard Levitte]
5893
5894 *) Bug fix: Montgomery multiplication could produce results with the
5895 wrong sign.
5896 [Ulf Möller]
5897
5898 *) Add RPM specification openssl.spec and modify it to build three
5899 packages. The default package contains applications, application
5900 documentation and run-time libraries. The devel package contains
5901 include files, static libraries and function documentation. The
5902 doc package contains the contents of the doc directory. The original
5903 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5904 [Richard Levitte]
5905
5906 *) Add a large number of documentation files for many SSL routines.
5907 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5908
5909 *) Add a configuration entry for Sony News 4.
5910 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5911
5912 *) Don't set the two most significant bits to one when generating a
5913 random number < q in the DSA library.
5914 [Ulf Möller]
5915
5916 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5917 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5918 the underlying transport is blocking) if a handshake took place.
5919 (The default behaviour is needed by applications such as s_client
5920 and s_server that use select() to determine when to use SSL_read;
5921 but for applications that know in advance when to expect data, it
5922 just makes things more complicated.)
5923 [Bodo Moeller]
5924
5925 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5926 from EGD.
5927 [Ben Laurie]
5928
5929 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5930 work better on such systems.
5931 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5932
5933 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5934 Update PKCS12_parse() so it copies the friendlyName and the
5935 keyid to the certificates aux info.
5936 [Steve Henson]
5937
5938 *) Fix bug in PKCS7_verify() which caused an infinite loop
5939 if there was more than one signature.
5940 [Sven Uszpelkat <su@celocom.de>]
5941
5942 *) Major change in util/mkdef.pl to include extra information
5943 about each symbol, as well as presentig variables as well
5944 as functions. This change means that there's n more need
5945 to rebuild the .num files when some algorithms are excluded.
5946 [Richard Levitte]
5947
5948 *) Allow the verify time to be set by an application,
5949 rather than always using the current time.
5950 [Steve Henson]
5951
5952 *) Phase 2 verify code reorganisation. The certificate
5953 verify code now looks up an issuer certificate by a
5954 number of criteria: subject name, authority key id
5955 and key usage. It also verifies self signed certificates
5956 by the same criteria. The main comparison function is
5957 X509_check_issued() which performs these checks.
5958
5959 Lot of changes were necessary in order to support this
5960 without completely rewriting the lookup code.
5961
5962 Authority and subject key identifier are now cached.
5963
5964 The LHASH 'certs' is X509_STORE has now been replaced
5965 by a STACK_OF(X509_OBJECT). This is mainly because an
5966 LHASH can't store or retrieve multiple objects with
5967 the same hash value.
5968
5969 As a result various functions (which were all internal
5970 use only) have changed to handle the new X509_STORE
5971 structure. This will break anything that messed round
5972 with X509_STORE internally.
5973
5974 The functions X509_STORE_add_cert() now checks for an
5975 exact match, rather than just subject name.
5976
5977 The X509_STORE API doesn't directly support the retrieval
5978 of multiple certificates matching a given criteria, however
5979 this can be worked round by performing a lookup first
5980 (which will fill the cache with candidate certificates)
5981 and then examining the cache for matches. This is probably
5982 the best we can do without throwing out X509_LOOKUP
5983 entirely (maybe later...).
5984
5985 The X509_VERIFY_CTX structure has been enhanced considerably.
5986
5987 All certificate lookup operations now go via a get_issuer()
5988 callback. Although this currently uses an X509_STORE it
5989 can be replaced by custom lookups. This is a simple way
5990 to bypass the X509_STORE hackery necessary to make this
5991 work and makes it possible to use more efficient techniques
5992 in future. A very simple version which uses a simple
5993 STACK for its trusted certificate store is also provided
5994 using X509_STORE_CTX_trusted_stack().
5995
5996 The verify_cb() and verify() callbacks now have equivalents
5997 in the X509_STORE_CTX structure.
5998
5999 X509_STORE_CTX also has a 'flags' field which can be used
6000 to customise the verify behaviour.
6001 [Steve Henson]
6002
6003 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6004 excludes S/MIME capabilities.
6005 [Steve Henson]
6006
6007 *) When a certificate request is read in keep a copy of the
6008 original encoding of the signed data and use it when outputing
6009 again. Signatures then use the original encoding rather than
6010 a decoded, encoded version which may cause problems if the
6011 request is improperly encoded.
6012 [Steve Henson]
6013
6014 *) For consistency with other BIO_puts implementations, call
6015 buffer_write(b, ...) directly in buffer_puts instead of calling
6016 BIO_write(b, ...).
6017
6018 In BIO_puts, increment b->num_write as in BIO_write.
6019 [Peter.Sylvester@EdelWeb.fr]
6020
6021 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6022 BN_zero, we may not return a BIGNUM with an array consisting of
6023 words set to zero.)
6024 [Bodo Moeller]
6025
6026 *) Avoid calling abort() from within the library when problems are
6027 detected, except if preprocessor symbols have been defined
6028 (such as REF_CHECK, BN_DEBUG etc.).
6029 [Bodo Moeller]
6030
6031 *) New openssl application 'rsautl'. This utility can be
6032 used for low level RSA operations. DER public key
6033 BIO/fp routines also added.
6034 [Steve Henson]
6035
6036 *) New Configure entry and patches for compiling on QNX 4.
6037 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6038
6039 *) A demo state-machine implementation was sponsored by
6040 Nuron (http://www.nuron.com/) and is now available in
6041 demos/state_machine.
6042 [Ben Laurie]
6043
6044 *) New options added to the 'dgst' utility for signature
6045 generation and verification.
6046 [Steve Henson]
6047
6048 *) Unrecognized PKCS#7 content types are now handled via a
6049 catch all ASN1_TYPE structure. This allows unsupported
6050 types to be stored as a "blob" and an application can
6051 encode and decode it manually.
6052 [Steve Henson]
6053
6054 *) Fix various signed/unsigned issues to make a_strex.c
6055 compile under VC++.
6056 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6057
6058 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6059 length if passed a buffer. ASN1_INTEGER_to_BN failed
6060 if passed a NULL BN and its argument was negative.
6061 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6062
6063 *) Modification to PKCS#7 encoding routines to output definite
6064 length encoding. Since currently the whole structures are in
6065 memory there's not real point in using indefinite length
6066 constructed encoding. However if OpenSSL is compiled with
6067 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6068 [Steve Henson]
6069
6070 *) Added BIO_vprintf() and BIO_vsnprintf().
6071 [Richard Levitte]
6072
6073 *) Added more prefixes to parse for in the the strings written
6074 through a logging bio, to cover all the levels that are available
6075 through syslog. The prefixes are now:
6076
6077 PANIC, EMERG, EMR => LOG_EMERG
6078 ALERT, ALR => LOG_ALERT
6079 CRIT, CRI => LOG_CRIT
6080 ERROR, ERR => LOG_ERR
6081 WARNING, WARN, WAR => LOG_WARNING
6082 NOTICE, NOTE, NOT => LOG_NOTICE
6083 INFO, INF => LOG_INFO
6084 DEBUG, DBG => LOG_DEBUG
6085
6086 and as before, if none of those prefixes are present at the
6087 beginning of the string, LOG_ERR is chosen.
6088
6089 On Win32, the LOG_* levels are mapped according to this:
6090
6091 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6092 LOG_WARNING => EVENTLOG_WARNING_TYPE
6093 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6094
6095 [Richard Levitte]
6096
6097 *) Made it possible to reconfigure with just the configuration
6098 argument "reconf" or "reconfigure". The command line arguments
6099 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6100 and are retrieved from there when reconfiguring.
6101 [Richard Levitte]
6102
6103 *) MD4 implemented.
6104 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6105
6106 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6107 [Richard Levitte]
6108
6109 *) The obj_dat.pl script was messing up the sorting of object
6110 names. The reason was that it compared the quoted version
6111 of strings as a result "OCSP" > "OCSP Signing" because
6112 " > SPACE. Changed script to store unquoted versions of
6113 names and add quotes on output. It was also omitting some
6114 names from the lookup table if they were given a default
6115 value (that is if SN is missing it is given the same
6116 value as LN and vice versa), these are now added on the
6117 grounds that if an object has a name we should be able to
6118 look it up. Finally added warning output when duplicate
6119 short or long names are found.
6120 [Steve Henson]
6121
6122 *) Changes needed for Tandem NSK.
6123 [Scott Uroff <scott@xypro.com>]
6124
6125 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6126 RSA_padding_check_SSLv23(), special padding was never detected
6127 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6128 version rollback attacks was not effective.
6129
6130 In s23_clnt.c, don't use special rollback-attack detection padding
6131 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6132 client; similarly, in s23_srvr.c, don't do the rollback check if
6133 SSL 2.0 is the only protocol enabled in the server.
6134 [Bodo Moeller]
6135
6136 *) Make it possible to get hexdumps of unprintable data with 'openssl
6137 asn1parse'. By implication, the functions ASN1_parse_dump() and
6138 BIO_dump_indent() are added.
6139 [Richard Levitte]
6140
6141 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6142 these print out strings and name structures based on various
6143 flags including RFC2253 support and proper handling of
6144 multibyte characters. Added options to the 'x509' utility
6145 to allow the various flags to be set.
6146 [Steve Henson]
6147
6148 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6149 Also change the functions X509_cmp_current_time() and
6150 X509_gmtime_adj() work with an ASN1_TIME structure,
6151 this will enable certificates using GeneralizedTime in validity
6152 dates to be checked.
6153 [Steve Henson]
6154
6155 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6156 negative public key encodings) on by default,
6157 NO_NEG_PUBKEY_BUG can be set to disable it.
6158 [Steve Henson]
6159
6160 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6161 content octets. An i2c_ASN1_OBJECT is unnecessary because
6162 the encoding can be trivially obtained from the structure.
6163 [Steve Henson]
6164
6165 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6166 not read locks (CRYPTO_r_[un]lock).
6167 [Bodo Moeller]
6168
6169 *) A first attempt at creating official support for shared
6170 libraries through configuration. I've kept it so the
6171 default is static libraries only, and the OpenSSL programs
6172 are always statically linked for now, but there are
6173 preparations for dynamic linking in place.
6174 This has been tested on Linux and Tru64.
6175 [Richard Levitte]
6176
6177 *) Randomness polling function for Win9x, as described in:
6178 Peter Gutmann, Software Generation of Practically Strong
6179 Random Numbers.
6180 [Ulf Möller]
6181
6182 *) Fix so PRNG is seeded in req if using an already existing
6183 DSA key.
6184 [Steve Henson]
6185
6186 *) New options to smime application. -inform and -outform
6187 allow alternative formats for the S/MIME message including
6188 PEM and DER. The -content option allows the content to be
6189 specified separately. This should allow things like Netscape
6190 form signing output easier to verify.
6191 [Steve Henson]
6192
6193 *) Fix the ASN1 encoding of tags using the 'long form'.
6194 [Steve Henson]
6195
6196 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6197 STRING types. These convert content octets to and from the
6198 underlying type. The actual tag and length octets are
6199 already assumed to have been read in and checked. These
6200 are needed because all other string types have virtually
6201 identical handling apart from the tag. By having versions
6202 of the ASN1 functions that just operate on content octets
6203 IMPLICIT tagging can be handled properly. It also allows
6204 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6205 and ASN1_INTEGER are identical apart from the tag.
6206 [Steve Henson]
6207
6208 *) Change the handling of OID objects as follows:
6209
6210 - New object identifiers are inserted in objects.txt, following
6211 the syntax given in objects.README.
6212 - objects.pl is used to process obj_mac.num and create a new
6213 obj_mac.h.
6214 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6215 obj_mac.h.
6216
6217 This is currently kind of a hack, and the perl code in objects.pl
6218 isn't very elegant, but it works as I intended. The simplest way
6219 to check that it worked correctly is to look in obj_dat.h and
6220 check the array nid_objs and make sure the objects haven't moved
6221 around (this is important!). Additions are OK, as well as
6222 consistent name changes.
6223 [Richard Levitte]
6224
6225 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6226 [Bodo Moeller]
6227
6228 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6229 The given file adds to whatever has already been seeded into the
6230 random pool through the RANDFILE configuration file option or
6231 environment variable, or the default random state file.
6232 [Richard Levitte]
6233
6234 *) mkstack.pl now sorts each macro group into lexical order.
6235 Previously the output order depended on the order the files
6236 appeared in the directory, resulting in needless rewriting
6237 of safestack.h .
6238 [Steve Henson]
6239
6240 *) Patches to make OpenSSL compile under Win32 again. Mostly
6241 work arounds for the VC++ problem that it treats func() as
6242 func(void). Also stripped out the parts of mkdef.pl that
6243 added extra typesafe functions: these no longer exist.
6244 [Steve Henson]
6245
6246 *) Reorganisation of the stack code. The macros are now all
6247 collected in safestack.h . Each macro is defined in terms of
6248 a "stack macro" of the form SKM_<name>(type, a, b). The
6249 DEBUG_SAFESTACK is now handled in terms of function casts,
6250 this has the advantage of retaining type safety without the
6251 use of additional functions. If DEBUG_SAFESTACK is not defined
6252 then the non typesafe macros are used instead. Also modified the
6253 mkstack.pl script to handle the new form. Needs testing to see
6254 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6255 the default if no major problems. Similar behaviour for ASN1_SET_OF
6256 and PKCS12_STACK_OF.
6257 [Steve Henson]
6258
6259 *) When some versions of IIS use the 'NET' form of private key the
6260 key derivation algorithm is different. Normally MD5(password) is
6261 used as a 128 bit RC4 key. In the modified case
6262 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6263 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6264 as the old Netscape_RSA functions except they have an additional
6265 'sgckey' parameter which uses the modified algorithm. Also added
6266 an -sgckey command line option to the rsa utility. Thanks to
6267 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6268 algorithm to openssl-dev.
6269 [Steve Henson]
6270
6271 *) The evp_local.h macros were using 'c.##kname' which resulted in
6272 invalid expansion on some systems (SCO 5.0.5 for example).
6273 Corrected to 'c.kname'.
6274 [Phillip Porch <root@theporch.com>]
6275
6276 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6277 a STACK of email addresses from a certificate or request, these look
6278 in the subject name and the subject alternative name extensions and
6279 omit any duplicate addresses.
6280 [Steve Henson]
6281
6282 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6283 This makes DSA verification about 2 % faster.
6284 [Bodo Moeller]
6285
6286 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6287 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6288 plus overhead for 1024 bit moduli).
6289 This makes exponentiations about 0.5 % faster for 1024 bit
6290 exponents (as measured by "openssl speed rsa2048").
6291 [Bodo Moeller]
6292
6293 *) Rename memory handling macros to avoid conflicts with other
6294 software:
6295 Malloc => OPENSSL_malloc
6296 Malloc_locked => OPENSSL_malloc_locked
6297 Realloc => OPENSSL_realloc
6298 Free => OPENSSL_free
6299 [Richard Levitte]
6300
6301 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6302 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6303 [Bodo Moeller]
6304
6305 *) CygWin32 support.
6306 [John Jarvie <jjarvie@newsguy.com>]
6307
6308 *) The type-safe stack code has been rejigged. It is now only compiled
6309 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6310 by default all type-specific stack functions are "#define"d back to
6311 standard stack functions. This results in more streamlined output
6312 but retains the type-safety checking possibilities of the original
6313 approach.
6314 [Geoff Thorpe]
6315
6316 *) The STACK code has been cleaned up, and certain type declarations
6317 that didn't make a lot of sense have been brought in line. This has
6318 also involved a cleanup of sorts in safestack.h to more correctly
6319 map type-safe stack functions onto their plain stack counterparts.
6320 This work has also resulted in a variety of "const"ifications of
6321 lots of the code, especially "_cmp" operations which should normally
6322 be prototyped with "const" parameters anyway.
6323 [Geoff Thorpe]
6324
6325 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6326 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6327 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6328 where all of 'md' is used each time the PRNG is used, but 'state'
6329 is used only indexed by a cyclic counter. As entropy may not be
6330 well distributed from the beginning, 'md' is important as a
6331 chaining variable. However, the output function chains only half
6332 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6333 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6334 in all of 'state' being rewritten, with the new values depending
6335 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6336 [Bodo Moeller]
6337
6338 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6339 the handshake is continued after ssl_verify_cert_chain();
6340 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6341 can lead to 'unexplainable' connection aborts later.
6342 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6343
6344 *) Major EVP API cipher revision.
6345 Add hooks for extra EVP features. This allows various cipher
6346 parameters to be set in the EVP interface. Support added for variable
6347 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6348 setting of RC2 and RC5 parameters.
6349
6350 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6351 ciphers.
6352
6353 Remove lots of duplicated code from the EVP library. For example *every*
6354 cipher init() function handles the 'iv' in the same way according to the
6355 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6356 for CFB and OFB modes they zero ctx->num.
6357
6358 New functionality allows removal of S/MIME code RC2 hack.
6359
6360 Most of the routines have the same form and so can be declared in terms
6361 of macros.
6362
6363 By shifting this to the top level EVP_CipherInit() it can be removed from
6364 all individual ciphers. If the cipher wants to handle IVs or keys
6365 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6366 flags.
6367
6368 Change lots of functions like EVP_EncryptUpdate() to now return a
6369 value: although software versions of the algorithms cannot fail
6370 any installed hardware versions can.
6371 [Steve Henson]
6372
6373 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6374 this option is set, tolerate broken clients that send the negotiated
6375 protocol version number instead of the requested protocol version
6376 number.
6377 [Bodo Moeller]
6378
6379 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6380 i.e. non-zero for export ciphersuites, zero otherwise.
6381 Previous versions had this flag inverted, inconsistent with
6382 rsa_tmp_cb (..._TMP_RSA_CB).
6383 [Bodo Moeller; problem reported by Amit Chopra]
6384
6385 *) Add missing DSA library text string. Work around for some IIS
6386 key files with invalid SEQUENCE encoding.
6387 [Steve Henson]
6388
6389 *) Add a document (doc/standards.txt) that list all kinds of standards
6390 and so on that are implemented in OpenSSL.
6391 [Richard Levitte]
6392
6393 *) Enhance c_rehash script. Old version would mishandle certificates
6394 with the same subject name hash and wouldn't handle CRLs at all.
6395 Added -fingerprint option to crl utility, to support new c_rehash
6396 features.
6397 [Steve Henson]
6398
6399 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6400 [Ulf Möller]
6401
6402 *) Fix for SSL server purpose checking. Server checking was
6403 rejecting certificates which had extended key usage present
6404 but no ssl client purpose.
6405 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6406
6407 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6408 is a little unclear about how a blank password is handled.
6409 Since the password in encoded as a BMPString with terminating
6410 double NULL a zero length password would end up as just the
6411 double NULL. However no password at all is different and is
6412 handled differently in the PKCS#12 key generation code. NS
6413 treats a blank password as zero length. MSIE treats it as no
6414 password on export: but it will try both on import. We now do
6415 the same: PKCS12_parse() tries zero length and no password if
6416 the password is set to "" or NULL (NULL is now a valid password:
6417 it wasn't before) as does the pkcs12 application.
6418 [Steve Henson]
6419
6420 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6421 perror when PEM_read_bio_X509_REQ fails, the error message must
6422 be obtained from the error queue.
6423 [Bodo Moeller]
6424
6425 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6426 it in ERR_remove_state if appropriate, and change ERR_get_state
6427 accordingly to avoid race conditions (this is necessary because
6428 thread_hash is no longer constant once set).
6429 [Bodo Moeller]
6430
6431 *) Bugfix for linux-elf makefile.one.
6432 [Ulf Möller]
6433
6434 *) RSA_get_default_method() will now cause a default
6435 RSA_METHOD to be chosen if one doesn't exist already.
6436 Previously this was only set during a call to RSA_new()
6437 or RSA_new_method(NULL) meaning it was possible for
6438 RSA_get_default_method() to return NULL.
6439 [Geoff Thorpe]
6440
6441 *) Added native name translation to the existing DSO code
6442 that will convert (if the flag to do so is set) filenames
6443 that are sufficiently small and have no path information
6444 into a canonical native form. Eg. "blah" converted to
6445 "libblah.so" or "blah.dll" etc.
6446 [Geoff Thorpe]
6447
6448 *) New function ERR_error_string_n(e, buf, len) which is like
6449 ERR_error_string(e, buf), but writes at most 'len' bytes
6450 including the 0 terminator. For ERR_error_string_n, 'buf'
6451 may not be NULL.
6452 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
6453
6454 *) CONF library reworked to become more general. A new CONF
6455 configuration file reader "class" is implemented as well as a
6456 new functions (NCONF_*, for "New CONF") to handle it. The now
6457 old CONF_* functions are still there, but are reimplemented to
6458 work in terms of the new functions. Also, a set of functions
6459 to handle the internal storage of the configuration data is
6460 provided to make it easier to write new configuration file
6461 reader "classes" (I can definitely see something reading a
6462 configuration file in XML format, for example), called _CONF_*,
6463 or "the configuration storage API"...
6464
6465 The new configuration file reading functions are:
6466
6467 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
6468 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
6469
6470 NCONF_default, NCONF_WIN32
6471
6472 NCONF_dump_fp, NCONF_dump_bio
6473
6474 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
6475 NCONF_new creates a new CONF object. This works in the same way
6476 as other interfaces in OpenSSL, like the BIO interface.
6477 NCONF_dump_* dump the internal storage of the configuration file,
6478 which is useful for debugging. All other functions take the same
6479 arguments as the old CONF_* functions wth the exception of the
6480 first that must be a `CONF *' instead of a `LHASH *'.
6481
6482 To make it easer to use the new classes with the old CONF_* functions,
6483 the function CONF_set_default_method is provided.
6484 [Richard Levitte]
6485
6486 *) Add '-tls1' option to 'openssl ciphers', which was already
6487 mentioned in the documentation but had not been implemented.
6488 (This option is not yet really useful because even the additional
6489 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
6490 [Bodo Moeller]
6491
6492 *) Initial DSO code added into libcrypto for letting OpenSSL (and
6493 OpenSSL-based applications) load shared libraries and bind to
6494 them in a portable way.
6495 [Geoff Thorpe, with contributions from Richard Levitte]
6496
6497 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
6498
6499 *) Make sure _lrotl and _lrotr are only used with MSVC.
6500
6501 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
6502 (the default implementation of RAND_status).
6503
6504 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
6505 to '-clrext' (= clear extensions), as intended and documented.
6506 [Bodo Moeller; inconsistency pointed out by Michael Attili
6507 <attili@amaxo.com>]
6508
6509 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
6510 was larger than the MD block size.
6511 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
6512
6513 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
6514 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
6515 using the passed key: if the passed key was a private key the result
6516 of X509_print(), for example, would be to print out all the private key
6517 components.
6518 [Steve Henson]
6519
6520 *) des_quad_cksum() byte order bug fix.
6521 [Ulf Möller, using the problem description in krb4-0.9.7, where
6522 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
6523
6524 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
6525 discouraged.
6526 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
6527
6528 *) For easily testing in shell scripts whether some command
6529 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
6530 returns with exit code 0 iff no command of the given name is available.
6531 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
6532 the output goes to stdout and nothing is printed to stderr.
6533 Additional arguments are always ignored.
6534
6535 Since for each cipher there is a command of the same name,
6536 the 'no-cipher' compilation switches can be tested this way.
6537
6538 ('openssl no-XXX' is not able to detect pseudo-commands such
6539 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
6540 [Bodo Moeller]
6541
6542 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
6543 [Bodo Moeller]
6544
6545 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
6546 is set; it will be thrown away anyway because each handshake creates
6547 its own key.
6548 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
6549 to parameters -- in previous versions (since OpenSSL 0.9.3) the
6550 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
6551 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
6552 [Bodo Moeller]
6553
6554 *) New s_client option -ign_eof: EOF at stdin is ignored, and
6555 'Q' and 'R' lose their special meanings (quit/renegotiate).
6556 This is part of what -quiet does; unlike -quiet, -ign_eof
6557 does not suppress any output.
6558 [Richard Levitte]
6559
6560 *) Add compatibility options to the purpose and trust code. The
6561 purpose X509_PURPOSE_ANY is "any purpose" which automatically
6562 accepts a certificate or CA, this was the previous behaviour,
6563 with all the associated security issues.
6564
6565 X509_TRUST_COMPAT is the old trust behaviour: only and
6566 automatically trust self signed roots in certificate store. A
6567 new trust setting X509_TRUST_DEFAULT is used to specify that
6568 a purpose has no associated trust setting and it should instead
6569 use the value in the default purpose.
6570 [Steve Henson]
6571
6572 *) Fix the PKCS#8 DSA private key code so it decodes keys again
6573 and fix a memory leak.
6574 [Steve Henson]
6575
6576 *) In util/mkerr.pl (which implements 'make errors'), preserve
6577 reason strings from the previous version of the .c file, as
6578 the default to have only downcase letters (and digits) in
6579 automatically generated reasons codes is not always appropriate.
6580 [Bodo Moeller]
6581
6582 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
6583 using strerror. Previously, ERR_reason_error_string() returned
6584 library names as reason strings for SYSerr; but SYSerr is a special
6585 case where small numbers are errno values, not library numbers.
6586 [Bodo Moeller]
6587
6588 *) Add '-dsaparam' option to 'openssl dhparam' application. This
6589 converts DSA parameters into DH parameters. (When creating parameters,
6590 DSA_generate_parameters is used.)
6591 [Bodo Moeller]
6592
6593 *) Include 'length' (recommended exponent length) in C code generated
6594 by 'openssl dhparam -C'.
6595 [Bodo Moeller]
6596
6597 *) The second argument to set_label in perlasm was already being used
6598 so couldn't be used as a "file scope" flag. Moved to third argument
6599 which was free.
6600 [Steve Henson]
6601
6602 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
6603 instead of RAND_bytes for encryption IVs and salts.
6604 [Bodo Moeller]
6605
6606 *) Include RAND_status() into RAND_METHOD instead of implementing
6607 it only for md_rand.c Otherwise replacing the PRNG by calling
6608 RAND_set_rand_method would be impossible.
6609 [Bodo Moeller]
6610
6611 *) Don't let DSA_generate_key() enter an infinite loop if the random
6612 number generation fails.
6613 [Bodo Moeller]
6614
6615 *) New 'rand' application for creating pseudo-random output.
6616 [Bodo Moeller]
6617
6618 *) Added configuration support for Linux/IA64
6619 [Rolf Haberrecker <rolf@suse.de>]
6620
6621 *) Assembler module support for Mingw32.
6622 [Ulf Möller]
6623
6624 *) Shared library support for HPUX (in shlib/).
6625 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
6626
6627 *) Shared library support for Solaris gcc.
6628 [Lutz Behnke <behnke@trustcenter.de>]
6629
6630 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
6631
6632 *) PKCS7_encrypt() was adding text MIME headers twice because they
6633 were added manually and by SMIME_crlf_copy().
6634 [Steve Henson]
6635
6636 *) In bntest.c don't call BN_rand with zero bits argument.
6637 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
6638
6639 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
6640 case was implemented. This caused BN_div_recp() to fail occasionally.
6641 [Ulf Möller]
6642
6643 *) Add an optional second argument to the set_label() in the perl
6644 assembly language builder. If this argument exists and is set
6645 to 1 it signals that the assembler should use a symbol whose
6646 scope is the entire file, not just the current function. This
6647 is needed with MASM which uses the format label:: for this scope.
6648 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
6649
6650 *) Change the ASN1 types so they are typedefs by default. Before
6651 almost all types were #define'd to ASN1_STRING which was causing
6652 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
6653 for example.
6654 [Steve Henson]
6655
6656 *) Change names of new functions to the new get1/get0 naming
6657 convention: After 'get1', the caller owns a reference count
6658 and has to call ..._free; 'get0' returns a pointer to some
6659 data structure without incrementing reference counters.
6660 (Some of the existing 'get' functions increment a reference
6661 counter, some don't.)
6662 Similarly, 'set1' and 'add1' functions increase reference
6663 counters or duplicate objects.
6664 [Steve Henson]
6665
6666 *) Allow for the possibility of temp RSA key generation failure:
6667 the code used to assume it always worked and crashed on failure.
6668 [Steve Henson]
6669
6670 *) Fix potential buffer overrun problem in BIO_printf().
6671 [Ulf Möller, using public domain code by Patrick Powell; problem
6672 pointed out by David Sacerdote <das33@cornell.edu>]
6673
6674 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
6675 RAND_egd() and RAND_status(). In the command line application,
6676 the EGD socket can be specified like a seed file using RANDFILE
6677 or -rand.
6678 [Ulf Möller]
6679
6680 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
6681 Some CAs (e.g. Verisign) distribute certificates in this form.
6682 [Steve Henson]
6683
6684 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
6685 list to exclude them. This means that no special compilation option
6686 is needed to use anonymous DH: it just needs to be included in the
6687 cipher list.
6688 [Steve Henson]
6689
6690 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
6691 EVP_MD_type. The old functionality is available in a new macro called
6692 EVP_MD_md(). Change code that uses it and update docs.
6693 [Steve Henson]
6694
6695 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
6696 where the 'void *' argument is replaced by a function pointer argument.
6697 Previously 'void *' was abused to point to functions, which works on
6698 many platforms, but is not correct. As these functions are usually
6699 called by macros defined in OpenSSL header files, most source code
6700 should work without changes.
6701 [Richard Levitte]
6702
6703 *) <openssl/opensslconf.h> (which is created by Configure) now contains
6704 sections with information on -D... compiler switches used for
6705 compiling the library so that applications can see them. To enable
6706 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
6707 must be defined. E.g.,
6708 #define OPENSSL_ALGORITHM_DEFINES
6709 #include <openssl/opensslconf.h>
6710 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
6711 [Richard Levitte, Ulf and Bodo Möller]
6712
6713 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
6714 record layer.
6715 [Bodo Moeller]
6716
6717 *) Change the 'other' type in certificate aux info to a STACK_OF
6718 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
6719 the required ASN1 format: arbitrary types determined by an OID.
6720 [Steve Henson]
6721
6722 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
6723 argument to 'req'. This is not because the function is newer or
6724 better than others it just uses the work 'NEW' in the certificate
6725 request header lines. Some software needs this.
6726 [Steve Henson]
6727
6728 *) Reorganise password command line arguments: now passwords can be
6729 obtained from various sources. Delete the PEM_cb function and make
6730 it the default behaviour: i.e. if the callback is NULL and the
6731 usrdata argument is not NULL interpret it as a null terminated pass
6732 phrase. If usrdata and the callback are NULL then the pass phrase
6733 is prompted for as usual.
6734 [Steve Henson]
6735
6736 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
6737 the support is automatically enabled. The resulting binaries will
6738 autodetect the card and use it if present.
6739 [Ben Laurie and Compaq Inc.]
6740
6741 *) Work around for Netscape hang bug. This sends certificate request
6742 and server done in one record. Since this is perfectly legal in the
6743 SSL/TLS protocol it isn't a "bug" option and is on by default. See
6744 the bugs/SSLv3 entry for more info.
6745 [Steve Henson]
6746
6747 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
6748 [Andy Polyakov]
6749
6750 *) Add -rand argument to smime and pkcs12 applications and read/write
6751 of seed file.
6752 [Steve Henson]
6753
6754 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
6755 [Bodo Moeller]
6756
6757 *) Add command line password options to the remaining applications.
6758 [Steve Henson]
6759
6760 *) Bug fix for BN_div_recp() for numerators with an even number of
6761 bits.
6762 [Ulf Möller]
6763
6764 *) More tests in bntest.c, and changed test_bn output.
6765 [Ulf Möller]
6766
6767 *) ./config recognizes MacOS X now.
6768 [Andy Polyakov]
6769
6770 *) Bug fix for BN_div() when the first words of num and divsor are
6771 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
6772 [Ulf Möller]
6773
6774 *) Add support for various broken PKCS#8 formats, and command line
6775 options to produce them.
6776 [Steve Henson]
6777
6778 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
6779 get temporary BIGNUMs from a BN_CTX.
6780 [Ulf Möller]
6781
6782 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
6783 for p == 0.
6784 [Ulf Möller]
6785
6786 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
6787 include a #define from the old name to the new. The original intent
6788 was that statically linked binaries could for example just call
6789 SSLeay_add_all_ciphers() to just add ciphers to the table and not
6790 link with digests. This never worked becayse SSLeay_add_all_digests()
6791 and SSLeay_add_all_ciphers() were in the same source file so calling
6792 one would link with the other. They are now in separate source files.
6793 [Steve Henson]
6794
6795 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
6796 [Steve Henson]
6797
6798 *) Use a less unusual form of the Miller-Rabin primality test (it used
6799 a binary algorithm for exponentiation integrated into the Miller-Rabin
6800 loop, our standard modexp algorithms are faster).
6801 [Bodo Moeller]
6802
6803 *) Support for the EBCDIC character set completed.
6804 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
6805
6806 *) Source code cleanups: use const where appropriate, eliminate casts,
6807 use void * instead of char * in lhash.
6808 [Ulf Möller]
6809
6810 *) Bugfix: ssl3_send_server_key_exchange was not restartable
6811 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
6812 this the server could overwrite ephemeral keys that the client
6813 has already seen).
6814 [Bodo Moeller]
6815
6816 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
6817 using 50 iterations of the Rabin-Miller test.
6818
6819 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
6820 iterations of the Rabin-Miller test as required by the appendix
6821 to FIPS PUB 186[-1]) instead of DSA_is_prime.
6822 As BN_is_prime_fasttest includes trial division, DSA parameter
6823 generation becomes much faster.
6824
6825 This implies a change for the callback functions in DSA_is_prime
6826 and DSA_generate_parameters: The callback function is called once
6827 for each positive witness in the Rabin-Miller test, not just
6828 occasionally in the inner loop; and the parameters to the
6829 callback function now provide an iteration count for the outer
6830 loop rather than for the current invocation of the inner loop.
6831 DSA_generate_parameters additionally can call the callback
6832 function with an 'iteration count' of -1, meaning that a
6833 candidate has passed the trial division test (when q is generated
6834 from an application-provided seed, trial division is skipped).
6835 [Bodo Moeller]
6836
6837 *) New function BN_is_prime_fasttest that optionally does trial
6838 division before starting the Rabin-Miller test and has
6839 an additional BN_CTX * argument (whereas BN_is_prime always
6840 has to allocate at least one BN_CTX).
6841 'callback(1, -1, cb_arg)' is called when a number has passed the
6842 trial division stage.
6843 [Bodo Moeller]
6844
6845 *) Fix for bug in CRL encoding. The validity dates weren't being handled
6846 as ASN1_TIME.
6847 [Steve Henson]
6848
6849 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
6850 [Steve Henson]
6851
6852 *) New function BN_pseudo_rand().
6853 [Ulf Möller]
6854
6855 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
6856 bignum version of BN_from_montgomery() with the working code from
6857 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
6858 the comments.
6859 [Ulf Möller]
6860
6861 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
6862 made it impossible to use the same SSL_SESSION data structure in
6863 SSL2 clients in multiple threads.
6864 [Bodo Moeller]
6865
6866 *) The return value of RAND_load_file() no longer counts bytes obtained
6867 by stat(). RAND_load_file(..., -1) is new and uses the complete file
6868 to seed the PRNG (previously an explicit byte count was required).
6869 [Ulf Möller, Bodo Möller]
6870
6871 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
6872 used (char *) instead of (void *) and had casts all over the place.
6873 [Steve Henson]
6874
6875 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
6876 [Ulf Möller]
6877
6878 *) Retain source code compatibility for BN_prime_checks macro:
6879 BN_is_prime(..., BN_prime_checks, ...) now uses
6880 BN_prime_checks_for_size to determine the appropriate number of
6881 Rabin-Miller iterations.
6882 [Ulf Möller]
6883
6884 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
6885 DH_CHECK_P_NOT_SAFE_PRIME.
6886 (Check if this is true? OpenPGP calls them "strong".)
6887 [Ulf Möller]
6888
6889 *) Merge the functionality of "dh" and "gendh" programs into a new program
6890 "dhparam". The old programs are retained for now but will handle DH keys
6891 (instead of parameters) in future.
6892 [Steve Henson]
6893
6894 *) Make the ciphers, s_server and s_client programs check the return values
6895 when a new cipher list is set.
6896 [Steve Henson]
6897
6898 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6899 ciphers. Before when the 56bit ciphers were enabled the sorting was
6900 wrong.
6901
6902 The syntax for the cipher sorting has been extended to support sorting by
6903 cipher-strength (using the strength_bits hard coded in the tables).
6904 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6905
6906 Fix a bug in the cipher-command parser: when supplying a cipher command
6907 string with an "undefined" symbol (neither command nor alphanumeric
6908 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6909 an error is flagged.
6910
6911 Due to the strength-sorting extension, the code of the
6912 ssl_create_cipher_list() function was completely rearranged. I hope that
6913 the readability was also increased :-)
6914 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6915
6916 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6917 for the first serial number and places 2 in the serial number file. This
6918 avoids problems when the root CA is created with serial number zero and
6919 the first user certificate has the same issuer name and serial number
6920 as the root CA.
6921 [Steve Henson]
6922
6923 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6924 the new code. Add documentation for this stuff.
6925 [Steve Henson]
6926
6927 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6928 X509_*() to X509at_*() on the grounds that they don't handle X509
6929 structures and behave in an analagous way to the X509v3 functions:
6930 they shouldn't be called directly but wrapper functions should be used
6931 instead.
6932
6933 So we also now have some wrapper functions that call the X509at functions
6934 when passed certificate requests. (TO DO: similar things can be done with
6935 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6936 things. Some of these need some d2i or i2d and print functionality
6937 because they handle more complex structures.)
6938 [Steve Henson]
6939
6940 *) Add missing #ifndefs that caused missing symbols when building libssl
6941 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6942 NO_RSA in ssl/s2*.c.
6943 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6944
6945 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6946 has a return value which indicates the quality of the random data
6947 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6948 error queue. New function RAND_pseudo_bytes() generates output that is
6949 guaranteed to be unique but not unpredictable. RAND_add is like
6950 RAND_seed, but takes an extra argument for an entropy estimate
6951 (RAND_seed always assumes full entropy).
6952 [Ulf Möller]
6953
6954 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6955 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6956 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6957 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6958 false-positive rate of at most 2^-80 for random input.
6959 [Bodo Moeller]
6960
6961 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6962 [Bodo Moeller]
6963
6964 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6965 in the 0.9.5 release), this returns the chain
6966 from an X509_CTX structure with a dup of the stack and all
6967 the X509 reference counts upped: so the stack will exist
6968 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6969 to use this.
6970
6971 Also make SSL_SESSION_print() print out the verify return
6972 code.
6973 [Steve Henson]
6974
6975 *) Add manpage for the pkcs12 command. Also change the default
6976 behaviour so MAC iteration counts are used unless the new
6977 -nomaciter option is used. This improves file security and
6978 only older versions of MSIE (4.0 for example) need it.
6979 [Steve Henson]
6980
6981 *) Honor the no-xxx Configure options when creating .DEF files.
6982 [Ulf Möller]
6983
6984 *) Add PKCS#10 attributes to field table: challengePassword,
6985 unstructuredName and unstructuredAddress. These are taken from
6986 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6987 international characters are used.
6988
6989 More changes to X509_ATTRIBUTE code: allow the setting of types
6990 based on strings. Remove the 'loc' parameter when adding
6991 attributes because these will be a SET OF encoding which is sorted
6992 in ASN1 order.
6993 [Steve Henson]
6994
6995 *) Initial changes to the 'req' utility to allow request generation
6996 automation. This will allow an application to just generate a template
6997 file containing all the field values and have req construct the
6998 request.
6999
7000 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7001 used all over the place including certificate requests and PKCS#7
7002 structures. They are currently handled manually where necessary with
7003 some primitive wrappers for PKCS#7. The new functions behave in a
7004 manner analogous to the X509 extension functions: they allow
7005 attributes to be looked up by NID and added.
7006
7007 Later something similar to the X509V3 code would be desirable to
7008 automatically handle the encoding, decoding and printing of the
7009 more complex types. The string types like challengePassword can
7010 be handled by the string table functions.
7011
7012 Also modified the multi byte string table handling. Now there is
7013 a 'global mask' which masks out certain types. The table itself
7014 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7015 is useful when for example there is only one permissible type
7016 (as in countryName) and using the mask might result in no valid
7017 types at all.
7018 [Steve Henson]
7019
7020 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7021 SSL_get_peer_finished to allow applications to obtain the latest
7022 Finished messages sent to the peer or expected from the peer,
7023 respectively. (SSL_get_peer_finished is usually the Finished message
7024 actually received from the peer, otherwise the protocol will be aborted.)
7025
7026 As the Finished message are message digests of the complete handshake
7027 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7028 be used for external authentication procedures when the authentication
7029 provided by SSL/TLS is not desired or is not enough.
7030 [Bodo Moeller]
7031
7032 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7033 the host supports BWX extension and if Compaq C is present on the
7034 $PATH. Just exploiting of the BWX extension results in 20-30%
7035 performance kick for some algorithms, e.g. DES and RC4 to mention
7036 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7037 SHA1.
7038 [Andy Polyakov]
7039
7040 *) Add support for MS "fast SGC". This is arguably a violation of the
7041 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7042 weak crypto and after checking the certificate is SGC a second one
7043 with strong crypto. MS SGC stops the first handshake after receiving
7044 the server certificate message and sends a second client hello. Since
7045 a server will typically do all the time consuming operations before
7046 expecting any further messages from the client (server key exchange
7047 is the most expensive) there is little difference between the two.
7048
7049 To get OpenSSL to support MS SGC we have to permit a second client
7050 hello message after we have sent server done. In addition we have to
7051 reset the MAC if we do get this second client hello.
7052 [Steve Henson]
7053
7054 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7055 if a DER encoded private key is RSA or DSA traditional format. Changed
7056 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7057 format DER encoded private key. Newer code should use PKCS#8 format which
7058 has the key type encoded in the ASN1 structure. Added DER private key
7059 support to pkcs8 application.
7060 [Steve Henson]
7061
7062 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7063 ciphersuites has been selected (as required by the SSL 3/TLS 1
7064 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7065 is set, we interpret this as a request to violate the specification
7066 (the worst that can happen is a handshake failure, and 'correct'
7067 behaviour would result in a handshake failure anyway).
7068 [Bodo Moeller]
7069
7070 *) In SSL_CTX_add_session, take into account that there might be multiple
7071 SSL_SESSION structures with the same session ID (e.g. when two threads
7072 concurrently obtain them from an external cache).
7073 The internal cache can handle only one SSL_SESSION with a given ID,
7074 so if there's a conflict, we now throw out the old one to achieve
7075 consistency.
7076 [Bodo Moeller]
7077
7078 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7079 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7080 some routines that use cipher OIDs: some ciphers do not have OIDs
7081 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7082 example.
7083 [Steve Henson]
7084
7085 *) Simplify the trust setting structure and code. Now we just have
7086 two sequences of OIDs for trusted and rejected settings. These will
7087 typically have values the same as the extended key usage extension
7088 and any application specific purposes.
7089
7090 The trust checking code now has a default behaviour: it will just
7091 check for an object with the same NID as the passed id. Functions can
7092 be provided to override either the default behaviour or the behaviour
7093 for a given id. SSL client, server and email already have functions
7094 in place for compatibility: they check the NID and also return "trusted"
7095 if the certificate is self signed.
7096 [Steve Henson]
7097
7098 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7099 traditional format into an EVP_PKEY structure.
7100 [Steve Henson]
7101
7102 *) Add a password callback function PEM_cb() which either prompts for
7103 a password if usr_data is NULL or otherwise assumes it is a null
7104 terminated password. Allow passwords to be passed on command line
7105 environment or config files in a few more utilities.
7106 [Steve Henson]
7107
7108 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7109 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7110 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7111 Update documentation.
7112 [Steve Henson]
7113
7114 *) Support for ASN1 "NULL" type. This could be handled before by using
7115 ASN1_TYPE but there wasn't any function that would try to read a NULL
7116 and produce an error if it couldn't. For compatibility we also have
7117 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7118 don't allocate anything because they don't need to.
7119 [Steve Henson]
7120
7121 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7122 for details.
7123 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7124
7125 *) Rebuild of the memory allocation routines used by OpenSSL code and
7126 possibly others as well. The purpose is to make an interface that
7127 provide hooks so anyone can build a separate set of allocation and
7128 deallocation routines to be used by OpenSSL, for example memory
7129 pool implementations, or something else, which was previously hard
7130 since Malloc(), Realloc() and Free() were defined as macros having
7131 the values malloc, realloc and free, respectively (except for Win32
7132 compilations). The same is provided for memory debugging code.
7133 OpenSSL already comes with functionality to find memory leaks, but
7134 this gives people a chance to debug other memory problems.
7135
7136 With these changes, a new set of functions and macros have appeared:
7137
7138 CRYPTO_set_mem_debug_functions() [F]
7139 CRYPTO_get_mem_debug_functions() [F]
7140 CRYPTO_dbg_set_options() [F]
7141 CRYPTO_dbg_get_options() [F]
7142 CRYPTO_malloc_debug_init() [M]
7143
7144 The memory debug functions are NULL by default, unless the library
7145 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7146 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7147 gives the standard debugging functions that come with OpenSSL) or
7148 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7149 provided by the library user) must be used. When the standard
7150 debugging functions are used, CRYPTO_dbg_set_options can be used to
7151 request additional information:
7152 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7153 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7154
7155 Also, things like CRYPTO_set_mem_functions will always give the
7156 expected result (the new set of functions is used for allocation
7157 and deallocation) at all times, regardless of platform and compiler
7158 options.
7159
7160 To finish it up, some functions that were never use in any other
7161 way than through macros have a new API and new semantic:
7162
7163 CRYPTO_dbg_malloc()
7164 CRYPTO_dbg_realloc()
7165 CRYPTO_dbg_free()
7166
7167 All macros of value have retained their old syntax.
7168 [Richard Levitte and Bodo Moeller]
7169
7170 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7171 ordering of SMIMECapabilities wasn't in "strength order" and there
7172 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7173 algorithm.
7174 [Steve Henson]
7175
7176 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7177 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7178 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7179
7180 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7181 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7182 functionality to handle multipart/signed properly) and a utility
7183 called 'smime' to call all this stuff. This is based on code I
7184 originally wrote for Celo who have kindly allowed it to be
7185 included in OpenSSL.
7186 [Steve Henson]
7187
7188 *) Add variants des_set_key_checked and des_set_key_unchecked of
7189 des_set_key (aka des_key_sched). Global variable des_check_key
7190 decides which of these is called by des_set_key; this way
7191 des_check_key behaves as it always did, but applications and
7192 the library itself, which was buggy for des_check_key == 1,
7193 have a cleaner way to pick the version they need.
7194 [Bodo Moeller]
7195
7196 *) New function PKCS12_newpass() which changes the password of a
7197 PKCS12 structure.
7198 [Steve Henson]
7199
7200 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7201 dynamic mix. In both cases the ids can be used as an index into the
7202 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7203 functions so they accept a list of the field values and the
7204 application doesn't need to directly manipulate the X509_TRUST
7205 structure.
7206 [Steve Henson]
7207
7208 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7209 need initialising.
7210 [Steve Henson]
7211
7212 *) Modify the way the V3 extension code looks up extensions. This now
7213 works in a similar way to the object code: we have some "standard"
7214 extensions in a static table which is searched with OBJ_bsearch()
7215 and the application can add dynamic ones if needed. The file
7216 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7217 updated whenever a new extension is added to the core code and kept
7218 in ext_nid order. There is a simple program 'tabtest.c' which checks
7219 this. New extensions are not added too often so this file can readily
7220 be maintained manually.
7221
7222 There are two big advantages in doing things this way. The extensions
7223 can be looked up immediately and no longer need to be "added" using
7224 X509V3_add_standard_extensions(): this function now does nothing.
7225 [Side note: I get *lots* of email saying the extension code doesn't
7226 work because people forget to call this function]
7227 Also no dynamic allocation is done unless new extensions are added:
7228 so if we don't add custom extensions there is no need to call
7229 X509V3_EXT_cleanup().
7230 [Steve Henson]
7231
7232 *) Modify enc utility's salting as follows: make salting the default. Add a
7233 magic header, so unsalted files fail gracefully instead of just decrypting
7234 to garbage. This is because not salting is a big security hole, so people
7235 should be discouraged from doing it.
7236 [Ben Laurie]
7237
7238 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7239 digest to be passed on the command line but it only used this
7240 parameter when signing a certificate. Modified so all relevant
7241 operations are affected by the digest parameter including the
7242 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7243 DSA key was used because it didn't fix the digest.
7244 [Steve Henson]
7245
7246 *) Initial certificate chain verify code. Currently tests the untrusted
7247 certificates for consistency with the verify purpose (which is set
7248 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7249
7250 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7251 this is because it will reject chains with invalid extensions whereas
7252 every previous version of OpenSSL and SSLeay made no checks at all.
7253
7254 Trust code: checks the root CA for the relevant trust settings. Trust
7255 settings have an initial value consistent with the verify purpose: e.g.
7256 if the verify purpose is for SSL client use it expects the CA to be
7257 trusted for SSL client use. However the default value can be changed to
7258 permit custom trust settings: one example of this would be to only trust
7259 certificates from a specific "secure" set of CAs.
7260
7261 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7262 which should be used for version portability: especially since the
7263 verify structure is likely to change more often now.
7264
7265 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7266 to set them. If not set then assume SSL clients will verify SSL servers
7267 and vice versa.
7268
7269 Two new options to the verify program: -untrusted allows a set of
7270 untrusted certificates to be passed in and -purpose which sets the
7271 intended purpose of the certificate. If a purpose is set then the
7272 new chain verify code is used to check extension consistency.
7273 [Steve Henson]
7274
7275 *) Support for the authority information access extension.
7276 [Steve Henson]
7277
7278 *) Modify RSA and DSA PEM read routines to transparently handle
7279 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7280 public keys in a format compatible with certificate
7281 SubjectPublicKeyInfo structures. Unfortunately there were already
7282 functions called *_PublicKey_* which used various odd formats so
7283 these are retained for compatibility: however the DSA variants were
7284 never in a public release so they have been deleted. Changed dsa/rsa
7285 utilities to handle the new format: note no releases ever handled public
7286 keys so we should be OK.
7287
7288 The primary motivation for this change is to avoid the same fiasco
7289 that dogs private keys: there are several incompatible private key
7290 formats some of which are standard and some OpenSSL specific and
7291 require various evil hacks to allow partial transparent handling and
7292 even then it doesn't work with DER formats. Given the option anything
7293 other than PKCS#8 should be dumped: but the other formats have to
7294 stay in the name of compatibility.
7295
7296 With public keys and the benefit of hindsight one standard format
7297 is used which works with EVP_PKEY, RSA or DSA structures: though
7298 it clearly returns an error if you try to read the wrong kind of key.
7299
7300 Added a -pubkey option to the 'x509' utility to output the public key.
7301 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7302 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7303 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7304 that do the same as the EVP_PKEY_assign_*() except they up the
7305 reference count of the added key (they don't "swallow" the
7306 supplied key).
7307 [Steve Henson]
7308
7309 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7310 CRLs would fail if the file contained no certificates or no CRLs:
7311 added a new function to read in both types and return the number
7312 read: this means that if none are read it will be an error. The
7313 DER versions of the certificate and CRL reader would always fail
7314 because it isn't possible to mix certificates and CRLs in DER format
7315 without choking one or the other routine. Changed this to just read
7316 a certificate: this is the best we can do. Also modified the code
7317 in apps/verify.c to take notice of return codes: it was previously
7318 attempting to read in certificates from NULL pointers and ignoring
7319 any errors: this is one reason why the cert and CRL reader seemed
7320 to work. It doesn't check return codes from the default certificate
7321 routines: these may well fail if the certificates aren't installed.
7322 [Steve Henson]
7323
7324 *) Code to support otherName option in GeneralName.
7325 [Steve Henson]
7326
7327 *) First update to verify code. Change the verify utility
7328 so it warns if it is passed a self signed certificate:
7329 for consistency with the normal behaviour. X509_verify
7330 has been modified to it will now verify a self signed
7331 certificate if *exactly* the same certificate appears
7332 in the store: it was previously impossible to trust a
7333 single self signed certificate. This means that:
7334 openssl verify ss.pem
7335 now gives a warning about a self signed certificate but
7336 openssl verify -CAfile ss.pem ss.pem
7337 is OK.
7338 [Steve Henson]
7339
7340 *) For servers, store verify_result in SSL_SESSION data structure
7341 (and add it to external session representation).
7342 This is needed when client certificate verifications fails,
7343 but an application-provided verification callback (set by
7344 SSL_CTX_set_cert_verify_callback) allows accepting the session
7345 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7346 but returns 1): When the session is reused, we have to set
7347 ssl->verify_result to the appropriate error code to avoid
7348 security holes.
7349 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7350
7351 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7352 case in PKCS7_dataInit() where the signed PKCS7 structure
7353 didn't contain any existing data because it was being created.
7354 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7355
7356 *) Add a salt to the key derivation routines in enc.c. This
7357 forms the first 8 bytes of the encrypted file. Also add a
7358 -S option to allow a salt to be input on the command line.
7359 [Steve Henson]
7360
7361 *) New function X509_cmp(). Oddly enough there wasn't a function
7362 to compare two certificates. We do this by working out the SHA1
7363 hash and comparing that. X509_cmp() will be needed by the trust
7364 code.
7365 [Steve Henson]
7366
7367 *) SSL_get1_session() is like SSL_get_session(), but increments
7368 the reference count in the SSL_SESSION returned.
7369 [Geoff Thorpe <geoff@eu.c2.net>]
7370
7371 *) Fix for 'req': it was adding a null to request attributes.
7372 Also change the X509_LOOKUP and X509_INFO code to handle
7373 certificate auxiliary information.
7374 [Steve Henson]
7375
7376 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7377 the 'enc' command.
7378 [Steve Henson]
7379
7380 *) Add the possibility to add extra information to the memory leak
7381 detecting output, to form tracebacks, showing from where each
7382 allocation was originated: CRYPTO_push_info("constant string") adds
7383 the string plus current file name and line number to a per-thread
7384 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7385 is like calling CYRPTO_pop_info() until the stack is empty.
7386 Also updated memory leak detection code to be multi-thread-safe.
7387 [Richard Levitte]
7388
7389 *) Add options -text and -noout to pkcs7 utility and delete the
7390 encryption options which never did anything. Update docs.
7391 [Steve Henson]
7392
7393 *) Add options to some of the utilities to allow the pass phrase
7394 to be included on either the command line (not recommended on
7395 OSes like Unix) or read from the environment. Update the
7396 manpages and fix a few bugs.
7397 [Steve Henson]
7398
7399 *) Add a few manpages for some of the openssl commands.
7400 [Steve Henson]
7401
7402 *) Fix the -revoke option in ca. It was freeing up memory twice,
7403 leaking and not finding already revoked certificates.
7404 [Steve Henson]
7405
7406 *) Extensive changes to support certificate auxiliary information.
7407 This involves the use of X509_CERT_AUX structure and X509_AUX
7408 functions. An X509_AUX function such as PEM_read_X509_AUX()
7409 can still read in a certificate file in the usual way but it
7410 will also read in any additional "auxiliary information". By
7411 doing things this way a fair degree of compatibility can be
7412 retained: existing certificates can have this information added
7413 using the new 'x509' options.
7414
7415 Current auxiliary information includes an "alias" and some trust
7416 settings. The trust settings will ultimately be used in enhanced
7417 certificate chain verification routines: currently a certificate
7418 can only be trusted if it is self signed and then it is trusted
7419 for all purposes.
7420 [Steve Henson]
7421
7422 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7423 The problem was that one of the replacement routines had not been working
7424 since SSLeay releases. For now the offending routine has been replaced
7425 with non-optimised assembler. Even so, this now gives around 95%
7426 performance improvement for 1024 bit RSA signs.
7427 [Mark Cox]
7428
7429 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7430 handling. Most clients have the effective key size in bits equal to
7431 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
7432 A few however don't do this and instead use the size of the decrypted key
7433 to determine the RC2 key length and the AlgorithmIdentifier to determine
7434 the effective key length. In this case the effective key length can still
7435 be 40 bits but the key length can be 168 bits for example. This is fixed
7436 by manually forcing an RC2 key into the EVP_PKEY structure because the
7437 EVP code can't currently handle unusual RC2 key sizes: it always assumes
7438 the key length and effective key length are equal.
7439 [Steve Henson]
7440
7441 *) Add a bunch of functions that should simplify the creation of
7442 X509_NAME structures. Now you should be able to do:
7443 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
7444 and have it automatically work out the correct field type and fill in
7445 the structures. The more adventurous can try:
7446 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
7447 and it will (hopefully) work out the correct multibyte encoding.
7448 [Steve Henson]
7449
7450 *) Change the 'req' utility to use the new field handling and multibyte
7451 copy routines. Before the DN field creation was handled in an ad hoc
7452 way in req, ca, and x509 which was rather broken and didn't support
7453 BMPStrings or UTF8Strings. Since some software doesn't implement
7454 BMPStrings or UTF8Strings yet, they can be enabled using the config file
7455 using the dirstring_type option. See the new comment in the default
7456 openssl.cnf for more info.
7457 [Steve Henson]
7458
7459 *) Make crypto/rand/md_rand.c more robust:
7460 - Assure unique random numbers after fork().
7461 - Make sure that concurrent threads access the global counter and
7462 md serializably so that we never lose entropy in them
7463 or use exactly the same state in multiple threads.
7464 Access to the large state is not always serializable because
7465 the additional locking could be a performance killer, and
7466 md should be large enough anyway.
7467 [Bodo Moeller]
7468
7469 *) New file apps/app_rand.c with commonly needed functionality
7470 for handling the random seed file.
7471
7472 Use the random seed file in some applications that previously did not:
7473 ca,
7474 dsaparam -genkey (which also ignored its '-rand' option),
7475 s_client,
7476 s_server,
7477 x509 (when signing).
7478 Except on systems with /dev/urandom, it is crucial to have a random
7479 seed file at least for key creation, DSA signing, and for DH exchanges;
7480 for RSA signatures we could do without one.
7481
7482 gendh and gendsa (unlike genrsa) used to read only the first byte
7483 of each file listed in the '-rand' option. The function as previously
7484 found in genrsa is now in app_rand.c and is used by all programs
7485 that support '-rand'.
7486 [Bodo Moeller]
7487
7488 *) In RAND_write_file, use mode 0600 for creating files;
7489 don't just chmod when it may be too late.
7490 [Bodo Moeller]
7491
7492 *) Report an error from X509_STORE_load_locations
7493 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
7494 [Bill Perry]
7495
7496 *) New function ASN1_mbstring_copy() this copies a string in either
7497 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
7498 into an ASN1_STRING type. A mask of permissible types is passed
7499 and it chooses the "minimal" type to use or an error if not type
7500 is suitable.
7501 [Steve Henson]
7502
7503 *) Add function equivalents to the various macros in asn1.h. The old
7504 macros are retained with an M_ prefix. Code inside the library can
7505 use the M_ macros. External code (including the openssl utility)
7506 should *NOT* in order to be "shared library friendly".
7507 [Steve Henson]
7508
7509 *) Add various functions that can check a certificate's extensions
7510 to see if it usable for various purposes such as SSL client,
7511 server or S/MIME and CAs of these types. This is currently
7512 VERY EXPERIMENTAL but will ultimately be used for certificate chain
7513 verification. Also added a -purpose flag to x509 utility to
7514 print out all the purposes.
7515 [Steve Henson]
7516
7517 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
7518 functions.
7519 [Steve Henson]
7520
7521 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
7522 for, obtain and decode and extension and obtain its critical flag.
7523 This allows all the necessary extension code to be handled in a
7524 single function call.
7525 [Steve Henson]
7526
7527 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
7528 platforms. See crypto/rc4/rc4_enc.c for further details.
7529 [Andy Polyakov]
7530
7531 *) New -noout option to asn1parse. This causes no output to be produced
7532 its main use is when combined with -strparse and -out to extract data
7533 from a file (which may not be in ASN.1 format).
7534 [Steve Henson]
7535
7536 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
7537 when producing the local key id.
7538 [Richard Levitte <levitte@stacken.kth.se>]
7539
7540 *) New option -dhparam in s_server. This allows a DH parameter file to be
7541 stated explicitly. If it is not stated then it tries the first server
7542 certificate file. The previous behaviour hard coded the filename
7543 "server.pem".
7544 [Steve Henson]
7545
7546 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
7547 a public key to be input or output. For example:
7548 openssl rsa -in key.pem -pubout -out pubkey.pem
7549 Also added necessary DSA public key functions to handle this.
7550 [Steve Henson]
7551
7552 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
7553 in the message. This was handled by allowing
7554 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
7555 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
7556
7557 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
7558 to the end of the strings whereas this didn't. This would cause problems
7559 if strings read with d2i_ASN1_bytes() were later modified.
7560 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
7561
7562 *) Fix for base64 decode bug. When a base64 bio reads only one line of
7563 data and it contains EOF it will end up returning an error. This is
7564 caused by input 46 bytes long. The cause is due to the way base64
7565 BIOs find the start of base64 encoded data. They do this by trying a
7566 trial decode on each line until they find one that works. When they
7567 do a flag is set and it starts again knowing it can pass all the
7568 data directly through the decoder. Unfortunately it doesn't reset
7569 the context it uses. This means that if EOF is reached an attempt
7570 is made to pass two EOFs through the context and this causes the
7571 resulting error. This can also cause other problems as well. As is
7572 usual with these problems it takes *ages* to find and the fix is
7573 trivial: move one line.
7574 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
7575
7576 *) Ugly workaround to get s_client and s_server working under Windows. The
7577 old code wouldn't work because it needed to select() on sockets and the
7578 tty (for keypresses and to see if data could be written). Win32 only
7579 supports select() on sockets so we select() with a 1s timeout on the
7580 sockets and then see if any characters are waiting to be read, if none
7581 are present then we retry, we also assume we can always write data to
7582 the tty. This isn't nice because the code then blocks until we've
7583 received a complete line of data and it is effectively polling the
7584 keyboard at 1s intervals: however it's quite a bit better than not
7585 working at all :-) A dedicated Windows application might handle this
7586 with an event loop for example.
7587 [Steve Henson]
7588
7589 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
7590 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
7591 will be called when RSA_sign() and RSA_verify() are used. This is useful
7592 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
7593 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
7594 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
7595 This necessitated the support of an extra signature type NID_md5_sha1
7596 for SSL signatures and modifications to the SSL library to use it instead
7597 of calling RSA_public_decrypt() and RSA_private_encrypt().
7598 [Steve Henson]
7599
7600 *) Add new -verify -CAfile and -CApath options to the crl program, these
7601 will lookup a CRL issuers certificate and verify the signature in a
7602 similar way to the verify program. Tidy up the crl program so it
7603 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
7604 less strict. It will now permit CRL extensions even if it is not
7605 a V2 CRL: this will allow it to tolerate some broken CRLs.
7606 [Steve Henson]
7607
7608 *) Initialize all non-automatic variables each time one of the openssl
7609 sub-programs is started (this is necessary as they may be started
7610 multiple times from the "OpenSSL>" prompt).
7611 [Lennart Bang, Bodo Moeller]
7612
7613 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
7614 removing all other RSA functionality (this is what NO_RSA does). This
7615 is so (for example) those in the US can disable those operations covered
7616 by the RSA patent while allowing storage and parsing of RSA keys and RSA
7617 key generation.
7618 [Steve Henson]
7619
7620 *) Non-copying interface to BIO pairs.
7621 (still largely untested)
7622 [Bodo Moeller]
7623
7624 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
7625 ASCII string. This was handled independently in various places before.
7626 [Steve Henson]
7627
7628 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
7629 UTF8 strings a character at a time.
7630 [Steve Henson]
7631
7632 *) Use client_version from client hello to select the protocol
7633 (s23_srvr.c) and for RSA client key exchange verification
7634 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
7635 [Bodo Moeller]
7636
7637 *) Add various utility functions to handle SPKACs, these were previously
7638 handled by poking round in the structure internals. Added new function
7639 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
7640 print, verify and generate SPKACs. Based on an original idea from
7641 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
7642 [Steve Henson]
7643
7644 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
7645 [Andy Polyakov]
7646
7647 *) Allow the config file extension section to be overwritten on the
7648 command line. Based on an original idea from Massimiliano Pala
7649 <madwolf@comune.modena.it>. The new option is called -extensions
7650 and can be applied to ca, req and x509. Also -reqexts to override
7651 the request extensions in req and -crlexts to override the crl extensions
7652 in ca.
7653 [Steve Henson]
7654
7655 *) Add new feature to the SPKAC handling in ca. Now you can include
7656 the same field multiple times by preceding it by "XXXX." for example:
7657 1.OU="Unit name 1"
7658 2.OU="Unit name 2"
7659 this is the same syntax as used in the req config file.
7660 [Steve Henson]
7661
7662 *) Allow certificate extensions to be added to certificate requests. These
7663 are specified in a 'req_extensions' option of the req section of the
7664 config file. They can be printed out with the -text option to req but
7665 are otherwise ignored at present.
7666 [Steve Henson]
7667
7668 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
7669 data read consists of only the final block it would not decrypted because
7670 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
7671 A misplaced 'break' also meant the decrypted final block might not be
7672 copied until the next read.
7673 [Steve Henson]
7674
7675 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
7676 a few extra parameters to the DH structure: these will be useful if
7677 for example we want the value of 'q' or implement X9.42 DH.
7678 [Steve Henson]
7679
7680 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
7681 provides hooks that allow the default DSA functions or functions on a
7682 "per key" basis to be replaced. This allows hardware acceleration and
7683 hardware key storage to be handled without major modification to the
7684 library. Also added low level modexp hooks and CRYPTO_EX structure and
7685 associated functions.
7686 [Steve Henson]
7687
7688 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
7689 as "read only": it can't be written to and the buffer it points to will
7690 not be freed. Reading from a read only BIO is much more efficient than
7691 a normal memory BIO. This was added because there are several times when
7692 an area of memory needs to be read from a BIO. The previous method was
7693 to create a memory BIO and write the data to it, this results in two
7694 copies of the data and an O(n^2) reading algorithm. There is a new
7695 function BIO_new_mem_buf() which creates a read only memory BIO from
7696 an area of memory. Also modified the PKCS#7 routines to use read only
7697 memory BIOs.
7698 [Steve Henson]
7699
7700 *) Bugfix: ssl23_get_client_hello did not work properly when called in
7701 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
7702 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
7703 but a retry condition occured while trying to read the rest.
7704 [Bodo Moeller]
7705
7706 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
7707 NID_pkcs7_encrypted by default: this was wrong since this should almost
7708 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
7709 the encrypted data type: this is a more sensible place to put it and it
7710 allows the PKCS#12 code to be tidied up that duplicated this
7711 functionality.
7712 [Steve Henson]
7713
7714 *) Changed obj_dat.pl script so it takes its input and output files on
7715 the command line. This should avoid shell escape redirection problems
7716 under Win32.
7717 [Steve Henson]
7718
7719 *) Initial support for certificate extension requests, these are included
7720 in things like Xenroll certificate requests. Included functions to allow
7721 extensions to be obtained and added.
7722 [Steve Henson]
7723
7724 *) -crlf option to s_client and s_server for sending newlines as
7725 CRLF (as required by many protocols).
7726 [Bodo Moeller]
7727
7728 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
7729
7730 *) Install libRSAglue.a when OpenSSL is built with RSAref.
7731 [Ralf S. Engelschall]
7732
7733 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
7734 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
7735
7736 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
7737 program.
7738 [Steve Henson]
7739
7740 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
7741 DH parameters/keys (q is lost during that conversion, but the resulting
7742 DH parameters contain its length).
7743
7744 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
7745 much faster than DH_generate_parameters (which creates parameters
7746 where p = 2*q + 1), and also the smaller q makes DH computations
7747 much more efficient (160-bit exponentiation instead of 1024-bit
7748 exponentiation); so this provides a convenient way to support DHE
7749 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
7750 utter importance to use
7751 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7752 or
7753 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7754 when such DH parameters are used, because otherwise small subgroup
7755 attacks may become possible!
7756 [Bodo Moeller]
7757
7758 *) Avoid memory leak in i2d_DHparams.
7759 [Bodo Moeller]
7760
7761 *) Allow the -k option to be used more than once in the enc program:
7762 this allows the same encrypted message to be read by multiple recipients.
7763 [Steve Henson]
7764
7765 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
7766 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
7767 it will always use the numerical form of the OID, even if it has a short
7768 or long name.
7769 [Steve Henson]
7770
7771 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
7772 method only got called if p,q,dmp1,dmq1,iqmp components were present,
7773 otherwise bn_mod_exp was called. In the case of hardware keys for example
7774 no private key components need be present and it might store extra data
7775 in the RSA structure, which cannot be accessed from bn_mod_exp.
7776 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
7777 private key operations.
7778 [Steve Henson]
7779
7780 *) Added support for SPARC Linux.
7781 [Andy Polyakov]
7782
7783 *) pem_password_cb function type incompatibly changed from
7784 typedef int pem_password_cb(char *buf, int size, int rwflag);
7785 to
7786 ....(char *buf, int size, int rwflag, void *userdata);
7787 so that applications can pass data to their callbacks:
7788 The PEM[_ASN1]_{read,write}... functions and macros now take an
7789 additional void * argument, which is just handed through whenever
7790 the password callback is called.
7791 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
7792
7793 New function SSL_CTX_set_default_passwd_cb_userdata.
7794
7795 Compatibility note: As many C implementations push function arguments
7796 onto the stack in reverse order, the new library version is likely to
7797 interoperate with programs that have been compiled with the old
7798 pem_password_cb definition (PEM_whatever takes some data that
7799 happens to be on the stack as its last argument, and the callback
7800 just ignores this garbage); but there is no guarantee whatsoever that
7801 this will work.
7802
7803 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
7804 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
7805 problems not only on Windows, but also on some Unix platforms.
7806 To avoid problematic command lines, these definitions are now in an
7807 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
7808 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
7809 [Bodo Moeller]
7810
7811 *) MIPS III/IV assembler module is reimplemented.
7812 [Andy Polyakov]
7813
7814 *) More DES library cleanups: remove references to srand/rand and
7815 delete an unused file.
7816 [Ulf Möller]
7817
7818 *) Add support for the the free Netwide assembler (NASM) under Win32,
7819 since not many people have MASM (ml) and it can be hard to obtain.
7820 This is currently experimental but it seems to work OK and pass all
7821 the tests. Check out INSTALL.W32 for info.
7822 [Steve Henson]
7823
7824 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
7825 without temporary keys kept an extra copy of the server key,
7826 and connections with temporary keys did not free everything in case
7827 of an error.
7828 [Bodo Moeller]
7829
7830 *) New function RSA_check_key and new openssl rsa option -check
7831 for verifying the consistency of RSA keys.
7832 [Ulf Moeller, Bodo Moeller]
7833
7834 *) Various changes to make Win32 compile work:
7835 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
7836 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
7837 comparison" warnings.
7838 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
7839 [Steve Henson]
7840
7841 *) Add a debugging option to PKCS#5 v2 key generation function: when
7842 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
7843 derived keys are printed to stderr.
7844 [Steve Henson]
7845
7846 *) Copy the flags in ASN1_STRING_dup().
7847 [Roman E. Pavlov <pre@mo.msk.ru>]
7848
7849 *) The x509 application mishandled signing requests containing DSA
7850 keys when the signing key was also DSA and the parameters didn't match.
7851
7852 It was supposed to omit the parameters when they matched the signing key:
7853 the verifying software was then supposed to automatically use the CA's
7854 parameters if they were absent from the end user certificate.
7855
7856 Omitting parameters is no longer recommended. The test was also
7857 the wrong way round! This was probably due to unusual behaviour in
7858 EVP_cmp_parameters() which returns 1 if the parameters match.
7859 This meant that parameters were omitted when they *didn't* match and
7860 the certificate was useless. Certificates signed with 'ca' didn't have
7861 this bug.
7862 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
7863
7864 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
7865 The interface is as follows:
7866 Applications can use
7867 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
7868 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
7869 "off" is now the default.
7870 The library internally uses
7871 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
7872 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
7873 to disable memory-checking temporarily.
7874
7875 Some inconsistent states that previously were possible (and were
7876 even the default) are now avoided.
7877
7878 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
7879 with each memory chunk allocated; this is occasionally more helpful
7880 than just having a counter.
7881
7882 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
7883
7884 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
7885 extensions.
7886 [Bodo Moeller]
7887
7888 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
7889 which largely parallels "options", but is for changing API behaviour,
7890 whereas "options" are about protocol behaviour.
7891 Initial "mode" flags are:
7892
7893 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7894 a single record has been written.
7895 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7896 retries use the same buffer location.
7897 (But all of the contents must be
7898 copied!)
7899 [Bodo Moeller]
7900
7901 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7902 worked.
7903
7904 *) Fix problems with no-hmac etc.
7905 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7906
7907 *) New functions RSA_get_default_method(), RSA_set_method() and
7908 RSA_get_method(). These allows replacement of RSA_METHODs without having
7909 to mess around with the internals of an RSA structure.
7910 [Steve Henson]
7911
7912 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7913 Also really enable memory leak checks in openssl.c and in some
7914 test programs.
7915 [Chad C. Mulligan, Bodo Moeller]
7916
7917 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7918 up the length of negative integers. This has now been simplified to just
7919 store the length when it is first determined and use it later, rather
7920 than trying to keep track of where data is copied and updating it to
7921 point to the end.
7922 [Steve Henson, reported by Brien Wheeler
7923 <bwheeler@authentica-security.com>]
7924
7925 *) Add a new function PKCS7_signatureVerify. This allows the verification
7926 of a PKCS#7 signature but with the signing certificate passed to the
7927 function itself. This contrasts with PKCS7_dataVerify which assumes the
7928 certificate is present in the PKCS#7 structure. This isn't always the
7929 case: certificates can be omitted from a PKCS#7 structure and be
7930 distributed by "out of band" means (such as a certificate database).
7931 [Steve Henson]
7932
7933 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7934 function prototypes in pem.h, also change util/mkdef.pl to add the
7935 necessary function names.
7936 [Steve Henson]
7937
7938 *) mk1mf.pl (used by Windows builds) did not properly read the
7939 options set by Configure in the top level Makefile, and Configure
7940 was not even able to write more than one option correctly.
7941 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7942 [Bodo Moeller]
7943
7944 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7945 file to be loaded from a BIO or FILE pointer. The BIO version will
7946 for example allow memory BIOs to contain config info.
7947 [Steve Henson]
7948
7949 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7950 Whoever hopes to achieve shared-library compatibility across versions
7951 must use this, not the compile-time macro.
7952 (Exercise 0.9.4: Which is the minimum library version required by
7953 such programs?)
7954 Note: All this applies only to multi-threaded programs, others don't
7955 need locks.
7956 [Bodo Moeller]
7957
7958 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7959 through a BIO pair triggered the default case, i.e.
7960 SSLerr(...,SSL_R_UNKNOWN_STATE).
7961 [Bodo Moeller]
7962
7963 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7964 can use the SSL library even if none of the specific BIOs is
7965 appropriate.
7966 [Bodo Moeller]
7967
7968 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7969 for the encoded length.
7970 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7971
7972 *) Add initial documentation of the X509V3 functions.
7973 [Steve Henson]
7974
7975 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7976 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7977 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7978 secure PKCS#8 private key format with a high iteration count.
7979 [Steve Henson]
7980
7981 *) Fix determination of Perl interpreter: A perl or perl5
7982 _directory_ in $PATH was also accepted as the interpreter.
7983 [Ralf S. Engelschall]
7984
7985 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7986 wrong with it but it was very old and did things like calling
7987 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7988 unusual formatting.
7989 [Steve Henson]
7990
7991 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7992 to use the new extension code.
7993 [Steve Henson]
7994
7995 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7996 with macros. This should make it easier to change their form, add extra
7997 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7998 constant.
7999 [Steve Henson]
8000
8001 *) Add to configuration table a new entry that can specify an alternative
8002 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8003 according to Mark Crispin <MRC@Panda.COM>.
8004 [Bodo Moeller]
8005
8006 #if 0
8007 *) DES CBC did not update the IV. Weird.
8008 [Ben Laurie]
8009 #else
8010 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8011 Changing the behaviour of the former might break existing programs --
8012 where IV updating is needed, des_ncbc_encrypt can be used.
8013 #endif
8014
8015 *) When bntest is run from "make test" it drives bc to check its
8016 calculations, as well as internally checking them. If an internal check
8017 fails, it needs to cause bc to give a non-zero result or make test carries
8018 on without noticing the failure. Fixed.
8019 [Ben Laurie]
8020
8021 *) DES library cleanups.
8022 [Ulf Möller]
8023
8024 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8025 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8026 ciphers. NOTE: although the key derivation function has been verified
8027 against some published test vectors it has not been extensively tested
8028 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8029 of v2.0.
8030 [Steve Henson]
8031
8032 *) Instead of "mkdir -p", which is not fully portable, use new
8033 Perl script "util/mkdir-p.pl".
8034 [Bodo Moeller]
8035
8036 *) Rewrite the way password based encryption (PBE) is handled. It used to
8037 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8038 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8039 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8040 the 'parameter' field of the AlgorithmIdentifier is passed to the
8041 underlying key generation function so it must do its own ASN1 parsing.
8042 This has also changed the EVP_PBE_CipherInit() function which now has a
8043 'parameter' argument instead of literal salt and iteration count values
8044 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8045 [Steve Henson]
8046
8047 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8048 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8049 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8050 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8051 value was just used as a "magic string" and not used directly its
8052 value doesn't matter.
8053 [Steve Henson]
8054
8055 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8056 support mutable.
8057 [Ben Laurie]
8058
8059 *) "linux-sparc64" configuration (ultrapenguin).
8060 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8061 "linux-sparc" configuration.
8062 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8063
8064 *) config now generates no-xxx options for missing ciphers.
8065 [Ulf Möller]
8066
8067 *) Support the EBCDIC character set (work in progress).
8068 File ebcdic.c not yet included because it has a different license.
8069 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8070
8071 *) Support BS2000/OSD-POSIX.
8072 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8073
8074 *) Make callbacks for key generation use void * instead of char *.
8075 [Ben Laurie]
8076
8077 *) Make S/MIME samples compile (not yet tested).
8078 [Ben Laurie]
8079
8080 *) Additional typesafe stacks.
8081 [Ben Laurie]
8082
8083 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8084 [Bodo Moeller]
8085
8086
8087 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8088
8089 *) New configuration variant "sco5-gcc".
8090
8091 *) Updated some demos.
8092 [Sean O Riordain, Wade Scholine]
8093
8094 *) Add missing BIO_free at exit of pkcs12 application.
8095 [Wu Zhigang]
8096
8097 *) Fix memory leak in conf.c.
8098 [Steve Henson]
8099
8100 *) Updates for Win32 to assembler version of MD5.
8101 [Steve Henson]
8102
8103 *) Set #! path to perl in apps/der_chop to where we found it
8104 instead of using a fixed path.
8105 [Bodo Moeller]
8106
8107 *) SHA library changes for irix64-mips4-cc.
8108 [Andy Polyakov]
8109
8110 *) Improvements for VMS support.
8111 [Richard Levitte]
8112
8113
8114 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8115
8116 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8117 This also avoids the problems with SC4.2 and unpatched SC5.
8118 [Andy Polyakov <appro@fy.chalmers.se>]
8119
8120 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8121 These are required because of the typesafe stack would otherwise break
8122 existing code. If old code used a structure member which used to be STACK
8123 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8124 sk_num or sk_value it would produce an error because the num, data members
8125 are not present in STACK_OF. Now it just produces a warning. sk_set
8126 replaces the old method of assigning a value to sk_value
8127 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8128 that does this will no longer work (and should use sk_set instead) but
8129 this could be regarded as a "questionable" behaviour anyway.
8130 [Steve Henson]
8131
8132 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8133 correctly handle encrypted S/MIME data.
8134 [Steve Henson]
8135
8136 *) Change type of various DES function arguments from des_cblock
8137 (which means, in function argument declarations, pointer to char)
8138 to des_cblock * (meaning pointer to array with 8 char elements),
8139 which allows the compiler to do more typechecking; it was like
8140 that back in SSLeay, but with lots of ugly casts.
8141
8142 Introduce new type const_des_cblock.
8143 [Bodo Moeller]
8144
8145 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8146 problems: find RecipientInfo structure that matches recipient certificate
8147 and initialise the ASN1 structures properly based on passed cipher.
8148 [Steve Henson]
8149
8150 *) Belatedly make the BN tests actually check the results.
8151 [Ben Laurie]
8152
8153 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8154 to and from BNs: it was completely broken. New compilation option
8155 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8156 key elements as negative integers.
8157 [Steve Henson]
8158
8159 *) Reorganize and speed up MD5.
8160 [Andy Polyakov <appro@fy.chalmers.se>]
8161
8162 *) VMS support.
8163 [Richard Levitte <richard@levitte.org>]
8164
8165 *) New option -out to asn1parse to allow the parsed structure to be
8166 output to a file. This is most useful when combined with the -strparse
8167 option to examine the output of things like OCTET STRINGS.
8168 [Steve Henson]
8169
8170 *) Make SSL library a little more fool-proof by not requiring any longer
8171 that SSL_set_{accept,connect}_state be called before
8172 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8173 in many applications because usually everything *appeared* to work as
8174 intended anyway -- now it really works as intended).
8175 [Bodo Moeller]
8176
8177 *) Move openssl.cnf out of lib/.
8178 [Ulf Möller]
8179
8180 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8181 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8182 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8183 [Ralf S. Engelschall]
8184
8185 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8186 handle PKCS#7 enveloped data properly.
8187 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8188
8189 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8190 copying pointers. The cert_st handling is changed by this in
8191 various ways (and thus what used to be known as ctx->default_cert
8192 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8193 any longer when s->cert does not give us what we need).
8194 ssl_cert_instantiate becomes obsolete by this change.
8195 As soon as we've got the new code right (possibly it already is?),
8196 we have solved a couple of bugs of the earlier code where s->cert
8197 was used as if it could not have been shared with other SSL structures.
8198
8199 Note that using the SSL API in certain dirty ways now will result
8200 in different behaviour than observed with earlier library versions:
8201 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8202 does not influence s as it used to.
8203
8204 In order to clean up things more thoroughly, inside SSL_SESSION
8205 we don't use CERT any longer, but a new structure SESS_CERT
8206 that holds per-session data (if available); currently, this is
8207 the peer's certificate chain and, for clients, the server's certificate
8208 and temporary key. CERT holds only those values that can have
8209 meaningful defaults in an SSL_CTX.
8210 [Bodo Moeller]
8211
8212 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8213 from the internal representation. Various PKCS#7 fixes: remove some
8214 evil casts and set the enc_dig_alg field properly based on the signing
8215 key type.
8216 [Steve Henson]
8217
8218 *) Allow PKCS#12 password to be set from the command line or the
8219 environment. Let 'ca' get its config file name from the environment
8220 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8221 and 'x509').
8222 [Steve Henson]
8223
8224 *) Allow certificate policies extension to use an IA5STRING for the
8225 organization field. This is contrary to the PKIX definition but
8226 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8227 extension option.
8228 [Steve Henson]
8229
8230 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8231 without disallowing inline assembler and the like for non-pedantic builds.
8232 [Ben Laurie]
8233
8234 *) Support Borland C++ builder.
8235 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8236
8237 *) Support Mingw32.
8238 [Ulf Möller]
8239
8240 *) SHA-1 cleanups and performance enhancements.
8241 [Andy Polyakov <appro@fy.chalmers.se>]
8242
8243 *) Sparc v8plus assembler for the bignum library.
8244 [Andy Polyakov <appro@fy.chalmers.se>]
8245
8246 *) Accept any -xxx and +xxx compiler options in Configure.
8247 [Ulf Möller]
8248
8249 *) Update HPUX configuration.
8250 [Anonymous]
8251
8252 *) Add missing sk_<type>_unshift() function to safestack.h
8253 [Ralf S. Engelschall]
8254
8255 *) New function SSL_CTX_use_certificate_chain_file that sets the
8256 "extra_cert"s in addition to the certificate. (This makes sense
8257 only for "PEM" format files, as chains as a whole are not
8258 DER-encoded.)
8259 [Bodo Moeller]
8260
8261 *) Support verify_depth from the SSL API.
8262 x509_vfy.c had what can be considered an off-by-one-error:
8263 Its depth (which was not part of the external interface)
8264 was actually counting the number of certificates in a chain;
8265 now it really counts the depth.
8266 [Bodo Moeller]
8267
8268 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8269 instead of X509err, which often resulted in confusing error
8270 messages since the error codes are not globally unique
8271 (e.g. an alleged error in ssl3_accept when a certificate
8272 didn't match the private key).
8273
8274 *) New function SSL_CTX_set_session_id_context that allows to set a default
8275 value (so that you don't need SSL_set_session_id_context for each
8276 connection using the SSL_CTX).
8277 [Bodo Moeller]
8278
8279 *) OAEP decoding bug fix.
8280 [Ulf Möller]
8281
8282 *) Support INSTALL_PREFIX for package builders, as proposed by
8283 David Harris.
8284 [Bodo Moeller]
8285
8286 *) New Configure options "threads" and "no-threads". For systems
8287 where the proper compiler options are known (currently Solaris
8288 and Linux), "threads" is the default.
8289 [Bodo Moeller]
8290
8291 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8292 [Bodo Moeller]
8293
8294 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8295 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8296 such as /usr/local/bin.
8297 [Bodo Moeller]
8298
8299 *) "make linux-shared" to build shared libraries.
8300 [Niels Poppe <niels@netbox.org>]
8301
8302 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8303 [Ulf Möller]
8304
8305 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8306 extension adding in x509 utility.
8307 [Steve Henson]
8308
8309 *) Remove NOPROTO sections and error code comments.
8310 [Ulf Möller]
8311
8312 *) Partial rewrite of the DEF file generator to now parse the ANSI
8313 prototypes.
8314 [Steve Henson]
8315
8316 *) New Configure options --prefix=DIR and --openssldir=DIR.
8317 [Ulf Möller]
8318
8319 *) Complete rewrite of the error code script(s). It is all now handled
8320 by one script at the top level which handles error code gathering,
8321 header rewriting and C source file generation. It should be much better
8322 than the old method: it now uses a modified version of Ulf's parser to
8323 read the ANSI prototypes in all header files (thus the old K&R definitions
8324 aren't needed for error creation any more) and do a better job of
8325 translating function codes into names. The old 'ASN1 error code imbedded
8326 in a comment' is no longer necessary and it doesn't use .err files which
8327 have now been deleted. Also the error code call doesn't have to appear all
8328 on one line (which resulted in some large lines...).
8329 [Steve Henson]
8330
8331 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8332 [Bodo Moeller]
8333
8334 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8335 0 (which usually indicates a closed connection), but continue reading.
8336 [Bodo Moeller]
8337
8338 *) Fix some race conditions.
8339 [Bodo Moeller]
8340
8341 *) Add support for CRL distribution points extension. Add Certificate
8342 Policies and CRL distribution points documentation.
8343 [Steve Henson]
8344
8345 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8346 [Ulf Möller]
8347
8348 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8349 8 of keying material. Merlin has also confirmed interop with this fix
8350 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8351 [Merlin Hughes <merlin@baltimore.ie>]
8352
8353 *) Fix lots of warnings.
8354 [Richard Levitte <levitte@stacken.kth.se>]
8355
8356 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8357 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8358 [Richard Levitte <levitte@stacken.kth.se>]
8359
8360 *) Fix problems with sizeof(long) == 8.
8361 [Andy Polyakov <appro@fy.chalmers.se>]
8362
8363 *) Change functions to ANSI C.
8364 [Ulf Möller]
8365
8366 *) Fix typos in error codes.
8367 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8368
8369 *) Remove defunct assembler files from Configure.
8370 [Ulf Möller]
8371
8372 *) SPARC v8 assembler BIGNUM implementation.
8373 [Andy Polyakov <appro@fy.chalmers.se>]
8374
8375 *) Support for Certificate Policies extension: both print and set.
8376 Various additions to support the r2i method this uses.
8377 [Steve Henson]
8378
8379 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8380 return a const string when you are expecting an allocated buffer.
8381 [Ben Laurie]
8382
8383 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8384 types DirectoryString and DisplayText.
8385 [Steve Henson]
8386
8387 *) Add code to allow r2i extensions to access the configuration database,
8388 add an LHASH database driver and add several ctx helper functions.
8389 [Steve Henson]
8390
8391 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8392 fail when they extended the size of a BIGNUM.
8393 [Steve Henson]
8394
8395 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8396 support typesafe stack.
8397 [Steve Henson]
8398
8399 *) Fix typo in SSL_[gs]et_options().
8400 [Nils Frostberg <nils@medcom.se>]
8401
8402 *) Delete various functions and files that belonged to the (now obsolete)
8403 old X509V3 handling code.
8404 [Steve Henson]
8405
8406 *) New Configure option "rsaref".
8407 [Ulf Möller]
8408
8409 *) Don't auto-generate pem.h.
8410 [Bodo Moeller]
8411
8412 *) Introduce type-safe ASN.1 SETs.
8413 [Ben Laurie]
8414
8415 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8416 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8417
8418 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8419 that links with OpenSSL (well at least cause lots of warnings), but fear
8420 not: the conversion is trivial, and it eliminates loads of evil casts. A
8421 few STACKed things have been converted already. Feel free to convert more.
8422 In the fullness of time, I'll do away with the STACK type altogether.
8423 [Ben Laurie]
8424
8425 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8426 specified in <certfile> by updating the entry in the index.txt file.
8427 This way one no longer has to edit the index.txt file manually for
8428 revoking a certificate. The -revoke option does the gory details now.
8429 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
8430
8431 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
8432 `-text' option at all and this way the `-noout -text' combination was
8433 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
8434 [Ralf S. Engelschall]
8435
8436 *) Make sure a corresponding plain text error message exists for the
8437 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
8438 verify callback function determined that a certificate was revoked.
8439 [Ralf S. Engelschall]
8440
8441 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
8442 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
8443 all available cipers including rc5, which was forgotten until now.
8444 In order to let the testing shell script know which algorithms
8445 are available, a new (up to now undocumented) command
8446 "openssl list-cipher-commands" is used.
8447 [Bodo Moeller]
8448
8449 *) Bugfix: s_client occasionally would sleep in select() when
8450 it should have checked SSL_pending() first.
8451 [Bodo Moeller]
8452
8453 *) New functions DSA_do_sign and DSA_do_verify to provide access to
8454 the raw DSA values prior to ASN.1 encoding.
8455 [Ulf Möller]
8456
8457 *) Tweaks to Configure
8458 [Niels Poppe <niels@netbox.org>]
8459
8460 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
8461 yet...
8462 [Steve Henson]
8463
8464 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
8465 [Ulf Möller]
8466
8467 *) New config option to avoid instructions that are illegal on the 80386.
8468 The default code is faster, but requires at least a 486.
8469 [Ulf Möller]
8470
8471 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
8472 SSL2_SERVER_VERSION (not used at all) macros, which are now the
8473 same as SSL2_VERSION anyway.
8474 [Bodo Moeller]
8475
8476 *) New "-showcerts" option for s_client.
8477 [Bodo Moeller]
8478
8479 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
8480 application. Various cleanups and fixes.
8481 [Steve Henson]
8482
8483 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
8484 modify error routines to work internally. Add error codes and PBE init
8485 to library startup routines.
8486 [Steve Henson]
8487
8488 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
8489 packing functions to asn1 and evp. Changed function names and error
8490 codes along the way.
8491 [Steve Henson]
8492
8493 *) PKCS12 integration: and so it begins... First of several patches to
8494 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
8495 objects to objects.h
8496 [Steve Henson]
8497
8498 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
8499 and display support for Thawte strong extranet extension.
8500 [Steve Henson]
8501
8502 *) Add LinuxPPC support.
8503 [Jeff Dubrule <igor@pobox.org>]
8504
8505 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
8506 bn_div_words in alpha.s.
8507 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
8508
8509 *) Make sure the RSA OAEP test is skipped under -DRSAref because
8510 OAEP isn't supported when OpenSSL is built with RSAref.
8511 [Ulf Moeller <ulf@fitug.de>]
8512
8513 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
8514 so they no longer are missing under -DNOPROTO.
8515 [Soren S. Jorvang <soren@t.dk>]
8516
8517
8518 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
8519
8520 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
8521 doesn't work when the session is reused. Coming soon!
8522 [Ben Laurie]
8523
8524 *) Fix a security hole, that allows sessions to be reused in the wrong
8525 context thus bypassing client cert protection! All software that uses
8526 client certs and session caches in multiple contexts NEEDS PATCHING to
8527 allow session reuse! A fuller solution is in the works.
8528 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
8529
8530 *) Some more source tree cleanups (removed obsolete files
8531 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
8532 permission on "config" script to be executable) and a fix for the INSTALL
8533 document.
8534 [Ulf Moeller <ulf@fitug.de>]
8535
8536 *) Remove some legacy and erroneous uses of malloc, free instead of
8537 Malloc, Free.
8538 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
8539
8540 *) Make rsa_oaep_test return non-zero on error.
8541 [Ulf Moeller <ulf@fitug.de>]
8542
8543 *) Add support for native Solaris shared libraries. Configure
8544 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
8545 if someone would make that last step automatic.
8546 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
8547
8548 *) ctx_size was not built with the right compiler during "make links". Fixed.
8549 [Ben Laurie]
8550
8551 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
8552 except NULL ciphers". This means the default cipher list will no longer
8553 enable NULL ciphers. They need to be specifically enabled e.g. with
8554 the string "DEFAULT:eNULL".
8555 [Steve Henson]
8556
8557 *) Fix to RSA private encryption routines: if p < q then it would
8558 occasionally produce an invalid result. This will only happen with
8559 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
8560 [Steve Henson]
8561
8562 *) Be less restrictive and allow also `perl util/perlpath.pl
8563 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
8564 because this way one can also use an interpreter named `perl5' (which is
8565 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
8566 installed as `perl').
8567 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8568
8569 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
8570 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8571
8572 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
8573 advapi32.lib to Win32 build and change the pem test comparision
8574 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
8575 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
8576 and crypto/des/ede_cbcm_enc.c.
8577 [Steve Henson]
8578
8579 *) DES quad checksum was broken on big-endian architectures. Fixed.
8580 [Ben Laurie]
8581
8582 *) Comment out two functions in bio.h that aren't implemented. Fix up the
8583 Win32 test batch file so it (might) work again. The Win32 test batch file
8584 is horrible: I feel ill....
8585 [Steve Henson]
8586
8587 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
8588 in e_os.h. Audit of header files to check ANSI and non ANSI
8589 sections: 10 functions were absent from non ANSI section and not exported
8590 from Windows DLLs. Fixed up libeay.num for new functions.
8591 [Steve Henson]
8592
8593 *) Make `openssl version' output lines consistent.
8594 [Ralf S. Engelschall]
8595
8596 *) Fix Win32 symbol export lists for BIO functions: Added
8597 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
8598 to ms/libeay{16,32}.def.
8599 [Ralf S. Engelschall]
8600
8601 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
8602 fine under Unix and passes some trivial tests I've now added. But the
8603 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
8604 added to make sure no one expects that this stuff really works in the
8605 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
8606 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
8607 openssl_bio.xs.
8608 [Ralf S. Engelschall]
8609
8610 *) Fix the generation of two part addresses in perl.
8611 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
8612
8613 *) Add config entry for Linux on MIPS.
8614 [John Tobey <jtobey@channel1.com>]
8615
8616 *) Make links whenever Configure is run, unless we are on Windoze.
8617 [Ben Laurie]
8618
8619 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
8620 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
8621 in CRLs.
8622 [Steve Henson]
8623
8624 *) Add a useful kludge to allow package maintainers to specify compiler and
8625 other platforms details on the command line without having to patch the
8626 Configure script everytime: One now can use ``perl Configure
8627 <id>:<details>'', i.e. platform ids are allowed to have details appended
8628 to them (seperated by colons). This is treated as there would be a static
8629 pre-configured entry in Configure's %table under key <id> with value
8630 <details> and ``perl Configure <id>'' is called. So, when you want to
8631 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
8632 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
8633 now, which overrides the FreeBSD-elf entry on-the-fly.
8634 [Ralf S. Engelschall]
8635
8636 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
8637 [Ben Laurie]
8638
8639 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
8640 on the `perl Configure ...' command line. This way one can compile
8641 OpenSSL libraries with Position Independent Code (PIC) which is needed
8642 for linking it into DSOs.
8643 [Ralf S. Engelschall]
8644
8645 *) Remarkably, export ciphers were totally broken and no-one had noticed!
8646 Fixed.
8647 [Ben Laurie]
8648
8649 *) Cleaned up the LICENSE document: The official contact for any license
8650 questions now is the OpenSSL core team under openssl-core@openssl.org.
8651 And add a paragraph about the dual-license situation to make sure people
8652 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
8653 to the OpenSSL toolkit.
8654 [Ralf S. Engelschall]
8655
8656 *) General source tree makefile cleanups: Made `making xxx in yyy...'
8657 display consistent in the source tree and replaced `/bin/rm' by `rm'.
8658 Additonally cleaned up the `make links' target: Remove unnecessary
8659 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
8660 to speed processing and no longer clutter the display with confusing
8661 stuff. Instead only the actually done links are displayed.
8662 [Ralf S. Engelschall]
8663
8664 *) Permit null encryption ciphersuites, used for authentication only. It used
8665 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
8666 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
8667 encryption.
8668 [Ben Laurie]
8669
8670 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
8671 signed attributes when verifying signatures (this would break them),
8672 the detached data encoding was wrong and public keys obtained using
8673 X509_get_pubkey() weren't freed.
8674 [Steve Henson]
8675
8676 *) Add text documentation for the BUFFER functions. Also added a work around
8677 to a Win95 console bug. This was triggered by the password read stuff: the
8678 last character typed gets carried over to the next fread(). If you were
8679 generating a new cert request using 'req' for example then the last
8680 character of the passphrase would be CR which would then enter the first
8681 field as blank.
8682 [Steve Henson]
8683
8684 *) Added the new `Includes OpenSSL Cryptography Software' button as
8685 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
8686 button and can be used by applications based on OpenSSL to show the
8687 relationship to the OpenSSL project.
8688 [Ralf S. Engelschall]
8689
8690 *) Remove confusing variables in function signatures in files
8691 ssl/ssl_lib.c and ssl/ssl.h.
8692 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8693
8694 *) Don't install bss_file.c under PREFIX/include/
8695 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8696
8697 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
8698 functions that return function pointers and has support for NT specific
8699 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
8700 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
8701 unsigned to signed types: this was killing the Win32 compile.
8702 [Steve Henson]
8703
8704 *) Add new certificate file to stack functions,
8705 SSL_add_dir_cert_subjects_to_stack() and
8706 SSL_add_file_cert_subjects_to_stack(). These largely supplant
8707 SSL_load_client_CA_file(), and can be used to add multiple certs easily
8708 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
8709 This means that Apache-SSL and similar packages don't have to mess around
8710 to add as many CAs as they want to the preferred list.
8711 [Ben Laurie]
8712
8713 *) Experiment with doxygen documentation. Currently only partially applied to
8714 ssl/ssl_lib.c.
8715 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
8716 openssl.doxy as the configuration file.
8717 [Ben Laurie]
8718
8719 *) Get rid of remaining C++-style comments which strict C compilers hate.
8720 [Ralf S. Engelschall, pointed out by Carlos Amengual]
8721
8722 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
8723 compiled in by default: it has problems with large keys.
8724 [Steve Henson]
8725
8726 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
8727 DH private keys and/or callback functions which directly correspond to
8728 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
8729 is needed for applications which have to configure certificates on a
8730 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
8731 (e.g. s_server).
8732 For the RSA certificate situation is makes no difference, but
8733 for the DSA certificate situation this fixes the "no shared cipher"
8734 problem where the OpenSSL cipher selection procedure failed because the
8735 temporary keys were not overtaken from the context and the API provided
8736 no way to reconfigure them.
8737 The new functions now let applications reconfigure the stuff and they
8738 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
8739 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
8740 non-public-API function ssl_cert_instantiate() is used as a helper
8741 function and also to reduce code redundancy inside ssl_rsa.c.
8742 [Ralf S. Engelschall]
8743
8744 *) Move s_server -dcert and -dkey options out of the undocumented feature
8745 area because they are useful for the DSA situation and should be
8746 recognized by the users.
8747 [Ralf S. Engelschall]
8748
8749 *) Fix the cipher decision scheme for export ciphers: the export bits are
8750 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
8751 SSL_EXP_MASK. So, the original variable has to be used instead of the
8752 already masked variable.
8753 [Richard Levitte <levitte@stacken.kth.se>]
8754
8755 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
8756 [Richard Levitte <levitte@stacken.kth.se>]
8757
8758 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
8759 from `int' to `unsigned int' because it's a length and initialized by
8760 EVP_DigestFinal() which expects an `unsigned int *'.
8761 [Richard Levitte <levitte@stacken.kth.se>]
8762
8763 *) Don't hard-code path to Perl interpreter on shebang line of Configure
8764 script. Instead use the usual Shell->Perl transition trick.
8765 [Ralf S. Engelschall]
8766
8767 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
8768 (in addition to RSA certificates) to match the behaviour of `openssl dsa
8769 -noout -modulus' as it's already the case for `openssl rsa -noout
8770 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
8771 currently the public key is printed (a decision which was already done by
8772 `openssl dsa -modulus' in the past) which serves a similar purpose.
8773 Additionally the NO_RSA no longer completely removes the whole -modulus
8774 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
8775 now, too.
8776 [Ralf S. Engelschall]
8777
8778 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
8779 BIO. See the source (crypto/evp/bio_ok.c) for more info.
8780 [Arne Ansper <arne@ats.cyber.ee>]
8781
8782 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
8783 to be added. Now both 'req' and 'ca' can use new objects defined in the
8784 config file.
8785 [Steve Henson]
8786
8787 *) Add cool BIO that does syslog (or event log on NT).
8788 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
8789
8790 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
8791 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
8792 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
8793 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
8794 [Ben Laurie]
8795
8796 *) Add preliminary config info for new extension code.
8797 [Steve Henson]
8798
8799 *) Make RSA_NO_PADDING really use no padding.
8800 [Ulf Moeller <ulf@fitug.de>]
8801
8802 *) Generate errors when private/public key check is done.
8803 [Ben Laurie]
8804
8805 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
8806 for some CRL extensions and new objects added.
8807 [Steve Henson]
8808
8809 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
8810 key usage extension and fuller support for authority key id.
8811 [Steve Henson]
8812
8813 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
8814 padding method for RSA, which is recommended for new applications in PKCS
8815 #1 v2.0 (RFC 2437, October 1998).
8816 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
8817 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
8818 against Bleichbacher's attack on RSA.
8819 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
8820 Ben Laurie]
8821
8822 *) Updates to the new SSL compression code
8823 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8824
8825 *) Fix so that the version number in the master secret, when passed
8826 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
8827 (because the server will not accept higher), that the version number
8828 is 0x03,0x01, not 0x03,0x00
8829 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8830
8831 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
8832 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
8833 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
8834 [Steve Henson]
8835
8836 *) Support for RAW extensions where an arbitrary extension can be
8837 created by including its DER encoding. See apps/openssl.cnf for
8838 an example.
8839 [Steve Henson]
8840
8841 *) Make sure latest Perl versions don't interpret some generated C array
8842 code as Perl array code in the crypto/err/err_genc.pl script.
8843 [Lars Weber <3weber@informatik.uni-hamburg.de>]
8844
8845 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
8846 not many people have the assembler. Various Win32 compilation fixes and
8847 update to the INSTALL.W32 file with (hopefully) more accurate Win32
8848 build instructions.
8849 [Steve Henson]
8850
8851 *) Modify configure script 'Configure' to automatically create crypto/date.h
8852 file under Win32 and also build pem.h from pem.org. New script
8853 util/mkfiles.pl to create the MINFO file on environments that can't do a
8854 'make files': perl util/mkfiles.pl >MINFO should work.
8855 [Steve Henson]
8856
8857 *) Major rework of DES function declarations, in the pursuit of correctness
8858 and purity. As a result, many evil casts evaporated, and some weirdness,
8859 too. You may find this causes warnings in your code. Zapping your evil
8860 casts will probably fix them. Mostly.
8861 [Ben Laurie]
8862
8863 *) Fix for a typo in asn1.h. Bug fix to object creation script
8864 obj_dat.pl. It considered a zero in an object definition to mean
8865 "end of object": none of the objects in objects.h have any zeros
8866 so it wasn't spotted.
8867 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
8868
8869 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
8870 Masking (CBCM). In the absence of test vectors, the best I have been able
8871 to do is check that the decrypt undoes the encrypt, so far. Send me test
8872 vectors if you have them.
8873 [Ben Laurie]
8874
8875 *) Correct calculation of key length for export ciphers (too much space was
8876 allocated for null ciphers). This has not been tested!
8877 [Ben Laurie]
8878
8879 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
8880 message is now correct (it understands "crypto" and "ssl" on its
8881 command line). There is also now an "update" option. This will update
8882 the util/ssleay.num and util/libeay.num files with any new functions.
8883 If you do a:
8884 perl util/mkdef.pl crypto ssl update
8885 it will update them.
8886 [Steve Henson]
8887
8888 *) Overhauled the Perl interface (perl/*):
8889 - ported BN stuff to OpenSSL's different BN library
8890 - made the perl/ source tree CVS-aware
8891 - renamed the package from SSLeay to OpenSSL (the files still contain
8892 their history because I've copied them in the repository)
8893 - removed obsolete files (the test scripts will be replaced
8894 by better Test::Harness variants in the future)
8895 [Ralf S. Engelschall]
8896
8897 *) First cut for a very conservative source tree cleanup:
8898 1. merge various obsolete readme texts into doc/ssleay.txt
8899 where we collect the old documents and readme texts.
8900 2. remove the first part of files where I'm already sure that we no
8901 longer need them because of three reasons: either they are just temporary
8902 files which were left by Eric or they are preserved original files where
8903 I've verified that the diff is also available in the CVS via "cvs diff
8904 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8905 the crypto/md/ stuff).
8906 [Ralf S. Engelschall]
8907
8908 *) More extension code. Incomplete support for subject and issuer alt
8909 name, issuer and authority key id. Change the i2v function parameters
8910 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8911 what that's for :-) Fix to ASN1 macro which messed up
8912 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8913 [Steve Henson]
8914
8915 *) Preliminary support for ENUMERATED type. This is largely copied from the
8916 INTEGER code.
8917 [Steve Henson]
8918
8919 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8920 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8921
8922 *) Make sure `make rehash' target really finds the `openssl' program.
8923 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8924
8925 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8926 like to hear about it if this slows down other processors.
8927 [Ben Laurie]
8928
8929 *) Add CygWin32 platform information to Configure script.
8930 [Alan Batie <batie@aahz.jf.intel.com>]
8931
8932 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8933 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8934
8935 *) New program nseq to manipulate netscape certificate sequences
8936 [Steve Henson]
8937
8938 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8939 few typos.
8940 [Steve Henson]
8941
8942 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8943 but the BN code had some problems that would cause failures when
8944 doing certificate verification and some other functions.
8945 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8946
8947 *) Add ASN1 and PEM code to support netscape certificate sequences.
8948 [Steve Henson]
8949
8950 *) Add ASN1 and PEM code to support netscape certificate sequences.
8951 [Steve Henson]
8952
8953 *) Add several PKIX and private extended key usage OIDs.
8954 [Steve Henson]
8955
8956 *) Modify the 'ca' program to handle the new extension code. Modify
8957 openssl.cnf for new extension format, add comments.
8958 [Steve Henson]
8959
8960 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8961 and add a sample to openssl.cnf so req -x509 now adds appropriate
8962 CA extensions.
8963 [Steve Henson]
8964
8965 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8966 error code, add initial support to X509_print() and x509 application.
8967 [Steve Henson]
8968
8969 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8970 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8971 stuff is currently isolated and isn't even compiled yet.
8972 [Steve Henson]
8973
8974 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8975 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8976 Removed the versions check from X509 routines when loading extensions:
8977 this allows certain broken certificates that don't set the version
8978 properly to be processed.
8979 [Steve Henson]
8980
8981 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8982 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8983 can still be regenerated with "make depend".
8984 [Ben Laurie]
8985
8986 *) Spelling mistake in C version of CAST-128.
8987 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8988
8989 *) Changes to the error generation code. The perl script err-code.pl
8990 now reads in the old error codes and retains the old numbers, only
8991 adding new ones if necessary. It also only changes the .err files if new
8992 codes are added. The makefiles have been modified to only insert errors
8993 when needed (to avoid needlessly modifying header files). This is done
8994 by only inserting errors if the .err file is newer than the auto generated
8995 C file. To rebuild all the error codes from scratch (the old behaviour)
8996 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8997 or delete all the .err files.
8998 [Steve Henson]
8999
9000 *) CAST-128 was incorrectly implemented for short keys. The C version has
9001 been fixed, but is untested. The assembler versions are also fixed, but
9002 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9003 to regenerate it if needed.
9004 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9005 Hagino <itojun@kame.net>]
9006
9007 *) File was opened incorrectly in randfile.c.
9008 [Ulf Möller <ulf@fitug.de>]
9009
9010 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9011 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9012 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9013 al: it's just almost always a UTCTime. Note this patch adds new error
9014 codes so do a "make errors" if there are problems.
9015 [Steve Henson]
9016
9017 *) Correct Linux 1 recognition in config.
9018 [Ulf Möller <ulf@fitug.de>]
9019
9020 *) Remove pointless MD5 hash when using DSA keys in ca.
9021 [Anonymous <nobody@replay.com>]
9022
9023 *) Generate an error if given an empty string as a cert directory. Also
9024 generate an error if handed NULL (previously returned 0 to indicate an
9025 error, but didn't set one).
9026 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9027
9028 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9029 [Ben Laurie]
9030
9031 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9032 parameters. This was causing a warning which killed off the Win32 compile.
9033 [Steve Henson]
9034
9035 *) Remove C++ style comments from crypto/bn/bn_local.h.
9036 [Neil Costigan <neil.costigan@celocom.com>]
9037
9038 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9039 based on a text string, looking up short and long names and finally
9040 "dot" format. The "dot" format stuff didn't work. Added new function
9041 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9042 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9043 OID is not part of the table.
9044 [Steve Henson]
9045
9046 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9047 X509_LOOKUP_by_alias().
9048 [Ben Laurie]
9049
9050 *) Sort openssl functions by name.
9051 [Ben Laurie]
9052
9053 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9054 encryption from sample DSA keys (in case anyone is interested the password
9055 was "1234").
9056 [Steve Henson]
9057
9058 *) Make _all_ *_free functions accept a NULL pointer.
9059 [Frans Heymans <fheymans@isaserver.be>]
9060
9061 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9062 NULL pointers.
9063 [Anonymous <nobody@replay.com>]
9064
9065 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9066 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9067
9068 *) Don't blow it for numeric -newkey arguments to apps/req.
9069 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9070
9071 *) Temp key "for export" tests were wrong in s3_srvr.c.
9072 [Anonymous <nobody@replay.com>]
9073
9074 *) Add prototype for temp key callback functions
9075 SSL_CTX_set_tmp_{rsa,dh}_callback().
9076 [Ben Laurie]
9077
9078 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9079 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9080 [Steve Henson]
9081
9082 *) X509_name_add_entry() freed the wrong thing after an error.
9083 [Arne Ansper <arne@ats.cyber.ee>]
9084
9085 *) rsa_eay.c would attempt to free a NULL context.
9086 [Arne Ansper <arne@ats.cyber.ee>]
9087
9088 *) BIO_s_socket() had a broken should_retry() on Windoze.
9089 [Arne Ansper <arne@ats.cyber.ee>]
9090
9091 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9092 [Arne Ansper <arne@ats.cyber.ee>]
9093
9094 *) Make sure the already existing X509_STORE->depth variable is initialized
9095 in X509_STORE_new(), but document the fact that this variable is still
9096 unused in the certificate verification process.
9097 [Ralf S. Engelschall]
9098
9099 *) Fix the various library and apps files to free up pkeys obtained from
9100 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9101 [Steve Henson]
9102
9103 *) Fix reference counting in X509_PUBKEY_get(). This makes
9104 demos/maurice/example2.c work, amongst others, probably.
9105 [Steve Henson and Ben Laurie]
9106
9107 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9108 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9109 are no longer created. This way we have a single and consistent command
9110 line interface `openssl <command>', similar to `cvs <command>'.
9111 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9112
9113 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9114 BIT STRING wrapper always have zero unused bits.
9115 [Steve Henson]
9116
9117 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9118 [Steve Henson]
9119
9120 *) Make the top-level INSTALL documentation easier to understand.
9121 [Paul Sutton]
9122
9123 *) Makefiles updated to exit if an error occurs in a sub-directory
9124 make (including if user presses ^C) [Paul Sutton]
9125
9126 *) Make Montgomery context stuff explicit in RSA data structure.
9127 [Ben Laurie]
9128
9129 *) Fix build order of pem and err to allow for generated pem.h.
9130 [Ben Laurie]
9131
9132 *) Fix renumbering bug in X509_NAME_delete_entry().
9133 [Ben Laurie]
9134
9135 *) Enhanced the err-ins.pl script so it makes the error library number
9136 global and can add a library name. This is needed for external ASN1 and
9137 other error libraries.
9138 [Steve Henson]
9139
9140 *) Fixed sk_insert which never worked properly.
9141 [Steve Henson]
9142
9143 *) Fix ASN1 macros so they can handle indefinite length construted
9144 EXPLICIT tags. Some non standard certificates use these: they can now
9145 be read in.
9146 [Steve Henson]
9147
9148 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9149 into a single doc/ssleay.txt bundle. This way the information is still
9150 preserved but no longer messes up this directory. Now it's new room for
9151 the new set of documenation files.
9152 [Ralf S. Engelschall]
9153
9154 *) SETs were incorrectly DER encoded. This was a major pain, because they
9155 shared code with SEQUENCEs, which aren't coded the same. This means that
9156 almost everything to do with SETs or SEQUENCEs has either changed name or
9157 number of arguments.
9158 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9159
9160 *) Fix test data to work with the above.
9161 [Ben Laurie]
9162
9163 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9164 was already fixed by Eric for 0.9.1 it seems.
9165 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9166
9167 *) Autodetect FreeBSD3.
9168 [Ben Laurie]
9169
9170 *) Fix various bugs in Configure. This affects the following platforms:
9171 nextstep
9172 ncr-scde
9173 unixware-2.0
9174 unixware-2.0-pentium
9175 sco5-cc.
9176 [Ben Laurie]
9177
9178 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9179 before they are needed.
9180 [Ben Laurie]
9181
9182 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9183 [Ben Laurie]
9184
9185
9186 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9187
9188 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9189 changed SSLeay to OpenSSL in version strings.
9190 [Ralf S. Engelschall]
9191
9192 *) Some fixups to the top-level documents.
9193 [Paul Sutton]
9194
9195 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9196 because the symlink to include/ was missing.
9197 [Ralf S. Engelschall]
9198
9199 *) Incorporated the popular no-RSA/DSA-only patches
9200 which allow to compile a RSA-free SSLeay.
9201 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9202
9203 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9204 when "ssleay" is still not found.
9205 [Ralf S. Engelschall]
9206
9207 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9208 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9209
9210 *) Updated the README file.
9211 [Ralf S. Engelschall]
9212
9213 *) Added various .cvsignore files in the CVS repository subdirs
9214 to make a "cvs update" really silent.
9215 [Ralf S. Engelschall]
9216
9217 *) Recompiled the error-definition header files and added
9218 missing symbols to the Win32 linker tables.
9219 [Ralf S. Engelschall]
9220
9221 *) Cleaned up the top-level documents;
9222 o new files: CHANGES and LICENSE
9223 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9224 o merged COPYRIGHT into LICENSE
9225 o removed obsolete TODO file
9226 o renamed MICROSOFT to INSTALL.W32
9227 [Ralf S. Engelschall]
9228
9229 *) Removed dummy files from the 0.9.1b source tree:
9230 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9231 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9232 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9233 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9234 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9235 [Ralf S. Engelschall]
9236
9237 *) Added various platform portability fixes.
9238 [Mark J. Cox]
9239
9240 *) The Genesis of the OpenSSL rpject:
9241 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9242 Young and Tim J. Hudson created while they were working for C2Net until
9243 summer 1998.
9244 [The OpenSSL Project]
9245
9246
9247 Changes between 0.9.0b and 0.9.1b [not released]
9248
9249 *) Updated a few CA certificates under certs/
9250 [Eric A. Young]
9251
9252 *) Changed some BIGNUM api stuff.
9253 [Eric A. Young]
9254
9255 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9256 DGUX x86, Linux Alpha, etc.
9257 [Eric A. Young]
9258
9259 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9260 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9261 available).
9262 [Eric A. Young]
9263
9264 *) Add -strparse option to asn1pars program which parses nested
9265 binary structures
9266 [Dr Stephen Henson <shenson@bigfoot.com>]
9267
9268 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9269 [Eric A. Young]
9270
9271 *) DSA fix for "ca" program.
9272 [Eric A. Young]
9273
9274 *) Added "-genkey" option to "dsaparam" program.
9275 [Eric A. Young]
9276
9277 *) Added RIPE MD160 (rmd160) message digest.
9278 [Eric A. Young]
9279
9280 *) Added -a (all) option to "ssleay version" command.
9281 [Eric A. Young]
9282
9283 *) Added PLATFORM define which is the id given to Configure.
9284 [Eric A. Young]
9285
9286 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9287 [Eric A. Young]
9288
9289 *) Extended the ASN.1 parser routines.
9290 [Eric A. Young]
9291
9292 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9293 [Eric A. Young]
9294
9295 *) Added a BN_CTX to the BN library.
9296 [Eric A. Young]
9297
9298 *) Fixed the weak key values in DES library
9299 [Eric A. Young]
9300
9301 *) Changed API in EVP library for cipher aliases.
9302 [Eric A. Young]
9303
9304 *) Added support for RC2/64bit cipher.
9305 [Eric A. Young]
9306
9307 *) Converted the lhash library to the crypto/mem.c functions.
9308 [Eric A. Young]
9309
9310 *) Added more recognized ASN.1 object ids.
9311 [Eric A. Young]
9312
9313 *) Added more RSA padding checks for SSL/TLS.
9314 [Eric A. Young]
9315
9316 *) Added BIO proxy/filter functionality.
9317 [Eric A. Young]
9318
9319 *) Added extra_certs to SSL_CTX which can be used
9320 send extra CA certificates to the client in the CA cert chain sending
9321 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9322 [Eric A. Young]
9323
9324 *) Now Fortezza is denied in the authentication phase because
9325 this is key exchange mechanism is not supported by SSLeay at all.
9326 [Eric A. Young]
9327
9328 *) Additional PKCS1 checks.
9329 [Eric A. Young]
9330
9331 *) Support the string "TLSv1" for all TLS v1 ciphers.
9332 [Eric A. Young]
9333
9334 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9335 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9336 [Eric A. Young]
9337
9338 *) Fixed a few memory leaks.
9339 [Eric A. Young]
9340
9341 *) Fixed various code and comment typos.
9342 [Eric A. Young]
9343
9344 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9345 bytes sent in the client random.
9346 [Edward Bishop <ebishop@spyglass.com>]
9347