]> git.ipfire.org Git - thirdparty/openssl.git/blame - CHANGES.md
Add deprecation macro for 3.1 and deprecate OPENSSL_LH_stats
[thirdparty/openssl.git] / CHANGES.md
CommitLineData
5f8e6c50
DMSP
1OpenSSL CHANGES
2===============
3
4This is a high-level summary of the most important changes.
5For a full list of changes, see the [git commit log][log] and
6pick the appropriate release branch.
7
8 [log]: https://github.com/openssl/openssl/commits/
9
44652c16
DMSP
10OpenSSL Releases
11----------------
12
27272657 13 - [OpenSSL 3.1](#openssl-31)
44652c16
DMSP
14 - [OpenSSL 3.0](#openssl-30)
15 - [OpenSSL 1.1.1](#openssl-111)
16 - [OpenSSL 1.1.0](#openssl-110)
17 - [OpenSSL 1.0.2](#openssl-102)
18 - [OpenSSL 1.0.1](#openssl-101)
19 - [OpenSSL 1.0.0](#openssl-100)
20 - [OpenSSL 0.9.x](#openssl-09x)
21
27272657
RL
22OpenSSL 3.1
23-----------
24
25### Changes between 3.0 and 3.1 [xx XXX xxxx]
26
a425c0fe
KK
27 * Add more SRTP protection profiles from RFC8723 and RFC8269.
28
29 *Kijin Kim*
30
7c78932b
DU
31 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
32
33 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
34
a3e53d56
TS
35 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
36 supported and enabled.
37
38 *Todd Short*
39
b139a956
NT
40 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
41 to the list of ciphersuites providing Perfect Forward Secrecy as
42 required by SECLEVEL >= 3.
43
44 *Dmitry Belyavskiy, Nicola Tuveri*
45
13a53fbf
PL
46 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting. The
47 SSL_CTRL_GET_IANA_GROUPS control code, exposed as the SSL_get0_iana_groups()
48 function-like macro, retrieves the list of supported groups sent by the peer,
49 and the function SSL_client_hello_get_extension_order() populates a caller-supplied
50 array with the list of extension types present in the ClientHello, in order of
51 appearance.
52
53 *Phus Lu*
54
1d28ada1
DS
55 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
56 to make it possible to use empty passphrase strings.
59ccb72c
DS
57
58 *Darshan Sen*
59
e8b597f3
OT
60 * RNDR and RNDRRS support in provider functions to provide
61 random number generation for Arm CPUs (aarch64).
62
63 *Orr Toledano*
64
af5e63e1
FG
65 * s_client and s_server apps now explicitly say when the TLS version
66 does not include the renegotiation mechanism. This avoids confusion
67 between that scenario versus when the TLS version includes secure
68 renegotiation but the peer lacks support for it.
69
70 *Felipe Gasper*
71
63b996e7
AM
72 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
73
74 *Tomasz Kantecki, Andrey Matyukov*
75
a4c4090c
MC
76 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
77 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
78 of 160 bits and above and less than 224 bits were previously accepted by
79 default but are now no longer allowed. By default TLS compression was
80 already disabled in previous OpenSSL versions. At security level 2 it cannot
81 be enabled.
82
83 *Matt Caswell*
84
d1b26ddb
EL
85 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
86 IANA standard names.
87
88 *Erik Lax*
89
c8ffd220
P
90 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
91 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
92 will need to load the legacy crypto provider.
93
94 *Paul Dale*
95
63d0f4d2
P
96 * The various OBJ_* functions have been made thread safe.
97
98 *Paul Dale*
99
e0710222
P
100 * CCM8 cipher suites in TLS have been downgraded to security level zero
101 because they use a short authentication tag which lowers their strength.
1a473d1c
P
102
103 *Paul Dale*
104
537976de
DB
105 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
106 by default.
107
108 *Dmitry Belyavskiy*
27272657 109
f87b4c4e
AM
110 * Parallel dual-prime 1536/2048-bit modular exponentiation for
111 AVX512_IFMA capable processors.
112
113 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
114
5317b6ee
HL
115 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
116 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
117 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
118 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
119 `OPENSSL_NO_DEPRECATED_3_1`.
120
121 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
122 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
123 definitions for these functions regardless of whether
124 `OPENSSL_NO_DEPRECATED_3_1` is defined.
125
126 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
127 functions regardless of whether they are using them. It is recommended that
128 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
129
130 *Hugo Landau*
131
c868d1f9
TM
132OpenSSL 3.0
133-----------
134
135For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
136listed here are only a brief description.
137The migration guide contains more detailed information related to new features,
138breaking changes, and mappings for the large list of deprecated functions.
139
140[Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
141
de85a9de
MC
142### Changes between 3.0.3 and 3.0.4 [21 June 2022]
143
144 * In addition to the c_rehash shell command injection identified in
145 CVE-2022-1292, further bugs where the c_rehash script does not
146 properly sanitise shell metacharacters to prevent command injection have been
147 fixed.
148
149 When the CVE-2022-1292 was fixed it was not discovered that there
150 are other places in the script where the file names of certificates
151 being hashed were possibly passed to a command executed through the shell.
152
153 This script is distributed by some operating systems in a manner where
154 it is automatically executed. On such operating systems, an attacker
155 could execute arbitrary commands with the privileges of the script.
156
157 Use of the c_rehash script is considered obsolete and should be replaced
158 by the OpenSSL rehash command line tool.
159 (CVE-2022-2068)
160
161 *Daniel Fiala, Tomáš Mráz*
8a66b2f9
P
162
163 * Case insensitive string comparison no longer uses locales. It has instead
164 been directly implemented.
165
166 *Paul Dale*
167
de85a9de 168### Changes between 3.0.2 and 3.0.3 [3 May 2022]
77d7b6ee 169
8b97bfcc
DB
170 * Case insensitive string comparison is reimplemented via new locale-agnostic
171 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
172 comparison. The previous implementation had problems when the Turkish locale
173 was used.
174
175 *Dmitry Belyavskiy*
176
73e044bd
MC
177 * Fixed a bug in the c_rehash script which was not properly sanitising shell
178 metacharacters to prevent command injection. This script is distributed by
179 some operating systems in a manner where it is automatically executed. On
180 such operating systems, an attacker could execute arbitrary commands with the
181 privileges of the script.
182
183 Use of the c_rehash script is considered obsolete and should be replaced
184 by the OpenSSL rehash command line tool.
185 (CVE-2022-1292)
186
187 *Tomáš Mráz*
188
189 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
190 certificate on an OCSP response. The bug caused the function in the case
191 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
192 response (meaning a successful verification) even in the case where the
193 response signing certificate fails to verify.
194
195 It is anticipated that most users of `OCSP_basic_verify` will not use the
196 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
197 a negative value (indicating a fatal error) in the case of a certificate
198 verification failure. The normal expected return value in this case would be
199 0.
200
201 This issue also impacts the command line OpenSSL "ocsp" application. When
202 verifying an ocsp response with the "-no_cert_checks" option the command line
203 application will report that the verification is successful even though it
204 has in fact failed. In this case the incorrect successful response will also
205 be accompanied by error messages showing the failure and contradicting the
206 apparently successful result.
207 ([CVE-2022-1343])
208
209 *Matt Caswell*
210
211 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
212 AAD data as the MAC key. This made the MAC key trivially predictable.
213
214 An attacker could exploit this issue by performing a man-in-the-middle attack
215 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
216 that the modified data would still pass the MAC integrity check.
217
218 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
219 endpoint will always be rejected by the recipient and the connection will
220 fail at that point. Many application protocols require data to be sent from
221 the client to the server first. Therefore, in such a case, only an OpenSSL
222 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
223
224 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
225 sent in both directions. In this case both clients and servers could be
226 affected, regardless of the application protocol.
227
228 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
229 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
230 the handshake when using this ciphersuite.
231
232 The confidentiality of data is not impacted by this issue, i.e. an attacker
233 cannot decrypt data that has been encrypted using this ciphersuite - they can
234 only modify it.
235
236 In order for this attack to work both endpoints must legitimately negotiate
237 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
238 OpenSSL 3.0, and is not available within the default provider or the default
239 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
240 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
241 following must have occurred:
242
243 1) OpenSSL must have been compiled with the (non-default) compile time option
244 enable-weak-ssl-ciphers
245
246 2) OpenSSL must have had the legacy provider explicitly loaded (either
247 through application code or via configuration)
248
249 3) The ciphersuite must have been explicitly added to the ciphersuite list
250
251 4) The libssl security level must have been set to 0 (default is 1)
252
253 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
254
255 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
256 others that both endpoints have in common
257 (CVE-2022-1434)
258
cac25075 259 *Matt Caswell*
73e044bd
MC
260
261 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
262 occuppied by the removed hash table entries.
263
264 This function is used when decoding certificates or keys. If a long lived
265 process periodically decodes certificates or keys its memory usage will
266 expand without bounds and the process might be terminated by the operating
267 system causing a denial of service. Also traversing the empty hash table
268 entries will take increasingly more time.
269
270 Typically such long lived processes might be TLS clients or TLS servers
271 configured to accept client certificate authentication.
272 (CVE-2022-1473)
273
cac25075 274 *Hugo Landau, Aliaksei Levin*
73e044bd 275
77d7b6ee
HL
276 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
277 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
278 statistics are no longer supported. For compatibility, these statistics are
279 still listed in the output but are now always reported as zero.
280
281 *Hugo Landau*
282
de85a9de 283### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
a40398a1
MC
284
285 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
286 for non-prime moduli.
287
288 Internally this function is used when parsing certificates that contain
289 elliptic curve public keys in compressed form or explicit elliptic curve
290 parameters with a base point encoded in compressed form.
291
292 It is possible to trigger the infinite loop by crafting a certificate that
293 has invalid explicit curve parameters.
294
295 Since certificate parsing happens prior to verification of the certificate
296 signature, any process that parses an externally supplied certificate may thus
297 be subject to a denial of service attack. The infinite loop can also be
298 reached when parsing crafted private keys as they can contain explicit
299 elliptic curve parameters.
300
301 Thus vulnerable situations include:
302
303 - TLS clients consuming server certificates
304 - TLS servers consuming client certificates
305 - Hosting providers taking certificates or private keys from customers
306 - Certificate authorities parsing certification requests from subscribers
307 - Anything else which parses ASN.1 elliptic curve parameters
308
309 Also any other applications that use the BN_mod_sqrt() where the attacker
310 can control the parameter values are vulnerable to this DoS issue.
311 ([CVE-2022-0778])
312
313 *Tomáš Mráz*
314
315 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
316 to the list of ciphersuites providing Perfect Forward Secrecy as
317 required by SECLEVEL >= 3.
318
319 *Dmitry Belyavskiy, Nicola Tuveri*
e180bf64
P
320
321 * Made the AES constant time code for no-asm configurations
322 optional due to the resulting 95% performance degradation.
323 The AES constant time code can be enabled, for no assembly
324 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
325
326 *Paul Dale*
327
a40398a1
MC
328 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
329 passphrase strings.
330
331 *Darshan Sen*
332
dfb39f73
TM
333 * The negative return value handling of the certificate verification callback
334 was reverted. The replacement is to set the verification retry state with
335 the SSL_set_retry_verify() function.
336
337 *Tomáš Mráz*
338
de85a9de 339### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
32a3b9b7 340
5eef9e1d
MC
341 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
342 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
343 verify a certificate supplied by a server. That function may return a
344 negative return value to indicate an internal error (for example out of
345 memory). Such a negative return value is mishandled by OpenSSL and will cause
346 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
347 success and a subsequent call to SSL_get_error() to return the value
348 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
349 returned by OpenSSL if the application has previously called
350 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
351 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
352 totally unexpected and applications may not behave correctly as a result. The
353 exact behaviour will depend on the application but it could result in
354 crashes, infinite loops or other similar incorrect responses.
355
356 This issue is made more serious in combination with a separate bug in OpenSSL
357 3.0 that will cause X509_verify_cert() to indicate an internal error when
358 processing a certificate chain. This will occur where a certificate does not
359 include the Subject Alternative Name extension but where a Certificate
360 Authority has enforced name constraints. This issue can occur even with valid
361 chains.
362 ([CVE-2021-4044])
363
364 *Matt Caswell*
365
32a3b9b7
RL
366 * Corrected a few file name and file reference bugs in the build,
367 installation and setup scripts, which lead to installation verification
368 failures. Slightly enhanced the installation verification script.
369
370 *Richard Levitte*
371
c868d1f9
TM
372 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
373 keys.
44652c16 374
c868d1f9 375 *Richard Levitte*
b7140b06 376
c868d1f9
TM
377 * Fixed PVK encoder to properly query for the passphrase.
378
379 *Tomáš Mráz*
380
381 * Multiple fixes in the OSSL_HTTP API functions.
382
383 *David von Oheimb*
384
385 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
386 OSSL_PARAM_INTEGER data type and return error on negative numbers
387 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
388 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
389
390 *Richard Levitte*
391
392 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
393
394 *Tomáš Mráz*
395
396 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
397
398 *Allan Jude*
399
c868d1f9
TM
400 * Multiple threading fixes.
401
402 *Matt Caswell*
403
404 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
405
406 *Tomáš Mráz*
407
408 * Allow fetching an operation from the provider that owns an unexportable key
409 as a fallback if that is still allowed by the property query.
410
411 *Richard Levitte*
b7140b06 412
de85a9de 413### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
c7d4d032 414
95a444c9
TM
415 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
416 deprecated.
417
418 *Matt Caswell*
419
420 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
421 S390X capability vector to zero. This simplifies testing of different code
422 paths on S390X architecture.
423
424 *Patrick Steuer*
425
426 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
427 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
428 SP 800-38D". The communication will fail at this point.
429
430 *Paul Dale*
431
432 * The EC_GROUP_clear_free() function is deprecated as there is nothing
433 confidential in EC_GROUP data.
434
435 *Nicola Tuveri*
436
437 * The byte order mark (BOM) character is ignored if encountered at the
438 beginning of a PEM-formatted file.
439
440 *Dmitry Belyavskiy*
441
442 * Added CMS support for the Russian GOST algorithms.
443
444 *Dmitry Belyavskiy*
445
6f242d22
TM
446 * Due to move of the implementation of cryptographic operations
447 to the providers, validation of various operation parameters can
448 be postponed until the actual operation is executed where previously
449 it happened immediately when an operation parameter was set.
450
451 For example when setting an unsupported curve with
452 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
453 fail but later keygen operations with the EVP_PKEY_CTX will fail.
454
455 *OpenSSL team members and many third party contributors*
456
69222552 457 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
458 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
459 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
460 instead to retrieve these algorithms from a provider.
461
462 *Shane Lontis*
463
bd32bdb8
TM
464 * On build targets where the multilib postfix is set in the build
465 configuration the libdir directory was changing based on whether
466 the lib directory with the multilib postfix exists on the system
467 or not. This unpredictable behavior was removed and eventual
468 multilib postfix is now always added to the default libdir. Use
469 `--libdir=lib` to override the libdir if adding the postfix is
470 undesirable.
471
472 *Jan Lána*
473
e5f8935c
P
474 * The triple DES key wrap functionality now conforms to RFC 3217 but is
475 no longer interoperable with OpenSSL 1.1.1.
476
477 *Paul Dale*
478
0f71b1eb
P
479 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
480 function codes, this function can only cause problems for calling
481 applications.
482
483 *Paul Dale*
484
8c5bff22
WE
485 * Add a configurable flag to output date formats as ISO 8601. Does not
486 change the default date format.
487
488 *William Edmisten*
489
f8ab78f6
RS
490 * Version of MSVC earlier than 1300 could get link warnings, which could
491 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
492 Support for this flag has been removed.
493
494 *Rich Salz*
495
a935791d
RS
496 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
497 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
498 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
499 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
500 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
501
502 *Rich Salz*
503
f04bb0bc
RS
504 * The signatures of the functions to get and set options on SSL and
505 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
506 Some source code changes may be required.
507
a935791d 508 *Rich Salz*
f04bb0bc 509
ff234c68
RS
510 * The public definitions of conf_method_st and conf_st have been
511 deprecated. They will be made opaque in a future release.
512
b3c2ed70 513 *Rich Salz and Tomáš Mráz*
ff234c68 514
55373bfd
RS
515 * Client-initiated renegotiation is disabled by default. To allow it, use
516 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
517 flag, or the "ClientRenegotiation" config parameter as appropriate.
518
a935791d 519 *Rich Salz*
55373bfd 520
f7050588
RS
521 * Add "abspath" and "includedir" pragma's to config files, to prevent,
522 or modify relative pathname inclusion.
3fb985fd 523
a935791d 524 *Rich Salz*
3fb985fd 525
3b9e4769 526 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
b7140b06 527 validated. Please consult the README-FIPS and
3b9e4769
DMSP
528 README-PROVIDERS files, as well as the migration guide.
529
3b9e4769
DMSP
530 *OpenSSL team members and many third party contributors*
531
f1ffaaee 532 * For the key types DH and DHX the allowed settable parameters are now different.
f1ffaaee
SL
533
534 *Shane Lontis*
535
bee3f389 536 * The openssl commands that read keys, certificates, and CRLs now
b7140b06 537 automatically detect the PEM or DER format of the input files.
bee3f389
TM
538
539 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
540
b7140b06 541 * Added enhanced PKCS#12 APIs which accept a library context.
b536880c
JS
542
543 *Jon Spillett*
544
ae6f65ae
MC
545 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
546
547 *Matt Caswell*
548
b7140b06 549 * Added support for Kernel TLS (KTLS).
6878f430
MC
550
551 *Boris Pismenny, John Baldwin and Andrew Gallatin*
552
72d2670b 553 * Support for RFC 5746 secure renegotiation is now required by default for
b7140b06 554 SSL or TLS connections to succeed.
72d2670b
BK
555
556 *Benjamin Kaduk*
557
9ac653d8
TM
558 * The signature of the `copy` functional parameter of the
559 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
560 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
561 the signature of the `pub_decode` functional parameter of the
562 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
563 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
564
565 *David von Oheimb*
566
9c1b19eb 567 * The error return values from some control calls (ctrl) have changed.
9c1b19eb
P
568
569 *Paul Dale*
570
e454a393 571 * A public key check is now performed during EVP_PKEY_derive_set_peer().
e454a393
SL
572
573 *Shane Lontis*
574
31b7f23d
TM
575 * Many functions in the EVP_ namespace that are getters of values from
576 implementations or contexts were renamed to include get or get0 in their
577 names. Old names are provided as macro aliases for compatibility and
578 are not deprecated.
579
580 *Tomáš Mráz*
581
0cfbc828
TM
582 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
583 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
584 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
b7140b06 585 are deprecated.
0cfbc828
TM
586
587 *Tomáš Mráz*
588
2db5834c 589 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
b7140b06 590 more key types.
2db5834c 591
28a8d07d 592 * The output from the command line applications may have minor
b7140b06 593 changes.
28a8d07d
P
594
595 *Paul Dale*
596
b7140b06 597 * The output from numerous "printing" may have minor changes.
2db5834c
MC
598
599 *David von Oheimb*
600
f70863d9
VD
601 * Windows thread synchronization uses read/write primitives (SRWLock) when
602 supported by the OS, otherwise CriticalSection continues to be used.
603
604 *Vincent Drake*
605
a30823c8
SL
606 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
607 work on read only BIO source/sinks that do not support these functions.
608 This allows piping or redirection of a file BIO using stdin to be buffered
609 into memory. This is used internally in OSSL_DECODER_from_bio().
610
611 *Shane Lontis*
612
f74f416b
MC
613 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
614 this function would return one of the values OSSL_STORE_INFO_NAME,
615 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
616 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
617 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
618 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
619 using this function should be amended to handle the changed return value.
620
621 *Richard Levitte*
622
6b937ae3 623 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
63b64f19 624 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
bef876f9 625 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
6b937ae3
DDO
626 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
627 contains more than one certificate identifier: This means that all
628 certificates referenced there MUST be part of the validation chain.
629
630 *David von Oheimb*
631
b7140b06
SL
632 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
633 RC5, DESX and DES have been moved to the legacy provider.
c7d4d032
MC
634
635 *Matt Caswell*
636
637 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
b7140b06 638 RIPEMD-160 have been moved to the legacy provider.
c7d4d032
MC
639
640 *Matt Caswell*
641
896dcda1
DB
642 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
643 provided key.
8e53d94d 644
896dcda1
DB
645 *Dmitry Belyavskiy*
646
647 * The deprecated functions EVP_PKEY_get0_RSA(),
cc57dc96
MC
648 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
649 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
b7140b06
SL
650 well as the similarly named "get1" functions behave differently in
651 OpenSSL 3.0.
7bc0fdd3 652
cc57dc96
MC
653 *Matt Caswell*
654
4d49b685 655 * A number of functions handling low-level keys or engines were deprecated
8e53d94d
MC
656 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
657 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
b7140b06 658 EVP_PKEY_get0_siphash().
8e53d94d
MC
659
660 *Matt Caswell*
661
0f183675
JS
662 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
663 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
664 will need to load the legacy crypto provider. This includes these PBE
665 algorithms which use this KDF:
666 - NID_pbeWithMD2AndDES_CBC
667 - NID_pbeWithMD5AndDES_CBC
668 - NID_pbeWithSHA1AndRC2_CBC
669 - NID_pbeWithMD2AndRC2_CBC
670 - NID_pbeWithMD5AndRC2_CBC
671 - NID_pbeWithSHA1AndDES_CBC
672
673 *Jon Spillett*
674
0800318a
TM
675 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
676 BIO_debug_callback() functions.
677
678 *Tomáš Mráz*
679
76e48c9d 680 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
b7140b06 681 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
44652c16 682
76e48c9d
TM
683 *Tomáš Mráz*
684
b7140b06 685 * The RAND_METHOD APIs have been deprecated.
12631540
P
686
687 *Paul Dale*
8e53d94d 688
b7140b06 689 * The SRP APIs have been deprecated.
13888e79
MC
690
691 *Matt Caswell*
692
7dd5a00f
P
693 * Add a compile time option to prevent the caching of provider fetched
694 algorithms. This is enabled by including the no-cached-fetch option
695 at configuration time.
696
697 *Paul Dale*
76e48c9d 698
b7140b06
SL
699 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
700 count of PKCS12_DEFAULT_ITER.
762970bd
TM
701
702 *Tomáš Mráz and Sahana Prasad*
703
b7140b06 704 * The openssl speed command does not use low-level API calls anymore.
f3ccfc76
TM
705
706 *Tomáš Mráz*
707
c781eb1c
AM
708 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
709 capable processors.
710
711 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
712
a763ca11 713 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
a763ca11
MC
714
715 *Matt Caswell*
716
f5680cd0
MC
717 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
718 providers may supply their own group implementations (using either the "key
719 exchange" or the "key encapsulation" methods) which will automatically be
720 detected and used by libssl.
721
722 *Matt Caswell, Nicola Tuveri*
723
7ff9fdd4 724 * The undocumented function X509_certificate_type() has been deprecated;
7ff9fdd4
RS
725
726 *Rich Salz*
727
b7140b06 728 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
4d2a6159
TM
729
730 *Tomáš Mráz*
731
b0aae913
RS
732 * Removed RSA padding mode for SSLv23 (which was only used for
733 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
734 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
735 `rsautl` command.
736
737 *Rich Salz*
738
b7140b06 739 * Deprecated the obsolete X9.31 RSA key generation related functions.
c27e7922 740
4672e5de
DDO
741 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
742 is not allowed to return a value > 1, this is no more taken as failure.
743
744 *Viktor Dukhovni and David von Oheimb*
745
746 * Deprecated the obsolete X9.31 RSA key generation related functions
747 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
748 BN_X931_generate_prime_ex().
749
66194839 750 *Tomáš Mráz*
c27e7922 751
93b39c85 752 * The default key generation method for the regular 2-prime RSA keys was
b7140b06 753 changed to the FIPS 186-4 B.3.6 method.
93b39c85
TM
754
755 *Shane Lontis*
756
757 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
93b39c85
TM
758
759 *Kurt Roeckx*
760
b7140b06 761 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1409b5f6
RS
762
763 *Rich Salz*
764
b7140b06
SL
765 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
766 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
83b6dc8d 767
8f965908 768 *Rich Salz, Richard Levitte, and David von Oheimb*
83b6dc8d 769
b7140b06 770 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
a07b0bfb
DDO
771
772 *David von Oheimb*
773
b7140b06 774 * Deprecated `OCSP_parse_url()`.
7932982b
DDO
775
776 *David von Oheimb*
777
9e49aff2 778 * Validation of SM2 keys has been separated from the validation of regular EC
b7140b06 779 keys.
9e49aff2
NT
780
781 *Nicola Tuveri*
782
ed37336b
NT
783 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
784 switches: a validation failure triggers an early exit, returning a failure
785 exit status to the parent process.
786
787 *Nicola Tuveri*
788
1c47539a
OH
789 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
790 to ignore unknown ciphers.
791
792 *Otto Hollmann*
793
ec2bfb7d
DDO
794 * The `-cipher-commands` and `-digest-commands` options
795 of the command line utility `list` have been deprecated.
796 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
a08489e2
DB
797
798 *Dmitry Belyavskiy*
799
f9253152
DDO
800 * Added convenience functions for generating asymmetric key pairs:
801 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
802 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
803
804 *David von Oheimb*
805
b7140b06 806 * All of the low level EC_KEY functions have been deprecated.
5b5eea4b 807
66194839 808 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
5b5eea4b 809
f5a46ed7 810 * Deprecated all the libcrypto and libssl error string loading
b7140b06 811 functions.
f5a46ed7
RL
812
813 *Richard Levitte*
814
1b2a55ff
MC
815 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
816 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
b7140b06 817 deprecated.
1b2a55ff
MC
818
819 *Matt Caswell*
820
ec2bfb7d 821 * The `-crypt` option to the `passwd` command line tool has been removed.
c87a7f31
P
822
823 *Paul Dale*
824
ec2bfb7d 825 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
a18cf8fc 826 were removed.
1696b890
RS
827
828 *Rich Salz*
829
8ea761bf 830 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
8ea761bf
SL
831
832 *Shane Lontis*
833
0a737e16 834 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
b7140b06 835 EVP_PKEY_get1_tls_encodedpoint().
0a737e16
MC
836
837 *Matt Caswell*
838
372e72b1 839 * The security callback, which can be customised by application code, supports
b7140b06
SL
840 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
841 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
372e72b1
MC
842
843 *Matt Caswell*
844
db554ae1
JM
845 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
846 interface. Their functionality remains unchanged.
847
848 *Jordan Montgomery*
849
f4bd5105
P
850 * Added new option for 'openssl list', '-providers', which will display the
851 list of loaded providers, their names, version and status. It optionally
852 displays their gettable parameters.
853
854 *Paul Dale*
855
b7140b06 856 * Removed EVP_PKEY_set_alias_type().
14711fff
RL
857
858 *Richard Levitte*
859
ec2bfb7d
DDO
860 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
861 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
3786d748 862
863 *Jeremy Walch*
864
31605414
MC
865 * Changed all "STACK" functions to be macros instead of inline functions. Macro
866 parameters are still checked for type safety at compile time via helper
867 inline functions.
868
869 *Matt Caswell*
870
7d615e21
P
871 * Remove the RAND_DRBG API
872
7d615e21
P
873 *Paul Dale and Matthias St. Pierre*
874
ec2bfb7d 875 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
37d898df
DW
876 as well as actual hostnames.
877
878 *David Woodhouse*
879
77174598
VD
880 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
881 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
882 conversely, silently ignore DTLS protocol version bounds when configuring
883 TLS-based contexts. The commands can be repeated to set bounds of both
884 types. The same applies with the corresponding "min_protocol" and
885 "max_protocol" command-line switches, in case some application uses both TLS
886 and DTLS.
887
888 SSL_CTX instances that are created for a fixed protocol version (e.g.
ec2bfb7d 889 `TLSv1_server_method()`) also silently ignore version bounds. Previously
77174598
VD
890 attempts to apply bounds to these protocol versions would result in an
891 error. Now only the "version-flexible" SSL_CTX instances are subject to
892 limits in configuration files in command-line options.
893
894 *Viktor Dukhovni*
895
8dab4de5
RL
896 * Deprecated the `ENGINE` API. Engines should be replaced with providers
897 going forward.
898
899 *Paul Dale*
900
901 * Reworked the recorded ERR codes to make better space for system errors.
902 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
903 given code is a system error (true) or an OpenSSL error (false).
904
905 *Richard Levitte*
906
907 * Reworked the test perl framework to better allow parallel testing.
908
909 *Nicola Tuveri and David von Oheimb*
910
7cc355c2
SL
911 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
912 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
913
914 *Shane Lontis*
915
16b0e0fc
RL
916 * 'Configure' has been changed to figure out the configuration target if
917 none is given on the command line. Consequently, the 'config' script is
918 now only a mere wrapper. All documentation is changed to only mention
919 'Configure'.
920
921 *Rich Salz and Richard Levitte*
922
b4250010
DMSP
923 * Added a library context `OSSL_LIB_CTX` that applications as well as
924 other libraries can use to form a separate context within which
925 libcrypto operations are performed.
3bd65f9b 926
3bd65f9b
RL
927 *Richard Levitte*
928
95a444c9
TM
929 * Added various `_ex` functions to the OpenSSL API that support using
930 a non-default `OSSL_LIB_CTX`.
931
932 *OpenSSL team*
933
11d3235e
TM
934 * Handshake now fails if Extended Master Secret extension is dropped
935 on renegotiation.
936
66194839 937 *Tomáš Mráz*
11d3235e 938
b7140b06 939 * Dropped interactive mode from the `openssl` program.
eca47139
RL
940
941 *Richard Levitte*
942
b7140b06 943 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
987e3a0e 944
c85c5e1a 945 *David von Oheimb and Shane Lontis*
987e3a0e 946
b7140b06 947 * Deprecated `EC_METHOD_get_field_type()`.
23ccae80
BB
948
949 *Billy Bob Brumley*
950
951 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
952 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
953 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
23ccae80
BB
954
955 *Billy Bob Brumley*
956
957 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
23ccae80
BB
958
959 *Billy Bob Brumley*
960
9e3c510b
F
961 * Add CAdES-BES signature verification support, mostly derived
962 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
963
964 *Filipe Raimundo da Silva*
965
966 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
967
968 *Antonio Iacono*
969
34347512 970 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
b7140b06 971 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
34347512
JZ
972
973 *Jakub Zelenka*
974
b7140b06 975 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
6b4eb933 976
c2f2db9b
BB
977 *Billy Bob Brumley*
978
979 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
b7140b06 980 EC_KEY_precompute_mult().
c2f2db9b
BB
981
982 *Billy Bob Brumley*
6b4eb933 983
b7140b06 984 * Deprecated EC_POINTs_mul().
4fcd15c1
BB
985
986 *Billy Bob Brumley*
987
b7140b06 988 * Removed FIPS_mode() and FIPS_mode_set().
31b069ec
SL
989
990 *Shane Lontis*
991
b7140b06 992 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
09b90e0e
DB
993
994 *Dmitry Belyavskiy*
995
07caec83 996 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
b7140b06 997 EC_POINT_get_Jprojective_coordinates_GFp().
07caec83
BB
998
999 *Billy Bob Brumley*
1000
be19d3ca
P
1001 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1002 arrays to be more easily constructed via a series of utility functions.
1003 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1004 the various push functions and finally convert to a passable OSSL_PARAM
1005 array using OSSL_PARAM_BLD_to_param().
1006
ccb8f0c8 1007 *Paul Dale*
be19d3ca 1008
aba03ae5 1009 * The security strength of SHA1 and MD5 based signatures in TLS has been
b7140b06 1010 reduced.
aba03ae5
KR
1011
1012 *Kurt Roeckx*
1013
8243d8d1
RL
1014 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1015 contain a provider side internal key.
1016
1017 *Richard Levitte*
1018
ccb8f0c8 1019 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
12d99aac
RL
1020
1021 *Richard Levitte*
c50604eb 1022
036cbb6b 1023 * Project text documents not yet having a proper file name extension
1dc1ea18
DDO
1024 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1025 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
036cbb6b
DDO
1026
1027 *David von Oheimb*
1028
1dc1ea18 1029 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
c50604eb
DMSP
1030 have been converted to Markdown with the goal to produce documents
1031 which not only look pretty when viewed online in the browser, but
1032 remain well readable inside a plain text editor.
1033
1034 To achieve this goal, a 'minimalistic' Markdown style has been applied
1035 which avoids formatting elements that interfere too much with the
1036 reading flow in the text file. For example, it
1037
1038 * avoids [ATX headings][] and uses [setext headings][] instead
1039 (which works for `<h1>` and `<h2>` headings only).
1040 * avoids [inline links][] and uses [reference links][] instead.
1041 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1042
1043 [ATX headings]: https://github.github.com/gfm/#atx-headings
1044 [setext headings]: https://github.github.com/gfm/#setext-headings
1045 [inline links]: https://github.github.com/gfm/#inline-link
1046 [reference links]: https://github.github.com/gfm/#reference-link
1047 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1048 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1049
1050 *Matthias St. Pierre*
1051
44652c16
DMSP
1052 * The test suite is changed to preserve results of each test recipe.
1053 A new directory test-runs/ with subdirectories named like the
1054 test recipes are created in the build tree for this purpose.
1055
1056 *Richard Levitte*
1057
e7774c28 1058 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
ec2bfb7d 1059 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
8d9a4d83 1060 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
e7774c28 1061
8d9a4d83 1062 *David von Oheimb, Martin Peylo*
e7774c28 1063
ec2bfb7d 1064 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
8f965908
DDO
1065 It supports arbitrary request and response content types, GET redirection,
1066 TLS, connections via HTTP(S) proxies, connections and exchange via
1067 user-defined BIOs (allowing implicit connections), persistent connections,
1068 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1069 The legacy OCSP-focused (and only partly documented) API
1070 is retained for backward compatibility, while most of it is deprecated.
e7774c28
DDO
1071
1072 *David von Oheimb*
1073
16c6534b
DDO
1074 * Added `util/check-format.pl`, a tool for checking adherence to the
1075 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1076 The checks performed are incomplete and yield some false positives.
1077 Still the tool should be useful for detecting most typical glitches.
1078
1079 *David von Oheimb*
1080
ec2bfb7d 1081 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
59131529 1082 If domain name resolution yields multiple IP addresses all of them are tried
ec2bfb7d 1083 after `connect()` failures.
59131529
DDO
1084
1085 *David von Oheimb*
1086
b7140b06 1087 * All of the low level RSA functions have been deprecated.
b47e7bbc 1088
44652c16
DMSP
1089 *Paul Dale*
1090
1091 * X509 certificates signed using SHA1 are no longer allowed at security
1092 level 1 and above.
44652c16
DMSP
1093
1094 *Kurt Roeckx*
1095
1096 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
b304f856
P
1097 modified to use PKEY APIs. These commands are now in maintenance mode
1098 and no new features will be added to them.
1099
1100 *Paul Dale*
1101
1102 * The command line utility rsautl has been deprecated.
b304f856
P
1103
1104 *Paul Dale*
1105
1106 * The command line utilities genrsa and rsa have been modified to use PKEY
10203a34
KR
1107 APIs. They now write PKCS#8 keys by default. These commands are now in
1108 maintenance mode and no new features will be added to them.
44652c16
DMSP
1109
1110 *Paul Dale*
1111
b7140b06 1112 * All of the low level DH functions have been deprecated.
59d7ad07
MC
1113
1114 *Paul Dale and Matt Caswell*
44652c16 1115
b7140b06 1116 * All of the low level DSA functions have been deprecated.
8e53d94d 1117
44652c16
DMSP
1118 *Paul Dale*
1119
1120 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
b7140b06 1121 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
44652c16
DMSP
1122
1123 *Richard Levitte*
1124
b7140b06 1125 * Deprecated low level ECDH and ECDSA functions.
44652c16
DMSP
1126
1127 *Paul Dale*
1128
b7140b06 1129 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
44652c16
DMSP
1130
1131 *Richard Levitte*
1132
ed576acd
TM
1133 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1134 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
44652c16
DMSP
1135 a new formulation to include all the things it can be used for,
1136 as well as words of caution.
1137
1138 *Richard Levitte*
1139
1140 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
44652c16
DMSP
1141
1142 *Paul Dale*
1143
b7140b06 1144 * All of the low level HMAC functions have been deprecated.
44652c16 1145
0a8a6afd 1146 *Paul Dale and David von Oheimb*
44652c16
DMSP
1147
1148 * Over two thousand fixes were made to the documentation, including:
1149 - Common options (such as -rand/-writerand, TLS version control, etc)
1150 were refactored and point to newly-enhanced descriptions in openssl.pod.
1151 - Added style conformance for all options (with help from Richard Levitte),
1152 documented all reported missing options, added a CI build to check
1153 that all options are documented and that no unimplemented options
1154 are documented.
1155 - Documented some internals, such as all use of environment variables.
1156 - Addressed all internal broken L<> references.
1157
1158 *Rich Salz*
1159
b7140b06 1160 * All of the low level CMAC functions have been deprecated.
44652c16
DMSP
1161
1162 *Paul Dale*
1163
1dc8eb5b
P
1164 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1165 functions have been deprecated.
44652c16 1166
4d49b685 1167 *Paul Dale and David von Oheimb*
44652c16 1168
257e9d03 1169 * Corrected the documentation of the return values from the `EVP_DigestSign*`
44652c16
DMSP
1170 set of functions. The documentation mentioned negative values for some
1171 errors, but this was never the case, so the mention of negative values
1172 was removed.
1173
1174 Code that followed the documentation and thereby check with something
1175 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1176
1177 *Richard Levitte*
1178
b7140b06 1179 * All of the low level cipher functions have been deprecated.
44652c16
DMSP
1180
1181 *Matt Caswell and Paul Dale*
5f8e6c50
DMSP
1182
1183 * Removed include/openssl/opensslconf.h.in and replaced it with
1184 include/openssl/configuration.h.in, which differs in not including
1185 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1186 was added to include both.
44652c16 1187
5f8e6c50
DMSP
1188 This allows internal hacks where one might need to modify the set
1189 of configured macros, for example this if deprecated symbols are
1190 still supposed to be available internally:
44652c16 1191
5f8e6c50 1192 #include <openssl/configuration.h>
44652c16 1193
5f8e6c50
DMSP
1194 #undef OPENSSL_NO_DEPRECATED
1195 #define OPENSSL_SUPPRESS_DEPRECATED
44652c16 1196
5f8e6c50 1197 #include <openssl/macros.h>
44652c16 1198
5f8e6c50
DMSP
1199 This should not be used by applications that use the exported
1200 symbols, as that will lead to linking errors.
44652c16 1201
5f8e6c50
DMSP
1202 *Richard Levitte*
1203
44652c16
DMSP
1204 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1205 used in exponentiation with 512-bit moduli. No EC algorithms are
1206 affected. Analysis suggests that attacks against 2-prime RSA1024,
1207 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1208 difficult to perform and are not believed likely. Attacks against DH512
1209 are considered just feasible. However, for an attack the target would
1210 have to re-use the DH512 private key, which is not recommended anyway.
4d49b685 1211 Also applications directly using the low-level API BN_mod_exp may be
44652c16 1212 affected if they use BN_FLG_CONSTTIME.
d8dc8538 1213 ([CVE-2019-1551])
44652c16
DMSP
1214
1215 *Andy Polyakov*
5f8e6c50 1216
44652c16
DMSP
1217 * Most memory-debug features have been deprecated, and the functionality
1218 replaced with no-ops.
5f8e6c50 1219
44652c16 1220 *Rich Salz*
257e9d03 1221
31605414 1222 * Added documentation for the STACK API.
257e9d03 1223
852c2ed2 1224 *Rich Salz*
5f8e6c50 1225
02649104
RL
1226 * Introduced a new method type and API, OSSL_ENCODER, to represent
1227 generic encoders. These do the same sort of job that PEM writers
1228 and d2i functions do, but with support for methods supplied by
1229 providers, and the possibility for providers to support other
1230 formats as well.
1231
1232 *Richard Levitte*
1233
1234 * Introduced a new method type and API, OSSL_DECODER, to represent
1235 generic decoders. These do the same sort of job that PEM readers
1236 and i2d functions do, but with support for methods supplied by
1237 providers, and the possibility for providers to support other
1238 formats as well.
5f8e6c50
DMSP
1239
1240 *Richard Levitte*
1241
1242 * Added a .pragma directive to the syntax of configuration files, to
1243 allow varying behavior in a supported and predictable manner.
1244 Currently added pragma:
1245
1246 .pragma dollarid:on
1247
1248 This allows dollar signs to be a keyword character unless it's
1249 followed by a opening brace or parenthesis. This is useful for
1250 platforms where dollar signs are commonly used in names, such as
1251 volume names and system directory names on VMS.
1252
1253 *Richard Levitte*
1254
b7140b06 1255 * Added functionality to create an EVP_PKEY from user data.
5f8e6c50
DMSP
1256
1257 *Richard Levitte*
536454e5 1258
5f8e6c50
DMSP
1259 * Change the interpretation of the '--api' configuration option to
1260 mean that this is a desired API compatibility level with no
1261 further meaning. The previous interpretation, that this would
1262 also mean to remove all deprecated symbols up to and including
1263 the given version, no requires that 'no-deprecated' is also used
1264 in the configuration.
1265
1266 When building applications, the desired API compatibility level
1267 can be set with the OPENSSL_API_COMPAT macro like before. For
1268 API compatibility version below 3.0, the old style numerical
1269 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1270 For version 3.0 and on, the value is expected to be the decimal
1271 value calculated from the major and minor version like this:
38c65481 1272
5f8e6c50 1273 MAJOR * 10000 + MINOR * 100
38c65481 1274
5f8e6c50 1275 Examples:
ea8c77a5 1276
5f8e6c50
DMSP
1277 -DOPENSSL_API_COMPAT=30000 For 3.0
1278 -DOPENSSL_API_COMPAT=30200 For 3.2
1279
1280 To hide declarations that are deprecated up to and including the
1281 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1282 given when building the application as well.
390c5795 1283
5f8e6c50 1284 *Richard Levitte*
e5641d7f 1285
5f8e6c50
DMSP
1286 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1287 access to certificate and CRL stores via URIs and OSSL_STORE
1288 loaders.
e5641d7f 1289
5f8e6c50 1290 This adds the following functions:
3ddc06f0 1291
5f8e6c50
DMSP
1292 - X509_LOOKUP_store()
1293 - X509_STORE_load_file()
1294 - X509_STORE_load_path()
1295 - X509_STORE_load_store()
1296 - SSL_add_store_cert_subjects_to_stack()
1297 - SSL_CTX_set_default_verify_store()
1298 - SSL_CTX_load_verify_file()
1299 - SSL_CTX_load_verify_dir()
1300 - SSL_CTX_load_verify_store()
e66cb363 1301
5f8e6c50 1302 *Richard Levitte*
732d31be 1303
5f8e6c50
DMSP
1304 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1305 The presence of this system service is determined at run-time.
223c59ea 1306
5f8e6c50 1307 *Richard Levitte*
173350bc 1308
5f8e6c50
DMSP
1309 * Added functionality to create an EVP_PKEY context based on data
1310 for methods from providers. This takes an algorithm name and a
1311 property query string and simply stores them, with the intent
1312 that any operation that uses this context will use those strings
1313 to fetch the needed methods implicitly, thereby making the port
1314 of application written for pre-3.0 OpenSSL easier.
acf20c7d 1315
5f8e6c50 1316 *Richard Levitte*
3d63b396 1317
5f8e6c50
DMSP
1318 * The undocumented function NCONF_WIN32() has been deprecated; for
1319 conversion details see the HISTORY section of doc/man5/config.pod
c2bf7208 1320
5f8e6c50 1321 *Rich Salz*
ba64ae6c 1322
5f8e6c50
DMSP
1323 * Introduced the new functions EVP_DigestSignInit_ex() and
1324 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1325 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1326 pages for further details.
0e0c6821 1327
5f8e6c50 1328 *Matt Caswell*
e6f418bc 1329
5f8e6c50
DMSP
1330 * Over two thousand fixes were made to the documentation, including:
1331 adding missing command flags, better style conformance, documentation
1332 of internals, etc.
3d63b396 1333
5f8e6c50 1334 *Rich Salz, Richard Levitte*
3d63b396 1335
5f8e6c50
DMSP
1336 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1337 X25519, X448, Ed25519 and Ed448.
a25f33d2 1338
5f8e6c50 1339 *Patrick Steuer*
17716680 1340
5f8e6c50
DMSP
1341 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1342 the first value.
0e4bc563 1343
5f8e6c50 1344 *Jon Spillett*
e30dd20c 1345
ec2bfb7d
DDO
1346 * Deprecated the public definition of `ERR_STATE` as well as the function
1347 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
5f8e6c50 1348 opaque type.
c05353c5 1349
5f8e6c50 1350 *Richard Levitte*
d741ccad 1351
5f8e6c50
DMSP
1352 * Added ERR functionality to give callers access to the stored function
1353 names that have replaced the older function code based functions.
aaf35f11 1354
af2f14ac
RL
1355 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1356 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1357 ERR_peek_error_all() and ERR_peek_last_error_all().
1358
b7140b06
SL
1359 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1360 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1361 ERR_func_error_string().
aaf35f11 1362
5f8e6c50 1363 *Richard Levitte*
3ff55e96 1364
5f8e6c50
DMSP
1365 * Extended testing to be verbose for failing tests only. The make variables
1366 VERBOSE_FAILURE or VF can be used to enable this:
b6af2c7e 1367
5f8e6c50
DMSP
1368 $ make VF=1 test # Unix
1369 $ mms /macro=(VF=1) test ! OpenVMS
1370 $ nmake VF=1 test # Windows
77202a85 1371
5f8e6c50 1372 *Richard Levitte*
57f39cc8 1373
b9fbacaa
DDO
1374 * Added the `-copy_extensions` option to the `x509` command for use with
1375 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1376 all extensions in the request are copied to the certificate or vice versa.
1377
1378 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1379
1380 * Added the `-copy_extensions` option to the `req` command for use with
1381 `-x509`. When given with the `copy` or `copyall` argument,
1382 all extensions in the certification request are copied to the certificate.
b65c5ec8
DDO
1383
1384 *David von Oheimb*
1385
b9fbacaa
DDO
1386 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1387 they generate are by default RFC 5280 compliant in the following sense:
1388 There is a subjectKeyIdentifier extension with a hash value of the public key
1389 and for not self-signed certs there is an authorityKeyIdentifier extension
1390 with a keyIdentifier field or issuer information identifying the signing key.
ec2bfb7d 1391 This is done unless some configuration overrides the new default behavior,
b9fbacaa 1392 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
ec2bfb7d
DDO
1393
1394 *David von Oheimb*
1395
1396 * Added several checks to `X509_verify_cert()` according to requirements in
0e071fbc
DO
1397 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1398 (which may be done by using the CLI option `-x509_strict`):
1399 * The basicConstraints of CA certificates must be marked critical.
1400 * CA certificates must explicitly include the keyUsage extension.
1401 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1402 * The issuer name of any certificate must not be empty.
1403 * The subject name of CA certs, certs with keyUsage crlSign,
1404 and certs without subjectAlternativeName must not be empty.
1405 * If a subjectAlternativeName extension is given it must not be empty.
1406 * The signatureAlgorithm field and the cert signature must be consistent.
1407 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1408 must not be marked critical.
1409 * The authorityKeyIdentifier must be given for X.509v3 certs
1410 unless they are self-signed.
1411 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1412
1413 *David von Oheimb*
1414
ec2bfb7d 1415 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
0e071fbc
DO
1416 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1417
66194839 1418 *Tomáš Mráz*
0e071fbc 1419
5f8e6c50 1420 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
ece9304c 1421 used even when parsing explicit parameters, when loading a encoded key
5f8e6c50
DMSP
1422 or calling `EC_GROUP_new_from_ecpkparameters()`/
1423 `EC_GROUP_new_from_ecparameters()`.
1424 This prevents bypass of security hardening and performance gains,
1425 especially for curves with specialized EC_METHODs.
1426 By default, if a key encoded with explicit parameters is loaded and later
ece9304c 1427 encoded, the output is still encoded with explicit parameters, even if
5f8e6c50 1428 internally a "named" EC_GROUP is used for computation.
480af99e 1429
5f8e6c50 1430 *Nicola Tuveri*
480af99e 1431
5f8e6c50
DMSP
1432 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1433 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1434 NULL. After this change, only the cofactor parameter can be NULL. It also
1435 does some minimal sanity checks on the passed order.
d8dc8538 1436 ([CVE-2019-1547])
bab53405 1437
5f8e6c50 1438 *Billy Bob Brumley*
31636a3e 1439
5f8e6c50
DMSP
1440 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1441 An attack is simple, if the first CMS_recipientInfo is valid but the
1442 second CMS_recipientInfo is chosen ciphertext. If the second
1443 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1444 encryption key will be replaced by garbage, and the message cannot be
1445 decoded, but if the RSA decryption fails, the correct encryption key is
1446 used and the recipient will not notice the attack.
1447 As a work around for this potential attack the length of the decrypted
1448 key must be equal to the cipher default key length, in case the
1449 certifiate is not given and all recipientInfo are tried out.
1450 The old behaviour can be re-enabled in the CMS code by setting the
1451 CMS_DEBUG_DECRYPT flag.
60aee6ce 1452
5f8e6c50 1453 *Bernd Edlinger*
31636a3e 1454
5f8e6c50
DMSP
1455 * Early start up entropy quality from the DEVRANDOM seed source has been
1456 improved for older Linux systems. The RAND subsystem will wait for
1457 /dev/random to be producing output before seeding from /dev/urandom.
1458 The seeded state is stored for future library initialisations using
1459 a system global shared memory segment. The shared memory identifier
1460 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1461 the desired value. The default identifier is 114.
31636a3e 1462
5f8e6c50 1463 *Paul Dale*
7a762197 1464
5f8e6c50
DMSP
1465 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1466 when primes for RSA keys are computed.
1467 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1468 the 2-prime and 3-prime RSA modules were easy to distinguish, since
257e9d03 1469 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
5f8e6c50
DMSP
1470 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1471 This avoids possible fingerprinting of newly generated RSA modules.
6caa4edd 1472
5f8e6c50 1473 *Bernd Edlinger*
28b6d502 1474
5f8e6c50
DMSP
1475 * Correct the extended master secret constant on EBCDIC systems. Without this
1476 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1477 negotiate EMS will fail. Unfortunately this also means that TLS connections
1478 between EBCDIC systems with this fix, and EBCDIC systems without this
1479 fix will fail if they negotiate EMS.
d5bbead4 1480
5f8e6c50 1481 *Matt Caswell*
837f2fc7 1482
5f8e6c50
DMSP
1483 * Changed the library initialisation so that the config file is now loaded
1484 by default. This was already the case for libssl. It now occurs for both
1485 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
ec2bfb7d 1486 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
6bf79e30 1487
5f8e6c50 1488 *Matt Caswell*
480af99e 1489
ec2bfb7d
DDO
1490 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1491 where the former acts as a replacement for `ERR_put_error()`, and the
1492 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1493 `ERR_raise_data()` adds more flexibility by taking a format string and
5f8e6c50 1494 an arbitrary number of arguments following it, to be processed with
ec2bfb7d 1495 `BIO_snprintf()`.
e65bcbce 1496
5f8e6c50 1497 *Richard Levitte*
db99c525 1498
ec2bfb7d 1499 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
5f8e6c50
DMSP
1500 to check if a named provider is loaded and available. When called, it
1501 will also activate all fallback providers if such are still present.
db99c525 1502
5f8e6c50 1503 *Richard Levitte*
db99c525 1504
5f8e6c50 1505 * Enforce a minimum DH modulus size of 512 bits.
f8d6be3f 1506
5f8e6c50 1507 *Bernd Edlinger*
f8d6be3f 1508
5f8e6c50
DMSP
1509 * Changed DH parameters to generate the order q subgroup instead of 2q.
1510 Previously generated DH parameters are still accepted by DH_check
1511 but DH_generate_key works around that by clearing bit 0 of the
1512 private key for those. This avoids leaking bit 0 of the private key.
f8d6be3f 1513
5f8e6c50 1514 *Bernd Edlinger*
f8d6be3f 1515
5f8e6c50 1516 * Significantly reduce secure memory usage by the randomness pools.
f8d6be3f 1517
5f8e6c50 1518 *Paul Dale*
f8d6be3f 1519
257e9d03 1520 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
5f8e6c50 1521 deprecated.
1a489c9a 1522
5f8e6c50 1523 *Rich Salz*
8528128b 1524
5f8e6c50
DMSP
1525 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1526 algorithms. An implementation of a key exchange algorithm can be obtained
1527 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1528 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1529 the older EVP_PKEY_derive_init() function. See the man pages for the new
1530 functions for further details.
8228fd89 1531
5f8e6c50 1532 *Matt Caswell*
adb92d56 1533
5f8e6c50 1534 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
6bf79e30 1535
5f8e6c50 1536 *Matt Caswell*
adb92d56 1537
5f8e6c50
DMSP
1538 * Removed the function names from error messages and deprecated the
1539 xxx_F_xxx define's.
6bf79e30 1540
0f71b1eb
P
1541 *Richard Levitte*
1542
5f8e6c50 1543 * Removed NextStep support and the macro OPENSSL_UNISTD
db99c525 1544
5f8e6c50 1545 *Rich Salz*
94fd382f 1546
5f8e6c50
DMSP
1547 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1548 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1549 Also removed "export var as function" capability; we do not export
1550 variables, only functions.
e194fe8f 1551
5f8e6c50 1552 *Rich Salz*
40a70628 1553
5f8e6c50
DMSP
1554 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1555 an error and 1 indicating success. In previous versions of OpenSSL this
1556 was a void type. If a key was set longer than the maximum possible this
1557 would crash.
c2c2e7a4 1558
5f8e6c50 1559 *Matt Caswell*
c2c2e7a4 1560
5f8e6c50 1561 * Support SM2 signing and verification schemes with X509 certificate.
d18ef847 1562
5f8e6c50 1563 *Paul Yang*
d357be38 1564
ec2bfb7d 1565 * Use SHA256 as the default digest for TS query in the `ts` app.
b615ad90 1566
66194839 1567 *Tomáš Mráz*
0ebfcc8f 1568
5f8e6c50 1569 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
62bad771 1570
5f8e6c50 1571 *Shane Lontis*
1ad2ecb6 1572
5f8e6c50
DMSP
1573 * Default cipher lists/suites are now available via a function, the
1574 #defines are deprecated.
bd3576d2 1575
5f8e6c50 1576 *Todd Short*
b64f8256 1577
5f8e6c50
DMSP
1578 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1579 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1580 for Windows Store apps easier. Also, the "no-uplink" option has been added.
a9be3af5 1581
5f8e6c50 1582 *Kenji Mouri*
47339f61 1583
5f8e6c50 1584 * Join the directories crypto/x509 and crypto/x509v3
b0b7b1c5 1585
5f8e6c50 1586 *Richard Levitte*
6d311938 1587
5f8e6c50 1588 * Added command 'openssl kdf' that uses the EVP_KDF API.
90b8bbb8 1589
5f8e6c50 1590 *Shane Lontis*
22a4f969 1591
5f8e6c50 1592 * Added command 'openssl mac' that uses the EVP_MAC API.
41b731f2 1593
5f8e6c50 1594 *Shane Lontis*
e778802f 1595
5f8e6c50
DMSP
1596 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1597 as default directories. Also added the command 'openssl info'
1598 for scripting purposes.
1d48dd00 1599
5f8e6c50 1600 *Richard Levitte*
28a98809 1601
5f8e6c50 1602 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
b7140b06 1603 deprecated.
8f7de4f0 1604
5f8e6c50 1605 *Matt Caswell*
5fbe91d8 1606
5f8e6c50 1607 * Add prediction resistance to the DRBG reseeding process.
9263e882 1608
5f8e6c50 1609 *Paul Dale*
f73e07cf 1610
5f8e6c50
DMSP
1611 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1612 mandated by IEEE Std 1619-2018.
f9a25931 1613
5f8e6c50 1614 *Paul Dale*
2f0cd195 1615
5f8e6c50 1616 * Added newline escaping functionality to a filename when using openssl dgst.
257e9d03 1617 This output format is to replicate the output format found in the `*sum`
5f8e6c50 1618 checksum programs. This aims to preserve backward compatibility.
268c2102 1619
5f8e6c50 1620 *Matt Eaton, Richard Levitte, and Paul Dale*
fc8ee06b 1621
5f8e6c50
DMSP
1622 * Removed the heartbeat message in DTLS feature, as it has very
1623 little usage and doesn't seem to fulfill a valuable purpose.
1624 The configuration option is now deprecated.
c7ac31e2 1625
5f8e6c50 1626 *Richard Levitte*
9d892e28 1627
5f8e6c50
DMSP
1628 * Changed the output of 'openssl {digestname} < file' to display the
1629 digest name in its output.
9d892e28 1630
5f8e6c50 1631 *Richard Levitte*
ee13f9b1 1632
5f8e6c50 1633 * Added a new generic trace API which provides support for enabling
b7140b06 1634 instrumentation through trace output.
cb0f35d7 1635
5f8e6c50 1636 *Richard Levitte & Matthias St. Pierre*
cfcf6453 1637
5f8e6c50
DMSP
1638 * Added build tests for C++. These are generated files that only do one
1639 thing, to include one public OpenSSL head file each. This tests that
1640 the public header files can be usefully included in a C++ application.
cdbb8c2f 1641
5f8e6c50
DMSP
1642 This test isn't enabled by default. It can be enabled with the option
1643 'enable-buildtest-c++'.
06d5b162 1644
5f8e6c50 1645 *Richard Levitte*
c35f549e 1646
95a444c9
TM
1647 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1648
1649 *Robbie Harwood*
1650
1651 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1652
1653 *Simo Sorce*
1654
1655 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
ebc828ca 1656
5f8e6c50 1657 *Shane Lontis*
79e259e3 1658
95a444c9 1659 * Added KMAC to EVP_MAC.
56ee3117 1660
5f8e6c50 1661 *Shane Lontis*
6063b27b 1662
5f8e6c50
DMSP
1663 * Added property based algorithm implementation selection framework to
1664 the core.
6063b27b 1665
5f8e6c50 1666 *Paul Dale*
6063b27b 1667
5f8e6c50
DMSP
1668 * Added SCA hardening for modular field inversion in EC_GROUP through
1669 a new dedicated field_inv() pointer in EC_METHOD.
1670 This also addresses a leakage affecting conversions from projective
1671 to affine coordinates.
792a9002 1672
5f8e6c50 1673 *Billy Bob Brumley, Nicola Tuveri*
792a9002 1674
5f8e6c50
DMSP
1675 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1676 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1677 those algorithms that were already supported through the EVP_PKEY API
1678 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1679 and scrypt are now wrappers that call EVP_KDF.
792a9002 1680
5f8e6c50 1681 *David Makepeace*
ce72df1c 1682
5f8e6c50 1683 * Build devcrypto engine as a dynamic engine.
4098e89c 1684
5f8e6c50 1685 *Eneas U de Queiroz*
4098e89c 1686
5f8e6c50 1687 * Add keyed BLAKE2 to EVP_MAC.
03f8b042 1688
5f8e6c50 1689 *Antoine Salon*
5dcdcd47 1690
5f8e6c50
DMSP
1691 * Fix a bug in the computation of the endpoint-pair shared secret used
1692 by DTLS over SCTP. This breaks interoperability with older versions
1693 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1694 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1695 interoperability with such broken implementations. However, enabling
1696 this switch breaks interoperability with correct implementations.
ae82b46f 1697
5f8e6c50
DMSP
1698 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1699 re-used X509_PUBKEY object if the second PUBKEY is malformed.
8d7ed6ff 1700
5f8e6c50 1701 *Bernd Edlinger*
8d7ed6ff 1702
5f8e6c50 1703 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1b24cca9 1704
5f8e6c50 1705 *Richard Levitte*
9ce5db45 1706
18fdebf1 1707 * Changed the license to the Apache License v2.0.
7f111b8b 1708
5f8e6c50 1709 *Richard Levitte*
651d0aff 1710
5f8e6c50 1711 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
13e91dd3 1712
5f8e6c50
DMSP
1713 - Major releases (indicated by incrementing the MAJOR release number)
1714 may introduce incompatible API/ABI changes.
1715 - Minor releases (indicated by incrementing the MINOR release number)
1716 may introduce new features but retain API/ABI compatibility.
1717 - Patch releases (indicated by incrementing the PATCH number)
1718 are intended for bug fixes and other improvements of existing
1719 features only (like improving performance or adding documentation)
1720 and retain API/ABI compatibility.
13e91dd3 1721
5f8e6c50 1722 *Richard Levitte*
13e91dd3 1723
5f8e6c50 1724 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
13e91dd3 1725
5f8e6c50 1726 *Todd Short*
651d0aff 1727
5f8e6c50
DMSP
1728 * Remove the 'dist' target and add a tarball building script. The
1729 'dist' target has fallen out of use, and it shouldn't be
1730 necessary to configure just to create a source distribution.
651d0aff 1731
5f8e6c50 1732 *Richard Levitte*
651d0aff 1733
5f8e6c50
DMSP
1734 * Recreate the OS390-Unix config target. It no longer relies on a
1735 special script like it did for OpenSSL pre-1.1.0.
651d0aff 1736
5f8e6c50 1737 *Richard Levitte*
651d0aff 1738
5f8e6c50
DMSP
1739 * Instead of having the source directories listed in Configure, add
1740 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1741 look into.
651d0aff 1742
5f8e6c50 1743 *Richard Levitte*
7f111b8b 1744
5f8e6c50 1745 * Add GMAC to EVP_MAC.
1b24cca9 1746
5f8e6c50 1747 *Paul Dale*
651d0aff 1748
5f8e6c50 1749 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
651d0aff 1750
5f8e6c50 1751 *Richard Levitte*
651d0aff 1752
5f8e6c50
DMSP
1753 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1754 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1755 to facilitate the continued use of MACs through raw private keys in
1dc1ea18 1756 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
651d0aff 1757
5f8e6c50 1758 *Richard Levitte*
651d0aff 1759
b7140b06 1760 * Deprecate ECDH_KDF_X9_62().
651d0aff 1761
5f8e6c50 1762 *Antoine Salon*
651d0aff 1763
5f8e6c50
DMSP
1764 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1765 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1766 are retained for backwards compatibility.
651d0aff 1767
5f8e6c50 1768 *Antoine Salon*
651d0aff 1769
5f8e6c50
DMSP
1770 * AES-XTS mode now enforces that its two keys are different to mitigate
1771 the attacked described in "Efficient Instantiations of Tweakable
1772 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1773 Details of this attack can be obtained from:
257e9d03 1774 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
651d0aff 1775
5f8e6c50 1776 *Paul Dale*
651d0aff 1777
5f8e6c50
DMSP
1778 * Rename the object files, i.e. give them other names than in previous
1779 versions. Their names now include the name of the final product, as
1780 well as its type mnemonic (bin, lib, shlib).
651d0aff 1781
5f8e6c50 1782 *Richard Levitte*
651d0aff 1783
5f8e6c50
DMSP
1784 * Added new option for 'openssl list', '-objects', which will display the
1785 list of built in objects, i.e. OIDs with names.
651d0aff 1786
5f8e6c50 1787 *Richard Levitte*
651d0aff 1788
64713cb1
CN
1789 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1790 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1791 be set explicitly.
1792
1793 *Chris Novakovic*
1794
5f8e6c50
DMSP
1795 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1796 improves application performance by removing data copies and providing
1797 applications with zero-copy system calls such as sendfile and splice.
651d0aff 1798
5f8e6c50 1799 *Boris Pismenny*
651d0aff 1800
b7140b06 1801 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
163b8016
ME
1802
1803 *Martin Elshuber*
1804
fc0aae73
DDO
1805 * `PKCS12_parse` now maintains the order of the parsed certificates
1806 when outputting them via `*ca` (rather than reversing it).
1807
1808 *David von Oheimb*
1809
b7140b06 1810 * Deprecated pthread fork support methods.
9750b4d3
RB
1811
1812 *Randall S. Becker*
1813
fc5245a9
HK
1814 * Added support for FFDHE key exchange in TLS 1.3.
1815
1816 *Raja Ashok*
1817
8e7d941a
RL
1818 * Added a new concept for OpenSSL plugability: providers. This
1819 functionality is designed to replace the ENGINE API and ENGINE
1820 implementations, and to be much more dynamic, allowing provider
1821 authors to introduce new algorithms among other things, as long as
1822 there's an API that supports the algorithm type.
1823
1824 With this concept comes a new core API for interaction between
1825 libcrypto and provider implementations. Public libcrypto functions
1826 that want to use providers do so through this core API.
1827
1828 The main documentation for this core API is found in
1829 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1830 refer to other manuals describing the API specific for supported
1831 algorithm types (also called operations).
1832
1833 *The OpenSSL team*
1834
44652c16
DMSP
1835OpenSSL 1.1.1
1836-------------
1837
522a32ef
OP
1838### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
1839
e0d00d79 1840### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
0e4e4e27
RL
1841
1842 * Avoid loading of a dynamic engine twice.
1843
1844 *Bernd Edlinger*
1845
1846 * Prioritise DANE TLSA issuer certs over peer certs
1847
1848 *Viktor Dukhovni*
1849
1850 * Fixed random API for MacOS prior to 10.12
1851
1852 These MacOS versions don't support the CommonCrypto APIs
1853
1854 *Lenny Primak*
1855
796f4f70
MC
1856### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1857
1858 * Fixed an SM2 Decryption Buffer Overflow.
1859
fdd43643
P
1860 In order to decrypt SM2 encrypted data an application is expected to
1861 call the API function EVP_PKEY_decrypt(). Typically an application will
1862 call this function twice. The first time, on entry, the "out" parameter
1863 can be NULL and, on exit, the "outlen" parameter is populated with the
1864 buffer size required to hold the decrypted plaintext. The application
1865 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1866 again, but this time passing a non-NULL value for the "out" parameter.
796f4f70
MC
1867
1868 A bug in the implementation of the SM2 decryption code means that the
fdd43643
P
1869 calculation of the buffer size required to hold the plaintext returned
1870 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1871 size required by the second call. This can lead to a buffer overflow
1872 when EVP_PKEY_decrypt() is called by the application a second time with
1873 a buffer that is too small.
1874
1875 A malicious attacker who is able present SM2 content for decryption to
1876 an application could cause attacker chosen data to overflow the buffer
1877 by up to a maximum of 62 bytes altering the contents of other data held
1878 after the buffer, possibly changing application behaviour or causing
1879 the application to crash. The location of the buffer is application
1880 dependent but is typically heap allocated.
796f4f70
MC
1881 ([CVE-2021-3711])
1882
1883 *Matt Caswell*
1884
fdd43643
P
1885 * Fixed various read buffer overruns processing ASN.1 strings
1886
1887 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1888 structure which contains a buffer holding the string data and a field
1889 holding the buffer length. This contrasts with normal C strings which
1890 are repesented as a buffer for the string data which is terminated
1891 with a NUL (0) byte.
1892
1893 Although not a strict requirement, ASN.1 strings that are parsed using
1894 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1895 well as any string whose value has been set with the ASN1_STRING_set()
1896 function will additionally NUL terminate the byte array in the
1897 ASN1_STRING structure.
1898
1899 However, it is possible for applications to directly construct valid
1900 ASN1_STRING structures which do not NUL terminate the byte array by
1901 directly setting the "data" and "length" fields in the ASN1_STRING
1902 array. This can also happen by using the ASN1_STRING_set0() function.
1903
1904 Numerous OpenSSL functions that print ASN.1 data have been found to
1905 assume that the ASN1_STRING byte array will be NUL terminated, even
1906 though this is not guaranteed for strings that have been directly
1907 constructed. Where an application requests an ASN.1 structure to be
1908 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
1909 been directly constructed by the application without NUL terminating
1910 the "data" field, then a read buffer overrun can occur.
1911
1912 The same thing can also occur during name constraints processing
1913 of certificates (for example if a certificate has been directly
1914 constructed by the application instead of loading it via the OpenSSL
1915 parsing functions, and the certificate contains non NUL terminated
1916 ASN1_STRING structures). It can also occur in the X509_get1_email(),
1917 X509_REQ_get1_email() and X509_get1_ocsp() functions.
1918
1919 If a malicious actor can cause an application to directly construct an
1920 ASN1_STRING and then process it through one of the affected OpenSSL
1921 functions then this issue could be hit. This might result in a crash
1922 (causing a Denial of Service attack). It could also result in the
1923 disclosure of private memory contents (such as private keys, or
1924 sensitive plaintext).
1925 ([CVE-2021-3712])
796f4f70
MC
1926
1927 *Matt Caswell*
1928
1929### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
5b57aa24 1930
468d9d55
MC
1931 * Fixed a problem with verifying a certificate chain when using the
1932 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1933 the certificates present in a certificate chain. It is not set by default.
1934
1935 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1936 the chain that have explicitly encoded elliptic curve parameters was added
1937 as an additional strict check.
1938
1939 An error in the implementation of this check meant that the result of a
1940 previous check to confirm that certificates in the chain are valid CA
1941 certificates was overwritten. This effectively bypasses the check
1942 that non-CA certificates must not be able to issue other certificates.
1943
1944 If a "purpose" has been configured then there is a subsequent opportunity
1945 for checks that the certificate is a valid CA. All of the named "purpose"
1946 values implemented in libcrypto perform this check. Therefore, where
1947 a purpose is set the certificate chain will still be rejected even when the
1948 strict flag has been used. A purpose is set by default in libssl client and
1949 server certificate verification routines, but it can be overridden or
1950 removed by an application.
1951
1952 In order to be affected, an application must explicitly set the
1953 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1954 for the certificate verification or, in the case of TLS client or server
1955 applications, override the default purpose.
1956 ([CVE-2021-3450])
1957
1958 *Tomáš Mráz*
1959
1960 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1961 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1962 renegotiation ClientHello omits the signature_algorithms extension (where it
1963 was present in the initial ClientHello), but includes a
1964 signature_algorithms_cert extension then a NULL pointer dereference will
1965 result, leading to a crash and a denial of service attack.
1966
1967 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1968 (which is the default configuration). OpenSSL TLS clients are not impacted by
1969 this issue.
1970 ([CVE-2021-3449])
1971
1972 *Peter Kästle and Samuel Sapalski*
1973
c913dbd7
MC
1974### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1975
1976 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1977 create a unique hash value based on the issuer and serial number data
1978 contained within an X509 certificate. However it was failing to correctly
1979 handle any errors that may occur while parsing the issuer field (which might
1980 occur if the issuer field is maliciously constructed). This may subsequently
1981 result in a NULL pointer deref and a crash leading to a potential denial of
1982 service attack.
1983 ([CVE-2021-23841])
1984
1985 *Matt Caswell*
1986
1987 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1988 padding mode to correctly check for rollback attacks. This is considered a
1989 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1990 CVE-2021-23839.
1991
1992 *Matt Caswell*
1993
1994 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1995 functions. Previously they could overflow the output length argument in some
1996 cases where the input length is close to the maximum permissable length for
1997 an integer on the platform. In such cases the return value from the function
1998 call would be 1 (indicating success), but the output length value would be
1999 negative. This could cause applications to behave incorrectly or crash.
2000 ([CVE-2021-23840])
2001
2002 *Matt Caswell*
2003
2004 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
5b57aa24
MC
2005 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2006 could be exploited in a side channel attack to recover the password. Since
2007 the attack is local host only this is outside of the current OpenSSL
2008 threat model and therefore no CVE is assigned.
2009
2010 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2011 issue.
2012
2013 *Matt Caswell*
2014
2015### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
6ffc3127 2016
1e13198f
MC
2017 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2018 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2019 If an attacker can control both items being compared then this could lead
2020 to a possible denial of service attack. OpenSSL itself uses the
2021 GENERAL_NAME_cmp function for two purposes:
2022 1) Comparing CRL distribution point names between an available CRL and a
2023 CRL distribution point embedded in an X509 certificate
2024 2) When verifying that a timestamp response token signer matches the
2025 timestamp authority name (exposed via the API functions
2026 TS_RESP_verify_response and TS_RESP_verify_token)
2027 ([CVE-2020-1971])
2028
2029 *Matt Caswell*
6ffc3127
DMSP
2030
2031### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2032
2033 * Certificates with explicit curve parameters are now disallowed in
2034 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2035
66194839 2036 *Tomáš Mráz*
6ffc3127
DMSP
2037
2038 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2039 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2040 conversely, silently ignore DTLS protocol version bounds when configuring
2041 TLS-based contexts. The commands can be repeated to set bounds of both
2042 types. The same applies with the corresponding "min_protocol" and
2043 "max_protocol" command-line switches, in case some application uses both TLS
2044 and DTLS.
2045
2046 SSL_CTX instances that are created for a fixed protocol version (e.g.
2047 TLSv1_server_method()) also silently ignore version bounds. Previously
2048 attempts to apply bounds to these protocol versions would result in an
2049 error. Now only the "version-flexible" SSL_CTX instances are subject to
2050 limits in configuration files in command-line options.
2051
2052 *Viktor Dukhovni*
2053
2054 * Handshake now fails if Extended Master Secret extension is dropped
2055 on renegotiation.
2056
66194839 2057 *Tomáš Mráz*
6ffc3127
DMSP
2058
2059 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2060
2061### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2062
2063 * Fixed segmentation fault in SSL_check_chain()
2064 Server or client applications that call the SSL_check_chain() function
2065 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2066 dereference as a result of incorrect handling of the
2067 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2068 or unrecognised signature algorithm is received from the peer. This could
2069 be exploited by a malicious peer in a Denial of Service attack.
d8dc8538 2070 ([CVE-2020-1967])
6ffc3127
DMSP
2071
2072 *Benjamin Kaduk*
2073
2074 * Added AES consttime code for no-asm configurations
2075 an optional constant time support for AES was added
2076 when building openssl for no-asm.
2077 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2078 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2079 At this time this feature is by default disabled.
2080 It will be enabled by default in 3.0.
2081
2082 *Bernd Edlinger*
2083
2084### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2085
2086 * Revert the change of EOF detection while reading in libssl to avoid
2087 regressions in applications depending on the current way of reporting
2088 the EOF. As the existing method is not fully accurate the change to
2089 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2090 branch and will be present in the 3.0 release.
2091
66194839 2092 *Tomáš Mráz*
6ffc3127
DMSP
2093
2094 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2095 when primes for RSA keys are computed.
2096 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2097 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2098 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
2099 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2100 This avoids possible fingerprinting of newly generated RSA modules.
2101
2102 *Bernd Edlinger*
8658fedd 2103
257e9d03 2104### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
8658fedd
DMSP
2105
2106 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2107 while reading in libssl then we would report an error back to the
2108 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2109 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2110 therefore give a hint as to what went wrong.
2111
2112 *Matt Caswell*
2113
2114 * Check that ed25519 and ed448 are allowed by the security level. Previously
2115 signature algorithms not using an MD were not being checked that they were
2116 allowed by the security level.
2117
2118 *Kurt Roeckx*
2119
2120 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2121 was not quite right. The behaviour was not consistent between resumption
2122 and normal handshakes, and also not quite consistent with historical
2123 behaviour. The behaviour in various scenarios has been clarified and
2124 it has been updated to make it match historical behaviour as closely as
2125 possible.
2126
2127 *Matt Caswell*
44652c16 2128
f33ca114
RL
2129 * *[VMS only]* The header files that the VMS compilers include automatically,
2130 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2131 that the C++ compiler doesn't understand. This is a shortcoming in the
2132 compiler, but can be worked around with `__cplusplus` guards.
2133
2134 C++ applications that use OpenSSL libraries must be compiled using the
2135 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2136 functions. Otherwise, only functions with symbols of less than 31
2137 characters can be used, as the linker will not be able to successfully
2138 resolve symbols with longer names.
2139
2140 *Richard Levitte*
2141
44652c16
DMSP
2142 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2143 The presence of this system service is determined at run-time.
2144
2145 *Richard Levitte*
2146
44652c16
DMSP
2147 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2148 the first value.
2149
2150 *Jon Spillett*
2151
257e9d03 2152### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
44652c16
DMSP
2153
2154 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2155 number generator (RNG). This was intended to include protection in the
2156 event of a fork() system call in order to ensure that the parent and child
2157 processes did not share the same RNG state. However this protection was not
2158 being used in the default case.
2159
2160 A partial mitigation for this issue is that the output from a high
2161 precision timer is mixed into the RNG state so the likelihood of a parent
2162 and child process sharing state is significantly reduced.
2163
2164 If an application already calls OPENSSL_init_crypto() explicitly using
2165 OPENSSL_INIT_ATFORK then this problem does not occur at all.
d8dc8538 2166 ([CVE-2019-1549])
44652c16
DMSP
2167
2168 *Matthias St. Pierre*
2169
2170 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
ece9304c 2171 used even when parsing explicit parameters, when loading a encoded key
44652c16
DMSP
2172 or calling `EC_GROUP_new_from_ecpkparameters()`/
2173 `EC_GROUP_new_from_ecparameters()`.
2174 This prevents bypass of security hardening and performance gains,
2175 especially for curves with specialized EC_METHODs.
2176 By default, if a key encoded with explicit parameters is loaded and later
ece9304c 2177 encoded, the output is still encoded with explicit parameters, even if
44652c16
DMSP
2178 internally a "named" EC_GROUP is used for computation.
2179
2180 *Nicola Tuveri*
2181
2182 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2183 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2184 NULL. After this change, only the cofactor parameter can be NULL. It also
2185 does some minimal sanity checks on the passed order.
d8dc8538 2186 ([CVE-2019-1547])
44652c16
DMSP
2187
2188 *Billy Bob Brumley*
2189
2190 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2191 An attack is simple, if the first CMS_recipientInfo is valid but the
2192 second CMS_recipientInfo is chosen ciphertext. If the second
2193 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2194 encryption key will be replaced by garbage, and the message cannot be
2195 decoded, but if the RSA decryption fails, the correct encryption key is
2196 used and the recipient will not notice the attack.
2197 As a work around for this potential attack the length of the decrypted
2198 key must be equal to the cipher default key length, in case the
2199 certifiate is not given and all recipientInfo are tried out.
2200 The old behaviour can be re-enabled in the CMS code by setting the
2201 CMS_DEBUG_DECRYPT flag.
d8dc8538 2202 ([CVE-2019-1563])
44652c16
DMSP
2203
2204 *Bernd Edlinger*
2205
2206 * Early start up entropy quality from the DEVRANDOM seed source has been
2207 improved for older Linux systems. The RAND subsystem will wait for
2208 /dev/random to be producing output before seeding from /dev/urandom.
2209 The seeded state is stored for future library initialisations using
2210 a system global shared memory segment. The shared memory identifier
2211 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2212 the desired value. The default identifier is 114.
2213
2214 *Paul Dale*
2215
2216 * Correct the extended master secret constant on EBCDIC systems. Without this
2217 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2218 negotiate EMS will fail. Unfortunately this also means that TLS connections
2219 between EBCDIC systems with this fix, and EBCDIC systems without this
2220 fix will fail if they negotiate EMS.
2221
2222 *Matt Caswell*
2223
2224 * Use Windows installation paths in the mingw builds
2225
2226 Mingw isn't a POSIX environment per se, which means that Windows
2227 paths should be used for installation.
d8dc8538 2228 ([CVE-2019-1552])
44652c16
DMSP
2229
2230 *Richard Levitte*
2231
2232 * Changed DH_check to accept parameters with order q and 2q subgroups.
2233 With order 2q subgroups the bit 0 of the private key is not secret
2234 but DH_generate_key works around that by clearing bit 0 of the
2235 private key for those. This avoids leaking bit 0 of the private key.
2236
2237 *Bernd Edlinger*
2238
2239 * Significantly reduce secure memory usage by the randomness pools.
2240
2241 *Paul Dale*
2242
2243 * Revert the DEVRANDOM_WAIT feature for Linux systems
2244
2245 The DEVRANDOM_WAIT feature added a select() call to wait for the
2246 /dev/random device to become readable before reading from the
2247 /dev/urandom device.
2248
2249 It turned out that this change had negative side effects on
2250 performance which were not acceptable. After some discussion it
2251 was decided to revert this feature and leave it up to the OS
2252 resp. the platform maintainer to ensure a proper initialization
2253 during early boot time.
2254
2255 *Matthias St. Pierre*
2256
257e9d03 2257### Changes between 1.1.1b and 1.1.1c [28 May 2019]
44652c16
DMSP
2258
2259 * Add build tests for C++. These are generated files that only do one
2260 thing, to include one public OpenSSL head file each. This tests that
2261 the public header files can be usefully included in a C++ application.
2262
2263 This test isn't enabled by default. It can be enabled with the option
2264 'enable-buildtest-c++'.
2265
2266 *Richard Levitte*
2267
2268 * Enable SHA3 pre-hashing for ECDSA and DSA.
2269
2270 *Patrick Steuer*
2271
2272 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
ec2bfb7d
DDO
2273 This changes the size when using the `genpkey` command when no size is given.
2274 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2275 generation commands to use 2048 bits by default.
44652c16
DMSP
2276
2277 *Kurt Roeckx*
2278
2279 * Reorganize the manual pages to consistently have RETURN VALUES,
2280 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2281 util/fix-doc-nits accordingly.
2282
2283 *Paul Yang, Joshua Lock*
2284
2285 * Add the missing accessor EVP_PKEY_get0_engine()
2286
2287 *Matt Caswell*
2288
ec2bfb7d 2289 * Have commands like `s_client` and `s_server` output the signature scheme
44652c16
DMSP
2290 along with other cipher suite parameters when debugging.
2291
2292 *Lorinczy Zsigmond*
2293
2294 * Make OPENSSL_config() error agnostic again.
2295
2296 *Richard Levitte*
2297
2298 * Do the error handling in RSA decryption constant time.
2299
2300 *Bernd Edlinger*
2301
2302 * Prevent over long nonces in ChaCha20-Poly1305.
2303
2304 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2305 for every encryption operation. RFC 7539 specifies that the nonce value
2306 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2307 and front pads the nonce with 0 bytes if it is less than 12
2308 bytes. However it also incorrectly allows a nonce to be set of up to 16
2309 bytes. In this case only the last 12 bytes are significant and any
2310 additional leading bytes are ignored.
2311
2312 It is a requirement of using this cipher that nonce values are
2313 unique. Messages encrypted using a reused nonce value are susceptible to
2314 serious confidentiality and integrity attacks. If an application changes
2315 the default nonce length to be longer than 12 bytes and then makes a
2316 change to the leading bytes of the nonce expecting the new value to be a
2317 new unique nonce then such an application could inadvertently encrypt
2318 messages with a reused nonce.
2319
2320 Additionally the ignored bytes in a long nonce are not covered by the
2321 integrity guarantee of this cipher. Any application that relies on the
2322 integrity of these ignored leading bytes of a long nonce may be further
2323 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2324 is safe because no such use sets such a long nonce value. However user
2325 applications that use this cipher directly and set a non-default nonce
2326 length to be longer than 12 bytes may be vulnerable.
2327
2328 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2329 Greef of Ronomon.
d8dc8538 2330 ([CVE-2019-1543])
44652c16
DMSP
2331
2332 *Matt Caswell*
2333
2334 * Add DEVRANDOM_WAIT feature for Linux systems
2335
2336 On older Linux systems where the getrandom() system call is not available,
2337 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2338 Contrary to getrandom(), the /dev/urandom device will not block during
2339 early boot when the kernel CSPRNG has not been seeded yet.
2340
2341 To mitigate this known weakness, use select() to wait for /dev/random to
2342 become readable before reading from /dev/urandom.
2343
2344 * Ensure that SM2 only uses SM3 as digest algorithm
2345
2346 *Paul Yang*
2347
257e9d03 2348### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
651d0aff 2349
5f8e6c50
DMSP
2350 * Change the info callback signals for the start and end of a post-handshake
2351 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2352 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2353 confused by this and assume that a TLSv1.2 renegotiation has started. This
2354 can break KeyUpdate handling. Instead we no longer signal the start and end
2355 of a post handshake message exchange (although the messages themselves are
2356 still signalled). This could break some applications that were expecting
2357 the old signals. However without this KeyUpdate is not usable for many
2358 applications.
651d0aff 2359
5f8e6c50 2360 *Matt Caswell*
651d0aff 2361
257e9d03 2362### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
651d0aff 2363
5f8e6c50 2364 * Timing vulnerability in DSA signature generation
651d0aff 2365
5f8e6c50
DMSP
2366 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2367 timing side channel attack. An attacker could use variations in the signing
2368 algorithm to recover the private key.
651d0aff 2369
5f8e6c50 2370 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
d8dc8538 2371 ([CVE-2018-0734])
651d0aff 2372
5f8e6c50 2373 *Paul Dale*
651d0aff 2374
5f8e6c50 2375 * Timing vulnerability in ECDSA signature generation
651d0aff 2376
5f8e6c50
DMSP
2377 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2378 timing side channel attack. An attacker could use variations in the signing
2379 algorithm to recover the private key.
651d0aff 2380
5f8e6c50 2381 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
d8dc8538 2382 ([CVE-2018-0735])
651d0aff 2383
5f8e6c50 2384 *Paul Dale*
651d0aff 2385
5f8e6c50
DMSP
2386 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2387 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2388 of two gigabytes and the error handling improved.
651d0aff 2389
5f8e6c50
DMSP
2390 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2391 categorized as a normal bug, not a security issue, because the DRBG reseeds
2392 automatically and is fully functional even without additional randomness
2393 provided by the application.
2394
257e9d03 2395### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
5f8e6c50
DMSP
2396
2397 * Add a new ClientHello callback. Provides a callback interface that gives
2398 the application the ability to adjust the nascent SSL object at the
2399 earliest stage of ClientHello processing, immediately after extensions have
2400 been collected but before they have been processed. In particular, this
2401 callback can adjust the supported TLS versions in response to the contents
2402 of the ClientHello
2403
2404 *Benjamin Kaduk*
2405
2406 * Add SM2 base algorithm support.
2407
2408 *Jack Lloyd*
2409
2410 * s390x assembly pack: add (improved) hardware-support for the following
2411 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2412 aes-cfb/cfb8, aes-ecb.
2413
2414 *Patrick Steuer*
2415
2416 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2417 parameter is no longer accepted, as it leads to a corrupt table. NULL
2418 pem_str is reserved for alias entries only.
2419
2420 *Richard Levitte*
2421
2422 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2423 step for prime curves. The new implementation is based on formulae from
2424 differential addition-and-doubling in homogeneous projective coordinates
2425 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2426 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2427 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2428 to work in projective coordinates.
2429
2430 *Billy Bob Brumley, Nicola Tuveri*
2431
2432 * Change generating and checking of primes so that the error rate of not
2433 being prime depends on the intended use based on the size of the input.
2434 For larger primes this will result in more rounds of Miller-Rabin.
2435 The maximal error rate for primes with more than 1080 bits is lowered
2436 to 2^-128.
2437
2438 *Kurt Roeckx, Annie Yousar*
2439
2440 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2441
2442 *Kurt Roeckx*
2443
2444 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2445 moving between systems, and to avoid confusion when a Windows build is
2446 done with mingw vs with MSVC. For POSIX installs, there's still a
2447 symlink or copy named 'tsget' to avoid that confusion as well.
2448
2449 *Richard Levitte*
2450
2451 * Revert blinding in ECDSA sign and instead make problematic addition
2452 length-invariant. Switch even to fixed-length Montgomery multiplication.
2453
2454 *Andy Polyakov*
2455
2456 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2457 step for binary curves. The new implementation is based on formulae from
2458 differential addition-and-doubling in mixed Lopez-Dahab projective
2459 coordinates, modified to independently blind the operands.
2460
2461 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2462
2463 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2464 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2465 EC_METHODs to implement their own specialized "ladder step", to take
2466 advantage of more favorable coordinate systems or more efficient
2467 differential addition-and-doubling algorithms.
2468
2469 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2470
2471 * Modified the random device based seed sources to keep the relevant
2472 file descriptors open rather than reopening them on each access.
2473 This allows such sources to operate in a chroot() jail without
2474 the associated device nodes being available. This behaviour can be
2475 controlled using RAND_keep_random_devices_open().
2476
2477 *Paul Dale*
2478
2479 * Numerous side-channel attack mitigations have been applied. This may have
2480 performance impacts for some algorithms for the benefit of improved
2481 security. Specific changes are noted in this change log by their respective
2482 authors.
2483
2484 *Matt Caswell*
2485
2486 * AIX shared library support overhaul. Switch to AIX "natural" way of
2487 handling shared libraries, which means collecting shared objects of
2488 different versions and bitnesses in one common archive. This allows to
2489 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2490 doesn't affect the way 3rd party applications are linked, only how
2491 multi-version installation is managed.
2492
2493 *Andy Polyakov*
2494
2495 * Make ec_group_do_inverse_ord() more robust and available to other
2496 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2497 mitigations are applied to the fallback BN_mod_inverse().
2498 When using this function rather than BN_mod_inverse() directly, new
2499 EC cryptosystem implementations are then safer-by-default.
2500
2501 *Billy Bob Brumley*
2502
2503 * Add coordinate blinding for EC_POINT and implement projective
2504 coordinate blinding for generic prime curves as a countermeasure to
2505 chosen point SCA attacks.
2506
2507 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2508
2509 * Add blinding to ECDSA and DSA signatures to protect against side channel
2510 attacks discovered by Keegan Ryan (NCC Group).
2511
2512 *Matt Caswell*
2513
ec2bfb7d 2514 * Enforce checking in the `pkeyutl` command to ensure that the input
5f8e6c50
DMSP
2515 length does not exceed the maximum supported digest length when performing
2516 a sign, verify or verifyrecover operation.
2517
2518 *Matt Caswell*
2519
2520 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2521 I/O in combination with something like select() or poll() will hang. This
2522 can be turned off again using SSL_CTX_clear_mode().
2523 Many applications do not properly handle non-application data records, and
2524 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2525 around the problems in those applications, but can also break some.
2526 It's recommended to read the manpages about SSL_read(), SSL_write(),
2527 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2528 SSL_CTX_set_read_ahead() again.
2529
2530 *Kurt Roeckx*
2531
2532 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2533 now allow empty (zero character) pass phrases.
2534
2535 *Richard Levitte*
2536
2537 * Apply blinding to binary field modular inversion and remove patent
2538 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2539
2540 *Billy Bob Brumley*
2541
2542 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2543 binary and prime elliptic curves.
2544
2545 *Billy Bob Brumley*
2546
2547 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2548 constant time fixed point multiplication.
2549
2550 *Billy Bob Brumley*
2551
2552 * Revise elliptic curve scalar multiplication with timing attack
2553 defenses: ec_wNAF_mul redirects to a constant time implementation
2554 when computing fixed point and variable point multiplication (which
2555 in OpenSSL are mostly used with secret scalars in keygen, sign,
2556 ECDH derive operations).
2557 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2558 Sohaib ul Hassan*
2559
2560 * Updated CONTRIBUTING
2561
2562 *Rich Salz*
2563
2564 * Updated DRBG / RAND to request nonce and additional low entropy
2565 randomness from the system.
2566
2567 *Matthias St. Pierre*
2568
2569 * Updated 'openssl rehash' to use OpenSSL consistent default.
2570
2571 *Richard Levitte*
2572
2573 * Moved the load of the ssl_conf module to libcrypto, which helps
2574 loading engines that libssl uses before libssl is initialised.
2575
2576 *Matt Caswell*
2577
2578 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2579
2580 *Matt Caswell*
2581
2582 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2583
2584 *Ingo Schwarze, Rich Salz*
2585
2586 * Added output of accepting IP address and port for 'openssl s_server'
2587
2588 *Richard Levitte*
2589
2590 * Added a new API for TLSv1.3 ciphersuites:
2591 SSL_CTX_set_ciphersuites()
2592 SSL_set_ciphersuites()
2593
2594 *Matt Caswell*
2595
2596 * Memory allocation failures consistently add an error to the error
2597 stack.
2598
2599 *Rich Salz*
2600
2601 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2602 in libcrypto when run as setuid/setgid.
2603
2604 *Bernd Edlinger*
2605
2606 * Load any config file by default when libssl is used.
2607
2608 *Matt Caswell*
2609
2610 * Added new public header file <openssl/rand_drbg.h> and documentation
2611 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2612
2613 *Matthias St. Pierre*
2614
2615 * QNX support removed (cannot find contributors to get their approval
2616 for the license change).
2617
2618 *Rich Salz*
2619
2620 * TLSv1.3 replay protection for early data has been implemented. See the
2621 SSL_read_early_data() man page for further details.
2622
2623 *Matt Caswell*
2624
2625 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2626 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2627 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2628 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2629 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2630 configuration has been separated out. See the ciphers man page or the
2631 SSL_CTX_set_ciphersuites() man page for more information.
2632
2633 *Matt Caswell*
2634
2635 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2636 in responder mode now supports the new "-multi" option, which
2637 spawns the specified number of child processes to handle OCSP
2638 requests. The "-timeout" option now also limits the OCSP
2639 responder's patience to wait to receive the full client request
2640 on a newly accepted connection. Child processes are respawned
2641 as needed, and the CA index file is automatically reloaded
2642 when changed. This makes it possible to run the "ocsp" responder
2643 as a long-running service, making the OpenSSL CA somewhat more
2644 feature-complete. In this mode, most diagnostic messages logged
2645 after entering the event loop are logged via syslog(3) rather than
2646 written to stderr.
2647
2648 *Viktor Dukhovni*
2649
2650 * Added support for X448 and Ed448. Heavily based on original work by
2651 Mike Hamburg.
2652
2653 *Matt Caswell*
2654
2655 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2656 objects loaded. This adds the functions OSSL_STORE_expect() and
2657 OSSL_STORE_find() as well as needed tools to construct searches and
2658 get the search data out of them.
2659
2660 *Richard Levitte*
2661
2662 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2663 version of OpenSSL should review their configuration settings to ensure
2664 that they are still appropriate for TLSv1.3. For further information see:
257e9d03 2665 <https://wiki.openssl.org/index.php/TLS1.3>
5f8e6c50
DMSP
2666
2667 *Matt Caswell*
2668
2669 * Grand redesign of the OpenSSL random generator
2670
2671 The default RAND method now utilizes an AES-CTR DRBG according to
2672 NIST standard SP 800-90Ar1. The new random generator is essentially
2673 a port of the default random generator from the OpenSSL FIPS 2.0
2674 object module. It is a hybrid deterministic random bit generator
2675 using an AES-CTR bit stream and which seeds and reseeds itself
2676 automatically using trusted system entropy sources.
2677
2678 Some of its new features are:
2679 - Support for multiple DRBG instances with seed chaining.
2680 - The default RAND method makes use of a DRBG.
2681 - There is a public and private DRBG instance.
2682 - The DRBG instances are fork-safe.
2683 - Keep all global DRBG instances on the secure heap if it is enabled.
2684 - The public and private DRBG instance are per thread for lock free
2685 operation
2686
2687 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2688
2689 * Changed Configure so it only says what it does and doesn't dump
2690 so much data. Instead, ./configdata.pm should be used as a script
2691 to display all sorts of configuration data.
2692
2693 *Richard Levitte*
2694
2695 * Added processing of "make variables" to Configure.
2696
2697 *Richard Levitte*
2698
2699 * Added SHA512/224 and SHA512/256 algorithm support.
2700
2701 *Paul Dale*
2702
2703 * The last traces of Netware support, first removed in 1.1.0, have
2704 now been removed.
2705
2706 *Rich Salz*
2707
2708 * Get rid of Makefile.shared, and in the process, make the processing
2709 of certain files (rc.obj, or the .def/.map/.opt files produced from
2710 the ordinal files) more visible and hopefully easier to trace and
2711 debug (or make silent).
2712
2713 *Richard Levitte*
2714
2715 * Make it possible to have environment variable assignments as
2716 arguments to config / Configure.
2717
2718 *Richard Levitte*
2719
2720 * Add multi-prime RSA (RFC 8017) support.
2721
2722 *Paul Yang*
2723
2724 * Add SM3 implemented according to GB/T 32905-2016
1dc1ea18
DDO
2725 *Jack Lloyd <jack.lloyd@ribose.com>,*
2726 *Ronald Tse <ronald.tse@ribose.com>,*
2727 *Erick Borsboom <erick.borsboom@ribose.com>*
5f8e6c50
DMSP
2728
2729 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2730 as documented in RFC6066.
2731 Based on a patch from Tomasz Moń
2732
2733 *Filipe Raimundo da Silva*
2734
2735 * Add SM4 implemented according to GB/T 32907-2016.
1dc1ea18
DDO
2736 *Jack Lloyd <jack.lloyd@ribose.com>,*
2737 *Ronald Tse <ronald.tse@ribose.com>,*
2738 *Erick Borsboom <erick.borsboom@ribose.com>*
5f8e6c50
DMSP
2739
2740 * Reimplement -newreq-nodes and ERR_error_string_n; the
2741 original author does not agree with the license change.
2742
2743 *Rich Salz*
2744
2745 * Add ARIA AEAD TLS support.
2746
2747 *Jon Spillett*
2748
2749 * Some macro definitions to support VS6 have been removed. Visual
2750 Studio 6 has not worked since 1.1.0
2751
2752 *Rich Salz*
2753
2754 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2755 without clearing the errors.
2756
2757 *Richard Levitte*
2758
2759 * Add "atfork" functions. If building on a system that without
2760 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2761 requirements. The RAND facility now uses/requires this.
2762
2763 *Rich Salz*
2764
2765 * Add SHA3.
2766
2767 *Andy Polyakov*
2768
2769 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2770 not possible to disable entirely. However, it's still possible to
2771 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2772 as a fallback).
2773
2774 To disable, configure with 'no-ui-console'. 'no-ui' is still
2775 possible to use as an alias. Check at compile time with the
2776 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2777 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2778
2779 *Richard Levitte*
2780
2781 * Add a STORE module, which implements a uniform and URI based reader of
2782 stores that can contain keys, certificates, CRLs and numerous other
2783 objects. The main API is loosely based on a few stdio functions,
2784 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2785 OSSL_STORE_error and OSSL_STORE_close.
2786 The implementation uses backends called "loaders" to implement arbitrary
2787 URI schemes. There is one built in "loader" for the 'file' scheme.
2788
2789 *Richard Levitte*
2790
2791 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2792 then adjusted to work on FreeBSD 8.4 as well.
2793 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2794 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2795
2796 *Richard Levitte*
2797
2798 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2799 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2800 error code calls like this:
2801
2802 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2803
2804 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2805 that can be encoded in C. For the foreseeable future, this will only
2806 affect new modules.
2807
2808 *Richard Levitte and Tim Hudson*
2809
2810 * Removed BSD cryptodev engine.
2811
2812 *Rich Salz*
2813
2814 * Add a build target 'build_all_generated', to build all generated files
2815 and only that. This can be used to prepare everything that requires
2816 things like perl for a system that lacks perl and then move everything
2817 to that system and do the rest of the build there.
2818
2819 *Richard Levitte*
2820
2821 * In the UI interface, make it possible to duplicate the user data. This
2822 can be used by engines that need to retain the data for a longer time
2823 than just the call where this user data is passed.
2824
2825 *Richard Levitte*
2826
2827 * Ignore the '-named_curve auto' value for compatibility of applications
2828 with OpenSSL 1.0.2.
2829
66194839 2830 *Tomáš Mráz <tmraz@fedoraproject.org>*
5f8e6c50
DMSP
2831
2832 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2833 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2834 alerts across multiple records (some of which could be empty). In practice
2835 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
44652c16 2836 prohibits this altogether and other libraries (BoringSSL, NSS) do not
5f8e6c50 2837 support this at all. Supporting it adds significant complexity to the
44652c16 2838 record layer, and its removal is unlikely to cause interoperability
5f8e6c50
DMSP
2839 issues.
2840
2841 *Matt Caswell*
2842
2843 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2844 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2845 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2846 in OpenSSL 1.2.0.
2847
2848 *Richard Levitte*
2849
2850 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2851 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2852
2853 *Richard Levitte, Andy Polyakov*
2854
2855 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2856 does for RSA, etc.
2857
2858 *Richard Levitte*
2859
2860 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2861 platform rather than 'mingw'.
2862
2863 *Richard Levitte*
2864
2865 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2866 success if they are asked to add an object which already exists
2867 in the store. This change cascades to other functions which load
2868 certificates and CRLs.
2869
2870 *Paul Dale*
2871
2872 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2873 facilitate stack unwinding even from assembly subroutines.
2874
2875 *Andy Polyakov*
2876
2877 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2878 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2879
2880 *Richard Levitte*
2881
2882 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2883 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2884 which is the minimum version we support.
2885
2886 *Richard Levitte*
2887
2888 * Certificate time validation (X509_cmp_time) enforces stricter
2889 compliance with RFC 5280. Fractional seconds and timezone offsets
2890 are no longer allowed.
2891
2892 *Emilia Käsper*
2893
2894 * Add support for ARIA
2895
2896 *Paul Dale*
2897
2898 * s_client will now send the Server Name Indication (SNI) extension by
2899 default unless the new "-noservername" option is used. The server name is
2900 based on the host provided to the "-connect" option unless overridden by
2901 using "-servername".
2902
2903 *Matt Caswell*
2904
2905 * Add support for SipHash
2906
2907 *Todd Short*
2908
2909 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2910 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2911 prevent issues where no progress is being made and the peer continually
2912 sends unrecognised record types, using up resources processing them.
2913
2914 *Matt Caswell*
2915
2916 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2917 using the algorithm defined in
257e9d03 2918 <https://www.akkadia.org/drepper/SHA-crypt.txt>
5f8e6c50
DMSP
2919
2920 *Richard Levitte*
2921
2922 * Heartbeat support has been removed; the ABI is changed for now.
2923
2924 *Richard Levitte, Rich Salz*
2925
2926 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2927
2928 *Emilia Käsper*
2929
2930 * The RSA "null" method, which was partially supported to avoid patent
2931 issues, has been replaced to always returns NULL.
2932
2933 *Rich Salz*
2934
44652c16
DMSP
2935OpenSSL 1.1.0
2936-------------
5f8e6c50 2937
257e9d03 2938### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
5f8e6c50 2939
44652c16 2940 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
ece9304c 2941 used even when parsing explicit parameters, when loading a encoded key
44652c16
DMSP
2942 or calling `EC_GROUP_new_from_ecpkparameters()`/
2943 `EC_GROUP_new_from_ecparameters()`.
2944 This prevents bypass of security hardening and performance gains,
2945 especially for curves with specialized EC_METHODs.
2946 By default, if a key encoded with explicit parameters is loaded and later
ece9304c 2947 encoded, the output is still encoded with explicit parameters, even if
44652c16 2948 internally a "named" EC_GROUP is used for computation.
5f8e6c50 2949
44652c16 2950 *Nicola Tuveri*
5f8e6c50 2951
44652c16
DMSP
2952 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2953 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2954 NULL. After this change, only the cofactor parameter can be NULL. It also
2955 does some minimal sanity checks on the passed order.
d8dc8538 2956 ([CVE-2019-1547])
5f8e6c50 2957
44652c16 2958 *Billy Bob Brumley*
5f8e6c50 2959
44652c16
DMSP
2960 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2961 An attack is simple, if the first CMS_recipientInfo is valid but the
2962 second CMS_recipientInfo is chosen ciphertext. If the second
2963 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2964 encryption key will be replaced by garbage, and the message cannot be
2965 decoded, but if the RSA decryption fails, the correct encryption key is
2966 used and the recipient will not notice the attack.
2967 As a work around for this potential attack the length of the decrypted
2968 key must be equal to the cipher default key length, in case the
2969 certifiate is not given and all recipientInfo are tried out.
2970 The old behaviour can be re-enabled in the CMS code by setting the
2971 CMS_DEBUG_DECRYPT flag.
d8dc8538 2972 ([CVE-2019-1563])
44652c16
DMSP
2973
2974 *Bernd Edlinger*
2975
2976 * Use Windows installation paths in the mingw builds
2977
2978 Mingw isn't a POSIX environment per se, which means that Windows
2979 paths should be used for installation.
d8dc8538 2980 ([CVE-2019-1552])
44652c16
DMSP
2981
2982 *Richard Levitte*
2983
257e9d03 2984### Changes between 1.1.0j and 1.1.0k [28 May 2019]
44652c16
DMSP
2985
2986 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
ec2bfb7d
DDO
2987 This changes the size when using the `genpkey` command when no size is given.
2988 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2989 generation commands to use 2048 bits by default.
44652c16
DMSP
2990
2991 *Kurt Roeckx*
2992
2993 * Prevent over long nonces in ChaCha20-Poly1305.
2994
2995 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2996 for every encryption operation. RFC 7539 specifies that the nonce value
2997 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2998 and front pads the nonce with 0 bytes if it is less than 12
2999 bytes. However it also incorrectly allows a nonce to be set of up to 16
3000 bytes. In this case only the last 12 bytes are significant and any
3001 additional leading bytes are ignored.
3002
3003 It is a requirement of using this cipher that nonce values are
3004 unique. Messages encrypted using a reused nonce value are susceptible to
3005 serious confidentiality and integrity attacks. If an application changes
3006 the default nonce length to be longer than 12 bytes and then makes a
3007 change to the leading bytes of the nonce expecting the new value to be a
3008 new unique nonce then such an application could inadvertently encrypt
3009 messages with a reused nonce.
3010
3011 Additionally the ignored bytes in a long nonce are not covered by the
3012 integrity guarantee of this cipher. Any application that relies on the
3013 integrity of these ignored leading bytes of a long nonce may be further
3014 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3015 is safe because no such use sets such a long nonce value. However user
3016 applications that use this cipher directly and set a non-default nonce
3017 length to be longer than 12 bytes may be vulnerable.
3018
3019 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3020 Greef of Ronomon.
d8dc8538 3021 ([CVE-2019-1543])
44652c16
DMSP
3022
3023 *Matt Caswell*
3024
3025 * Added SCA hardening for modular field inversion in EC_GROUP through
3026 a new dedicated field_inv() pointer in EC_METHOD.
3027 This also addresses a leakage affecting conversions from projective
3028 to affine coordinates.
3029
3030 *Billy Bob Brumley, Nicola Tuveri*
3031
3032 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3033 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3034
3035 *Bernd Edlinger*
3036
3037 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3038
3039 *Richard Levitte*
3040
3041 * Remove the 'dist' target and add a tarball building script. The
3042 'dist' target has fallen out of use, and it shouldn't be
3043 necessary to configure just to create a source distribution.
3044
3045 *Richard Levitte*
3046
257e9d03 3047### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
44652c16
DMSP
3048
3049 * Timing vulnerability in DSA signature generation
3050
3051 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3052 timing side channel attack. An attacker could use variations in the signing
3053 algorithm to recover the private key.
3054
3055 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
d8dc8538 3056 ([CVE-2018-0734])
44652c16
DMSP
3057
3058 *Paul Dale*
3059
3060 * Timing vulnerability in ECDSA signature generation
3061
3062 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3063 timing side channel attack. An attacker could use variations in the signing
3064 algorithm to recover the private key.
3065
3066 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
d8dc8538 3067 ([CVE-2018-0735])
44652c16
DMSP
3068
3069 *Paul Dale*
3070
3071 * Add coordinate blinding for EC_POINT and implement projective
3072 coordinate blinding for generic prime curves as a countermeasure to
3073 chosen point SCA attacks.
3074
3075 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3076
257e9d03 3077### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
44652c16
DMSP
3078
3079 * Client DoS due to large DH parameter
3080
3081 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3082 malicious server can send a very large prime value to the client. This will
3083 cause the client to spend an unreasonably long period of time generating a
3084 key for this prime resulting in a hang until the client has finished. This
3085 could be exploited in a Denial Of Service attack.
3086
3087 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
d8dc8538 3088 ([CVE-2018-0732])
44652c16
DMSP
3089
3090 *Guido Vranken*
3091
3092 * Cache timing vulnerability in RSA Key Generation
3093
3094 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3095 a cache timing side channel attack. An attacker with sufficient access to
3096 mount cache timing attacks during the RSA key generation process could
3097 recover the private key.
5f8e6c50
DMSP
3098
3099 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3100 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
d8dc8538 3101 ([CVE-2018-0737])
5f8e6c50
DMSP
3102
3103 *Billy Brumley*
3104
3105 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3106 parameter is no longer accepted, as it leads to a corrupt table. NULL
3107 pem_str is reserved for alias entries only.
3108
3109 *Richard Levitte*
3110
3111 * Revert blinding in ECDSA sign and instead make problematic addition
3112 length-invariant. Switch even to fixed-length Montgomery multiplication.
3113
3114 *Andy Polyakov*
3115
3116 * Change generating and checking of primes so that the error rate of not
3117 being prime depends on the intended use based on the size of the input.
3118 For larger primes this will result in more rounds of Miller-Rabin.
3119 The maximal error rate for primes with more than 1080 bits is lowered
3120 to 2^-128.
3121
3122 *Kurt Roeckx, Annie Yousar*
3123
3124 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3125
3126 *Kurt Roeckx*
3127
3128 * Add blinding to ECDSA and DSA signatures to protect against side channel
3129 attacks discovered by Keegan Ryan (NCC Group).
3130
3131 *Matt Caswell*
3132
3133 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3134 now allow empty (zero character) pass phrases.
3135
3136 *Richard Levitte*
3137
3138 * Certificate time validation (X509_cmp_time) enforces stricter
3139 compliance with RFC 5280. Fractional seconds and timezone offsets
3140 are no longer allowed.
3141
3142 *Emilia Käsper*
3143
3144 * Fixed a text canonicalisation bug in CMS
3145
3146 Where a CMS detached signature is used with text content the text goes
3147 through a canonicalisation process first prior to signing or verifying a
3148 signature. This process strips trailing space at the end of lines, converts
3149 line terminators to CRLF and removes additional trailing line terminators
3150 at the end of a file. A bug in the canonicalisation process meant that
3151 some characters, such as form-feed, were incorrectly treated as whitespace
3152 and removed. This is contrary to the specification (RFC5485). This fix
3153 could mean that detached text data signed with an earlier version of
3154 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3155 signed with a fixed OpenSSL may fail to verify with an earlier version of
3156 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3157 and use the "-binary" flag (for the "cms" command line application) or set
3158 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3159
3160 *Matt Caswell*
3161
257e9d03 3162### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
5f8e6c50
DMSP
3163
3164 * Constructed ASN.1 types with a recursive definition could exceed the stack
3165
3166 Constructed ASN.1 types with a recursive definition (such as can be found
3167 in PKCS7) could eventually exceed the stack given malicious input with
3168 excessive recursion. This could result in a Denial Of Service attack. There
3169 are no such structures used within SSL/TLS that come from untrusted sources
3170 so this is considered safe.
3171
3172 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3173 project.
d8dc8538 3174 ([CVE-2018-0739])
5f8e6c50
DMSP
3175
3176 *Matt Caswell*
3177
3178 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3179
3180 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3181 effectively reduced to only comparing the least significant bit of each
3182 byte. This allows an attacker to forge messages that would be considered as
3183 authenticated in an amount of tries lower than that guaranteed by the
3184 security claims of the scheme. The module can only be compiled by the
3185 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3186
3187 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3188 (IBM).
d8dc8538 3189 ([CVE-2018-0733])
5f8e6c50
DMSP
3190
3191 *Andy Polyakov*
3192
3193 * Add a build target 'build_all_generated', to build all generated files
3194 and only that. This can be used to prepare everything that requires
3195 things like perl for a system that lacks perl and then move everything
3196 to that system and do the rest of the build there.
3197
3198 *Richard Levitte*
3199
3200 * Backport SSL_OP_NO_RENGOTIATION
3201
3202 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3203 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3204 changes this is no longer possible in 1.1.0. Therefore the new
3205 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3206 1.1.0 to provide equivalent functionality.
3207
3208 Note that if an application built against 1.1.0h headers (or above) is run
3209 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3210 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3211
3212 *Matt Caswell*
3213
3214 * Removed the OS390-Unix config target. It relied on a script that doesn't
3215 exist.
3216
3217 *Rich Salz*
3218
3219 * rsaz_1024_mul_avx2 overflow bug on x86_64
3220
3221 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3222 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3223 Analysis suggests that attacks against RSA and DSA as a result of this
3224 defect would be very difficult to perform and are not believed likely.
3225 Attacks against DH1024 are considered just feasible, because most of the
3226 work necessary to deduce information about a private key may be performed
3227 offline. The amount of resources required for such an attack would be
3228 significant. However, for an attack on TLS to be meaningful, the server
3229 would have to share the DH1024 private key among multiple clients, which is
3230 no longer an option since CVE-2016-0701.
3231
3232 This only affects processors that support the AVX2 but not ADX extensions
3233 like Intel Haswell (4th generation).
3234
3235 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3236 was originally found via the OSS-Fuzz project.
d8dc8538 3237 ([CVE-2017-3738])
5f8e6c50
DMSP
3238
3239 *Andy Polyakov*
3240
257e9d03 3241### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
5f8e6c50
DMSP
3242
3243 * bn_sqrx8x_internal carry bug on x86_64
3244
3245 There is a carry propagating bug in the x86_64 Montgomery squaring
3246 procedure. No EC algorithms are affected. Analysis suggests that attacks
3247 against RSA and DSA as a result of this defect would be very difficult to
3248 perform and are not believed likely. Attacks against DH are considered just
3249 feasible (although very difficult) because most of the work necessary to
3250 deduce information about a private key may be performed offline. The amount
3251 of resources required for such an attack would be very significant and
3252 likely only accessible to a limited number of attackers. An attacker would
3253 additionally need online access to an unpatched system using the target
3254 private key in a scenario with persistent DH parameters and a private
3255 key that is shared between multiple clients.
3256
3257 This only affects processors that support the BMI1, BMI2 and ADX extensions
3258 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3259
3260 This issue was reported to OpenSSL by the OSS-Fuzz project.
d8dc8538 3261 ([CVE-2017-3736])
5f8e6c50
DMSP
3262
3263 *Andy Polyakov*
3264
3265 * Malformed X.509 IPAddressFamily could cause OOB read
3266
3267 If an X.509 certificate has a malformed IPAddressFamily extension,
3268 OpenSSL could do a one-byte buffer overread. The most likely result
3269 would be an erroneous display of the certificate in text format.
3270
3271 This issue was reported to OpenSSL by the OSS-Fuzz project.
d8dc8538 3272 ([CVE-2017-3735])
5f8e6c50
DMSP
3273
3274 *Rich Salz*
3275
257e9d03 3276### Changes between 1.1.0e and 1.1.0f [25 May 2017]
5f8e6c50
DMSP
3277
3278 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3279 platform rather than 'mingw'.
3280
3281 *Richard Levitte*
3282
3283 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3284 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3285 which is the minimum version we support.
3286
3287 *Richard Levitte*
3288
257e9d03 3289### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
5f8e6c50
DMSP
3290
3291 * Encrypt-Then-Mac renegotiation crash
3292
3293 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3294 negotiated where it was not in the original handshake (or vice-versa) then
3295 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3296 and servers are affected.
3297
3298 This issue was reported to OpenSSL by Joe Orton (Red Hat).
d8dc8538 3299 ([CVE-2017-3733])
5f8e6c50
DMSP
3300
3301 *Matt Caswell*
3302
257e9d03 3303### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
5f8e6c50
DMSP
3304
3305 * Truncated packet could crash via OOB read
3306
3307 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3308 cipher is being used, then a truncated packet can cause that host to
3309 perform an out-of-bounds read, usually resulting in a crash.
3310
3311 This issue was reported to OpenSSL by Robert Święcki of Google.
d8dc8538 3312 ([CVE-2017-3731])
5f8e6c50
DMSP
3313
3314 *Andy Polyakov*
3315
3316 * Bad (EC)DHE parameters cause a client crash
3317
3318 If a malicious server supplies bad parameters for a DHE or ECDHE key
3319 exchange then this can result in the client attempting to dereference a
3320 NULL pointer leading to a client crash. This could be exploited in a Denial
3321 of Service attack.
3322
3323 This issue was reported to OpenSSL by Guido Vranken.
d8dc8538 3324 ([CVE-2017-3730])
5f8e6c50
DMSP
3325
3326 *Matt Caswell*
3327
3328 * BN_mod_exp may produce incorrect results on x86_64
3329
3330 There is a carry propagating bug in the x86_64 Montgomery squaring
3331 procedure. No EC algorithms are affected. Analysis suggests that attacks
3332 against RSA and DSA as a result of this defect would be very difficult to
3333 perform and are not believed likely. Attacks against DH are considered just
3334 feasible (although very difficult) because most of the work necessary to
3335 deduce information about a private key may be performed offline. The amount
3336 of resources required for such an attack would be very significant and
3337 likely only accessible to a limited number of attackers. An attacker would
3338 additionally need online access to an unpatched system using the target
3339 private key in a scenario with persistent DH parameters and a private
3340 key that is shared between multiple clients. For example this can occur by
3341 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3342 similar to CVE-2015-3193 but must be treated as a separate problem.
3343
3344 This issue was reported to OpenSSL by the OSS-Fuzz project.
d8dc8538 3345 ([CVE-2017-3732])
5f8e6c50
DMSP
3346
3347 *Andy Polyakov*
3348
257e9d03 3349### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
5f8e6c50
DMSP
3350
3351 * ChaCha20/Poly1305 heap-buffer-overflow
3352
257e9d03 3353 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
5f8e6c50
DMSP
3354 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3355 crash. This issue is not considered to be exploitable beyond a DoS.
3356
3357 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
d8dc8538 3358 ([CVE-2016-7054])
5f8e6c50
DMSP
3359
3360 *Richard Levitte*
3361
3362 * CMS Null dereference
3363
3364 Applications parsing invalid CMS structures can crash with a NULL pointer
3365 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3366 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3367 structure callback if an attempt is made to free certain invalid encodings.
3368 Only CHOICE structures using a callback which do not handle NULL value are
3369 affected.
3370
3371 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
d8dc8538 3372 ([CVE-2016-7053])
5f8e6c50
DMSP
3373
3374 *Stephen Henson*
3375
3376 * Montgomery multiplication may produce incorrect results
3377
3378 There is a carry propagating bug in the Broadwell-specific Montgomery
3379 multiplication procedure that handles input lengths divisible by, but
3380 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3381 and DH private keys are impossible. This is because the subroutine in
3382 question is not used in operations with the private key itself and an input
3383 of the attacker's direct choice. Otherwise the bug can manifest itself as
3384 transient authentication and key negotiation failures or reproducible
3385 erroneous outcome of public-key operations with specially crafted input.
3386 Among EC algorithms only Brainpool P-512 curves are affected and one
3387 presumably can attack ECDH key negotiation. Impact was not analyzed in
3388 detail, because pre-requisites for attack are considered unlikely. Namely
3389 multiple clients have to choose the curve in question and the server has to
3390 share the private key among them, neither of which is default behaviour.
3391 Even then only clients that chose the curve will be affected.
3392
3393 This issue was publicly reported as transient failures and was not
3394 initially recognized as a security issue. Thanks to Richard Morgan for
3395 providing reproducible case.
d8dc8538 3396 ([CVE-2016-7055])
5f8e6c50
DMSP
3397
3398 *Andy Polyakov*
3399
3400 * Removed automatic addition of RPATH in shared libraries and executables,
3401 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3402
3403 *Richard Levitte*
3404
257e9d03 3405### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
5f8e6c50
DMSP
3406
3407 * Fix Use After Free for large message sizes
3408
3409 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3410 message larger than approx 16k is received then the underlying buffer to
3411 store the incoming message is reallocated and moved. Unfortunately a
3412 dangling pointer to the old location is left which results in an attempt to
3413 write to the previously freed location. This is likely to result in a
3414 crash, however it could potentially lead to execution of arbitrary code.
3415
3416 This issue only affects OpenSSL 1.1.0a.
3417
3418 This issue was reported to OpenSSL by Robert Święcki.
d8dc8538 3419 ([CVE-2016-6309])
5f8e6c50
DMSP
3420
3421 *Matt Caswell*
3422
257e9d03 3423### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
5f8e6c50
DMSP
3424
3425 * OCSP Status Request extension unbounded memory growth
3426
3427 A malicious client can send an excessively large OCSP Status Request
3428 extension. If that client continually requests renegotiation, sending a
3429 large OCSP Status Request extension each time, then there will be unbounded
3430 memory growth on the server. This will eventually lead to a Denial Of
3431 Service attack through memory exhaustion. Servers with a default
3432 configuration are vulnerable even if they do not support OCSP. Builds using
3433 the "no-ocsp" build time option are not affected.
3434
3435 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
d8dc8538 3436 ([CVE-2016-6304])
5f8e6c50
DMSP
3437
3438 *Matt Caswell*
3439
3440 * SSL_peek() hang on empty record
3441
3442 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3443 sends an empty record. This could be exploited by a malicious peer in a
3444 Denial Of Service attack.
3445
3446 This issue was reported to OpenSSL by Alex Gaynor.
d8dc8538 3447 ([CVE-2016-6305])
5f8e6c50
DMSP
3448
3449 *Matt Caswell*
3450
3451 * Excessive allocation of memory in tls_get_message_header() and
3452 dtls1_preprocess_fragment()
3453
3454 A (D)TLS message includes 3 bytes for its length in the header for the
3455 message. This would allow for messages up to 16Mb in length. Messages of
3456 this length are excessive and OpenSSL includes a check to ensure that a
3457 peer is sending reasonably sized messages in order to avoid too much memory
3458 being consumed to service a connection. A flaw in the logic of version
3459 1.1.0 means that memory for the message is allocated too early, prior to
3460 the excessive message length check. Due to way memory is allocated in
3461 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3462 to service a connection. This could lead to a Denial of Service through
3463 memory exhaustion. However, the excessive message length check still takes
3464 place, and this would cause the connection to immediately fail. Assuming
3465 that the application calls SSL_free() on the failed connection in a timely
3466 manner then the 21Mb of allocated memory will then be immediately freed
3467 again. Therefore the excessive memory allocation will be transitory in
3468 nature. This then means that there is only a security impact if:
3469
3470 1) The application does not call SSL_free() in a timely manner in the event
3471 that the connection fails
3472 or
3473 2) The application is working in a constrained environment where there is
3474 very little free memory
3475 or
3476 3) The attacker initiates multiple connection attempts such that there are
3477 multiple connections in a state where memory has been allocated for the
3478 connection; SSL_free() has not yet been called; and there is insufficient
3479 memory to service the multiple requests.
3480
3481 Except in the instance of (1) above any Denial Of Service is likely to be
3482 transitory because as soon as the connection fails the memory is
3483 subsequently freed again in the SSL_free() call. However there is an
3484 increased risk during this period of application crashes due to the lack of
3485 memory - which would then mean a more serious Denial of Service.
3486
3487 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3488 (CVE-2016-6307 and CVE-2016-6308)
3489
3490 *Matt Caswell*
3491
3492 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3493 had to be removed. Primary reason is that vendor assembler can't
3494 assemble our modules with -KPIC flag. As result it, assembly
3495 support, was not even available as option. But its lack means
3496 lack of side-channel resistant code, which is incompatible with
3497 security by todays standards. Fortunately gcc is readily available
3498 prepackaged option, which we firmly point at...
3499
3500 *Andy Polyakov*
3501
257e9d03 3502### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
5f8e6c50
DMSP
3503
3504 * Windows command-line tool supports UTF-8 opt-in option for arguments
3505 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3506 (to any value) allows Windows user to access PKCS#12 file generated
3507 with Windows CryptoAPI and protected with non-ASCII password, as well
3508 as files generated under UTF-8 locale on Linux also protected with
3509 non-ASCII password.
3510
3511 *Andy Polyakov*
3512
d8dc8538 3513 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
5f8e6c50
DMSP
3514 have been disabled by default and removed from DEFAULT, just like RC4.
3515 See the RC4 item below to re-enable both.
3516
3517 *Rich Salz*
3518
3519 * The method for finding the storage location for the Windows RAND seed file
3520 has changed. First we check %RANDFILE%. If that is not set then we check
3521 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3522 all else fails we fall back to C:\.
3523
3524 *Matt Caswell*
3525
3526 * The EVP_EncryptUpdate() function has had its return type changed from void
3527 to int. A return of 0 indicates and error while a return of 1 indicates
3528 success.
3529
3530 *Matt Caswell*
3531
3532 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3533 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3534 off the constant time implementation for RSA, DSA and DH have been made
3535 no-ops and deprecated.
3536
3537 *Matt Caswell*
3538
3539 * Windows RAND implementation was simplified to only get entropy by
3540 calling CryptGenRandom(). Various other RAND-related tickets
3541 were also closed.
3542
3543 *Joseph Wylie Yandle, Rich Salz*
3544
257e9d03
RS
3545 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3546 and `OPENSSL_LH_`, respectively. The old names are available
5f8e6c50
DMSP
3547 with API compatibility. They new names are now completely documented.
3548
3549 *Rich Salz*
3550
3551 * Unify TYPE_up_ref(obj) methods signature.
3552 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3553 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3554 int (instead of void) like all others TYPE_up_ref() methods.
3555 So now these methods also check the return value of CRYPTO_atomic_add(),
3556 and the validity of object reference counter.
3557
3558 *fdasilvayy@gmail.com*
3559
3560 * With Windows Visual Studio builds, the .pdb files are installed
3561 alongside the installed libraries and executables. For a static
3562 library installation, ossl_static.pdb is the associate compiler
3563 generated .pdb file to be used when linking programs.
3564
3565 *Richard Levitte*
3566
3567 * Remove openssl.spec. Packaging files belong with the packagers.
3568
3569 *Richard Levitte*
3570
3571 * Automatic Darwin/OSX configuration has had a refresh, it will now
3572 recognise x86_64 architectures automatically. You can still decide
3573 to build for a different bitness with the environment variable
3574 KERNEL_BITS (can be 32 or 64), for example:
3575
3576 KERNEL_BITS=32 ./config
3577
3578 *Richard Levitte*
3579
3580 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3581 256 bit AES and HMAC with SHA256.
3582
3583 *Steve Henson*
3584
3585 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3586
3587 *Andy Polyakov*
3588
3589 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3590
3591 *Rich Salz*
3592
3593 * To enable users to have their own config files and build file templates,
3594 Configure looks in the directory indicated by the environment variable
3595 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3596 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3597 name and is used as is.
3598
3599 *Richard Levitte*
3600
3601 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3602 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3603 X509_CERT_FILE_CTX was removed.
3604
3605 *Rich Salz*
3606
3607 * "shared" builds are now the default. To create only static libraries use
3608 the "no-shared" Configure option.
3609
3610 *Matt Caswell*
3611
3612 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3613 All of these option have not worked for some while and are fundamental
3614 algorithms.
3615
3616 *Matt Caswell*
3617
3618 * Make various cleanup routines no-ops and mark them as deprecated. Most
3619 global cleanup functions are no longer required because they are handled
3620 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3621 Explicitly de-initing can cause problems (e.g. where a library that uses
3622 OpenSSL de-inits, but an application is still using it). The affected
3623 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3624 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3625 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3626 COMP_zlib_cleanup().
3627
3628 *Matt Caswell*
3629
3630 * --strict-warnings no longer enables runtime debugging options
3631 such as REF_DEBUG. Instead, debug options are automatically
3632 enabled with '--debug' builds.
3633
3634 *Andy Polyakov, Emilia Käsper*
3635
3636 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3637 have been moved out of the public header files. New functions for managing
3638 these have been added.
3639
3640 *Matt Caswell*
3641
3642 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3643 objects have been moved out of the public header files. New
3644 functions for managing these have been added.
3645
3646 *Richard Levitte*
3647
3648 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3649 have been moved out of the public header files. New functions for managing
3650 these have been added.
3651
3652 *Matt Caswell*
3653
3654 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3655 moved out of the public header files. New functions for managing these
3656 have been added.
3657
3658 *Matt Caswell*
3659
3660 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3661
3662 *Matt Caswell*
3663
3664 * Removed the mk1mf build scripts.
3665
3666 *Richard Levitte*
3667
3668 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3669 it is always safe to #include a header now.
3670
3671 *Rich Salz*
3672
3673 * Removed the aged BC-32 config and all its supporting scripts
3674
3675 *Richard Levitte*
3676
3677 * Removed support for Ultrix, Netware, and OS/2.
3678
3679 *Rich Salz*
3680
3681 * Add support for HKDF.
3682
3683 *Alessandro Ghedini*
3684
3685 * Add support for blake2b and blake2s
3686
3687 *Bill Cox*
3688
3689 * Added support for "pipelining". Ciphers that have the
3690 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3691 encryptions/decryptions simultaneously. There are currently no built-in
3692 ciphers with this property but the expectation is that engines will be able
3693 to offer it to significantly improve throughput. Support has been extended
3694 into libssl so that multiple records for a single connection can be
3695 processed in one go (for >=TLS 1.1).
3696
3697 *Matt Caswell*
3698
3699 * Added the AFALG engine. This is an async capable engine which is able to
3700 offload work to the Linux kernel. In this initial version it only supports
3701 AES128-CBC. The kernel must be version 4.1.0 or greater.
3702
3703 *Catriona Lucey*
3704
3705 * OpenSSL now uses a new threading API. It is no longer necessary to
3706 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3707 are two supported threading models: pthreads and windows threads. It is
3708 also possible to configure OpenSSL at compile time for "no-threads". The
3709 old threading API should no longer be used. The functions have been
3710 replaced with "no-op" compatibility macros.
3711
3712 *Alessandro Ghedini, Matt Caswell*
3713
3714 * Modify behavior of ALPN to invoke callback after SNI/servername
3715 callback, such that updates to the SSL_CTX affect ALPN.
3716
3717 *Todd Short*
3718
3719 * Add SSL_CIPHER queries for authentication and key-exchange.
3720
3721 *Todd Short*
3722
3723 * Changes to the DEFAULT cipherlist:
257e9d03
RS
3724 - Prefer (EC)DHE handshakes over plain RSA.
3725 - Prefer AEAD ciphers over legacy ciphers.
3726 - Prefer ECDSA over RSA when both certificates are available.
3727 - Prefer TLSv1.2 ciphers/PRF.
3728 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3729 default cipherlist.
5f8e6c50
DMSP
3730
3731 *Emilia Käsper*
3732
3733 * Change the ECC default curve list to be this, in order: x25519,
3734 secp256r1, secp521r1, secp384r1.
3735
3736 *Rich Salz*
3737
3738 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3739 disabled by default. They can be re-enabled using the
3740 enable-weak-ssl-ciphers option to Configure.
3741
3742 *Matt Caswell*
3743
3744 * If the server has ALPN configured, but supports no protocols that the
3745 client advertises, send a fatal "no_application_protocol" alert.
3746 This behaviour is SHALL in RFC 7301, though it isn't universally
3747 implemented by other servers.
3748
3749 *Emilia Käsper*
3750
3751 * Add X25519 support.
3752 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3753 for public and private key encoding using the format documented in
3754 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3755 key generation and key derivation.
3756
3757 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3758 X25519(29).
3759
3760 *Steve Henson*
3761
3762 * Deprecate SRP_VBASE_get_by_user.
3763 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
d8dc8538 3764 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
5f8e6c50
DMSP
3765 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3766 seed, even if the seed is configured.
3767
3768 Users should use SRP_VBASE_get1_by_user instead. Note that in
3769 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3770 also that even though configuring the SRP seed attempts to hide
3771 invalid usernames by continuing the handshake with fake
3772 credentials, this behaviour is not constant time and no strong
3773 guarantees are made that the handshake is indistinguishable from
3774 that of a valid user.
3775
3776 *Emilia Käsper*
3777
3778 * Configuration change; it's now possible to build dynamic engines
3779 without having to build shared libraries and vice versa. This
ec2bfb7d 3780 only applies to the engines in `engines/`, those in `crypto/engine/`
5f8e6c50
DMSP
3781 will always be built into libcrypto (i.e. "static").
3782
3783 Building dynamic engines is enabled by default; to disable, use
3784 the configuration option "disable-dynamic-engine".
3785
3786 The only requirements for building dynamic engines are the
3787 presence of the DSO module and building with position independent
3788 code, so they will also automatically be disabled if configuring
3789 with "disable-dso" or "disable-pic".
3790
3791 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3792 are also taken away from openssl/opensslconf.h, as they are
3793 irrelevant.
3794
3795 *Richard Levitte*
3796
3797 * Configuration change; if there is a known flag to compile
3798 position independent code, it will always be applied on the
3799 libcrypto and libssl object files, and never on the application
3800 object files. This means other libraries that use routines from
3801 libcrypto / libssl can be made into shared libraries regardless
3802 of how OpenSSL was configured.
3803
3804 If this isn't desirable, the configuration options "disable-pic"
3805 or "no-pic" can be used to disable the use of PIC. This will
3806 also disable building shared libraries and dynamic engines.
3807
3808 *Richard Levitte*
3809
3810 * Removed JPAKE code. It was experimental and has no wide use.
3811
3812 *Rich Salz*
3813
3814 * The INSTALL_PREFIX Makefile variable has been renamed to
3815 DESTDIR. That makes for less confusion on what this variable
3816 is for. Also, the configuration option --install_prefix is
3817 removed.
3818
3819 *Richard Levitte*
3820
3821 * Heartbeat for TLS has been removed and is disabled by default
3822 for DTLS; configure with enable-heartbeats. Code that uses the
3823 old #define's might need to be updated.
3824
3825 *Emilia Käsper, Rich Salz*
3826
3827 * Rename REF_CHECK to REF_DEBUG.
3828
3829 *Rich Salz*
3830
3831 * New "unified" build system
3832
3833 The "unified" build system is aimed to be a common system for all
3834 platforms we support. With it comes new support for VMS.
3835
3836 This system builds supports building in a different directory tree
3837 than the source tree. It produces one Makefile (for unix family
3838 or lookalikes), or one descrip.mms (for VMS).
3839
3840 The source of information to make the Makefile / descrip.mms is
3841 small files called 'build.info', holding the necessary
3842 information for each directory with source to compile, and a
3843 template in Configurations, like unix-Makefile.tmpl or
3844 descrip.mms.tmpl.
3845
3846 With this change, the library names were also renamed on Windows
3847 and on VMS. They now have names that are closer to the standard
3848 on Unix, and include the major version number, and in certain
3849 cases, the architecture they are built for. See "Notes on shared
3850 libraries" in INSTALL.
3851
3852 We rely heavily on the perl module Text::Template.
3853
3854 *Richard Levitte*
3855
3856 * Added support for auto-initialisation and de-initialisation of the library.
3857 OpenSSL no longer requires explicit init or deinit routines to be called,
3858 except in certain circumstances. See the OPENSSL_init_crypto() and
3859 OPENSSL_init_ssl() man pages for further information.
3860
3861 *Matt Caswell*
3862
3863 * The arguments to the DTLSv1_listen function have changed. Specifically the
3864 "peer" argument is now expected to be a BIO_ADDR object.
3865
3866 * Rewrite of BIO networking library. The BIO library lacked consistent
3867 support of IPv6, and adding it required some more extensive
3868 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3869 which hold all types of addresses and chains of address information.
3870 It also introduces a new API, with functions like BIO_socket,
3871 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3872 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3873 have been adapted accordingly.
3874
3875 *Richard Levitte*
3876
3877 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3878 the leading 0-byte.
3879
3880 *Emilia Käsper*
3881
3882 * CRIME protection: disable compression by default, even if OpenSSL is
3883 compiled with zlib enabled. Applications can still enable compression
3884 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3885 using the SSL_CONF library to configure compression.
3886
3887 *Emilia Käsper*
3888
3889 * The signature of the session callback configured with
3890 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
257e9d03
RS
3891 was explicitly marked as `const unsigned char*` instead of
3892 `unsigned char*`.
5f8e6c50
DMSP
3893
3894 *Emilia Käsper*
3895
3896 * Always DPURIFY. Remove the use of uninitialized memory in the
3897 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3898
3899 *Emilia Käsper*
3900
3901 * Removed many obsolete configuration items, including
3902 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3903 MD2_CHAR, MD2_INT, MD2_LONG
3904 BF_PTR, BF_PTR2
3905 IDEA_SHORT, IDEA_LONG
3906 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3907
3908 *Rich Salz, with advice from Andy Polyakov*
3909
3910 * Many BN internals have been moved to an internal header file.
3911
3912 *Rich Salz with help from Andy Polyakov*
3913
3914 * Configuration and writing out the results from it has changed.
3915 Files such as Makefile include/openssl/opensslconf.h and are now
3916 produced through general templates, such as Makefile.in and
3917 crypto/opensslconf.h.in and some help from the perl module
3918 Text::Template.
3919
3920 Also, the center of configuration information is no longer
3921 Makefile. Instead, Configure produces a perl module in
3922 configdata.pm which holds most of the config data (in the hash
3923 table %config), the target data that comes from the target
1dc1ea18 3924 configuration in one of the `Configurations/*.conf` files (in
5f8e6c50
DMSP
3925 %target).
3926
3927 *Richard Levitte*
3928
3929 * To clarify their intended purposes, the Configure options
3930 --prefix and --openssldir change their semantics, and become more
3931 straightforward and less interdependent.
3932
3933 --prefix shall be used exclusively to give the location INSTALLTOP
3934 where programs, scripts, libraries, include files and manuals are
3935 going to be installed. The default is now /usr/local.
3936
3937 --openssldir shall be used exclusively to give the default
3938 location OPENSSLDIR where certificates, private keys, CRLs are
3939 managed. This is also where the default openssl.cnf gets
3940 installed.
3941 If the directory given with this option is a relative path, the
3942 values of both the --prefix value and the --openssldir value will
3943 be combined to become OPENSSLDIR.
3944 The default for --openssldir is INSTALLTOP/ssl.
3945
3946 Anyone who uses --openssldir to specify where OpenSSL is to be
3947 installed MUST change to use --prefix instead.
3948
3949 *Richard Levitte*
3950
3951 * The GOST engine was out of date and therefore it has been removed. An up
3952 to date GOST engine is now being maintained in an external repository.
257e9d03 3953 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
5f8e6c50
DMSP
3954 support for GOST ciphersuites (these are only activated if a GOST engine
3955 is present).
3956
3957 *Matt Caswell*
3958
3959 * EGD is no longer supported by default; use enable-egd when
3960 configuring.
3961
3962 *Ben Kaduk and Rich Salz*
3963
3964 * The distribution now has Makefile.in files, which are used to
3965 create Makefile's when Configure is run. *Configure must be run
3966 before trying to build now.*
3967
3968 *Rich Salz*
3969
3970 * The return value for SSL_CIPHER_description() for error conditions
3971 has changed.
3972
3973 *Rich Salz*
3974
3975 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3976
3977 Obtaining and performing DNSSEC validation of TLSA records is
3978 the application's responsibility. The application provides
3979 the TLSA records of its choice to OpenSSL, and these are then
3980 used to authenticate the peer.
3981
3982 The TLSA records need not even come from DNS. They can, for
3983 example, be used to implement local end-entity certificate or
3984 trust-anchor "pinning", where the "pin" data takes the form
3985 of TLSA records, which can augment or replace verification
3986 based on the usual WebPKI public certification authorities.
3987
3988 *Viktor Dukhovni*
3989
3990 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3991 continues to support deprecated interfaces in default builds.
3992 However, applications are strongly advised to compile their
3993 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3994 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3995 or the 1.1.0 releases.
3996
3997 In environments in which all applications have been ported to
3998 not use any deprecated interfaces OpenSSL's Configure script
3999 should be used with the --api=1.1.0 option to entirely remove
4000 support for the deprecated features from the library and
4001 unconditionally disable them in the installed headers.
4002 Essentially the same effect can be achieved with the "no-deprecated"
4003 argument to Configure, except that this will always restrict
4004 the build to just the latest API, rather than a fixed API
4005 version.
4006
4007 As applications are ported to future revisions of the API,
4008 they should update their compile-time OPENSSL_API_COMPAT define
4009 accordingly, but in most cases should be able to continue to
4010 compile with later releases.
4011
4012 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4013 0x10000000L and 0x00908000L, respectively. However those
4014 versions did not support the OPENSSL_API_COMPAT feature, and
4015 so applications are not typically tested for explicit support
4016 of just the undeprecated features of either release.
4017
4018 *Viktor Dukhovni*
4019
4020 * Add support for setting the minimum and maximum supported protocol.
4021 It can bet set via the SSL_set_min_proto_version() and
4022 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4023 MaxProtocol. It's recommended to use the new APIs to disable
4024 protocols instead of disabling individual protocols using
4025 SSL_set_options() or SSL_CONF's Protocol. This change also
4026 removes support for disabling TLS 1.2 in the OpenSSL TLS
4027 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4028
4029 *Kurt Roeckx*
4030
4031 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4032
4033 *Andy Polyakov*
4034
4035 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4036 and integrates ECDSA and ECDH functionality into EC. Implementations can
4037 now redirect key generation and no longer need to convert to or from
4038 ECDSA_SIG format.
4039
4040 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4041 include the ec.h header file instead.
4042
4043 *Steve Henson*
4044
4045 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4046 ciphers who are no longer supported and drops support the ephemeral RSA key
4047 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4048
4049 *Kurt Roeckx*
4050
4051 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4052 opaque. For HMAC_CTX, the following constructors and destructors
4053 were added:
4054
1dc1ea18
DDO
4055 HMAC_CTX *HMAC_CTX_new(void);
4056 void HMAC_CTX_free(HMAC_CTX *ctx);
5f8e6c50
DMSP
4057
4058 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4059 destroy such methods has been added. See EVP_MD_meth_new(3) and
4060 EVP_CIPHER_meth_new(3) for documentation.
4061
4062 Additional changes:
1dc1ea18
DDO
4063 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4064 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4065 `EVP_MD_CTX_reset()` should be called instead to reinitialise
5f8e6c50
DMSP
4066 an already created structure.
4067 2) For consistency with the majority of our object creators and
1dc1ea18
DDO
4068 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4069 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
5f8e6c50
DMSP
4070 for deprecated builds.
4071
4072 *Richard Levitte*
4073
4074 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4075 cryptographic operations to be performed asynchronously as long as an
4076 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4077 further details. Libssl has also had this capability integrated with the
4078 introduction of the new mode SSL_MODE_ASYNC and associated error
4079 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4080 pages. This work was developed in partnership with Intel Corp.
4081
4082 *Matt Caswell*
4083
4084 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4085 always enabled now. If you want to disable the support you should
4086 exclude it using the list of supported ciphers. This also means that the
4087 "-no_ecdhe" option has been removed from s_server.
4088
4089 *Kurt Roeckx*
4090
4091 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4092 SSL_{CTX_}set1_curves() which can set a list.
4093
4094 *Kurt Roeckx*
4095
4096 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4097 curve you want to support using SSL_{CTX_}set1_curves().
4098
4099 *Kurt Roeckx*
4100
4101 * State machine rewrite. The state machine code has been significantly
4102 refactored in order to remove much duplication of code and solve issues
036cbb6b
DDO
4103 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4104 further details). This change does have some associated API changes.
4105 Notably the SSL_state() function has been removed and replaced by
4106 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4107 SSL_set_state() has been removed altogether. The previous handshake states
4108 defined in ssl.h and ssl3.h have also been removed.
5f8e6c50
DMSP
4109
4110 *Matt Caswell*
4111
4112 * All instances of the string "ssleay" in the public API were replaced
4113 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4114 Some error codes related to internal RSA_eay API's were renamed.
4115
4116 *Rich Salz*
4117
4118 * The demo files in crypto/threads were moved to demo/threads.
4119
4120 *Rich Salz*
4121
4122 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4123 sureware and ubsec.
4124
4125 *Matt Caswell, Rich Salz*
4126
4127 * New ASN.1 embed macro.
4128
4129 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4130 structure is not allocated: it is part of the parent. That is instead of
4131
4132 FOO *x;
4133
4134 it must be:
4135
4136 FOO x;
4137
4138 This reduces memory fragmentation and make it impossible to accidentally
4139 set a mandatory field to NULL.
4140
4141 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4142 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4143 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4144 SEQUENCE OF.
4145
4146 *Steve Henson*
4147
4148 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4149
4150 *Emilia Käsper*
4151
4152 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4153 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4154 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4155 DES and RC4 ciphersuites.
4156
4157 *Matt Caswell*
4158
4159 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4160 This changes the decoding behaviour for some invalid messages,
4161 though the change is mostly in the more lenient direction, and
4162 legacy behaviour is preserved as much as possible.
4163
4164 *Emilia Käsper*
4165
4166 * Fix no-stdio build.
1dc1ea18
DDO
4167 *David Woodhouse <David.Woodhouse@intel.com> and also*
4168 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
5f8e6c50
DMSP
4169
4170 * New testing framework
4171 The testing framework has been largely rewritten and is now using
4172 perl and the perl modules Test::Harness and an extended variant of
4173 Test::More called OpenSSL::Test to do its work. All test scripts in
4174 test/ have been rewritten into test recipes, and all direct calls to
4175 executables in test/Makefile have become individual recipes using the
4176 simplified testing OpenSSL::Test::Simple.
4177
4178 For documentation on our testing modules, do:
4179
4180 perldoc test/testlib/OpenSSL/Test/Simple.pm
4181 perldoc test/testlib/OpenSSL/Test.pm
4182
4183 *Richard Levitte*
4184
4185 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4186 are used; the latter aborts on memory leaks (usually checked on exit).
4187 Some undocumented "set malloc, etc., hooks" functions were removed
4188 and others were changed. All are now documented.
4189
4190 *Rich Salz*
4191
4192 * In DSA_generate_parameters_ex, if the provided seed is too short,
4193 return an error
4194
4195 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4196
4197 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4198 from RFC4279, RFC4785, RFC5487, RFC5489.
4199
4200 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4201 original RSA_PSK patch.
4202
4203 *Steve Henson*
4204
4205 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4206 era flag was never set throughout the codebase (only read). Also removed
4207 SSL3_FLAGS_POP_BUFFER which was only used if
4208 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4209
4210 *Matt Caswell*
4211
4212 * Changed the default name options in the "ca", "crl", "req" and "x509"
4213 to be "oneline" instead of "compat".
4214
4215 *Richard Levitte*
4216
4217 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4218 not aware of clients that still exhibit this bug, and the workaround
4219 hasn't been working properly for a while.
4220
4221 *Emilia Käsper*
4222
4223 * The return type of BIO_number_read() and BIO_number_written() as well as
4224 the corresponding num_read and num_write members in the BIO structure has
4225 changed from unsigned long to uint64_t. On platforms where an unsigned
4226 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4227 transferred.
4228
4229 *Matt Caswell*
4230
4231 * Given the pervasive nature of TLS extensions it is inadvisable to run
4232 OpenSSL without support for them. It also means that maintaining
4233 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4234 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
4235
4236 *Matt Caswell*
4237
4238 * Removed support for the two export grade static DH ciphersuites
4239 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4240 were newly added (along with a number of other static DH ciphersuites) to
4241 1.0.2. However the two export ones have *never* worked since they were
4242 introduced. It seems strange in any case to be adding new export
4243 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4244
4245 *Matt Caswell*
4246
4247 * Version negotiation has been rewritten. In particular SSLv23_method(),
4248 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4249 and turned into macros which simply call the new preferred function names
4250 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4251 should use the new names instead. Also as part of this change the ssl23.h
4252 header file has been removed.
4253
4254 *Matt Caswell*
4255
4256 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4257 code and the associated standard is no longer considered fit-for-purpose.
4258
4259 *Matt Caswell*
4260
4261 * RT2547 was closed. When generating a private key, try to make the
4262 output file readable only by the owner. This behavior change might
4263 be noticeable when interacting with other software.
4264
4265 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4266 Added a test.
4267
4268 *Rich Salz*
4269
4270 * Added HTTP GET support to the ocsp command.
4271
4272 *Rich Salz*
4273
4274 * Changed default digest for the dgst and enc commands from MD5 to
4275 sha256
4276
4277 *Rich Salz*
4278
4279 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4280
4281 *Matt Caswell*
4282
4283 * Added support for TLS extended master secret from
4284 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4285 initial patch which was a great help during development.
4286
4287 *Steve Henson*
4288
4289 * All libssl internal structures have been removed from the public header
4290 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4291 now redundant). Users should not attempt to access internal structures
4292 directly. Instead they should use the provided API functions.
4293
4294 *Matt Caswell*
4295
4296 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4297 Access to deprecated functions can be re-enabled by running config with
4298 "enable-deprecated". In addition applications wishing to use deprecated
4299 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4300 will, by default, disable some transitive includes that previously existed
4301 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4302
4303 *Matt Caswell*
4304
4305 * Added support for OCB mode. OpenSSL has been granted a patent license
4306 compatible with the OpenSSL license for use of OCB. Details are available
257e9d03 4307 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
5f8e6c50
DMSP
4308 for OCB can be removed by calling config with no-ocb.
4309
4310 *Matt Caswell*
4311
4312 * SSLv2 support has been removed. It still supports receiving a SSLv2
4313 compatible client hello.
4314
4315 *Kurt Roeckx*
4316
4317 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4318 done while fixing the error code for the key-too-small case.
4319
4320 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4321
4322 * CA.sh has been removed; use CA.pl instead.
4323
4324 *Rich Salz*
4325
4326 * Removed old DES API.
4327
4328 *Rich Salz*
4329
4330 * Remove various unsupported platforms:
4331 Sony NEWS4
4332 BEOS and BEOS_R5
4333 NeXT
4334 SUNOS
4335 MPE/iX
4336 Sinix/ReliantUNIX RM400
4337 DGUX
4338 NCR
4339 Tandem
4340 Cray
4341 16-bit platforms such as WIN16
4342
4343 *Rich Salz*
4344
4345 * Clean up OPENSSL_NO_xxx #define's
257e9d03
RS
4346 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4347 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4348 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4349 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4350 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4351 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4352 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4353 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4354 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4355 - Remove MS_STATIC; it's a relic from platforms <32 bits.
5f8e6c50
DMSP
4356
4357 *Rich Salz*
4358
4359 * Cleaned up dead code
4360 Remove all but one '#ifdef undef' which is to be looked at.
4361
4362 *Rich Salz*
4363
4364 * Clean up calling of xxx_free routines.
4365 Just like free(), fix most of the xxx_free routines to accept
4366 NULL. Remove the non-null checks from callers. Save much code.
4367
4368 *Rich Salz*
4369
4370 * Add secure heap for storage of private keys (when possible).
4371 Add BIO_s_secmem(), CBIGNUM, etc.
4372 Contributed by Akamai Technologies under our Corporate CLA.
4373
4374 *Rich Salz*
4375
4376 * Experimental support for a new, fast, unbiased prime candidate generator,
4377 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4378
4379 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4380
4381 * New output format NSS in the sess_id command line tool. This allows
4382 exporting the session id and the master key in NSS keylog format.
4383
4384 *Martin Kaiser <martin@kaiser.cx>*
4385
4386 * Harmonize version and its documentation. -f flag is used to display
4387 compilation flags.
4388
4389 *mancha <mancha1@zoho.com>*
4390
4391 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4392 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4393
4394 *mancha <mancha1@zoho.com>*
4395
4396 * Fix some double frees. These are not thought to be exploitable.
4397
4398 *mancha <mancha1@zoho.com>*
4399
4400 * A missing bounds check in the handling of the TLS heartbeat extension
4401 can be used to reveal up to 64k of memory to a connected client or
4402 server.
4403
4404 Thanks for Neel Mehta of Google Security for discovering this bug and to
4405 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
d8dc8538 4406 preparing the fix ([CVE-2014-0160])
5f8e6c50
DMSP
4407
4408 *Adam Langley, Bodo Moeller*
4409
4410 * Fix for the attack described in the paper "Recovering OpenSSL
4411 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4412 by Yuval Yarom and Naomi Benger. Details can be obtained from:
257e9d03 4413 <http://eprint.iacr.org/2014/140>
5f8e6c50
DMSP
4414
4415 Thanks to Yuval Yarom and Naomi Benger for discovering this
d8dc8538 4416 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
5f8e6c50
DMSP
4417
4418 *Yuval Yarom and Naomi Benger*
4419
4420 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4421 this fixes a limitation in previous versions of OpenSSL.
4422
4423 *Steve Henson*
4424
4425 * Experimental encrypt-then-mac support.
4426
4427 Experimental support for encrypt then mac from
4428 draft-gutmann-tls-encrypt-then-mac-02.txt
4429
4430 To enable it set the appropriate extension number (0x42 for the test
4431 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4432
4433 For non-compliant peers (i.e. just about everything) this should have no
4434 effect.
4435
4436 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4437
5f8e6c50
DMSP
4438 *Steve Henson*
4439
4440 * Add EVP support for key wrapping algorithms, to avoid problems with
4441 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4442 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4443 algorithms and include tests cases.
4444
4445 *Steve Henson*
4446
4447 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4448 enveloped data.
4449
4450 *Steve Henson*
4451
4452 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4453 MGF1 digest and OAEP label.
4454
4455 *Steve Henson*
4456
4457 * Make openssl verify return errors.
4458
4459 *Chris Palmer <palmer@google.com> and Ben Laurie*
4460
4461 * New function ASN1_TIME_diff to calculate the difference between two
4462 ASN1_TIME structures or one structure and the current time.
4463
4464 *Steve Henson*
4465
4466 * Update fips_test_suite to support multiple command line options. New
4467 test to induce all self test errors in sequence and check expected
4468 failures.
4469
4470 *Steve Henson*
4471
4472 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4473 sign or verify all in one operation.
4474
4475 *Steve Henson*
4476
4477 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4478 test programs and fips_test_suite. Includes functionality to parse
4479 the minimal script output of fipsalgest.pl directly.
4480
4481 *Steve Henson*
4482
4483 * Add authorisation parameter to FIPS_module_mode_set().
4484
4485 *Steve Henson*
4486
4487 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4488
4489 *Steve Henson*
4490
4491 * Use separate DRBG fields for internal and external flags. New function
4492 FIPS_drbg_health_check() to perform on demand health checking. Add
4493 generation tests to fips_test_suite with reduced health check interval to
4494 demonstrate periodic health checking. Add "nodh" option to
4495 fips_test_suite to skip very slow DH test.
4496
4497 *Steve Henson*
4498
4499 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4500 based on NID.
4501
4502 *Steve Henson*
4503
4504 * More extensive health check for DRBG checking many more failure modes.
4505 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4506 combination: call this in fips_test_suite.
4507
4508 *Steve Henson*
4509
4510 * Add support for canonical generation of DSA parameter 'g'. See
4511 FIPS 186-3 A.2.3.
4512
4513 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4514 POST to handle HMAC cases.
4515
4516 *Steve Henson*
4517
4518 * Add functions FIPS_module_version() and FIPS_module_version_text()
4519 to return numerical and string versions of the FIPS module number.
4520
4521 *Steve Henson*
4522
4523 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4524 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4525 outside the validated module in the FIPS capable OpenSSL.
4526
4527 *Steve Henson*
4528
4529 * Minor change to DRBG entropy callback semantics. In some cases
4530 there is no multiple of the block length between min_len and
4531 max_len. Allow the callback to return more than max_len bytes
4532 of entropy but discard any extra: it is the callback's responsibility
4533 to ensure that the extra data discarded does not impact the
4534 requested amount of entropy.
4535
4536 *Steve Henson*
4537
4538 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4539 information in FIPS186-3, SP800-57 and SP800-131A.
4540
4541 *Steve Henson*
4542
4543 * CCM support via EVP. Interface is very similar to GCM case except we
4544 must supply all data in one chunk (i.e. no update, final) and the
4545 message length must be supplied if AAD is used. Add algorithm test
4546 support.
4547
4548 *Steve Henson*
4549
4550 * Initial version of POST overhaul. Add POST callback to allow the status
4551 of POST to be monitored and/or failures induced. Modify fips_test_suite
4552 to use callback. Always run all selftests even if one fails.
4553
4554 *Steve Henson*
4555
4556 * XTS support including algorithm test driver in the fips_gcmtest program.
4557 Note: this does increase the maximum key length from 32 to 64 bytes but
4558 there should be no binary compatibility issues as existing applications
4559 will never use XTS mode.
4560
4561 *Steve Henson*
4562
4563 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4564 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4565 performs algorithm blocking for unapproved PRNG types. Also do not
4566 set PRNG type in FIPS_mode_set(): leave this to the application.
4567 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4568 the standard OpenSSL PRNG: set additional data to a date time vector.
4569
4570 *Steve Henson*
4571
1dc1ea18 4572 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5f8e6c50
DMSP
4573 This shouldn't present any incompatibility problems because applications
4574 shouldn't be using these directly and any that are will need to rethink
4575 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4576
4577 *Steve Henson*
4578
4579 * Extensive self tests and health checking required by SP800-90 DRBG.
4580 Remove strength parameter from FIPS_drbg_instantiate and always
4581 instantiate at maximum supported strength.
4582
4583 *Steve Henson*
4584
4585 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4586
4587 *Steve Henson*
4588
4589 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4590
4591 *Steve Henson*
4592
4593 * New function DH_compute_key_padded() to compute a DH key and pad with
4594 leading zeroes if needed: this complies with SP800-56A et al.
4595
4596 *Steve Henson*
4597
4598 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4599 anything, incomplete, subject to change and largely untested at present.
4600
4601 *Steve Henson*
4602
4603 * Modify fipscanisteronly build option to only build the necessary object
4604 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4605
4606 *Steve Henson*
4607
4608 * Add experimental option FIPSSYMS to give all symbols in
4609 fipscanister.o and FIPS or fips prefix. This will avoid
4610 conflicts with future versions of OpenSSL. Add perl script
4611 util/fipsas.pl to preprocess assembly language source files
4612 and rename any affected symbols.
4613
4614 *Steve Henson*
4615
4616 * Add selftest checks and algorithm block of non-fips algorithms in
4617 FIPS mode. Remove DES2 from selftests.
4618
4619 *Steve Henson*
4620
4621 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4622 return internal method without any ENGINE dependencies. Add new
4623 tiny fips sign and verify functions.
4624
4625 *Steve Henson*
4626
4627 * New build option no-ec2m to disable characteristic 2 code.
4628
4629 *Steve Henson*
4630
4631 * New build option "fipscanisteronly". This only builds fipscanister.o
4632 and (currently) associated fips utilities. Uses the file Makefile.fips
4633 instead of Makefile.org as the prototype.
4634
4635 *Steve Henson*
4636
4637 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4638 Update fips_gcmtest to use IV generator.
4639
4640 *Steve Henson*
4641
4642 * Initial, experimental EVP support for AES-GCM. AAD can be input by
257e9d03 4643 setting output buffer to NULL. The `*Final` function must be
5f8e6c50
DMSP
4644 called although it will not retrieve any additional data. The tag
4645 can be set or retrieved with a ctrl. The IV length is by default 12
4646 bytes (96 bits) but can be set to an alternative value. If the IV
4647 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4648 set before the key.
4649
4650 *Steve Henson*
4651
4652 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4653 underlying do_cipher function handles all cipher semantics itself
4654 including padding and finalisation. This is useful if (for example)
4655 an ENGINE cipher handles block padding itself. The behaviour of
4656 do_cipher is subtly changed if this flag is set: the return value
4657 is the number of characters written to the output buffer (zero is
4658 no longer an error code) or a negative error code. Also if the
4659 input buffer is NULL and length 0 finalisation should be performed.
4660
4661 *Steve Henson*
4662
4663 * If a candidate issuer certificate is already part of the constructed
4664 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4665
4666 *Steve Henson*
4667
4668 * Improve forward-security support: add functions
4669
4670 void SSL_CTX_set_not_resumable_session_callback(
4671 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4672 void SSL_set_not_resumable_session_callback(
4673 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4674
4675 for use by SSL/TLS servers; the callback function will be called whenever a
4676 new session is created, and gets to decide whether the session may be
4677 cached to make it resumable (return 0) or not (return 1). (As by the
4678 SSL/TLS protocol specifications, the session_id sent by the server will be
4679 empty to indicate that the session is not resumable; also, the server will
4680 not generate RFC 4507 (RFC 5077) session tickets.)
4681
4682 A simple reasonable callback implementation is to return is_forward_secure.
4683 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4684 by the SSL/TLS server library, indicating whether it can provide forward
4685 security.
4686
4687 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4688
4689 * New -verify_name option in command line utilities to set verification
4690 parameters by name.
4691
4692 *Steve Henson*
4693
4694 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4695 Add CMAC pkey methods.
4696
4697 *Steve Henson*
4698
4699 * Experimental renegotiation in s_server -www mode. If the client
4700 browses /reneg connection is renegotiated. If /renegcert it is
4701 renegotiated requesting a certificate.
4702
4703 *Steve Henson*
4704
4705 * Add an "external" session cache for debugging purposes to s_server. This
4706 should help trace issues which normally are only apparent in deployed
4707 multi-process servers.
4708
4709 *Steve Henson*
4710
4711 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4712 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4713 BIO_set_cipher() and some obscure PEM functions were changed so they
4714 can now return an error. The RAND changes required a change to the
4715 RAND_METHOD structure.
4716
4717 *Steve Henson*
4718
44652c16 4719 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5f8e6c50
DMSP
4720 a gcc attribute to warn if the result of a function is ignored. This
4721 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4722 whose return value is often ignored.
4723
4724 *Steve Henson*
4725
4726 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4727 These allow SCTs (signed certificate timestamps) to be requested and
4728 validated when establishing a connection.
4729
4730 *Rob Percival <robpercival@google.com>*
4731
44652c16
DMSP
4732OpenSSL 1.0.2
4733-------------
5f8e6c50 4734
257e9d03 4735### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5f8e6c50 4736
44652c16 4737 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
ece9304c 4738 used even when parsing explicit parameters, when loading a encoded key
44652c16
DMSP
4739 or calling `EC_GROUP_new_from_ecpkparameters()`/
4740 `EC_GROUP_new_from_ecparameters()`.
4741 This prevents bypass of security hardening and performance gains,
4742 especially for curves with specialized EC_METHODs.
4743 By default, if a key encoded with explicit parameters is loaded and later
ece9304c 4744 encoded, the output is still encoded with explicit parameters, even if
44652c16 4745 internally a "named" EC_GROUP is used for computation.
5f8e6c50 4746
44652c16 4747 *Nicola Tuveri*
5f8e6c50 4748
44652c16
DMSP
4749 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4750 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4751 NULL. After this change, only the cofactor parameter can be NULL. It also
4752 does some minimal sanity checks on the passed order.
d8dc8538 4753 ([CVE-2019-1547])
5f8e6c50 4754
44652c16 4755 *Billy Bob Brumley*
5f8e6c50 4756
44652c16
DMSP
4757 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4758 An attack is simple, if the first CMS_recipientInfo is valid but the
4759 second CMS_recipientInfo is chosen ciphertext. If the second
4760 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4761 encryption key will be replaced by garbage, and the message cannot be
4762 decoded, but if the RSA decryption fails, the correct encryption key is
4763 used and the recipient will not notice the attack.
4764 As a work around for this potential attack the length of the decrypted
4765 key must be equal to the cipher default key length, in case the
4766 certifiate is not given and all recipientInfo are tried out.
4767 The old behaviour can be re-enabled in the CMS code by setting the
4768 CMS_DEBUG_DECRYPT flag.
d8dc8538 4769 ([CVE-2019-1563])
5f8e6c50 4770
44652c16 4771 *Bernd Edlinger*
5f8e6c50 4772
44652c16 4773 * Document issue with installation paths in diverse Windows builds
5f8e6c50 4774
44652c16
DMSP
4775 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4776 binaries and run-time config file.
d8dc8538 4777 ([CVE-2019-1552])
5f8e6c50 4778
44652c16 4779 *Richard Levitte*
5f8e6c50 4780
257e9d03 4781### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5f8e6c50 4782
44652c16 4783 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
ec2bfb7d
DDO
4784 This changes the size when using the `genpkey` command when no size is given.
4785 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4786 generation commands to use 2048 bits by default.
5f8e6c50 4787
44652c16 4788 *Kurt Roeckx*
5f8e6c50 4789
44652c16 4790 * Add FIPS support for Android Arm 64-bit
5f8e6c50 4791
44652c16
DMSP
4792 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4793 Module in Version 2.0.10. For some reason, the corresponding target
4794 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4795 built with FIPS support on Android Arm 64-bit. This omission has been
4796 fixed.
5f8e6c50 4797
44652c16 4798 *Matthias St. Pierre*
5f8e6c50 4799
257e9d03 4800### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5f8e6c50 4801
44652c16 4802 * 0-byte record padding oracle
5f8e6c50 4803
44652c16
DMSP
4804 If an application encounters a fatal protocol error and then calls
4805 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4806 then OpenSSL can respond differently to the calling application if a 0 byte
4807 record is received with invalid padding compared to if a 0 byte record is
4808 received with an invalid MAC. If the application then behaves differently
4809 based on that in a way that is detectable to the remote peer, then this
4810 amounts to a padding oracle that could be used to decrypt data.
5f8e6c50 4811
44652c16
DMSP
4812 In order for this to be exploitable "non-stitched" ciphersuites must be in
4813 use. Stitched ciphersuites are optimised implementations of certain
4814 commonly used ciphersuites. Also the application must call SSL_shutdown()
4815 twice even if a protocol error has occurred (applications should not do
4816 this but some do anyway).
5f8e6c50 4817
44652c16
DMSP
4818 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4819 Aviram, with additional investigation by Steven Collison and Andrew
4820 Hourselt. It was reported to OpenSSL on 10th December 2018.
d8dc8538 4821 ([CVE-2019-1559])
5f8e6c50
DMSP
4822
4823 *Matt Caswell*
4824
44652c16 4825 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5f8e6c50 4826
44652c16 4827 *Richard Levitte*
5f8e6c50 4828
257e9d03 4829### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5f8e6c50 4830
44652c16 4831 * Microarchitecture timing vulnerability in ECC scalar multiplication
5f8e6c50 4832
44652c16
DMSP
4833 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4834 shown to be vulnerable to a microarchitecture timing side channel attack.
4835 An attacker with sufficient access to mount local timing attacks during
4836 ECDSA signature generation could recover the private key.
5f8e6c50 4837
44652c16
DMSP
4838 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4839 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4840 Nicola Tuveri.
d8dc8538 4841 ([CVE-2018-5407])
5f8e6c50 4842
44652c16 4843 *Billy Brumley*
5f8e6c50 4844
44652c16 4845 * Timing vulnerability in DSA signature generation
5f8e6c50 4846
44652c16
DMSP
4847 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4848 timing side channel attack. An attacker could use variations in the signing
4849 algorithm to recover the private key.
5f8e6c50 4850
44652c16 4851 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
d8dc8538 4852 ([CVE-2018-0734])
5f8e6c50 4853
44652c16 4854 *Paul Dale*
5f8e6c50 4855
44652c16
DMSP
4856 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4857 Module, accidentally introduced while backporting security fixes from the
4858 development branch and hindering the use of ECC in FIPS mode.
5f8e6c50 4859
44652c16 4860 *Nicola Tuveri*
5f8e6c50 4861
257e9d03 4862### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5f8e6c50 4863
44652c16 4864 * Client DoS due to large DH parameter
5f8e6c50 4865
44652c16
DMSP
4866 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4867 malicious server can send a very large prime value to the client. This will
4868 cause the client to spend an unreasonably long period of time generating a
4869 key for this prime resulting in a hang until the client has finished. This
4870 could be exploited in a Denial Of Service attack.
5f8e6c50 4871
44652c16 4872 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
d8dc8538 4873 ([CVE-2018-0732])
5f8e6c50 4874
44652c16 4875 *Guido Vranken*
5f8e6c50 4876
44652c16 4877 * Cache timing vulnerability in RSA Key Generation
5f8e6c50 4878
44652c16
DMSP
4879 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4880 a cache timing side channel attack. An attacker with sufficient access to
4881 mount cache timing attacks during the RSA key generation process could
4882 recover the private key.
5f8e6c50 4883
44652c16
DMSP
4884 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4885 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
d8dc8538 4886 ([CVE-2018-0737])
5f8e6c50 4887
44652c16 4888 *Billy Brumley*
5f8e6c50 4889
44652c16
DMSP
4890 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4891 parameter is no longer accepted, as it leads to a corrupt table. NULL
4892 pem_str is reserved for alias entries only.
5f8e6c50 4893
44652c16 4894 *Richard Levitte*
5f8e6c50 4895
44652c16
DMSP
4896 * Revert blinding in ECDSA sign and instead make problematic addition
4897 length-invariant. Switch even to fixed-length Montgomery multiplication.
5f8e6c50 4898
44652c16 4899 *Andy Polyakov*
5f8e6c50 4900
44652c16
DMSP
4901 * Change generating and checking of primes so that the error rate of not
4902 being prime depends on the intended use based on the size of the input.
4903 For larger primes this will result in more rounds of Miller-Rabin.
4904 The maximal error rate for primes with more than 1080 bits is lowered
4905 to 2^-128.
5f8e6c50 4906
44652c16 4907 *Kurt Roeckx, Annie Yousar*
5f8e6c50 4908
44652c16 4909 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5f8e6c50 4910
44652c16 4911 *Kurt Roeckx*
5f8e6c50 4912
44652c16
DMSP
4913 * Add blinding to ECDSA and DSA signatures to protect against side channel
4914 attacks discovered by Keegan Ryan (NCC Group).
5f8e6c50 4915
44652c16 4916 *Matt Caswell*
5f8e6c50 4917
44652c16
DMSP
4918 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4919 now allow empty (zero character) pass phrases.
5f8e6c50 4920
44652c16 4921 *Richard Levitte*
5f8e6c50 4922
44652c16
DMSP
4923 * Certificate time validation (X509_cmp_time) enforces stricter
4924 compliance with RFC 5280. Fractional seconds and timezone offsets
4925 are no longer allowed.
5f8e6c50 4926
44652c16 4927 *Emilia Käsper*
5f8e6c50 4928
257e9d03 4929### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5f8e6c50 4930
44652c16 4931 * Constructed ASN.1 types with a recursive definition could exceed the stack
5f8e6c50 4932
44652c16
DMSP
4933 Constructed ASN.1 types with a recursive definition (such as can be found
4934 in PKCS7) could eventually exceed the stack given malicious input with
4935 excessive recursion. This could result in a Denial Of Service attack. There
4936 are no such structures used within SSL/TLS that come from untrusted sources
4937 so this is considered safe.
5f8e6c50 4938
44652c16
DMSP
4939 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4940 project.
d8dc8538 4941 ([CVE-2018-0739])
5f8e6c50 4942
44652c16 4943 *Matt Caswell*
5f8e6c50 4944
257e9d03 4945### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5f8e6c50 4946
44652c16 4947 * Read/write after SSL object in error state
5f8e6c50 4948
44652c16
DMSP
4949 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4950 mechanism. The intent was that if a fatal error occurred during a handshake
4951 then OpenSSL would move into the error state and would immediately fail if
4952 you attempted to continue the handshake. This works as designed for the
4953 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4954 SSL_connect()), however due to a bug it does not work correctly if
4955 SSL_read() or SSL_write() is called directly. In that scenario, if the
4956 handshake fails then a fatal error will be returned in the initial function
4957 call. If SSL_read()/SSL_write() is subsequently called by the application
4958 for the same SSL object then it will succeed and the data is passed without
4959 being decrypted/encrypted directly from the SSL/TLS record layer.
5f8e6c50 4960
44652c16
DMSP
4961 In order to exploit this issue an application bug would have to be present
4962 that resulted in a call to SSL_read()/SSL_write() being issued after having
4963 already received a fatal error.
5f8e6c50 4964
44652c16 4965 This issue was reported to OpenSSL by David Benjamin (Google).
d8dc8538 4966 ([CVE-2017-3737])
5f8e6c50
DMSP
4967
4968 *Matt Caswell*
4969
44652c16 4970 * rsaz_1024_mul_avx2 overflow bug on x86_64
5f8e6c50 4971
44652c16
DMSP
4972 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4973 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4974 Analysis suggests that attacks against RSA and DSA as a result of this
4975 defect would be very difficult to perform and are not believed likely.
4976 Attacks against DH1024 are considered just feasible, because most of the
4977 work necessary to deduce information about a private key may be performed
4978 offline. The amount of resources required for such an attack would be
4979 significant. However, for an attack on TLS to be meaningful, the server
4980 would have to share the DH1024 private key among multiple clients, which is
4981 no longer an option since CVE-2016-0701.
5f8e6c50 4982
44652c16
DMSP
4983 This only affects processors that support the AVX2 but not ADX extensions
4984 like Intel Haswell (4th generation).
5f8e6c50 4985
44652c16
DMSP
4986 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4987 was originally found via the OSS-Fuzz project.
d8dc8538 4988 ([CVE-2017-3738])
5f8e6c50 4989
44652c16 4990 *Andy Polyakov*
5f8e6c50 4991
257e9d03 4992### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
44652c16
DMSP
4993
4994 * bn_sqrx8x_internal carry bug on x86_64
5f8e6c50
DMSP
4995
4996 There is a carry propagating bug in the x86_64 Montgomery squaring
4997 procedure. No EC algorithms are affected. Analysis suggests that attacks
4998 against RSA and DSA as a result of this defect would be very difficult to
4999 perform and are not believed likely. Attacks against DH are considered just
5000 feasible (although very difficult) because most of the work necessary to
5001 deduce information about a private key may be performed offline. The amount
5002 of resources required for such an attack would be very significant and
5003 likely only accessible to a limited number of attackers. An attacker would
5004 additionally need online access to an unpatched system using the target
5005 private key in a scenario with persistent DH parameters and a private
44652c16 5006 key that is shared between multiple clients.
5f8e6c50 5007
44652c16
DMSP
5008 This only affects processors that support the BMI1, BMI2 and ADX extensions
5009 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5010
5011 This issue was reported to OpenSSL by the OSS-Fuzz project.
d8dc8538 5012 ([CVE-2017-3736])
5f8e6c50
DMSP
5013
5014 *Andy Polyakov*
5015
44652c16 5016 * Malformed X.509 IPAddressFamily could cause OOB read
5f8e6c50 5017
44652c16
DMSP
5018 If an X.509 certificate has a malformed IPAddressFamily extension,
5019 OpenSSL could do a one-byte buffer overread. The most likely result
5020 would be an erroneous display of the certificate in text format.
5f8e6c50 5021
44652c16 5022 This issue was reported to OpenSSL by the OSS-Fuzz project.
5f8e6c50 5023
44652c16 5024 *Rich Salz*
5f8e6c50 5025
257e9d03 5026### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5f8e6c50 5027
44652c16
DMSP
5028 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5029 platform rather than 'mingw'.
5f8e6c50 5030
44652c16 5031 *Richard Levitte*
5f8e6c50 5032
257e9d03 5033### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5f8e6c50 5034
44652c16 5035 * Truncated packet could crash via OOB read
5f8e6c50 5036
44652c16
DMSP
5037 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5038 cipher is being used, then a truncated packet can cause that host to
5039 perform an out-of-bounds read, usually resulting in a crash.
5f8e6c50 5040
44652c16 5041 This issue was reported to OpenSSL by Robert Święcki of Google.
d8dc8538 5042 ([CVE-2017-3731])
5f8e6c50 5043
44652c16 5044 *Andy Polyakov*
5f8e6c50 5045
44652c16 5046 * BN_mod_exp may produce incorrect results on x86_64
5f8e6c50 5047
44652c16
DMSP
5048 There is a carry propagating bug in the x86_64 Montgomery squaring
5049 procedure. No EC algorithms are affected. Analysis suggests that attacks
5050 against RSA and DSA as a result of this defect would be very difficult to
5051 perform and are not believed likely. Attacks against DH are considered just
5052 feasible (although very difficult) because most of the work necessary to
5053 deduce information about a private key may be performed offline. The amount
5054 of resources required for such an attack would be very significant and
5055 likely only accessible to a limited number of attackers. An attacker would
5056 additionally need online access to an unpatched system using the target
5057 private key in a scenario with persistent DH parameters and a private
5058 key that is shared between multiple clients. For example this can occur by
5059 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5060 similar to CVE-2015-3193 but must be treated as a separate problem.
5f8e6c50 5061
44652c16 5062 This issue was reported to OpenSSL by the OSS-Fuzz project.
d8dc8538 5063 ([CVE-2017-3732])
5f8e6c50 5064
44652c16 5065 *Andy Polyakov*
5f8e6c50 5066
44652c16 5067 * Montgomery multiplication may produce incorrect results
5f8e6c50 5068
44652c16
DMSP
5069 There is a carry propagating bug in the Broadwell-specific Montgomery
5070 multiplication procedure that handles input lengths divisible by, but
5071 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5072 and DH private keys are impossible. This is because the subroutine in
5073 question is not used in operations with the private key itself and an input
5074 of the attacker's direct choice. Otherwise the bug can manifest itself as
5075 transient authentication and key negotiation failures or reproducible
5076 erroneous outcome of public-key operations with specially crafted input.
5077 Among EC algorithms only Brainpool P-512 curves are affected and one
5078 presumably can attack ECDH key negotiation. Impact was not analyzed in
5079 detail, because pre-requisites for attack are considered unlikely. Namely
5080 multiple clients have to choose the curve in question and the server has to
5081 share the private key among them, neither of which is default behaviour.
5082 Even then only clients that chose the curve will be affected.
5f8e6c50 5083
44652c16
DMSP
5084 This issue was publicly reported as transient failures and was not
5085 initially recognized as a security issue. Thanks to Richard Morgan for
5086 providing reproducible case.
d8dc8538 5087 ([CVE-2016-7055])
44652c16
DMSP
5088
5089 *Andy Polyakov*
5090
5091 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5092 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5093 prevent issues where no progress is being made and the peer continually
5094 sends unrecognised record types, using up resources processing them.
5f8e6c50
DMSP
5095
5096 *Matt Caswell*
5097
257e9d03 5098### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5f8e6c50 5099
44652c16 5100 * Missing CRL sanity check
5f8e6c50 5101
44652c16
DMSP
5102 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5103 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5104 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5f8e6c50 5105
44652c16 5106 This issue only affects the OpenSSL 1.0.2i
d8dc8538 5107 ([CVE-2016-7052])
5f8e6c50 5108
44652c16 5109 *Matt Caswell*
5f8e6c50 5110
257e9d03 5111### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5f8e6c50 5112
44652c16 5113 * OCSP Status Request extension unbounded memory growth
5f8e6c50 5114
44652c16
DMSP
5115 A malicious client can send an excessively large OCSP Status Request
5116 extension. If that client continually requests renegotiation, sending a
5117 large OCSP Status Request extension each time, then there will be unbounded
5118 memory growth on the server. This will eventually lead to a Denial Of
5119 Service attack through memory exhaustion. Servers with a default
5120 configuration are vulnerable even if they do not support OCSP. Builds using
5121 the "no-ocsp" build time option are not affected.
5f8e6c50 5122
44652c16 5123 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
d8dc8538 5124 ([CVE-2016-6304])
5f8e6c50 5125
44652c16 5126 *Matt Caswell*
5f8e6c50 5127
44652c16
DMSP
5128 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5129 HIGH to MEDIUM.
5f8e6c50 5130
44652c16
DMSP
5131 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5132 Leurent (INRIA)
d8dc8538 5133 ([CVE-2016-2183])
5f8e6c50 5134
44652c16 5135 *Rich Salz*
5f8e6c50 5136
44652c16 5137 * OOB write in MDC2_Update()
5f8e6c50 5138
44652c16
DMSP
5139 An overflow can occur in MDC2_Update() either if called directly or
5140 through the EVP_DigestUpdate() function using MDC2. If an attacker
5141 is able to supply very large amounts of input data after a previous
5142 call to EVP_EncryptUpdate() with a partial block then a length check
5143 can overflow resulting in a heap corruption.
5f8e6c50 5144
44652c16
DMSP
5145 The amount of data needed is comparable to SIZE_MAX which is impractical
5146 on most platforms.
5f8e6c50 5147
44652c16 5148 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
d8dc8538 5149 ([CVE-2016-6303])
5f8e6c50
DMSP
5150
5151 *Stephen Henson*
5152
44652c16 5153 * Malformed SHA512 ticket DoS
5f8e6c50 5154
44652c16
DMSP
5155 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5156 DoS attack where a malformed ticket will result in an OOB read which will
5157 ultimately crash.
5f8e6c50 5158
44652c16
DMSP
5159 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5160 a custom server callback and ticket lookup mechanism.
5f8e6c50 5161
44652c16 5162 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
d8dc8538 5163 ([CVE-2016-6302])
5f8e6c50 5164
44652c16 5165 *Stephen Henson*
5f8e6c50 5166
44652c16 5167 * OOB write in BN_bn2dec()
5f8e6c50 5168
44652c16
DMSP
5169 The function BN_bn2dec() does not check the return value of BN_div_word().
5170 This can cause an OOB write if an application uses this function with an
5171 overly large BIGNUM. This could be a problem if an overly large certificate
5172 or CRL is printed out from an untrusted source. TLS is not affected because
5173 record limits will reject an oversized certificate before it is parsed.
5f8e6c50 5174
44652c16 5175 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
d8dc8538 5176 ([CVE-2016-2182])
5f8e6c50 5177
44652c16 5178 *Stephen Henson*
5f8e6c50 5179
44652c16 5180 * OOB read in TS_OBJ_print_bio()
5f8e6c50 5181
44652c16
DMSP
5182 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5183 the total length the OID text representation would use and not the amount
5184 of data written. This will result in OOB reads when large OIDs are
5185 presented.
5f8e6c50 5186
44652c16 5187 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
d8dc8538 5188 ([CVE-2016-2180])
5f8e6c50 5189
44652c16 5190 *Stephen Henson*
5f8e6c50 5191
44652c16 5192 * Pointer arithmetic undefined behaviour
5f8e6c50 5193
44652c16 5194 Avoid some undefined pointer arithmetic
5f8e6c50 5195
44652c16
DMSP
5196 A common idiom in the codebase is to check limits in the following manner:
5197 "p + len > limit"
5f8e6c50 5198
44652c16
DMSP
5199 Where "p" points to some malloc'd data of SIZE bytes and
5200 limit == p + SIZE
5f8e6c50 5201
44652c16
DMSP
5202 "len" here could be from some externally supplied data (e.g. from a TLS
5203 message).
5f8e6c50 5204
44652c16
DMSP
5205 The rules of C pointer arithmetic are such that "p + len" is only well
5206 defined where len <= SIZE. Therefore the above idiom is actually
5207 undefined behaviour.
5f8e6c50 5208
44652c16
DMSP
5209 For example this could cause problems if some malloc implementation
5210 provides an address for "p" such that "p + len" actually overflows for
5211 values of len that are too big and therefore p + len < limit.
5f8e6c50 5212
44652c16 5213 This issue was reported to OpenSSL by Guido Vranken
d8dc8538 5214 ([CVE-2016-2177])
5f8e6c50 5215
44652c16 5216 *Matt Caswell*
5f8e6c50 5217
44652c16 5218 * Constant time flag not preserved in DSA signing
5f8e6c50 5219
44652c16
DMSP
5220 Operations in the DSA signing algorithm should run in constant time in
5221 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5222 implementation means that a non-constant time codepath is followed for
5223 certain operations. This has been demonstrated through a cache-timing
5224 attack to be sufficient for an attacker to recover the private DSA key.
5f8e6c50 5225
44652c16
DMSP
5226 This issue was reported by César Pereida (Aalto University), Billy Brumley
5227 (Tampere University of Technology), and Yuval Yarom (The University of
5228 Adelaide and NICTA).
d8dc8538 5229 ([CVE-2016-2178])
5f8e6c50 5230
44652c16 5231 *César Pereida*
5f8e6c50 5232
44652c16 5233 * DTLS buffered message DoS
5f8e6c50 5234
44652c16
DMSP
5235 In a DTLS connection where handshake messages are delivered out-of-order
5236 those messages that OpenSSL is not yet ready to process will be buffered
5237 for later use. Under certain circumstances, a flaw in the logic means that
5238 those messages do not get removed from the buffer even though the handshake
5239 has been completed. An attacker could force up to approx. 15 messages to
5240 remain in the buffer when they are no longer required. These messages will
5241 be cleared when the DTLS connection is closed. The default maximum size for
5242 a message is 100k. Therefore the attacker could force an additional 1500k
5243 to be consumed per connection. By opening many simulataneous connections an
5244 attacker could cause a DoS attack through memory exhaustion.
5f8e6c50 5245
44652c16 5246 This issue was reported to OpenSSL by Quan Luo.
d8dc8538 5247 ([CVE-2016-2179])
5f8e6c50 5248
44652c16 5249 *Matt Caswell*
5f8e6c50 5250
44652c16 5251 * DTLS replay protection DoS
5f8e6c50 5252
44652c16
DMSP
5253 A flaw in the DTLS replay attack protection mechanism means that records
5254 that arrive for future epochs update the replay protection "window" before
5255 the MAC for the record has been validated. This could be exploited by an
5256 attacker by sending a record for the next epoch (which does not have to
5257 decrypt or have a valid MAC), with a very large sequence number. This means
5258 that all subsequent legitimate packets are dropped causing a denial of
5259 service for a specific DTLS connection.
5f8e6c50 5260
44652c16 5261 This issue was reported to OpenSSL by the OCAP audit team.
d8dc8538 5262 ([CVE-2016-2181])
5f8e6c50 5263
44652c16 5264 *Matt Caswell*
5f8e6c50 5265
44652c16 5266 * Certificate message OOB reads
5f8e6c50 5267
44652c16
DMSP
5268 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5269 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5270 theoretical DoS risk but this has not been observed in practice on common
5271 platforms.
5f8e6c50 5272
44652c16
DMSP
5273 The messages affected are client certificate, client certificate request
5274 and server certificate. As a result the attack can only be performed
5275 against a client or a server which enables client authentication.
5f8e6c50 5276
44652c16 5277 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
d8dc8538 5278 ([CVE-2016-6306])
5f8e6c50 5279
44652c16 5280 *Stephen Henson*
5f8e6c50 5281
257e9d03 5282### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5f8e6c50 5283
44652c16 5284 * Prevent padding oracle in AES-NI CBC MAC check
5f8e6c50 5285
44652c16
DMSP
5286 A MITM attacker can use a padding oracle attack to decrypt traffic
5287 when the connection uses an AES CBC cipher and the server support
5288 AES-NI.
5f8e6c50 5289
44652c16 5290 This issue was introduced as part of the fix for Lucky 13 padding
d8dc8538 5291 attack ([CVE-2013-0169]). The padding check was rewritten to be in
44652c16
DMSP
5292 constant time by making sure that always the same bytes are read and
5293 compared against either the MAC or padding bytes. But it no longer
5294 checked that there was enough data to have both the MAC and padding
5295 bytes.
5f8e6c50 5296
44652c16 5297 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5f8e6c50 5298
44652c16 5299 *Kurt Roeckx*
5f8e6c50 5300
44652c16
DMSP
5301 * Fix EVP_EncodeUpdate overflow
5302
5303 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5304 Base64 encoding of binary data. If an attacker is able to supply very large
5305 amounts of input data then a length check can overflow resulting in a heap
5306 corruption.
5307
5308 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
257e9d03 5309 the `PEM_write_bio*` family of functions. These are mainly used within the
44652c16
DMSP
5310 OpenSSL command line applications, so any application which processes data
5311 from an untrusted source and outputs it as a PEM file should be considered
5312 vulnerable to this issue. User applications that call these APIs directly
5313 with large amounts of untrusted data may also be vulnerable.
5314
5315 This issue was reported by Guido Vranken.
d8dc8538 5316 ([CVE-2016-2105])
5f8e6c50
DMSP
5317
5318 *Matt Caswell*
5319
44652c16 5320 * Fix EVP_EncryptUpdate overflow
5f8e6c50 5321
44652c16
DMSP
5322 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5323 is able to supply very large amounts of input data after a previous call to
5324 EVP_EncryptUpdate() with a partial block then a length check can overflow
5325 resulting in a heap corruption. Following an analysis of all OpenSSL
5326 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5327 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5328 the first called function after an EVP_EncryptInit(), and therefore that
5329 specific call must be safe. The second form is where the length passed to
5330 EVP_EncryptUpdate() can be seen from the code to be some small value and
5331 therefore there is no possibility of an overflow. Since all instances are
5332 one of these two forms, it is believed that there can be no overflows in
5333 internal code due to this problem. It should be noted that
5334 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5335 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5336 of these calls have also been analysed too and it is believed there are no
5337 instances in internal usage where an overflow could occur.
5f8e6c50 5338
44652c16 5339 This issue was reported by Guido Vranken.
d8dc8538 5340 ([CVE-2016-2106])
5f8e6c50
DMSP
5341
5342 *Matt Caswell*
5343
44652c16 5344 * Prevent ASN.1 BIO excessive memory allocation
5f8e6c50 5345
44652c16
DMSP
5346 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5347 a short invalid encoding can cause allocation of large amounts of memory
5348 potentially consuming excessive resources or exhausting memory.
5f8e6c50 5349
44652c16
DMSP
5350 Any application parsing untrusted data through d2i BIO functions is
5351 affected. The memory based functions such as d2i_X509() are *not* affected.
5352 Since the memory based functions are used by the TLS library, TLS
5353 applications are not affected.
5354
5355 This issue was reported by Brian Carpenter.
d8dc8538 5356 ([CVE-2016-2109])
5f8e6c50
DMSP
5357
5358 *Stephen Henson*
5359
44652c16 5360 * EBCDIC overread
5f8e6c50 5361
44652c16
DMSP
5362 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5363 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5364 in arbitrary stack data being returned in the buffer.
5f8e6c50 5365
44652c16 5366 This issue was reported by Guido Vranken.
d8dc8538 5367 ([CVE-2016-2176])
5f8e6c50 5368
44652c16 5369 *Matt Caswell*
5f8e6c50 5370
44652c16
DMSP
5371 * Modify behavior of ALPN to invoke callback after SNI/servername
5372 callback, such that updates to the SSL_CTX affect ALPN.
5f8e6c50 5373
44652c16 5374 *Todd Short*
5f8e6c50 5375
44652c16
DMSP
5376 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5377 default.
5378
5379 *Kurt Roeckx*
5380
5381 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5382 methods are enabled and ssl2 is disabled the methods return NULL.
5383
5384 *Kurt Roeckx*
5385
257e9d03 5386### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
44652c16
DMSP
5387
5388* Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5389 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5390 provide any "EXPORT" or "LOW" strength ciphers.
5391
5392 *Viktor Dukhovni*
5393
5394* Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5395 is by default disabled at build-time. Builds that are not configured with
5396 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5397 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5398 will need to explicitly call either of:
5399
5400 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5401 or
5402 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5403
5404 as appropriate. Even if either of those is used, or the application
5405 explicitly uses the version-specific SSLv2_method() or its client and
5406 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5407 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5408 ciphers, and SSLv2 56-bit DES are no longer available.
d8dc8538 5409 ([CVE-2016-0800])
44652c16
DMSP
5410
5411 *Viktor Dukhovni*
5412
5413 * Fix a double-free in DSA code
5414
5415 A double free bug was discovered when OpenSSL parses malformed DSA private
5416 keys and could lead to a DoS attack or memory corruption for applications
5417 that receive DSA private keys from untrusted sources. This scenario is
5418 considered rare.
5419
5420 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5421 libFuzzer.
d8dc8538 5422 ([CVE-2016-0705])
44652c16
DMSP
5423
5424 *Stephen Henson*
5425
5426 * Disable SRP fake user seed to address a server memory leak.
5427
5428 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5429
5430 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5431 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5432 was changed to ignore the "fake user" SRP seed, even if the seed
5433 is configured.
5434
5435 Users should use SRP_VBASE_get1_by_user instead. Note that in
5436 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5437 also that even though configuring the SRP seed attempts to hide
5438 invalid usernames by continuing the handshake with fake
5439 credentials, this behaviour is not constant time and no strong
5440 guarantees are made that the handshake is indistinguishable from
5441 that of a valid user.
d8dc8538 5442 ([CVE-2016-0798])
44652c16
DMSP
5443
5444 *Emilia Käsper*
5445
5446 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5447
5448 In the BN_hex2bn function the number of hex digits is calculated using an
1dc1ea18
DDO
5449 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5450 large values of `i` this can result in `bn_expand` not allocating any
5451 memory because `i * 4` is negative. This can leave the internal BIGNUM data
44652c16 5452 field as NULL leading to a subsequent NULL ptr deref. For very large values
1dc1ea18 5453 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
44652c16
DMSP
5454 In this case memory is allocated to the internal BIGNUM data field, but it
5455 is insufficiently sized leading to heap corruption. A similar issue exists
5456 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5457 is ever called by user applications with very large untrusted hex/dec data.
5458 This is anticipated to be a rare occurrence.
5459
5460 All OpenSSL internal usage of these functions use data that is not expected
5461 to be untrusted, e.g. config file data or application command line
5462 arguments. If user developed applications generate config file data based
5463 on untrusted data then it is possible that this could also lead to security
5464 consequences. This is also anticipated to be rare.
5465
5466 This issue was reported to OpenSSL by Guido Vranken.
d8dc8538 5467 ([CVE-2016-0797])
44652c16
DMSP
5468
5469 *Matt Caswell*
5470
257e9d03 5471 * Fix memory issues in `BIO_*printf` functions
44652c16 5472
1dc1ea18 5473 The internal `fmtstr` function used in processing a "%s" format string in
257e9d03 5474 the `BIO_*printf` functions could overflow while calculating the length of a
44652c16
DMSP
5475 string and cause an OOB read when printing very long strings.
5476
1dc1ea18 5477 Additionally the internal `doapr_outch` function can attempt to write to an
44652c16
DMSP
5478 OOB memory location (at an offset from the NULL pointer) in the event of a
5479 memory allocation failure. In 1.0.2 and below this could be caused where
5480 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5481 could be in processing a very long "%s" format string. Memory leaks can
5482 also occur.
5483
5484 The first issue may mask the second issue dependent on compiler behaviour.
5485 These problems could enable attacks where large amounts of untrusted data
257e9d03 5486 is passed to the `BIO_*printf` functions. If applications use these functions
44652c16
DMSP
5487 in this way then they could be vulnerable. OpenSSL itself uses these
5488 functions when printing out human-readable dumps of ASN.1 data. Therefore
5489 applications that print this data could be vulnerable if the data is from
5490 untrusted sources. OpenSSL command line applications could also be
5491 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5492 as command line arguments.
5493
5494 Libssl is not considered directly vulnerable. Additionally certificates etc
5495 received via remote connections via libssl are also unlikely to be able to
5496 trigger these issues because of message size limits enforced within libssl.
5497
5498 This issue was reported to OpenSSL Guido Vranken.
d8dc8538 5499 ([CVE-2016-0799])
44652c16
DMSP
5500
5501 *Matt Caswell*
5502
5503 * Side channel attack on modular exponentiation
5504
5505 A side-channel attack was found which makes use of cache-bank conflicts on
5506 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5507 of RSA keys. The ability to exploit this issue is limited as it relies on
5508 an attacker who has control of code in a thread running on the same
5509 hyper-threaded core as the victim thread which is performing decryptions.
5510
5511 This issue was reported to OpenSSL by Yuval Yarom, The University of
5512 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5513 Nadia Heninger, University of Pennsylvania with more information at
257e9d03 5514 <http://cachebleed.info>.
d8dc8538 5515 ([CVE-2016-0702])
44652c16
DMSP
5516
5517 *Andy Polyakov*
5518
ec2bfb7d 5519 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
44652c16
DMSP
5520 if no keysize is specified with default_bits. This fixes an
5521 omission in an earlier change that changed all RSA/DSA key generation
ec2bfb7d 5522 commands to use 2048 bits by default.
44652c16
DMSP
5523
5524 *Emilia Käsper*
5525
257e9d03
RS
5526### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5527
44652c16
DMSP
5528 * DH small subgroups
5529
5530 Historically OpenSSL only ever generated DH parameters based on "safe"
5531 primes. More recently (in version 1.0.2) support was provided for
5532 generating X9.42 style parameter files such as those required for RFC 5114
5533 support. The primes used in such files may not be "safe". Where an
5534 application is using DH configured with parameters based on primes that are
5535 not "safe" then an attacker could use this fact to find a peer's private
5536 DH exponent. This attack requires that the attacker complete multiple
5537 handshakes in which the peer uses the same private DH exponent. For example
5538 this could be used to discover a TLS server's private DH exponent if it's
5539 reusing the private DH exponent or it's using a static DH ciphersuite.
5540
5541 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5542 TLS. It is not on by default. If the option is not set then the server
5543 reuses the same private DH exponent for the life of the server process and
5544 would be vulnerable to this attack. It is believed that many popular
5545 applications do set this option and would therefore not be at risk.
5546
5547 The fix for this issue adds an additional check where a "q" parameter is
5548 available (as is the case in X9.42 based parameters). This detects the
5549 only known attack, and is the only possible defense for static DH
5550 ciphersuites. This could have some performance impact.
5551
5552 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5553 default and cannot be disabled. This could have some performance impact.
5554
5555 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
d8dc8538 5556 ([CVE-2016-0701])
44652c16
DMSP
5557
5558 *Matt Caswell*
5559
5560 * SSLv2 doesn't block disabled ciphers
5561
5562 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5563 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5564 been disabled, provided that the SSLv2 protocol was not also disabled via
5565 SSL_OP_NO_SSLv2.
5566
5567 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5568 and Sebastian Schinzel.
d8dc8538 5569 ([CVE-2015-3197])
44652c16
DMSP
5570
5571 *Viktor Dukhovni*
5572
257e9d03 5573### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
44652c16
DMSP
5574
5575 * BN_mod_exp may produce incorrect results on x86_64
5576
5577 There is a carry propagating bug in the x86_64 Montgomery squaring
5578 procedure. No EC algorithms are affected. Analysis suggests that attacks
5579 against RSA and DSA as a result of this defect would be very difficult to
5580 perform and are not believed likely. Attacks against DH are considered just
5581 feasible (although very difficult) because most of the work necessary to
5582 deduce information about a private key may be performed offline. The amount
5583 of resources required for such an attack would be very significant and
5584 likely only accessible to a limited number of attackers. An attacker would
5585 additionally need online access to an unpatched system using the target
5586 private key in a scenario with persistent DH parameters and a private
5587 key that is shared between multiple clients. For example this can occur by
5588 default in OpenSSL DHE based SSL/TLS ciphersuites.
5589
5590 This issue was reported to OpenSSL by Hanno Böck.
d8dc8538 5591 ([CVE-2015-3193])
44652c16
DMSP
5592
5593 *Andy Polyakov*
5594
5595 * Certificate verify crash with missing PSS parameter
5596
5597 The signature verification routines will crash with a NULL pointer
5598 dereference if presented with an ASN.1 signature using the RSA PSS
5599 algorithm and absent mask generation function parameter. Since these
5600 routines are used to verify certificate signature algorithms this can be
5601 used to crash any certificate verification operation and exploited in a
5602 DoS attack. Any application which performs certificate verification is
5603 vulnerable including OpenSSL clients and servers which enable client
5604 authentication.
5605
5606 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
d8dc8538 5607 ([CVE-2015-3194])
44652c16
DMSP
5608
5609 *Stephen Henson*
5610
5611 * X509_ATTRIBUTE memory leak
5612
5613 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5614 memory. This structure is used by the PKCS#7 and CMS routines so any
5615 application which reads PKCS#7 or CMS data from untrusted sources is
5616 affected. SSL/TLS is not affected.
5617
5618 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5619 libFuzzer.
d8dc8538 5620 ([CVE-2015-3195])
44652c16
DMSP
5621
5622 *Stephen Henson*
5623
5624 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5625 This changes the decoding behaviour for some invalid messages,
5626 though the change is mostly in the more lenient direction, and
5627 legacy behaviour is preserved as much as possible.
5628
5629 *Emilia Käsper*
5630
5631 * In DSA_generate_parameters_ex, if the provided seed is too short,
5632 return an error
5633
5634 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5635
257e9d03 5636### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
44652c16
DMSP
5637
5638 * Alternate chains certificate forgery
5639
5640 During certificate verification, OpenSSL will attempt to find an
5641 alternative certificate chain if the first attempt to build such a chain
5642 fails. An error in the implementation of this logic can mean that an
5643 attacker could cause certain checks on untrusted certificates to be
5644 bypassed, such as the CA flag, enabling them to use a valid leaf
5645 certificate to act as a CA and "issue" an invalid certificate.
5646
5647 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5648 (Google/BoringSSL).
5649
5650 *Matt Caswell*
5651
257e9d03 5652### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
44652c16
DMSP
5653
5654 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5655 incompatibility in the handling of HMAC. The previous ABI has now been
5656 restored.
5657
5658 *Matt Caswell*
5659
257e9d03 5660### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
44652c16
DMSP
5661
5662 * Malformed ECParameters causes infinite loop
5663
5664 When processing an ECParameters structure OpenSSL enters an infinite loop
5665 if the curve specified is over a specially malformed binary polynomial
5666 field.
5667
5668 This can be used to perform denial of service against any
5669 system which processes public keys, certificate requests or
5670 certificates. This includes TLS clients and TLS servers with
5671 client authentication enabled.
5672
5673 This issue was reported to OpenSSL by Joseph Barr-Pixton.
d8dc8538 5674 ([CVE-2015-1788])
44652c16
DMSP
5675
5676 *Andy Polyakov*
5677
5678 * Exploitable out-of-bounds read in X509_cmp_time
5679
5680 X509_cmp_time does not properly check the length of the ASN1_TIME
5681 string and can read a few bytes out of bounds. In addition,
5682 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5683 time string.
5684
5685 An attacker can use this to craft malformed certificates and CRLs of
5686 various sizes and potentially cause a segmentation fault, resulting in
5687 a DoS on applications that verify certificates or CRLs. TLS clients
5688 that verify CRLs are affected. TLS clients and servers with client
5689 authentication enabled may be affected if they use custom verification
5690 callbacks.
5691
5692 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5693 independently by Hanno Böck.
d8dc8538 5694 ([CVE-2015-1789])
44652c16
DMSP
5695
5696 *Emilia Käsper*
5697
5698 * PKCS7 crash with missing EnvelopedContent
5699
5700 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5701 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5702 with missing content and trigger a NULL pointer dereference on parsing.
5703
5704 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5705 structures from untrusted sources are affected. OpenSSL clients and
5706 servers are not affected.
5707
5708 This issue was reported to OpenSSL by Michal Zalewski (Google).
d8dc8538 5709 ([CVE-2015-1790])
44652c16
DMSP
5710
5711 *Emilia Käsper*
5712
5713 * CMS verify infinite loop with unknown hash function
5714
5715 When verifying a signedData message the CMS code can enter an infinite loop
5716 if presented with an unknown hash function OID. This can be used to perform
5717 denial of service against any system which verifies signedData messages using
5718 the CMS code.
5719 This issue was reported to OpenSSL by Johannes Bauer.
d8dc8538 5720 ([CVE-2015-1792])
44652c16
DMSP
5721
5722 *Stephen Henson*
5723
5724 * Race condition handling NewSessionTicket
5725
5726 If a NewSessionTicket is received by a multi-threaded client when attempting to
5727 reuse a previous ticket then a race condition can occur potentially leading to
5728 a double free of the ticket data.
d8dc8538 5729 ([CVE-2015-1791])
44652c16
DMSP
5730
5731 *Matt Caswell*
5732
5733 * Only support 256-bit or stronger elliptic curves with the
5734 'ecdh_auto' setting (server) or by default (client). Of supported
5735 curves, prefer P-256 (both).
5736
5737 *Emilia Kasper*
5738
257e9d03 5739### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
44652c16
DMSP
5740
5741 * ClientHello sigalgs DoS fix
5742
5743 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5744 invalid signature algorithms extension a NULL pointer dereference will
5745 occur. This can be exploited in a DoS attack against the server.
5746
5747 This issue was was reported to OpenSSL by David Ramos of Stanford
5748 University.
d8dc8538 5749 ([CVE-2015-0291])
44652c16
DMSP
5750
5751 *Stephen Henson and Matt Caswell*
5752
5753 * Multiblock corrupted pointer fix
5754
5755 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5756 feature only applies on 64 bit x86 architecture platforms that support AES
5757 NI instructions. A defect in the implementation of "multiblock" can cause
5758 OpenSSL's internal write buffer to become incorrectly set to NULL when
5759 using non-blocking IO. Typically, when the user application is using a
5760 socket BIO for writing, this will only result in a failed connection.
5761 However if some other BIO is used then it is likely that a segmentation
5762 fault will be triggered, thus enabling a potential DoS attack.
5763
5764 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
d8dc8538 5765 ([CVE-2015-0290])
44652c16
DMSP
5766
5767 *Matt Caswell*
5768
5769 * Segmentation fault in DTLSv1_listen fix
5770
5771 The DTLSv1_listen function is intended to be stateless and processes the
5772 initial ClientHello from many peers. It is common for user code to loop
5773 over the call to DTLSv1_listen until a valid ClientHello is received with
5774 an associated cookie. A defect in the implementation of DTLSv1_listen means
5775 that state is preserved in the SSL object from one invocation to the next
5776 that can lead to a segmentation fault. Errors processing the initial
5777 ClientHello can trigger this scenario. An example of such an error could be
5778 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5779 server.
5780
5781 This issue was reported to OpenSSL by Per Allansson.
d8dc8538 5782 ([CVE-2015-0207])
44652c16
DMSP
5783
5784 *Matt Caswell*
5785
5786 * Segmentation fault in ASN1_TYPE_cmp fix
5787
5788 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5789 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5790 certificate signature algorithm consistency this can be used to crash any
5791 certificate verification operation and exploited in a DoS attack. Any
5792 application which performs certificate verification is vulnerable including
5793 OpenSSL clients and servers which enable client authentication.
d8dc8538 5794 ([CVE-2015-0286])
44652c16
DMSP
5795
5796 *Stephen Henson*
5797
5798 * Segmentation fault for invalid PSS parameters fix
5799
5800 The signature verification routines will crash with a NULL pointer
5801 dereference if presented with an ASN.1 signature using the RSA PSS
5802 algorithm and invalid parameters. Since these routines are used to verify
5803 certificate signature algorithms this can be used to crash any
5804 certificate verification operation and exploited in a DoS attack. Any
5805 application which performs certificate verification is vulnerable including
5806 OpenSSL clients and servers which enable client authentication.
5807
5808 This issue was was reported to OpenSSL by Brian Carpenter.
d8dc8538 5809 ([CVE-2015-0208])
44652c16
DMSP
5810
5811 *Stephen Henson*
5812
5813 * ASN.1 structure reuse memory corruption fix
5814
5815 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5816 memory corruption via an invalid write. Such reuse is and has been
5817 strongly discouraged and is believed to be rare.
5818
5819 Applications that parse structures containing CHOICE or ANY DEFINED BY
5820 components may be affected. Certificate parsing (d2i_X509 and related
5821 functions) are however not affected. OpenSSL clients and servers are
5822 not affected.
d8dc8538 5823 ([CVE-2015-0287])
44652c16
DMSP
5824
5825 *Stephen Henson*
5826
5827 * PKCS7 NULL pointer dereferences fix
5828
5829 The PKCS#7 parsing code does not handle missing outer ContentInfo
5830 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5831 missing content and trigger a NULL pointer dereference on parsing.
5832
5833 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5834 otherwise parse PKCS#7 structures from untrusted sources are
5835 affected. OpenSSL clients and servers are not affected.
5836
5837 This issue was reported to OpenSSL by Michal Zalewski (Google).
d8dc8538 5838 ([CVE-2015-0289])
44652c16
DMSP
5839
5840 *Emilia Käsper*
5841
5842 * DoS via reachable assert in SSLv2 servers fix
5843
5844 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5845 servers that both support SSLv2 and enable export cipher suites by sending
5846 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5847
5848 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5849 (OpenSSL development team).
d8dc8538 5850 ([CVE-2015-0293])
44652c16
DMSP
5851
5852 *Emilia Käsper*
5853
5854 * Empty CKE with client auth and DHE fix
5855
5856 If client auth is used then a server can seg fault in the event of a DHE
5857 ciphersuite being selected and a zero length ClientKeyExchange message
5858 being sent by the client. This could be exploited in a DoS attack.
d8dc8538 5859 ([CVE-2015-1787])
44652c16
DMSP
5860
5861 *Matt Caswell*
5862
5863 * Handshake with unseeded PRNG fix
5864
5865 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5866 with an unseeded PRNG. The conditions are:
5867 - The client is on a platform where the PRNG has not been seeded
5868 automatically, and the user has not seeded manually
5869 - A protocol specific client method version has been used (i.e. not
5870 SSL_client_methodv23)
5871 - A ciphersuite is used that does not require additional random data from
5872 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5873
5874 If the handshake succeeds then the client random that has been used will
5875 have been generated from a PRNG with insufficient entropy and therefore the
5876 output may be predictable.
5877
5878 For example using the following command with an unseeded openssl will
5879 succeed on an unpatched platform:
5880
5881 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
d8dc8538 5882 ([CVE-2015-0285])
44652c16
DMSP
5883
5884 *Matt Caswell*
5885
5886 * Use After Free following d2i_ECPrivatekey error fix
5887
5888 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5889 could cause a use after free condition. This, in turn, could cause a double
5890 free in several private key parsing functions (such as d2i_PrivateKey
5891 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5892 for applications that receive EC private keys from untrusted
5893 sources. This scenario is considered rare.
5894
5895 This issue was discovered by the BoringSSL project and fixed in their
5896 commit 517073cd4b.
d8dc8538 5897 ([CVE-2015-0209])
44652c16
DMSP
5898
5899 *Matt Caswell*
5900
5901 * X509_to_X509_REQ NULL pointer deref fix
5902
5903 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5904 the certificate key is invalid. This function is rarely used in practice.
5905
5906 This issue was discovered by Brian Carpenter.
d8dc8538 5907 ([CVE-2015-0288])
44652c16
DMSP
5908
5909 *Stephen Henson*
5910
5911 * Removed the export ciphers from the DEFAULT ciphers
5912
5913 *Kurt Roeckx*
5914
257e9d03 5915### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
44652c16
DMSP
5916
5917 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5918 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5919 So far those who have to target multiple platforms would compromise
5920 and argue that binary targeting say ARMv5 would still execute on
5921 ARMv8. "Universal" build resolves this compromise by providing
5922 near-optimal performance even on newer platforms.
5923
5924 *Andy Polyakov*
5925
5926 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5927 (other platforms pending).
5928
5929 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5f8e6c50
DMSP
5930
5931 * Add support for the SignedCertificateTimestampList certificate and
5932 OCSP response extensions from RFC6962.
5933
44652c16
DMSP
5934 *Rob Stradling*
5935
5936 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5937 for corner cases. (Certain input points at infinity could lead to
5938 bogus results, with non-infinity inputs mapped to infinity too.)
5939
5940 *Bodo Moeller*
5941
5942 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5943 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5944 common cases are optimized and there still is room for further
5945 improvements. Vector Permutation AES for Altivec is also added.
5946
5947 *Andy Polyakov*
5948
5949 * Add support for little-endian ppc64 Linux target.
5950
5951 *Marcelo Cerri (IBM)*
5952
5953 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5954 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5955 are optimized and there still is room for further improvements.
5956 Both 32- and 64-bit modes are supported.
5957
5958 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5959
5960 * Improved ARMv7 NEON support.
5961
5962 *Andy Polyakov*
5963
5964 * Support for SPARC Architecture 2011 crypto extensions, first
5965 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5966 SHA256/512, MD5, GHASH and modular exponentiation.
5967
5968 *Andy Polyakov, David Miller*
5969
5970 * Accelerated modular exponentiation for Intel processors, a.k.a.
5971 RSAZ.
5972
5973 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5974
5975 * Support for new and upcoming Intel processors, including AVX2,
5976 BMI and SHA ISA extensions. This includes additional "stitched"
5977 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5978 for TLS encrypt.
5979
5980 This work was sponsored by Intel Corp.
5981
5982 *Andy Polyakov*
5983
5984 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5985 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5986 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5987
5988 *Steve Henson*
5989
5990 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5991 this fixes a limitation in previous versions of OpenSSL.
5992
5993 *Steve Henson*
5994
5995 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5996 MGF1 digest and OAEP label.
5997
5998 *Steve Henson*
5999
6000 * Add EVP support for key wrapping algorithms, to avoid problems with
6001 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6002 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6003 algorithms and include tests cases.
6004
6005 *Steve Henson*
6006
6007 * Add functions to allocate and set the fields of an ECDSA_METHOD
6008 structure.
6009
6010 *Douglas E. Engert, Steve Henson*
6011
6012 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6013 difference in days and seconds between two tm or ASN1_TIME structures.
6014
6015 *Steve Henson*
6016
6017 * Add -rev test option to s_server to just reverse order of characters
6018 received by client and send back to server. Also prints an abbreviated
6019 summary of the connection parameters.
6020
6021 *Steve Henson*
6022
6023 * New option -brief for s_client and s_server to print out a brief summary
6024 of connection parameters.
6025
6026 *Steve Henson*
6027
6028 * Add callbacks for arbitrary TLS extensions.
6029
6030 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6031
6032 * New option -crl_download in several openssl utilities to download CRLs
6033 from CRLDP extension in certificates.
6034
6035 *Steve Henson*
6036
6037 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6038
6039 *Steve Henson*
6040
6041 * New function X509_CRL_diff to generate a delta CRL from the difference
6042 of two full CRLs. Add support to "crl" utility.
6043
6044 *Steve Henson*
6045
6046 * New functions to set lookup_crls function and to retrieve
6047 X509_STORE from X509_STORE_CTX.
6048
6049 *Steve Henson*
6050
6051 * Print out deprecated issuer and subject unique ID fields in
6052 certificates.
6053
6054 *Steve Henson*
6055
6056 * Extend OCSP I/O functions so they can be used for simple general purpose
6057 HTTP as well as OCSP. New wrapper function which can be used to download
6058 CRLs using the OCSP API.
6059
6060 *Steve Henson*
6061
6062 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6063
6064 *Steve Henson*
6065
257e9d03 6066 * `SSL_CONF*` functions. These provide a common framework for application
44652c16
DMSP
6067 configuration using configuration files or command lines.
6068
6069 *Steve Henson*
6070
6071 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6072 message callback and prints the results. Needs compile time option
6073 "enable-ssl-trace". New options to s_client and s_server to enable
6074 tracing.
6075
6076 *Steve Henson*
6077
6078 * New ctrl and macro to retrieve supported points extensions.
6079 Print out extension in s_server and s_client.
6080
6081 *Steve Henson*
6082
6083 * New functions to retrieve certificate signature and signature
6084 OID NID.
6085
6086 *Steve Henson*
6087
6088 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6089 client to OpenSSL.
6090
6091 *Steve Henson*
6092
6093 * New Suite B modes for TLS code. These use and enforce the requirements
6094 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6095 only use Suite B curves. The Suite B modes can be set by using the
6096 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6097
6098 *Steve Henson*
6099
6100 * New chain verification flags for Suite B levels of security. Check
6101 algorithms are acceptable when flags are set in X509_verify_cert.
6102
6103 *Steve Henson*
6104
6105 * Make tls1_check_chain return a set of flags indicating checks passed
6106 by a certificate chain. Add additional tests to handle client
6107 certificates: checks for matching certificate type and issuer name
6108 comparison.
6109
6110 *Steve Henson*
6111
6112 * If an attempt is made to use a signature algorithm not in the peer
6113 preference list abort the handshake. If client has no suitable
6114 signature algorithms in response to a certificate request do not
6115 use the certificate.
6116
6117 *Steve Henson*
6118
6119 * If server EC tmp key is not in client preference list abort handshake.
6120
6121 *Steve Henson*
6122
6123 * Add support for certificate stores in CERT structure. This makes it
6124 possible to have different stores per SSL structure or one store in
6125 the parent SSL_CTX. Include distinct stores for certificate chain
6126 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6127 to build and store a certificate chain in CERT structure: returning
6128 an error if the chain cannot be built: this will allow applications
6129 to test if a chain is correctly configured.
6130
6131 Note: if the CERT based stores are not set then the parent SSL_CTX
6132 store is used to retain compatibility with existing behaviour.
6133
44652c16
DMSP
6134 *Steve Henson*
6135
6136 * New function ssl_set_client_disabled to set a ciphersuite disabled
6137 mask based on the current session, check mask when sending client
6138 hello and checking the requested ciphersuite.
6139
6140 *Steve Henson*
6141
6142 * New ctrls to retrieve and set certificate types in a certificate
6143 request message. Print out received values in s_client. If certificate
6144 types is not set with custom values set sensible values based on
6145 supported signature algorithms.
6146
6147 *Steve Henson*
6148
6149 * Support for distinct client and server supported signature algorithms.
6150
6151 *Steve Henson*
6152
6153 * Add certificate callback. If set this is called whenever a certificate
6154 is required by client or server. An application can decide which
6155 certificate chain to present based on arbitrary criteria: for example
6156 supported signature algorithms. Add very simple example to s_server.
6157 This fixes many of the problems and restrictions of the existing client
6158 certificate callback: for example you can now clear an existing
6159 certificate and specify the whole chain.
6160
6161 *Steve Henson*
6162
6163 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6164 the certificate can be used for (if anything). Set valid_flags field
6165 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6166 to have similar checks in it.
6167
6168 Add new "cert_flags" field to CERT structure and include a "strict mode".
6169 This enforces some TLS certificate requirements (such as only permitting
6170 certificate signature algorithms contained in the supported algorithms
6171 extension) which some implementations ignore: this option should be used
6172 with caution as it could cause interoperability issues.
6173
6174 *Steve Henson*
6175
6176 * Update and tidy signature algorithm extension processing. Work out
6177 shared signature algorithms based on preferences and peer algorithms
6178 and print them out in s_client and s_server. Abort handshake if no
6179 shared signature algorithms.
6180
6181 *Steve Henson*
6182
6183 * Add new functions to allow customised supported signature algorithms
6184 for SSL and SSL_CTX structures. Add options to s_client and s_server
6185 to support them.
6186
6187 *Steve Henson*
6188
6189 * New function SSL_certs_clear() to delete all references to certificates
6190 from an SSL structure. Before this once a certificate had been added
6191 it couldn't be removed.
6192
6193 *Steve Henson*
6194
6195 * Integrate hostname, email address and IP address checking with certificate
6196 verification. New verify options supporting checking in openssl utility.
6197
6198 *Steve Henson*
6199
6200 * Fixes and wildcard matching support to hostname and email checking
6201 functions. Add manual page.
6202
6203 *Florian Weimer (Red Hat Product Security Team)*
6204
6205 * New functions to check a hostname email or IP address against a
6206 certificate. Add options x509 utility to print results of checks against
6207 a certificate.
6208
6209 *Steve Henson*
6210
6211 * Fix OCSP checking.
6212
6213 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6214
6215 * Initial experimental support for explicitly trusted non-root CAs.
6216 OpenSSL still tries to build a complete chain to a root but if an
6217 intermediate CA has a trust setting included that is used. The first
6218 setting is used: whether to trust (e.g., -addtrust option to the x509
6219 utility) or reject.
6220
6221 *Steve Henson*
6222
6223 * Add -trusted_first option which attempts to find certificates in the
6224 trusted store even if an untrusted chain is also supplied.
6225
6226 *Steve Henson*
6227
6228 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6229 platform support for Linux and Android.
6230
6231 *Andy Polyakov*
6232
6233 * Support for linux-x32, ILP32 environment in x86_64 framework.
6234
6235 *Andy Polyakov*
6236
6237 * Experimental multi-implementation support for FIPS capable OpenSSL.
6238 When in FIPS mode the approved implementations are used as normal,
6239 when not in FIPS mode the internal unapproved versions are used instead.
6240 This means that the FIPS capable OpenSSL isn't forced to use the
6241 (often lower performance) FIPS implementations outside FIPS mode.
6242
6243 *Steve Henson*
6244
6245 * Transparently support X9.42 DH parameters when calling
6246 PEM_read_bio_DHparameters. This means existing applications can handle
6247 the new parameter format automatically.
6248
6249 *Steve Henson*
6250
6251 * Initial experimental support for X9.42 DH parameter format: mainly
6252 to support use of 'q' parameter for RFC5114 parameters.
6253
6254 *Steve Henson*
6255
6256 * Add DH parameters from RFC5114 including test data to dhtest.
6257
6258 *Steve Henson*
6259
6260 * Support for automatic EC temporary key parameter selection. If enabled
6261 the most preferred EC parameters are automatically used instead of
6262 hardcoded fixed parameters. Now a server just has to call:
6263 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6264 support ECDH and use the most appropriate parameters.
6265
6266 *Steve Henson*
6267
6268 * Enhance and tidy EC curve and point format TLS extension code. Use
6269 static structures instead of allocation if default values are used.
6270 New ctrls to set curves we wish to support and to retrieve shared curves.
6271 Print out shared curves in s_server. New options to s_server and s_client
6272 to set list of supported curves.
6273
6274 *Steve Henson*
6275
6276 * New ctrls to retrieve supported signature algorithms and
6277 supported curve values as an array of NIDs. Extend openssl utility
6278 to print out received values.
6279
6280 *Steve Henson*
6281
6282 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6283 between NIDs and the more common NIST names such as "P-256". Enhance
6284 ecparam utility and ECC method to recognise the NIST names for curves.
6285
6286 *Steve Henson*
6287
6288 * Enhance SSL/TLS certificate chain handling to support different
6289 chains for each certificate instead of one chain in the parent SSL_CTX.
6290
6291 *Steve Henson*
6292
6293 * Support for fixed DH ciphersuite client authentication: where both
6294 server and client use DH certificates with common parameters.
6295
6296 *Steve Henson*
6297
6298 * Support for fixed DH ciphersuites: those requiring DH server
6299 certificates.
6300
6301 *Steve Henson*
6302
6303 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6304 the certificate.
6305 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6306 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6307 X509_CINF_get_signature were reverted post internal team review.
6308
44652c16
DMSP
6309OpenSSL 1.0.1
6310-------------
6311
257e9d03 6312### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
44652c16
DMSP
6313
6314 * OCSP Status Request extension unbounded memory growth
6315
6316 A malicious client can send an excessively large OCSP Status Request
6317 extension. If that client continually requests renegotiation, sending a
6318 large OCSP Status Request extension each time, then there will be unbounded
6319 memory growth on the server. This will eventually lead to a Denial Of
6320 Service attack through memory exhaustion. Servers with a default
6321 configuration are vulnerable even if they do not support OCSP. Builds using
6322 the "no-ocsp" build time option are not affected.
6323
6324 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
d8dc8538 6325 ([CVE-2016-6304])
44652c16
DMSP
6326
6327 *Matt Caswell*
6328
6329 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6330 HIGH to MEDIUM.
6331
6332 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6333 Leurent (INRIA)
d8dc8538 6334 ([CVE-2016-2183])
44652c16
DMSP
6335
6336 *Rich Salz*
6337
6338 * OOB write in MDC2_Update()
6339
6340 An overflow can occur in MDC2_Update() either if called directly or
6341 through the EVP_DigestUpdate() function using MDC2. If an attacker
6342 is able to supply very large amounts of input data after a previous
6343 call to EVP_EncryptUpdate() with a partial block then a length check
6344 can overflow resulting in a heap corruption.
6345
6346 The amount of data needed is comparable to SIZE_MAX which is impractical
6347 on most platforms.
6348
6349 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
d8dc8538 6350 ([CVE-2016-6303])
44652c16
DMSP
6351
6352 *Stephen Henson*
6353
6354 * Malformed SHA512 ticket DoS
6355
6356 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6357 DoS attack where a malformed ticket will result in an OOB read which will
6358 ultimately crash.
6359
6360 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6361 a custom server callback and ticket lookup mechanism.
6362
6363 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
d8dc8538 6364 ([CVE-2016-6302])
44652c16
DMSP
6365
6366 *Stephen Henson*
6367
6368 * OOB write in BN_bn2dec()
6369
6370 The function BN_bn2dec() does not check the return value of BN_div_word().
6371 This can cause an OOB write if an application uses this function with an
6372 overly large BIGNUM. This could be a problem if an overly large certificate
6373 or CRL is printed out from an untrusted source. TLS is not affected because
6374 record limits will reject an oversized certificate before it is parsed.
6375
6376 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
d8dc8538 6377 ([CVE-2016-2182])
44652c16
DMSP
6378
6379 *Stephen Henson*
6380
6381 * OOB read in TS_OBJ_print_bio()
6382
6383 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6384 the total length the OID text representation would use and not the amount
6385 of data written. This will result in OOB reads when large OIDs are
6386 presented.
6387
6388 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
d8dc8538 6389 ([CVE-2016-2180])
44652c16
DMSP
6390
6391 *Stephen Henson*
6392
6393 * Pointer arithmetic undefined behaviour
6394
6395 Avoid some undefined pointer arithmetic
6396
6397 A common idiom in the codebase is to check limits in the following manner:
6398 "p + len > limit"
6399
6400 Where "p" points to some malloc'd data of SIZE bytes and
6401 limit == p + SIZE
6402
6403 "len" here could be from some externally supplied data (e.g. from a TLS
6404 message).
6405
6406 The rules of C pointer arithmetic are such that "p + len" is only well
6407 defined where len <= SIZE. Therefore the above idiom is actually
6408 undefined behaviour.
6409
6410 For example this could cause problems if some malloc implementation
6411 provides an address for "p" such that "p + len" actually overflows for
6412 values of len that are too big and therefore p + len < limit.
6413
6414 This issue was reported to OpenSSL by Guido Vranken
d8dc8538 6415 ([CVE-2016-2177])
44652c16
DMSP
6416
6417 *Matt Caswell*
6418
6419 * Constant time flag not preserved in DSA signing
6420
6421 Operations in the DSA signing algorithm should run in constant time in
6422 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6423 implementation means that a non-constant time codepath is followed for
6424 certain operations. This has been demonstrated through a cache-timing
6425 attack to be sufficient for an attacker to recover the private DSA key.
6426
6427 This issue was reported by César Pereida (Aalto University), Billy Brumley
6428 (Tampere University of Technology), and Yuval Yarom (The University of
6429 Adelaide and NICTA).
d8dc8538 6430 ([CVE-2016-2178])
44652c16
DMSP
6431
6432 *César Pereida*
6433
6434 * DTLS buffered message DoS
6435
6436 In a DTLS connection where handshake messages are delivered out-of-order
6437 those messages that OpenSSL is not yet ready to process will be buffered
6438 for later use. Under certain circumstances, a flaw in the logic means that
6439 those messages do not get removed from the buffer even though the handshake
6440 has been completed. An attacker could force up to approx. 15 messages to
6441 remain in the buffer when they are no longer required. These messages will
6442 be cleared when the DTLS connection is closed. The default maximum size for
6443 a message is 100k. Therefore the attacker could force an additional 1500k
6444 to be consumed per connection. By opening many simulataneous connections an
6445 attacker could cause a DoS attack through memory exhaustion.
6446
6447 This issue was reported to OpenSSL by Quan Luo.
d8dc8538 6448 ([CVE-2016-2179])
44652c16
DMSP
6449
6450 *Matt Caswell*
6451
6452 * DTLS replay protection DoS
6453
6454 A flaw in the DTLS replay attack protection mechanism means that records
6455 that arrive for future epochs update the replay protection "window" before
6456 the MAC for the record has been validated. This could be exploited by an
6457 attacker by sending a record for the next epoch (which does not have to
6458 decrypt or have a valid MAC), with a very large sequence number. This means
6459 that all subsequent legitimate packets are dropped causing a denial of
6460 service for a specific DTLS connection.
6461
6462 This issue was reported to OpenSSL by the OCAP audit team.
d8dc8538 6463 ([CVE-2016-2181])
44652c16
DMSP
6464
6465 *Matt Caswell*
6466
6467 * Certificate message OOB reads
6468
6469 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6470 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6471 theoretical DoS risk but this has not been observed in practice on common
6472 platforms.
6473
6474 The messages affected are client certificate, client certificate request
6475 and server certificate. As a result the attack can only be performed
6476 against a client or a server which enables client authentication.
6477
6478 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
d8dc8538 6479 ([CVE-2016-6306])
44652c16
DMSP
6480
6481 *Stephen Henson*
6482
257e9d03 6483### Changes between 1.0.1s and 1.0.1t [3 May 2016]
44652c16
DMSP
6484
6485 * Prevent padding oracle in AES-NI CBC MAC check
6486
6487 A MITM attacker can use a padding oracle attack to decrypt traffic
6488 when the connection uses an AES CBC cipher and the server support
6489 AES-NI.
6490
6491 This issue was introduced as part of the fix for Lucky 13 padding
d8dc8538 6492 attack ([CVE-2013-0169]). The padding check was rewritten to be in
44652c16
DMSP
6493 constant time by making sure that always the same bytes are read and
6494 compared against either the MAC or padding bytes. But it no longer
6495 checked that there was enough data to have both the MAC and padding
6496 bytes.
6497
6498 This issue was reported by Juraj Somorovsky using TLS-Attacker.
d8dc8538 6499 ([CVE-2016-2107])
44652c16
DMSP
6500
6501 *Kurt Roeckx*
6502
6503 * Fix EVP_EncodeUpdate overflow
6504
6505 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6506 Base64 encoding of binary data. If an attacker is able to supply very large
6507 amounts of input data then a length check can overflow resulting in a heap
6508 corruption.
6509
6510 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
257e9d03 6511 the `PEM_write_bio*` family of functions. These are mainly used within the
44652c16
DMSP
6512 OpenSSL command line applications, so any application which processes data
6513 from an untrusted source and outputs it as a PEM file should be considered
6514 vulnerable to this issue. User applications that call these APIs directly
6515 with large amounts of untrusted data may also be vulnerable.
6516
6517 This issue was reported by Guido Vranken.
d8dc8538 6518 ([CVE-2016-2105])
44652c16
DMSP
6519
6520 *Matt Caswell*
6521
6522 * Fix EVP_EncryptUpdate overflow
6523
6524 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6525 is able to supply very large amounts of input data after a previous call to
6526 EVP_EncryptUpdate() with a partial block then a length check can overflow
6527 resulting in a heap corruption. Following an analysis of all OpenSSL
6528 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6529 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6530 the first called function after an EVP_EncryptInit(), and therefore that
6531 specific call must be safe. The second form is where the length passed to
6532 EVP_EncryptUpdate() can be seen from the code to be some small value and
6533 therefore there is no possibility of an overflow. Since all instances are
6534 one of these two forms, it is believed that there can be no overflows in
6535 internal code due to this problem. It should be noted that
6536 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6537 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6538 of these calls have also been analysed too and it is believed there are no
6539 instances in internal usage where an overflow could occur.
6540
6541 This issue was reported by Guido Vranken.
d8dc8538 6542 ([CVE-2016-2106])
44652c16
DMSP
6543
6544 *Matt Caswell*
6545
6546 * Prevent ASN.1 BIO excessive memory allocation
6547
6548 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6549 a short invalid encoding can casuse allocation of large amounts of memory
6550 potentially consuming excessive resources or exhausting memory.
6551
6552 Any application parsing untrusted data through d2i BIO functions is
6553 affected. The memory based functions such as d2i_X509() are *not* affected.
6554 Since the memory based functions are used by the TLS library, TLS
6555 applications are not affected.
6556
6557 This issue was reported by Brian Carpenter.
d8dc8538 6558 ([CVE-2016-2109])
44652c16
DMSP
6559
6560 *Stephen Henson*
6561
6562 * EBCDIC overread
6563
6564 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6565 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6566 in arbitrary stack data being returned in the buffer.
6567
6568 This issue was reported by Guido Vranken.
d8dc8538 6569 ([CVE-2016-2176])
44652c16
DMSP
6570
6571 *Matt Caswell*
6572
6573 * Modify behavior of ALPN to invoke callback after SNI/servername
6574 callback, such that updates to the SSL_CTX affect ALPN.
6575
6576 *Todd Short*
6577
6578 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6579 default.
6580
6581 *Kurt Roeckx*
6582
6583 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6584 methods are enabled and ssl2 is disabled the methods return NULL.
6585
6586 *Kurt Roeckx*
6587
257e9d03 6588### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
44652c16
DMSP
6589
6590* Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6591 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6592 provide any "EXPORT" or "LOW" strength ciphers.
6593
6594 *Viktor Dukhovni*
6595
6596* Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6597 is by default disabled at build-time. Builds that are not configured with
6598 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6599 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6600 will need to explicitly call either of:
6601
6602 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6603 or
6604 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6605
6606 as appropriate. Even if either of those is used, or the application
6607 explicitly uses the version-specific SSLv2_method() or its client and
6608 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6609 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6610 ciphers, and SSLv2 56-bit DES are no longer available.
d8dc8538 6611 ([CVE-2016-0800])
44652c16
DMSP
6612
6613 *Viktor Dukhovni*
6614
6615 * Fix a double-free in DSA code
6616
6617 A double free bug was discovered when OpenSSL parses malformed DSA private
6618 keys and could lead to a DoS attack or memory corruption for applications
6619 that receive DSA private keys from untrusted sources. This scenario is
6620 considered rare.
6621
6622 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6623 libFuzzer.
d8dc8538 6624 ([CVE-2016-0705])
44652c16
DMSP
6625
6626 *Stephen Henson*
6627
6628 * Disable SRP fake user seed to address a server memory leak.
6629
6630 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6631
6632 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6633 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6634 was changed to ignore the "fake user" SRP seed, even if the seed
6635 is configured.
6636
6637 Users should use SRP_VBASE_get1_by_user instead. Note that in
6638 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6639 also that even though configuring the SRP seed attempts to hide
6640 invalid usernames by continuing the handshake with fake
6641 credentials, this behaviour is not constant time and no strong
6642 guarantees are made that the handshake is indistinguishable from
6643 that of a valid user.
d8dc8538 6644 ([CVE-2016-0798])
44652c16
DMSP
6645
6646 *Emilia Käsper*
6647
6648 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6649
6650 In the BN_hex2bn function the number of hex digits is calculated using an
1dc1ea18
DDO
6651 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6652 large values of `i` this can result in `bn_expand` not allocating any
6653 memory because `i * 4` is negative. This can leave the internal BIGNUM data
44652c16 6654 field as NULL leading to a subsequent NULL ptr deref. For very large values
1dc1ea18 6655 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
44652c16
DMSP
6656 In this case memory is allocated to the internal BIGNUM data field, but it
6657 is insufficiently sized leading to heap corruption. A similar issue exists
6658 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6659 is ever called by user applications with very large untrusted hex/dec data.
6660 This is anticipated to be a rare occurrence.
6661
6662 All OpenSSL internal usage of these functions use data that is not expected
6663 to be untrusted, e.g. config file data or application command line
6664 arguments. If user developed applications generate config file data based
6665 on untrusted data then it is possible that this could also lead to security
6666 consequences. This is also anticipated to be rare.
6667
6668 This issue was reported to OpenSSL by Guido Vranken.
d8dc8538 6669 ([CVE-2016-0797])
44652c16
DMSP
6670
6671 *Matt Caswell*
6672
257e9d03 6673 * Fix memory issues in `BIO_*printf` functions
44652c16 6674
1dc1ea18 6675 The internal `fmtstr` function used in processing a "%s" format string in
257e9d03 6676 the `BIO_*printf` functions could overflow while calculating the length of a
44652c16
DMSP
6677 string and cause an OOB read when printing very long strings.
6678
1dc1ea18 6679 Additionally the internal `doapr_outch` function can attempt to write to an
44652c16
DMSP
6680 OOB memory location (at an offset from the NULL pointer) in the event of a
6681 memory allocation failure. In 1.0.2 and below this could be caused where
6682 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6683 could be in processing a very long "%s" format string. Memory leaks can
6684 also occur.
6685
6686 The first issue may mask the second issue dependent on compiler behaviour.
6687 These problems could enable attacks where large amounts of untrusted data
257e9d03 6688 is passed to the `BIO_*printf` functions. If applications use these functions
44652c16
DMSP
6689 in this way then they could be vulnerable. OpenSSL itself uses these
6690 functions when printing out human-readable dumps of ASN.1 data. Therefore
6691 applications that print this data could be vulnerable if the data is from
6692 untrusted sources. OpenSSL command line applications could also be
6693 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6694 as command line arguments.
6695
6696 Libssl is not considered directly vulnerable. Additionally certificates etc
6697 received via remote connections via libssl are also unlikely to be able to
6698 trigger these issues because of message size limits enforced within libssl.
6699
6700 This issue was reported to OpenSSL Guido Vranken.
d8dc8538 6701 ([CVE-2016-0799])
44652c16
DMSP
6702
6703 *Matt Caswell*
6704
6705 * Side channel attack on modular exponentiation
6706
6707 A side-channel attack was found which makes use of cache-bank conflicts on
6708 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6709 of RSA keys. The ability to exploit this issue is limited as it relies on
6710 an attacker who has control of code in a thread running on the same
6711 hyper-threaded core as the victim thread which is performing decryptions.
6712
6713 This issue was reported to OpenSSL by Yuval Yarom, The University of
6714 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6715 Nadia Heninger, University of Pennsylvania with more information at
257e9d03 6716 <http://cachebleed.info>.
d8dc8538 6717 ([CVE-2016-0702])
44652c16
DMSP
6718
6719 *Andy Polyakov*
6720
ec2bfb7d 6721 * Change the req command to generate a 2048-bit RSA/DSA key by default,
44652c16
DMSP
6722 if no keysize is specified with default_bits. This fixes an
6723 omission in an earlier change that changed all RSA/DSA key generation
ec2bfb7d 6724 commands to use 2048 bits by default.
44652c16
DMSP
6725
6726 *Emilia Käsper*
6727
257e9d03 6728### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
44652c16
DMSP
6729
6730 * Protection for DH small subgroup attacks
6731
6732 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6733 switched on by default and cannot be disabled. This could have some
6734 performance impact.
6735
6736 *Matt Caswell*
6737
6738 * SSLv2 doesn't block disabled ciphers
6739
6740 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6741 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6742 been disabled, provided that the SSLv2 protocol was not also disabled via
6743 SSL_OP_NO_SSLv2.
6744
6745 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6746 and Sebastian Schinzel.
d8dc8538 6747 ([CVE-2015-3197])
44652c16
DMSP
6748
6749 *Viktor Dukhovni*
6750
6751 * Reject DH handshakes with parameters shorter than 1024 bits.
6752
6753 *Kurt Roeckx*
6754
257e9d03 6755### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
44652c16
DMSP
6756
6757 * Certificate verify crash with missing PSS parameter
6758
6759 The signature verification routines will crash with a NULL pointer
6760 dereference if presented with an ASN.1 signature using the RSA PSS
6761 algorithm and absent mask generation function parameter. Since these
6762 routines are used to verify certificate signature algorithms this can be
6763 used to crash any certificate verification operation and exploited in a
6764 DoS attack. Any application which performs certificate verification is
6765 vulnerable including OpenSSL clients and servers which enable client
6766 authentication.
6767
6768 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
d8dc8538 6769 ([CVE-2015-3194])
44652c16
DMSP
6770
6771 *Stephen Henson*
6772
6773 * X509_ATTRIBUTE memory leak
6774
6775 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6776 memory. This structure is used by the PKCS#7 and CMS routines so any
6777 application which reads PKCS#7 or CMS data from untrusted sources is
6778 affected. SSL/TLS is not affected.
6779
6780 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6781 libFuzzer.
d8dc8538 6782 ([CVE-2015-3195])
44652c16
DMSP
6783
6784 *Stephen Henson*
6785
6786 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6787 This changes the decoding behaviour for some invalid messages,
6788 though the change is mostly in the more lenient direction, and
6789 legacy behaviour is preserved as much as possible.
6790
6791 *Emilia Käsper*
6792
6793 * In DSA_generate_parameters_ex, if the provided seed is too short,
6794 use a random seed, as already documented.
6795
6796 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6797
257e9d03 6798### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
44652c16
DMSP
6799
6800 * Alternate chains certificate forgery
6801
6802 During certificate verfification, OpenSSL will attempt to find an
6803 alternative certificate chain if the first attempt to build such a chain
6804 fails. An error in the implementation of this logic can mean that an
6805 attacker could cause certain checks on untrusted certificates to be
6806 bypassed, such as the CA flag, enabling them to use a valid leaf
6807 certificate to act as a CA and "issue" an invalid certificate.
6808
6809 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6810 (Google/BoringSSL).
d8dc8538 6811 ([CVE-2015-1793])
44652c16
DMSP
6812
6813 *Matt Caswell*
6814
6815 * Race condition handling PSK identify hint
6816
6817 If PSK identity hints are received by a multi-threaded client then
6818 the values are wrongly updated in the parent SSL_CTX structure. This can
6819 result in a race condition potentially leading to a double free of the
6820 identify hint data.
d8dc8538 6821 ([CVE-2015-3196])
44652c16
DMSP
6822
6823 *Stephen Henson*
6824
257e9d03
RS
6825### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6826
44652c16
DMSP
6827 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6828 incompatibility in the handling of HMAC. The previous ABI has now been
6829 restored.
6830
257e9d03 6831### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
44652c16
DMSP
6832
6833 * Malformed ECParameters causes infinite loop
6834
6835 When processing an ECParameters structure OpenSSL enters an infinite loop
6836 if the curve specified is over a specially malformed binary polynomial
6837 field.
6838
6839 This can be used to perform denial of service against any
6840 system which processes public keys, certificate requests or
6841 certificates. This includes TLS clients and TLS servers with
6842 client authentication enabled.
6843
6844 This issue was reported to OpenSSL by Joseph Barr-Pixton.
d8dc8538 6845 ([CVE-2015-1788])
44652c16
DMSP
6846
6847 *Andy Polyakov*
6848
6849 * Exploitable out-of-bounds read in X509_cmp_time
6850
6851 X509_cmp_time does not properly check the length of the ASN1_TIME
6852 string and can read a few bytes out of bounds. In addition,
6853 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6854 time string.
6855
6856 An attacker can use this to craft malformed certificates and CRLs of
6857 various sizes and potentially cause a segmentation fault, resulting in
6858 a DoS on applications that verify certificates or CRLs. TLS clients
6859 that verify CRLs are affected. TLS clients and servers with client
6860 authentication enabled may be affected if they use custom verification
6861 callbacks.
6862
6863 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6864 independently by Hanno Böck.
d8dc8538 6865 ([CVE-2015-1789])
44652c16
DMSP
6866
6867 *Emilia Käsper*
6868
6869 * PKCS7 crash with missing EnvelopedContent
6870
6871 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6872 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6873 with missing content and trigger a NULL pointer dereference on parsing.
6874
6875 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6876 structures from untrusted sources are affected. OpenSSL clients and
6877 servers are not affected.
5f8e6c50 6878
44652c16 6879 This issue was reported to OpenSSL by Michal Zalewski (Google).
d8dc8538 6880 ([CVE-2015-1790])
5f8e6c50 6881
44652c16 6882 *Emilia Käsper*
5f8e6c50 6883
44652c16
DMSP
6884 * CMS verify infinite loop with unknown hash function
6885
6886 When verifying a signedData message the CMS code can enter an infinite loop
6887 if presented with an unknown hash function OID. This can be used to perform
6888 denial of service against any system which verifies signedData messages using
6889 the CMS code.
6890 This issue was reported to OpenSSL by Johannes Bauer.
d8dc8538 6891 ([CVE-2015-1792])
44652c16
DMSP
6892
6893 *Stephen Henson*
6894
6895 * Race condition handling NewSessionTicket
6896
6897 If a NewSessionTicket is received by a multi-threaded client when attempting to
6898 reuse a previous ticket then a race condition can occur potentially leading to
6899 a double free of the ticket data.
d8dc8538 6900 ([CVE-2015-1791])
44652c16
DMSP
6901
6902 *Matt Caswell*
6903
6904 * Reject DH handshakes with parameters shorter than 768 bits.
6905
6906 *Kurt Roeckx and Emilia Kasper*
6907
6908 * dhparam: generate 2048-bit parameters by default.
6909
6910 *Kurt Roeckx and Emilia Kasper*
6911
257e9d03 6912### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
44652c16
DMSP
6913
6914 * Segmentation fault in ASN1_TYPE_cmp fix
6915
6916 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6917 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6918 certificate signature algorithm consistency this can be used to crash any
6919 certificate verification operation and exploited in a DoS attack. Any
6920 application which performs certificate verification is vulnerable including
6921 OpenSSL clients and servers which enable client authentication.
d8dc8538 6922 ([CVE-2015-0286])
44652c16
DMSP
6923
6924 *Stephen Henson*
6925
6926 * ASN.1 structure reuse memory corruption fix
6927
6928 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6929 memory corruption via an invalid write. Such reuse is and has been
6930 strongly discouraged and is believed to be rare.
6931
6932 Applications that parse structures containing CHOICE or ANY DEFINED BY
6933 components may be affected. Certificate parsing (d2i_X509 and related
6934 functions) are however not affected. OpenSSL clients and servers are
6935 not affected.
d8dc8538 6936 ([CVE-2015-0287])
44652c16
DMSP
6937
6938 *Stephen Henson*
6939
6940 * PKCS7 NULL pointer dereferences fix
6941
6942 The PKCS#7 parsing code does not handle missing outer ContentInfo
6943 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6944 missing content and trigger a NULL pointer dereference on parsing.
6945
6946 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6947 otherwise parse PKCS#7 structures from untrusted sources are
6948 affected. OpenSSL clients and servers are not affected.
6949
6950 This issue was reported to OpenSSL by Michal Zalewski (Google).
d8dc8538 6951 ([CVE-2015-0289])
44652c16
DMSP
6952
6953 *Emilia Käsper*
6954
6955 * DoS via reachable assert in SSLv2 servers fix
6956
6957 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6958 servers that both support SSLv2 and enable export cipher suites by sending
6959 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6960
6961 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6962 (OpenSSL development team).
d8dc8538 6963 ([CVE-2015-0293])
44652c16
DMSP
6964
6965 *Emilia Käsper*
6966
6967 * Use After Free following d2i_ECPrivatekey error fix
6968
6969 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6970 could cause a use after free condition. This, in turn, could cause a double
6971 free in several private key parsing functions (such as d2i_PrivateKey
6972 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6973 for applications that receive EC private keys from untrusted
6974 sources. This scenario is considered rare.
6975
6976 This issue was discovered by the BoringSSL project and fixed in their
6977 commit 517073cd4b.
d8dc8538 6978 ([CVE-2015-0209])
44652c16
DMSP
6979
6980 *Matt Caswell*
6981
6982 * X509_to_X509_REQ NULL pointer deref fix
6983
6984 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6985 the certificate key is invalid. This function is rarely used in practice.
6986
6987 This issue was discovered by Brian Carpenter.
d8dc8538 6988 ([CVE-2015-0288])
44652c16
DMSP
6989
6990 *Stephen Henson*
6991
6992 * Removed the export ciphers from the DEFAULT ciphers
6993
6994 *Kurt Roeckx*
6995
257e9d03 6996### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
44652c16
DMSP
6997
6998 * Build fixes for the Windows and OpenVMS platforms
6999
7000 *Matt Caswell and Richard Levitte*
7001
257e9d03 7002### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
44652c16
DMSP
7003
7004 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7005 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7006 dereference. This could lead to a Denial Of Service attack. Thanks to
7007 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
d8dc8538 7008 ([CVE-2014-3571])
44652c16
DMSP
7009
7010 *Steve Henson*
7011
7012 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7013 dtls1_buffer_record function under certain conditions. In particular this
7014 could occur if an attacker sent repeated DTLS records with the same
7015 sequence number but for the next epoch. The memory leak could be exploited
7016 by an attacker in a Denial of Service attack through memory exhaustion.
7017 Thanks to Chris Mueller for reporting this issue.
d8dc8538 7018 ([CVE-2015-0206])
44652c16
DMSP
7019
7020 *Matt Caswell*
7021
7022 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7023 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7024 method would be set to NULL which could later result in a NULL pointer
7025 dereference. Thanks to Frank Schmirler for reporting this issue.
d8dc8538 7026 ([CVE-2014-3569])
44652c16
DMSP
7027
7028 *Kurt Roeckx*
7029
7030 * Abort handshake if server key exchange message is omitted for ephemeral
7031 ECDH ciphersuites.
7032
7033 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7034 reporting this issue.
d8dc8538 7035 ([CVE-2014-3572])
44652c16
DMSP
7036
7037 *Steve Henson*
7038
7039 * Remove non-export ephemeral RSA code on client and server. This code
7040 violated the TLS standard by allowing the use of temporary RSA keys in
7041 non-export ciphersuites and could be used by a server to effectively
7042 downgrade the RSA key length used to a value smaller than the server
7043 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7044 INRIA or reporting this issue.
d8dc8538 7045 ([CVE-2015-0204])
44652c16
DMSP
7046
7047 *Steve Henson*
7048
7049 * Fixed issue where DH client certificates are accepted without verification.
7050 An OpenSSL server will accept a DH certificate for client authentication
7051 without the certificate verify message. This effectively allows a client to
7052 authenticate without the use of a private key. This only affects servers
7053 which trust a client certificate authority which issues certificates
7054 containing DH keys: these are extremely rare and hardly ever encountered.
7055 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7056 this issue.
d8dc8538 7057 ([CVE-2015-0205])
44652c16
DMSP
7058
7059 *Steve Henson*
7060
7061 * Ensure that the session ID context of an SSL is updated when its
7062 SSL_CTX is updated via SSL_set_SSL_CTX.
7063
7064 The session ID context is typically set from the parent SSL_CTX,
7065 and can vary with the CTX.
7066
7067 *Adam Langley*
7068
7069 * Fix various certificate fingerprint issues.
7070
7071 By using non-DER or invalid encodings outside the signed portion of a
7072 certificate the fingerprint can be changed without breaking the signature.
7073 Although no details of the signed portion of the certificate can be changed
7074 this can cause problems with some applications: e.g. those using the
7075 certificate fingerprint for blacklists.
7076
7077 1. Reject signatures with non zero unused bits.
7078
7079 If the BIT STRING containing the signature has non zero unused bits reject
7080 the signature. All current signature algorithms require zero unused bits.
7081
7082 2. Check certificate algorithm consistency.
7083
7084 Check the AlgorithmIdentifier inside TBS matches the one in the
7085 certificate signature. NB: this will result in signature failure
7086 errors for some broken certificates.
7087
7088 Thanks to Konrad Kraszewski from Google for reporting this issue.
7089
7090 3. Check DSA/ECDSA signatures use DER.
7091
7092 Re-encode DSA/ECDSA signatures and compare with the original received
7093 signature. Return an error if there is a mismatch.
7094
7095 This will reject various cases including garbage after signature
7096 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7097 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7098 (negative or with leading zeroes).
7099
7100 Further analysis was conducted and fixes were developed by Stephen Henson
7101 of the OpenSSL core team.
7102
d8dc8538 7103 ([CVE-2014-8275])
44652c16
DMSP
7104
7105 *Steve Henson*
7106
43a70f02
RS
7107 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7108 results on some platforms, including x86_64. This bug occurs at random
7109 with a very low probability, and is not known to be exploitable in any
7110 way, though its exact impact is difficult to determine. Thanks to Pieter
7111 Wuille (Blockstream) who reported this issue and also suggested an initial
7112 fix. Further analysis was conducted by the OpenSSL development team and
7113 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7114 the OpenSSL core team.
d8dc8538 7115 ([CVE-2014-3570])
5f8e6c50
DMSP
7116
7117 *Andy Polyakov*
7118
43a70f02
RS
7119 * Do not resume sessions on the server if the negotiated protocol
7120 version does not match the session's version. Resuming with a different
7121 version, while not strictly forbidden by the RFC, is of questionable
7122 sanity and breaks all known clients.
5f8e6c50 7123
44652c16
DMSP
7124 *David Benjamin, Emilia Käsper*
7125
43a70f02
RS
7126 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7127 early CCS messages during renegotiation. (Note that because
7128 renegotiation is encrypted, this early CCS was not exploitable.)
44652c16
DMSP
7129
7130 *Emilia Käsper*
7131
43a70f02
RS
7132 * Tighten client-side session ticket handling during renegotiation:
7133 ensure that the client only accepts a session ticket if the server sends
7134 the extension anew in the ServerHello. Previously, a TLS client would
7135 reuse the old extension state and thus accept a session ticket if one was
7136 announced in the initial ServerHello.
44652c16 7137
43a70f02
RS
7138 Similarly, ensure that the client requires a session ticket if one
7139 was advertised in the ServerHello. Previously, a TLS client would
7140 ignore a missing NewSessionTicket message.
44652c16
DMSP
7141
7142 *Emilia Käsper*
7143
257e9d03 7144### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
44652c16
DMSP
7145
7146 * SRTP Memory Leak.
7147
7148 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7149 sends a carefully crafted handshake message, to cause OpenSSL to fail
7150 to free up to 64k of memory causing a memory leak. This could be
7151 exploited in a Denial Of Service attack. This issue affects OpenSSL
7152 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7153 whether SRTP is used or configured. Implementations of OpenSSL that
7154 have been compiled with OPENSSL_NO_SRTP defined are not affected.
5f8e6c50 7155
44652c16 7156 The fix was developed by the OpenSSL team.
d8dc8538 7157 ([CVE-2014-3513])
5f8e6c50 7158
44652c16 7159 *OpenSSL team*
5f8e6c50 7160
44652c16 7161 * Session Ticket Memory Leak.
5f8e6c50 7162
44652c16
DMSP
7163 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7164 integrity of that ticket is first verified. In the event of a session
7165 ticket integrity check failing, OpenSSL will fail to free memory
7166 causing a memory leak. By sending a large number of invalid session
7167 tickets an attacker could exploit this issue in a Denial Of Service
7168 attack.
d8dc8538 7169 ([CVE-2014-3567])
5f8e6c50 7170
44652c16 7171 *Steve Henson*
5f8e6c50 7172
44652c16 7173 * Build option no-ssl3 is incomplete.
5f8e6c50 7174
44652c16
DMSP
7175 When OpenSSL is configured with "no-ssl3" as a build option, servers
7176 could accept and complete a SSL 3.0 handshake, and clients could be
7177 configured to send them.
d8dc8538 7178 ([CVE-2014-3568])
5f8e6c50 7179
44652c16 7180 *Akamai and the OpenSSL team*
5f8e6c50 7181
44652c16
DMSP
7182 * Add support for TLS_FALLBACK_SCSV.
7183 Client applications doing fallback retries should call
7184 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
d8dc8538 7185 ([CVE-2014-3566])
5f8e6c50 7186
44652c16 7187 *Adam Langley, Bodo Moeller*
5f8e6c50 7188
44652c16 7189 * Add additional DigestInfo checks.
5f8e6c50 7190
44652c16
DMSP
7191 Re-encode DigestInto in DER and check against the original when
7192 verifying RSA signature: this will reject any improperly encoded
7193 DigestInfo structures.
5f8e6c50 7194
44652c16 7195 Note: this is a precautionary measure and no attacks are currently known.
5f8e6c50 7196
5f8e6c50
DMSP
7197 *Steve Henson*
7198
257e9d03 7199### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
5f8e6c50 7200
44652c16
DMSP
7201 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7202 SRP code can be overrun an internal buffer. Add sanity check that
7203 g, A, B < N to SRP code.
5f8e6c50 7204
44652c16
DMSP
7205 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7206 Group for discovering this issue.
d8dc8538 7207 ([CVE-2014-3512])
5f8e6c50
DMSP
7208
7209 *Steve Henson*
7210
44652c16
DMSP
7211 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7212 TLS 1.0 instead of higher protocol versions when the ClientHello message
7213 is badly fragmented. This allows a man-in-the-middle attacker to force a
7214 downgrade to TLS 1.0 even if both the server and the client support a
7215 higher protocol version, by modifying the client's TLS records.
5f8e6c50 7216
44652c16
DMSP
7217 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7218 researching this issue.
d8dc8538 7219 ([CVE-2014-3511])
5f8e6c50 7220
44652c16 7221 *David Benjamin*
5f8e6c50 7222
44652c16
DMSP
7223 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7224 to a denial of service attack. A malicious server can crash the client
7225 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7226 ciphersuite and sending carefully crafted handshake messages.
5f8e6c50 7227
44652c16
DMSP
7228 Thanks to Felix Gröbert (Google) for discovering and researching this
7229 issue.
d8dc8538 7230 ([CVE-2014-3510])
5f8e6c50 7231
44652c16 7232 *Emilia Käsper*
5f8e6c50 7233
44652c16
DMSP
7234 * By sending carefully crafted DTLS packets an attacker could cause openssl
7235 to leak memory. This can be exploited through a Denial of Service attack.
7236 Thanks to Adam Langley for discovering and researching this issue.
d8dc8538 7237 ([CVE-2014-3507])
5f8e6c50 7238
44652c16 7239 *Adam Langley*
5f8e6c50 7240
44652c16
DMSP
7241 * An attacker can force openssl to consume large amounts of memory whilst
7242 processing DTLS handshake messages. This can be exploited through a
7243 Denial of Service attack.
7244 Thanks to Adam Langley for discovering and researching this issue.
d8dc8538 7245 ([CVE-2014-3506])
5f8e6c50 7246
44652c16 7247 *Adam Langley*
5f8e6c50 7248
44652c16
DMSP
7249 * An attacker can force an error condition which causes openssl to crash
7250 whilst processing DTLS packets due to memory being freed twice. This
7251 can be exploited through a Denial of Service attack.
7252 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7253 this issue.
d8dc8538 7254 ([CVE-2014-3505])
5f8e6c50 7255
44652c16 7256 *Adam Langley*
5f8e6c50 7257
44652c16
DMSP
7258 * If a multithreaded client connects to a malicious server using a resumed
7259 session and the server sends an ec point format extension it could write
7260 up to 255 bytes to freed memory.
5f8e6c50 7261
44652c16
DMSP
7262 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7263 issue.
d8dc8538 7264 ([CVE-2014-3509])
5f8e6c50 7265
44652c16 7266 *Gabor Tyukasz*
5f8e6c50 7267
44652c16
DMSP
7268 * A malicious server can crash an OpenSSL client with a null pointer
7269 dereference (read) by specifying an SRP ciphersuite even though it was not
7270 properly negotiated with the client. This can be exploited through a
7271 Denial of Service attack.
5f8e6c50 7272
44652c16
DMSP
7273 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7274 discovering and researching this issue.
d8dc8538 7275 ([CVE-2014-5139])
5f8e6c50
DMSP
7276
7277 *Steve Henson*
7278
44652c16
DMSP
7279 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7280 X509_name_oneline, X509_name_print_ex et al. to leak some information
7281 from the stack. Applications may be affected if they echo pretty printing
7282 output to the attacker.
5f8e6c50 7283
44652c16 7284 Thanks to Ivan Fratric (Google) for discovering this issue.
d8dc8538 7285 ([CVE-2014-3508])
5f8e6c50 7286
44652c16 7287 *Emilia Käsper, and Steve Henson*
5f8e6c50 7288
44652c16
DMSP
7289 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7290 for corner cases. (Certain input points at infinity could lead to
7291 bogus results, with non-infinity inputs mapped to infinity too.)
5f8e6c50 7292
44652c16 7293 *Bodo Moeller*
5f8e6c50 7294
257e9d03 7295### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
5f8e6c50 7296
44652c16
DMSP
7297 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7298 handshake can force the use of weak keying material in OpenSSL
7299 SSL/TLS clients and servers.
5f8e6c50 7300
44652c16 7301 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
d8dc8538 7302 researching this issue. ([CVE-2014-0224])
5f8e6c50 7303
44652c16 7304 *KIKUCHI Masashi, Steve Henson*
5f8e6c50 7305
44652c16
DMSP
7306 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7307 OpenSSL DTLS client the code can be made to recurse eventually crashing
7308 in a DoS attack.
5f8e6c50 7309
44652c16 7310 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
d8dc8538 7311 ([CVE-2014-0221])
5f8e6c50 7312
44652c16 7313 *Imre Rad, Steve Henson*
5f8e6c50 7314
44652c16
DMSP
7315 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7316 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7317 client or server. This is potentially exploitable to run arbitrary
7318 code on a vulnerable client or server.
5f8e6c50 7319
d8dc8538 7320 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
5f8e6c50 7321
44652c16 7322 *Jüri Aedla, Steve Henson*
5f8e6c50 7323
44652c16
DMSP
7324 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7325 are subject to a denial of service attack.
5f8e6c50 7326
44652c16 7327 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
d8dc8538 7328 this issue. ([CVE-2014-3470])
5f8e6c50 7329
44652c16 7330 *Felix Gröbert, Ivan Fratric, Steve Henson*
5f8e6c50 7331
44652c16
DMSP
7332 * Harmonize version and its documentation. -f flag is used to display
7333 compilation flags.
5f8e6c50 7334
44652c16 7335 *mancha <mancha1@zoho.com>*
5f8e6c50 7336
44652c16
DMSP
7337 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7338 in i2d_ECPrivateKey.
5f8e6c50 7339
44652c16 7340 *mancha <mancha1@zoho.com>*
5f8e6c50 7341
44652c16 7342 * Fix some double frees. These are not thought to be exploitable.
5f8e6c50 7343
44652c16 7344 *mancha <mancha1@zoho.com>*
5f8e6c50 7345
257e9d03 7346### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
5f8e6c50 7347
44652c16
DMSP
7348 * A missing bounds check in the handling of the TLS heartbeat extension
7349 can be used to reveal up to 64k of memory to a connected client or
7350 server.
5f8e6c50 7351
44652c16
DMSP
7352 Thanks for Neel Mehta of Google Security for discovering this bug and to
7353 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
d8dc8538 7354 preparing the fix ([CVE-2014-0160])
5f8e6c50 7355
44652c16 7356 *Adam Langley, Bodo Moeller*
5f8e6c50 7357
44652c16
DMSP
7358 * Fix for the attack described in the paper "Recovering OpenSSL
7359 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7360 by Yuval Yarom and Naomi Benger. Details can be obtained from:
257e9d03 7361 <http://eprint.iacr.org/2014/140>
5f8e6c50 7362
44652c16 7363 Thanks to Yuval Yarom and Naomi Benger for discovering this
d8dc8538 7364 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
5f8e6c50 7365
44652c16 7366 *Yuval Yarom and Naomi Benger*
5f8e6c50 7367
44652c16 7368 * TLS pad extension: draft-agl-tls-padding-03
5f8e6c50 7369
44652c16
DMSP
7370 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7371 TLS client Hello record length value would otherwise be > 255 and
7372 less that 512 pad with a dummy extension containing zeroes so it
7373 is at least 512 bytes long.
5f8e6c50 7374
44652c16 7375 *Adam Langley, Steve Henson*
5f8e6c50 7376
257e9d03 7377### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
5f8e6c50 7378
44652c16
DMSP
7379 * Fix for TLS record tampering bug. A carefully crafted invalid
7380 handshake could crash OpenSSL with a NULL pointer exception.
7381 Thanks to Anton Johansson for reporting this issues.
d8dc8538 7382 ([CVE-2013-4353])
5f8e6c50 7383
44652c16
DMSP
7384 * Keep original DTLS digest and encryption contexts in retransmission
7385 structures so we can use the previous session parameters if they need
d8dc8538 7386 to be resent. ([CVE-2013-6450])
5f8e6c50 7387
44652c16 7388 *Steve Henson*
5f8e6c50 7389
44652c16
DMSP
7390 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7391 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7392 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7393 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7394 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7395 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
5f8e6c50 7396
44652c16 7397 *Rob Stradling, Adam Langley*
5f8e6c50 7398
257e9d03 7399### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
5f8e6c50 7400
44652c16
DMSP
7401 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7402 supporting platforms or when small records were transferred.
5f8e6c50 7403
44652c16 7404 *Andy Polyakov, Steve Henson*
5f8e6c50 7405
257e9d03 7406### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
5f8e6c50 7407
44652c16 7408 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
5f8e6c50 7409
44652c16
DMSP
7410 This addresses the flaw in CBC record processing discovered by
7411 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
257e9d03 7412 at: <http://www.isg.rhul.ac.uk/tls/>
5f8e6c50 7413
44652c16
DMSP
7414 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7415 Security Group at Royal Holloway, University of London
7416 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7417 Emilia Käsper for the initial patch.
d8dc8538 7418 ([CVE-2013-0169])
5f8e6c50 7419
44652c16 7420 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
5f8e6c50 7421
44652c16
DMSP
7422 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7423 ciphersuites which can be exploited in a denial of service attack.
7424 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7425 and detecting this bug and to Wolfgang Ettlinger
7426 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
d8dc8538 7427 ([CVE-2012-2686])
5f8e6c50 7428
44652c16 7429 *Adam Langley*
5f8e6c50 7430
44652c16 7431 * Return an error when checking OCSP signatures when key is NULL.
d8dc8538 7432 This fixes a DoS attack. ([CVE-2013-0166])
5f8e6c50
DMSP
7433
7434 *Steve Henson*
7435
44652c16 7436 * Make openssl verify return errors.
5f8e6c50 7437
44652c16 7438 *Chris Palmer <palmer@google.com> and Ben Laurie*
5f8e6c50 7439
44652c16
DMSP
7440 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7441 the right response is stapled. Also change SSL_get_certificate()
7442 so it returns the certificate actually sent.
257e9d03 7443 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
5f8e6c50 7444
44652c16 7445 *Rob Stradling <rob.stradling@comodo.com>*
5f8e6c50 7446
44652c16 7447 * Fix possible deadlock when decoding public keys.
5f8e6c50
DMSP
7448
7449 *Steve Henson*
7450
44652c16
DMSP
7451 * Don't use TLS 1.0 record version number in initial client hello
7452 if renegotiating.
5f8e6c50 7453
44652c16 7454 *Steve Henson*
5f8e6c50 7455
257e9d03 7456### Changes between 1.0.1b and 1.0.1c [10 May 2012]
5f8e6c50 7457
44652c16
DMSP
7458 * Sanity check record length before skipping explicit IV in TLS
7459 1.2, 1.1 and DTLS to fix DoS attack.
5f8e6c50 7460
44652c16
DMSP
7461 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7462 fuzzing as a service testing platform.
d8dc8538 7463 ([CVE-2012-2333])
5f8e6c50
DMSP
7464
7465 *Steve Henson*
7466
44652c16
DMSP
7467 * Initialise tkeylen properly when encrypting CMS messages.
7468 Thanks to Solar Designer of Openwall for reporting this issue.
5f8e6c50
DMSP
7469
7470 *Steve Henson*
7471
44652c16
DMSP
7472 * In FIPS mode don't try to use composite ciphers as they are not
7473 approved.
5f8e6c50
DMSP
7474
7475 *Steve Henson*
7476
257e9d03 7477### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
44652c16
DMSP
7478
7479 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7480 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7481 mean any application compiled against OpenSSL 1.0.0 headers setting
7482 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7483 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7484 0x10000000L Any application which was previously compiled against
7485 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7486 will need to be recompiled as a result. Letting be results in
7487 inability to disable specifically TLS 1.1 and in client context,
7488 in unlike event, limit maximum offered version to TLS 1.0 [see below].
5f8e6c50
DMSP
7489
7490 *Steve Henson*
7491
44652c16
DMSP
7492 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7493 disable just protocol X, but all protocols above X *if* there are
7494 protocols *below* X still enabled. In more practical terms it means
7495 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1dc1ea18
DDO
7496 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7497 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
44652c16 7498 client side.
5f8e6c50 7499
44652c16 7500 *Andy Polyakov*
5f8e6c50 7501
257e9d03 7502### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
5f8e6c50 7503
44652c16
DMSP
7504 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7505 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7506 in CRYPTO_realloc_clean.
5f8e6c50 7507
44652c16
DMSP
7508 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7509 issue and to Adam Langley <agl@chromium.org> for fixing it.
d8dc8538 7510 ([CVE-2012-2110])
5f8e6c50 7511
44652c16 7512 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
5f8e6c50 7513
44652c16 7514 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
5f8e6c50 7515
44652c16 7516 *Adam Langley*
5f8e6c50 7517
44652c16
DMSP
7518 * Workarounds for some broken servers that "hang" if a client hello
7519 record length exceeds 255 bytes.
7520
7521 1. Do not use record version number > TLS 1.0 in initial client
7522 hello: some (but not all) hanging servers will now work.
7523 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7524 the number of ciphers sent in the client hello. This should be
7525 set to an even number, such as 50, for example by passing:
7526 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7527 Most broken servers should now work.
7528 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7529 TLS 1.2 client support entirely.
5f8e6c50
DMSP
7530
7531 *Steve Henson*
7532
44652c16 7533 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
5f8e6c50 7534
44652c16 7535 *Andy Polyakov*
5f8e6c50 7536
257e9d03 7537### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
44652c16
DMSP
7538
7539 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7540 STRING form instead of a DigestInfo.
5f8e6c50
DMSP
7541
7542 *Steve Henson*
7543
44652c16
DMSP
7544 * The format used for MDC2 RSA signatures is inconsistent between EVP
7545 and the RSA_sign/RSA_verify functions. This was made more apparent when
7546 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7547 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7548 the correct format in RSA_verify so both forms transparently work.
5f8e6c50 7549
44652c16 7550 *Steve Henson*
5f8e6c50 7551
44652c16
DMSP
7552 * Some servers which support TLS 1.0 can choke if we initially indicate
7553 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7554 encrypted premaster secret. As a workaround use the maximum permitted
7555 client version in client hello, this should keep such servers happy
7556 and still work with previous versions of OpenSSL.
5f8e6c50 7557
44652c16 7558 *Steve Henson*
5f8e6c50 7559
44652c16 7560 * Add support for TLS/DTLS heartbeats.
5f8e6c50 7561
44652c16 7562 *Robin Seggelmann <seggelmann@fh-muenster.de>*
5f8e6c50 7563
44652c16 7564 * Add support for SCTP.
5f8e6c50 7565
44652c16 7566 *Robin Seggelmann <seggelmann@fh-muenster.de>*
5f8e6c50 7567
44652c16 7568 * Improved PRNG seeding for VOS.
5f8e6c50 7569
44652c16 7570 *Paul Green <Paul.Green@stratus.com>*
5f8e6c50 7571
44652c16 7572 * Extensive assembler packs updates, most notably:
5f8e6c50 7573
257e9d03
RS
7574 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7575 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7576 - x86_64: bit-sliced AES implementation;
7577 - ARM: NEON support, contemporary platforms optimizations;
7578 - s390x: z196 support;
7579 - `*`: GHASH and GF(2^m) multiplication implementations;
5f8e6c50 7580
44652c16 7581 *Andy Polyakov*
5f8e6c50 7582
44652c16
DMSP
7583 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7584 (removal of unnecessary code)
5f8e6c50 7585
44652c16 7586 *Peter Sylvester <peter.sylvester@edelweb.fr>*
5f8e6c50 7587
44652c16 7588 * Add TLS key material exporter from RFC 5705.
5f8e6c50 7589
44652c16 7590 *Eric Rescorla*
5f8e6c50 7591
44652c16 7592 * Add DTLS-SRTP negotiation from RFC 5764.
5f8e6c50 7593
44652c16 7594 *Eric Rescorla*
5f8e6c50 7595
44652c16 7596 * Add Next Protocol Negotiation,
257e9d03 7597 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
44652c16
DMSP
7598 disabled with a no-npn flag to config or Configure. Code donated
7599 by Google.
5f8e6c50 7600
44652c16 7601 *Adam Langley <agl@google.com> and Ben Laurie*
5f8e6c50 7602
44652c16
DMSP
7603 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7604 NIST-P256, NIST-P521, with constant-time single point multiplication on
7605 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7606 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7607 Code made available under Apache License version 2.0.
5f8e6c50 7608
44652c16
DMSP
7609 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7610 line to include this in your build of OpenSSL, and run "make depend" (or
7611 "make update"). This enables the following EC_METHODs:
5f8e6c50 7612
44652c16
DMSP
7613 EC_GFp_nistp224_method()
7614 EC_GFp_nistp256_method()
7615 EC_GFp_nistp521_method()
5f8e6c50 7616
44652c16
DMSP
7617 EC_GROUP_new_by_curve_name() will automatically use these (while
7618 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7619 implementations).
5f8e6c50 7620
44652c16 7621 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
5f8e6c50 7622
44652c16
DMSP
7623 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7624 all platforms. Move ssize_t definition from e_os.h to the public
7625 header file e_os2.h as it now appears in public header file cms.h
5f8e6c50 7626
44652c16 7627 *Steve Henson*
5f8e6c50 7628
44652c16
DMSP
7629 * New -sigopt option to the ca, req and x509 utilities. Additional
7630 signature parameters can be passed using this option and in
7631 particular PSS.
5f8e6c50 7632
44652c16 7633 *Steve Henson*
5f8e6c50 7634
44652c16
DMSP
7635 * Add RSA PSS signing function. This will generate and set the
7636 appropriate AlgorithmIdentifiers for PSS based on those in the
7637 corresponding EVP_MD_CTX structure. No application support yet.
5f8e6c50 7638
44652c16 7639 *Steve Henson*
5f8e6c50 7640
44652c16
DMSP
7641 * Support for companion algorithm specific ASN1 signing routines.
7642 New function ASN1_item_sign_ctx() signs a pre-initialised
7643 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7644 the appropriate parameters.
5f8e6c50
DMSP
7645
7646 *Steve Henson*
7647
44652c16
DMSP
7648 * Add new algorithm specific ASN1 verification initialisation function
7649 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7650 handling will be the same no matter what EVP_PKEY_METHOD is used.
7651 Add a PSS handler to support verification of PSS signatures: checked
7652 against a number of sample certificates.
5f8e6c50 7653
44652c16 7654 *Steve Henson*
5f8e6c50 7655
44652c16 7656 * Add signature printing for PSS. Add PSS OIDs.
5f8e6c50 7657
44652c16 7658 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
5f8e6c50 7659
44652c16
DMSP
7660 * Add algorithm specific signature printing. An individual ASN1 method
7661 can now print out signatures instead of the standard hex dump.
5f8e6c50 7662
44652c16
DMSP
7663 More complex signatures (e.g. PSS) can print out more meaningful
7664 information. Include DSA version that prints out the signature
7665 parameters r, s.
5f8e6c50 7666
44652c16 7667 *Steve Henson*
5f8e6c50 7668
44652c16
DMSP
7669 * Password based recipient info support for CMS library: implementing
7670 RFC3211.
5f8e6c50 7671
44652c16 7672 *Steve Henson*
5f8e6c50 7673
44652c16
DMSP
7674 * Split password based encryption into PBES2 and PBKDF2 functions. This
7675 neatly separates the code into cipher and PBE sections and is required
7676 for some algorithms that split PBES2 into separate pieces (such as
7677 password based CMS).
5f8e6c50 7678
44652c16 7679 *Steve Henson*
5f8e6c50 7680
44652c16
DMSP
7681 * Session-handling fixes:
7682 - Fix handling of connections that are resuming with a session ID,
7683 but also support Session Tickets.
7684 - Fix a bug that suppressed issuing of a new ticket if the client
7685 presented a ticket with an expired session.
7686 - Try to set the ticket lifetime hint to something reasonable.
7687 - Make tickets shorter by excluding irrelevant information.
7688 - On the client side, don't ignore renewed tickets.
5f8e6c50 7689
44652c16 7690 *Adam Langley, Bodo Moeller (Google)*
5f8e6c50 7691
44652c16 7692 * Fix PSK session representation.
5f8e6c50 7693
44652c16 7694 *Bodo Moeller*
5f8e6c50 7695
44652c16 7696 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
5f8e6c50 7697
44652c16 7698 This work was sponsored by Intel.
5f8e6c50 7699
44652c16 7700 *Andy Polyakov*
5f8e6c50 7701
44652c16
DMSP
7702 * Add GCM support to TLS library. Some custom code is needed to split
7703 the IV between the fixed (from PRF) and explicit (from TLS record)
7704 portions. This adds all GCM ciphersuites supported by RFC5288 and
257e9d03 7705 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
44652c16 7706 add a special AESGCM string for GCM only.
5f8e6c50 7707
44652c16 7708 *Steve Henson*
5f8e6c50 7709
44652c16
DMSP
7710 * Expand range of ctrls for AES GCM. Permit setting invocation
7711 field on decrypt and retrieval of invocation field only on encrypt.
5f8e6c50 7712
44652c16 7713 *Steve Henson*
5f8e6c50 7714
44652c16
DMSP
7715 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7716 As required by RFC5289 these ciphersuites cannot be used if for
7717 versions of TLS earlier than 1.2.
5f8e6c50 7718
44652c16 7719 *Steve Henson*
5f8e6c50 7720
44652c16
DMSP
7721 * For FIPS capable OpenSSL interpret a NULL default public key method
7722 as unset and return the appropriate default but do *not* set the default.
7723 This means we can return the appropriate method in applications that
7724 switch between FIPS and non-FIPS modes.
7725
7726 *Steve Henson*
5f8e6c50 7727
44652c16
DMSP
7728 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7729 ENGINE is used then we cannot handle that in the FIPS module so we
7730 keep original code iff non-FIPS operations are allowed.
5f8e6c50
DMSP
7731
7732 *Steve Henson*
7733
44652c16 7734 * Add -attime option to openssl utilities.
5f8e6c50 7735
44652c16 7736 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
5f8e6c50 7737
44652c16 7738 * Redirect DSA and DH operations to FIPS module in FIPS mode.
5f8e6c50
DMSP
7739
7740 *Steve Henson*
7741
44652c16
DMSP
7742 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7743 FIPS EC methods unconditionally for now.
5f8e6c50 7744
44652c16 7745 *Steve Henson*
5f8e6c50 7746
44652c16 7747 * New build option no-ec2m to disable characteristic 2 code.
5f8e6c50 7748
44652c16 7749 *Steve Henson*
5f8e6c50 7750
44652c16
DMSP
7751 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7752 all cases can be covered as some introduce binary incompatibilities.
5f8e6c50 7753
44652c16 7754 *Steve Henson*
5f8e6c50 7755
44652c16
DMSP
7756 * Redirect RSA operations to FIPS module including keygen,
7757 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
5f8e6c50 7758
44652c16 7759 *Steve Henson*
5f8e6c50 7760
4d49b685 7761 * Add similar low-level API blocking to ciphers.
5f8e6c50 7762
44652c16 7763 *Steve Henson*
5f8e6c50 7764
4d49b685 7765 * low-level digest APIs are not approved in FIPS mode: any attempt
44652c16 7766 to use these will cause a fatal error. Applications that *really* want
257e9d03 7767 to use them can use the `private_*` version instead.
5f8e6c50 7768
44652c16 7769 *Steve Henson*
5f8e6c50 7770
44652c16 7771 * Redirect cipher operations to FIPS module for FIPS builds.
5f8e6c50 7772
44652c16 7773 *Steve Henson*
5f8e6c50 7774
44652c16 7775 * Redirect digest operations to FIPS module for FIPS builds.
5f8e6c50 7776
44652c16
DMSP
7777 *Steve Henson*
7778
7779 * Update build system to add "fips" flag which will link in fipscanister.o
7780 for static and shared library builds embedding a signature if needed.
5f8e6c50
DMSP
7781
7782 *Steve Henson*
7783
44652c16
DMSP
7784 * Output TLS supported curves in preference order instead of numerical
7785 order. This is currently hardcoded for the highest order curves first.
7786 This should be configurable so applications can judge speed vs strength.
5f8e6c50 7787
44652c16 7788 *Steve Henson*
5f8e6c50 7789
44652c16 7790 * Add TLS v1.2 server support for client authentication.
5f8e6c50 7791
44652c16 7792 *Steve Henson*
5f8e6c50 7793
44652c16
DMSP
7794 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7795 and enable MD5.
5f8e6c50 7796
44652c16 7797 *Steve Henson*
5f8e6c50 7798
44652c16
DMSP
7799 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7800 FIPS modules versions.
5f8e6c50 7801
44652c16 7802 *Steve Henson*
5f8e6c50 7803
44652c16
DMSP
7804 * Add TLS v1.2 client side support for client authentication. Keep cache
7805 of handshake records longer as we don't know the hash algorithm to use
7806 until after the certificate request message is received.
5f8e6c50 7807
44652c16 7808 *Steve Henson*
5f8e6c50 7809
44652c16
DMSP
7810 * Initial TLS v1.2 client support. Add a default signature algorithms
7811 extension including all the algorithms we support. Parse new signature
7812 format in client key exchange. Relax some ECC signing restrictions for
7813 TLS v1.2 as indicated in RFC5246.
5f8e6c50 7814
44652c16 7815 *Steve Henson*
5f8e6c50 7816
44652c16
DMSP
7817 * Add server support for TLS v1.2 signature algorithms extension. Switch
7818 to new signature format when needed using client digest preference.
7819 All server ciphersuites should now work correctly in TLS v1.2. No client
7820 support yet and no support for client certificates.
5f8e6c50 7821
44652c16 7822 *Steve Henson*
5f8e6c50 7823
44652c16
DMSP
7824 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7825 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7826 ciphersuites. At present only RSA key exchange ciphersuites work with
7827 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7828 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7829 and version checking.
5f8e6c50 7830
44652c16 7831 *Steve Henson*
5f8e6c50 7832
44652c16
DMSP
7833 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7834 with this defined it will not be affected by any changes to ssl internal
7835 structures. Add several utility functions to allow openssl application
7836 to work with OPENSSL_NO_SSL_INTERN defined.
5f8e6c50 7837
44652c16 7838 *Steve Henson*
5f8e6c50 7839
44652c16
DMSP
7840 * A long standing patch to add support for SRP from EdelWeb (Peter
7841 Sylvester and Christophe Renou) was integrated.
7842 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7843 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7844 Ben Laurie*
5f8e6c50 7845
44652c16 7846 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
5f8e6c50 7847
44652c16 7848 *Steve Henson*
5f8e6c50 7849
44652c16
DMSP
7850 * Permit abbreviated handshakes when renegotiating using the function
7851 SSL_renegotiate_abbreviated().
5f8e6c50 7852
44652c16 7853 *Robin Seggelmann <seggelmann@fh-muenster.de>*
5f8e6c50 7854
44652c16
DMSP
7855 * Add call to ENGINE_register_all_complete() to
7856 ENGINE_load_builtin_engines(), so some implementations get used
7857 automatically instead of needing explicit application support.
5f8e6c50 7858
44652c16 7859 *Steve Henson*
5f8e6c50 7860
44652c16 7861 * Add support for TLS key exporter as described in RFC5705.
5f8e6c50 7862
44652c16 7863 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
5f8e6c50 7864
44652c16
DMSP
7865 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7866 a few changes are required:
5f8e6c50 7867
44652c16
DMSP
7868 Add SSL_OP_NO_TLSv1_1 flag.
7869 Add TLSv1_1 methods.
7870 Update version checking logic to handle version 1.1.
7871 Add explicit IV handling (ported from DTLS code).
7872 Add command line options to s_client/s_server.
5f8e6c50 7873
44652c16 7874 *Steve Henson*
5f8e6c50 7875
44652c16
DMSP
7876OpenSSL 1.0.0
7877-------------
5f8e6c50 7878
257e9d03 7879### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
5f8e6c50 7880
44652c16 7881 * X509_ATTRIBUTE memory leak
5f8e6c50 7882
44652c16
DMSP
7883 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7884 memory. This structure is used by the PKCS#7 and CMS routines so any
7885 application which reads PKCS#7 or CMS data from untrusted sources is
7886 affected. SSL/TLS is not affected.
5f8e6c50 7887
44652c16
DMSP
7888 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7889 libFuzzer.
d8dc8538 7890 ([CVE-2015-3195])
5f8e6c50 7891
44652c16 7892 *Stephen Henson*
5f8e6c50 7893
44652c16 7894 * Race condition handling PSK identify hint
5f8e6c50 7895
44652c16
DMSP
7896 If PSK identity hints are received by a multi-threaded client then
7897 the values are wrongly updated in the parent SSL_CTX structure. This can
7898 result in a race condition potentially leading to a double free of the
7899 identify hint data.
d8dc8538 7900 ([CVE-2015-3196])
5f8e6c50 7901
44652c16 7902 *Stephen Henson*
5f8e6c50 7903
257e9d03 7904### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
5f8e6c50 7905
44652c16 7906 * Malformed ECParameters causes infinite loop
5f8e6c50 7907
44652c16
DMSP
7908 When processing an ECParameters structure OpenSSL enters an infinite loop
7909 if the curve specified is over a specially malformed binary polynomial
7910 field.
5f8e6c50 7911
44652c16
DMSP
7912 This can be used to perform denial of service against any
7913 system which processes public keys, certificate requests or
7914 certificates. This includes TLS clients and TLS servers with
7915 client authentication enabled.
5f8e6c50 7916
44652c16 7917 This issue was reported to OpenSSL by Joseph Barr-Pixton.
d8dc8538 7918 ([CVE-2015-1788])
5f8e6c50 7919
44652c16 7920 *Andy Polyakov*
5f8e6c50 7921
44652c16 7922 * Exploitable out-of-bounds read in X509_cmp_time
5f8e6c50 7923
44652c16
DMSP
7924 X509_cmp_time does not properly check the length of the ASN1_TIME
7925 string and can read a few bytes out of bounds. In addition,
7926 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7927 time string.
5f8e6c50 7928
44652c16
DMSP
7929 An attacker can use this to craft malformed certificates and CRLs of
7930 various sizes and potentially cause a segmentation fault, resulting in
7931 a DoS on applications that verify certificates or CRLs. TLS clients
7932 that verify CRLs are affected. TLS clients and servers with client
7933 authentication enabled may be affected if they use custom verification
7934 callbacks.
5f8e6c50 7935
44652c16
DMSP
7936 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7937 independently by Hanno Böck.
d8dc8538 7938 ([CVE-2015-1789])
5f8e6c50 7939
44652c16 7940 *Emilia Käsper*
5f8e6c50 7941
44652c16 7942 * PKCS7 crash with missing EnvelopedContent
5f8e6c50 7943
44652c16
DMSP
7944 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7945 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7946 with missing content and trigger a NULL pointer dereference on parsing.
5f8e6c50 7947
44652c16
DMSP
7948 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7949 structures from untrusted sources are affected. OpenSSL clients and
7950 servers are not affected.
5f8e6c50 7951
44652c16 7952 This issue was reported to OpenSSL by Michal Zalewski (Google).
d8dc8538 7953 ([CVE-2015-1790])
5f8e6c50 7954
44652c16 7955 *Emilia Käsper*
5f8e6c50 7956
44652c16 7957 * CMS verify infinite loop with unknown hash function
5f8e6c50 7958
44652c16
DMSP
7959 When verifying a signedData message the CMS code can enter an infinite loop
7960 if presented with an unknown hash function OID. This can be used to perform
7961 denial of service against any system which verifies signedData messages using
7962 the CMS code.
7963 This issue was reported to OpenSSL by Johannes Bauer.
d8dc8538 7964 ([CVE-2015-1792])
5f8e6c50 7965
44652c16 7966 *Stephen Henson*
5f8e6c50 7967
44652c16 7968 * Race condition handling NewSessionTicket
5f8e6c50 7969
44652c16
DMSP
7970 If a NewSessionTicket is received by a multi-threaded client when attempting to
7971 reuse a previous ticket then a race condition can occur potentially leading to
7972 a double free of the ticket data.
d8dc8538 7973 ([CVE-2015-1791])
5f8e6c50 7974
44652c16 7975 *Matt Caswell*
5f8e6c50 7976
257e9d03 7977### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
5f8e6c50 7978
44652c16
DMSP
7979 * Segmentation fault in ASN1_TYPE_cmp fix
7980
7981 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7982 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7983 certificate signature algorithm consistency this can be used to crash any
7984 certificate verification operation and exploited in a DoS attack. Any
7985 application which performs certificate verification is vulnerable including
7986 OpenSSL clients and servers which enable client authentication.
d8dc8538 7987 ([CVE-2015-0286])
5f8e6c50 7988
44652c16 7989 *Stephen Henson*
5f8e6c50 7990
44652c16 7991 * ASN.1 structure reuse memory corruption fix
5f8e6c50 7992
44652c16
DMSP
7993 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7994 memory corruption via an invalid write. Such reuse is and has been
7995 strongly discouraged and is believed to be rare.
5f8e6c50 7996
44652c16
DMSP
7997 Applications that parse structures containing CHOICE or ANY DEFINED BY
7998 components may be affected. Certificate parsing (d2i_X509 and related
7999 functions) are however not affected. OpenSSL clients and servers are
8000 not affected.
d8dc8538 8001 ([CVE-2015-0287])
5f8e6c50 8002
44652c16 8003 *Stephen Henson*
5f8e6c50 8004
44652c16 8005 * PKCS7 NULL pointer dereferences fix
5f8e6c50 8006
44652c16
DMSP
8007 The PKCS#7 parsing code does not handle missing outer ContentInfo
8008 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8009 missing content and trigger a NULL pointer dereference on parsing.
5f8e6c50 8010
44652c16
DMSP
8011 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8012 otherwise parse PKCS#7 structures from untrusted sources are
8013 affected. OpenSSL clients and servers are not affected.
5f8e6c50 8014
44652c16 8015 This issue was reported to OpenSSL by Michal Zalewski (Google).
d8dc8538 8016 ([CVE-2015-0289])
5f8e6c50 8017
44652c16 8018 *Emilia Käsper*
5f8e6c50 8019
44652c16 8020 * DoS via reachable assert in SSLv2 servers fix
5f8e6c50 8021
44652c16
DMSP
8022 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8023 servers that both support SSLv2 and enable export cipher suites by sending
8024 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5f8e6c50 8025
44652c16
DMSP
8026 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8027 (OpenSSL development team).
d8dc8538 8028 ([CVE-2015-0293])
5f8e6c50 8029
44652c16 8030 *Emilia Käsper*
5f8e6c50 8031
44652c16 8032 * Use After Free following d2i_ECPrivatekey error fix
5f8e6c50 8033
44652c16
DMSP
8034 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8035 could cause a use after free condition. This, in turn, could cause a double
8036 free in several private key parsing functions (such as d2i_PrivateKey
8037 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8038 for applications that receive EC private keys from untrusted
8039 sources. This scenario is considered rare.
5f8e6c50 8040
44652c16
DMSP
8041 This issue was discovered by the BoringSSL project and fixed in their
8042 commit 517073cd4b.
d8dc8538 8043 ([CVE-2015-0209])
5f8e6c50 8044
44652c16 8045 *Matt Caswell*
5f8e6c50 8046
44652c16 8047 * X509_to_X509_REQ NULL pointer deref fix
5f8e6c50 8048
44652c16
DMSP
8049 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8050 the certificate key is invalid. This function is rarely used in practice.
5f8e6c50 8051
44652c16 8052 This issue was discovered by Brian Carpenter.
d8dc8538 8053 ([CVE-2015-0288])
5f8e6c50 8054
44652c16 8055 *Stephen Henson*
5f8e6c50 8056
44652c16 8057 * Removed the export ciphers from the DEFAULT ciphers
5f8e6c50 8058
44652c16 8059 *Kurt Roeckx*
5f8e6c50 8060
257e9d03 8061### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
5f8e6c50 8062
44652c16 8063 * Build fixes for the Windows and OpenVMS platforms
5f8e6c50 8064
44652c16 8065 *Matt Caswell and Richard Levitte*
5f8e6c50 8066
257e9d03 8067### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
44652c16
DMSP
8068
8069 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8070 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8071 dereference. This could lead to a Denial Of Service attack. Thanks to
8072 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
d8dc8538 8073 ([CVE-2014-3571])
5f8e6c50
DMSP
8074
8075 *Steve Henson*
8076
44652c16
DMSP
8077 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8078 dtls1_buffer_record function under certain conditions. In particular this
8079 could occur if an attacker sent repeated DTLS records with the same
8080 sequence number but for the next epoch. The memory leak could be exploited
8081 by an attacker in a Denial of Service attack through memory exhaustion.
8082 Thanks to Chris Mueller for reporting this issue.
d8dc8538 8083 ([CVE-2015-0206])
5f8e6c50 8084
44652c16 8085 *Matt Caswell*
5f8e6c50 8086
44652c16
DMSP
8087 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8088 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8089 method would be set to NULL which could later result in a NULL pointer
8090 dereference. Thanks to Frank Schmirler for reporting this issue.
d8dc8538 8091 ([CVE-2014-3569])
5f8e6c50 8092
44652c16 8093 *Kurt Roeckx*
5f8e6c50 8094
44652c16
DMSP
8095 * Abort handshake if server key exchange message is omitted for ephemeral
8096 ECDH ciphersuites.
5f8e6c50 8097
44652c16
DMSP
8098 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8099 reporting this issue.
d8dc8538 8100 ([CVE-2014-3572])
5f8e6c50 8101
44652c16 8102 *Steve Henson*
5f8e6c50 8103
44652c16
DMSP
8104 * Remove non-export ephemeral RSA code on client and server. This code
8105 violated the TLS standard by allowing the use of temporary RSA keys in
8106 non-export ciphersuites and could be used by a server to effectively
8107 downgrade the RSA key length used to a value smaller than the server
8108 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8109 INRIA or reporting this issue.
d8dc8538 8110 ([CVE-2015-0204])
5f8e6c50 8111
44652c16 8112 *Steve Henson*
5f8e6c50 8113
44652c16
DMSP
8114 * Fixed issue where DH client certificates are accepted without verification.
8115 An OpenSSL server will accept a DH certificate for client authentication
8116 without the certificate verify message. This effectively allows a client to
8117 authenticate without the use of a private key. This only affects servers
8118 which trust a client certificate authority which issues certificates
8119 containing DH keys: these are extremely rare and hardly ever encountered.
8120 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8121 this issue.
d8dc8538 8122 ([CVE-2015-0205])
5f8e6c50 8123
44652c16 8124 *Steve Henson*
5f8e6c50 8125
43a70f02
RS
8126 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8127 results on some platforms, including x86_64. This bug occurs at random
8128 with a very low probability, and is not known to be exploitable in any
8129 way, though its exact impact is difficult to determine. Thanks to Pieter
8130 Wuille (Blockstream) who reported this issue and also suggested an initial
8131 fix. Further analysis was conducted by the OpenSSL development team and
8132 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8133 the OpenSSL core team.
d8dc8538 8134 ([CVE-2014-3570])
5f8e6c50 8135
43a70f02 8136 *Andy Polyakov*
5f8e6c50 8137
43a70f02 8138 * Fix various certificate fingerprint issues.
5f8e6c50 8139
44652c16
DMSP
8140 By using non-DER or invalid encodings outside the signed portion of a
8141 certificate the fingerprint can be changed without breaking the signature.
8142 Although no details of the signed portion of the certificate can be changed
8143 this can cause problems with some applications: e.g. those using the
8144 certificate fingerprint for blacklists.
5f8e6c50 8145
44652c16 8146 1. Reject signatures with non zero unused bits.
5f8e6c50 8147
44652c16
DMSP
8148 If the BIT STRING containing the signature has non zero unused bits reject
8149 the signature. All current signature algorithms require zero unused bits.
5f8e6c50 8150
44652c16 8151 2. Check certificate algorithm consistency.
5f8e6c50 8152
44652c16
DMSP
8153 Check the AlgorithmIdentifier inside TBS matches the one in the
8154 certificate signature. NB: this will result in signature failure
8155 errors for some broken certificates.
5f8e6c50 8156
44652c16 8157 Thanks to Konrad Kraszewski from Google for reporting this issue.
5f8e6c50 8158
44652c16 8159 3. Check DSA/ECDSA signatures use DER.
5f8e6c50 8160
44652c16
DMSP
8161 Reencode DSA/ECDSA signatures and compare with the original received
8162 signature. Return an error if there is a mismatch.
5f8e6c50 8163
44652c16
DMSP
8164 This will reject various cases including garbage after signature
8165 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8166 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8167 (negative or with leading zeroes).
5f8e6c50 8168
44652c16
DMSP
8169 Further analysis was conducted and fixes were developed by Stephen Henson
8170 of the OpenSSL core team.
5f8e6c50 8171
d8dc8538 8172 ([CVE-2014-8275])
5f8e6c50
DMSP
8173
8174 *Steve Henson*
8175
257e9d03 8176### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
5f8e6c50 8177
44652c16 8178 * Session Ticket Memory Leak.
5f8e6c50 8179
44652c16
DMSP
8180 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8181 integrity of that ticket is first verified. In the event of a session
8182 ticket integrity check failing, OpenSSL will fail to free memory
8183 causing a memory leak. By sending a large number of invalid session
8184 tickets an attacker could exploit this issue in a Denial Of Service
8185 attack.
d8dc8538 8186 ([CVE-2014-3567])
5f8e6c50
DMSP
8187
8188 *Steve Henson*
8189
44652c16 8190 * Build option no-ssl3 is incomplete.
5f8e6c50 8191
44652c16
DMSP
8192 When OpenSSL is configured with "no-ssl3" as a build option, servers
8193 could accept and complete a SSL 3.0 handshake, and clients could be
8194 configured to send them.
d8dc8538 8195 ([CVE-2014-3568])
5f8e6c50 8196
44652c16
DMSP
8197 *Akamai and the OpenSSL team*
8198
8199 * Add support for TLS_FALLBACK_SCSV.
8200 Client applications doing fallback retries should call
8201 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
d8dc8538 8202 ([CVE-2014-3566])
5f8e6c50 8203
44652c16 8204 *Adam Langley, Bodo Moeller*
5f8e6c50 8205
44652c16 8206 * Add additional DigestInfo checks.
5f8e6c50 8207
44652c16
DMSP
8208 Reencode DigestInto in DER and check against the original when
8209 verifying RSA signature: this will reject any improperly encoded
8210 DigestInfo structures.
5f8e6c50 8211
44652c16 8212 Note: this is a precautionary measure and no attacks are currently known.
5f8e6c50 8213
5f8e6c50
DMSP
8214 *Steve Henson*
8215
257e9d03 8216### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
5f8e6c50 8217
44652c16
DMSP
8218 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8219 to a denial of service attack. A malicious server can crash the client
8220 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8221 ciphersuite and sending carefully crafted handshake messages.
5f8e6c50 8222
44652c16
DMSP
8223 Thanks to Felix Gröbert (Google) for discovering and researching this
8224 issue.
d8dc8538 8225 ([CVE-2014-3510])
5f8e6c50 8226
44652c16 8227 *Emilia Käsper*
5f8e6c50 8228
44652c16
DMSP
8229 * By sending carefully crafted DTLS packets an attacker could cause openssl
8230 to leak memory. This can be exploited through a Denial of Service attack.
8231 Thanks to Adam Langley for discovering and researching this issue.
d8dc8538 8232 ([CVE-2014-3507])
5f8e6c50 8233
44652c16 8234 *Adam Langley*
5f8e6c50 8235
44652c16
DMSP
8236 * An attacker can force openssl to consume large amounts of memory whilst
8237 processing DTLS handshake messages. This can be exploited through a
8238 Denial of Service attack.
8239 Thanks to Adam Langley for discovering and researching this issue.
d8dc8538 8240 ([CVE-2014-3506])
5f8e6c50 8241
44652c16 8242 *Adam Langley*
5f8e6c50 8243
44652c16
DMSP
8244 * An attacker can force an error condition which causes openssl to crash
8245 whilst processing DTLS packets due to memory being freed twice. This
8246 can be exploited through a Denial of Service attack.
8247 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8248 this issue.
d8dc8538 8249 ([CVE-2014-3505])
5f8e6c50 8250
44652c16 8251 *Adam Langley*
5f8e6c50 8252
44652c16
DMSP
8253 * If a multithreaded client connects to a malicious server using a resumed
8254 session and the server sends an ec point format extension it could write
8255 up to 255 bytes to freed memory.
5f8e6c50 8256
44652c16
DMSP
8257 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8258 issue.
d8dc8538 8259 ([CVE-2014-3509])
5f8e6c50 8260
44652c16 8261 *Gabor Tyukasz*
5f8e6c50 8262
44652c16
DMSP
8263 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8264 X509_name_oneline, X509_name_print_ex et al. to leak some information
8265 from the stack. Applications may be affected if they echo pretty printing
8266 output to the attacker.
5f8e6c50 8267
44652c16 8268 Thanks to Ivan Fratric (Google) for discovering this issue.
d8dc8538 8269 ([CVE-2014-3508])
5f8e6c50 8270
44652c16 8271 *Emilia Käsper, and Steve Henson*
5f8e6c50 8272
44652c16
DMSP
8273 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8274 for corner cases. (Certain input points at infinity could lead to
8275 bogus results, with non-infinity inputs mapped to infinity too.)
5f8e6c50 8276
44652c16 8277 *Bodo Moeller*
5f8e6c50 8278
257e9d03 8279### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
5f8e6c50 8280
44652c16
DMSP
8281 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8282 handshake can force the use of weak keying material in OpenSSL
8283 SSL/TLS clients and servers.
5f8e6c50 8284
44652c16 8285 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
d8dc8538 8286 researching this issue. ([CVE-2014-0224])
5f8e6c50 8287
44652c16 8288 *KIKUCHI Masashi, Steve Henson*
5f8e6c50 8289
44652c16
DMSP
8290 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8291 OpenSSL DTLS client the code can be made to recurse eventually crashing
8292 in a DoS attack.
5f8e6c50 8293
44652c16 8294 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
d8dc8538 8295 ([CVE-2014-0221])
5f8e6c50 8296
44652c16 8297 *Imre Rad, Steve Henson*
5f8e6c50 8298
44652c16
DMSP
8299 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8300 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8301 client or server. This is potentially exploitable to run arbitrary
8302 code on a vulnerable client or server.
5f8e6c50 8303
d8dc8538 8304 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
5f8e6c50 8305
44652c16 8306 *Jüri Aedla, Steve Henson*
5f8e6c50 8307
44652c16
DMSP
8308 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8309 are subject to a denial of service attack.
5f8e6c50 8310
44652c16 8311 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
d8dc8538 8312 this issue. ([CVE-2014-3470])
5f8e6c50 8313
44652c16 8314 *Felix Gröbert, Ivan Fratric, Steve Henson*
5f8e6c50 8315
44652c16
DMSP
8316 * Harmonize version and its documentation. -f flag is used to display
8317 compilation flags.
5f8e6c50 8318
44652c16 8319 *mancha <mancha1@zoho.com>*
5f8e6c50 8320
44652c16
DMSP
8321 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8322 in i2d_ECPrivateKey.
5f8e6c50 8323
44652c16 8324 *mancha <mancha1@zoho.com>*
5f8e6c50 8325
44652c16 8326 * Fix some double frees. These are not thought to be exploitable.
5f8e6c50 8327
44652c16 8328 *mancha <mancha1@zoho.com>*
5f8e6c50 8329
44652c16
DMSP
8330 * Fix for the attack described in the paper "Recovering OpenSSL
8331 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8332 by Yuval Yarom and Naomi Benger. Details can be obtained from:
257e9d03 8333 <http://eprint.iacr.org/2014/140>
5f8e6c50 8334
44652c16 8335 Thanks to Yuval Yarom and Naomi Benger for discovering this
d8dc8538 8336 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
5f8e6c50 8337
44652c16 8338 *Yuval Yarom and Naomi Benger*
5f8e6c50 8339
257e9d03 8340### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
5f8e6c50 8341
44652c16
DMSP
8342 * Keep original DTLS digest and encryption contexts in retransmission
8343 structures so we can use the previous session parameters if they need
d8dc8538 8344 to be resent. ([CVE-2013-6450])
5f8e6c50
DMSP
8345
8346 *Steve Henson*
8347
44652c16
DMSP
8348 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8349 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8350 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8351 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8352 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8353 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
5f8e6c50 8354
44652c16 8355 *Rob Stradling, Adam Langley*
5f8e6c50 8356
257e9d03 8357### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
5f8e6c50 8358
44652c16 8359 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
5f8e6c50 8360
44652c16
DMSP
8361 This addresses the flaw in CBC record processing discovered by
8362 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
257e9d03 8363 at: <http://www.isg.rhul.ac.uk/tls/>
5f8e6c50 8364
44652c16
DMSP
8365 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8366 Security Group at Royal Holloway, University of London
8367 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8368 Emilia Käsper for the initial patch.
d8dc8538 8369 ([CVE-2013-0169])
5f8e6c50 8370
44652c16 8371 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
5f8e6c50 8372
44652c16 8373 * Return an error when checking OCSP signatures when key is NULL.
d8dc8538 8374 This fixes a DoS attack. ([CVE-2013-0166])
5f8e6c50
DMSP
8375
8376 *Steve Henson*
8377
44652c16
DMSP
8378 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8379 the right response is stapled. Also change SSL_get_certificate()
8380 so it returns the certificate actually sent.
257e9d03 8381 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
44652c16 8382 (This is a backport)
5f8e6c50 8383
44652c16 8384 *Rob Stradling <rob.stradling@comodo.com>*
5f8e6c50 8385
44652c16 8386 * Fix possible deadlock when decoding public keys.
5f8e6c50
DMSP
8387
8388 *Steve Henson*
8389
257e9d03 8390### Changes between 1.0.0i and 1.0.0j [10 May 2012]
5f8e6c50 8391
44652c16
DMSP
8392[NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8393OpenSSL 1.0.1.]
5f8e6c50 8394
44652c16
DMSP
8395 * Sanity check record length before skipping explicit IV in DTLS
8396 to fix DoS attack.
5f8e6c50 8397
44652c16
DMSP
8398 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8399 fuzzing as a service testing platform.
d8dc8538 8400 ([CVE-2012-2333])
5f8e6c50
DMSP
8401
8402 *Steve Henson*
8403
44652c16
DMSP
8404 * Initialise tkeylen properly when encrypting CMS messages.
8405 Thanks to Solar Designer of Openwall for reporting this issue.
5f8e6c50
DMSP
8406
8407 *Steve Henson*
8408
257e9d03 8409### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
5f8e6c50 8410
44652c16
DMSP
8411 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8412 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8413 in CRYPTO_realloc_clean.
5f8e6c50 8414
44652c16
DMSP
8415 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8416 issue and to Adam Langley <agl@chromium.org> for fixing it.
d8dc8538 8417 ([CVE-2012-2110])
5f8e6c50 8418
44652c16 8419 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
5f8e6c50 8420
257e9d03 8421### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
5f8e6c50
DMSP
8422
8423 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8424 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8425 content decryption and always return the same error. Note: this attack
8426 needs on average 2^20 messages so it only affects automated senders. The
8427 old behaviour can be re-enabled in the CMS code by setting the
8428 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8429 an MMA defence is not necessary.
8430 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
d8dc8538 8431 this issue. ([CVE-2012-0884])
5f8e6c50
DMSP
8432
8433 *Steve Henson*
8434
8435 * Fix CVE-2011-4619: make sure we really are receiving a
8436 client hello before rejecting multiple SGC restarts. Thanks to
8437 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8438
8439 *Steve Henson*
8440
257e9d03 8441### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
5f8e6c50
DMSP
8442
8443 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8444 Thanks to Antonio Martin, Enterprise Secure Access Research and
8445 Development, Cisco Systems, Inc. for discovering this bug and
d8dc8538 8446 preparing a fix. ([CVE-2012-0050])
5f8e6c50
DMSP
8447
8448 *Antonio Martin*
8449
257e9d03 8450### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
5f8e6c50
DMSP
8451
8452 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8453 of the Vaudenay padding oracle attack on CBC mode encryption
8454 which enables an efficient plaintext recovery attack against
8455 the OpenSSL implementation of DTLS. Their attack exploits timing
8456 differences arising during decryption processing. A research
8457 paper describing this attack can be found at:
257e9d03 8458 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
5f8e6c50
DMSP
8459 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8460 Security Group at Royal Holloway, University of London
8461 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8462 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
d8dc8538 8463 for preparing the fix. ([CVE-2011-4108])
5f8e6c50
DMSP
8464
8465 *Robin Seggelmann, Michael Tuexen*
8466
8467 * Clear bytes used for block padding of SSL 3.0 records.
d8dc8538 8468 ([CVE-2011-4576])
5f8e6c50
DMSP
8469
8470 *Adam Langley (Google)*
8471
8472 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8473 Kadianakis <desnacked@gmail.com> for discovering this issue and
d8dc8538 8474 Adam Langley for preparing the fix. ([CVE-2011-4619])
5f8e6c50
DMSP
8475
8476 *Adam Langley (Google)*
8477
d8dc8538 8478 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
5f8e6c50
DMSP
8479
8480 *Andrey Kulikov <amdeich@gmail.com>*
8481
8482 * Prevent malformed RFC3779 data triggering an assertion failure.
8483 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
d8dc8538 8484 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
5f8e6c50
DMSP
8485
8486 *Rob Austein <sra@hactrn.net>*
8487
8488 * Improved PRNG seeding for VOS.
8489
8490 *Paul Green <Paul.Green@stratus.com>*
8491
8492 * Fix ssl_ciph.c set-up race.
8493
8494 *Adam Langley (Google)*
8495
8496 * Fix spurious failures in ecdsatest.c.
8497
8498 *Emilia Käsper (Google)*
8499
8500 * Fix the BIO_f_buffer() implementation (which was mixing different
257e9d03 8501 interpretations of the `..._len` fields).
5f8e6c50
DMSP
8502
8503 *Adam Langley (Google)*
8504
8505 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8506 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8507 threads won't reuse the same blinding coefficients.
8508
8509 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8510 lock to call BN_BLINDING_invert_ex, and avoids one use of
8511 BN_BLINDING_update for each BN_BLINDING structure (previously,
8512 the last update always remained unused).
8513
8514 *Emilia Käsper (Google)*
8515
8516 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8517
8518 *Bob Buckholz (Google)*
8519
257e9d03 8520### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
5f8e6c50
DMSP
8521
8522 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
d8dc8538 8523 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
5f8e6c50
DMSP
8524
8525 *Kaspar Brand <ossl@velox.ch>*
8526
8527 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
d8dc8538 8528 for multi-threaded use of ECDH. ([CVE-2011-3210])
5f8e6c50
DMSP
8529
8530 *Adam Langley (Google)*
8531
8532 * Fix x509_name_ex_d2i memory leak on bad inputs.
8533
8534 *Bodo Moeller*
8535
8536 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8537 signature public key algorithm by using OID xref utilities instead.
8538 Before this you could only use some ECC ciphersuites with SHA1 only.
8539
8540 *Steve Henson*
8541
8542 * Add protection against ECDSA timing attacks as mentioned in the paper
8543 by Billy Bob Brumley and Nicola Tuveri, see:
257e9d03 8544 <http://eprint.iacr.org/2011/232.pdf>
5f8e6c50
DMSP
8545
8546 *Billy Bob Brumley and Nicola Tuveri*
8547
257e9d03 8548### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
5f8e6c50
DMSP
8549
8550 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8551
8552 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8553
8554 * Fix bug in string printing code: if *any* escaping is enabled we must
8555 escape the escape character (backslash) or the resulting string is
8556 ambiguous.
8557
8558 *Steve Henson*
8559
257e9d03 8560### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
5f8e6c50
DMSP
8561
8562 * Disable code workaround for ancient and obsolete Netscape browsers
8563 and servers: an attacker can use it in a ciphersuite downgrade attack.
8564 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8565
8566 *Steve Henson*
8567
8568 * Fixed J-PAKE implementation error, originally discovered by
8569 Sebastien Martini, further info and confirmation from Stefan
8570 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8571
8572 *Ben Laurie*
8573
257e9d03 8574### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
5f8e6c50
DMSP
8575
8576 * Fix extension code to avoid race conditions which can result in a buffer
8577 overrun vulnerability: resumed sessions must not be modified as they can
8578 be shared by multiple threads. CVE-2010-3864
8579
8580 *Steve Henson*
8581
8582 * Fix WIN32 build system to correctly link an ENGINE directory into
8583 a DLL.
8584
8585 *Steve Henson*
8586
257e9d03 8587### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
5f8e6c50
DMSP
8588
8589 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
d8dc8538 8590 ([CVE-2010-1633])
5f8e6c50
DMSP
8591
8592 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8593
257e9d03 8594### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
5f8e6c50
DMSP
8595
8596 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8597 context. The operation can be customised via the ctrl mechanism in
8598 case ENGINEs want to include additional functionality.
8599
8600 *Steve Henson*
8601
8602 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8603
8604 *Steve Henson*
8605
8606 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8607 output hashes compatible with older versions of OpenSSL.
8608
8609 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8610
8611 * Fix compression algorithm handling: if resuming a session use the
8612 compression algorithm of the resumed session instead of determining
8613 it from client hello again. Don't allow server to change algorithm.
8614
8615 *Steve Henson*
8616
ec2bfb7d 8617 * Add load_crls() function to commands tidying load_certs() too. Add option
5f8e6c50
DMSP
8618 to verify utility to allow additional CRLs to be included.
8619
8620 *Steve Henson*
8621
8622 * Update OCSP request code to permit adding custom headers to the request:
8623 some responders need this.
8624
8625 *Steve Henson*
8626
8627 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8628 correctly.
8629
8630 *Julia Lawall <julia@diku.dk>*
8631
ec2bfb7d 8632 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
5f8e6c50
DMSP
8633 needlessly dereferenced structures, used obsolete functions and
8634 didn't handle all updated verify codes correctly.
8635
8636 *Steve Henson*
8637
8638 * Disable MD2 in the default configuration.
8639
8640 *Steve Henson*
8641
8642 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8643 indicate the initial BIO being pushed or popped. This makes it possible
8644 to determine whether the BIO is the one explicitly called or as a result
8645 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8646 it handles reference counts correctly and doesn't zero out the I/O bio
8647 when it is not being explicitly popped. WARNING: applications which
8648 included workarounds for the old buggy behaviour will need to be modified
8649 or they could free up already freed BIOs.
8650
8651 *Steve Henson*
8652
8653 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8654 renaming to all platforms (within the 0.9.8 branch, this was
8655 done conditionally on Netware platforms to avoid a name clash).
8656
8657 *Guenter <lists@gknw.net>*
8658
8659 * Add ECDHE and PSK support to DTLS.
8660
8661 *Michael Tuexen <tuexen@fh-muenster.de>*
8662
8663 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8664 be used on C++.
8665
8666 *Steve Henson*
8667
8668 * Add "missing" function EVP_MD_flags() (without this the only way to
8669 retrieve a digest flags is by accessing the structure directly. Update
257e9d03 8670 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
5f8e6c50
DMSP
8671 or cipher is registered as in the "from" argument. Print out all
8672 registered digests in the dgst usage message instead of manually
8673 attempting to work them out.
8674
8675 *Steve Henson*
8676
8677 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8678 this allows the use of compression and extensions. Change default cipher
8679 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8680 by default unless an application cipher string requests it.
8681
8682 *Steve Henson*
8683
8684 * Alter match criteria in PKCS12_parse(). It used to try to use local
8685 key ids to find matching certificates and keys but some PKCS#12 files
8686 don't follow the (somewhat unwritten) rules and this strategy fails.
8687 Now just gather all certificates together and the first private key
8688 then look for the first certificate that matches the key.
8689
8690 *Steve Henson*
8691
8692 * Support use of registered digest and cipher names for dgst and cipher
8693 commands instead of having to add each one as a special case. So now
8694 you can do:
8695
8696 openssl sha256 foo
8697
8698 as well as:
8699
8700 openssl dgst -sha256 foo
8701
8702 and this works for ENGINE based algorithms too.
8703
5f8e6c50
DMSP
8704 *Steve Henson*
8705
8706 * Update Gost ENGINE to support parameter files.
8707
8708 *Victor B. Wagner <vitus@cryptocom.ru>*
8709
8710 * Support GeneralizedTime in ca utility.
8711
8712 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8713
8714 * Enhance the hash format used for certificate directory links. The new
8715 form uses the canonical encoding (meaning equivalent names will work
8716 even if they aren't identical) and uses SHA1 instead of MD5. This form
8717 is incompatible with the older format and as a result c_rehash should
8718 be used to rebuild symbolic links.
8719
8720 *Steve Henson*
8721
8722 * Make PKCS#8 the default write format for private keys, replacing the
8723 traditional format. This form is standardised, more secure and doesn't
8724 include an implicit MD5 dependency.
8725
8726 *Steve Henson*
8727
8728 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8729 committed to OpenSSL should pass this lot as a minimum.
8730
8731 *Steve Henson*
8732
8733 * Add session ticket override functionality for use by EAP-FAST.
8734
8735 *Jouni Malinen <j@w1.fi>*
8736
8737 * Modify HMAC functions to return a value. Since these can be implemented
8738 in an ENGINE errors can occur.
8739
8740 *Steve Henson*
8741
8742 * Type-checked OBJ_bsearch_ex.
8743
8744 *Ben Laurie*
8745
8746 * Type-checked OBJ_bsearch. Also some constification necessitated
8747 by type-checking. Still to come: TXT_DB, bsearch(?),
8748 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8749 CONF_VALUE.
8750
8751 *Ben Laurie*
8752
8753 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8754 seconds to a tm structure directly, instead of going through OS
8755 specific date routines. This avoids any issues with OS routines such
257e9d03 8756 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
5f8e6c50
DMSP
8757 and X509_time_adj_ex() to cover the extended range. The existing
8758 X509_time_adj() is still usable and will no longer have any date issues.
8759
8760 *Steve Henson*
8761
8762 * Delta CRL support. New use deltas option which will attempt to locate
8763 and search any appropriate delta CRLs available.
8764
8765 This work was sponsored by Google.
8766
8767 *Steve Henson*
8768
8769 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8770 code and add additional score elements. Validate alternate CRL paths
8771 as part of the CRL checking and indicate a new error "CRL path validation
8772 error" in this case. Applications wanting additional details can use
8773 the verify callback and check the new "parent" field. If this is not
8774 NULL CRL path validation is taking place. Existing applications won't
8775 see this because it requires extended CRL support which is off by
8776 default.
8777
8778 This work was sponsored by Google.
8779
8780 *Steve Henson*
8781
8782 * Support for freshest CRL extension.
8783
8784 This work was sponsored by Google.
8785
8786 *Steve Henson*
8787
8788 * Initial indirect CRL support. Currently only supported in the CRLs
8789 passed directly and not via lookup. Process certificate issuer
8790 CRL entry extension and lookup CRL entries by bother issuer name
8791 and serial number. Check and process CRL issuer entry in IDP extension.
8792
8793 This work was sponsored by Google.
8794
8795 *Steve Henson*
8796
8797 * Add support for distinct certificate and CRL paths. The CRL issuer
8798 certificate is validated separately in this case. Only enabled if
8799 an extended CRL support flag is set: this flag will enable additional
8800 CRL functionality in future.
8801
8802 This work was sponsored by Google.
8803
8804 *Steve Henson*
8805
8806 * Add support for policy mappings extension.
8807
8808 This work was sponsored by Google.
8809
8810 *Steve Henson*
8811
8812 * Fixes to pathlength constraint, self issued certificate handling,
8813 policy processing to align with RFC3280 and PKITS tests.
8814
8815 This work was sponsored by Google.
8816
8817 *Steve Henson*
8818
8819 * Support for name constraints certificate extension. DN, email, DNS
8820 and URI types are currently supported.
8821
8822 This work was sponsored by Google.
8823
8824 *Steve Henson*
8825
8826 * To cater for systems that provide a pointer-based thread ID rather
8827 than numeric, deprecate the current numeric thread ID mechanism and
8828 replace it with a structure and associated callback type. This
8829 mechanism allows a numeric "hash" to be extracted from a thread ID in
8830 either case, and on platforms where pointers are larger than 'long',
8831 mixing is done to help ensure the numeric 'hash' is usable even if it
8832 can't be guaranteed unique. The default mechanism is to use "&errno"
8833 as a pointer-based thread ID to distinguish between threads.
8834
8835 Applications that want to provide their own thread IDs should now use
8836 CRYPTO_THREADID_set_callback() to register a callback that will call
8837 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8838
8839 Note that ERR_remove_state() is now deprecated, because it is tied
8840 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8841 to free the current thread's error state should be replaced by
8842 ERR_remove_thread_state(NULL).
8843
8844 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8845 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8846 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8847 application was previously providing a numeric thread callback that
8848 was inappropriate for distinguishing threads, then uniqueness might
8849 have been obtained with &errno that happened immediately in the
8850 intermediate development versions of OpenSSL; this is no longer the
8851 case, the numeric thread callback will now override the automatic use
8852 of &errno.)
8853
8854 *Geoff Thorpe, with help from Bodo Moeller*
8855
8856 * Initial support for different CRL issuing certificates. This covers a
8857 simple case where the self issued certificates in the chain exist and
8858 the real CRL issuer is higher in the existing chain.
8859
8860 This work was sponsored by Google.
8861
8862 *Steve Henson*
8863
8864 * Removed effectively defunct crypto/store from the build.
8865
8866 *Ben Laurie*
8867
8868 * Revamp of STACK to provide stronger type-checking. Still to come:
8869 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8870 ASN1_STRING, CONF_VALUE.
8871
8872 *Ben Laurie*
8873
8874 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8875 RAM on SSL connections. This option can save about 34k per idle SSL.
8876
8877 *Nick Mathewson*
8878
8879 * Revamp of LHASH to provide stronger type-checking. Still to come:
8880 STACK, TXT_DB, bsearch, qsort.
8881
8882 *Ben Laurie*
8883
8884 * Initial support for Cryptographic Message Syntax (aka CMS) based
8885 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8886 support for data, signedData, compressedData, digestedData and
8887 encryptedData, envelopedData types included. Scripts to check against
8888 RFC4134 examples draft and interop and consistency checks of many
8889 content types and variants.
8890
8891 *Steve Henson*
8892
8893 * Add options to enc utility to support use of zlib compression BIO.
8894
8895 *Steve Henson*
8896
8897 * Extend mk1mf to support importing of options and assembly language
8898 files from Configure script, currently only included in VC-WIN32.
8899 The assembly language rules can now optionally generate the source
8900 files from the associated perl scripts.
8901
8902 *Steve Henson*
8903
8904 * Implement remaining functionality needed to support GOST ciphersuites.
8905 Interop testing has been performed using CryptoPro implementations.
8906
8907 *Victor B. Wagner <vitus@cryptocom.ru>*
8908
8909 * s390x assembler pack.
8910
8911 *Andy Polyakov*
8912
8913 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8914 "family."
8915
8916 *Andy Polyakov*
8917
8918 * Implement Opaque PRF Input TLS extension as specified in
8919 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8920 official specification yet and no extension type assignment by
8921 IANA exists, this extension (for now) will have to be explicitly
8922 enabled when building OpenSSL by providing the extension number
8923 to use. For example, specify an option
8924
8925 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8926
8927 to the "config" or "Configure" script to enable the extension,
8928 assuming extension number 0x9527 (which is a completely arbitrary
8929 and unofficial assignment based on the MD5 hash of the Internet
8930 Draft). Note that by doing so, you potentially lose
8931 interoperability with other TLS implementations since these might
8932 be using the same extension number for other purposes.
8933
8934 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8935 opaque PRF input value to use in the handshake. This will create
8936 an internal copy of the length-'len' string at 'src', and will
8937 return non-zero for success.
8938
8939 To get more control and flexibility, provide a callback function
8940 by using
8941
8942 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8943 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8944
8945 where
8946
8947 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8948 void *arg;
8949
8950 Callback function 'cb' will be called in handshakes, and is
8951 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8952 Argument 'arg' is for application purposes (the value as given to
8953 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8954 be provided to the callback function). The callback function
8955 has to return non-zero to report success: usually 1 to use opaque
8956 PRF input just if possible, or 2 to enforce use of the opaque PRF
8957 input. In the latter case, the library will abort the handshake
8958 if opaque PRF input is not successfully negotiated.
8959
8960 Arguments 'peerinput' and 'len' given to the callback function
8961 will always be NULL and 0 in the case of a client. A server will
8962 see the client's opaque PRF input through these variables if
8963 available (NULL and 0 otherwise). Note that if the server
8964 provides an opaque PRF input, the length must be the same as the
8965 length of the client's opaque PRF input.
8966
8967 Note that the callback function will only be called when creating
8968 a new session (session resumption can resume whatever was
8969 previously negotiated), and will not be called in SSL 2.0
8970 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8971 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8972 for applications that need to enforce opaque PRF input.
8973
5f8e6c50
DMSP
8974 *Bodo Moeller*
8975
8976 * Update ssl code to support digests other than SHA1+MD5 for handshake
8977 MAC.
8978
5f8e6c50
DMSP
8979 *Victor B. Wagner <vitus@cryptocom.ru>*
8980
8981 * Add RFC4507 support to OpenSSL. This includes the corrections in
8982 RFC4507bis. The encrypted ticket format is an encrypted encoded
8983 SSL_SESSION structure, that way new session features are automatically
8984 supported.
8985
8986 If a client application caches session in an SSL_SESSION structure
8987 support is transparent because tickets are now stored in the encoded
8988 SSL_SESSION.
8989
8990 The SSL_CTX structure automatically generates keys for ticket
8991 protection in servers so again support should be possible
8992 with no application modification.
8993
8994 If a client or server wishes to disable RFC4507 support then the option
8995 SSL_OP_NO_TICKET can be set.
8996
8997 Add a TLS extension debugging callback to allow the contents of any client
8998 or server extensions to be examined.
8999
9000 This work was sponsored by Google.
9001
9002 *Steve Henson*
9003
9004 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9005 OpenSSL should now compile cleanly on gcc 4.2
9006
9007 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9008
9009 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9010 support including streaming MAC support: this is required for GOST
9011 ciphersuite support.
9012
9013 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9014
9015 * Add option -stream to use PKCS#7 streaming in smime utility. New
9016 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9017 to output in BER and PEM format.
9018
9019 *Steve Henson*
9020
9021 * Experimental support for use of HMAC via EVP_PKEY interface. This
257e9d03 9022 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
5f8e6c50
DMSP
9023 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9024 ENGINE support for HMAC keys which are unextractable. New -mac and
9025 -macopt options to dgst utility.
9026
9027 *Steve Henson*
9028
9029 * New option -sigopt to dgst utility. Update dgst to use
257e9d03 9030 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
5f8e6c50
DMSP
9031 alternative signing parameters such as X9.31 or PSS in the dgst
9032 utility.
9033
9034 *Steve Henson*
9035
9036 * Change ssl_cipher_apply_rule(), the internal function that does
9037 the work each time a ciphersuite string requests enabling
9038 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9039 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9040 the order of disabled ciphersuites such that those ciphersuites
9041 that most recently went from enabled to disabled not only stay
9042 in order with respect to each other, but also have higher priority
9043 than other disabled ciphersuites the next time ciphersuites are
9044 enabled again.
9045
9046 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9047 the same ciphersuites as with "HIGH" alone, but in a specific
9048 order where the PSK ciphersuites come first (since they are the
9049 most recently disabled ciphersuites when "HIGH" is parsed).
9050
9051 Also, change ssl_create_cipher_list() (using this new
9052 functionality) such that between otherwise identical
9053 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9054 the default order.
9055
9056 *Bodo Moeller*
9057
9058 * Change ssl_create_cipher_list() so that it automatically
9059 arranges the ciphersuites in reasonable order before starting
9060 to process the rule string. Thus, the definition for "DEFAULT"
9061 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
ec2bfb7d 9062 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
5f8e6c50
DMSP
9063 This makes it much easier to arrive at a reasonable default order
9064 in applications for which anonymous ciphers are OK (meaning
9065 that you can't actually use DEFAULT).
9066
9067 *Bodo Moeller; suggested by Victor Duchovni*
9068
9069 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9070 processing) into multiple integers instead of setting
9071 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9072 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9073 (These masks as well as the individual bit definitions are hidden
9074 away into the non-exported interface ssl/ssl_locl.h, so this
9075 change to the definition of the SSL_CIPHER structure shouldn't
9076 affect applications.) This give us more bits for each of these
9077 categories, so there is no longer a need to coagulate AES128 and
9078 AES256 into a single algorithm bit, and to coagulate Camellia128
9079 and Camellia256 into a single algorithm bit, which has led to all
9080 kinds of kludges.
9081
9082 Thus, among other things, the kludge introduced in 0.9.7m and
9083 0.9.8e for masking out AES256 independently of AES128 or masking
9084 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9085
9086 With the change, we also introduce new ciphersuite aliases that
9087 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9088 "CAMELLIA256".
9089
9090 *Bodo Moeller*
9091
9092 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9093 Use the leftmost N bytes of the signature input if the input is
9094 larger than the prime q (with N being the size in bytes of q).
9095
9096 *Nils Larsch*
9097
9098 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9099 it yet and it is largely untested.
9100
9101 *Steve Henson*
9102
9103 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9104
9105 *Nils Larsch*
9106
9107 * Initial incomplete changes to avoid need for function casts in OpenSSL
9108 some compilers (gcc 4.2 and later) reject their use. Safestack is
9109 reimplemented. Update ASN1 to avoid use of legacy functions.
9110
9111 *Steve Henson*
9112
9113 * Win32/64 targets are linked with Winsock2.
9114
9115 *Andy Polyakov*
9116
9117 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9118 to external functions. This can be used to increase CRL handling
9119 efficiency especially when CRLs are very large by (for example) storing
9120 the CRL revoked certificates in a database.
9121
9122 *Steve Henson*
9123
9124 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9125 new CRLs added to a directory can be used. New command line option
9126 -verify_return_error to s_client and s_server. This causes real errors
9127 to be returned by the verify callback instead of carrying on no matter
9128 what. This reflects the way a "real world" verify callback would behave.
9129
9130 *Steve Henson*
9131
9132 * GOST engine, supporting several GOST algorithms and public key formats.
9133 Kindly donated by Cryptocom.
9134
9135 *Cryptocom*
9136
9137 * Partial support for Issuing Distribution Point CRL extension. CRLs
9138 partitioned by DP are handled but no indirect CRL or reason partitioning
9139 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9140 selected via a scoring technique which handles IDP and AKID in CRLs.
9141
9142 *Steve Henson*
9143
9144 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9145 will ultimately be used for all verify operations: this will remove the
9146 X509_STORE dependency on certificate verification and allow alternative
9147 lookup methods. X509_STORE based implementations of these two callbacks.
9148
9149 *Steve Henson*
9150
9151 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9152 Modify get_crl() to find a valid (unexpired) CRL if possible.
9153
9154 *Steve Henson*
9155
9156 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9157 this would be called X509_CRL_cmp() but that name is already used by
9158 a function that just compares CRL issuer names. Cache several CRL
9159 extensions in X509_CRL structure and cache CRLDP in X509.
9160
9161 *Steve Henson*
9162
9163 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9164 this maps equivalent X509_NAME structures into a consistent structure.
9165 Name comparison can then be performed rapidly using memcmp().
9166
9167 *Steve Henson*
9168
9169 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9170 utility.
9171
9172 *Steve Henson*
9173
9174 * Allow digests to supply their own micalg string for S/MIME type using
9175 the ctrl EVP_MD_CTRL_MICALG.
9176
9177 *Steve Henson*
9178
9179 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9180 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9181 ctrl. It can then customise the structure before and/or after signing
9182 if necessary.
9183
9184 *Steve Henson*
9185
9186 * New function OBJ_add_sigid() to allow application defined signature OIDs
9187 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9188 to free up any added signature OIDs.
9189
9190 *Steve Henson*
9191
9192 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9193 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9194 digest and cipher tables. New options added to openssl utility:
9195 list-message-digest-algorithms and list-cipher-algorithms.
9196
9197 *Steve Henson*
9198
9199 * Change the array representation of binary polynomials: the list
9200 of degrees of non-zero coefficients is now terminated with -1.
9201 Previously it was terminated with 0, which was also part of the
9202 value; thus, the array representation was not applicable to
9203 polynomials where t^0 has coefficient zero. This change makes
9204 the array representation useful in a more general context.
9205
9206 *Douglas Stebila*
9207
9208 * Various modifications and fixes to SSL/TLS cipher string
9209 handling. For ECC, the code now distinguishes between fixed ECDH
9210 with RSA certificates on the one hand and with ECDSA certificates
9211 on the other hand, since these are separate ciphersuites. The
9212 unused code for Fortezza ciphersuites has been removed.
9213
9214 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9215 (not "ECDHE"). For consistency with the code for DH
9216 certificates, use of ECDH certificates is now considered ECDH
9217 authentication, not RSA or ECDSA authentication (the latter is
9218 merely the CA's signing algorithm and not actively used in the
9219 protocol).
9220
9221 The temporary ciphersuite alias "ECCdraft" is no longer
9222 available, and ECC ciphersuites are no longer excluded from "ALL"
9223 and "DEFAULT". The following aliases now exist for RFC 4492
9224 ciphersuites, most of these by analogy with the DH case:
9225
9226 kECDHr - ECDH cert, signed with RSA
9227 kECDHe - ECDH cert, signed with ECDSA
9228 kECDH - ECDH cert (signed with either RSA or ECDSA)
9229 kEECDH - ephemeral ECDH
9230 ECDH - ECDH cert or ephemeral ECDH
9231
9232 aECDH - ECDH cert
9233 aECDSA - ECDSA cert
9234 ECDSA - ECDSA cert
9235
9236 AECDH - anonymous ECDH
9237 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9238
5f8e6c50
DMSP
9239 *Bodo Moeller*
9240
9241 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9242 Use correct micalg parameters depending on digest(s) in signed message.
9243
9244 *Steve Henson*
9245
9246 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9247 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9248
9249 *Steve Henson*
9250
9251 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9252 an engine to register a method. Add ENGINE lookups for methods and
9253 functional reference processing.
9254
9255 *Steve Henson*
9256
257e9d03
RS
9257 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9258 `EVP_{Sign,Verify}*` which allow an application to customise the signature
5f8e6c50
DMSP
9259 process.
9260
9261 *Steve Henson*
9262
9263 * New -resign option to smime utility. This adds one or more signers
9264 to an existing PKCS#7 signedData structure. Also -md option to use an
9265 alternative message digest algorithm for signing.
9266
9267 *Steve Henson*
9268
9269 * Tidy up PKCS#7 routines and add new functions to make it easier to
9270 create PKCS7 structures containing multiple signers. Update smime
9271 application to support multiple signers.
9272
9273 *Steve Henson*
9274
9275 * New -macalg option to pkcs12 utility to allow setting of an alternative
9276 digest MAC.
9277
9278 *Steve Henson*
9279
9280 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9281 Reorganize PBE internals to lookup from a static table using NIDs,
9282 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9283 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9284 PRF which will be automatically used with PBES2.
9285
9286 *Steve Henson*
9287
9288 * Replace the algorithm specific calls to generate keys in "req" with the
9289 new API.
9290
9291 *Steve Henson*
9292
9293 * Update PKCS#7 enveloped data routines to use new API. This is now
9294 supported by any public key method supporting the encrypt operation. A
9295 ctrl is added to allow the public key algorithm to examine or modify
9296 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9297 a no op.
9298
9299 *Steve Henson*
9300
9301 * Add a ctrl to asn1 method to allow a public key algorithm to express
9302 a default digest type to use. In most cases this will be SHA1 but some
9303 algorithms (such as GOST) need to specify an alternative digest. The
9304 return value indicates how strong the preference is 1 means optional and
9305 2 is mandatory (that is it is the only supported type). Modify
9306 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9307 use the default md. Update openssl utilities to use the default digest
9308 type for signing if it is not explicitly indicated.
9309
9310 *Steve Henson*
9311
9312 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9313 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9314 signing method from the key type. This effectively removes the link
9315 between digests and public key types.
9316
9317 *Steve Henson*
9318
9319 * Add an OID cross reference table and utility functions. Its purpose is to
9320 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9321 rsaEncryption. This will allow some of the algorithm specific hackery
9322 needed to use the correct OID to be removed.
9323
9324 *Steve Henson*
9325
9326 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9327 structures for PKCS7_sign(). They are now set up by the relevant public
9328 key ASN1 method.
9329
9330 *Steve Henson*
9331
9332 * Add provisional EC pkey method with support for ECDSA and ECDH.
9333
9334 *Steve Henson*
9335
9336 * Add support for key derivation (agreement) in the API, DH method and
9337 pkeyutl.
9338
9339 *Steve Henson*
9340
9341 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9342 public and private key formats. As a side effect these add additional
9343 command line functionality not previously available: DSA signatures can be
9344 generated and verified using pkeyutl and DH key support and generation in
9345 pkey, genpkey.
9346
9347 *Steve Henson*
9348
9349 * BeOS support.
9350
9351 *Oliver Tappe <zooey@hirschkaefer.de>*
9352
9353 * New make target "install_html_docs" installs HTML renditions of the
9354 manual pages.
9355
9356 *Oliver Tappe <zooey@hirschkaefer.de>*
9357
9358 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9359 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9360 support key and parameter generation and add initial key generation
9361 functionality for RSA.
9362
9363 *Steve Henson*
9364
9365 * Add functions for main EVP_PKEY_method operations. The undocumented
257e9d03
RS
9366 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9367 `EVP_PKEY_{encrypt,decrypt}_old`.
5f8e6c50
DMSP
9368
9369 *Steve Henson*
9370
9371 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9372 key API, doesn't do much yet.
9373
9374 *Steve Henson*
9375
9376 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9377 public key algorithms. New option to openssl utility:
9378 "list-public-key-algorithms" to print out info.
9379
9380 *Steve Henson*
9381
9382 * Implement the Supported Elliptic Curves Extension for
9383 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9384
9385 *Douglas Stebila*
9386
9387 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9388 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9389
9390 *Steve Henson*
9391
9392 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9393 utilities such as rsa, dsa, dsaparam etc except they process any key
9394 type.
9395
9396 *Steve Henson*
9397
9398 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9399 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9400 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9401 structure.
9402
9403 *Steve Henson*
9404
9405 * Initial support for pluggable public key ASN1.
9406 De-spaghettify the public key ASN1 handling. Move public and private
9407 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9408 algorithm specific handling to a single module within the relevant
9409 algorithm directory. Add functions to allow (near) opaque processing
9410 of public and private key structures.
9411
9412 *Steve Henson*
9413
9414 * Implement the Supported Point Formats Extension for
9415 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9416
9417 *Douglas Stebila*
9418
9419 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9420 for the psk identity [hint] and the psk callback functions to the
9421 SSL_SESSION, SSL and SSL_CTX structure.
9422
9423 New ciphersuites:
9424 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9425 PSK-AES256-CBC-SHA
9426
9427 New functions:
9428 SSL_CTX_use_psk_identity_hint
9429 SSL_get_psk_identity_hint
9430 SSL_get_psk_identity
9431 SSL_use_psk_identity_hint
9432
5f8e6c50
DMSP
9433 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9434
9435 * Add RFC 3161 compliant time stamp request creation, response generation
9436 and response verification functionality.
9437
9438 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9439
9440 * Add initial support for TLS extensions, specifically for the server_name
9441 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9442 have new members for a host name. The SSL data structure has an
257e9d03 9443 additional member `SSL_CTX *initial_ctx` so that new sessions can be
5f8e6c50
DMSP
9444 stored in that context to allow for session resumption, even after the
9445 SSL has been switched to a new SSL_CTX in reaction to a client's
9446 server_name extension.
9447
9448 New functions (subject to change):
9449
9450 SSL_get_servername()
9451 SSL_get_servername_type()
9452 SSL_set_SSL_CTX()
9453
9454 New CTRL codes and macros (subject to change):
9455
9456 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9457 - SSL_CTX_set_tlsext_servername_callback()
9458 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9459 - SSL_CTX_set_tlsext_servername_arg()
9460 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9461
9462 openssl s_client has a new '-servername ...' option.
9463
9464 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9465 '-key2 ...', '-servername_fatal' (subject to change). This allows
9466 testing the HostName extension for a specific single host name ('-cert'
9467 and '-key' remain fallbacks for handshakes without HostName
9468 negotiation). If the unrecognized_name alert has to be sent, this by
9469 default is a warning; it becomes fatal with the '-servername_fatal'
9470 option.
9471
5f8e6c50
DMSP
9472 *Peter Sylvester, Remy Allais, Christophe Renou*
9473
9474 * Whirlpool hash implementation is added.
9475
9476 *Andy Polyakov*
9477
9478 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9479 bn(64,32). Because of instruction set limitations it doesn't have
9480 any negative impact on performance. This was done mostly in order
9481 to make it possible to share assembler modules, such as bn_mul_mont
9482 implementations, between 32- and 64-bit builds without hassle.
9483
9484 *Andy Polyakov*
9485
9486 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9487 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9488 macro.
9489
9490 *Bodo Moeller*
9491
9492 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9493 dedicated Montgomery multiplication procedure, is introduced.
9494 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9495 "64-bit" performance on certain 32-bit targets.
9496
9497 *Andy Polyakov*
9498
9499 * New option SSL_OP_NO_COMP to disable use of compression selectively
9500 in SSL structures. New SSL ctrl to set maximum send fragment size.
9501 Save memory by setting the I/O buffer sizes dynamically instead of
9502 using the maximum available value.
9503
9504 *Steve Henson*
9505
9506 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9507 in addition to the text details.
9508
9509 *Bodo Moeller*
9510
9511 * Very, very preliminary EXPERIMENTAL support for printing of general
9512 ASN1 structures. This currently produces rather ugly output and doesn't
9513 handle several customised structures at all.
9514
9515 *Steve Henson*
9516
9517 * Integrated support for PVK file format and some related formats such
9518 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9519 these in the 'rsa' and 'dsa' utilities.
9520
9521 *Steve Henson*
9522
9523 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9524
9525 *Steve Henson*
9526
9527 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9528 place for the (very old) "NETSCAPE" format certificates which are now
9529 handled using new ASN1 code equivalents.
9530
9531 *Steve Henson*
9532
9533 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9534 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9535 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9536
9537 *Nils Larsch*
9538
9539 * Modify CRL distribution points extension code to print out previously
9540 unsupported fields. Enhance extension setting code to allow setting of
9541 all fields.
9542
9543 *Steve Henson*
9544
9545 * Add print and set support for Issuing Distribution Point CRL extension.
9546
9547 *Steve Henson*
9548
9549 * Change 'Configure' script to enable Camellia by default.
9550
9551 *NTT*
9552
44652c16
DMSP
9553OpenSSL 0.9.x
9554-------------
9555
257e9d03 9556### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
5f8e6c50
DMSP
9557
9558 * When rejecting SSL/TLS records due to an incorrect version number, never
9559 update s->server with a new major version number. As of
9560 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9561 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9562 the previous behavior could result in a read attempt at NULL when
9563 receiving specific incorrect SSL/TLS records once record payload
d8dc8538 9564 protection is active. ([CVE-2010-0740])
5f8e6c50
DMSP
9565
9566 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9567
9568 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9569 could be crashed if the relevant tables were not present (e.g. chrooted).
9570
9571 *Tomas Hoger <thoger@redhat.com>*
9572
257e9d03 9573### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
5f8e6c50 9574
d8dc8538 9575 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
5f8e6c50
DMSP
9576
9577 *Martin Olsson, Neel Mehta*
9578
9579 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9580 accommodate for stack sorting, always a write lock!).
9581
9582 *Bodo Moeller*
9583
9584 * On some versions of WIN32 Heap32Next is very slow. This can cause
9585 excessive delays in the RAND_poll(): over a minute. As a workaround
9586 include a time check in the inner Heap32Next loop too.
9587
9588 *Steve Henson*
9589
9590 * The code that handled flushing of data in SSL/TLS originally used the
9591 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9592 the problem outlined in PR#1949. The fix suggested there however can
9593 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9594 of Apache). So instead simplify the code to flush unconditionally.
9595 This should be fine since flushing with no data to flush is a no op.
9596
9597 *Steve Henson*
9598
9599 * Handle TLS versions 2.0 and later properly and correctly use the
9600 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9601 off ancient servers have a habit of sticking around for a while...
9602
9603 *Steve Henson*
9604
9605 * Modify compression code so it frees up structures without using the
9606 ex_data callbacks. This works around a problem where some applications
9607 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9608 restarting) then use compression (e.g. SSL with compression) later.
9609 This results in significant per-connection memory leaks and
9610 has caused some security issues including CVE-2008-1678 and
9611 CVE-2009-4355.
9612
9613 *Steve Henson*
9614
9615 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9616 change when encrypting or decrypting.
9617
9618 *Bodo Moeller*
9619
9620 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9621 connect and renegotiate with servers which do not support RI.
9622 Until RI is more widely deployed this option is enabled by default.
9623
9624 *Steve Henson*
9625
9626 * Add "missing" ssl ctrls to clear options and mode.
9627
9628 *Steve Henson*
9629
9630 * If client attempts to renegotiate and doesn't support RI respond with
9631 a no_renegotiation alert as required by RFC5746. Some renegotiating
9632 TLS clients will continue a connection gracefully when they receive
9633 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9634 waiting for a server hello which it will never receive. Now we treat a
9635 received no_renegotiation alert as a fatal error. This is because
9636 applications requesting a renegotiation might well expect it to succeed
9637 and would have no code in place to handle the server denying it so the
9638 only safe thing to do is to terminate the connection.
9639
9640 *Steve Henson*
9641
9642 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9643 peer supports secure renegotiation and 0 otherwise. Print out peer
9644 renegotiation support in s_client/s_server.
9645
9646 *Steve Henson*
9647
9648 * Replace the highly broken and deprecated SPKAC certification method with
9649 the updated NID creation version. This should correctly handle UTF8.
9650
9651 *Steve Henson*
9652
9653 * Implement RFC5746. Re-enable renegotiation but require the extension
9654 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9655 turns out to be a bad idea. It has been replaced by
9656 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9657 SSL_CTX_set_options(). This is really not recommended unless you
9658 know what you are doing.
9659
9660 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9661
9662 * Fixes to stateless session resumption handling. Use initial_ctx when
9663 issuing and attempting to decrypt tickets in case it has changed during
9664 servername handling. Use a non-zero length session ID when attempting
9665 stateless session resumption: this makes it possible to determine if
9666 a resumption has occurred immediately after receiving server hello
9667 (several places in OpenSSL subtly assume this) instead of later in
9668 the handshake.
9669
9670 *Steve Henson*
9671
9672 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9673 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9674 fixes for a few places where the return code is not checked
9675 correctly.
9676
9677 *Julia Lawall <julia@diku.dk>*
9678
9679 * Add --strict-warnings option to Configure script to include devteam
9680 warnings in other configurations.
9681
9682 *Steve Henson*
9683
9684 * Add support for --libdir option and LIBDIR variable in makefiles. This
9685 makes it possible to install openssl libraries in locations which
9686 have names other than "lib", for example "/usr/lib64" which some
9687 systems need.
9688
9689 *Steve Henson, based on patch from Jeremy Utley*
9690
9691 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9692 X690 8.9.12 and can produce some misleading textual output of OIDs.
9693
9694 *Steve Henson, reported by Dan Kaminsky*
9695
9696 * Delete MD2 from algorithm tables. This follows the recommendation in
9697 several standards that it is not used in new applications due to
9698 several cryptographic weaknesses. For binary compatibility reasons
9699 the MD2 API is still compiled in by default.
9700
9701 *Steve Henson*
9702
9703 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9704 and restored.
9705
9706 *Steve Henson*
9707
9708 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9709 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9710 clash.
9711
9712 *Guenter <lists@gknw.net>*
9713
9714 * Fix the server certificate chain building code to use X509_verify_cert(),
9715 it used to have an ad-hoc builder which was unable to cope with anything
9716 other than a simple chain.
9717
9718 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9719
9720 * Don't check self signed certificate signatures in X509_verify_cert()
9721 by default (a flag can override this): it just wastes time without
9722 adding any security. As a useful side effect self signed root CAs
9723 with non-FIPS digests are now usable in FIPS mode.
9724
9725 *Steve Henson*
9726
9727 * In dtls1_process_out_of_seq_message() the check if the current message
9728 is already buffered was missing. For every new message was memory
9729 allocated, allowing an attacker to perform an denial of service attack
9730 with sending out of seq handshake messages until there is no memory
9731 left. Additionally every future message was buffered, even if the
9732 sequence number made no sense and would be part of another handshake.
9733 So only messages with sequence numbers less than 10 in advance will be
d8dc8538 9734 buffered. ([CVE-2009-1378])
5f8e6c50
DMSP
9735
9736 *Robin Seggelmann, discovered by Daniel Mentz*
9737
9738 * Records are buffered if they arrive with a future epoch to be
9739 processed after finishing the corresponding handshake. There is
9740 currently no limitation to this buffer allowing an attacker to perform
9741 a DOS attack with sending records with future epochs until there is no
9742 memory left. This patch adds the pqueue_size() function to determine
9743 the size of a buffer and limits the record buffer to 100 entries.
d8dc8538 9744 ([CVE-2009-1377])
5f8e6c50
DMSP
9745
9746 *Robin Seggelmann, discovered by Daniel Mentz*
9747
9748 * Keep a copy of frag->msg_header.frag_len so it can be used after the
d8dc8538 9749 parent structure is freed. ([CVE-2009-1379])
5f8e6c50
DMSP
9750
9751 *Daniel Mentz*
9752
9753 * Handle non-blocking I/O properly in SSL_shutdown() call.
9754
9755 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9756
257e9d03 9757 * Add `2.5.4.*` OIDs
5f8e6c50
DMSP
9758
9759 *Ilya O. <vrghost@gmail.com>*
9760
257e9d03 9761### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
5f8e6c50
DMSP
9762
9763 * Disable renegotiation completely - this fixes a severe security
d8dc8538 9764 problem ([CVE-2009-3555]) at the cost of breaking all
5f8e6c50
DMSP
9765 renegotiation. Renegotiation can be re-enabled by setting
9766 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9767 run-time. This is really not recommended unless you know what
9768 you're doing.
9769
9770 *Ben Laurie*
9771
257e9d03 9772### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
5f8e6c50
DMSP
9773
9774 * Don't set val to NULL when freeing up structures, it is freed up by
257e9d03 9775 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
d8dc8538 9776 zeroing past the valid field. ([CVE-2009-0789])
5f8e6c50
DMSP
9777
9778 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9779
9780 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9781 checked correctly. This would allow some invalid signed attributes to
d8dc8538 9782 appear to verify correctly. ([CVE-2009-0591])
5f8e6c50
DMSP
9783
9784 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9785
9786 * Reject UniversalString and BMPString types with invalid lengths. This
9787 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
d8dc8538 9788 a legal length. ([CVE-2009-0590])
5f8e6c50
DMSP
9789
9790 *Steve Henson*
9791
9792 * Set S/MIME signing as the default purpose rather than setting it
9793 unconditionally. This allows applications to override it at the store
9794 level.
9795
9796 *Steve Henson*
9797
9798 * Permit restricted recursion of ASN1 strings. This is needed in practice
9799 to handle some structures.
9800
9801 *Steve Henson*
9802
9803 * Improve efficiency of mem_gets: don't search whole buffer each time
9804 for a '\n'
9805
9806 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9807
9808 * New -hex option for openssl rand.
9809
9810 *Matthieu Herrb*
9811
9812 * Print out UTF8String and NumericString when parsing ASN1.
9813
9814 *Steve Henson*
9815
9816 * Support NumericString type for name components.
9817
9818 *Steve Henson*
9819
9820 * Allow CC in the environment to override the automatically chosen
9821 compiler. Note that nothing is done to ensure flags work with the
9822 chosen compiler.
9823
9824 *Ben Laurie*
9825
257e9d03 9826### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
5f8e6c50
DMSP
9827
9828 * Properly check EVP_VerifyFinal() and similar return values
d8dc8538 9829 ([CVE-2008-5077]).
5f8e6c50
DMSP
9830
9831 *Ben Laurie, Bodo Moeller, Google Security Team*
9832
9833 * Enable TLS extensions by default.
9834
9835 *Ben Laurie*
9836
9837 * Allow the CHIL engine to be loaded, whether the application is
9838 multithreaded or not. (This does not release the developer from the
9839 obligation to set up the dynamic locking callbacks.)
9840
9841 *Sander Temme <sander@temme.net>*
9842
9843 * Use correct exit code if there is an error in dgst command.
9844
9845 *Steve Henson; problem pointed out by Roland Dirlewanger*
9846
9847 * Tweak Configure so that you need to say "experimental-jpake" to enable
9848 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9849
9850 *Bodo Moeller*
9851
9852 * Add experimental JPAKE support, including demo authentication in
9853 s_client and s_server.
9854
9855 *Ben Laurie*
9856
9857 * Set the comparison function in v3_addr_canonize().
9858
9859 *Rob Austein <sra@hactrn.net>*
9860
9861 * Add support for XMPP STARTTLS in s_client.
9862
9863 *Philip Paeps <philip@freebsd.org>*
9864
9865 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9866 to ensure that even with this option, only ciphersuites in the
9867 server's preference list will be accepted. (Note that the option
9868 applies only when resuming a session, so the earlier behavior was
9869 just about the algorithm choice for symmetric cryptography.)
9870
9871 *Bodo Moeller*
9872
257e9d03 9873### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
5f8e6c50
DMSP
9874
9875 * Fix NULL pointer dereference if a DTLS server received
d8dc8538 9876 ChangeCipherSpec as first record ([CVE-2009-1386]).
5f8e6c50
DMSP
9877
9878 *PR #1679*
9879
9880 * Fix a state transition in s3_srvr.c and d1_srvr.c
257e9d03 9881 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
5f8e6c50
DMSP
9882
9883 *Nagendra Modadugu*
9884
9885 * The fix in 0.9.8c that supposedly got rid of unsafe
9886 double-checked locking was incomplete for RSA blinding,
9887 addressing just one layer of what turns out to have been
9888 doubly unsafe triple-checked locking.
9889
9890 So now fix this for real by retiring the MONT_HELPER macro
9891 in crypto/rsa/rsa_eay.c.
9892
5f8e6c50
DMSP
9893 *Bodo Moeller; problem pointed out by Marius Schilder*
9894
9895 * Various precautionary measures:
9896
9897 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9898
9899 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9900 (NB: This would require knowledge of the secret session ticket key
9901 to exploit, in which case you'd be SOL either way.)
9902
9903 - Change bn_nist.c so that it will properly handle input BIGNUMs
9904 outside the expected range.
9905
9906 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9907 builds.
9908
5f8e6c50
DMSP
9909 *Neel Mehta, Bodo Moeller*
9910
9911 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9912 the load fails. Useful for distros.
9913
9914 *Ben Laurie and the FreeBSD team*
9915
9916 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9917
9918 *Steve Henson*
9919
9920 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9921
9922 *Huang Ying*
9923
9924 * Expand ENGINE to support engine supplied SSL client certificate functions.
9925
9926 This work was sponsored by Logica.
9927
9928 *Steve Henson*
9929
9930 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9931 keystores. Support for SSL/TLS client authentication too.
9932 Not compiled unless enable-capieng specified to Configure.
9933
9934 This work was sponsored by Logica.
9935
9936 *Steve Henson*
9937
9938 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9939 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9940 attribute creation routines such as certificate requests and PKCS#12
9941 files.
9942
9943 *Steve Henson*
9944
257e9d03 9945### Changes between 0.9.8g and 0.9.8h [28 May 2008]
5f8e6c50
DMSP
9946
9947 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9948 handshake which could lead to a client crash as found using the
d8dc8538 9949 Codenomicon TLS test suite ([CVE-2008-1672])
5f8e6c50
DMSP
9950
9951 *Steve Henson, Mark Cox*
9952
9953 * Fix double free in TLS server name extensions which could lead to
d8dc8538 9954 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
5f8e6c50
DMSP
9955
9956 *Joe Orton*
9957
9958 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9959
9960 Clear the error queue to ensure that error entries left from
9961 older function calls do not interfere with the correct operation.
9962
9963 *Lutz Jaenicke, Erik de Castro Lopo*
9964
9965 * Remove root CA certificates of commercial CAs:
9966
9967 The OpenSSL project does not recommend any specific CA and does not
9968 have any policy with respect to including or excluding any CA.
9969 Therefore it does not make any sense to ship an arbitrary selection
9970 of root CA certificates with the OpenSSL software.
9971
9972 *Lutz Jaenicke*
9973
9974 * RSA OAEP patches to fix two separate invalid memory reads.
9975 The first one involves inputs when 'lzero' is greater than
9976 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9977 before the beginning of from). The second one involves inputs where
9978 the 'db' section contains nothing but zeroes (there is a one-byte
9979 invalid read after the end of 'db').
9980
9981 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9982
9983 * Partial backport from 0.9.9-dev:
9984
9985 Introduce bn_mul_mont (dedicated Montgomery multiplication
9986 procedure) as a candidate for BIGNUM assembler implementation.
9987 While 0.9.9-dev uses assembler for various architectures, only
9988 x86_64 is available by default here in the 0.9.8 branch, and
9989 32-bit x86 is available through a compile-time setting.
9990
9991 To try the 32-bit x86 assembler implementation, use Configure
9992 option "enable-montasm" (which exists only for this backport).
9993
9994 As "enable-montasm" for 32-bit x86 disclaims code stability
9995 anyway, in this constellation we activate additional code
9996 backported from 0.9.9-dev for further performance improvements,
9997 namely BN_from_montgomery_word. (To enable this otherwise,
44652c16 9998 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
5f8e6c50 9999
5f8e6c50
DMSP
10000 *Andy Polyakov (backport partially by Bodo Moeller)*
10001
10002 * Add TLS session ticket callback. This allows an application to set
10003 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10004 values. This is useful for key rollover for example where several key
10005 sets may exist with different names.
10006
10007 *Steve Henson*
10008
10009 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10010 This was broken until now in 0.9.8 releases, such that the only way
10011 a registered ENGINE could be used (assuming it initialises
10012 successfully on the host) was to explicitly set it as the default
10013 for the relevant algorithms. This is in contradiction with 0.9.7
10014 behaviour and the documentation. With this fix, when an ENGINE is
10015 registered into a given algorithm's table of implementations, the
10016 'uptodate' flag is reset so that auto-discovery will be used next
10017 time a new context for that algorithm attempts to select an
10018 implementation.
10019
10020 *Ian Lister (tweaked by Geoff Thorpe)*
10021
10022 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10023 implementation in the following ways:
10024
10025 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10026 hard coded.
10027
10028 Lack of BER streaming support means one pass streaming processing is
10029 only supported if data is detached: setting the streaming flag is
10030 ignored for embedded content.
10031
10032 CMS support is disabled by default and must be explicitly enabled
10033 with the enable-cms configuration option.
10034
10035 *Steve Henson*
10036
10037 * Update the GMP engine glue to do direct copies between BIGNUM and
10038 mpz_t when openssl and GMP use the same limb size. Otherwise the
10039 existing "conversion via a text string export" trick is still used.
10040
10041 *Paul Sheer <paulsheer@gmail.com>*
10042
10043 * Zlib compression BIO. This is a filter BIO which compressed and
10044 uncompresses any data passed through it.
10045
10046 *Steve Henson*
10047
10048 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10049 RFC3394 compatible AES key wrapping.
10050
10051 *Steve Henson*
10052
10053 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10054 sets string data without copying. X509_ALGOR_set0() and
10055 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10056 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10057 from an X509_ATTRIBUTE structure optionally checking it occurs only
10058 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10059 data.
10060
10061 *Steve Henson*
10062
10063 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10064 to get the expected BN_FLG_CONSTTIME behavior.
10065
10066 *Bodo Moeller (Google)*
10067
10068 * Netware support:
10069
10070 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10071 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10072 - added some more tests to do_tests.pl
10073 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10074 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10075 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10076 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10077 - various changes to netware.pl to enable gcc-cross builds on Win32
10078 platform
10079 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10080 - various changes to fix missing prototype warnings
10081 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10082 - added AES, WHIRLPOOL and CPUID assembler code to build files
10083 - added missing AES assembler make rules to mk1mf.pl
ec2bfb7d 10084 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
5f8e6c50
DMSP
10085
10086 *Guenter Knauf <eflash@gmx.net>*
10087
10088 * Implement certificate status request TLS extension defined in RFC3546.
10089 A client can set the appropriate parameters and receive the encoded
10090 OCSP response via a callback. A server can query the supplied parameters
10091 and set the encoded OCSP response in the callback. Add simplified examples
10092 to s_client and s_server.
10093
10094 *Steve Henson*
10095
257e9d03 10096### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5f8e6c50
DMSP
10097
10098 * Fix various bugs:
10099 + Binary incompatibility of ssl_ctx_st structure
10100 + DTLS interoperation with non-compliant servers
10101 + Don't call get_session_cb() without proposed session
10102 + Fix ia64 assembler code
10103
10104 *Andy Polyakov, Steve Henson*
10105
257e9d03 10106### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5f8e6c50
DMSP
10107
10108 * DTLS Handshake overhaul. There were longstanding issues with
10109 OpenSSL DTLS implementation, which were making it impossible for
10110 RFC 4347 compliant client to communicate with OpenSSL server.
10111 Unfortunately just fixing these incompatibilities would "cut off"
10112 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10113 server keeps tolerating non RFC compliant syntax. The opposite is
10114 not true, 0.9.8f client can not communicate with earlier server.
10115 This update even addresses CVE-2007-4995.
10116
10117 *Andy Polyakov*
10118
10119 * Changes to avoid need for function casts in OpenSSL: some compilers
10120 (gcc 4.2 and later) reject their use.
10121 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10122 Steve Henson*
10123
10124 * Add RFC4507 support to OpenSSL. This includes the corrections in
10125 RFC4507bis. The encrypted ticket format is an encrypted encoded
10126 SSL_SESSION structure, that way new session features are automatically
10127 supported.
10128
10129 If a client application caches session in an SSL_SESSION structure
10130 support is transparent because tickets are now stored in the encoded
10131 SSL_SESSION.
10132
10133 The SSL_CTX structure automatically generates keys for ticket
10134 protection in servers so again support should be possible
10135 with no application modification.
10136
10137 If a client or server wishes to disable RFC4507 support then the option
10138 SSL_OP_NO_TICKET can be set.
10139
10140 Add a TLS extension debugging callback to allow the contents of any client
10141 or server extensions to be examined.
10142
10143 This work was sponsored by Google.
10144
10145 *Steve Henson*
10146
10147 * Add initial support for TLS extensions, specifically for the server_name
10148 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10149 have new members for a host name. The SSL data structure has an
257e9d03 10150 additional member `SSL_CTX *initial_ctx` so that new sessions can be
5f8e6c50
DMSP
10151 stored in that context to allow for session resumption, even after the
10152 SSL has been switched to a new SSL_CTX in reaction to a client's
10153 server_name extension.
10154
10155 New functions (subject to change):
10156
10157 SSL_get_servername()
10158 SSL_get_servername_type()
10159 SSL_set_SSL_CTX()
10160
10161 New CTRL codes and macros (subject to change):
10162
10163 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10164 - SSL_CTX_set_tlsext_servername_callback()
10165 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10166 - SSL_CTX_set_tlsext_servername_arg()
10167 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10168
10169 openssl s_client has a new '-servername ...' option.
10170
10171 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10172 '-key2 ...', '-servername_fatal' (subject to change). This allows
10173 testing the HostName extension for a specific single host name ('-cert'
10174 and '-key' remain fallbacks for handshakes without HostName
10175 negotiation). If the unrecognized_name alert has to be sent, this by
10176 default is a warning; it becomes fatal with the '-servername_fatal'
10177 option.
10178
5f8e6c50
DMSP
10179 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10180
10181 * Add AES and SSE2 assembly language support to VC++ build.
10182
10183 *Steve Henson*
10184
10185 * Mitigate attack on final subtraction in Montgomery reduction.
10186
10187 *Andy Polyakov*
10188
10189 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10190 (which previously caused an internal error).
10191
10192 *Bodo Moeller*
10193
10194 * Squeeze another 10% out of IGE mode when in != out.
10195
10196 *Ben Laurie*
10197
10198 * AES IGE mode speedup.
10199
10200 *Dean Gaudet (Google)*
10201
10202 * Add the Korean symmetric 128-bit cipher SEED (see
257e9d03 10203 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
5f8e6c50
DMSP
10204 add SEED ciphersuites from RFC 4162:
10205
10206 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10207 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10208 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10209 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10210
10211 To minimize changes between patchlevels in the OpenSSL 0.9.8
10212 series, SEED remains excluded from compilation unless OpenSSL
10213 is configured with 'enable-seed'.
10214
10215 *KISA, Bodo Moeller*
10216
10217 * Mitigate branch prediction attacks, which can be practical if a
10218 single processor is shared, allowing a spy process to extract
10219 information. For detailed background information, see
257e9d03 10220 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
5f8e6c50
DMSP
10221 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10222 and Necessary Software Countermeasures"). The core of the change
10223 are new versions BN_div_no_branch() and
10224 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10225 respectively, which are slower, but avoid the security-relevant
10226 conditional branches. These are automatically called by BN_div()
10227 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10228 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10229 remove a conditional branch.
10230
10231 BN_FLG_CONSTTIME is the new name for the previous
10232 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10233 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10234 in the exponent causes BN_mod_exp_mont() to use the alternative
10235 implementation in BN_mod_exp_mont_consttime().) The old name
10236 remains as a deprecated alias.
10237
10238 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10239 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10240 constant-time implementations for more than just exponentiation.
10241 Here too the old name is kept as a deprecated alias.
10242
10243 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10244 the BN_BLINDING structure gets an independent copy of the
257e9d03 10245 modulus. This means that the previous `BIGNUM *m` argument to
5f8e6c50 10246 BN_BLINDING_new() and to BN_BLINDING_create_param() now
257e9d03 10247 essentially becomes `const BIGNUM *m`, although we can't actually
5f8e6c50
DMSP
10248 change this in the header file before 0.9.9. It allows
10249 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10250 enable BN_FLG_CONSTTIME.
10251
5f8e6c50
DMSP
10252 *Matthew D Wood (Intel Corp)*
10253
10254 * In the SSL/TLS server implementation, be strict about session ID
10255 context matching (which matters if an application uses a single
10256 external cache for different purposes). Previously,
10257 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10258 set. This did ensure strict client verification, but meant that,
10259 with applications using a single external cache for quite
10260 different requirements, clients could circumvent ciphersuite
10261 restrictions for a given session ID context by starting a session
10262 in a different context.
10263
10264 *Bodo Moeller*
10265
10266 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10267 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10268 authentication-only ciphersuites.
10269
10270 *Bodo Moeller*
10271
10272 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10273 not complete and could lead to a possible single byte overflow
d8dc8538 10274 ([CVE-2007-5135]) [Ben Laurie]
5f8e6c50 10275
257e9d03 10276### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
5f8e6c50
DMSP
10277
10278 * Since AES128 and AES256 (and similarly Camellia128 and
10279 Camellia256) share a single mask bit in the logic of
10280 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10281 kludge to work properly if AES128 is available and AES256 isn't
10282 (or if Camellia128 is available and Camellia256 isn't).
10283
10284 *Victor Duchovni*
10285
10286 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10287 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10288 When a point or a seed is encoded in a BIT STRING, we need to
10289 prevent the removal of trailing zero bits to get the proper DER
10290 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10291 of a NamedBitList, for which trailing 0 bits need to be removed.)
10292
10293 *Bodo Moeller*
10294
10295 * Have SSL/TLS server implementation tolerate "mismatched" record
10296 protocol version while receiving ClientHello even if the
10297 ClientHello is fragmented. (The server can't insist on the
10298 particular protocol version it has chosen before the ServerHello
10299 message has informed the client about his choice.)
10300
10301 *Bodo Moeller*
10302
10303 * Add RFC 3779 support.
10304
10305 *Rob Austein for ARIN, Ben Laurie*
10306
10307 * Load error codes if they are not already present instead of using a
10308 static variable. This allows them to be cleanly unloaded and reloaded.
10309 Improve header file function name parsing.
10310
10311 *Steve Henson*
10312
10313 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10314 or CAPABILITY handshake as required by RFCs.
10315
10316 *Goetz Babin-Ebell*
10317
257e9d03 10318### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5f8e6c50
DMSP
10319
10320 * Introduce limits to prevent malicious keys being able to
d8dc8538 10321 cause a denial of service. ([CVE-2006-2940])
5f8e6c50
DMSP
10322
10323 *Steve Henson, Bodo Moeller*
10324
10325 * Fix ASN.1 parsing of certain invalid structures that can result
d8dc8538 10326 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
5f8e6c50
DMSP
10327
10328 * Fix buffer overflow in SSL_get_shared_ciphers() function.
d8dc8538 10329 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
5f8e6c50
DMSP
10330
10331 * Fix SSL client code which could crash if connecting to a
d8dc8538 10332 malicious SSLv2 server. ([CVE-2006-4343])
5f8e6c50
DMSP
10333
10334 *Tavis Ormandy and Will Drewry, Google Security Team*
10335
10336 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10337 match only those. Before that, "AES256-SHA" would be interpreted
10338 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10339 the same strength classification in 0.9.7h) as we currently only
10340 have a single AES bit in the ciphersuite description bitmap.
10341 That change, however, also applied to ciphersuite strings such as
10342 "RC4-MD5" that intentionally matched multiple ciphersuites --
10343 namely, SSL 2.0 ciphersuites in addition to the more common ones
10344 from SSL 3.0/TLS 1.0.
10345
10346 So we change the selection algorithm again: Naming an explicit
10347 ciphersuite selects this one ciphersuite, and any other similar
10348 ciphersuite (same bitmap) from *other* protocol versions.
10349 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10350 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10351
10352 Since SSL 2.0 does not have any ciphersuites for which the
10353 128/256 bit distinction would be relevant, this works for now.
10354 The proper fix will be to use different bits for AES128 and
10355 AES256, which would have avoided the problems from the beginning;
10356 however, bits are scarce, so we can only do this in a new release
10357 (not just a patchlevel) when we can change the SSL_CIPHER
10358 definition to split the single 'unsigned long mask' bitmap into
10359 multiple values to extend the available space.
10360
5f8e6c50
DMSP
10361 *Bodo Moeller*
10362
257e9d03 10363### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5f8e6c50
DMSP
10364
10365 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
d8dc8538 10366 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
5f8e6c50
DMSP
10367
10368 * Add AES IGE and biIGE modes.
10369
10370 *Ben Laurie*
10371
10372 * Change the Unix randomness entropy gathering to use poll() when
10373 possible instead of select(), since the latter has some
10374 undesirable limitations.
10375
10376 *Darryl Miles via Richard Levitte and Bodo Moeller*
10377
10378 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10379 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10380 cannot be implicitly activated as part of, e.g., the "AES" alias.
10381 However, please upgrade to OpenSSL 0.9.9[-dev] for
10382 non-experimental use of the ECC ciphersuites to get TLS extension
10383 support, which is required for curve and point format negotiation
10384 to avoid potential handshake problems.
10385
10386 *Bodo Moeller*
10387
10388 * Disable rogue ciphersuites:
10389
257e9d03
RS
10390 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10391 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10392 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5f8e6c50
DMSP
10393
10394 The latter two were purportedly from
10395 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10396 appear there.
10397
10398 Also deactivate the remaining ciphersuites from
10399 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10400 unofficial, and the ID has long expired.
10401
10402 *Bodo Moeller*
10403
10404 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10405 dual-core machines) and other potential thread-safety issues.
10406
10407 *Bodo Moeller*
10408
10409 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10410 versions), which is now available for royalty-free use
257e9d03 10411 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
5f8e6c50
DMSP
10412 Also, add Camellia TLS ciphersuites from RFC 4132.
10413
10414 To minimize changes between patchlevels in the OpenSSL 0.9.8
10415 series, Camellia remains excluded from compilation unless OpenSSL
10416 is configured with 'enable-camellia'.
10417
10418 *NTT*
10419
10420 * Disable the padding bug check when compression is in use. The padding
10421 bug check assumes the first packet is of even length, this is not
10422 necessarily true if compression is enabled and can result in false
10423 positives causing handshake failure. The actual bug test is ancient
10424 code so it is hoped that implementations will either have fixed it by
10425 now or any which still have the bug do not support compression.
10426
10427 *Steve Henson*
10428
257e9d03 10429### Changes between 0.9.8a and 0.9.8b [04 May 2006]
5f8e6c50
DMSP
10430
10431 * When applying a cipher rule check to see if string match is an explicit
10432 cipher suite and only match that one cipher suite if it is.
10433
10434 *Steve Henson*
10435
10436 * Link in manifests for VC++ if needed.
10437
10438 *Austin Ziegler <halostatue@gmail.com>*
10439
10440 * Update support for ECC-based TLS ciphersuites according to
10441 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10442 TLS extensions, which are supported starting with the 0.9.9
10443 branch, not in the OpenSSL 0.9.8 branch).
10444
10445 *Douglas Stebila*
10446
10447 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10448 opaque EVP_CIPHER_CTX handling.
10449
10450 *Steve Henson*
10451
10452 * Fixes and enhancements to zlib compression code. We now only use
44652c16 10453 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
5f8e6c50 10454 to conform with the standards mentioned here:
257e9d03 10455 <http://www.zlib.net/DLL_FAQ.txt>
5f8e6c50
DMSP
10456 Static zlib linking now works on Windows and the new --with-zlib-include
10457 --with-zlib-lib options to Configure can be used to supply the location
10458 of the headers and library. Gracefully handle case where zlib library
10459 can't be loaded.
10460
10461 *Steve Henson*
10462
10463 * Several fixes and enhancements to the OID generation code. The old code
10464 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10465 handle numbers larger than ULONG_MAX, truncated printing and had a
10466 non standard OBJ_obj2txt() behaviour.
10467
10468 *Steve Henson*
10469
10470 * Add support for building of engines under engine/ as shared libraries
10471 under VC++ build system.
10472
10473 *Steve Henson*
10474
10475 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10476 Hopefully, we will not see any false combination of paths any more.
10477
10478 *Richard Levitte*
10479
257e9d03 10480### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5f8e6c50
DMSP
10481
10482 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10483 (part of SSL_OP_ALL). This option used to disable the
10484 countermeasure against man-in-the-middle protocol-version
10485 rollback in the SSL 2.0 server implementation, which is a bad
d8dc8538 10486 idea. ([CVE-2005-2969])
5f8e6c50
DMSP
10487
10488 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10489 for Information Security, National Institute of Advanced Industrial
257e9d03 10490 Science and Technology [AIST], Japan)*
5f8e6c50
DMSP
10491
10492 * Add two function to clear and return the verify parameter flags.
10493
10494 *Steve Henson*
10495
10496 * Keep cipherlists sorted in the source instead of sorting them at
10497 runtime, thus removing the need for a lock.
10498
10499 *Nils Larsch*
10500
10501 * Avoid some small subgroup attacks in Diffie-Hellman.
10502
10503 *Nick Mathewson and Ben Laurie*
10504
10505 * Add functions for well-known primes.
10506
10507 *Nick Mathewson*
10508
10509 * Extended Windows CE support.
10510
10511 *Satoshi Nakamura and Andy Polyakov*
10512
10513 * Initialize SSL_METHOD structures at compile time instead of during
10514 runtime, thus removing the need for a lock.
10515
10516 *Steve Henson*
10517
10518 * Make PKCS7_decrypt() work even if no certificate is supplied by
10519 attempting to decrypt each encrypted key in turn. Add support to
10520 smime utility.
10521
10522 *Steve Henson*
10523
257e9d03 10524### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5f8e6c50
DMSP
10525
10526[NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10527OpenSSL 0.9.8.]
10528
10529 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10530
10531 *Richard Levitte*
10532
10533 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10534 key into the same file any more.
10535
10536 *Richard Levitte*
10537
10538 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10539
10540 *Andy Polyakov*
10541
10542 * Add -utf8 command line and config file option to 'ca'.
10543
10544 *Stefan <stf@udoma.org*
10545
10546 * Removed the macro des_crypt(), as it seems to conflict with some
10547 libraries. Use DES_crypt().
10548
10549 *Richard Levitte*
10550
10551 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10552 involves renaming the source and generated shared-libs for
10553 both. The engines will accept the corrected or legacy ids
10554 ('ncipher' and '4758_cca' respectively) when binding. NB,
10555 this only applies when building 'shared'.
10556
10557 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10558
10559 * Add attribute functions to EVP_PKEY structure. Modify
10560 PKCS12_create() to recognize a CSP name attribute and
10561 use it. Make -CSP option work again in pkcs12 utility.
10562
10563 *Steve Henson*
10564
10565 * Add new functionality to the bn blinding code:
10566 - automatic re-creation of the BN_BLINDING parameters after
10567 a fixed number of uses (currently 32)
10568 - add new function for parameter creation
10569 - introduce flags to control the update behaviour of the
10570 BN_BLINDING parameters
10571 - hide BN_BLINDING structure
10572 Add a second BN_BLINDING slot to the RSA structure to improve
10573 performance when a single RSA object is shared among several
10574 threads.
10575
10576 *Nils Larsch*
10577
10578 * Add support for DTLS.
10579
10580 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10581
10582 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10583 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10584
10585 *Walter Goulet*
10586
10587 * Remove buggy and incomplete DH cert support from
10588 ssl/ssl_rsa.c and ssl/s3_both.c
10589
10590 *Nils Larsch*
10591
10592 * Use SHA-1 instead of MD5 as the default digest algorithm for
ec2bfb7d 10593 the `apps/openssl` commands.
5f8e6c50
DMSP
10594
10595 *Nils Larsch*
10596
10597 * Compile clean with "-Wall -Wmissing-prototypes
10598 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10599 DEBUG_SAFESTACK must also be set.
10600
10601 *Ben Laurie*
10602
10603 * Change ./Configure so that certain algorithms can be disabled by default.
10604 The new counterpiece to "no-xxx" is "enable-xxx".
10605
10606 The patented RC5 and MDC2 algorithms will now be disabled unless
10607 "enable-rc5" and "enable-mdc2", respectively, are specified.
10608
10609 (IDEA remains enabled despite being patented. This is because IDEA
10610 is frequently required for interoperability, and there is no license
10611 fee for non-commercial use. As before, "no-idea" can be used to
10612 avoid this algorithm.)
10613
5f8e6c50
DMSP
10614 *Bodo Moeller*
10615
10616 * Add processing of proxy certificates (see RFC 3820). This work was
10617 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10618 EGEE (Enabling Grids for E-science in Europe).
10619
10620 *Richard Levitte*
10621
10622 * RC4 performance overhaul on modern architectures/implementations, such
10623 as Intel P4, IA-64 and AMD64.
10624
10625 *Andy Polyakov*
10626
10627 * New utility extract-section.pl. This can be used specify an alternative
10628 section number in a pod file instead of having to treat each file as
10629 a separate case in Makefile. This can be done by adding two lines to the
10630 pod file:
10631
10632 =for comment openssl_section:XXX
10633
10634 The blank line is mandatory.
10635
5f8e6c50
DMSP
10636 *Steve Henson*
10637
10638 * New arguments -certform, -keyform and -pass for s_client and s_server
10639 to allow alternative format key and certificate files and passphrase
10640 sources.
10641
10642 *Steve Henson*
10643
10644 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10645 update associated structures and add various utility functions.
10646
10647 Add new policy related verify parameters, include policy checking in
10648 standard verify code. Enhance 'smime' application with extra parameters
10649 to support policy checking and print out.
10650
10651 *Steve Henson*
10652
10653 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10654 Nehemiah processors. These extensions support AES encryption in hardware
10655 as well as RNG (though RNG support is currently disabled).
10656
10657 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10658
257e9d03 10659 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
5f8e6c50
DMSP
10660
10661 *Geoff Thorpe*
10662
10663 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10664
10665 *Andy Polyakov and a number of other people*
10666
10667 * Improved PowerPC platform support. Most notably BIGNUM assembler
10668 implementation contributed by IBM.
10669
10670 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10671
10672 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10673 exponent rather than 'unsigned long'. There is a corresponding change to
10674 the new 'rsa_keygen' element of the RSA_METHOD structure.
10675
10676 *Jelte Jansen, Geoff Thorpe*
10677
10678 * Functionality for creating the initial serial number file is now
10679 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10680
10681 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10682 number file to 1, which is bound to cause problems. To avoid
10683 the problems while respecting compatibility between different 0.9.7
10684 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10685 CA.pl for serial number initialization. With the new release 0.9.8,
10686 we can fix the problem directly in the 'ca' utility.)
10687
10688 *Steve Henson*
10689
10690 * Reduced header interdependencies by declaring more opaque objects in
10691 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10692 give fewer recursive includes, which could break lazy source code - so
10693 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10694 developers should define this symbol when building and using openssl to
10695 ensure they track the recommended behaviour, interfaces, [etc], but
10696 backwards-compatible behaviour prevails when this isn't defined.
10697
10698 *Geoff Thorpe*
10699
10700 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10701
10702 *Steve Henson*
10703
10704 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10705 This will generate a random key of the appropriate length based on the
10706 cipher context. The EVP_CIPHER can provide its own random key generation
10707 routine to support keys of a specific form. This is used in the des and
10708 3des routines to generate a key of the correct parity. Update S/MIME
10709 code to use new functions and hence generate correct parity DES keys.
10710 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10711 valid (weak or incorrect parity).
10712
10713 *Steve Henson*
10714
10715 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10716 as looking them up. This is useful when the verified structure may contain
10717 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10718 present unless the new PKCS7_NO_CRL flag is asserted.
10719
10720 *Steve Henson*
10721
10722 * Extend ASN1 oid configuration module. It now additionally accepts the
10723 syntax:
10724
10725 shortName = some long name, 1.2.3.4
10726
10727 *Steve Henson*
10728
10729 * Reimplemented the BN_CTX implementation. There is now no more static
10730 limitation on the number of variables it can handle nor the depth of the
10731 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10732 information can now expand as required, and rather than having a single
10733 static array of bignums, BN_CTX now uses a linked-list of such arrays
10734 allowing it to expand on demand whilst maintaining the usefulness of
10735 BN_CTX's "bundling".
10736
10737 *Geoff Thorpe*
10738
10739 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10740 to allow all RSA operations to function using a single BN_CTX.
10741
10742 *Geoff Thorpe*
10743
10744 * Preliminary support for certificate policy evaluation and checking. This
10745 is initially intended to pass the tests outlined in "Conformance Testing
10746 of Relying Party Client Certificate Path Processing Logic" v1.07.
10747
10748 *Steve Henson*
10749
10750 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10751 remained unused and not that useful. A variety of other little bignum
10752 tweaks and fixes have also been made continuing on from the audit (see
10753 below).
10754
10755 *Geoff Thorpe*
10756
10757 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10758 associated ASN1, EVP and SSL functions and old ASN1 macros.
10759
10760 *Richard Levitte*
10761
10762 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10763 and this should never fail. So the return value from the use of
10764 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10765 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10766
10767 *Geoff Thorpe*
10768
10769 * BN_CTX_get() should return zero-valued bignums, providing the same
10770 initialised value as BN_new().
10771
10772 *Geoff Thorpe, suggested by Ulf Möller*
10773
10774 * Support for inhibitAnyPolicy certificate extension.
10775
10776 *Steve Henson*
10777
10778 * An audit of the BIGNUM code is underway, for which debugging code is
10779 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10780 is considered valid when processing BIGNUMs, and causes execution to
10781 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10782 further steps are taken to deliberately pollute unused data in BIGNUM
10783 structures to try and expose faulty code further on. For now, openssl will
10784 (in its default mode of operation) continue to tolerate the inconsistent
10785 forms that it has tolerated in the past, but authors and packagers should
10786 consider trying openssl and their own applications when compiled with
10787 these debugging symbols defined. It will help highlight potential bugs in
10788 their own code, and will improve the test coverage for OpenSSL itself. At
10789 some point, these tighter rules will become openssl's default to improve
10790 maintainability, though the assert()s and other overheads will remain only
10791 in debugging configurations. See bn.h for more details.
10792
10793 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10794
10795 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10796 that can only be obtained through BN_CTX_new() (which implicitly
10797 initialises it). The presence of this function only made it possible
10798 to overwrite an existing structure (and cause memory leaks).
10799
10800 *Geoff Thorpe*
10801
10802 * Because of the callback-based approach for implementing LHASH as a
10803 template type, lh_insert() adds opaque objects to hash-tables and
10804 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10805 to clean up those corresponding objects before destroying the hash table
10806 (and losing the object pointers). So some over-zealous constifications in
10807 LHASH have been relaxed so that lh_insert() does not take (nor store) the
257e9d03 10808 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
5f8e6c50
DMSP
10809 prototyped to have "const" restrictions on the object pointers they are
10810 given (and so aren't required to cast them away any more).
10811
10812 *Geoff Thorpe*
10813
10814 * The tmdiff.h API was so ugly and minimal that our own timing utility
10815 (speed) prefers to use its own implementation. The two implementations
10816 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
257e9d03
RS
10817 its object type properly exposed (MS_TM) instead of casting to/from
10818 `char *`. This may still change yet if someone realises MS_TM and
10819 `ms_time_***`
5f8e6c50
DMSP
10820 aren't necessarily the greatest nomenclatures - but this is what was used
10821 internally to the implementation so I've used that for now.
10822
10823 *Geoff Thorpe*
10824
10825 * Ensure that deprecated functions do not get compiled when
10826 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10827 the self-tests were still using deprecated key-generation functions so
10828 these have been updated also.
10829
10830 *Geoff Thorpe*
10831
10832 * Reorganise PKCS#7 code to separate the digest location functionality
10833 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10834 New function PKCS7_set_digest() to set the digest type for PKCS#7
10835 digestedData type. Add additional code to correctly generate the
10836 digestedData type and add support for this type in PKCS7 initialization
10837 functions.
10838
10839 *Steve Henson*
10840
10841 * New function PKCS7_set0_type_other() this initializes a PKCS7
10842 structure of type "other".
10843
10844 *Steve Henson*
10845
10846 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10847 sure the loop does correctly stop and breaking ("division by zero")
10848 modulus operations are not performed. The (pre-generated) prime
10849 table crypto/bn/bn_prime.h was already correct, but it could not be
10850 re-generated on some platforms because of the "division by zero"
10851 situation in the script.
10852
10853 *Ralf S. Engelschall*
10854
10855 * Update support for ECC-based TLS ciphersuites according to
10856 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10857 SHA-1 now is only used for "small" curves (where the
10858 representation of a field element takes up to 24 bytes); for
10859 larger curves, the field element resulting from ECDH is directly
10860 used as premaster secret.
10861
10862 *Douglas Stebila (Sun Microsystems Laboratories)*
10863
10864 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10865 curve secp160r1 to the tests.
10866
10867 *Douglas Stebila (Sun Microsystems Laboratories)*
10868
10869 * Add the possibility to load symbols globally with DSO.
10870
10871 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10872
10873 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10874 control of the error stack.
10875
10876 *Richard Levitte*
10877
10878 * Add support for STORE in ENGINE.
10879
10880 *Richard Levitte*
10881
10882 * Add the STORE type. The intention is to provide a common interface
10883 to certificate and key stores, be they simple file-based stores, or
10884 HSM-type store, or LDAP stores, or...
10885 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10886
10887 *Richard Levitte*
10888
10889 * Add a generic structure called OPENSSL_ITEM. This can be used to
10890 pass a list of arguments to any function as well as provide a way
10891 for a function to pass data back to the caller.
10892
10893 *Richard Levitte*
10894
10895 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10896 works like BUF_strdup() but can be used to duplicate a portion of
10897 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10898 a memory area.
10899
10900 *Richard Levitte*
10901
10902 * Add the function sk_find_ex() which works like sk_find(), but will
10903 return an index to an element even if an exact match couldn't be
10904 found. The index is guaranteed to point at the element where the
10905 searched-for key would be inserted to preserve sorting order.
10906
10907 *Richard Levitte*
10908
10909 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10910 takes an extra flags argument for optional functionality. Currently,
10911 the following flags are defined:
10912
10913 OBJ_BSEARCH_VALUE_ON_NOMATCH
10914 This one gets OBJ_bsearch_ex() to return a pointer to the first
10915 element where the comparing function returns a negative or zero
10916 number.
10917
10918 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10919 This one gets OBJ_bsearch_ex() to return a pointer to the first
10920 element where the comparing function returns zero. This is useful
10921 if there are more than one element where the comparing function
10922 returns zero.
10923
10924 *Richard Levitte*
10925
10926 * Make it possible to create self-signed certificates with 'openssl ca'
10927 in such a way that the self-signed certificate becomes part of the
10928 CA database and uses the same mechanisms for serial number generation
10929 as all other certificate signing. The new flag '-selfsign' enables
10930 this functionality. Adapt CA.sh and CA.pl.in.
10931
10932 *Richard Levitte*
10933
10934 * Add functionality to check the public key of a certificate request
10935 against a given private. This is useful to check that a certificate
10936 request can be signed by that key (self-signing).
10937
10938 *Richard Levitte*
10939
10940 * Make it possible to have multiple active certificates with the same
10941 subject in the CA index file. This is done only if the keyword
10942 'unique_subject' is set to 'no' in the main CA section (default
10943 if 'CA_default') of the configuration file. The value is saved
10944 with the database itself in a separate index attribute file,
10945 named like the index file with '.attr' appended to the name.
10946
10947 *Richard Levitte*
10948
10949 * Generate multi-valued AVAs using '+' notation in config files for
10950 req and dirName.
10951
10952 *Steve Henson*
10953
10954 * Support for nameConstraints certificate extension.
10955
10956 *Steve Henson*
10957
10958 * Support for policyConstraints certificate extension.
10959
10960 *Steve Henson*
10961
10962 * Support for policyMappings certificate extension.
10963
10964 *Steve Henson*
10965
10966 * Make sure the default DSA_METHOD implementation only uses its
10967 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10968 and change its own handlers to be NULL so as to remove unnecessary
10969 indirection. This lets alternative implementations fallback to the
10970 default implementation more easily.
10971
10972 *Geoff Thorpe*
10973
10974 * Support for directoryName in GeneralName related extensions
10975 in config files.
10976
10977 *Steve Henson*
10978
10979 * Make it possible to link applications using Makefile.shared.
10980 Make that possible even when linking against static libraries!
10981
10982 *Richard Levitte*
10983
10984 * Support for single pass processing for S/MIME signing. This now
10985 means that S/MIME signing can be done from a pipe, in addition
10986 cleartext signing (multipart/signed type) is effectively streaming
10987 and the signed data does not need to be all held in memory.
10988
10989 This is done with a new flag PKCS7_STREAM. When this flag is set
10990 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10991 is done after the data is output (and digests calculated) in
10992 SMIME_write_PKCS7().
10993
10994 *Steve Henson*
10995
10996 * Add full support for -rpath/-R, both in shared libraries and
10997 applications, at least on the platforms where it's known how
10998 to do it.
10999
11000 *Richard Levitte*
11001
11002 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11003 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11004 will now compute a table of multiples of the generator that
11005 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11006 faster (notably in the case of a single point multiplication,
11007 scalar * generator).
11008
11009 *Nils Larsch, Bodo Moeller*
11010
11011 * IPv6 support for certificate extensions. The various extensions
11012 which use the IP:a.b.c.d can now take IPv6 addresses using the
11013 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11014 correctly.
11015
11016 *Steve Henson*
11017
11018 * Added an ENGINE that implements RSA by performing private key
11019 exponentiations with the GMP library. The conversions to and from
11020 GMP's mpz_t format aren't optimised nor are any montgomery forms
11021 cached, and on x86 it appears OpenSSL's own performance has caught up.
11022 However there are likely to be other architectures where GMP could
11023 provide a boost. This ENGINE is not built in by default, but it can be
11024 specified at Configure time and should be accompanied by the necessary
11025 linker additions, eg;
11026 ./config -DOPENSSL_USE_GMP -lgmp
11027
11028 *Geoff Thorpe*
11029
11030 * "openssl engine" will not display ENGINE/DSO load failure errors when
11031 testing availability of engines with "-t" - the old behaviour is
11032 produced by increasing the feature's verbosity with "-tt".
11033
11034 *Geoff Thorpe*
11035
11036 * ECDSA routines: under certain error conditions uninitialized BN objects
11037 could be freed. Solution: make sure initialization is performed early
11038 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11039 via PR#459)
11040
11041 *Lutz Jaenicke*
11042
11043 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11044 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11045 software implementations. For DSA and DH, parameter generation can
11046 also be overridden by providing the appropriate method callbacks.
11047
11048 *Geoff Thorpe*
11049
11050 * Change the "progress" mechanism used in key-generation and
11051 primality testing to functions that take a new BN_GENCB pointer in
257e9d03 11052 place of callback/argument pairs. The new API functions have `_ex`
5f8e6c50
DMSP
11053 postfixes and the older functions are reimplemented as wrappers for
11054 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11055 declarations of the old functions to help (graceful) attempts to
11056 migrate to the new functions. Also, the new key-generation API
11057 functions operate on a caller-supplied key-structure and return
11058 success/failure rather than returning a key or NULL - this is to
11059 help make "keygen" another member function of RSA_METHOD etc.
11060
11061 Example for using the new callback interface:
11062
11063 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11064 void *my_arg = ...;
11065 BN_GENCB my_cb;
11066
11067 BN_GENCB_set(&my_cb, my_callback, my_arg);
11068
11069 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11070 /* For the meaning of a, b in calls to my_callback(), see the
11071 * documentation of the function that calls the callback.
11072 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11073 * my_callback should return 1 if it wants BN_is_prime_ex()
11074 * to continue, or 0 to stop.
11075 */
11076
11077 *Geoff Thorpe*
11078
11079 * Change the ZLIB compression method to be stateful, and make it
11080 available to TLS with the number defined in
11081 draft-ietf-tls-compression-04.txt.
11082
11083 *Richard Levitte*
11084
11085 * Add the ASN.1 structures and functions for CertificatePair, which
11086 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11087
11088 CertificatePair ::= SEQUENCE {
11089 forward [0] Certificate OPTIONAL,
11090 reverse [1] Certificate OPTIONAL,
11091 -- at least one of the pair shall be present -- }
11092
11093 Also implement the PEM functions to read and write certificate
11094 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11095
11096 This needed to be defined, mostly for the sake of the LDAP
11097 attribute crossCertificatePair, but may prove useful elsewhere as
11098 well.
11099
11100 *Richard Levitte*
11101
11102 * Make it possible to inhibit symlinking of shared libraries in
11103 Makefile.shared, for Cygwin's sake.
11104
11105 *Richard Levitte*
11106
11107 * Extend the BIGNUM API by creating a function
11108 void BN_set_negative(BIGNUM *a, int neg);
11109 and a macro that behave like
11110 int BN_is_negative(const BIGNUM *a);
11111
11112 to avoid the need to access 'a->neg' directly in applications.
11113
11114 *Nils Larsch*
11115
11116 * Implement fast modular reduction for pseudo-Mersenne primes
11117 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11118 EC_GROUP_new_curve_GFp() will now automatically use this
11119 if applicable.
11120
11121 *Nils Larsch <nla@trustcenter.de>*
11122
11123 * Add new lock type (CRYPTO_LOCK_BN).
11124
11125 *Bodo Moeller*
11126
11127 * Change the ENGINE framework to automatically load engines
11128 dynamically from specific directories unless they could be
11129 found to already be built in or loaded. Move all the
11130 current engines except for the cryptodev one to a new
11131 directory engines/.
11132 The engines in engines/ are built as shared libraries if
11133 the "shared" options was given to ./Configure or ./config.
11134 Otherwise, they are inserted in libcrypto.a.
11135 /usr/local/ssl/engines is the default directory for dynamic
11136 engines, but that can be overridden at configure time through
11137 the usual use of --prefix and/or --openssldir, and at run
11138 time with the environment variable OPENSSL_ENGINES.
11139
11140 *Geoff Thorpe and Richard Levitte*
11141
11142 * Add Makefile.shared, a helper makefile to build shared
11143 libraries. Adapt Makefile.org.
11144
11145 *Richard Levitte*
11146
11147 * Add version info to Win32 DLLs.
11148
11149 *Peter 'Luna' Runestig" <peter@runestig.com>*
11150
11151 * Add new 'medium level' PKCS#12 API. Certificates and keys
11152 can be added using this API to created arbitrary PKCS#12
4d49b685 11153 files while avoiding the low-level API.
5f8e6c50
DMSP
11154
11155 New options to PKCS12_create(), key or cert can be NULL and
11156 will then be omitted from the output file. The encryption
11157 algorithm NIDs can be set to -1 for no encryption, the mac
11158 iteration count can be set to 0 to omit the mac.
11159
11160 Enhance pkcs12 utility by making the -nokeys and -nocerts
11161 options work when creating a PKCS#12 file. New option -nomac
11162 to omit the mac, NONE can be set for an encryption algorithm.
11163 New code is modified to use the enhanced PKCS12_create()
4d49b685 11164 instead of the low-level API.
5f8e6c50
DMSP
11165
11166 *Steve Henson*
11167
11168 * Extend ASN1 encoder to support indefinite length constructed
11169 encoding. This can output sequences tags and octet strings in
11170 this form. Modify pk7_asn1.c to support indefinite length
11171 encoding. This is experimental and needs additional code to
11172 be useful, such as an ASN1 bio and some enhanced streaming
11173 PKCS#7 code.
11174
11175 Extend template encode functionality so that tagging is passed
11176 down to the template encoder.
11177
11178 *Steve Henson*
11179
11180 * Let 'openssl req' fail if an argument to '-newkey' is not
11181 recognized instead of using RSA as a default.
11182
11183 *Bodo Moeller*
11184
11185 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11186 As these are not official, they are not included in "ALL";
11187 the "ECCdraft" ciphersuite group alias can be used to select them.
11188
11189 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11190
11191 * Add ECDH engine support.
11192
11193 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11194
11195 * Add ECDH in new directory crypto/ecdh/.
11196
11197 *Douglas Stebila (Sun Microsystems Laboratories)*
11198
11199 * Let BN_rand_range() abort with an error after 100 iterations
11200 without success (which indicates a broken PRNG).
11201
11202 *Bodo Moeller*
11203
11204 * Change BN_mod_sqrt() so that it verifies that the input value
11205 is really the square of the return value. (Previously,
11206 BN_mod_sqrt would show GIGO behaviour.)
11207
11208 *Bodo Moeller*
11209
11210 * Add named elliptic curves over binary fields from X9.62, SECG,
11211 and WAP/WTLS; add OIDs that were still missing.
11212
257e9d03 11213 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
5f8e6c50
DMSP
11214
11215 * Extend the EC library for elliptic curves over binary fields
11216 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11217 New EC_METHOD:
11218
11219 EC_GF2m_simple_method
11220
11221 New API functions:
11222
11223 EC_GROUP_new_curve_GF2m
11224 EC_GROUP_set_curve_GF2m
11225 EC_GROUP_get_curve_GF2m
11226 EC_POINT_set_affine_coordinates_GF2m
11227 EC_POINT_get_affine_coordinates_GF2m
11228 EC_POINT_set_compressed_coordinates_GF2m
11229
11230 Point compression for binary fields is disabled by default for
11231 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11232 enable it).
11233
11234 As binary polynomials are represented as BIGNUMs, various members
11235 of the EC_GROUP and EC_POINT data structures can be shared
11236 between the implementations for prime fields and binary fields;
257e9d03
RS
11237 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11238 are essentially identical to their `..._GFp` counterparts.
11239 (For simplicity, the `..._GFp` prefix has been dropped from
5f8e6c50
DMSP
11240 various internal method names.)
11241
11242 An internal 'field_div' method (similar to 'field_mul' and
11243 'field_sqr') has been added; this is used only for binary fields.
11244
257e9d03 11245 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
5f8e6c50
DMSP
11246
11247 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11248 through methods ('mul', 'precompute_mult').
11249
11250 The generic implementations (now internally called 'ec_wNAF_mul'
11251 and 'ec_wNAF_precomputed_mult') remain the default if these
11252 methods are undefined.
11253
257e9d03 11254 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
5f8e6c50
DMSP
11255
11256 * New function EC_GROUP_get_degree, which is defined through
11257 EC_METHOD. For curves over prime fields, this returns the bit
11258 length of the modulus.
11259
257e9d03 11260 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
5f8e6c50
DMSP
11261
11262 * New functions EC_GROUP_dup, EC_POINT_dup.
11263 (These simply call ..._new and ..._copy).
11264
257e9d03 11265 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
5f8e6c50
DMSP
11266
11267 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11268 Polynomials are represented as BIGNUMs (where the sign bit is not
11269 used) in the following functions [macros]:
11270
11271 BN_GF2m_add
11272 BN_GF2m_sub [= BN_GF2m_add]
11273 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11274 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11275 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11276 BN_GF2m_mod_inv
11277 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11278 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11279 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11280 BN_GF2m_cmp [= BN_ucmp]
11281
11282 (Note that only the 'mod' functions are actually for fields GF(2^m).
11283 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11284
11285 For some functions, an the irreducible polynomial defining a
11286 field can be given as an 'unsigned int[]' with strictly
11287 decreasing elements giving the indices of those bits that are set;
11288 i.e., p[] represents the polynomial
11289 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11290 where
11291 p[0] > p[1] > ... > p[k] = 0.
11292 This applies to the following functions:
11293
11294 BN_GF2m_mod_arr
11295 BN_GF2m_mod_mul_arr
11296 BN_GF2m_mod_sqr_arr
11297 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11298 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11299 BN_GF2m_mod_exp_arr
11300 BN_GF2m_mod_sqrt_arr
11301 BN_GF2m_mod_solve_quad_arr
11302 BN_GF2m_poly2arr
11303 BN_GF2m_arr2poly
11304
11305 Conversion can be performed by the following functions:
11306
11307 BN_GF2m_poly2arr
11308 BN_GF2m_arr2poly
11309
11310 bntest.c has additional tests for binary polynomial arithmetic.
11311
11312 Two implementations for BN_GF2m_mod_div() are available.
11313 The default algorithm simply uses BN_GF2m_mod_inv() and
11314 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11315 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11316 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11317
257e9d03 11318 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
5f8e6c50
DMSP
11319
11320 * Add new error code 'ERR_R_DISABLED' that can be used when some
11321 functionality is disabled at compile-time.
11322
11323 *Douglas Stebila <douglas.stebila@sun.com>*
11324
11325 * Change default behaviour of 'openssl asn1parse' so that more
11326 information is visible when viewing, e.g., a certificate:
11327
11328 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11329 mode the content of non-printable OCTET STRINGs is output in a
11330 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11331 avoid the appearance of a printable string.
11332
11333 *Nils Larsch <nla@trustcenter.de>*
11334
11335 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11336 functions
11337 EC_GROUP_set_asn1_flag()
11338 EC_GROUP_get_asn1_flag()
11339 EC_GROUP_set_point_conversion_form()
11340 EC_GROUP_get_point_conversion_form()
11341 These control ASN1 encoding details:
11342 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11343 has been set to OPENSSL_EC_NAMED_CURVE.
11344 - Points are encoded in uncompressed form by default; options for
11345 asn1_for are as for point2oct, namely
11346 POINT_CONVERSION_COMPRESSED
11347 POINT_CONVERSION_UNCOMPRESSED
11348 POINT_CONVERSION_HYBRID
11349
11350 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11351 functions
11352 EC_GROUP_set_seed()
11353 EC_GROUP_get0_seed()
11354 EC_GROUP_get_seed_len()
11355 This is used only for ASN1 purposes (so far).
11356
11357 *Nils Larsch <nla@trustcenter.de>*
11358
11359 * Add 'field_type' member to EC_METHOD, which holds the NID
11360 of the appropriate field type OID. The new function
11361 EC_METHOD_get_field_type() returns this value.
11362
11363 *Nils Larsch <nla@trustcenter.de>*
11364
11365 * Add functions
11366 EC_POINT_point2bn()
11367 EC_POINT_bn2point()
11368 EC_POINT_point2hex()
11369 EC_POINT_hex2point()
11370 providing useful interfaces to EC_POINT_point2oct() and
11371 EC_POINT_oct2point().
11372
11373 *Nils Larsch <nla@trustcenter.de>*
11374
11375 * Change internals of the EC library so that the functions
11376 EC_GROUP_set_generator()
11377 EC_GROUP_get_generator()
11378 EC_GROUP_get_order()
11379 EC_GROUP_get_cofactor()
11380 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11381 to methods, which would lead to unnecessary code duplication when
11382 adding different types of curves.
11383
11384 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11385
11386 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11387 arithmetic, and such that modified wNAFs are generated
11388 (which avoid length expansion in many cases).
11389
11390 *Bodo Moeller*
11391
11392 * Add a function EC_GROUP_check_discriminant() (defined via
11393 EC_METHOD) that verifies that the curve discriminant is non-zero.
11394
11395 Add a function EC_GROUP_check() that makes some sanity tests
11396 on a EC_GROUP, its generator and order. This includes
11397 EC_GROUP_check_discriminant().
11398
11399 *Nils Larsch <nla@trustcenter.de>*
11400
11401 * Add ECDSA in new directory crypto/ecdsa/.
11402
11403 Add applications 'openssl ecparam' and 'openssl ecdsa'
11404 (these are based on 'openssl dsaparam' and 'openssl dsa').
11405
11406 ECDSA support is also included in various other files across the
11407 library. Most notably,
11408 - 'openssl req' now has a '-newkey ecdsa:file' option;
11409 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11410 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11411 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11412 them suitable for ECDSA where domain parameters must be
11413 extracted before the specific public key;
11414 - ECDSA engine support has been added.
11415
11416 *Nils Larsch <nla@trustcenter.de>*
11417
11418 * Include some named elliptic curves, and add OIDs from X9.62,
11419 SECG, and WAP/WTLS. Each curve can be obtained from the new
11420 function
11421 EC_GROUP_new_by_curve_name(),
11422 and the list of available named curves can be obtained with
11423 EC_get_builtin_curves().
11424 Also add a 'curve_name' member to EC_GROUP objects, which can be
11425 accessed via
11426 EC_GROUP_set_curve_name()
11427 EC_GROUP_get_curve_name()
11428
11429 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11430
11431 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11432 was actually never needed) and in BN_mul(). The removal in BN_mul()
11433 required a small change in bn_mul_part_recursive() and the addition
11434 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11435 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11436 bn_sub_words() and bn_add_words() except they take arrays with
11437 differing sizes.
11438
11439 *Richard Levitte*
11440
257e9d03 11441### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
5f8e6c50
DMSP
11442
11443 * Cleanse PEM buffers before freeing them since they may contain
11444 sensitive data.
11445
11446 *Benjamin Bennett <ben@psc.edu>*
11447
11448 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11449 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11450 authentication-only ciphersuites.
11451
11452 *Bodo Moeller*
11453
11454 * Since AES128 and AES256 share a single mask bit in the logic of
11455 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11456 kludge to work properly if AES128 is available and AES256 isn't.
11457
11458 *Victor Duchovni*
11459
11460 * Expand security boundary to match 1.1.1 module.
11461
11462 *Steve Henson*
11463
11464 * Remove redundant features: hash file source, editing of test vectors
11465 modify fipsld to use external fips_premain.c signature.
11466
11467 *Steve Henson*
11468
11469 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11470 run algorithm test programs.
11471
11472 *Steve Henson*
11473
11474 * Make algorithm test programs more tolerant of whitespace.
11475
11476 *Steve Henson*
11477
11478 * Have SSL/TLS server implementation tolerate "mismatched" record
11479 protocol version while receiving ClientHello even if the
11480 ClientHello is fragmented. (The server can't insist on the
11481 particular protocol version it has chosen before the ServerHello
11482 message has informed the client about his choice.)
11483
11484 *Bodo Moeller*
11485
11486 * Load error codes if they are not already present instead of using a
11487 static variable. This allows them to be cleanly unloaded and reloaded.
11488
11489 *Steve Henson*
11490
257e9d03 11491### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
5f8e6c50
DMSP
11492
11493 * Introduce limits to prevent malicious keys being able to
d8dc8538 11494 cause a denial of service. ([CVE-2006-2940])
5f8e6c50
DMSP
11495
11496 *Steve Henson, Bodo Moeller*
11497
11498 * Fix ASN.1 parsing of certain invalid structures that can result
d8dc8538 11499 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
5f8e6c50
DMSP
11500
11501 * Fix buffer overflow in SSL_get_shared_ciphers() function.
d8dc8538 11502 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
5f8e6c50
DMSP
11503
11504 * Fix SSL client code which could crash if connecting to a
d8dc8538 11505 malicious SSLv2 server. ([CVE-2006-4343])
5f8e6c50
DMSP
11506
11507 *Tavis Ormandy and Will Drewry, Google Security Team*
11508
11509 * Change ciphersuite string processing so that an explicit
11510 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11511 will no longer include "AES128-SHA"), and any other similar
11512 ciphersuite (same bitmap) from *other* protocol versions (so that
11513 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11514 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11515 changes from 0.9.8b and 0.9.8d.
11516
11517 *Bodo Moeller*
11518
257e9d03 11519### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
5f8e6c50
DMSP
11520
11521 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
d8dc8538 11522 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
5f8e6c50
DMSP
11523
11524 * Change the Unix randomness entropy gathering to use poll() when
11525 possible instead of select(), since the latter has some
11526 undesirable limitations.
11527
11528 *Darryl Miles via Richard Levitte and Bodo Moeller*
11529
11530 * Disable rogue ciphersuites:
11531
257e9d03
RS
11532 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11533 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11534 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5f8e6c50
DMSP
11535
11536 The latter two were purportedly from
11537 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11538 appear there.
11539
11540 Also deactivate the remaining ciphersuites from
11541 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11542 unofficial, and the ID has long expired.
11543
11544 *Bodo Moeller*
11545
11546 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11547 dual-core machines) and other potential thread-safety issues.
11548
11549 *Bodo Moeller*
11550
257e9d03 11551### Changes between 0.9.7i and 0.9.7j [04 May 2006]
5f8e6c50
DMSP
11552
11553 * Adapt fipsld and the build system to link against the validated FIPS
11554 module in FIPS mode.
11555
11556 *Steve Henson*
11557
11558 * Fixes for VC++ 2005 build under Windows.
11559
11560 *Steve Henson*
11561
11562 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11563 from a Windows bash shell such as MSYS. It is autodetected from the
11564 "config" script when run from a VC++ environment. Modify standard VC++
11565 build to use fipscanister.o from the GNU make build.
11566
11567 *Steve Henson*
11568
257e9d03 11569### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
5f8e6c50
DMSP
11570
11571 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11572 The value now differs depending on if you build for FIPS or not.
11573 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11574 safely run with a non-FIPSed libcrypto, as it may crash because of
11575 the difference induced by this change.
11576
11577 *Andy Polyakov*
11578
257e9d03 11579### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
5f8e6c50
DMSP
11580
11581 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11582 (part of SSL_OP_ALL). This option used to disable the
11583 countermeasure against man-in-the-middle protocol-version
11584 rollback in the SSL 2.0 server implementation, which is a bad
d8dc8538 11585 idea. ([CVE-2005-2969])
5f8e6c50
DMSP
11586
11587 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11588 for Information Security, National Institute of Advanced Industrial
257e9d03 11589 Science and Technology [AIST, Japan)]*
5f8e6c50
DMSP
11590
11591 * Minimal support for X9.31 signatures and PSS padding modes. This is
11592 mainly for FIPS compliance and not fully integrated at this stage.
11593
11594 *Steve Henson*
11595
11596 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11597 the exponentiation using a fixed-length exponent. (Otherwise,
11598 the information leaked through timing could expose the secret key
11599 after many signatures; cf. Bleichenbacher's attack on DSA with
11600 biased k.)
11601
11602 *Bodo Moeller*
11603
11604 * Make a new fixed-window mod_exp implementation the default for
11605 RSA, DSA, and DH private-key operations so that the sequence of
11606 squares and multiplies and the memory access pattern are
11607 independent of the particular secret key. This will mitigate
11608 cache-timing and potential related attacks.
11609
11610 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11611 and this is automatically used by BN_mod_exp_mont() if the new flag
11612 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11613 will use this BN flag for private exponents unless the flag
11614 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11615 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11616
5f8e6c50
DMSP
11617 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11618
11619 * Change the client implementation for SSLv23_method() and
11620 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11621 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11622 (Previously, the SSL 2.0 backwards compatible Client Hello
11623 message format would be used even with SSL_OP_NO_SSLv2.)
11624
11625 *Bodo Moeller*
11626
11627 * Add support for smime-type MIME parameter in S/MIME messages which some
11628 clients need.
11629
11630 *Steve Henson*
11631
11632 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11633 a threadsafe manner. Modify rsa code to use new function and add calls
11634 to dsa and dh code (which had race conditions before).
11635
11636 *Steve Henson*
11637
11638 * Include the fixed error library code in the C error file definitions
11639 instead of fixing them up at runtime. This keeps the error code
11640 structures constant.
11641
11642 *Steve Henson*
11643
257e9d03 11644### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
5f8e6c50
DMSP
11645
11646[NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11647OpenSSL 0.9.8.]
11648
11649 * Fixes for newer kerberos headers. NB: the casts are needed because
11650 the 'length' field is signed on one version and unsigned on another
11651 with no (?) obvious way to tell the difference, without these VC++
11652 complains. Also the "definition" of FAR (blank) is no longer included
11653 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11654 some needed definitions.
11655
11656 *Steve Henson*
11657
11658 * Undo Cygwin change.
11659
11660 *Ulf Möller*
11661
11662 * Added support for proxy certificates according to RFC 3820.
11663 Because they may be a security thread to unaware applications,
11664 they must be explicitly allowed in run-time. See
11665 docs/HOWTO/proxy_certificates.txt for further information.
11666
11667 *Richard Levitte*
11668
257e9d03 11669### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
5f8e6c50
DMSP
11670
11671 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11672 server and client random values. Previously
11673 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11674 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11675
11676 This change has negligible security impact because:
11677
11678 1. Server and client random values still have 24 bytes of pseudo random
11679 data.
11680
11681 2. Server and client random values are sent in the clear in the initial
11682 handshake.
11683
11684 3. The master secret is derived using the premaster secret (48 bytes in
11685 size for static RSA ciphersuites) as well as client server and random
11686 values.
11687
11688 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11689 to our attention.
11690
11691 *Stephen Henson, reported by UK NISCC*
11692
11693 * Use Windows randomness collection on Cygwin.
11694
11695 *Ulf Möller*
11696
11697 * Fix hang in EGD/PRNGD query when communication socket is closed
11698 prematurely by EGD/PRNGD.
11699
11700 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11701
11702 * Prompt for pass phrases when appropriate for PKCS12 input format.
11703
11704 *Steve Henson*
11705
11706 * Back-port of selected performance improvements from development
11707 branch, as well as improved support for PowerPC platforms.
11708
11709 *Andy Polyakov*
11710
11711 * Add lots of checks for memory allocation failure, error codes to indicate
11712 failure and freeing up memory if a failure occurs.
11713
11714 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11715
11716 * Add new -passin argument to dgst.
11717
11718 *Steve Henson*
11719
11720 * Perform some character comparisons of different types in X509_NAME_cmp:
11721 this is needed for some certificates that re-encode DNs into UTF8Strings
11722 (in violation of RFC3280) and can't or won't issue name rollover
11723 certificates.
11724
11725 *Steve Henson*
11726
11727 * Make an explicit check during certificate validation to see that
11728 the CA setting in each certificate on the chain is correct. As a
11729 side effect always do the following basic checks on extensions,
11730 not just when there's an associated purpose to the check:
11731
257e9d03
RS
11732 - if there is an unhandled critical extension (unless the user
11733 has chosen to ignore this fault)
11734 - if the path length has been exceeded (if one is set at all)
11735 - that certain extensions fit the associated purpose (if one has
11736 been given)
5f8e6c50
DMSP
11737
11738 *Richard Levitte*
11739
257e9d03 11740### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
5f8e6c50
DMSP
11741
11742 * Avoid a race condition when CRLs are checked in a multi threaded
11743 environment. This would happen due to the reordering of the revoked
11744 entries during signature checking and serial number lookup. Now the
11745 encoding is cached and the serial number sort performed under a lock.
11746 Add new STACK function sk_is_sorted().
11747
11748 *Steve Henson*
11749
11750 * Add Delta CRL to the extension code.
11751
11752 *Steve Henson*
11753
11754 * Various fixes to s3_pkt.c so alerts are sent properly.
11755
11756 *David Holmes <d.holmes@f5.com>*
11757
11758 * Reduce the chances of duplicate issuer name and serial numbers (in
11759 violation of RFC3280) using the OpenSSL certificate creation utilities.
11760 This is done by creating a random 64 bit value for the initial serial
11761 number when a serial number file is created or when a self signed
11762 certificate is created using 'openssl req -x509'. The initial serial
11763 number file is created using 'openssl x509 -next_serial' in CA.pl
11764 rather than being initialized to 1.
11765
11766 *Steve Henson*
11767
257e9d03 11768### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
5f8e6c50
DMSP
11769
11770 * Fix null-pointer assignment in do_change_cipher_spec() revealed
d8dc8538 11771 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
5f8e6c50
DMSP
11772
11773 *Joe Orton, Steve Henson*
11774
11775 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
d8dc8538 11776 ([CVE-2004-0112])
5f8e6c50
DMSP
11777
11778 *Joe Orton, Steve Henson*
11779
11780 * Make it possible to have multiple active certificates with the same
11781 subject in the CA index file. This is done only if the keyword
11782 'unique_subject' is set to 'no' in the main CA section (default
11783 if 'CA_default') of the configuration file. The value is saved
11784 with the database itself in a separate index attribute file,
11785 named like the index file with '.attr' appended to the name.
11786
11787 *Richard Levitte*
11788
11789 * X509 verify fixes. Disable broken certificate workarounds when
11790 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11791 keyUsage extension present. Don't accept CRLs with unhandled critical
11792 extensions: since verify currently doesn't process CRL extensions this
11793 rejects a CRL with *any* critical extensions. Add new verify error codes
11794 for these cases.
11795
11796 *Steve Henson*
11797
11798 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11799 A clarification of RFC2560 will require the use of OCTET STRINGs and
11800 some implementations cannot handle the current raw format. Since OpenSSL
11801 copies and compares OCSP nonces as opaque blobs without any attempt at
11802 parsing them this should not create any compatibility issues.
11803
11804 *Steve Henson*
11805
11806 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11807 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11808 this HMAC (and other) operations are several times slower than OpenSSL
11809 < 0.9.7.
11810
11811 *Steve Henson*
11812
11813 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11814
11815 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11816
11817 * Use the correct content when signing type "other".
11818
11819 *Steve Henson*
11820
257e9d03 11821### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5f8e6c50
DMSP
11822
11823 * Fix various bugs revealed by running the NISCC test suite:
11824
11825 Stop out of bounds reads in the ASN1 code when presented with
11826 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11827
d8dc8538 11828 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
5f8e6c50
DMSP
11829
11830 If verify callback ignores invalid public key errors don't try to check
11831 certificate signature with the NULL public key.
11832
5f8e6c50
DMSP
11833 *Steve Henson*
11834
11835 * New -ignore_err option in ocsp application to stop the server
11836 exiting on the first error in a request.
11837
11838 *Steve Henson*
11839
11840 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11841 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11842 specifications.
11843
11844 *Steve Henson*
11845
11846 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11847 extra data after the compression methods not only for TLS 1.0
11848 but also for SSL 3.0 (as required by the specification).
11849
11850 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11851
11852 * Change X509_certificate_type() to mark the key as exported/exportable
11853 when it's 512 *bits* long, not 512 bytes.
11854
11855 *Richard Levitte*
11856
11857 * Change AES_cbc_encrypt() so it outputs exact multiple of
11858 blocks during encryption.
11859
11860 *Richard Levitte*
11861
11862 * Various fixes to base64 BIO and non blocking I/O. On write
11863 flushes were not handled properly if the BIO retried. On read
11864 data was not being buffered properly and had various logic bugs.
11865 This also affects blocking I/O when the data being decoded is a
11866 certain size.
11867
11868 *Steve Henson*
11869
11870 * Various S/MIME bugfixes and compatibility changes:
11871 output correct application/pkcs7 MIME type if
11872 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11873 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11874 of files as .eml work). Correctly handle very long lines in MIME
11875 parser.
11876
11877 *Steve Henson*
11878
257e9d03 11879### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5f8e6c50
DMSP
11880
11881 * Countermeasure against the Klima-Pokorny-Rosa extension of
11882 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11883 a protocol version number mismatch like a decryption error
11884 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11885
11886 *Bodo Moeller*
11887
11888 * Turn on RSA blinding by default in the default implementation
11889 to avoid a timing attack. Applications that don't want it can call
11890 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11891 They would be ill-advised to do so in most cases.
11892
11893 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11894
11895 * Change RSA blinding code so that it works when the PRNG is not
11896 seeded (in this case, the secret RSA exponent is abused as
11897 an unpredictable seed -- if it is not unpredictable, there
11898 is no point in blinding anyway). Make RSA blinding thread-safe
11899 by remembering the creator's thread ID in rsa->blinding and
11900 having all other threads use local one-time blinding factors
11901 (this requires more computation than sharing rsa->blinding, but
11902 avoids excessive locking; and if an RSA object is not shared
11903 between threads, blinding will still be very fast).
11904
11905 *Bodo Moeller*
11906
11907 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11908 ENGINE as defaults for all supported algorithms irrespective of
11909 the 'flags' parameter. 'flags' is now honoured, so applications
11910 should make sure they are passing it correctly.
11911
11912 *Geoff Thorpe*
11913
11914 * Target "mingw" now allows native Windows code to be generated in
11915 the Cygwin environment as well as with the MinGW compiler.
11916
11917 *Ulf Moeller*
11918
257e9d03 11919### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5f8e6c50
DMSP
11920
11921 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11922 via timing by performing a MAC computation even if incorrect
11923 block cipher padding has been found. This is a countermeasure
11924 against active attacks where the attacker has to distinguish
d8dc8538 11925 between bad padding and a MAC verification error. ([CVE-2003-0078])
5f8e6c50
DMSP
11926
11927 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11928 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11929 Martin Vuagnoux (EPFL, Ilion)*
11930
11931 * Make the no-err option work as intended. The intention with no-err
11932 is not to have the whole error stack handling routines removed from
11933 libcrypto, it's only intended to remove all the function name and
11934 reason texts, thereby removing some of the footprint that may not
11935 be interesting if those errors aren't displayed anyway.
11936
11937 NOTE: it's still possible for any application or module to have its
11938 own set of error texts inserted. The routines are there, just not
11939 used by default when no-err is given.
11940
11941 *Richard Levitte*
11942
11943 * Add support for FreeBSD on IA64.
11944
11945 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11946
11947 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11948 Kerberos function mit_des_cbc_cksum(). Before this change,
11949 the value returned by DES_cbc_cksum() was like the one from
11950 mit_des_cbc_cksum(), except the bytes were swapped.
11951
11952 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11953
11954 * Allow an application to disable the automatic SSL chain building.
11955 Before this a rather primitive chain build was always performed in
11956 ssl3_output_cert_chain(): an application had no way to send the
11957 correct chain if the automatic operation produced an incorrect result.
11958
11959 Now the chain builder is disabled if either:
11960
11961 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11962
11963 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11964
11965 The reasoning behind this is that an application would not want the
11966 auto chain building to take place if extra chain certificates are
11967 present and it might also want a means of sending no additional
11968 certificates (for example the chain has two certificates and the
11969 root is omitted).
11970
11971 *Steve Henson*
11972
11973 * Add the possibility to build without the ENGINE framework.
11974
11975 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11976
11977 * Under Win32 gmtime() can return NULL: check return value in
11978 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11979
11980 *Steve Henson*
11981
11982 * DSA routines: under certain error conditions uninitialized BN objects
11983 could be freed. Solution: make sure initialization is performed early
11984 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11985 Nils Larsch <nla@trustcenter.de> via PR#459)
11986
11987 *Lutz Jaenicke*
11988
11989 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11990 checked on reconnect on the client side, therefore session resumption
11991 could still fail with a "ssl session id is different" error. This
11992 behaviour is masked when SSL_OP_ALL is used due to
11993 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11994 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11995 followup to PR #377.
11996
11997 *Lutz Jaenicke*
11998
11999 * IA-32 assembler support enhancements: unified ELF targets, support
12000 for SCO/Caldera platforms, fix for Cygwin shared build.
12001
12002 *Andy Polyakov*
12003
12004 * Add support for FreeBSD on sparc64. As a consequence, support for
12005 FreeBSD on non-x86 processors is separate from x86 processors on
12006 the config script, much like the NetBSD support.
12007
12008 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12009
257e9d03 12010### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5f8e6c50
DMSP
12011
12012[NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12013OpenSSL 0.9.7.]
12014
12015 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12016 code (06) was taken as the first octet of the session ID and the last
12017 octet was ignored consequently. As a result SSLv2 client side session
12018 caching could not have worked due to the session ID mismatch between
12019 client and server.
12020 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12021 PR #377.
12022
12023 *Lutz Jaenicke*
12024
12025 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12026 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12027 removed entirely.
12028
12029 *Richard Levitte*
12030
12031 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12032 seems that in spite of existing for more than a year, many application
12033 author have done nothing to provide the necessary callbacks, which
12034 means that this particular engine will not work properly anywhere.
12035 This is a very unfortunate situation which forces us, in the name
12036 of usability, to give the hw_ncipher.c a static lock, which is part
12037 of libcrypto.
12038 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12039 appear in 0.9.8 or later. We EXPECT application authors to have
12040 dealt properly with this when 0.9.8 is released (unless we actually
12041 make such changes in the libcrypto locking code that changes will
12042 have to be made anyway).
12043
12044 *Richard Levitte*
12045
12046 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12047 octets have been read, EOF or an error occurs. Without this change
12048 some truncated ASN1 structures will not produce an error.
12049
12050 *Steve Henson*
12051
12052 * Disable Heimdal support, since it hasn't been fully implemented.
12053 Still give the possibility to force the use of Heimdal, but with
12054 warnings and a request that patches get sent to openssl-dev.
12055
12056 *Richard Levitte*
12057
12058 * Add the VC-CE target, introduce the WINCE sysname, and add
12059 INSTALL.WCE and appropriate conditionals to make it build.
12060
12061 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12062
12063 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12064 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12065 edit numbers of the version.
12066
12067 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12068
12069 * Introduce safe string copy and catenation functions
12070 (BUF_strlcpy() and BUF_strlcat()).
12071
12072 *Ben Laurie (CHATS) and Richard Levitte*
12073
12074 * Avoid using fixed-size buffers for one-line DNs.
12075
12076 *Ben Laurie (CHATS)*
12077
12078 * Add BUF_MEM_grow_clean() to avoid information leakage when
12079 resizing buffers containing secrets, and use where appropriate.
12080
12081 *Ben Laurie (CHATS)*
12082
12083 * Avoid using fixed size buffers for configuration file location.
12084
12085 *Ben Laurie (CHATS)*
12086
12087 * Avoid filename truncation for various CA files.
12088
12089 *Ben Laurie (CHATS)*
12090
12091 * Use sizeof in preference to magic numbers.
12092
12093 *Ben Laurie (CHATS)*
12094
12095 * Avoid filename truncation in cert requests.
12096
12097 *Ben Laurie (CHATS)*
12098
12099 * Add assertions to check for (supposedly impossible) buffer
12100 overflows.
12101
12102 *Ben Laurie (CHATS)*
12103
12104 * Don't cache truncated DNS entries in the local cache (this could
12105 potentially lead to a spoofing attack).
12106
12107 *Ben Laurie (CHATS)*
12108
12109 * Fix various buffers to be large enough for hex/decimal
12110 representations in a platform independent manner.
12111
12112 *Ben Laurie (CHATS)*
12113
12114 * Add CRYPTO_realloc_clean() to avoid information leakage when
12115 resizing buffers containing secrets, and use where appropriate.
12116
12117 *Ben Laurie (CHATS)*
12118
12119 * Add BIO_indent() to avoid much slightly worrying code to do
12120 indents.
12121
12122 *Ben Laurie (CHATS)*
12123
12124 * Convert sprintf()/BIO_puts() to BIO_printf().
12125
12126 *Ben Laurie (CHATS)*
12127
12128 * buffer_gets() could terminate with the buffer only half
12129 full. Fixed.
12130
12131 *Ben Laurie (CHATS)*
12132
12133 * Add assertions to prevent user-supplied crypto functions from
12134 overflowing internal buffers by having large block sizes, etc.
12135
12136 *Ben Laurie (CHATS)*
12137
12138 * New OPENSSL_assert() macro (similar to assert(), but enabled
12139 unconditionally).
12140
12141 *Ben Laurie (CHATS)*
12142
12143 * Eliminate unused copy of key in RC4.
12144
12145 *Ben Laurie (CHATS)*
12146
12147 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12148
12149 *Ben Laurie (CHATS)*
12150
12151 * Fix off-by-one error in EGD path.
12152
12153 *Ben Laurie (CHATS)*
12154
12155 * If RANDFILE path is too long, ignore instead of truncating.
12156
12157 *Ben Laurie (CHATS)*
12158
12159 * Eliminate unused and incorrectly sized X.509 structure
12160 CBCParameter.
12161
12162 *Ben Laurie (CHATS)*
12163
12164 * Eliminate unused and dangerous function knumber().
12165
12166 *Ben Laurie (CHATS)*
12167
12168 * Eliminate unused and dangerous structure, KSSL_ERR.
12169
12170 *Ben Laurie (CHATS)*
12171
12172 * Protect against overlong session ID context length in an encoded
12173 session object. Since these are local, this does not appear to be
12174 exploitable.
12175
12176 *Ben Laurie (CHATS)*
12177
12178 * Change from security patch (see 0.9.6e below) that did not affect
12179 the 0.9.6 release series:
12180
12181 Remote buffer overflow in SSL3 protocol - an attacker could
12182 supply an oversized master key in Kerberos-enabled versions.
d8dc8538 12183 ([CVE-2002-0657])
5f8e6c50
DMSP
12184
12185 *Ben Laurie (CHATS)*
12186
12187 * Change the SSL kerb5 codes to match RFC 2712.
12188
12189 *Richard Levitte*
12190
12191 * Make -nameopt work fully for req and add -reqopt switch.
12192
12193 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12194
12195 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12196
12197 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12198
12199 * Make sure tests can be performed even if the corresponding algorithms
12200 have been removed entirely. This was also the last step to make
12201 OpenSSL compilable with DJGPP under all reasonable conditions.
12202
12203 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12204
12205 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12206 to allow version independent disabling of normally unselected ciphers,
12207 which may be activated as a side-effect of selecting a single cipher.
12208
12209 (E.g., cipher list string "RSA" enables ciphersuites that are left
12210 out of "ALL" because they do not provide symmetric encryption.
12211 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12212
12213 *Lutz Jaenicke, Bodo Moeller*
12214
12215 * Add appropriate support for separate platform-dependent build
12216 directories. The recommended way to make a platform-dependent
12217 build directory is the following (tested on Linux), maybe with
12218 some local tweaks:
12219
12220 # Place yourself outside of the OpenSSL source tree. In
12221 # this example, the environment variable OPENSSL_SOURCE
12222 # is assumed to contain the absolute OpenSSL source directory.
12223 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12224 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12225 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12226 mkdir -p `dirname $F`
12227 ln -s $OPENSSL_SOURCE/$F $F
12228 done
12229
12230 To be absolutely sure not to disturb the source tree, a "make clean"
12231 is a good thing. If it isn't successful, don't worry about it,
12232 it probably means the source directory is very clean.
12233
12234 *Richard Levitte*
12235
12236 * Make sure any ENGINE control commands make local copies of string
12237 pointers passed to them whenever necessary. Otherwise it is possible
12238 the caller may have overwritten (or deallocated) the original string
12239 data when a later ENGINE operation tries to use the stored values.
12240
12241 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12242
12243 * Improve diagnostics in file reading and command-line digests.
12244
12245 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12246
12247 * Add AES modes CFB and OFB to the object database. Correct an
12248 error in AES-CFB decryption.
12249
12250 *Richard Levitte*
12251
12252 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12253 allows existing EVP_CIPHER_CTX structures to be reused after
257e9d03 12254 calling `EVP_*Final()`. This behaviour is used by encryption
5f8e6c50
DMSP
12255 BIOs and some applications. This has the side effect that
12256 applications must explicitly clean up cipher contexts with
12257 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12258
12259 *Steve Henson*
12260
12261 * Check the values of dna and dnb in bn_mul_recursive before calling
12262 bn_mul_comba (a non zero value means the a or b arrays do not contain
12263 n2 elements) and fallback to bn_mul_normal if either is not zero.
12264
12265 *Steve Henson*
12266
12267 * Fix escaping of non-ASCII characters when using the -subj option
12268 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12269
12270 *Lutz Jaenicke*
12271
12272 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12273 form for "surname", serialNumber has no short form.
12274 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12275 therefore remove "mail" short name for "internet 7".
12276 The OID for unique identifiers in X509 certificates is
12277 x500UniqueIdentifier, not uniqueIdentifier.
12278 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12279
12280 *Lutz Jaenicke*
12281
12282 * Add an "init" command to the ENGINE config module and auto initialize
12283 ENGINEs. Without any "init" command the ENGINE will be initialized
12284 after all ctrl commands have been executed on it. If init=1 the
12285 ENGINE is initialized at that point (ctrls before that point are run
12286 on the uninitialized ENGINE and after on the initialized one). If
12287 init=0 then the ENGINE will not be initialized at all.
12288
12289 *Steve Henson*
12290
12291 * Fix the 'app_verify_callback' interface so that the user-defined
12292 argument is actually passed to the callback: In the
12293 SSL_CTX_set_cert_verify_callback() prototype, the callback
12294 declaration has been changed from
12295 int (*cb)()
12296 into
12297 int (*cb)(X509_STORE_CTX *,void *);
12298 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12299 i=s->ctx->app_verify_callback(&ctx)
12300 has been changed into
12301 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12302
12303 To update applications using SSL_CTX_set_cert_verify_callback(),
12304 a dummy argument can be added to their callback functions.
12305
12306 *D. K. Smetters <smetters@parc.xerox.com>*
12307
12308 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12309
12310 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12311
12312 * Add and OPENSSL_LOAD_CONF define which will cause
12313 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12314 This allows older applications to transparently support certain
12315 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12316 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12317 load the config file and OPENSSL_add_all_algorithms_conf() which will
12318 always load it have also been added.
12319
12320 *Steve Henson*
12321
12322 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12323 Adjust NIDs and EVP layer.
12324
12325 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12326
12327 * Config modules support in openssl utility.
12328
12329 Most commands now load modules from the config file,
12330 though in a few (such as version) this isn't done
12331 because it couldn't be used for anything.
12332
12333 In the case of ca and req the config file used is
12334 the same as the utility itself: that is the -config
12335 command line option can be used to specify an
12336 alternative file.
12337
12338 *Steve Henson*
12339
12340 * Move default behaviour from OPENSSL_config(). If appname is NULL
12341 use "openssl_conf" if filename is NULL use default openssl config file.
12342
12343 *Steve Henson*
12344
12345 * Add an argument to OPENSSL_config() to allow the use of an alternative
12346 config section name. Add a new flag to tolerate a missing config file
12347 and move code to CONF_modules_load_file().
12348
12349 *Steve Henson*
12350
12351 * Support for crypto accelerator cards from Accelerated Encryption
12352 Processing, www.aep.ie. (Use engine 'aep')
12353 The support was copied from 0.9.6c [engine] and adapted/corrected
12354 to work with the new engine framework.
12355
12356 *AEP Inc. and Richard Levitte*
12357
12358 * Support for SureWare crypto accelerator cards from Baltimore
12359 Technologies. (Use engine 'sureware')
12360 The support was copied from 0.9.6c [engine] and adapted
12361 to work with the new engine framework.
12362
12363 *Richard Levitte*
12364
12365 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12366 make the newer ENGINE framework commands for the CHIL engine work.
12367
12368 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12369
12370 * Make it possible to produce shared libraries on ReliantUNIX.
12371
12372 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12373
12374 * Add the configuration target debug-linux-ppro.
12375 Make 'openssl rsa' use the general key loading routines
ec2bfb7d 12376 implemented in `apps.c`, and make those routines able to
5f8e6c50
DMSP
12377 handle the key format FORMAT_NETSCAPE and the variant
12378 FORMAT_IISSGC.
12379
12380 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12381
12382 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12383
12384 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12385
12386 * Add -keyform to rsautl, and document -engine.
12387
12388 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12389
12390 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12391 BIO_R_NO_SUCH_FILE error code rather than the generic
12392 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12393
12394 *Ben Laurie*
12395
12396 * Add new functions
12397 ERR_peek_last_error
12398 ERR_peek_last_error_line
12399 ERR_peek_last_error_line_data.
12400 These are similar to
12401 ERR_peek_error
12402 ERR_peek_error_line
12403 ERR_peek_error_line_data,
12404 but report on the latest error recorded rather than the first one
12405 still in the error queue.
12406
12407 *Ben Laurie, Bodo Moeller*
12408
12409 * default_algorithms option in ENGINE config module. This allows things
12410 like:
12411 default_algorithms = ALL
12412 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12413
12414 *Steve Henson*
12415
12416 * Preliminary ENGINE config module.
12417
12418 *Steve Henson*
12419
12420 * New experimental application configuration code.
12421
12422 *Steve Henson*
12423
12424 * Change the AES code to follow the same name structure as all other
12425 symmetric ciphers, and behave the same way. Move everything to
12426 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12427
12428 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12429
12430 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12431
12432 *Ben Laurie and Theo de Raadt*
12433
12434 * Add option to output public keys in req command.
12435
12436 *Massimiliano Pala madwolf@openca.org*
12437
12438 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12439 (up to about 10% better than before for P-192 and P-224).
12440
12441 *Bodo Moeller*
12442
12443 * New functions/macros
12444
12445 SSL_CTX_set_msg_callback(ctx, cb)
12446 SSL_CTX_set_msg_callback_arg(ctx, arg)
12447 SSL_set_msg_callback(ssl, cb)
12448 SSL_set_msg_callback_arg(ssl, arg)
12449
12450 to request calling a callback function
12451
12452 void cb(int write_p, int version, int content_type,
12453 const void *buf, size_t len, SSL *ssl, void *arg)
12454
12455 whenever a protocol message has been completely received
12456 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12457 protocol version according to which the SSL library interprets
12458 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12459 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12460 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12461 specification (change_cipher_spec(20), alert(21), handshake(22)).
12462 'buf' and 'len' point to the actual message, 'ssl' to the
12463 SSL object, and 'arg' is the application-defined value set by
12464 SSL[_CTX]_set_msg_callback_arg().
12465
12466 'openssl s_client' and 'openssl s_server' have new '-msg' options
12467 to enable a callback that displays all protocol messages.
12468
12469 *Bodo Moeller*
12470
12471 * Change the shared library support so shared libraries are built as
12472 soon as the corresponding static library is finished, and thereby get
12473 openssl and the test programs linked against the shared library.
12474 This still only happens when the keyword "shard" has been given to
12475 the configuration scripts.
12476
12477 NOTE: shared library support is still an experimental thing, and
12478 backward binary compatibility is still not guaranteed.
12479
12480 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12481
12482 * Add support for Subject Information Access extension.
12483
12484 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12485
12486 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12487 additional bytes when new memory had to be allocated, not just
12488 when reusing an existing buffer.
12489
12490 *Bodo Moeller*
12491
12492 * New command line and configuration option 'utf8' for the req command.
12493 This allows field values to be specified as UTF8 strings.
12494
12495 *Steve Henson*
12496
12497 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12498 runs for the former and machine-readable output for the latter.
12499
12500 *Ben Laurie*
12501
12502 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12503 of the e-mail address in the DN (i.e., it will go into a certificate
12504 extension only). The new configuration file option 'email_in_dn = no'
12505 has the same effect.
12506
12507 *Massimiliano Pala madwolf@openca.org*
12508
257e9d03
RS
12509 * Change all functions with names starting with `des_` to be starting
12510 with `DES_` instead. Add wrappers that are compatible with libdes,
12511 but are named `_ossl_old_des_*`. Finally, add macros that map the
12512 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
5f8e6c50 12513 compatibility is desired. If OpenSSL 0.9.6c compatibility is
257e9d03 12514 desired, the `des_*` symbols will be mapped to `DES_*`, with one
5f8e6c50
DMSP
12515 exception.
12516
12517 Since we provide two compatibility mappings, the user needs to
12518 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12519 compatibility is desired. The default (i.e., when that macro
12520 isn't defined) is OpenSSL 0.9.6c compatibility.
12521
12522 There are also macros that enable and disable the support of old
12523 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12524 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12525 are defined, the default will apply: to support the old des routines.
12526
12527 In either case, one must include openssl/des.h to get the correct
12528 definitions. Do not try to just include openssl/des_old.h, that
12529 won't work.
12530
12531 NOTE: This is a major break of an old API into a new one. Software
257e9d03 12532 authors are encouraged to switch to the `DES_` style functions. Some
5f8e6c50
DMSP
12533 time in the future, des_old.h and the libdes compatibility functions
12534 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12535 default), and then completely removed.
12536
12537 *Richard Levitte*
12538
12539 * Test for certificates which contain unsupported critical extensions.
12540 If such a certificate is found during a verify operation it is
12541 rejected by default: this behaviour can be overridden by either
12542 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12543 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12544 X509_supported_extension() has also been added which returns 1 if a
12545 particular extension is supported.
12546
12547 *Steve Henson*
12548
12549 * Modify the behaviour of EVP cipher functions in similar way to digests
12550 to retain compatibility with existing code.
12551
12552 *Steve Henson*
12553
12554 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12555 compatibility with existing code. In particular the 'ctx' parameter does
12556 not have to be to be initialized before the call to EVP_DigestInit() and
12557 it is tidied up after a call to EVP_DigestFinal(). New function
12558 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12559 EVP_MD_CTX_copy() changed to not require the destination to be
12560 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12561 requires the destination to be valid.
12562
12563 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12564 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12565
12566 *Steve Henson*
12567
12568 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12569 so that complete 'Handshake' protocol structures are kept in memory
12570 instead of overwriting 'msg_type' and 'length' with 'body' data.
12571
12572 *Bodo Moeller*
12573
12574 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12575
12576 *Massimo Santin via Richard Levitte*
12577
12578 * Major restructuring to the underlying ENGINE code. This includes
12579 reduction of linker bloat, separation of pure "ENGINE" manipulation
12580 (initialisation, etc) from functionality dealing with implementations
12581 of specific crypto interfaces. This change also introduces integrated
12582 support for symmetric ciphers and digest implementations - so ENGINEs
12583 can now accelerate these by providing EVP_CIPHER and EVP_MD
036cbb6b
DDO
12584 implementations of their own. This is detailed in
12585 [crypto/engine/README.md](crypto/engine/README.md)
5f8e6c50
DMSP
12586 as it couldn't be adequately described here. However, there are a few
12587 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12588 were changed in the original introduction of ENGINE code have now
12589 reverted back - the hooking from this code to ENGINE is now a good
12590 deal more passive and at run-time, operations deal directly with
12591 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12592 dereferencing through an ENGINE pointer any more. Also, the ENGINE
257e9d03 12593 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
5f8e6c50
DMSP
12594 they were not being used by the framework as there is no concept of a
12595 BIGNUM_METHOD and they could not be generalised to the new
12596 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12597 ENGINE_cpy() has been removed as it cannot be consistently defined in
12598 the new code.
12599
12600 *Geoff Thorpe*
12601
12602 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12603
12604 *Steve Henson*
12605
12606 * Change mkdef.pl to sort symbols that get the same entry number,
257e9d03 12607 and make sure the automatically generated functions `ERR_load_*`
5f8e6c50
DMSP
12608 become part of libeay.num as well.
12609
12610 *Richard Levitte*
12611
12612 * New function SSL_renegotiate_pending(). This returns true once
12613 renegotiation has been requested (either SSL_renegotiate() call
12614 or HelloRequest/ClientHello received from the peer) and becomes
12615 false once a handshake has been completed.
12616 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12617 sends a HelloRequest, but does not ensure that a handshake takes
12618 place. SSL_renegotiate_pending() is useful for checking if the
12619 client has followed the request.)
12620
12621 *Bodo Moeller*
12622
12623 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12624 By default, clients may request session resumption even during
12625 renegotiation (if session ID contexts permit); with this option,
12626 session resumption is possible only in the first handshake.
12627
12628 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12629 more bits available for options that should not be part of
12630 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12631
12632 *Bodo Moeller*
12633
12634 * Add some demos for certificate and certificate request creation.
12635
12636 *Steve Henson*
12637
12638 * Make maximum certificate chain size accepted from the peer application
257e9d03 12639 settable (`SSL*_get/set_max_cert_list()`), as proposed by
5f8e6c50
DMSP
12640 "Douglas E. Engert" <deengert@anl.gov>.
12641
12642 *Lutz Jaenicke*
12643
12644 * Add support for shared libraries for Unixware-7
12645 (Boyd Lynn Gerber <gerberb@zenez.com>).
12646
12647 *Lutz Jaenicke*
12648
12649 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12650 be done prior to destruction. Use this to unload error strings from
12651 ENGINEs that load their own error strings. NB: This adds two new API
12652 functions to "get" and "set" this destroy handler in an ENGINE.
12653
12654 *Geoff Thorpe*
12655
12656 * Alter all existing ENGINE implementations (except "openssl" and
12657 "openbsd") to dynamically instantiate their own error strings. This
12658 makes them more flexible to be built both as statically-linked ENGINEs
12659 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12660 Also, add stub code to each that makes building them as self-contained
036cbb6b 12661 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
5f8e6c50
DMSP
12662
12663 *Geoff Thorpe*
12664
12665 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12666 implementations into applications that are completely implemented in
12667 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12668 commands that can be used to configure what shared-library to load and
12669 to control aspects of the way it is handled. Also, made an update to
036cbb6b
DDO
12670 the [README-Engine.md](README-Engine.md) file
12671 that brings its information up-to-date and
5f8e6c50
DMSP
12672 provides some information and instructions on the "dynamic" ENGINE
12673 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12674
12675 *Geoff Thorpe*
12676
12677 * Make it possible to unload ranges of ERR strings with a new
12678 "ERR_unload_strings" function.
12679
12680 *Geoff Thorpe*
12681
12682 * Add a copy() function to EVP_MD.
12683
12684 *Ben Laurie*
12685
12686 * Make EVP_MD routines take a context pointer instead of just the
12687 md_data void pointer.
12688
12689 *Ben Laurie*
12690
12691 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12692 that the digest can only process a single chunk of data
12693 (typically because it is provided by a piece of
12694 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12695 is only going to provide a single chunk of data, and hence the
12696 framework needn't accumulate the data for oneshot drivers.
12697
12698 *Ben Laurie*
12699
12700 * As with "ERR", make it possible to replace the underlying "ex_data"
12701 functions. This change also alters the storage and management of global
12702 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12703 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12704 index counters. The API functions that use this state have been changed
12705 to take a "class_index" rather than pointers to the class's local STACK
12706 and counter, and there is now an API function to dynamically create new
12707 classes. This centralisation allows us to (a) plug a lot of the
12708 thread-safety problems that existed, and (b) makes it possible to clean
12709 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12710 such data would previously have always leaked in application code and
12711 workarounds were in place to make the memory debugging turn a blind eye
12712 to it. Application code that doesn't use this new function will still
12713 leak as before, but their memory debugging output will announce it now
12714 rather than letting it slide.
12715
12716 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12717 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12718 has a return value to indicate success or failure.
12719
12720 *Geoff Thorpe*
12721
12722 * Make it possible to replace the underlying "ERR" functions such that the
12723 global state (2 LHASH tables and 2 locks) is only used by the "default"
12724 implementation. This change also adds two functions to "get" and "set"
12725 the implementation prior to it being automatically set the first time
12726 any other ERR function takes place. Ie. an application can call "get",
12727 pass the return value to a module it has just loaded, and that module
12728 can call its own "set" function using that value. This means the
12729 module's "ERR" operations will use (and modify) the error state in the
12730 application and not in its own statically linked copy of OpenSSL code.
12731
12732 *Geoff Thorpe*
12733
257e9d03 12734 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
5f8e6c50
DMSP
12735 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12736 the operation, and provides a more encapsulated way for external code
12737 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12738 to use these functions rather than manually incrementing the counts.
12739
12740 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12741
12742 *Geoff Thorpe*
12743
12744 * Add EVP test program.
12745
12746 *Ben Laurie*
12747
12748 * Add symmetric cipher support to ENGINE. Expect the API to change!
12749
12750 *Ben Laurie*
12751
12752 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12753 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12754 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12755 These allow a CRL to be built without having to access X509_CRL fields
12756 directly. Modify 'ca' application to use new functions.
12757
12758 *Steve Henson*
12759
12760 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12761 bug workarounds. Rollback attack detection is a security feature.
12762 The problem will only arise on OpenSSL servers when TLSv1 is not
12763 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12764 Software authors not wanting to support TLSv1 will have special reasons
12765 for their choice and can explicitly enable this option.
12766
12767 *Bodo Moeller, Lutz Jaenicke*
12768
12769 * Rationalise EVP so it can be extended: don't include a union of
12770 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12771 (similar to those existing for EVP_CIPHER_CTX).
12772 Usage example:
12773
12774 EVP_MD_CTX md;
12775
12776 EVP_MD_CTX_init(&md); /* new function call */
12777 EVP_DigestInit(&md, EVP_sha1());
12778 EVP_DigestUpdate(&md, in, len);
12779 EVP_DigestFinal(&md, out, NULL);
12780 EVP_MD_CTX_cleanup(&md); /* new function call */
12781
5f8e6c50
DMSP
12782 *Ben Laurie*
12783
12784 * Make DES key schedule conform to the usual scheme, as well as
12785 correcting its structure. This means that calls to DES functions
12786 now have to pass a pointer to a des_key_schedule instead of a
12787 plain des_key_schedule (which was actually always a pointer
12788 anyway): E.g.,
12789
12790 des_key_schedule ks;
12791
12792 des_set_key_checked(..., &ks);
12793 des_ncbc_encrypt(..., &ks, ...);
12794
12795 (Note that a later change renames 'des_...' into 'DES_...'.)
12796
12797 *Ben Laurie*
12798
12799 * Initial reduction of linker bloat: the use of some functions, such as
12800 PEM causes large amounts of unused functions to be linked in due to
12801 poor organisation. For example pem_all.c contains every PEM function
12802 which has a knock on effect of linking in large amounts of (unused)
12803 ASN1 code. Grouping together similar functions and splitting unrelated
12804 functions prevents this.
12805
12806 *Steve Henson*
12807
12808 * Cleanup of EVP macros.
12809
12810 *Ben Laurie*
12811
257e9d03
RS
12812 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12813 correct `_ecb suffix`.
5f8e6c50
DMSP
12814
12815 *Ben Laurie*
12816
12817 * Add initial OCSP responder support to ocsp application. The
12818 revocation information is handled using the text based index
12819 use by the ca application. The responder can either handle
12820 requests generated internally, supplied in files (for example
12821 via a CGI script) or using an internal minimal server.
12822
12823 *Steve Henson*
12824
12825 * Add configuration choices to get zlib compression for TLS.
12826
12827 *Richard Levitte*
12828
12829 * Changes to Kerberos SSL for RFC 2712 compliance:
257e9d03
RS
12830 1. Implemented real KerberosWrapper, instead of just using
12831 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12832 2. Implemented optional authenticator field of KerberosWrapper.
5f8e6c50
DMSP
12833
12834 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12835 and authenticator structs; see crypto/krb5/.
12836
12837 Generalized Kerberos calls to support multiple Kerberos libraries.
257e9d03
RS
12838 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12839 via Richard Levitte*
5f8e6c50
DMSP
12840
12841 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12842 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12843 values for each of the key sizes rather than having just
12844 parameters (and 'speed' generating keys each time).
12845
12846 *Geoff Thorpe*
12847
12848 * Speed up EVP routines.
12849 Before:
12850crypt
12851pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12852s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12853s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12854s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12855crypt
12856s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12857s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12858s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12859 After:
12860crypt
12861s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12862crypt
12863s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12864
12865 *Ben Laurie*
12866
12867 * Added the OS2-EMX target.
12868
12869 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12870
ec2bfb7d 12871 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
3e3ad3c5 12872 New functions to support `NCONF` routines in extension code.
ec2bfb7d
DDO
12873 New function `CONF_set_nconf()`
12874 to allow functions which take an `NCONF` to also handle the old `LHASH`
12875 structure: this means that the old `CONF` compatible routines can be
12876 retained (in particular w.rt. extensions) without having to duplicate the
12877 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
5f8e6c50
DMSP
12878
12879 *Steve Henson*
12880
12881 * Enhance the general user interface with mechanisms for inner control
12882 and with possibilities to have yes/no kind of prompts.
12883
12884 *Richard Levitte*
12885
4d49b685 12886 * Change all calls to low-level digest routines in the library and
5f8e6c50
DMSP
12887 applications to use EVP. Add missing calls to HMAC_cleanup() and
12888 don't assume HMAC_CTX can be copied using memcpy().
12889
12890 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12891
12892 * Add the possibility to control engines through control names but with
12893 arbitrary arguments instead of just a string.
12894 Change the key loaders to take a UI_METHOD instead of a callback
12895 function pointer. NOTE: this breaks binary compatibility with earlier
12896 versions of OpenSSL [engine].
12897 Adapt the nCipher code for these new conditions and add a card insertion
12898 callback.
12899
12900 *Richard Levitte*
12901
12902 * Enhance the general user interface with mechanisms to better support
12903 dialog box interfaces, application-defined prompts, the possibility
12904 to use defaults (for example default passwords from somewhere else)
12905 and interrupts/cancellations.
12906
12907 *Richard Levitte*
12908
12909 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12910 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12911
12912 *Steve Henson*
12913
12914 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12915 tidy up some unnecessarily weird code in 'sk_new()').
12916
12917 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12918
12919 * Change the key loading routines for ENGINEs to use the same kind
12920 callback (pem_password_cb) as all other routines that need this
12921 kind of callback.
12922
12923 *Richard Levitte*
12924
12925 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12926 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12927 than this minimum value is recommended.
12928
12929 *Lutz Jaenicke*
12930
12931 * New random seeder for OpenVMS, using the system process statistics
12932 that are easily reachable.
12933
12934 *Richard Levitte*
12935
12936 * Windows apparently can't transparently handle global
12937 variables defined in DLLs. Initialisations such as:
12938
12939 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12940
12941 won't compile. This is used by the any applications that need to
12942 declare their own ASN1 modules. This was fixed by adding the option
12943 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12944 needed for static libraries under Win32.
12945
12946 *Steve Henson*
12947
12948 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12949 setting of purpose and trust fields. New X509_STORE trust and
12950 purpose functions and tidy up setting in other SSL functions.
12951
12952 *Steve Henson*
12953
12954 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12955 structure. These are inherited by X509_STORE_CTX when it is
12956 initialised. This allows various defaults to be set in the
12957 X509_STORE structure (such as flags for CRL checking and custom
12958 purpose or trust settings) for functions which only use X509_STORE_CTX
12959 internally such as S/MIME.
12960
12961 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12962 trust settings if they are not set in X509_STORE. This allows X509_STORE
12963 purposes and trust (in S/MIME for example) to override any set by default.
12964
12965 Add command line options for CRL checking to smime, s_client and s_server
12966 applications.
12967
12968 *Steve Henson*
12969
12970 * Initial CRL based revocation checking. If the CRL checking flag(s)
12971 are set then the CRL is looked up in the X509_STORE structure and
12972 its validity and signature checked, then if the certificate is found
12973 in the CRL the verify fails with a revoked error.
12974
12975 Various new CRL related callbacks added to X509_STORE_CTX structure.
12976
12977 Command line options added to 'verify' application to support this.
12978
12979 This needs some additional work, such as being able to handle multiple
12980 CRLs with different times, extension based lookup (rather than just
12981 by subject name) and ultimately more complete V2 CRL extension
12982 handling.
12983
12984 *Steve Henson*
12985
12986 * Add a general user interface API (crypto/ui/). This is designed
12987 to replace things like des_read_password and friends (backward
12988 compatibility functions using this new API are provided).
12989 The purpose is to remove prompting functions from the DES code
12990 section as well as provide for prompting through dialog boxes in
12991 a window system and the like.
12992
12993 *Richard Levitte*
12994
12995 * Add "ex_data" support to ENGINE so implementations can add state at a
12996 per-structure level rather than having to store it globally.
12997
12998 *Geoff*
12999
13000 * Make it possible for ENGINE structures to be copied when retrieved by
13001 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13002 This causes the "original" ENGINE structure to act like a template,
13003 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13004 operational state can be localised to each ENGINE structure, despite the
13005 fact they all share the same "methods". New ENGINE structures returned in
13006 this case have no functional references and the return value is the single
13007 structural reference. This matches the single structural reference returned
13008 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13009 ENGINE structure.
13010
13011 *Geoff*
13012
13013 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13014 needs to match any other type at all we need to manually clear the
13015 tag cache.
13016
13017 *Steve Henson*
13018
13019 * Changes to the "openssl engine" utility to include;
13020 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13021 about an ENGINE's available control commands.
13022 - executing control commands from command line arguments using the
13023 '-pre' and '-post' switches. '-post' is only used if '-t' is
13024 specified and the ENGINE is successfully initialised. The syntax for
13025 the individual commands are colon-separated, for example;
13026 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13027
13028 *Geoff*
13029
13030 * New dynamic control command support for ENGINEs. ENGINEs can now
13031 declare their own commands (numbers), names (strings), descriptions,
13032 and input types for run-time discovery by calling applications. A
13033 subset of these commands are implicitly classed as "executable"
13034 depending on their input type, and only these can be invoked through
13035 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13036 can be based on user input, config files, etc). The distinction is
13037 that "executable" commands cannot return anything other than a boolean
13038 result and can only support numeric or string input, whereas some
13039 discoverable commands may only be for direct use through
13040 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13041 pointers, or other custom uses. The "executable" commands are to
13042 support parameterisations of ENGINE behaviour that can be
13043 unambiguously defined by ENGINEs and used consistently across any
13044 OpenSSL-based application. Commands have been added to all the
13045 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13046 control over shared-library paths without source code alterations.
13047
13048 *Geoff*
13049
13050 * Changed all ENGINE implementations to dynamically allocate their
13051 ENGINEs rather than declaring them statically. Apart from this being
13052 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13053 this also allows the implementations to compile without using the
13054 internal engine_int.h header.
13055
13056 *Geoff*
13057
13058 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13059 'const' value. Any code that should be able to modify a RAND_METHOD
13060 should already have non-const pointers to it (ie. they should only
13061 modify their own ones).
13062
13063 *Geoff*
13064
13065 * Made a variety of little tweaks to the ENGINE code.
13066 - "atalla" and "ubsec" string definitions were moved from header files
13067 to C code. "nuron" string definitions were placed in variables
13068 rather than hard-coded - allowing parameterisation of these values
13069 later on via ctrl() commands.
13070 - Removed unused "#if 0"'d code.
13071 - Fixed engine list iteration code so it uses ENGINE_free() to release
13072 structural references.
13073 - Constified the RAND_METHOD element of ENGINE structures.
13074 - Constified various get/set functions as appropriate and added
13075 missing functions (including a catch-all ENGINE_cpy that duplicates
13076 all ENGINE values onto a new ENGINE except reference counts/state).
13077 - Removed NULL parameter checks in get/set functions. Setting a method
13078 or function to NULL is a way of cancelling out a previously set
13079 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13080 and doesn't justify the extra error symbols and code.
13081 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13082 flags from engine_int.h to engine.h.
13083 - Changed prototypes for ENGINE handler functions (init(), finish(),
13084 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13085
13086 *Geoff*
13087
13088 * Implement binary inversion algorithm for BN_mod_inverse in addition
13089 to the algorithm using long division. The binary algorithm can be
13090 used only if the modulus is odd. On 32-bit systems, it is faster
13091 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13092 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13093 up to 450 bits. In 64-bit environments, the binary algorithm
13094 appears to be advantageous for much longer moduli; here we use it
13095 for moduli up to 2048 bits.
13096
13097 *Bodo Moeller*
13098
13099 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13100 could not support the combine flag in choice fields.
13101
13102 *Steve Henson*
13103
13104 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13105 extensions from a certificate request to the certificate.
13106
13107 *Steve Henson*
13108
13109 * Allow multiple 'certopt' and 'nameopt' options to be separated
13110 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13111 file: this allows the display of the certificate about to be
13112 signed to be customised, to allow certain fields to be included
13113 or excluded and extension details. The old system didn't display
13114 multicharacter strings properly, omitted fields not in the policy
13115 and couldn't display additional details such as extensions.
13116
13117 *Steve Henson*
13118
13119 * Function EC_POINTs_mul for multiple scalar multiplication
13120 of an arbitrary number of elliptic curve points
13121 \sum scalars[i]*points[i],
13122 optionally including the generator defined for the EC_GROUP:
13123 scalar*generator + \sum scalars[i]*points[i].
13124
13125 EC_POINT_mul is a simple wrapper function for the typical case
13126 that the point list has just one item (besides the optional
13127 generator).
13128
13129 *Bodo Moeller*
13130
13131 * First EC_METHODs for curves over GF(p):
13132
13133 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13134 operations and provides various method functions that can also
13135 operate with faster implementations of modular arithmetic.
13136
13137 EC_GFp_mont_method() reuses most functions that are part of
13138 EC_GFp_simple_method, but uses Montgomery arithmetic.
13139
13140 *Bodo Moeller; point addition and point doubling
13141 implementation directly derived from source code provided by
13142 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13143
13144 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13145 crypto/ec/ec_lib.c):
13146
13147 Curves are EC_GROUP objects (with an optional group generator)
13148 based on EC_METHODs that are built into the library.
13149
13150 Points are EC_POINT objects based on EC_GROUP objects.
13151
13152 Most of the framework would be able to handle curves over arbitrary
13153 finite fields, but as there are no obvious types for fields other
13154 than GF(p), some functions are limited to that for now.
13155
13156 *Bodo Moeller*
13157
13158 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13159 that the file contains a complete HTTP response.
13160
13161 *Richard Levitte*
13162
13163 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13164 change the def and num file printf format specifier from "%-40sXXX"
13165 to "%-39s XXX". The latter will always guarantee a space after the
13166 field while the former will cause them to run together if the field
13167 is 40 of more characters long.
13168
13169 *Steve Henson*
13170
13171 * Constify the cipher and digest 'method' functions and structures
13172 and modify related functions to take constant EVP_MD and EVP_CIPHER
13173 pointers.
13174
13175 *Steve Henson*
13176
13177 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13178 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13179
13180 *Bodo Moeller*
13181
257e9d03 13182 * Modify `EVP_Digest*()` routines so they now return values. Although the
5f8e6c50
DMSP
13183 internal software routines can never fail additional hardware versions
13184 might.
13185
13186 *Steve Henson*
13187
13188 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13189
13190 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13191 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13192
13193 ASN1 error codes
13194 ERR_R_NESTED_ASN1_ERROR
13195 ...
13196 ERR_R_MISSING_ASN1_EOS
13197 were 4 .. 9, conflicting with
13198 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13199 ...
13200 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13201 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13202
13203 Add new error code 'ERR_R_INTERNAL_ERROR'.
13204
13205 *Bodo Moeller*
13206
13207 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13208 suffices.
13209
13210 *Bodo Moeller*
13211
13212 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13213 sets the subject name for a new request or supersedes the
13214 subject name in a given request. Formats that can be parsed are
13215 'CN=Some Name, OU=myOU, C=IT'
13216 and
13217 'CN=Some Name/OU=myOU/C=IT'.
13218
13219 Add options '-batch' and '-verbose' to 'openssl req'.
13220
13221 *Massimiliano Pala <madwolf@hackmasters.net>*
13222
13223 * Introduce the possibility to access global variables through
13224 functions on platform were that's the best way to handle exporting
13225 global variables in shared libraries. To enable this functionality,
13226 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13227 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13228 is normally done by Configure or something similar).
13229
13230 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13231 in the source file (foo.c) like this:
13232
13233 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13234 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13235
13236 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13237 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13238
13239 OPENSSL_DECLARE_GLOBAL(int,foo);
13240 #define foo OPENSSL_GLOBAL_REF(foo)
13241 OPENSSL_DECLARE_GLOBAL(double,bar);
13242 #define bar OPENSSL_GLOBAL_REF(bar)
13243
13244 The #defines are very important, and therefore so is including the
13245 header file everywhere where the defined globals are used.
13246
13247 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13248 of ASN.1 items, but that structure is a bit different.
13249
13250 The largest change is in util/mkdef.pl which has been enhanced with
13251 better and easier to understand logic to choose which symbols should
13252 go into the Windows .def files as well as a number of fixes and code
13253 cleanup (among others, algorithm keywords are now sorted
13254 lexicographically to avoid constant rewrites).
13255
13256 *Richard Levitte*
13257
13258 * In BN_div() keep a copy of the sign of 'num' before writing the
13259 result to 'rm' because if rm==num the value will be overwritten
13260 and produce the wrong result if 'num' is negative: this caused
13261 problems with BN_mod() and BN_nnmod().
13262
13263 *Steve Henson*
13264
13265 * Function OCSP_request_verify(). This checks the signature on an
13266 OCSP request and verifies the signer certificate. The signer
13267 certificate is just checked for a generic purpose and OCSP request
13268 trust settings.
13269
13270 *Steve Henson*
13271
13272 * Add OCSP_check_validity() function to check the validity of OCSP
13273 responses. OCSP responses are prepared in real time and may only
13274 be a few seconds old. Simply checking that the current time lies
13275 between thisUpdate and nextUpdate max reject otherwise valid responses
13276 caused by either OCSP responder or client clock inaccuracy. Instead
13277 we allow thisUpdate and nextUpdate to fall within a certain period of
13278 the current time. The age of the response can also optionally be
13279 checked. Two new options -validity_period and -status_age added to
13280 ocsp utility.
13281
13282 *Steve Henson*
13283
13284 * If signature or public key algorithm is unrecognized print out its
13285 OID rather that just UNKNOWN.
13286
13287 *Steve Henson*
13288
13289 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13290 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13291 ID to be generated from the issuer certificate alone which can then be
13292 passed to OCSP_id_issuer_cmp().
13293
13294 *Steve Henson*
13295
13296 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13297 ASN1 modules to export functions returning ASN1_ITEM pointers
13298 instead of the ASN1_ITEM structures themselves. This adds several
13299 new macros which allow the underlying ASN1 function/structure to
13300 be accessed transparently. As a result code should not use ASN1_ITEM
13301 references directly (such as &X509_it) but instead use the relevant
13302 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13303 use of the new ASN1 code on platforms where exporting structures
13304 is problematical (for example in shared libraries) but exporting
13305 functions returning pointers to structures is not.
13306
13307 *Steve Henson*
13308
13309 * Add support for overriding the generation of SSL/TLS session IDs.
13310 These callbacks can be registered either in an SSL_CTX or per SSL.
13311 The purpose of this is to allow applications to control, if they wish,
13312 the arbitrary values chosen for use as session IDs, particularly as it
13313 can be useful for session caching in multiple-server environments. A
13314 command-line switch for testing this (and any client code that wishes
13315 to use such a feature) has been added to "s_server".
13316
13317 *Geoff Thorpe, Lutz Jaenicke*
13318
13319 * Modify mkdef.pl to recognise and parse preprocessor conditionals
1dc1ea18
DDO
13320 of the form `#if defined(...) || defined(...) || ...` and
13321 `#if !defined(...) && !defined(...) && ...`. This also avoids
5f8e6c50
DMSP
13322 the growing number of special cases it was previously handling.
13323
13324 *Richard Levitte*
13325
13326 * Make all configuration macros available for application by making
13327 sure they are available in opensslconf.h, by giving them names starting
257e9d03 13328 with `OPENSSL_` to avoid conflicts with other packages and by making
5f8e6c50
DMSP
13329 sure e_os2.h will cover all platform-specific cases together with
13330 opensslconf.h.
13331 Additionally, it is now possible to define configuration/platform-
13332 specific names (called "system identities"). In the C code, these
257e9d03
RS
13333 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13334 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13335 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
5f8e6c50
DMSP
13336 what is available.
13337
13338 *Richard Levitte*
13339
13340 * New option -set_serial to 'req' and 'x509' this allows the serial
13341 number to use to be specified on the command line. Previously self
13342 signed certificates were hard coded with serial number 0 and the
13343 CA options of 'x509' had to use a serial number in a file which was
13344 auto incremented.
13345
13346 *Steve Henson*
13347
13348 * New options to 'ca' utility to support V2 CRL entry extensions.
13349 Currently CRL reason, invalidity date and hold instruction are
13350 supported. Add new CRL extensions to V3 code and some new objects.
13351
13352 *Steve Henson*
13353
13354 * New function EVP_CIPHER_CTX_set_padding() this is used to
13355 disable standard block padding (aka PKCS#5 padding) in the EVP
13356 API, which was previously mandatory. This means that the data is
13357 not padded in any way and so the total length much be a multiple
13358 of the block size, otherwise an error occurs.
13359
13360 *Steve Henson*
13361
13362 * Initial (incomplete) OCSP SSL support.
13363
13364 *Steve Henson*
13365
13366 * New function OCSP_parse_url(). This splits up a URL into its host,
13367 port and path components: primarily to parse OCSP URLs. New -url
13368 option to ocsp utility.
13369
13370 *Steve Henson*
13371
13372 * New nonce behavior. The return value of OCSP_check_nonce() now
13373 reflects the various checks performed. Applications can decide
13374 whether to tolerate certain situations such as an absent nonce
13375 in a response when one was present in a request: the ocsp application
13376 just prints out a warning. New function OCSP_add1_basic_nonce()
13377 this is to allow responders to include a nonce in a response even if
13378 the request is nonce-less.
13379
13380 *Steve Henson*
13381
ec2bfb7d 13382 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
5f8e6c50 13383 skipped when using openssl x509 multiple times on a single input file,
257e9d03 13384 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
5f8e6c50
DMSP
13385
13386 *Bodo Moeller*
13387
13388 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13389 set string type: to handle setting ASN1_TIME structures. Fix ca
13390 utility to correctly initialize revocation date of CRLs.
13391
13392 *Steve Henson*
13393
13394 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13395 the clients preferred ciphersuites and rather use its own preferences.
13396 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13397 Internet Explorer by ensuring unchanged hash method during stepup.
13398 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13399
13400 *Lutz Jaenicke*
13401
13402 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13403 to aes and add a new 'exist' option to print out symbols that don't
13404 appear to exist.
13405
13406 *Steve Henson*
13407
13408 * Additional options to ocsp utility to allow flags to be set and
13409 additional certificates supplied.
13410
13411 *Steve Henson*
13412
13413 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13414 OCSP client a number of certificate to only verify the response
13415 signature against.
13416
13417 *Richard Levitte*
13418
13419 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13420 handle the new API. Currently only ECB, CBC modes supported. Add new
13421 AES OIDs.
13422
13423 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13424 Encryption Standard (AES) Ciphersuites for Transport Layer
13425 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13426 not enabled by default and were not part of the "ALL" ciphersuite
13427 alias because they were not yet official; they could be
13428 explicitly requested by specifying the "AESdraft" ciphersuite
13429 group alias. In the final release of OpenSSL 0.9.7, the group
13430 alias is called "AES" and is part of "ALL".)
13431
13432 *Ben Laurie, Steve Henson, Bodo Moeller*
13433
13434 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13435 request to response.
13436
13437 *Steve Henson*
13438
13439 * Functions for OCSP responders. OCSP_request_onereq_count(),
13440 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13441 extract information from a certificate request. OCSP_response_create()
13442 creates a response and optionally adds a basic response structure.
13443 OCSP_basic_add1_status() adds a complete single response to a basic
13444 response and returns the OCSP_SINGLERESP structure just added (to allow
13445 extensions to be included for example). OCSP_basic_add1_cert() adds a
13446 certificate to a basic response and OCSP_basic_sign() signs a basic
13447 response with various flags. New helper functions ASN1_TIME_check()
13448 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13449 (converts ASN1_TIME to GeneralizedTime).
13450
13451 *Steve Henson*
13452
13453 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13454 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13455 structure from a certificate. X509_pubkey_digest() digests the public_key
13456 contents: this is used in various key identifiers.
13457
13458 *Steve Henson*
13459
13460 * Make sk_sort() tolerate a NULL argument.
13461
13462 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13463
13464 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13465 passed by the function are trusted implicitly. If any of them signed the
13466 response then it is assumed to be valid and is not verified.
13467
13468 *Steve Henson*
13469
13470 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13471 to data. This was previously part of the PKCS7 ASN1 code. This
13472 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13473 *Steve Henson, reported by Kenneth R. Robinette
13474 <support@securenetterm.com>*
13475
13476 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13477 routines: without these tracing memory leaks is very painful.
13478 Fix leaks in PKCS12 and PKCS7 routines.
13479
13480 *Steve Henson*
13481
13482 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13483 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13484 effectively meant GeneralizedTime would never be used. Now it
13485 is initialised to -1 but X509_time_adj() now has to check the value
13486 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13487 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13488 *Steve Henson, reported by Kenneth R. Robinette
13489 <support@securenetterm.com>*
13490
13491 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13492 result in a zero length in the ASN1_INTEGER structure which was
13493 not consistent with the structure when d2i_ASN1_INTEGER() was used
13494 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13495 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13496 where it did not print out a minus for negative ASN1_INTEGER.
13497
13498 *Steve Henson*
13499
13500 * Add summary printout to ocsp utility. The various functions which
13501 convert status values to strings have been renamed to:
13502 OCSP_response_status_str(), OCSP_cert_status_str() and
13503 OCSP_crl_reason_str() and are no longer static. New options
13504 to verify nonce values and to disable verification. OCSP response
13505 printout format cleaned up.
13506
13507 *Steve Henson*
13508
13509 * Add additional OCSP certificate checks. These are those specified
13510 in RFC2560. This consists of two separate checks: the CA of the
13511 certificate being checked must either be the OCSP signer certificate
13512 or the issuer of the OCSP signer certificate. In the latter case the
13513 OCSP signer certificate must contain the OCSP signing extended key
13514 usage. This check is performed by attempting to match the OCSP
13515 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13516 in the OCSP_CERTID structures of the response.
13517
13518 *Steve Henson*
13519
13520 * Initial OCSP certificate verification added to OCSP_basic_verify()
13521 and related routines. This uses the standard OpenSSL certificate
13522 verify routines to perform initial checks (just CA validity) and
13523 to obtain the certificate chain. Then additional checks will be
13524 performed on the chain. Currently the root CA is checked to see
13525 if it is explicitly trusted for OCSP signing. This is used to set
13526 a root CA as a global signing root: that is any certificate that
13527 chains to that CA is an acceptable OCSP signing certificate.
13528
13529 *Steve Henson*
13530
13531 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13532 extensions from a separate configuration file.
13533 As when reading extensions from the main configuration file,
13534 the '-extensions ...' option may be used for specifying the
13535 section to use.
13536
13537 *Massimiliano Pala <madwolf@comune.modena.it>*
13538
13539 * New OCSP utility. Allows OCSP requests to be generated or
13540 read. The request can be sent to a responder and the output
44652c16 13541 parsed, outputted or printed in text form. Not complete yet:
5f8e6c50
DMSP
13542 still needs to check the OCSP response validity.
13543
13544 *Steve Henson*
13545
13546 * New subcommands for 'openssl ca':
257e9d03 13547 `openssl ca -status <serial>` prints the status of the cert with
5f8e6c50 13548 the given serial number (according to the index file).
257e9d03 13549 `openssl ca -updatedb` updates the expiry status of certificates
5f8e6c50
DMSP
13550 in the index file.
13551
13552 *Massimiliano Pala <madwolf@comune.modena.it>*
13553
13554 * New '-newreq-nodes' command option to CA.pl. This is like
13555 '-newreq', but calls 'openssl req' with the '-nodes' option
13556 so that the resulting key is not encrypted.
13557
13558 *Damien Miller <djm@mindrot.org>*
13559
13560 * New configuration for the GNU Hurd.
13561
13562 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13563
13564 * Initial code to implement OCSP basic response verify. This
13565 is currently incomplete. Currently just finds the signer's
13566 certificate and verifies the signature on the response.
13567
13568 *Steve Henson*
13569
13570 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13571 value of OPENSSLDIR. This is available via the new '-d' option
13572 to 'openssl version', and is also included in 'openssl version -a'.
13573
13574 *Bodo Moeller*
13575
13576 * Allowing defining memory allocation callbacks that will be given
13577 file name and line number information in additional arguments
257e9d03 13578 (a `const char*` and an int). The basic functionality remains, as
5f8e6c50
DMSP
13579 well as the original possibility to just replace malloc(),
13580 realloc() and free() by functions that do not know about these
13581 additional arguments. To register and find out the current
13582 settings for extended allocation functions, the following
13583 functions are provided:
13584
13585 CRYPTO_set_mem_ex_functions
13586 CRYPTO_set_locked_mem_ex_functions
13587 CRYPTO_get_mem_ex_functions
13588 CRYPTO_get_locked_mem_ex_functions
13589
13590 These work the same way as CRYPTO_set_mem_functions and friends.
257e9d03 13591 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
5f8e6c50 13592 extended allocation function is enabled.
257e9d03 13593 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
5f8e6c50
DMSP
13594 a conventional allocation function is enabled.
13595
13596 *Richard Levitte, Bodo Moeller*
13597
13598 * Finish off removing the remaining LHASH function pointer casts.
13599 There should no longer be any prototype-casting required when using
13600 the LHASH abstraction, and any casts that remain are "bugs". See
13601 the callback types and macros at the head of lhash.h for details
13602 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13603
13604 *Geoff Thorpe*
13605
13606 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13607 If /dev/[u]random devices are not available or do not return enough
13608 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13609 be queried.
13610 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13611 /etc/entropy will be queried once each in this sequence, querying stops
13612 when enough entropy was collected without querying more sockets.
13613
13614 *Lutz Jaenicke*
13615
13616 * Change the Unix RAND_poll() variant to be able to poll several
13617 random devices, as specified by DEVRANDOM, until a sufficient amount
13618 of data has been collected. We spend at most 10 ms on each file
13619 (select timeout) and read in non-blocking mode. DEVRANDOM now
13620 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13621 (previously it was just the string "/dev/urandom"), so on typical
13622 platforms the 10 ms delay will never occur.
13623 Also separate out the Unix variant to its own file, rand_unix.c.
13624 For VMS, there's a currently-empty rand_vms.c.
13625
13626 *Richard Levitte*
13627
13628 * Move OCSP client related routines to ocsp_cl.c. These
13629 provide utility functions which an application needing
13630 to issue a request to an OCSP responder and analyse the
13631 response will typically need: as opposed to those which an
13632 OCSP responder itself would need which will be added later.
13633
13634 OCSP_request_sign() signs an OCSP request with an API similar
13635 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13636 response. OCSP_response_get1_basic() extracts basic response
13637 from response. OCSP_resp_find_status(): finds and extracts status
13638 information from an OCSP_CERTID structure (which will be created
13639 when the request structure is built). These are built from lower
13640 level functions which work on OCSP_SINGLERESP structures but
13641 won't normally be used unless the application wishes to examine
13642 extensions in the OCSP response for example.
13643
13644 Replace nonce routines with a pair of functions.
13645 OCSP_request_add1_nonce() adds a nonce value and optionally
13646 generates a random value. OCSP_check_nonce() checks the
13647 validity of the nonce in an OCSP response.
13648
13649 *Steve Henson*
13650
13651 * Change function OCSP_request_add() to OCSP_request_add0_id().
13652 This doesn't copy the supplied OCSP_CERTID and avoids the
13653 need to free up the newly created id. Change return type
13654 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13655 This can then be used to add extensions to the request.
13656 Deleted OCSP_request_new(), since most of its functionality
13657 is now in OCSP_REQUEST_new() (and the case insensitive name
13658 clash) apart from the ability to set the request name which
13659 will be added elsewhere.
13660
13661 *Steve Henson*
13662
13663 * Update OCSP API. Remove obsolete extensions argument from
13664 various functions. Extensions are now handled using the new
13665 OCSP extension code. New simple OCSP HTTP function which
13666 can be used to send requests and parse the response.
13667
13668 *Steve Henson*
13669
13670 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13671 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13672 uses the special reorder version of SET OF to sort the attributes
13673 and reorder them to match the encoded order. This resolves a long
13674 standing problem: a verify on a PKCS7 structure just after signing
13675 it used to fail because the attribute order did not match the
13676 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13677 it uses the received order. This is necessary to tolerate some broken
13678 software that does not order SET OF. This is handled by encoding
13679 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13680 to produce the required SET OF.
13681
13682 *Steve Henson*
13683
13684 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13685 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13686 files to get correct declarations of the ASN.1 item variables.
13687
13688 *Richard Levitte*
13689
13690 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13691 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13692 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13693 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13694 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13695 ASN1_ITEM and no wrapper functions.
13696
13697 *Steve Henson*
13698
13699 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13700 replace the old function pointer based I/O routines. Change most of
257e9d03 13701 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
5f8e6c50
DMSP
13702
13703 *Steve Henson*
13704
13705 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13706 lines, recognize more "algorithms" that can be deselected, and make
13707 it complain about algorithm deselection that isn't recognised.
13708
13709 *Richard Levitte*
13710
13711 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13712 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13713 to use new functions. Add NO_ASN1_OLD which can be set to remove
13714 some old style ASN1 functions: this can be used to determine if old
13715 code will still work when these eventually go away.
13716
13717 *Steve Henson*
13718
13719 * New extension functions for OCSP structures, these follow the
13720 same conventions as certificates and CRLs.
13721
13722 *Steve Henson*
13723
13724 * New function X509V3_add1_i2d(). This automatically encodes and
13725 adds an extension. Its behaviour can be customised with various
13726 flags to append, replace or delete. Various wrappers added for
13727 certificates and CRLs.
13728
13729 *Steve Henson*
13730
13731 * Fix to avoid calling the underlying ASN1 print routine when
13732 an extension cannot be parsed. Correct a typo in the
13733 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13734
13735 *Steve Henson*
13736
13737 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13738 entries for variables.
13739
13740 *Steve Henson*
13741
ec2bfb7d 13742 * Add functionality to `apps/openssl.c` for detecting locking
5f8e6c50
DMSP
13743 problems: As the program is single-threaded, all we have
13744 to do is register a locking callback using an array for
13745 storing which locks are currently held by the program.
13746
13747 *Bodo Moeller*
13748
13749 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13750 SSL_get_ex_data_X509_STORE_idx(), which is used in
13751 ssl_verify_cert_chain() and thus can be called at any time
13752 during TLS/SSL handshakes so that thread-safety is essential.
13753 Unfortunately, the ex_data design is not at all suited
13754 for multi-threaded use, so it probably should be abolished.
13755
13756 *Bodo Moeller*
13757
13758 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13759
13760 *Broadcom, tweaked and integrated by Geoff Thorpe*
13761
13762 * Move common extension printing code to new function
13763 X509V3_print_extensions(). Reorganise OCSP print routines and
13764 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13765
13766 *Steve Henson*
13767
13768 * New function X509_signature_print() to remove duplication in some
13769 print routines.
13770
13771 *Steve Henson*
13772
13773 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13774 set (this was treated exactly the same as SET OF previously). This
13775 is used to reorder the STACK representing the structure to match the
13776 encoding. This will be used to get round a problem where a PKCS7
13777 structure which was signed could not be verified because the STACK
13778 order did not reflect the encoded order.
13779
13780 *Steve Henson*
13781
13782 * Reimplement the OCSP ASN1 module using the new code.
13783
13784 *Steve Henson*
13785
13786 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13787 for its ASN1 operations. The old style function pointers still exist
13788 for now but they will eventually go away.
13789
13790 *Steve Henson*
13791
13792 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13793 completely replaces the old ASN1 functionality with a table driven
13794 encoder and decoder which interprets an ASN1_ITEM structure describing
13795 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13796 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13797 has also been converted to the new form.
13798
13799 *Steve Henson*
13800
13801 * Change BN_mod_exp_recp so that negative moduli are tolerated
13802 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13803 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13804 for negative moduli.
13805
13806 *Bodo Moeller*
13807
13808 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13809 of not touching the result's sign bit.
13810
13811 *Bodo Moeller*
13812
13813 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13814 set.
13815
13816 *Bodo Moeller*
13817
13818 * Changed the LHASH code to use prototypes for callbacks, and created
13819 macros to declare and implement thin (optionally static) functions
13820 that provide type-safety and avoid function pointer casting for the
13821 type-specific callbacks.
13822
13823 *Geoff Thorpe*
13824
13825 * Added Kerberos Cipher Suites to be used with TLS, as written in
13826 RFC 2712.
13827 *Veers Staats <staatsvr@asc.hpc.mil>,
257e9d03 13828 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
5f8e6c50
DMSP
13829
13830 * Reformat the FAQ so the different questions and answers can be divided
13831 in sections depending on the subject.
13832
13833 *Richard Levitte*
13834
13835 * Have the zlib compression code load ZLIB.DLL dynamically under
13836 Windows.
13837
13838 *Richard Levitte*
13839
13840 * New function BN_mod_sqrt for computing square roots modulo a prime
13841 (using the probabilistic Tonelli-Shanks algorithm unless
13842 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13843 be handled deterministically).
13844
13845 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13846
13847 * Make BN_mod_inverse faster by explicitly handling small quotients
13848 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13849 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13850
13851 *Bodo Moeller*
13852
13853 * New function BN_kronecker.
13854
13855 *Bodo Moeller*
13856
13857 * Fix BN_gcd so that it works on negative inputs; the result is
13858 positive unless both parameters are zero.
13859 Previously something reasonably close to an infinite loop was
13860 possible because numbers could be growing instead of shrinking
13861 in the implementation of Euclid's algorithm.
13862
13863 *Bodo Moeller*
13864
13865 * Fix BN_is_word() and BN_is_one() macros to take into account the
13866 sign of the number in question.
13867
13868 Fix BN_is_word(a,w) to work correctly for w == 0.
13869
13870 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13871 because its test if the absolute value of 'a' equals 'w'.
13872 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13873 it exists mostly for use in the implementations of BN_is_zero(),
13874 BN_is_one(), and BN_is_word().
13875
13876 *Bodo Moeller*
13877
13878 * New function BN_swap.
13879
13880 *Bodo Moeller*
13881
13882 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13883 the exponentiation functions are more likely to produce reasonable
13884 results on negative inputs.
13885
13886 *Bodo Moeller*
13887
13888 * Change BN_mod_mul so that the result is always non-negative.
13889 Previously, it could be negative if one of the factors was negative;
13890 I don't think anyone really wanted that behaviour.
13891
13892 *Bodo Moeller*
13893
1dc1ea18
DDO
13894 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13895 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13896 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
5f8e6c50
DMSP
13897 and add new functions:
13898
13899 BN_nnmod
13900 BN_mod_sqr
13901 BN_mod_add
13902 BN_mod_add_quick
13903 BN_mod_sub
13904 BN_mod_sub_quick
13905 BN_mod_lshift1
13906 BN_mod_lshift1_quick
13907 BN_mod_lshift
13908 BN_mod_lshift_quick
13909
13910 These functions always generate non-negative results.
13911
1dc1ea18
DDO
13912 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13913 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
5f8e6c50 13914
1dc1ea18
DDO
13915 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13916 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13917 be reduced modulo `m`.
5f8e6c50
DMSP
13918
13919 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13920
1dc1ea18 13921<!--
5f8e6c50
DMSP
13922 The following entry accidentally appeared in the CHANGES file
13923 distributed with OpenSSL 0.9.7. The modifications described in
13924 it do *not* apply to OpenSSL 0.9.7.
13925
13926 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13927 was actually never needed) and in BN_mul(). The removal in BN_mul()
13928 required a small change in bn_mul_part_recursive() and the addition
13929 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13930 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13931 bn_sub_words() and bn_add_words() except they take arrays with
13932 differing sizes.
13933
13934 *Richard Levitte*
1dc1ea18 13935-->
5f8e6c50
DMSP
13936
13937 * In 'openssl passwd', verify passwords read from the terminal
13938 unless the '-salt' option is used (which usually means that
13939 verification would just waste user's time since the resulting
13940 hash is going to be compared with some given password hash)
13941 or the new '-noverify' option is used.
13942
13943 This is an incompatible change, but it does not affect
13944 non-interactive use of 'openssl passwd' (passwords on the command
13945 line, '-stdin' option, '-in ...' option) and thus should not
13946 cause any problems.
13947
13948 *Bodo Moeller*
13949
13950 * Remove all references to RSAref, since there's no more need for it.
13951
13952 *Richard Levitte*
13953
13954 * Make DSO load along a path given through an environment variable
13955 (SHLIB_PATH) with shl_load().
13956
13957 *Richard Levitte*
13958
13959 * Constify the ENGINE code as a result of BIGNUM constification.
13960 Also constify the RSA code and most things related to it. In a
13961 few places, most notable in the depth of the ASN.1 code, ugly
13962 casts back to non-const were required (to be solved at a later
13963 time)
13964
13965 *Richard Levitte*
13966
13967 * Make it so the openssl application has all engines loaded by default.
13968
13969 *Richard Levitte*
13970
13971 * Constify the BIGNUM routines a little more.
13972
13973 *Richard Levitte*
13974
13975 * Add the following functions:
13976
13977 ENGINE_load_cswift()
13978 ENGINE_load_chil()
13979 ENGINE_load_atalla()
13980 ENGINE_load_nuron()
13981 ENGINE_load_builtin_engines()
13982
13983 That way, an application can itself choose if external engines that
13984 are built-in in OpenSSL shall ever be used or not. The benefit is
13985 that applications won't have to be linked with libdl or other dso
13986 libraries unless it's really needed.
13987
13988 Changed 'openssl engine' to load all engines on demand.
13989 Changed the engine header files to avoid the duplication of some
13990 declarations (they differed!).
13991
13992 *Richard Levitte*
13993
13994 * 'openssl engine' can now list capabilities.
13995
13996 *Richard Levitte*
13997
13998 * Better error reporting in 'openssl engine'.
13999
14000 *Richard Levitte*
14001
14002 * Never call load_dh_param(NULL) in s_server.
14003
14004 *Bodo Moeller*
14005
14006 * Add engine application. It can currently list engines by name and
14007 identity, and test if they are actually available.
14008
14009 *Richard Levitte*
14010
14011 * Improve RPM specification file by forcing symbolic linking and making
14012 sure the installed documentation is also owned by root.root.
14013
14014 *Damien Miller <djm@mindrot.org>*
14015
14016 * Give the OpenSSL applications more possibilities to make use of
14017 keys (public as well as private) handled by engines.
14018
14019 *Richard Levitte*
14020
14021 * Add OCSP code that comes from CertCo.
14022
14023 *Richard Levitte*
14024
14025 * Add VMS support for the Rijndael code.
14026
14027 *Richard Levitte*
14028
14029 * Added untested support for Nuron crypto accelerator.
14030
14031 *Ben Laurie*
14032
14033 * Add support for external cryptographic devices. This code was
14034 previously distributed separately as the "engine" branch.
14035
14036 *Geoff Thorpe, Richard Levitte*
14037
14038 * Rework the filename-translation in the DSO code. It is now possible to
14039 have far greater control over how a "name" is turned into a filename
14040 depending on the operating environment and any oddities about the
14041 different shared library filenames on each system.
14042
14043 *Geoff Thorpe*
14044
14045 * Support threads on FreeBSD-elf in Configure.
14046
14047 *Richard Levitte*
14048
14049 * Fix for SHA1 assembly problem with MASM: it produces
14050 warnings about corrupt line number information when assembling
14051 with debugging information. This is caused by the overlapping
14052 of two sections.
14053
14054 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14055
14056 * NCONF changes.
14057 NCONF_get_number() has no error checking at all. As a replacement,
257e9d03 14058 NCONF_get_number_e() is defined (`_e` for "error checking") and is
5f8e6c50
DMSP
14059 promoted strongly. The old NCONF_get_number is kept around for
14060 binary backward compatibility.
14061 Make it possible for methods to load from something other than a BIO,
14062 by providing a function pointer that is given a name instead of a BIO.
14063 For example, this could be used to load configuration data from an
14064 LDAP server.
14065
14066 *Richard Levitte*
14067
14068 * Fix for non blocking accept BIOs. Added new I/O special reason
14069 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14070 with non blocking I/O was not possible because no retry code was
14071 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14072 this case.
14073
14074 *Steve Henson*
14075
14076 * Added the beginnings of Rijndael support.
14077
14078 *Ben Laurie*
14079
14080 * Fix for bug in DirectoryString mask setting. Add support for
14081 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14082 to allow certificate printing to more controllable, additional
14083 'certopt' option to 'x509' to allow new printing options to be
14084 set.
14085
14086 *Steve Henson*
14087
14088 * Clean old EAY MD5 hack from e_os.h.
14089
14090 *Richard Levitte*
14091
257e9d03 14092### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5f8e6c50
DMSP
14093
14094 * Fix null-pointer assignment in do_change_cipher_spec() revealed
d8dc8538 14095 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
5f8e6c50
DMSP
14096
14097 *Joe Orton, Steve Henson*
14098
257e9d03 14099### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5f8e6c50
DMSP
14100
14101 * Fix additional bug revealed by the NISCC test suite:
14102
14103 Stop bug triggering large recursion when presented with
d8dc8538 14104 certain ASN.1 tags ([CVE-2003-0851])
5f8e6c50
DMSP
14105
14106 *Steve Henson*
14107
257e9d03 14108### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5f8e6c50
DMSP
14109
14110 * Fix various bugs revealed by running the NISCC test suite:
14111
14112 Stop out of bounds reads in the ASN1 code when presented with
14113 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14114
14115 If verify callback ignores invalid public key errors don't try to check
14116 certificate signature with the NULL public key.
14117
5f8e6c50
DMSP
14118 *Steve Henson*
14119
14120 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14121 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14122 specifications.
14123
14124 *Steve Henson*
14125
14126 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14127 extra data after the compression methods not only for TLS 1.0
14128 but also for SSL 3.0 (as required by the specification).
14129
14130 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14131
14132 * Change X509_certificate_type() to mark the key as exported/exportable
14133 when it's 512 *bits* long, not 512 bytes.
14134
14135 *Richard Levitte*
14136
257e9d03 14137### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5f8e6c50
DMSP
14138
14139 * Countermeasure against the Klima-Pokorny-Rosa extension of
14140 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14141 a protocol version number mismatch like a decryption error
14142 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14143
14144 *Bodo Moeller*
14145
14146 * Turn on RSA blinding by default in the default implementation
14147 to avoid a timing attack. Applications that don't want it can call
14148 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14149 They would be ill-advised to do so in most cases.
14150
14151 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14152
14153 * Change RSA blinding code so that it works when the PRNG is not
14154 seeded (in this case, the secret RSA exponent is abused as
14155 an unpredictable seed -- if it is not unpredictable, there
14156 is no point in blinding anyway). Make RSA blinding thread-safe
14157 by remembering the creator's thread ID in rsa->blinding and
14158 having all other threads use local one-time blinding factors
14159 (this requires more computation than sharing rsa->blinding, but
14160 avoids excessive locking; and if an RSA object is not shared
14161 between threads, blinding will still be very fast).
14162
14163 *Bodo Moeller*
14164
257e9d03 14165### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5f8e6c50
DMSP
14166
14167 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14168 via timing by performing a MAC computation even if incorrect
14169 block cipher padding has been found. This is a countermeasure
14170 against active attacks where the attacker has to distinguish
d8dc8538 14171 between bad padding and a MAC verification error. ([CVE-2003-0078])
5f8e6c50
DMSP
14172
14173 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14174 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14175 Martin Vuagnoux (EPFL, Ilion)*
14176
257e9d03 14177### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5f8e6c50
DMSP
14178
14179 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14180 memory from its contents. This is done with a counter that will
14181 place alternating values in each byte. This can be used to solve
14182 two issues: 1) the removal of calls to memset() by highly optimizing
14183 compilers, and 2) cleansing with other values than 0, since those can
14184 be read through on certain media, for example a swap space on disk.
14185
14186 *Geoff Thorpe*
14187
14188 * Bugfix: client side session caching did not work with external caching,
14189 because the session->cipher setting was not restored when reloading
14190 from the external cache. This problem was masked, when
14191 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14192 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14193
14194 *Lutz Jaenicke*
14195
14196 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14197 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14198
14199 *Zeev Lieber <zeev-l@yahoo.com>*
14200
14201 * Undo an undocumented change introduced in 0.9.6e which caused
14202 repeated calls to OpenSSL_add_all_ciphers() and
14203 OpenSSL_add_all_digests() to be ignored, even after calling
14204 EVP_cleanup().
14205
14206 *Richard Levitte*
14207
14208 * Change the default configuration reader to deal with last line not
14209 being properly terminated.
14210
14211 *Richard Levitte*
14212
14213 * Change X509_NAME_cmp() so it applies the special rules on handling
14214 DN values that are of type PrintableString, as well as RDNs of type
14215 emailAddress where the value has the type ia5String.
14216
14217 *stefank@valicert.com via Richard Levitte*
14218
14219 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14220 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14221 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14222 the bitwise-OR of the two for use by the majority of applications
14223 wanting this behaviour, and update the docs. The documented
14224 behaviour and actual behaviour were inconsistent and had been
14225 changing anyway, so this is more a bug-fix than a behavioural
14226 change.
14227
14228 *Geoff Thorpe, diagnosed by Nadav Har'El*
14229
14230 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14231 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14232
14233 *Bodo Moeller*
14234
14235 * Fix initialization code race conditions in
14236 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14237 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14238 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14239 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14240 ssl2_get_cipher_by_char(),
14241 ssl3_get_cipher_by_char().
14242
14243 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14244
14245 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14246 the cached sessions are flushed, as the remove_cb() might use ex_data
14247 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14248 (see [openssl.org #212]).
14249
14250 *Geoff Thorpe, Lutz Jaenicke*
14251
14252 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14253 length, instead of the encoding length to d2i_ASN1_OBJECT.
14254
14255 *Steve Henson*
14256
257e9d03 14257### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5f8e6c50
DMSP
14258
14259 * [In 0.9.6g-engine release:]
257e9d03 14260 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
5f8e6c50
DMSP
14261
14262 *Lynn Gazis <lgazis@rainbow.com>*
14263
257e9d03 14264### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5f8e6c50
DMSP
14265
14266 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14267 and get fix the header length calculation.
14268 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
257e9d03 14269 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
5f8e6c50
DMSP
14270
14271 * Use proper error handling instead of 'assertions' in buffer
14272 overflow checks added in 0.9.6e. This prevents DoS (the
14273 assertions could call abort()).
14274
14275 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14276
257e9d03 14277### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5f8e6c50
DMSP
14278
14279 * Add various sanity checks to asn1_get_length() to reject
14280 the ASN1 length bytes if they exceed sizeof(long), will appear
14281 negative or the content length exceeds the length of the
14282 supplied buffer.
14283
14284 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14285
14286 * Fix cipher selection routines: ciphers without encryption had no flags
14287 for the cipher strength set and where therefore not handled correctly
14288 by the selection routines (PR #130).
14289
14290 *Lutz Jaenicke*
14291
14292 * Fix EVP_dsa_sha macro.
14293
14294 *Nils Larsch*
14295
14296 * New option
14297 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14298 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14299 that was added in OpenSSL 0.9.6d.
14300
14301 As the countermeasure turned out to be incompatible with some
14302 broken SSL implementations, the new option is part of SSL_OP_ALL.
14303 SSL_OP_ALL is usually employed when compatibility with weird SSL
14304 implementations is desired (e.g. '-bugs' option to 's_client' and
14305 's_server'), so the new option is automatically set in many
14306 applications.
14307
14308 *Bodo Moeller*
14309
14310 * Changes in security patch:
14311
14312 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14313 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14314 Air Force Materiel Command, USAF, under agreement number
14315 F30602-01-2-0537.
14316
14317 * Add various sanity checks to asn1_get_length() to reject
14318 the ASN1 length bytes if they exceed sizeof(long), will appear
14319 negative or the content length exceeds the length of the
d8dc8538 14320 supplied buffer. ([CVE-2002-0659])
5f8e6c50
DMSP
14321
14322 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14323
14324 * Assertions for various potential buffer overflows, not known to
14325 happen in practice.
14326
14327 *Ben Laurie (CHATS)*
14328
14329 * Various temporary buffers to hold ASCII versions of integers were
d8dc8538 14330 too small for 64 bit platforms. ([CVE-2002-0655])
257e9d03 14331 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
5f8e6c50
DMSP
14332
14333 * Remote buffer overflow in SSL3 protocol - an attacker could
d8dc8538 14334 supply an oversized session ID to a client. ([CVE-2002-0656])
5f8e6c50 14335
44652c16 14336 *Ben Laurie (CHATS)*
5f8e6c50
DMSP
14337
14338 * Remote buffer overflow in SSL2 protocol - an attacker could
d8dc8538 14339 supply an oversized client master key. ([CVE-2002-0656])
5f8e6c50
DMSP
14340
14341 *Ben Laurie (CHATS)*
14342
257e9d03 14343### Changes between 0.9.6c and 0.9.6d [9 May 2002]
5f8e6c50
DMSP
14344
14345 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14346 encoded as NULL) with id-dsa-with-sha1.
14347
14348 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14349
ec2bfb7d 14350 * Check various `X509_...()` return values in `apps/req.c`.
5f8e6c50
DMSP
14351
14352 *Nils Larsch <nla@trustcenter.de>*
14353
14354 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14355 an end-of-file condition would erroneously be flagged, when the CRLF
14356 was just at the end of a processed block. The bug was discovered when
14357 processing data through a buffering memory BIO handing the data to a
14358 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14359 <ptsekov@syntrex.com> and Nedelcho Stanev.
14360
14361 *Lutz Jaenicke*
14362
14363 * Implement a countermeasure against a vulnerability recently found
14364 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14365 before application data chunks to avoid the use of known IVs
14366 with data potentially chosen by the attacker.
14367
14368 *Bodo Moeller*
14369
14370 * Fix length checks in ssl3_get_client_hello().
14371
14372 *Bodo Moeller*
14373
14374 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14375 to prevent ssl3_read_internal() from incorrectly assuming that
14376 ssl3_read_bytes() found application data while handshake
14377 processing was enabled when in fact s->s3->in_read_app_data was
14378 merely automatically cleared during the initial handshake.
14379
14380 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14381
14382 * Fix object definitions for Private and Enterprise: they were not
14383 recognized in their shortname (=lowercase) representation. Extend
14384 obj_dat.pl to issue an error when using undefined keywords instead
14385 of silently ignoring the problem (Svenning Sorensen
14386 <sss@sss.dnsalias.net>).
14387
14388 *Lutz Jaenicke*
14389
14390 * Fix DH_generate_parameters() so that it works for 'non-standard'
14391 generators, i.e. generators other than 2 and 5. (Previously, the
14392 code did not properly initialise the 'add' and 'rem' values to
14393 BN_generate_prime().)
14394
14395 In the new general case, we do not insist that 'generator' is
14396 actually a primitive root: This requirement is rather pointless;
14397 a generator of the order-q subgroup is just as good, if not
14398 better.
14399
14400 *Bodo Moeller*
14401
14402 * Map new X509 verification errors to alerts. Discovered and submitted by
14403 Tom Wu <tom@arcot.com>.
14404
14405 *Lutz Jaenicke*
14406
14407 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14408 returning non-zero before the data has been completely received
14409 when using non-blocking I/O.
14410
14411 *Bodo Moeller; problem pointed out by John Hughes*
14412
14413 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14414
14415 *Ben Laurie, Lutz Jaenicke*
14416
14417 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14418 Yoram Zahavi <YoramZ@gilian.com>).
14419
14420 *Lutz Jaenicke*
14421
14422 * Add information about CygWin 1.3 and on, and preserve proper
14423 configuration for the versions before that.
14424
14425 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14426
14427 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14428 check whether we deal with a copy of a session and do not delete from
14429 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14430 <izhar@checkpoint.com>.
14431
14432 *Lutz Jaenicke*
14433
14434 * Do not store session data into the internal session cache, if it
14435 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14436 flag is set). Proposed by Aslam <aslam@funk.com>.
14437
14438 *Lutz Jaenicke*
14439
14440 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14441 value is 0.
14442
14443 *Richard Levitte*
14444
14445 * [In 0.9.6d-engine release:]
14446 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14447
14448 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14449
14450 * Add the configuration target linux-s390x.
14451
14452 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14453
14454 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14455 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14456 variable as an indication that a ClientHello message has been
14457 received. As the flag value will be lost between multiple
14458 invocations of ssl3_accept when using non-blocking I/O, the
14459 function may not be aware that a handshake has actually taken
14460 place, thus preventing a new session from being added to the
14461 session cache.
14462
14463 To avoid this problem, we now set s->new_session to 2 instead of
14464 using a local variable.
14465
14466 *Lutz Jaenicke, Bodo Moeller*
14467
14468 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14469 if the SSL_R_LENGTH_MISMATCH error is detected.
14470
14471 *Geoff Thorpe, Bodo Moeller*
14472
14473 * New 'shared_ldflag' column in Configure platform table.
14474
14475 *Richard Levitte*
14476
14477 * Fix EVP_CIPHER_mode macro.
14478
14479 *"Dan S. Camper" <dan@bti.net>*
14480
14481 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14482 type, we must throw them away by setting rr->length to 0.
14483
14484 *D P Chang <dpc@qualys.com>*
14485
257e9d03 14486### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5f8e6c50
DMSP
14487
14488 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14489 <Dominikus.Scherkl@biodata.com>. (The previous implementation
257e9d03
RS
14490 worked incorrectly for those cases where range = `10..._2` and
14491 `3*range` is two bits longer than range.)
5f8e6c50
DMSP
14492
14493 *Bodo Moeller*
14494
14495 * Only add signing time to PKCS7 structures if it is not already
14496 present.
14497
14498 *Steve Henson*
14499
14500 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14501 OBJ_ld_ce should be OBJ_id_ce.
14502 Also some ip-pda OIDs in crypto/objects/objects.txt were
14503 incorrect (cf. RFC 3039).
14504
14505 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14506
14507 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14508 returns early because it has nothing to do.
14509
14510 *Andy Schneider <andy.schneider@bjss.co.uk>*
14511
14512 * [In 0.9.6c-engine release:]
14513 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14514
14515 *Andy Schneider <andy.schneider@bjss.co.uk>*
14516
14517 * [In 0.9.6c-engine release:]
14518 Add support for Cryptographic Appliance's keyserver technology.
14519 (Use engine 'keyclient')
14520
14521 *Cryptographic Appliances and Geoff Thorpe*
14522
14523 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14524 is called via tools/c89.sh because arguments have to be
14525 rearranged (all '-L' options must appear before the first object
14526 modules).
14527
14528 *Richard Shapiro <rshapiro@abinitio.com>*
14529
14530 * [In 0.9.6c-engine release:]
14531 Add support for Broadcom crypto accelerator cards, backported
14532 from 0.9.7.
14533
14534 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14535
14536 * [In 0.9.6c-engine release:]
14537 Add support for SureWare crypto accelerator cards from
14538 Baltimore Technologies. (Use engine 'sureware')
14539
14540 *Baltimore Technologies and Mark Cox*
14541
14542 * [In 0.9.6c-engine release:]
14543 Add support for crypto accelerator cards from Accelerated
14544 Encryption Processing, www.aep.ie. (Use engine 'aep')
14545
14546 *AEP Inc. and Mark Cox*
14547
14548 * Add a configuration entry for gcc on UnixWare.
14549
14550 *Gary Benson <gbenson@redhat.com>*
14551
14552 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14553 messages are stored in a single piece (fixed-length part and
14554 variable-length part combined) and fix various bugs found on the way.
14555
14556 *Bodo Moeller*
14557
14558 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14559 instead. BIO_gethostbyname() does not know what timeouts are
14560 appropriate, so entries would stay in cache even when they have
14561 become invalid.
257e9d03 14562 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
5f8e6c50
DMSP
14563
14564 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14565 faced with a pathologically small ClientHello fragment that does
14566 not contain client_version: Instead of aborting with an error,
14567 simply choose the highest available protocol version (i.e.,
14568 TLS 1.0 unless it is disabled). In practice, ClientHello
14569 messages are never sent like this, but this change gives us
14570 strictly correct behaviour at least for TLS.
14571
44652c16 14572 *Bodo Moeller*
5f8e6c50
DMSP
14573
14574 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14575 never resets s->method to s->ctx->method when called from within
14576 one of the SSL handshake functions.
14577
14578 *Bodo Moeller; problem pointed out by Niko Baric*
14579
14580 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14581 (sent using the client's version number) if client_version is
14582 smaller than the protocol version in use. Also change
14583 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14584 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14585 the client will at least see that alert.
14586
14587 *Bodo Moeller*
14588
14589 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14590 correctly.
14591
14592 *Bodo Moeller*
14593
14594 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14595 client receives HelloRequest while in a handshake.
14596
14597 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14598
14599 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14600 should end in 'break', not 'goto end' which circumvents various
14601 cleanups done in state SSL_ST_OK. But session related stuff
14602 must be disabled for SSL_ST_OK in the case that we just sent a
14603 HelloRequest.
14604
14605 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14606 before just sending a HelloRequest.
14607
14608 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14609
14610 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14611 reveal whether illegal block cipher padding was found or a MAC
14612 verification error occurred. (Neither SSLerr() codes nor alerts
14613 are directly visible to potential attackers, but the information
14614 may leak via logfiles.)
14615
14616 Similar changes are not required for the SSL 2.0 implementation
14617 because the number of padding bytes is sent in clear for SSL 2.0,
14618 and the extra bytes are just ignored. However ssl/s2_pkt.c
14619 failed to verify that the purported number of padding bytes is in
14620 the legal range.
14621
14622 *Bodo Moeller*
14623
14624 * Add OpenUNIX-8 support including shared libraries
14625 (Boyd Lynn Gerber <gerberb@zenez.com>).
14626
14627 *Lutz Jaenicke*
14628
14629 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14630 'wristwatch attack' using huge encoding parameters (cf.
14631 James H. Manger's CRYPTO 2001 paper). Note that the
14632 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14633 encoding parameters and hence was not vulnerable.
14634
14635 *Bodo Moeller*
14636
14637 * BN_sqr() bug fix.
14638
14639 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14640
14641 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14642 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14643 followed by modular reduction.
14644
14645 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14646
14647 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14648 equivalent based on BN_pseudo_rand() instead of BN_rand().
14649
14650 *Bodo Moeller*
14651
14652 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14653 This function was broken, as the check for a new client hello message
14654 to handle SGC did not allow these large messages.
14655 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14656
14657 *Lutz Jaenicke*
14658
257e9d03 14659 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
5f8e6c50
DMSP
14660
14661 *Lutz Jaenicke*
14662
14663 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14664 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14665
14666 *Lutz Jaenicke*
14667
14668 * Rework the configuration and shared library support for Tru64 Unix.
14669 The configuration part makes use of modern compiler features and
14670 still retains old compiler behavior for those that run older versions
14671 of the OS. The shared library support part includes a variant that
14672 uses the RPATH feature, and is available through the special
14673 configuration target "alpha-cc-rpath", which will never be selected
14674 automatically.
14675
14676 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14677
14678 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14679 with the same message size as in ssl3_get_certificate_request().
14680 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14681 messages might inadvertently be reject as too long.
14682
14683 *Petr Lampa <lampa@fee.vutbr.cz>*
14684
14685 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14686
14687 *Andy Polyakov*
14688
14689 * Modified SSL library such that the verify_callback that has been set
44652c16 14690 specifically for an SSL object with SSL_set_verify() is actually being
5f8e6c50
DMSP
14691 used. Before the change, a verify_callback set with this function was
14692 ignored and the verify_callback() set in the SSL_CTX at the time of
14693 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14694 to allow the necessary settings.
14695
14696 *Lutz Jaenicke*
14697
14698 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14699 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14700 done automatically (in contradiction to the requirements of the C
14701 standard). This made problems when used from OpenSSH.
14702
14703 *Lutz Jaenicke*
14704
14705 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14706 dh->length and always used
14707
14708 BN_rand_range(priv_key, dh->p).
14709
14710 BN_rand_range() is not necessary for Diffie-Hellman, and this
14711 specific range makes Diffie-Hellman unnecessarily inefficient if
14712 dh->length (recommended exponent length) is much smaller than the
14713 length of dh->p. We could use BN_rand_range() if the order of
14714 the subgroup was stored in the DH structure, but we only have
14715 dh->length.
14716
14717 So switch back to
14718
14719 BN_rand(priv_key, l, ...)
14720
14721 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14722 otherwise.
14723
14724 *Bodo Moeller*
14725
14726 * In
14727
14728 RSA_eay_public_encrypt
14729 RSA_eay_private_decrypt
14730 RSA_eay_private_encrypt (signing)
14731 RSA_eay_public_decrypt (signature verification)
14732
14733 (default implementations for RSA_public_encrypt,
14734 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14735 always reject numbers >= n.
14736
14737 *Bodo Moeller*
14738
14739 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14740 to synchronize access to 'locking_thread'. This is necessary on
14741 systems where access to 'locking_thread' (an 'unsigned long'
14742 variable) is not atomic.
14743
14744 *Bodo Moeller*
14745
14746 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14747 *before* setting the 'crypto_lock_rand' flag. The previous code had
14748 a race condition if 0 is a valid thread ID.
14749
14750 *Travis Vitek <vitek@roguewave.com>*
14751
14752 * Add support for shared libraries under Irix.
14753
14754 *Albert Chin-A-Young <china@thewrittenword.com>*
14755
14756 * Add configuration option to build on Linux on both big-endian and
14757 little-endian MIPS.
14758
14759 *Ralf Baechle <ralf@uni-koblenz.de>*
14760
14761 * Add the possibility to create shared libraries on HP-UX.
14762
14763 *Richard Levitte*
14764
257e9d03 14765### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5f8e6c50
DMSP
14766
14767 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14768 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14769 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14770 PRNG state recovery was possible based on the output of
14771 one PRNG request appropriately sized to gain knowledge on
14772 'md' followed by enough consecutive 1-byte PRNG requests
14773 to traverse all of 'state'.
14774
14775 1. When updating 'md_local' (the current thread's copy of 'md')
14776 during PRNG output generation, hash all of the previous
14777 'md_local' value, not just the half used for PRNG output.
14778
14779 2. Make the number of bytes from 'state' included into the hash
14780 independent from the number of PRNG bytes requested.
14781
14782 The first measure alone would be sufficient to avoid
14783 Markku-Juhani's attack. (Actually it had never occurred
14784 to me that the half of 'md_local' used for chaining was the
14785 half from which PRNG output bytes were taken -- I had always
14786 assumed that the secret half would be used.) The second
14787 measure makes sure that additional data from 'state' is never
14788 mixed into 'md_local' in small portions; this heuristically
14789 further strengthens the PRNG.
14790
14791 *Bodo Moeller*
14792
14793 * Fix crypto/bn/asm/mips3.s.
14794
14795 *Andy Polyakov*
14796
14797 * When only the key is given to "enc", the IV is undefined. Print out
14798 an error message in this case.
14799
14800 *Lutz Jaenicke*
14801
14802 * Handle special case when X509_NAME is empty in X509 printing routines.
14803
14804 *Steve Henson*
14805
14806 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14807 positive and less than q.
14808
14809 *Bodo Moeller*
14810
257e9d03 14811 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
5f8e6c50
DMSP
14812 used: it isn't thread safe and the add_lock_callback should handle
14813 that itself.
14814
14815 *Paul Rose <Paul.Rose@bridge.com>*
14816
14817 * Verify that incoming data obeys the block size in
14818 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14819
14820 *Bodo Moeller*
14821
14822 * Fix OAEP check.
14823
14824 *Ulf Möller, Bodo Möller*
14825
14826 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14827 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14828 when fixing the server behaviour for backwards-compatible 'client
14829 hello' messages. (Note that the attack is impractical against
14830 SSL 3.0 and TLS 1.0 anyway because length and version checking
14831 means that the probability of guessing a valid ciphertext is
14832 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14833 paper.)
14834
14835 Before 0.9.5, the countermeasure (hide the error by generating a
14836 random 'decryption result') did not work properly because
14837 ERR_clear_error() was missing, meaning that SSL_get_error() would
14838 detect the supposedly ignored error.
14839
14840 Both problems are now fixed.
14841
14842 *Bodo Moeller*
14843
14844 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14845 (previously it was 1024).
14846
14847 *Bodo Moeller*
14848
14849 * Fix for compatibility mode trust settings: ignore trust settings
14850 unless some valid trust or reject settings are present.
14851
14852 *Steve Henson*
14853
14854 * Fix for blowfish EVP: its a variable length cipher.
14855
14856 *Steve Henson*
14857
14858 * Fix various bugs related to DSA S/MIME verification. Handle missing
14859 parameters in DSA public key structures and return an error in the
14860 DSA routines if parameters are absent.
14861
14862 *Steve Henson*
14863
14864 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14865 in the current directory if neither $RANDFILE nor $HOME was set.
14866 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14867 caused some confusion to Windows users who haven't defined $HOME.
14868 Thus RAND_file_name() is changed again: e_os.h can define a
14869 DEFAULT_HOME, which will be used if $HOME is not set.
14870 For Windows, we use "C:"; on other platforms, we still require
14871 environment variables.
14872
14873 * Move 'if (!initialized) RAND_poll()' into regions protected by
14874 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14875 having multiple threads call RAND_poll() concurrently.
14876
14877 *Bodo Moeller*
14878
14879 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14880 combination of a flag and a thread ID variable.
14881 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14882 flag), *other* threads can enter ssleay_add_bytes without obeying
14883 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14884 that they do not hold after the first thread unsets add_do_not_lock).
14885
14886 *Bodo Moeller*
14887
14888 * Change bctest again: '-x' expressions are not available in all
14889 versions of 'test'.
14890
14891 *Bodo Moeller*
14892
257e9d03 14893### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
5f8e6c50
DMSP
14894
14895 * Fix a couple of memory leaks in PKCS7_dataDecode()
14896
14897 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14898
14899 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14900 the default extension for executables, if any. Also, make the perl
14901 scripts that use symlink() to test if it really exists and use "cp"
14902 if it doesn't. All this made OpenSSL compilable and installable in
14903 CygWin.
14904
14905 *Richard Levitte*
14906
14907 * Fix for asn1_GetSequence() for indefinite length constructed data.
14908 If SEQUENCE is length is indefinite just set c->slen to the total
14909 amount of data available.
14910
14911 *Steve Henson, reported by shige@FreeBSD.org*
14912
14913 *This change does not apply to 0.9.7.*
14914
14915 * Change bctest to avoid here-documents inside command substitution
14916 (workaround for FreeBSD /bin/sh bug).
14917 For compatibility with Ultrix, avoid shell functions (introduced
14918 in the bctest version that searches along $PATH).
14919
14920 *Bodo Moeller*
14921
14922 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14923 with des_encrypt() defined on some operating systems, like Solaris
14924 and UnixWare.
14925
14926 *Richard Levitte*
14927
14928 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14929 On the Importance of Eliminating Errors in Cryptographic
14930 Computations, J. Cryptology 14 (2001) 2, 101-119,
257e9d03 14931 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
5f8e6c50
DMSP
14932
14933 *Ulf Moeller*
14934
14935 * MIPS assembler BIGNUM division bug fix.
14936
14937 *Andy Polyakov*
14938
14939 * Disabled incorrect Alpha assembler code.
14940
14941 *Richard Levitte*
14942
14943 * Fix PKCS#7 decode routines so they correctly update the length
14944 after reading an EOC for the EXPLICIT tag.
14945
14946 *Steve Henson*
14947
14948 *This change does not apply to 0.9.7.*
14949
14950 * Fix bug in PKCS#12 key generation routines. This was triggered
14951 if a 3DES key was generated with a 0 initial byte. Include
14952 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14953 (but broken) behaviour.
14954
14955 *Steve Henson*
14956
14957 * Enhance bctest to search for a working bc along $PATH and print
14958 it when found.
14959
14960 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14961
14962 * Fix memory leaks in err.c: free err_data string if necessary;
14963 don't write to the wrong index in ERR_set_error_data.
14964
14965 *Bodo Moeller*
14966
14967 * Implement ssl23_peek (analogous to ssl23_read), which previously
14968 did not exist.
14969
14970 *Bodo Moeller*
14971
257e9d03 14972 * Replace rdtsc with `_emit` statements for VC++ version 5.
5f8e6c50
DMSP
14973
14974 *Jeremy Cooper <jeremy@baymoo.org>*
14975
14976 * Make it possible to reuse SSLv2 sessions.
14977
14978 *Richard Levitte*
14979
14980 * In copy_email() check for >= 0 as a return value for
14981 X509_NAME_get_index_by_NID() since 0 is a valid index.
14982
14983 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14984
14985 * Avoid coredump with unsupported or invalid public keys by checking if
14986 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14987 PKCS7_verify() fails with non detached data.
14988
14989 *Steve Henson*
14990
14991 * Don't use getenv in library functions when run as setuid/setgid.
14992 New function OPENSSL_issetugid().
14993
14994 *Ulf Moeller*
14995
14996 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14997 due to incorrect handling of multi-threading:
14998
14999 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15000
15001 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15002
15003 3. Count how many times MemCheck_off() has been called so that
15004 nested use can be treated correctly. This also avoids
15005 inband-signalling in the previous code (which relied on the
15006 assumption that thread ID 0 is impossible).
15007
15008 *Bodo Moeller*
15009
15010 * Add "-rand" option also to s_client and s_server.
15011
15012 *Lutz Jaenicke*
15013
15014 * Fix CPU detection on Irix 6.x.
15015 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
257e9d03 15016 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
5f8e6c50
DMSP
15017
15018 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15019 was empty.
15020
15021 *Steve Henson*
15022
15023 *This change does not apply to 0.9.7.*
15024
15025 * Use the cached encoding of an X509_NAME structure rather than
15026 copying it. This is apparently the reason for the libsafe "errors"
15027 but the code is actually correct.
15028
15029 *Steve Henson*
15030
15031 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15032 Bleichenbacher's DSA attack.
15033 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15034 to be set and top=0 forces the highest bit to be set; top=-1 is new
15035 and leaves the highest bit random.
15036
15037 *Ulf Moeller, Bodo Moeller*
15038
257e9d03 15039 * In the `NCONF_...`-based implementations for `CONF_...` queries
5f8e6c50
DMSP
15040 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15041 a temporary CONF structure with the data component set to NULL
15042 (which gives segmentation faults in lh_retrieve).
15043 Instead, use NULL for the CONF pointer in CONF_get_string and
15044 CONF_get_number (which may use environment variables) and directly
15045 return NULL from CONF_get_section.
15046
15047 *Bodo Moeller*
15048
15049 * Fix potential buffer overrun for EBCDIC.
15050
15051 *Ulf Moeller*
15052
15053 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15054 keyUsage if basicConstraints absent for a CA.
15055
15056 *Steve Henson*
15057
15058 * Make SMIME_write_PKCS7() write mail header values with a format that
15059 is more generally accepted (no spaces before the semicolon), since
15060 some programs can't parse those values properly otherwise. Also make
15061 sure BIO's that break lines after each write do not create invalid
15062 headers.
15063
15064 *Richard Levitte*
15065
15066 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15067 macros previously used would not encode an empty SEQUENCE OF
15068 and break the signature.
15069
15070 *Steve Henson*
15071
15072 *This change does not apply to 0.9.7.*
15073
15074 * Zero the premaster secret after deriving the master secret in
15075 DH ciphersuites.
15076
15077 *Steve Henson*
15078
15079 * Add some EVP_add_digest_alias registrations (as found in
15080 OpenSSL_add_all_digests()) to SSL_library_init()
15081 aka OpenSSL_add_ssl_algorithms(). This provides improved
15082 compatibility with peers using X.509 certificates
15083 with unconventional AlgorithmIdentifier OIDs.
15084
15085 *Bodo Moeller*
15086
15087 * Fix for Irix with NO_ASM.
15088
15089 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15090
15091 * ./config script fixes.
15092
15093 *Ulf Moeller, Richard Levitte*
15094
15095 * Fix 'openssl passwd -1'.
15096
15097 *Bodo Moeller*
15098
15099 * Change PKCS12_key_gen_asc() so it can cope with non null
15100 terminated strings whose length is passed in the passlen
15101 parameter, for example from PEM callbacks. This was done
15102 by adding an extra length parameter to asc2uni().
15103
15104 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15105
15106 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15107 call failed, free the DSA structure.
15108
15109 *Bodo Moeller*
15110
15111 * Fix to uni2asc() to cope with zero length Unicode strings.
15112 These are present in some PKCS#12 files.
15113
15114 *Steve Henson*
15115
15116 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15117 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15118 when writing a 32767 byte record.
15119
15120 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15121
257e9d03
RS
15122 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15123 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
5f8e6c50
DMSP
15124
15125 (RSA objects have a reference count access to which is protected
15126 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15127 so they are meant to be shared between threads.)
15128 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15129 "Reddie, Steven" <Steven.Reddie@ca.com>*
15130
15131 * Fix a deadlock in CRYPTO_mem_leaks().
15132
15133 *Bodo Moeller*
15134
15135 * Use better test patterns in bntest.
15136
15137 *Ulf Möller*
15138
15139 * rand_win.c fix for Borland C.
15140
15141 *Ulf Möller*
15142
15143 * BN_rshift bugfix for n == 0.
15144
15145 *Bodo Moeller*
15146
15147 * Add a 'bctest' script that checks for some known 'bc' bugs
15148 so that 'make test' does not abort just because 'bc' is broken.
15149
15150 *Bodo Moeller*
15151
15152 * Store verify_result within SSL_SESSION also for client side to
15153 avoid potential security hole. (Re-used sessions on the client side
15154 always resulted in verify_result==X509_V_OK, not using the original
15155 result of the server certificate verification.)
15156
15157 *Lutz Jaenicke*
15158
15159 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15160 SSL3_RT_APPLICATION_DATA, return 0.
15161 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15162
15163 *Bodo Moeller*
15164
15165 * Fix SSL_peek:
15166 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15167 releases, have been re-implemented by renaming the previous
15168 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15169 and ssl3_read_internal, respectively, and adding 'peek' parameters
15170 to them. The new ssl[23]_{read,peek} functions are calls to
15171 ssl[23]_read_internal with the 'peek' flag set appropriately.
15172 A 'peek' parameter has also been added to ssl3_read_bytes, which
15173 does the actual work for ssl3_read_internal.
15174
15175 *Bodo Moeller*
15176
15177 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15178 the method-specific "init()" handler. Also clean up ex_data after
15179 calling the method-specific "finish()" handler. Previously, this was
15180 happening the other way round.
15181
15182 *Geoff Thorpe*
15183
15184 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15185 The previous value, 12, was not always sufficient for BN_mod_exp().
15186
15187 *Bodo Moeller*
15188
15189 * Make sure that shared libraries get the internal name engine with
15190 the full version number and not just 0. This should mark the
15191 shared libraries as not backward compatible. Of course, this should
15192 be changed again when we can guarantee backward binary compatibility.
15193
15194 *Richard Levitte*
15195
15196 * Fix typo in get_cert_by_subject() in by_dir.c
15197
15198 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15199
15200 * Rework the system to generate shared libraries:
15201
15202 - Make note of the expected extension for the shared libraries and
15203 if there is a need for symbolic links from for example libcrypto.so.0
15204 to libcrypto.so.0.9.7. There is extended info in Configure for
15205 that.
15206
15207 - Make as few rebuilds of the shared libraries as possible.
15208
15209 - Still avoid linking the OpenSSL programs with the shared libraries.
15210
15211 - When installing, install the shared libraries separately from the
15212 static ones.
15213
15214 *Richard Levitte*
15215
15216 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15217
15218 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15219 and not in SSL_clear because the latter is also used by the
15220 accept/connect functions; previously, the settings made by
15221 SSL_set_read_ahead would be lost during the handshake.
15222
15223 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15224
15225 * Correct util/mkdef.pl to be selective about disabled algorithms.
15226 Previously, it would create entries for disabled algorithms no
15227 matter what.
15228
15229 *Richard Levitte*
15230
15231 * Added several new manual pages for SSL_* function.
15232
15233 *Lutz Jaenicke*
15234
257e9d03 15235### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
5f8e6c50
DMSP
15236
15237 * In ssl23_get_client_hello, generate an error message when faced
15238 with an initial SSL 3.0/TLS record that is too small to contain the
15239 first two bytes of the ClientHello message, i.e. client_version.
15240 (Note that this is a pathologic case that probably has never happened
15241 in real life.) The previous approach was to use the version number
15242 from the record header as a substitute; but our protocol choice
15243 should not depend on that one because it is not authenticated
15244 by the Finished messages.
15245
15246 *Bodo Moeller*
15247
15248 * More robust randomness gathering functions for Windows.
15249
15250 *Jeffrey Altman <jaltman@columbia.edu>*
15251
15252 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15253 not set then we don't setup the error code for issuer check errors
15254 to avoid possibly overwriting other errors which the callback does
15255 handle. If an application does set the flag then we assume it knows
15256 what it is doing and can handle the new informational codes
15257 appropriately.
15258
15259 *Steve Henson*
15260
15261 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15262 a general "ANY" type, as such it should be able to decode anything
15263 including tagged types. However it didn't check the class so it would
15264 wrongly interpret tagged types in the same way as their universal
15265 counterpart and unknown types were just rejected. Changed so that the
15266 tagged and unknown types are handled in the same way as a SEQUENCE:
15267 that is the encoding is stored intact. There is also a new type
15268 "V_ASN1_OTHER" which is used when the class is not universal, in this
15269 case we have no idea what the actual type is so we just lump them all
15270 together.
15271
15272 *Steve Henson*
15273
15274 * On VMS, stdout may very well lead to a file that is written to
15275 in a record-oriented fashion. That means that every write() will
15276 write a separate record, which will be read separately by the
15277 programs trying to read from it. This can be very confusing.
15278
15279 The solution is to put a BIO filter in the way that will buffer
15280 text until a linefeed is reached, and then write everything a
15281 line at a time, so every record written will be an actual line,
15282 not chunks of lines and not (usually doesn't happen, but I've
15283 seen it once) several lines in one record. BIO_f_linebuffer() is
15284 the answer.
15285
15286 Currently, it's a VMS-only method, because that's where it has
15287 been tested well enough.
15288
15289 *Richard Levitte*
15290
15291 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15292 it can return incorrect results.
15293 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15294 but it was in 0.9.6-beta[12].)
15295
15296 *Bodo Moeller*
15297
15298 * Disable the check for content being present when verifying detached
15299 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15300 include zero length content when signing messages.
15301
15302 *Steve Henson*
15303
15304 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15305 BIO_ctrl (for BIO pairs).
15306
15307 *Bodo Möller*
15308
15309 * Add DSO method for VMS.
15310
15311 *Richard Levitte*
15312
15313 * Bug fix: Montgomery multiplication could produce results with the
15314 wrong sign.
15315
15316 *Ulf Möller*
15317
15318 * Add RPM specification openssl.spec and modify it to build three
15319 packages. The default package contains applications, application
15320 documentation and run-time libraries. The devel package contains
15321 include files, static libraries and function documentation. The
15322 doc package contains the contents of the doc directory. The original
15323 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15324
15325 *Richard Levitte*
15326
15327 * Add a large number of documentation files for many SSL routines.
15328
15329 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15330
15331 * Add a configuration entry for Sony News 4.
15332
15333 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15334
15335 * Don't set the two most significant bits to one when generating a
15336 random number < q in the DSA library.
15337
15338 *Ulf Möller*
15339
15340 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15341 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15342 the underlying transport is blocking) if a handshake took place.
15343 (The default behaviour is needed by applications such as s_client
15344 and s_server that use select() to determine when to use SSL_read;
15345 but for applications that know in advance when to expect data, it
15346 just makes things more complicated.)
15347
15348 *Bodo Moeller*
15349
15350 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15351 from EGD.
15352
15353 *Ben Laurie*
15354
257e9d03 15355 * Add a few more EBCDIC conditionals that make `req` and `x509`
5f8e6c50
DMSP
15356 work better on such systems.
15357
15358 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15359
15360 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15361 Update PKCS12_parse() so it copies the friendlyName and the
15362 keyid to the certificates aux info.
15363
15364 *Steve Henson*
15365
15366 * Fix bug in PKCS7_verify() which caused an infinite loop
15367 if there was more than one signature.
15368
15369 *Sven Uszpelkat <su@celocom.de>*
15370
15371 * Major change in util/mkdef.pl to include extra information
15372 about each symbol, as well as presenting variables as well
15373 as functions. This change means that there's n more need
15374 to rebuild the .num files when some algorithms are excluded.
15375
15376 *Richard Levitte*
15377
15378 * Allow the verify time to be set by an application,
15379 rather than always using the current time.
15380
15381 *Steve Henson*
15382
15383 * Phase 2 verify code reorganisation. The certificate
15384 verify code now looks up an issuer certificate by a
15385 number of criteria: subject name, authority key id
15386 and key usage. It also verifies self signed certificates
15387 by the same criteria. The main comparison function is
15388 X509_check_issued() which performs these checks.
15389
15390 Lot of changes were necessary in order to support this
15391 without completely rewriting the lookup code.
15392
15393 Authority and subject key identifier are now cached.
15394
15395 The LHASH 'certs' is X509_STORE has now been replaced
15396 by a STACK_OF(X509_OBJECT). This is mainly because an
15397 LHASH can't store or retrieve multiple objects with
15398 the same hash value.
15399
15400 As a result various functions (which were all internal
15401 use only) have changed to handle the new X509_STORE
15402 structure. This will break anything that messed round
15403 with X509_STORE internally.
15404
15405 The functions X509_STORE_add_cert() now checks for an
15406 exact match, rather than just subject name.
15407
15408 The X509_STORE API doesn't directly support the retrieval
15409 of multiple certificates matching a given criteria, however
15410 this can be worked round by performing a lookup first
15411 (which will fill the cache with candidate certificates)
15412 and then examining the cache for matches. This is probably
15413 the best we can do without throwing out X509_LOOKUP
15414 entirely (maybe later...).
15415
15416 The X509_VERIFY_CTX structure has been enhanced considerably.
15417
15418 All certificate lookup operations now go via a get_issuer()
15419 callback. Although this currently uses an X509_STORE it
15420 can be replaced by custom lookups. This is a simple way
15421 to bypass the X509_STORE hackery necessary to make this
15422 work and makes it possible to use more efficient techniques
15423 in future. A very simple version which uses a simple
15424 STACK for its trusted certificate store is also provided
15425 using X509_STORE_CTX_trusted_stack().
15426
15427 The verify_cb() and verify() callbacks now have equivalents
15428 in the X509_STORE_CTX structure.
15429
15430 X509_STORE_CTX also has a 'flags' field which can be used
15431 to customise the verify behaviour.
15432
15433 *Steve Henson*
15434
15435 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15436 excludes S/MIME capabilities.
15437
15438 *Steve Henson*
15439
15440 * When a certificate request is read in keep a copy of the
15441 original encoding of the signed data and use it when outputting
15442 again. Signatures then use the original encoding rather than
15443 a decoded, encoded version which may cause problems if the
15444 request is improperly encoded.
15445
15446 *Steve Henson*
15447
15448 * For consistency with other BIO_puts implementations, call
15449 buffer_write(b, ...) directly in buffer_puts instead of calling
15450 BIO_write(b, ...).
15451
15452 In BIO_puts, increment b->num_write as in BIO_write.
15453
15454 *Peter.Sylvester@EdelWeb.fr*
15455
15456 * Fix BN_mul_word for the case where the word is 0. (We have to use
15457 BN_zero, we may not return a BIGNUM with an array consisting of
15458 words set to zero.)
15459
15460 *Bodo Moeller*
15461
15462 * Avoid calling abort() from within the library when problems are
15463 detected, except if preprocessor symbols have been defined
15464 (such as REF_CHECK, BN_DEBUG etc.).
15465
15466 *Bodo Moeller*
15467
15468 * New openssl application 'rsautl'. This utility can be
4d49b685 15469 used for low-level RSA operations. DER public key
5f8e6c50
DMSP
15470 BIO/fp routines also added.
15471
15472 *Steve Henson*
15473
15474 * New Configure entry and patches for compiling on QNX 4.
15475
15476 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15477
15478 * A demo state-machine implementation was sponsored by
257e9d03 15479 Nuron (<http://www.nuron.com/>) and is now available in
5f8e6c50
DMSP
15480 demos/state_machine.
15481
15482 *Ben Laurie*
15483
15484 * New options added to the 'dgst' utility for signature
15485 generation and verification.
15486
15487 *Steve Henson*
15488
15489 * Unrecognized PKCS#7 content types are now handled via a
15490 catch all ASN1_TYPE structure. This allows unsupported
15491 types to be stored as a "blob" and an application can
15492 encode and decode it manually.
15493
15494 *Steve Henson*
15495
15496 * Fix various signed/unsigned issues to make a_strex.c
15497 compile under VC++.
15498
15499 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15500
15501 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15502 length if passed a buffer. ASN1_INTEGER_to_BN failed
15503 if passed a NULL BN and its argument was negative.
15504
15505 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15506
15507 * Modification to PKCS#7 encoding routines to output definite
15508 length encoding. Since currently the whole structures are in
15509 memory there's not real point in using indefinite length
15510 constructed encoding. However if OpenSSL is compiled with
15511 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15512
15513 *Steve Henson*
15514
15515 * Added BIO_vprintf() and BIO_vsnprintf().
15516
15517 *Richard Levitte*
15518
15519 * Added more prefixes to parse for in the strings written
15520 through a logging bio, to cover all the levels that are available
15521 through syslog. The prefixes are now:
15522
15523 PANIC, EMERG, EMR => LOG_EMERG
15524 ALERT, ALR => LOG_ALERT
15525 CRIT, CRI => LOG_CRIT
15526 ERROR, ERR => LOG_ERR
15527 WARNING, WARN, WAR => LOG_WARNING
15528 NOTICE, NOTE, NOT => LOG_NOTICE
15529 INFO, INF => LOG_INFO
15530 DEBUG, DBG => LOG_DEBUG
15531
15532 and as before, if none of those prefixes are present at the
15533 beginning of the string, LOG_ERR is chosen.
15534
257e9d03 15535 On Win32, the `LOG_*` levels are mapped according to this:
5f8e6c50
DMSP
15536
15537 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15538 LOG_WARNING => EVENTLOG_WARNING_TYPE
15539 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15540
5f8e6c50
DMSP
15541 *Richard Levitte*
15542
15543 * Made it possible to reconfigure with just the configuration
15544 argument "reconf" or "reconfigure". The command line arguments
15545 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15546 and are retrieved from there when reconfiguring.
15547
15548 *Richard Levitte*
15549
15550 * MD4 implemented.
15551
15552 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15553
15554 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15555
15556 *Richard Levitte*
15557
15558 * The obj_dat.pl script was messing up the sorting of object
15559 names. The reason was that it compared the quoted version
15560 of strings as a result "OCSP" > "OCSP Signing" because
15561 " > SPACE. Changed script to store unquoted versions of
15562 names and add quotes on output. It was also omitting some
15563 names from the lookup table if they were given a default
15564 value (that is if SN is missing it is given the same
15565 value as LN and vice versa), these are now added on the
15566 grounds that if an object has a name we should be able to
15567 look it up. Finally added warning output when duplicate
15568 short or long names are found.
15569
15570 *Steve Henson*
15571
15572 * Changes needed for Tandem NSK.
15573
15574 *Scott Uroff <scott@xypro.com>*
15575
15576 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15577 RSA_padding_check_SSLv23(), special padding was never detected
15578 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15579 version rollback attacks was not effective.
15580
15581 In s23_clnt.c, don't use special rollback-attack detection padding
15582 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15583 client; similarly, in s23_srvr.c, don't do the rollback check if
15584 SSL 2.0 is the only protocol enabled in the server.
15585
15586 *Bodo Moeller*
15587
15588 * Make it possible to get hexdumps of unprintable data with 'openssl
15589 asn1parse'. By implication, the functions ASN1_parse_dump() and
15590 BIO_dump_indent() are added.
15591
15592 *Richard Levitte*
15593
15594 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15595 these print out strings and name structures based on various
15596 flags including RFC2253 support and proper handling of
15597 multibyte characters. Added options to the 'x509' utility
15598 to allow the various flags to be set.
15599
15600 *Steve Henson*
15601
15602 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15603 Also change the functions X509_cmp_current_time() and
15604 X509_gmtime_adj() work with an ASN1_TIME structure,
15605 this will enable certificates using GeneralizedTime in validity
15606 dates to be checked.
15607
15608 *Steve Henson*
15609
15610 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15611 negative public key encodings) on by default,
15612 NO_NEG_PUBKEY_BUG can be set to disable it.
15613
15614 *Steve Henson*
15615
15616 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15617 content octets. An i2c_ASN1_OBJECT is unnecessary because
15618 the encoding can be trivially obtained from the structure.
15619
15620 *Steve Henson*
15621
257e9d03
RS
15622 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15623 not read locks (`CRYPTO_r_[un]lock`).
5f8e6c50
DMSP
15624
15625 *Bodo Moeller*
15626
15627 * A first attempt at creating official support for shared
15628 libraries through configuration. I've kept it so the
15629 default is static libraries only, and the OpenSSL programs
15630 are always statically linked for now, but there are
15631 preparations for dynamic linking in place.
15632 This has been tested on Linux and Tru64.
15633
15634 *Richard Levitte*
15635
15636 * Randomness polling function for Win9x, as described in:
15637 Peter Gutmann, Software Generation of Practically Strong
15638 Random Numbers.
15639
15640 *Ulf Möller*
15641
15642 * Fix so PRNG is seeded in req if using an already existing
15643 DSA key.
15644
15645 *Steve Henson*
15646
15647 * New options to smime application. -inform and -outform
15648 allow alternative formats for the S/MIME message including
15649 PEM and DER. The -content option allows the content to be
15650 specified separately. This should allow things like Netscape
15651 form signing output easier to verify.
15652
15653 *Steve Henson*
15654
15655 * Fix the ASN1 encoding of tags using the 'long form'.
15656
15657 *Steve Henson*
15658
257e9d03 15659 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
5f8e6c50
DMSP
15660 STRING types. These convert content octets to and from the
15661 underlying type. The actual tag and length octets are
15662 already assumed to have been read in and checked. These
15663 are needed because all other string types have virtually
15664 identical handling apart from the tag. By having versions
15665 of the ASN1 functions that just operate on content octets
15666 IMPLICIT tagging can be handled properly. It also allows
15667 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15668 and ASN1_INTEGER are identical apart from the tag.
15669
15670 *Steve Henson*
15671
15672 * Change the handling of OID objects as follows:
15673
15674 - New object identifiers are inserted in objects.txt, following
1dc1ea18 15675 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
5f8e6c50
DMSP
15676 - objects.pl is used to process obj_mac.num and create a new
15677 obj_mac.h.
15678 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15679 obj_mac.h.
15680
15681 This is currently kind of a hack, and the perl code in objects.pl
15682 isn't very elegant, but it works as I intended. The simplest way
15683 to check that it worked correctly is to look in obj_dat.h and
15684 check the array nid_objs and make sure the objects haven't moved
15685 around (this is important!). Additions are OK, as well as
15686 consistent name changes.
15687
15688 *Richard Levitte*
15689
15690 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15691
15692 *Bodo Moeller*
15693
15694 * Addition of the command line parameter '-rand file' to 'openssl req'.
15695 The given file adds to whatever has already been seeded into the
15696 random pool through the RANDFILE configuration file option or
15697 environment variable, or the default random state file.
15698
15699 *Richard Levitte*
15700
15701 * mkstack.pl now sorts each macro group into lexical order.
15702 Previously the output order depended on the order the files
15703 appeared in the directory, resulting in needless rewriting
15704 of safestack.h .
15705
15706 *Steve Henson*
15707
15708 * Patches to make OpenSSL compile under Win32 again. Mostly
15709 work arounds for the VC++ problem that it treats func() as
15710 func(void). Also stripped out the parts of mkdef.pl that
15711 added extra typesafe functions: these no longer exist.
15712
15713 *Steve Henson*
15714
15715 * Reorganisation of the stack code. The macros are now all
15716 collected in safestack.h . Each macro is defined in terms of
257e9d03 15717 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
5f8e6c50
DMSP
15718 DEBUG_SAFESTACK is now handled in terms of function casts,
15719 this has the advantage of retaining type safety without the
15720 use of additional functions. If DEBUG_SAFESTACK is not defined
15721 then the non typesafe macros are used instead. Also modified the
15722 mkstack.pl script to handle the new form. Needs testing to see
15723 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15724 the default if no major problems. Similar behaviour for ASN1_SET_OF
15725 and PKCS12_STACK_OF.
15726
15727 *Steve Henson*
15728
15729 * When some versions of IIS use the 'NET' form of private key the
15730 key derivation algorithm is different. Normally MD5(password) is
15731 used as a 128 bit RC4 key. In the modified case
15732 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15733 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15734 as the old Netscape_RSA functions except they have an additional
15735 'sgckey' parameter which uses the modified algorithm. Also added
15736 an -sgckey command line option to the rsa utility. Thanks to
15737 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15738 algorithm to openssl-dev.
15739
15740 *Steve Henson*
15741
15742 * The evp_local.h macros were using 'c.##kname' which resulted in
15743 invalid expansion on some systems (SCO 5.0.5 for example).
15744 Corrected to 'c.kname'.
15745
15746 *Phillip Porch <root@theporch.com>*
15747
15748 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15749 a STACK of email addresses from a certificate or request, these look
15750 in the subject name and the subject alternative name extensions and
15751 omit any duplicate addresses.
15752
15753 *Steve Henson*
15754
15755 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15756 This makes DSA verification about 2 % faster.
15757
15758 *Bodo Moeller*
15759
257e9d03 15760 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
5f8e6c50
DMSP
15761 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15762 plus overhead for 1024 bit moduli).
15763 This makes exponentiations about 0.5 % faster for 1024 bit
15764 exponents (as measured by "openssl speed rsa2048").
15765
15766 *Bodo Moeller*
15767
15768 * Rename memory handling macros to avoid conflicts with other
15769 software:
15770 Malloc => OPENSSL_malloc
15771 Malloc_locked => OPENSSL_malloc_locked
15772 Realloc => OPENSSL_realloc
15773 Free => OPENSSL_free
15774
15775 *Richard Levitte*
15776
15777 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15778 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15779
15780 *Bodo Moeller*
15781
15782 * CygWin32 support.
15783
15784 *John Jarvie <jjarvie@newsguy.com>*
15785
15786 * The type-safe stack code has been rejigged. It is now only compiled
15787 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15788 by default all type-specific stack functions are "#define"d back to
15789 standard stack functions. This results in more streamlined output
15790 but retains the type-safety checking possibilities of the original
15791 approach.
15792
15793 *Geoff Thorpe*
15794
15795 * The STACK code has been cleaned up, and certain type declarations
15796 that didn't make a lot of sense have been brought in line. This has
15797 also involved a cleanup of sorts in safestack.h to more correctly
15798 map type-safe stack functions onto their plain stack counterparts.
15799 This work has also resulted in a variety of "const"ifications of
257e9d03 15800 lots of the code, especially `_cmp` operations which should normally
5f8e6c50
DMSP
15801 be prototyped with "const" parameters anyway.
15802
15803 *Geoff Thorpe*
15804
15805 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15806 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15807 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15808 where all of 'md' is used each time the PRNG is used, but 'state'
15809 is used only indexed by a cyclic counter. As entropy may not be
15810 well distributed from the beginning, 'md' is important as a
15811 chaining variable. However, the output function chains only half
15812 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15813 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15814 in all of 'state' being rewritten, with the new values depending
15815 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15816
15817 *Bodo Moeller*
15818
15819 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15820 the handshake is continued after ssl_verify_cert_chain();
15821 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15822 can lead to 'unexplainable' connection aborts later.
15823
15824 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15825
15826 * Major EVP API cipher revision.
15827 Add hooks for extra EVP features. This allows various cipher
15828 parameters to be set in the EVP interface. Support added for variable
15829 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15830 setting of RC2 and RC5 parameters.
15831
15832 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15833 ciphers.
15834
15835 Remove lots of duplicated code from the EVP library. For example *every*
15836 cipher init() function handles the 'iv' in the same way according to the
15837 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15838 for CFB and OFB modes they zero ctx->num.
15839
15840 New functionality allows removal of S/MIME code RC2 hack.
15841
15842 Most of the routines have the same form and so can be declared in terms
15843 of macros.
15844
15845 By shifting this to the top level EVP_CipherInit() it can be removed from
15846 all individual ciphers. If the cipher wants to handle IVs or keys
15847 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15848 flags.
15849
15850 Change lots of functions like EVP_EncryptUpdate() to now return a
15851 value: although software versions of the algorithms cannot fail
15852 any installed hardware versions can.
15853
15854 *Steve Henson*
15855
15856 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15857 this option is set, tolerate broken clients that send the negotiated
15858 protocol version number instead of the requested protocol version
15859 number.
15860
15861 *Bodo Moeller*
15862
257e9d03 15863 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
5f8e6c50
DMSP
15864 i.e. non-zero for export ciphersuites, zero otherwise.
15865 Previous versions had this flag inverted, inconsistent with
15866 rsa_tmp_cb (..._TMP_RSA_CB).
15867
15868 *Bodo Moeller; problem reported by Amit Chopra*
15869
15870 * Add missing DSA library text string. Work around for some IIS
15871 key files with invalid SEQUENCE encoding.
15872
15873 *Steve Henson*
15874
15875 * Add a document (doc/standards.txt) that list all kinds of standards
15876 and so on that are implemented in OpenSSL.
15877
15878 *Richard Levitte*
15879
15880 * Enhance c_rehash script. Old version would mishandle certificates
15881 with the same subject name hash and wouldn't handle CRLs at all.
15882 Added -fingerprint option to crl utility, to support new c_rehash
15883 features.
15884
15885 *Steve Henson*
15886
15887 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15888
15889 *Ulf Möller*
15890
15891 * Fix for SSL server purpose checking. Server checking was
15892 rejecting certificates which had extended key usage present
15893 but no ssl client purpose.
15894
15895 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15896
15897 * Make PKCS#12 code work with no password. The PKCS#12 spec
15898 is a little unclear about how a blank password is handled.
15899 Since the password in encoded as a BMPString with terminating
15900 double NULL a zero length password would end up as just the
15901 double NULL. However no password at all is different and is
15902 handled differently in the PKCS#12 key generation code. NS
15903 treats a blank password as zero length. MSIE treats it as no
15904 password on export: but it will try both on import. We now do
15905 the same: PKCS12_parse() tries zero length and no password if
15906 the password is set to "" or NULL (NULL is now a valid password:
15907 it wasn't before) as does the pkcs12 application.
15908
15909 *Steve Henson*
15910
ec2bfb7d 15911 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
5f8e6c50
DMSP
15912 perror when PEM_read_bio_X509_REQ fails, the error message must
15913 be obtained from the error queue.
15914
15915 *Bodo Moeller*
15916
15917 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15918 it in ERR_remove_state if appropriate, and change ERR_get_state
15919 accordingly to avoid race conditions (this is necessary because
15920 thread_hash is no longer constant once set).
15921
15922 *Bodo Moeller*
15923
15924 * Bugfix for linux-elf makefile.one.
15925
15926 *Ulf Möller*
15927
15928 * RSA_get_default_method() will now cause a default
15929 RSA_METHOD to be chosen if one doesn't exist already.
15930 Previously this was only set during a call to RSA_new()
15931 or RSA_new_method(NULL) meaning it was possible for
15932 RSA_get_default_method() to return NULL.
15933
15934 *Geoff Thorpe*
15935
15936 * Added native name translation to the existing DSO code
15937 that will convert (if the flag to do so is set) filenames
15938 that are sufficiently small and have no path information
15939 into a canonical native form. Eg. "blah" converted to
15940 "libblah.so" or "blah.dll" etc.
15941
15942 *Geoff Thorpe*
15943
15944 * New function ERR_error_string_n(e, buf, len) which is like
15945 ERR_error_string(e, buf), but writes at most 'len' bytes
15946 including the 0 terminator. For ERR_error_string_n, 'buf'
15947 may not be NULL.
15948
15949 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15950
15951 * CONF library reworked to become more general. A new CONF
15952 configuration file reader "class" is implemented as well as a
257e9d03
RS
15953 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15954 old `CONF_*` functions are still there, but are reimplemented to
5f8e6c50
DMSP
15955 work in terms of the new functions. Also, a set of functions
15956 to handle the internal storage of the configuration data is
15957 provided to make it easier to write new configuration file
15958 reader "classes" (I can definitely see something reading a
257e9d03 15959 configuration file in XML format, for example), called `_CONF_*`,
5f8e6c50
DMSP
15960 or "the configuration storage API"...
15961
15962 The new configuration file reading functions are:
15963
15964 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15965 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15966
15967 NCONF_default, NCONF_WIN32
15968
15969 NCONF_dump_fp, NCONF_dump_bio
15970
15971 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15972 NCONF_new creates a new CONF object. This works in the same way
15973 as other interfaces in OpenSSL, like the BIO interface.
257e9d03 15974 `NCONF_dump_*` dump the internal storage of the configuration file,
5f8e6c50 15975 which is useful for debugging. All other functions take the same
257e9d03
RS
15976 arguments as the old `CONF_*` functions with the exception of the
15977 first that must be a `CONF *` instead of a `LHASH *`.
5f8e6c50 15978
257e9d03 15979 To make it easier to use the new classes with the old `CONF_*` functions,
5f8e6c50
DMSP
15980 the function CONF_set_default_method is provided.
15981
15982 *Richard Levitte*
15983
15984 * Add '-tls1' option to 'openssl ciphers', which was already
15985 mentioned in the documentation but had not been implemented.
15986 (This option is not yet really useful because even the additional
15987 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15988
15989 *Bodo Moeller*
15990
15991 * Initial DSO code added into libcrypto for letting OpenSSL (and
15992 OpenSSL-based applications) load shared libraries and bind to
15993 them in a portable way.
15994
15995 *Geoff Thorpe, with contributions from Richard Levitte*
15996
257e9d03 15997### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5f8e6c50
DMSP
15998
15999 * Make sure _lrotl and _lrotr are only used with MSVC.
16000
16001 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16002 (the default implementation of RAND_status).
16003
16004 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16005 to '-clrext' (= clear extensions), as intended and documented.
16006 *Bodo Moeller; inconsistency pointed out by Michael Attili
16007 <attili@amaxo.com>*
16008
16009 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16010 was larger than the MD block size.
16011
16012 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16013
16014 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16015 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16016 using the passed key: if the passed key was a private key the result
16017 of X509_print(), for example, would be to print out all the private key
16018 components.
16019
16020 *Steve Henson*
16021
16022 * des_quad_cksum() byte order bug fix.
16023 *Ulf Möller, using the problem description in krb4-0.9.7, where
257e9d03 16024 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
5f8e6c50
DMSP
16025
16026 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16027 discouraged.
16028
16029 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16030
16031 * For easily testing in shell scripts whether some command
16032 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16033 returns with exit code 0 iff no command of the given name is available.
16034 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16035 the output goes to stdout and nothing is printed to stderr.
16036 Additional arguments are always ignored.
16037
16038 Since for each cipher there is a command of the same name,
16039 the 'no-cipher' compilation switches can be tested this way.
16040
16041 ('openssl no-XXX' is not able to detect pseudo-commands such
16042 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16043
16044 *Bodo Moeller*
16045
16046 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16047
16048 *Bodo Moeller*
16049
16050 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16051 is set; it will be thrown away anyway because each handshake creates
16052 its own key.
16053 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16054 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16055 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16056 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16057
16058 *Bodo Moeller*
16059
16060 * New s_client option -ign_eof: EOF at stdin is ignored, and
16061 'Q' and 'R' lose their special meanings (quit/renegotiate).
16062 This is part of what -quiet does; unlike -quiet, -ign_eof
16063 does not suppress any output.
16064
16065 *Richard Levitte*
16066
16067 * Add compatibility options to the purpose and trust code. The
16068 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16069 accepts a certificate or CA, this was the previous behaviour,
16070 with all the associated security issues.
16071
16072 X509_TRUST_COMPAT is the old trust behaviour: only and
16073 automatically trust self signed roots in certificate store. A
16074 new trust setting X509_TRUST_DEFAULT is used to specify that
16075 a purpose has no associated trust setting and it should instead
16076 use the value in the default purpose.
16077
16078 *Steve Henson*
16079
16080 * Fix the PKCS#8 DSA private key code so it decodes keys again
16081 and fix a memory leak.
16082
16083 *Steve Henson*
16084
16085 * In util/mkerr.pl (which implements 'make errors'), preserve
16086 reason strings from the previous version of the .c file, as
16087 the default to have only downcase letters (and digits) in
16088 automatically generated reasons codes is not always appropriate.
16089
16090 *Bodo Moeller*
16091
16092 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16093 using strerror. Previously, ERR_reason_error_string() returned
16094 library names as reason strings for SYSerr; but SYSerr is a special
16095 case where small numbers are errno values, not library numbers.
16096
16097 *Bodo Moeller*
16098
16099 * Add '-dsaparam' option to 'openssl dhparam' application. This
16100 converts DSA parameters into DH parameters. (When creating parameters,
16101 DSA_generate_parameters is used.)
16102
16103 *Bodo Moeller*
16104
16105 * Include 'length' (recommended exponent length) in C code generated
16106 by 'openssl dhparam -C'.
16107
16108 *Bodo Moeller*
16109
16110 * The second argument to set_label in perlasm was already being used
16111 so couldn't be used as a "file scope" flag. Moved to third argument
16112 which was free.
16113
16114 *Steve Henson*
16115
16116 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16117 instead of RAND_bytes for encryption IVs and salts.
16118
16119 *Bodo Moeller*
16120
16121 * Include RAND_status() into RAND_METHOD instead of implementing
16122 it only for md_rand.c Otherwise replacing the PRNG by calling
16123 RAND_set_rand_method would be impossible.
16124
16125 *Bodo Moeller*
16126
16127 * Don't let DSA_generate_key() enter an infinite loop if the random
16128 number generation fails.
16129
16130 *Bodo Moeller*
16131
16132 * New 'rand' application for creating pseudo-random output.
16133
16134 *Bodo Moeller*
16135
16136 * Added configuration support for Linux/IA64
16137
16138 *Rolf Haberrecker <rolf@suse.de>*
16139
16140 * Assembler module support for Mingw32.
16141
16142 *Ulf Möller*
16143
16144 * Shared library support for HPUX (in shlib/).
16145
16146 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16147
16148 * Shared library support for Solaris gcc.
16149
16150 *Lutz Behnke <behnke@trustcenter.de>*
16151
257e9d03 16152### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5f8e6c50
DMSP
16153
16154 * PKCS7_encrypt() was adding text MIME headers twice because they
16155 were added manually and by SMIME_crlf_copy().
16156
16157 *Steve Henson*
16158
16159 * In bntest.c don't call BN_rand with zero bits argument.
16160
16161 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16162
16163 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16164 case was implemented. This caused BN_div_recp() to fail occasionally.
16165
16166 *Ulf Möller*
16167
16168 * Add an optional second argument to the set_label() in the perl
16169 assembly language builder. If this argument exists and is set
16170 to 1 it signals that the assembler should use a symbol whose
16171 scope is the entire file, not just the current function. This
16172 is needed with MASM which uses the format label:: for this scope.
16173
16174 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16175
16176 * Change the ASN1 types so they are typedefs by default. Before
16177 almost all types were #define'd to ASN1_STRING which was causing
16178 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16179 for example.
16180
16181 *Steve Henson*
16182
16183 * Change names of new functions to the new get1/get0 naming
16184 convention: After 'get1', the caller owns a reference count
257e9d03 16185 and has to call `..._free`; 'get0' returns a pointer to some
5f8e6c50
DMSP
16186 data structure without incrementing reference counters.
16187 (Some of the existing 'get' functions increment a reference
16188 counter, some don't.)
16189 Similarly, 'set1' and 'add1' functions increase reference
16190 counters or duplicate objects.
16191
16192 *Steve Henson*
16193
16194 * Allow for the possibility of temp RSA key generation failure:
16195 the code used to assume it always worked and crashed on failure.
16196
16197 *Steve Henson*
16198
16199 * Fix potential buffer overrun problem in BIO_printf().
16200 *Ulf Möller, using public domain code by Patrick Powell; problem
257e9d03 16201 pointed out by David Sacerdote <das33@cornell.edu>*
5f8e6c50
DMSP
16202
16203 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16204 RAND_egd() and RAND_status(). In the command line application,
16205 the EGD socket can be specified like a seed file using RANDFILE
16206 or -rand.
16207
16208 *Ulf Möller*
16209
16210 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16211 Some CAs (e.g. Verisign) distribute certificates in this form.
16212
16213 *Steve Henson*
16214
16215 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16216 list to exclude them. This means that no special compilation option
16217 is needed to use anonymous DH: it just needs to be included in the
16218 cipher list.
16219
16220 *Steve Henson*
16221
16222 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16223 EVP_MD_type. The old functionality is available in a new macro called
16224 EVP_MD_md(). Change code that uses it and update docs.
16225
16226 *Steve Henson*
16227
257e9d03
RS
16228 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16229 where the `void *` argument is replaced by a function pointer argument.
16230 Previously `void *` was abused to point to functions, which works on
5f8e6c50
DMSP
16231 many platforms, but is not correct. As these functions are usually
16232 called by macros defined in OpenSSL header files, most source code
16233 should work without changes.
16234
16235 *Richard Levitte*
16236
257e9d03 16237 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
5f8e6c50
DMSP
16238 sections with information on -D... compiler switches used for
16239 compiling the library so that applications can see them. To enable
257e9d03 16240 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
5f8e6c50
DMSP
16241 must be defined. E.g.,
16242 #define OPENSSL_ALGORITHM_DEFINES
16243 #include <openssl/opensslconf.h>
257e9d03 16244 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
5f8e6c50
DMSP
16245
16246 *Richard Levitte, Ulf and Bodo Möller*
16247
16248 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16249 record layer.
16250
16251 *Bodo Moeller*
16252
16253 * Change the 'other' type in certificate aux info to a STACK_OF
16254 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16255 the required ASN1 format: arbitrary types determined by an OID.
16256
16257 *Steve Henson*
16258
16259 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16260 argument to 'req'. This is not because the function is newer or
16261 better than others it just uses the work 'NEW' in the certificate
16262 request header lines. Some software needs this.
16263
16264 *Steve Henson*
16265
16266 * Reorganise password command line arguments: now passwords can be
16267 obtained from various sources. Delete the PEM_cb function and make
16268 it the default behaviour: i.e. if the callback is NULL and the
16269 usrdata argument is not NULL interpret it as a null terminated pass
16270 phrase. If usrdata and the callback are NULL then the pass phrase
16271 is prompted for as usual.
16272
16273 *Steve Henson*
16274
16275 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16276 the support is automatically enabled. The resulting binaries will
16277 autodetect the card and use it if present.
16278
16279 *Ben Laurie and Compaq Inc.*
16280
16281 * Work around for Netscape hang bug. This sends certificate request
16282 and server done in one record. Since this is perfectly legal in the
16283 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16284 the bugs/SSLv3 entry for more info.
16285
16286 *Steve Henson*
16287
16288 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16289
16290 *Andy Polyakov*
16291
16292 * Add -rand argument to smime and pkcs12 applications and read/write
16293 of seed file.
16294
16295 *Steve Henson*
16296
16297 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16298
16299 *Bodo Moeller*
16300
16301 * Add command line password options to the remaining applications.
16302
16303 *Steve Henson*
16304
16305 * Bug fix for BN_div_recp() for numerators with an even number of
16306 bits.
16307
16308 *Ulf Möller*
16309
16310 * More tests in bntest.c, and changed test_bn output.
16311
16312 *Ulf Möller*
16313
16314 * ./config recognizes MacOS X now.
16315
16316 *Andy Polyakov*
16317
16318 * Bug fix for BN_div() when the first words of num and divisor are
257e9d03 16319 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
5f8e6c50
DMSP
16320
16321 *Ulf Möller*
16322
16323 * Add support for various broken PKCS#8 formats, and command line
16324 options to produce them.
16325
16326 *Steve Henson*
16327
16328 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16329 get temporary BIGNUMs from a BN_CTX.
16330
16331 *Ulf Möller*
16332
16333 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16334 for p == 0.
16335
16336 *Ulf Möller*
16337
257e9d03 16338 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
5f8e6c50
DMSP
16339 include a #define from the old name to the new. The original intent
16340 was that statically linked binaries could for example just call
16341 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16342 link with digests. This never worked because SSLeay_add_all_digests()
16343 and SSLeay_add_all_ciphers() were in the same source file so calling
16344 one would link with the other. They are now in separate source files.
16345
16346 *Steve Henson*
16347
16348 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16349
16350 *Steve Henson*
16351
16352 * Use a less unusual form of the Miller-Rabin primality test (it used
16353 a binary algorithm for exponentiation integrated into the Miller-Rabin
16354 loop, our standard modexp algorithms are faster).
16355
16356 *Bodo Moeller*
16357
16358 * Support for the EBCDIC character set completed.
16359
16360 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16361
16362 * Source code cleanups: use const where appropriate, eliminate casts,
257e9d03 16363 use `void *` instead of `char *` in lhash.
5f8e6c50
DMSP
16364
16365 *Ulf Möller*
16366
16367 * Bugfix: ssl3_send_server_key_exchange was not restartable
16368 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16369 this the server could overwrite ephemeral keys that the client
16370 has already seen).
16371
16372 *Bodo Moeller*
16373
16374 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16375 using 50 iterations of the Rabin-Miller test.
16376
16377 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16378 iterations of the Rabin-Miller test as required by the appendix
16379 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16380 As BN_is_prime_fasttest includes trial division, DSA parameter
16381 generation becomes much faster.
16382
16383 This implies a change for the callback functions in DSA_is_prime
16384 and DSA_generate_parameters: The callback function is called once
16385 for each positive witness in the Rabin-Miller test, not just
16386 occasionally in the inner loop; and the parameters to the
16387 callback function now provide an iteration count for the outer
16388 loop rather than for the current invocation of the inner loop.
16389 DSA_generate_parameters additionally can call the callback
16390 function with an 'iteration count' of -1, meaning that a
16391 candidate has passed the trial division test (when q is generated
16392 from an application-provided seed, trial division is skipped).
16393
16394 *Bodo Moeller*
16395
16396 * New function BN_is_prime_fasttest that optionally does trial
16397 division before starting the Rabin-Miller test and has
16398 an additional BN_CTX * argument (whereas BN_is_prime always
16399 has to allocate at least one BN_CTX).
16400 'callback(1, -1, cb_arg)' is called when a number has passed the
16401 trial division stage.
16402
16403 *Bodo Moeller*
16404
16405 * Fix for bug in CRL encoding. The validity dates weren't being handled
16406 as ASN1_TIME.
16407
16408 *Steve Henson*
16409
16410 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16411
16412 *Steve Henson*
16413
16414 * New function BN_pseudo_rand().
16415
16416 *Ulf Möller*
16417
16418 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16419 bignum version of BN_from_montgomery() with the working code from
16420 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16421 the comments.
16422
16423 *Ulf Möller*
16424
16425 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16426 made it impossible to use the same SSL_SESSION data structure in
16427 SSL2 clients in multiple threads.
16428
16429 *Bodo Moeller*
16430
16431 * The return value of RAND_load_file() no longer counts bytes obtained
16432 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16433 to seed the PRNG (previously an explicit byte count was required).
16434
16435 *Ulf Möller, Bodo Möller*
16436
16437 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
257e9d03 16438 used `char *` instead of `void *` and had casts all over the place.
5f8e6c50
DMSP
16439
16440 *Steve Henson*
16441
16442 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16443
16444 *Ulf Möller*
16445
16446 * Retain source code compatibility for BN_prime_checks macro:
16447 BN_is_prime(..., BN_prime_checks, ...) now uses
16448 BN_prime_checks_for_size to determine the appropriate number of
16449 Rabin-Miller iterations.
16450
16451 *Ulf Möller*
16452
16453 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16454 DH_CHECK_P_NOT_SAFE_PRIME.
16455 (Check if this is true? OpenPGP calls them "strong".)
16456
16457 *Ulf Möller*
16458
16459 * Merge the functionality of "dh" and "gendh" programs into a new program
16460 "dhparam". The old programs are retained for now but will handle DH keys
16461 (instead of parameters) in future.
16462
16463 *Steve Henson*
16464
16465 * Make the ciphers, s_server and s_client programs check the return values
16466 when a new cipher list is set.
16467
16468 *Steve Henson*
16469
16470 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16471 ciphers. Before when the 56bit ciphers were enabled the sorting was
16472 wrong.
16473
16474 The syntax for the cipher sorting has been extended to support sorting by
16475 cipher-strength (using the strength_bits hard coded in the tables).
ec2bfb7d 16476 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
5f8e6c50
DMSP
16477
16478 Fix a bug in the cipher-command parser: when supplying a cipher command
16479 string with an "undefined" symbol (neither command nor alphanumeric
16480 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16481 an error is flagged.
16482
16483 Due to the strength-sorting extension, the code of the
16484 ssl_create_cipher_list() function was completely rearranged. I hope that
16485 the readability was also increased :-)
16486
16487 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16488
16489 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16490 for the first serial number and places 2 in the serial number file. This
16491 avoids problems when the root CA is created with serial number zero and
16492 the first user certificate has the same issuer name and serial number
16493 as the root CA.
16494
16495 *Steve Henson*
16496
16497 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16498 the new code. Add documentation for this stuff.
16499
16500 *Steve Henson*
16501
16502 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
257e9d03 16503 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
5f8e6c50
DMSP
16504 structures and behave in an analogous way to the X509v3 functions:
16505 they shouldn't be called directly but wrapper functions should be used
16506 instead.
16507
16508 So we also now have some wrapper functions that call the X509at functions
16509 when passed certificate requests. (TO DO: similar things can be done with
16510 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16511 things. Some of these need some d2i or i2d and print functionality
16512 because they handle more complex structures.)
16513
16514 *Steve Henson*
16515
16516 * Add missing #ifndefs that caused missing symbols when building libssl
16517 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
257e9d03 16518 NO_RSA in `ssl/s2*.c`.
5f8e6c50
DMSP
16519
16520 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16521
16522 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16523 has a return value which indicates the quality of the random data
16524 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16525 error queue. New function RAND_pseudo_bytes() generates output that is
16526 guaranteed to be unique but not unpredictable. RAND_add is like
16527 RAND_seed, but takes an extra argument for an entropy estimate
16528 (RAND_seed always assumes full entropy).
16529
16530 *Ulf Möller*
16531
16532 * Do more iterations of Rabin-Miller probable prime test (specifically,
16533 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16534 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16535 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16536 false-positive rate of at most 2^-80 for random input.
16537
16538 *Bodo Moeller*
16539
16540 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16541
16542 *Bodo Moeller*
16543
16544 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16545 in the 0.9.5 release), this returns the chain
16546 from an X509_CTX structure with a dup of the stack and all
16547 the X509 reference counts upped: so the stack will exist
16548 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16549 to use this.
16550
16551 Also make SSL_SESSION_print() print out the verify return
16552 code.
16553
16554 *Steve Henson*
16555
16556 * Add manpage for the pkcs12 command. Also change the default
16557 behaviour so MAC iteration counts are used unless the new
16558 -nomaciter option is used. This improves file security and
16559 only older versions of MSIE (4.0 for example) need it.
16560
16561 *Steve Henson*
16562
16563 * Honor the no-xxx Configure options when creating .DEF files.
16564
16565 *Ulf Möller*
16566
16567 * Add PKCS#10 attributes to field table: challengePassword,
16568 unstructuredName and unstructuredAddress. These are taken from
16569 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16570 international characters are used.
16571
16572 More changes to X509_ATTRIBUTE code: allow the setting of types
16573 based on strings. Remove the 'loc' parameter when adding
16574 attributes because these will be a SET OF encoding which is sorted
16575 in ASN1 order.
16576
16577 *Steve Henson*
16578
16579 * Initial changes to the 'req' utility to allow request generation
16580 automation. This will allow an application to just generate a template
16581 file containing all the field values and have req construct the
16582 request.
16583
16584 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16585 used all over the place including certificate requests and PKCS#7
16586 structures. They are currently handled manually where necessary with
16587 some primitive wrappers for PKCS#7. The new functions behave in a
16588 manner analogous to the X509 extension functions: they allow
16589 attributes to be looked up by NID and added.
16590
16591 Later something similar to the X509V3 code would be desirable to
16592 automatically handle the encoding, decoding and printing of the
16593 more complex types. The string types like challengePassword can
16594 be handled by the string table functions.
16595
16596 Also modified the multi byte string table handling. Now there is
16597 a 'global mask' which masks out certain types. The table itself
16598 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16599 is useful when for example there is only one permissible type
16600 (as in countryName) and using the mask might result in no valid
16601 types at all.
16602
16603 *Steve Henson*
16604
16605 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16606 SSL_get_peer_finished to allow applications to obtain the latest
16607 Finished messages sent to the peer or expected from the peer,
16608 respectively. (SSL_get_peer_finished is usually the Finished message
16609 actually received from the peer, otherwise the protocol will be aborted.)
16610
16611 As the Finished message are message digests of the complete handshake
16612 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16613 be used for external authentication procedures when the authentication
16614 provided by SSL/TLS is not desired or is not enough.
16615
16616 *Bodo Moeller*
16617
16618 * Enhanced support for Alpha Linux is added. Now ./config checks if
16619 the host supports BWX extension and if Compaq C is present on the
16620 $PATH. Just exploiting of the BWX extension results in 20-30%
16621 performance kick for some algorithms, e.g. DES and RC4 to mention
16622 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16623 SHA1.
16624
16625 *Andy Polyakov*
16626
16627 * Add support for MS "fast SGC". This is arguably a violation of the
16628 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16629 weak crypto and after checking the certificate is SGC a second one
16630 with strong crypto. MS SGC stops the first handshake after receiving
16631 the server certificate message and sends a second client hello. Since
16632 a server will typically do all the time consuming operations before
16633 expecting any further messages from the client (server key exchange
16634 is the most expensive) there is little difference between the two.
16635
16636 To get OpenSSL to support MS SGC we have to permit a second client
16637 hello message after we have sent server done. In addition we have to
16638 reset the MAC if we do get this second client hello.
16639
16640 *Steve Henson*
16641
16642 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16643 if a DER encoded private key is RSA or DSA traditional format. Changed
16644 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16645 format DER encoded private key. Newer code should use PKCS#8 format which
16646 has the key type encoded in the ASN1 structure. Added DER private key
16647 support to pkcs8 application.
16648
16649 *Steve Henson*
16650
16651 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16652 ciphersuites has been selected (as required by the SSL 3/TLS 1
16653 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16654 is set, we interpret this as a request to violate the specification
16655 (the worst that can happen is a handshake failure, and 'correct'
16656 behaviour would result in a handshake failure anyway).
16657
16658 *Bodo Moeller*
16659
16660 * In SSL_CTX_add_session, take into account that there might be multiple
16661 SSL_SESSION structures with the same session ID (e.g. when two threads
16662 concurrently obtain them from an external cache).
16663 The internal cache can handle only one SSL_SESSION with a given ID,
16664 so if there's a conflict, we now throw out the old one to achieve
16665 consistency.
16666
16667 *Bodo Moeller*
16668
16669 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16670 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16671 some routines that use cipher OIDs: some ciphers do not have OIDs
16672 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16673 example.
16674
16675 *Steve Henson*
16676
16677 * Simplify the trust setting structure and code. Now we just have
16678 two sequences of OIDs for trusted and rejected settings. These will
16679 typically have values the same as the extended key usage extension
16680 and any application specific purposes.
16681
16682 The trust checking code now has a default behaviour: it will just
16683 check for an object with the same NID as the passed id. Functions can
16684 be provided to override either the default behaviour or the behaviour
16685 for a given id. SSL client, server and email already have functions
16686 in place for compatibility: they check the NID and also return "trusted"
16687 if the certificate is self signed.
16688
16689 *Steve Henson*
16690
16691 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16692 traditional format into an EVP_PKEY structure.
16693
16694 *Steve Henson*
16695
16696 * Add a password callback function PEM_cb() which either prompts for
16697 a password if usr_data is NULL or otherwise assumes it is a null
16698 terminated password. Allow passwords to be passed on command line
16699 environment or config files in a few more utilities.
16700
16701 *Steve Henson*
16702
16703 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16704 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16705 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16706 Update documentation.
16707
16708 *Steve Henson*
16709
16710 * Support for ASN1 "NULL" type. This could be handled before by using
16711 ASN1_TYPE but there wasn't any function that would try to read a NULL
16712 and produce an error if it couldn't. For compatibility we also have
16713 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16714 don't allocate anything because they don't need to.
16715
16716 *Steve Henson*
16717
16718 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16719 for details.
16720
16721 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16722
16723 * Rebuild of the memory allocation routines used by OpenSSL code and
16724 possibly others as well. The purpose is to make an interface that
16725 provide hooks so anyone can build a separate set of allocation and
16726 deallocation routines to be used by OpenSSL, for example memory
16727 pool implementations, or something else, which was previously hard
16728 since Malloc(), Realloc() and Free() were defined as macros having
16729 the values malloc, realloc and free, respectively (except for Win32
16730 compilations). The same is provided for memory debugging code.
16731 OpenSSL already comes with functionality to find memory leaks, but
16732 this gives people a chance to debug other memory problems.
16733
16734 With these changes, a new set of functions and macros have appeared:
16735
16736 CRYPTO_set_mem_debug_functions() [F]
16737 CRYPTO_get_mem_debug_functions() [F]
16738 CRYPTO_dbg_set_options() [F]
16739 CRYPTO_dbg_get_options() [F]
16740 CRYPTO_malloc_debug_init() [M]
16741
16742 The memory debug functions are NULL by default, unless the library
16743 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16744 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16745 gives the standard debugging functions that come with OpenSSL) or
16746 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16747 provided by the library user) must be used. When the standard
16748 debugging functions are used, CRYPTO_dbg_set_options can be used to
16749 request additional information:
16750 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16751 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16752
16753 Also, things like CRYPTO_set_mem_functions will always give the
16754 expected result (the new set of functions is used for allocation
16755 and deallocation) at all times, regardless of platform and compiler
16756 options.
16757
16758 To finish it up, some functions that were never use in any other
16759 way than through macros have a new API and new semantic:
16760
16761 CRYPTO_dbg_malloc()
16762 CRYPTO_dbg_realloc()
16763 CRYPTO_dbg_free()
16764
16765 All macros of value have retained their old syntax.
16766
16767 *Richard Levitte and Bodo Moeller*
16768
16769 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16770 ordering of SMIMECapabilities wasn't in "strength order" and there
16771 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16772 algorithm.
16773
16774 *Steve Henson*
16775
16776 * Some ASN1 types with illegal zero length encoding (INTEGER,
16777 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16778
16779 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16780
16781 * Merge in my S/MIME library for OpenSSL. This provides a simple
16782 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16783 functionality to handle multipart/signed properly) and a utility
16784 called 'smime' to call all this stuff. This is based on code I
16785 originally wrote for Celo who have kindly allowed it to be
16786 included in OpenSSL.
16787
16788 *Steve Henson*
16789
16790 * Add variants des_set_key_checked and des_set_key_unchecked of
16791 des_set_key (aka des_key_sched). Global variable des_check_key
16792 decides which of these is called by des_set_key; this way
16793 des_check_key behaves as it always did, but applications and
16794 the library itself, which was buggy for des_check_key == 1,
16795 have a cleaner way to pick the version they need.
16796
16797 *Bodo Moeller*
16798
16799 * New function PKCS12_newpass() which changes the password of a
16800 PKCS12 structure.
16801
16802 *Steve Henson*
16803
16804 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16805 dynamic mix. In both cases the ids can be used as an index into the
16806 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16807 functions so they accept a list of the field values and the
16808 application doesn't need to directly manipulate the X509_TRUST
16809 structure.
16810
16811 *Steve Henson*
16812
16813 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16814 need initialising.
16815
16816 *Steve Henson*
16817
16818 * Modify the way the V3 extension code looks up extensions. This now
16819 works in a similar way to the object code: we have some "standard"
16820 extensions in a static table which is searched with OBJ_bsearch()
16821 and the application can add dynamic ones if needed. The file
16822 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16823 updated whenever a new extension is added to the core code and kept
16824 in ext_nid order. There is a simple program 'tabtest.c' which checks
16825 this. New extensions are not added too often so this file can readily
16826 be maintained manually.
16827
16828 There are two big advantages in doing things this way. The extensions
16829 can be looked up immediately and no longer need to be "added" using
16830 X509V3_add_standard_extensions(): this function now does nothing.
257e9d03
RS
16831 Side note: I get *lots* of email saying the extension code doesn't
16832 work because people forget to call this function.
5f8e6c50
DMSP
16833 Also no dynamic allocation is done unless new extensions are added:
16834 so if we don't add custom extensions there is no need to call
16835 X509V3_EXT_cleanup().
16836
16837 *Steve Henson*
16838
16839 * Modify enc utility's salting as follows: make salting the default. Add a
16840 magic header, so unsalted files fail gracefully instead of just decrypting
16841 to garbage. This is because not salting is a big security hole, so people
16842 should be discouraged from doing it.
16843
16844 *Ben Laurie*
16845
16846 * Fixes and enhancements to the 'x509' utility. It allowed a message
16847 digest to be passed on the command line but it only used this
16848 parameter when signing a certificate. Modified so all relevant
16849 operations are affected by the digest parameter including the
16850 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16851 DSA key was used because it didn't fix the digest.
16852
16853 *Steve Henson*
16854
16855 * Initial certificate chain verify code. Currently tests the untrusted
16856 certificates for consistency with the verify purpose (which is set
16857 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16858
16859 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16860 this is because it will reject chains with invalid extensions whereas
16861 every previous version of OpenSSL and SSLeay made no checks at all.
16862
16863 Trust code: checks the root CA for the relevant trust settings. Trust
16864 settings have an initial value consistent with the verify purpose: e.g.
16865 if the verify purpose is for SSL client use it expects the CA to be
16866 trusted for SSL client use. However the default value can be changed to
16867 permit custom trust settings: one example of this would be to only trust
16868 certificates from a specific "secure" set of CAs.
16869
16870 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16871 which should be used for version portability: especially since the
16872 verify structure is likely to change more often now.
16873
16874 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16875 to set them. If not set then assume SSL clients will verify SSL servers
16876 and vice versa.
16877
16878 Two new options to the verify program: -untrusted allows a set of
16879 untrusted certificates to be passed in and -purpose which sets the
16880 intended purpose of the certificate. If a purpose is set then the
16881 new chain verify code is used to check extension consistency.
16882
16883 *Steve Henson*
16884
16885 * Support for the authority information access extension.
16886
16887 *Steve Henson*
16888
16889 * Modify RSA and DSA PEM read routines to transparently handle
16890 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16891 public keys in a format compatible with certificate
16892 SubjectPublicKeyInfo structures. Unfortunately there were already
16893 functions called *_PublicKey_* which used various odd formats so
16894 these are retained for compatibility: however the DSA variants were
16895 never in a public release so they have been deleted. Changed dsa/rsa
16896 utilities to handle the new format: note no releases ever handled public
16897 keys so we should be OK.
16898
16899 The primary motivation for this change is to avoid the same fiasco
16900 that dogs private keys: there are several incompatible private key
16901 formats some of which are standard and some OpenSSL specific and
16902 require various evil hacks to allow partial transparent handling and
16903 even then it doesn't work with DER formats. Given the option anything
16904 other than PKCS#8 should be dumped: but the other formats have to
16905 stay in the name of compatibility.
16906
16907 With public keys and the benefit of hindsight one standard format
16908 is used which works with EVP_PKEY, RSA or DSA structures: though
16909 it clearly returns an error if you try to read the wrong kind of key.
16910
16911 Added a -pubkey option to the 'x509' utility to output the public key.
257e9d03
RS
16912 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16913 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16914 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16915 that do the same as the `EVP_PKEY_assign_*()` except they up the
5f8e6c50
DMSP
16916 reference count of the added key (they don't "swallow" the
16917 supplied key).
16918
16919 *Steve Henson*
16920
16921 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16922 CRLs would fail if the file contained no certificates or no CRLs:
16923 added a new function to read in both types and return the number
16924 read: this means that if none are read it will be an error. The
16925 DER versions of the certificate and CRL reader would always fail
16926 because it isn't possible to mix certificates and CRLs in DER format
16927 without choking one or the other routine. Changed this to just read
16928 a certificate: this is the best we can do. Also modified the code
ec2bfb7d 16929 in `apps/verify.c` to take notice of return codes: it was previously
5f8e6c50
DMSP
16930 attempting to read in certificates from NULL pointers and ignoring
16931 any errors: this is one reason why the cert and CRL reader seemed
16932 to work. It doesn't check return codes from the default certificate
16933 routines: these may well fail if the certificates aren't installed.
16934
16935 *Steve Henson*
16936
16937 * Code to support otherName option in GeneralName.
16938
16939 *Steve Henson*
16940
16941 * First update to verify code. Change the verify utility
16942 so it warns if it is passed a self signed certificate:
16943 for consistency with the normal behaviour. X509_verify
16944 has been modified to it will now verify a self signed
16945 certificate if *exactly* the same certificate appears
16946 in the store: it was previously impossible to trust a
16947 single self signed certificate. This means that:
16948 openssl verify ss.pem
16949 now gives a warning about a self signed certificate but
16950 openssl verify -CAfile ss.pem ss.pem
16951 is OK.
16952
16953 *Steve Henson*
16954
16955 * For servers, store verify_result in SSL_SESSION data structure
16956 (and add it to external session representation).
16957 This is needed when client certificate verifications fails,
16958 but an application-provided verification callback (set by
16959 SSL_CTX_set_cert_verify_callback) allows accepting the session
16960 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16961 but returns 1): When the session is reused, we have to set
16962 ssl->verify_result to the appropriate error code to avoid
16963 security holes.
16964
16965 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16966
16967 * Fix a bug in the new PKCS#7 code: it didn't consider the
16968 case in PKCS7_dataInit() where the signed PKCS7 structure
16969 didn't contain any existing data because it was being created.
16970
16971 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16972
16973 * Add a salt to the key derivation routines in enc.c. This
16974 forms the first 8 bytes of the encrypted file. Also add a
16975 -S option to allow a salt to be input on the command line.
16976
16977 *Steve Henson*
16978
16979 * New function X509_cmp(). Oddly enough there wasn't a function
16980 to compare two certificates. We do this by working out the SHA1
16981 hash and comparing that. X509_cmp() will be needed by the trust
16982 code.
16983
16984 *Steve Henson*
16985
16986 * SSL_get1_session() is like SSL_get_session(), but increments
16987 the reference count in the SSL_SESSION returned.
16988
16989 *Geoff Thorpe <geoff@eu.c2.net>*
16990
16991 * Fix for 'req': it was adding a null to request attributes.
16992 Also change the X509_LOOKUP and X509_INFO code to handle
16993 certificate auxiliary information.
16994
16995 *Steve Henson*
16996
16997 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16998 the 'enc' command.
16999
17000 *Steve Henson*
17001
17002 * Add the possibility to add extra information to the memory leak
17003 detecting output, to form tracebacks, showing from where each
17004 allocation was originated: CRYPTO_push_info("constant string") adds
17005 the string plus current file name and line number to a per-thread
17006 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17007 is like calling CYRPTO_pop_info() until the stack is empty.
17008 Also updated memory leak detection code to be multi-thread-safe.
17009
17010 *Richard Levitte*
17011
17012 * Add options -text and -noout to pkcs7 utility and delete the
17013 encryption options which never did anything. Update docs.
17014
17015 *Steve Henson*
17016
17017 * Add options to some of the utilities to allow the pass phrase
17018 to be included on either the command line (not recommended on
17019 OSes like Unix) or read from the environment. Update the
17020 manpages and fix a few bugs.
17021
17022 *Steve Henson*
17023
17024 * Add a few manpages for some of the openssl commands.
17025
17026 *Steve Henson*
17027
17028 * Fix the -revoke option in ca. It was freeing up memory twice,
17029 leaking and not finding already revoked certificates.
17030
17031 *Steve Henson*
17032
17033 * Extensive changes to support certificate auxiliary information.
17034 This involves the use of X509_CERT_AUX structure and X509_AUX
17035 functions. An X509_AUX function such as PEM_read_X509_AUX()
17036 can still read in a certificate file in the usual way but it
17037 will also read in any additional "auxiliary information". By
17038 doing things this way a fair degree of compatibility can be
17039 retained: existing certificates can have this information added
17040 using the new 'x509' options.
17041
17042 Current auxiliary information includes an "alias" and some trust
17043 settings. The trust settings will ultimately be used in enhanced
17044 certificate chain verification routines: currently a certificate
17045 can only be trusted if it is self signed and then it is trusted
17046 for all purposes.
17047
17048 *Steve Henson*
17049
257e9d03 17050 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
5f8e6c50
DMSP
17051 The problem was that one of the replacement routines had not been working
17052 since SSLeay releases. For now the offending routine has been replaced
17053 with non-optimised assembler. Even so, this now gives around 95%
17054 performance improvement for 1024 bit RSA signs.
17055
17056 *Mark Cox*
17057
17058 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17059 handling. Most clients have the effective key size in bits equal to
17060 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17061 A few however don't do this and instead use the size of the decrypted key
17062 to determine the RC2 key length and the AlgorithmIdentifier to determine
17063 the effective key length. In this case the effective key length can still
17064 be 40 bits but the key length can be 168 bits for example. This is fixed
17065 by manually forcing an RC2 key into the EVP_PKEY structure because the
17066 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17067 the key length and effective key length are equal.
17068
17069 *Steve Henson*
17070
17071 * Add a bunch of functions that should simplify the creation of
17072 X509_NAME structures. Now you should be able to do:
17073 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17074 and have it automatically work out the correct field type and fill in
17075 the structures. The more adventurous can try:
17076 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17077 and it will (hopefully) work out the correct multibyte encoding.
17078
17079 *Steve Henson*
17080
17081 * Change the 'req' utility to use the new field handling and multibyte
17082 copy routines. Before the DN field creation was handled in an ad hoc
17083 way in req, ca, and x509 which was rather broken and didn't support
17084 BMPStrings or UTF8Strings. Since some software doesn't implement
17085 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17086 using the dirstring_type option. See the new comment in the default
17087 openssl.cnf for more info.
17088
17089 *Steve Henson*
17090
17091 * Make crypto/rand/md_rand.c more robust:
17092 - Assure unique random numbers after fork().
17093 - Make sure that concurrent threads access the global counter and
17094 md serializably so that we never lose entropy in them
17095 or use exactly the same state in multiple threads.
17096 Access to the large state is not always serializable because
17097 the additional locking could be a performance killer, and
17098 md should be large enough anyway.
17099
17100 *Bodo Moeller*
17101
ec2bfb7d 17102 * New file `apps/app_rand.c` with commonly needed functionality
5f8e6c50
DMSP
17103 for handling the random seed file.
17104
17105 Use the random seed file in some applications that previously did not:
17106 ca,
17107 dsaparam -genkey (which also ignored its '-rand' option),
17108 s_client,
17109 s_server,
17110 x509 (when signing).
17111 Except on systems with /dev/urandom, it is crucial to have a random
17112 seed file at least for key creation, DSA signing, and for DH exchanges;
17113 for RSA signatures we could do without one.
17114
17115 gendh and gendsa (unlike genrsa) used to read only the first byte
17116 of each file listed in the '-rand' option. The function as previously
17117 found in genrsa is now in app_rand.c and is used by all programs
17118 that support '-rand'.
17119
17120 *Bodo Moeller*
17121
17122 * In RAND_write_file, use mode 0600 for creating files;
17123 don't just chmod when it may be too late.
17124
17125 *Bodo Moeller*
17126
17127 * Report an error from X509_STORE_load_locations
17128 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17129
17130 *Bill Perry*
17131
17132 * New function ASN1_mbstring_copy() this copies a string in either
17133 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17134 into an ASN1_STRING type. A mask of permissible types is passed
17135 and it chooses the "minimal" type to use or an error if not type
17136 is suitable.
17137
17138 *Steve Henson*
17139
17140 * Add function equivalents to the various macros in asn1.h. The old
257e9d03
RS
17141 macros are retained with an `M_` prefix. Code inside the library can
17142 use the `M_` macros. External code (including the openssl utility)
5f8e6c50
DMSP
17143 should *NOT* in order to be "shared library friendly".
17144
17145 *Steve Henson*
17146
17147 * Add various functions that can check a certificate's extensions
17148 to see if it usable for various purposes such as SSL client,
17149 server or S/MIME and CAs of these types. This is currently
17150 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17151 verification. Also added a -purpose flag to x509 utility to
17152 print out all the purposes.
17153
17154 *Steve Henson*
17155
17156 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17157 functions.
17158
17159 *Steve Henson*
17160
257e9d03 17161 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
5f8e6c50
DMSP
17162 for, obtain and decode and extension and obtain its critical flag.
17163 This allows all the necessary extension code to be handled in a
17164 single function call.
17165
17166 *Steve Henson*
17167
17168 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17169 platforms. See crypto/rc4/rc4_enc.c for further details.
17170
17171 *Andy Polyakov*
17172
17173 * New -noout option to asn1parse. This causes no output to be produced
17174 its main use is when combined with -strparse and -out to extract data
17175 from a file (which may not be in ASN.1 format).
17176
17177 *Steve Henson*
17178
17179 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17180 when producing the local key id.
17181
17182 *Richard Levitte <levitte@stacken.kth.se>*
17183
17184 * New option -dhparam in s_server. This allows a DH parameter file to be
17185 stated explicitly. If it is not stated then it tries the first server
17186 certificate file. The previous behaviour hard coded the filename
17187 "server.pem".
17188
17189 *Steve Henson*
17190
17191 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17192 a public key to be input or output. For example:
17193 openssl rsa -in key.pem -pubout -out pubkey.pem
17194 Also added necessary DSA public key functions to handle this.
17195
17196 *Steve Henson*
17197
17198 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17199 in the message. This was handled by allowing
17200 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17201
17202 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17203
17204 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17205 to the end of the strings whereas this didn't. This would cause problems
17206 if strings read with d2i_ASN1_bytes() were later modified.
17207
17208 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17209
17210 * Fix for base64 decode bug. When a base64 bio reads only one line of
17211 data and it contains EOF it will end up returning an error. This is
17212 caused by input 46 bytes long. The cause is due to the way base64
17213 BIOs find the start of base64 encoded data. They do this by trying a
17214 trial decode on each line until they find one that works. When they
17215 do a flag is set and it starts again knowing it can pass all the
17216 data directly through the decoder. Unfortunately it doesn't reset
17217 the context it uses. This means that if EOF is reached an attempt
17218 is made to pass two EOFs through the context and this causes the
17219 resulting error. This can also cause other problems as well. As is
17220 usual with these problems it takes *ages* to find and the fix is
17221 trivial: move one line.
17222
257e9d03 17223 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
5f8e6c50
DMSP
17224
17225 * Ugly workaround to get s_client and s_server working under Windows. The
17226 old code wouldn't work because it needed to select() on sockets and the
17227 tty (for keypresses and to see if data could be written). Win32 only
17228 supports select() on sockets so we select() with a 1s timeout on the
17229 sockets and then see if any characters are waiting to be read, if none
17230 are present then we retry, we also assume we can always write data to
17231 the tty. This isn't nice because the code then blocks until we've
17232 received a complete line of data and it is effectively polling the
17233 keyboard at 1s intervals: however it's quite a bit better than not
17234 working at all :-) A dedicated Windows application might handle this
17235 with an event loop for example.
17236
17237 *Steve Henson*
17238
17239 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17240 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17241 will be called when RSA_sign() and RSA_verify() are used. This is useful
17242 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17243 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17244 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17245 This necessitated the support of an extra signature type NID_md5_sha1
17246 for SSL signatures and modifications to the SSL library to use it instead
17247 of calling RSA_public_decrypt() and RSA_private_encrypt().
17248
17249 *Steve Henson*
17250
17251 * Add new -verify -CAfile and -CApath options to the crl program, these
17252 will lookup a CRL issuers certificate and verify the signature in a
17253 similar way to the verify program. Tidy up the crl program so it
17254 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17255 less strict. It will now permit CRL extensions even if it is not
17256 a V2 CRL: this will allow it to tolerate some broken CRLs.
17257
17258 *Steve Henson*
17259
17260 * Initialize all non-automatic variables each time one of the openssl
17261 sub-programs is started (this is necessary as they may be started
17262 multiple times from the "OpenSSL>" prompt).
17263
17264 *Lennart Bang, Bodo Moeller*
17265
17266 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17267 removing all other RSA functionality (this is what NO_RSA does). This
17268 is so (for example) those in the US can disable those operations covered
17269 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17270 key generation.
17271
17272 *Steve Henson*
17273
17274 * Non-copying interface to BIO pairs.
17275 (still largely untested)
17276
17277 *Bodo Moeller*
17278
17279 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17280 ASCII string. This was handled independently in various places before.
17281
17282 *Steve Henson*
17283
17284 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17285 UTF8 strings a character at a time.
17286
17287 *Steve Henson*
17288
17289 * Use client_version from client hello to select the protocol
17290 (s23_srvr.c) and for RSA client key exchange verification
17291 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17292
17293 *Bodo Moeller*
17294
17295 * Add various utility functions to handle SPKACs, these were previously
17296 handled by poking round in the structure internals. Added new function
17297 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17298 print, verify and generate SPKACs. Based on an original idea from
17299 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17300
17301 *Steve Henson*
17302
17303 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17304
17305 *Andy Polyakov*
17306
17307 * Allow the config file extension section to be overwritten on the
17308 command line. Based on an original idea from Massimiliano Pala
17309 <madwolf@comune.modena.it>. The new option is called -extensions
17310 and can be applied to ca, req and x509. Also -reqexts to override
17311 the request extensions in req and -crlexts to override the crl extensions
17312 in ca.
17313
17314 *Steve Henson*
17315
17316 * Add new feature to the SPKAC handling in ca. Now you can include
17317 the same field multiple times by preceding it by "XXXX." for example:
17318 1.OU="Unit name 1"
17319 2.OU="Unit name 2"
17320 this is the same syntax as used in the req config file.
17321
17322 *Steve Henson*
17323
17324 * Allow certificate extensions to be added to certificate requests. These
17325 are specified in a 'req_extensions' option of the req section of the
17326 config file. They can be printed out with the -text option to req but
17327 are otherwise ignored at present.
17328
17329 *Steve Henson*
17330
17331 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17332 data read consists of only the final block it would not decrypted because
17333 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17334 A misplaced 'break' also meant the decrypted final block might not be
17335 copied until the next read.
17336
17337 *Steve Henson*
17338
17339 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17340 a few extra parameters to the DH structure: these will be useful if
17341 for example we want the value of 'q' or implement X9.42 DH.
17342
17343 *Steve Henson*
17344
17345 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17346 provides hooks that allow the default DSA functions or functions on a
17347 "per key" basis to be replaced. This allows hardware acceleration and
17348 hardware key storage to be handled without major modification to the
4d49b685 17349 library. Also added low-level modexp hooks and CRYPTO_EX structure and
5f8e6c50
DMSP
17350 associated functions.
17351
17352 *Steve Henson*
17353
17354 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17355 as "read only": it can't be written to and the buffer it points to will
17356 not be freed. Reading from a read only BIO is much more efficient than
17357 a normal memory BIO. This was added because there are several times when
17358 an area of memory needs to be read from a BIO. The previous method was
17359 to create a memory BIO and write the data to it, this results in two
17360 copies of the data and an O(n^2) reading algorithm. There is a new
17361 function BIO_new_mem_buf() which creates a read only memory BIO from
17362 an area of memory. Also modified the PKCS#7 routines to use read only
17363 memory BIOs.
17364
17365 *Steve Henson*
17366
17367 * Bugfix: ssl23_get_client_hello did not work properly when called in
17368 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17369 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17370 but a retry condition occurred while trying to read the rest.
17371
17372 *Bodo Moeller*
17373
17374 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17375 NID_pkcs7_encrypted by default: this was wrong since this should almost
17376 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17377 the encrypted data type: this is a more sensible place to put it and it
17378 allows the PKCS#12 code to be tidied up that duplicated this
17379 functionality.
17380
17381 *Steve Henson*
17382
17383 * Changed obj_dat.pl script so it takes its input and output files on
17384 the command line. This should avoid shell escape redirection problems
17385 under Win32.
17386
17387 *Steve Henson*
17388
17389 * Initial support for certificate extension requests, these are included
17390 in things like Xenroll certificate requests. Included functions to allow
17391 extensions to be obtained and added.
17392
17393 *Steve Henson*
17394
17395 * -crlf option to s_client and s_server for sending newlines as
17396 CRLF (as required by many protocols).
17397
17398 *Bodo Moeller*
17399
257e9d03 17400### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
5f8e6c50
DMSP
17401
17402 * Install libRSAglue.a when OpenSSL is built with RSAref.
17403
17404 *Ralf S. Engelschall*
17405
257e9d03 17406 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
5f8e6c50
DMSP
17407
17408 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17409
17410 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17411 program.
17412
17413 *Steve Henson*
17414
17415 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17416 DH parameters/keys (q is lost during that conversion, but the resulting
17417 DH parameters contain its length).
17418
17419 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17420 much faster than DH_generate_parameters (which creates parameters
257e9d03 17421 where `p = 2*q + 1`), and also the smaller q makes DH computations
5f8e6c50
DMSP
17422 much more efficient (160-bit exponentiation instead of 1024-bit
17423 exponentiation); so this provides a convenient way to support DHE
17424 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17425 utter importance to use
17426 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17427 or
17428 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17429 when such DH parameters are used, because otherwise small subgroup
17430 attacks may become possible!
17431
17432 *Bodo Moeller*
17433
17434 * Avoid memory leak in i2d_DHparams.
17435
17436 *Bodo Moeller*
17437
17438 * Allow the -k option to be used more than once in the enc program:
17439 this allows the same encrypted message to be read by multiple recipients.
17440
17441 *Steve Henson*
17442
17443 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17444 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17445 it will always use the numerical form of the OID, even if it has a short
17446 or long name.
17447
17448 *Steve Henson*
17449
17450 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17451 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17452 otherwise bn_mod_exp was called. In the case of hardware keys for example
17453 no private key components need be present and it might store extra data
17454 in the RSA structure, which cannot be accessed from bn_mod_exp.
17455 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17456 private key operations.
17457
17458 *Steve Henson*
17459
17460 * Added support for SPARC Linux.
17461
17462 *Andy Polyakov*
17463
17464 * pem_password_cb function type incompatibly changed from
17465 typedef int pem_password_cb(char *buf, int size, int rwflag);
17466 to
17467 ....(char *buf, int size, int rwflag, void *userdata);
17468 so that applications can pass data to their callbacks:
257e9d03 17469 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
5f8e6c50
DMSP
17470 additional void * argument, which is just handed through whenever
17471 the password callback is called.
17472
17473 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17474
17475 New function SSL_CTX_set_default_passwd_cb_userdata.
17476
17477 Compatibility note: As many C implementations push function arguments
17478 onto the stack in reverse order, the new library version is likely to
17479 interoperate with programs that have been compiled with the old
17480 pem_password_cb definition (PEM_whatever takes some data that
17481 happens to be on the stack as its last argument, and the callback
17482 just ignores this garbage); but there is no guarantee whatsoever that
17483 this will work.
17484
17485 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17486 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17487 problems not only on Windows, but also on some Unix platforms.
17488 To avoid problematic command lines, these definitions are now in an
17489 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17490 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17491
17492 *Bodo Moeller*
17493
17494 * MIPS III/IV assembler module is reimplemented.
17495
17496 *Andy Polyakov*
17497
17498 * More DES library cleanups: remove references to srand/rand and
17499 delete an unused file.
17500
17501 *Ulf Möller*
17502
17503 * Add support for the free Netwide assembler (NASM) under Win32,
17504 since not many people have MASM (ml) and it can be hard to obtain.
17505 This is currently experimental but it seems to work OK and pass all
17506 the tests. Check out INSTALL.W32 for info.
17507
17508 *Steve Henson*
17509
17510 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17511 without temporary keys kept an extra copy of the server key,
17512 and connections with temporary keys did not free everything in case
17513 of an error.
17514
17515 *Bodo Moeller*
17516
17517 * New function RSA_check_key and new openssl rsa option -check
17518 for verifying the consistency of RSA keys.
17519
17520 *Ulf Moeller, Bodo Moeller*
17521
17522 * Various changes to make Win32 compile work:
17523 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17524 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17525 comparison" warnings.
257e9d03 17526 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
5f8e6c50
DMSP
17527
17528 *Steve Henson*
17529
17530 * Add a debugging option to PKCS#5 v2 key generation function: when
17531 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17532 derived keys are printed to stderr.
17533
17534 *Steve Henson*
17535
17536 * Copy the flags in ASN1_STRING_dup().
17537
17538 *Roman E. Pavlov <pre@mo.msk.ru>*
17539
17540 * The x509 application mishandled signing requests containing DSA
17541 keys when the signing key was also DSA and the parameters didn't match.
17542
17543 It was supposed to omit the parameters when they matched the signing key:
17544 the verifying software was then supposed to automatically use the CA's
17545 parameters if they were absent from the end user certificate.
17546
17547 Omitting parameters is no longer recommended. The test was also
17548 the wrong way round! This was probably due to unusual behaviour in
17549 EVP_cmp_parameters() which returns 1 if the parameters match.
17550 This meant that parameters were omitted when they *didn't* match and
17551 the certificate was useless. Certificates signed with 'ca' didn't have
17552 this bug.
17553
17554 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17555
17556 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17557 The interface is as follows:
17558 Applications can use
17559 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17560 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17561 "off" is now the default.
17562 The library internally uses
17563 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17564 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17565 to disable memory-checking temporarily.
17566
17567 Some inconsistent states that previously were possible (and were
17568 even the default) are now avoided.
17569
17570 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17571 with each memory chunk allocated; this is occasionally more helpful
17572 than just having a counter.
17573
17574 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17575
17576 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17577 extensions.
17578
17579 *Bodo Moeller*
17580
17581 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17582 which largely parallels "options", but is for changing API behaviour,
17583 whereas "options" are about protocol behaviour.
17584 Initial "mode" flags are:
17585
17586 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17587 a single record has been written.
17588 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17589 retries use the same buffer location.
17590 (But all of the contents must be
17591 copied!)
17592
17593 *Bodo Moeller*
17594
17595 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17596 worked.
17597
17598 * Fix problems with no-hmac etc.
17599
17600 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17601
17602 * New functions RSA_get_default_method(), RSA_set_method() and
17603 RSA_get_method(). These allows replacement of RSA_METHODs without having
17604 to mess around with the internals of an RSA structure.
17605
17606 *Steve Henson*
17607
17608 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17609 Also really enable memory leak checks in openssl.c and in some
17610 test programs.
17611
17612 *Chad C. Mulligan, Bodo Moeller*
17613
17614 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17615 up the length of negative integers. This has now been simplified to just
17616 store the length when it is first determined and use it later, rather
17617 than trying to keep track of where data is copied and updating it to
17618 point to the end.
257e9d03 17619 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
5f8e6c50
DMSP
17620
17621 * Add a new function PKCS7_signatureVerify. This allows the verification
17622 of a PKCS#7 signature but with the signing certificate passed to the
17623 function itself. This contrasts with PKCS7_dataVerify which assumes the
17624 certificate is present in the PKCS#7 structure. This isn't always the
17625 case: certificates can be omitted from a PKCS#7 structure and be
17626 distributed by "out of band" means (such as a certificate database).
17627
17628 *Steve Henson*
17629
257e9d03 17630 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
5f8e6c50
DMSP
17631 function prototypes in pem.h, also change util/mkdef.pl to add the
17632 necessary function names.
17633
17634 *Steve Henson*
17635
17636 * mk1mf.pl (used by Windows builds) did not properly read the
17637 options set by Configure in the top level Makefile, and Configure
17638 was not even able to write more than one option correctly.
17639 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17640
17641 *Bodo Moeller*
17642
17643 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17644 file to be loaded from a BIO or FILE pointer. The BIO version will
17645 for example allow memory BIOs to contain config info.
17646
17647 *Steve Henson*
17648
17649 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17650 Whoever hopes to achieve shared-library compatibility across versions
17651 must use this, not the compile-time macro.
17652 (Exercise 0.9.4: Which is the minimum library version required by
17653 such programs?)
17654 Note: All this applies only to multi-threaded programs, others don't
17655 need locks.
17656
17657 *Bodo Moeller*
17658
17659 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17660 through a BIO pair triggered the default case, i.e.
17661 SSLerr(...,SSL_R_UNKNOWN_STATE).
17662
17663 *Bodo Moeller*
17664
17665 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17666 can use the SSL library even if none of the specific BIOs is
17667 appropriate.
17668
17669 *Bodo Moeller*
17670
17671 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17672 for the encoded length.
17673
17674 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17675
17676 * Add initial documentation of the X509V3 functions.
17677
17678 *Steve Henson*
17679
17680 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17681 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17682 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17683 secure PKCS#8 private key format with a high iteration count.
17684
17685 *Steve Henson*
17686
17687 * Fix determination of Perl interpreter: A perl or perl5
257e9d03 17688 *directory* in $PATH was also accepted as the interpreter.
5f8e6c50
DMSP
17689
17690 *Ralf S. Engelschall*
17691
17692 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17693 wrong with it but it was very old and did things like calling
17694 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17695 unusual formatting.
17696
17697 *Steve Henson*
17698
17699 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17700 to use the new extension code.
17701
17702 *Steve Henson*
17703
17704 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17705 with macros. This should make it easier to change their form, add extra
17706 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17707 constant.
17708
17709 *Steve Henson*
17710
17711 * Add to configuration table a new entry that can specify an alternative
17712 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17713 according to Mark Crispin <MRC@Panda.COM>.
17714
17715 *Bodo Moeller*
17716
5f8e6c50
DMSP
17717 * DES CBC did not update the IV. Weird.
17718
17719 *Ben Laurie*
17720lse
17721 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17722 Changing the behaviour of the former might break existing programs --
17723 where IV updating is needed, des_ncbc_encrypt can be used.
17724ndif
17725
17726 * When bntest is run from "make test" it drives bc to check its
17727 calculations, as well as internally checking them. If an internal check
17728 fails, it needs to cause bc to give a non-zero result or make test carries
17729 on without noticing the failure. Fixed.
17730
17731 *Ben Laurie*
17732
17733 * DES library cleanups.
17734
17735 *Ulf Möller*
17736
17737 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17738 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17739 ciphers. NOTE: although the key derivation function has been verified
17740 against some published test vectors it has not been extensively tested
17741 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17742 of v2.0.
17743
17744 *Steve Henson*
17745
17746 * Instead of "mkdir -p", which is not fully portable, use new
17747 Perl script "util/mkdir-p.pl".
17748
17749 *Bodo Moeller*
17750
17751 * Rewrite the way password based encryption (PBE) is handled. It used to
17752 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17753 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17754 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17755 the 'parameter' field of the AlgorithmIdentifier is passed to the
17756 underlying key generation function so it must do its own ASN1 parsing.
17757 This has also changed the EVP_PBE_CipherInit() function which now has a
17758 'parameter' argument instead of literal salt and iteration count values
17759 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17760
17761 *Steve Henson*
17762
17763 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17764 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17765 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17766 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17767 value was just used as a "magic string" and not used directly its
17768 value doesn't matter.
17769
17770 *Steve Henson*
17771
17772 * Introduce some semblance of const correctness to BN. Shame C doesn't
17773 support mutable.
17774
17775 *Ben Laurie*
17776
17777 * "linux-sparc64" configuration (ultrapenguin).
17778
17779 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17780 "linux-sparc" configuration.
17781
17782 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17783
17784 * config now generates no-xxx options for missing ciphers.
17785
17786 *Ulf Möller*
17787
17788 * Support the EBCDIC character set (work in progress).
17789 File ebcdic.c not yet included because it has a different license.
17790
17791 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17792
17793 * Support BS2000/OSD-POSIX.
17794
17795 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17796
257e9d03 17797 * Make callbacks for key generation use `void *` instead of `char *`.
5f8e6c50
DMSP
17798
17799 *Ben Laurie*
17800
17801 * Make S/MIME samples compile (not yet tested).
17802
17803 *Ben Laurie*
17804
17805 * Additional typesafe stacks.
17806
17807 *Ben Laurie*
17808
17809 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17810
17811 *Bodo Moeller*
17812
257e9d03 17813### Changes between 0.9.3 and 0.9.3a [29 May 1999]
5f8e6c50
DMSP
17814
17815 * New configuration variant "sco5-gcc".
17816
17817 * Updated some demos.
17818
17819 *Sean O Riordain, Wade Scholine*
17820
17821 * Add missing BIO_free at exit of pkcs12 application.
17822
17823 *Wu Zhigang*
17824
17825 * Fix memory leak in conf.c.
17826
17827 *Steve Henson*
17828
17829 * Updates for Win32 to assembler version of MD5.
17830
17831 *Steve Henson*
17832
ec2bfb7d 17833 * Set #! path to perl in `apps/der_chop` to where we found it
5f8e6c50
DMSP
17834 instead of using a fixed path.
17835
17836 *Bodo Moeller*
17837
17838 * SHA library changes for irix64-mips4-cc.
17839
17840 *Andy Polyakov*
17841
17842 * Improvements for VMS support.
17843
17844 *Richard Levitte*
17845
257e9d03 17846### Changes between 0.9.2b and 0.9.3 [24 May 1999]
5f8e6c50
DMSP
17847
17848 * Bignum library bug fix. IRIX 6 passes "make test" now!
17849 This also avoids the problems with SC4.2 and unpatched SC5.
17850
17851 *Andy Polyakov <appro@fy.chalmers.se>*
17852
17853 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17854 These are required because of the typesafe stack would otherwise break
17855 existing code. If old code used a structure member which used to be STACK
17856 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17857 sk_num or sk_value it would produce an error because the num, data members
17858 are not present in STACK_OF. Now it just produces a warning. sk_set
17859 replaces the old method of assigning a value to sk_value
17860 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17861 that does this will no longer work (and should use sk_set instead) but
17862 this could be regarded as a "questionable" behaviour anyway.
17863
17864 *Steve Henson*
17865
17866 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17867 correctly handle encrypted S/MIME data.
17868
17869 *Steve Henson*
17870
17871 * Change type of various DES function arguments from des_cblock
17872 (which means, in function argument declarations, pointer to char)
17873 to des_cblock * (meaning pointer to array with 8 char elements),
17874 which allows the compiler to do more typechecking; it was like
17875 that back in SSLeay, but with lots of ugly casts.
17876
17877 Introduce new type const_des_cblock.
17878
17879 *Bodo Moeller*
17880
17881 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17882 problems: find RecipientInfo structure that matches recipient certificate
17883 and initialise the ASN1 structures properly based on passed cipher.
17884
17885 *Steve Henson*
17886
17887 * Belatedly make the BN tests actually check the results.
17888
17889 *Ben Laurie*
17890
17891 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17892 to and from BNs: it was completely broken. New compilation option
17893 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17894 key elements as negative integers.
17895
17896 *Steve Henson*
17897
17898 * Reorganize and speed up MD5.
17899
17900 *Andy Polyakov <appro@fy.chalmers.se>*
17901
17902 * VMS support.
17903
17904 *Richard Levitte <richard@levitte.org>*
17905
17906 * New option -out to asn1parse to allow the parsed structure to be
17907 output to a file. This is most useful when combined with the -strparse
17908 option to examine the output of things like OCTET STRINGS.
17909
17910 *Steve Henson*
17911
17912 * Make SSL library a little more fool-proof by not requiring any longer
257e9d03
RS
17913 that `SSL_set_{accept,connect}_state` be called before
17914 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
5f8e6c50
DMSP
17915 in many applications because usually everything *appeared* to work as
17916 intended anyway -- now it really works as intended).
17917
17918 *Bodo Moeller*
17919
17920 * Move openssl.cnf out of lib/.
17921
17922 *Ulf Möller*
17923
257e9d03 17924 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
5f8e6c50 17925 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
257e9d03 17926 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
5f8e6c50
DMSP
17927
17928 *Ralf S. Engelschall*
17929
17930 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17931 handle PKCS#7 enveloped data properly.
17932
17933 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17934
17935 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17936 copying pointers. The cert_st handling is changed by this in
17937 various ways (and thus what used to be known as ctx->default_cert
257e9d03 17938 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
5f8e6c50
DMSP
17939 any longer when s->cert does not give us what we need).
17940 ssl_cert_instantiate becomes obsolete by this change.
17941 As soon as we've got the new code right (possibly it already is?),
17942 we have solved a couple of bugs of the earlier code where s->cert
17943 was used as if it could not have been shared with other SSL structures.
17944
17945 Note that using the SSL API in certain dirty ways now will result
17946 in different behaviour than observed with earlier library versions:
257e9d03 17947 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
5f8e6c50
DMSP
17948 does not influence s as it used to.
17949
17950 In order to clean up things more thoroughly, inside SSL_SESSION
17951 we don't use CERT any longer, but a new structure SESS_CERT
17952 that holds per-session data (if available); currently, this is
17953 the peer's certificate chain and, for clients, the server's certificate
17954 and temporary key. CERT holds only those values that can have
17955 meaningful defaults in an SSL_CTX.
17956
17957 *Bodo Moeller*
17958
17959 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17960 from the internal representation. Various PKCS#7 fixes: remove some
17961 evil casts and set the enc_dig_alg field properly based on the signing
17962 key type.
17963
17964 *Steve Henson*
17965
17966 * Allow PKCS#12 password to be set from the command line or the
17967 environment. Let 'ca' get its config file name from the environment
17968 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17969 and 'x509').
17970
17971 *Steve Henson*
17972
17973 * Allow certificate policies extension to use an IA5STRING for the
17974 organization field. This is contrary to the PKIX definition but
17975 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17976 extension option.
17977
17978 *Steve Henson*
17979
17980 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17981 without disallowing inline assembler and the like for non-pedantic builds.
17982
17983 *Ben Laurie*
17984
17985 * Support Borland C++ builder.
17986
17987 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17988
17989 * Support Mingw32.
17990
17991 *Ulf Möller*
17992
17993 * SHA-1 cleanups and performance enhancements.
17994
17995 *Andy Polyakov <appro@fy.chalmers.se>*
17996
17997 * Sparc v8plus assembler for the bignum library.
17998
17999 *Andy Polyakov <appro@fy.chalmers.se>*
18000
18001 * Accept any -xxx and +xxx compiler options in Configure.
18002
18003 *Ulf Möller*
18004
18005 * Update HPUX configuration.
18006
18007 *Anonymous*
18008
257e9d03 18009 * Add missing `sk_<type>_unshift()` function to safestack.h
5f8e6c50
DMSP
18010
18011 *Ralf S. Engelschall*
18012
18013 * New function SSL_CTX_use_certificate_chain_file that sets the
18014 "extra_cert"s in addition to the certificate. (This makes sense
18015 only for "PEM" format files, as chains as a whole are not
18016 DER-encoded.)
18017
18018 *Bodo Moeller*
18019
18020 * Support verify_depth from the SSL API.
18021 x509_vfy.c had what can be considered an off-by-one-error:
18022 Its depth (which was not part of the external interface)
18023 was actually counting the number of certificates in a chain;
18024 now it really counts the depth.
18025
18026 *Bodo Moeller*
18027
18028 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18029 instead of X509err, which often resulted in confusing error
18030 messages since the error codes are not globally unique
18031 (e.g. an alleged error in ssl3_accept when a certificate
18032 didn't match the private key).
18033
18034 * New function SSL_CTX_set_session_id_context that allows to set a default
18035 value (so that you don't need SSL_set_session_id_context for each
18036 connection using the SSL_CTX).
18037
18038 *Bodo Moeller*
18039
18040 * OAEP decoding bug fix.
18041
18042 *Ulf Möller*
18043
18044 * Support INSTALL_PREFIX for package builders, as proposed by
18045 David Harris.
18046
18047 *Bodo Moeller*
18048
18049 * New Configure options "threads" and "no-threads". For systems
18050 where the proper compiler options are known (currently Solaris
18051 and Linux), "threads" is the default.
18052
18053 *Bodo Moeller*
18054
18055 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18056
18057 *Bodo Moeller*
18058
18059 * Install various scripts to $(OPENSSLDIR)/misc, not to
18060 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18061 such as /usr/local/bin.
18062
18063 *Bodo Moeller*
18064
18065 * "make linux-shared" to build shared libraries.
18066
18067 *Niels Poppe <niels@netbox.org>*
18068
257e9d03 18069 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
5f8e6c50
DMSP
18070
18071 *Ulf Möller*
18072
18073 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18074 extension adding in x509 utility.
18075
18076 *Steve Henson*
18077
18078 * Remove NOPROTO sections and error code comments.
18079
18080 *Ulf Möller*
18081
18082 * Partial rewrite of the DEF file generator to now parse the ANSI
18083 prototypes.
18084
18085 *Steve Henson*
18086
18087 * New Configure options --prefix=DIR and --openssldir=DIR.
18088
18089 *Ulf Möller*
18090
18091 * Complete rewrite of the error code script(s). It is all now handled
18092 by one script at the top level which handles error code gathering,
18093 header rewriting and C source file generation. It should be much better
18094 than the old method: it now uses a modified version of Ulf's parser to
18095 read the ANSI prototypes in all header files (thus the old K&R definitions
18096 aren't needed for error creation any more) and do a better job of
44652c16 18097 translating function codes into names. The old 'ASN1 error code embedded
5f8e6c50
DMSP
18098 in a comment' is no longer necessary and it doesn't use .err files which
18099 have now been deleted. Also the error code call doesn't have to appear all
18100 on one line (which resulted in some large lines...).
18101
18102 *Steve Henson*
18103
257e9d03 18104 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
5f8e6c50
DMSP
18105
18106 *Bodo Moeller*
18107
18108 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18109 0 (which usually indicates a closed connection), but continue reading.
18110
18111 *Bodo Moeller*
18112
18113 * Fix some race conditions.
18114
18115 *Bodo Moeller*
18116
18117 * Add support for CRL distribution points extension. Add Certificate
18118 Policies and CRL distribution points documentation.
18119
18120 *Steve Henson*
18121
18122 * Move the autogenerated header file parts to crypto/opensslconf.h.
18123
18124 *Ulf Möller*
18125
18126 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18127 8 of keying material. Merlin has also confirmed interop with this fix
18128 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18129
18130 *Merlin Hughes <merlin@baltimore.ie>*
18131
18132 * Fix lots of warnings.
18133
18134 *Richard Levitte <levitte@stacken.kth.se>*
18135
18136 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18137 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18138
18139 *Richard Levitte <levitte@stacken.kth.se>*
18140
18141 * Fix problems with sizeof(long) == 8.
18142
18143 *Andy Polyakov <appro@fy.chalmers.se>*
18144
18145 * Change functions to ANSI C.
18146
18147 *Ulf Möller*
18148
18149 * Fix typos in error codes.
18150
18151 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18152
18153 * Remove defunct assembler files from Configure.
18154
18155 *Ulf Möller*
18156
18157 * SPARC v8 assembler BIGNUM implementation.
18158
18159 *Andy Polyakov <appro@fy.chalmers.se>*
18160
18161 * Support for Certificate Policies extension: both print and set.
18162 Various additions to support the r2i method this uses.
18163
18164 *Steve Henson*
18165
18166 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18167 return a const string when you are expecting an allocated buffer.
18168
18169 *Ben Laurie*
18170
18171 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18172 types DirectoryString and DisplayText.
18173
18174 *Steve Henson*
18175
18176 * Add code to allow r2i extensions to access the configuration database,
18177 add an LHASH database driver and add several ctx helper functions.
18178
18179 *Steve Henson*
18180
18181 * Fix an evil bug in bn_expand2() which caused various BN functions to
18182 fail when they extended the size of a BIGNUM.
18183
18184 *Steve Henson*
18185
18186 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18187 support typesafe stack.
18188
18189 *Steve Henson*
18190
18191 * Fix typo in SSL_[gs]et_options().
18192
18193 *Nils Frostberg <nils@medcom.se>*
18194
18195 * Delete various functions and files that belonged to the (now obsolete)
18196 old X509V3 handling code.
18197
18198 *Steve Henson*
18199
18200 * New Configure option "rsaref".
18201
18202 *Ulf Möller*
18203
18204 * Don't auto-generate pem.h.
18205
18206 *Bodo Moeller*
18207
18208 * Introduce type-safe ASN.1 SETs.
18209
18210 *Ben Laurie*
18211
18212 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18213
18214 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18215
18216 * Introduce type-safe STACKs. This will almost certainly break lots of code
18217 that links with OpenSSL (well at least cause lots of warnings), but fear
18218 not: the conversion is trivial, and it eliminates loads of evil casts. A
18219 few STACKed things have been converted already. Feel free to convert more.
18220 In the fullness of time, I'll do away with the STACK type altogether.
18221
18222 *Ben Laurie*
18223
257e9d03
RS
18224 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18225 specified in `<certfile>` by updating the entry in the index.txt file.
5f8e6c50
DMSP
18226 This way one no longer has to edit the index.txt file manually for
18227 revoking a certificate. The -revoke option does the gory details now.
18228
18229 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18230
257e9d03
RS
18231 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18232 `-text` option at all and this way the `-noout -text` combination was
18233 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
5f8e6c50
DMSP
18234
18235 *Ralf S. Engelschall*
18236
18237 * Make sure a corresponding plain text error message exists for the
18238 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18239 verify callback function determined that a certificate was revoked.
18240
18241 *Ralf S. Engelschall*
18242
257e9d03 18243 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
5f8e6c50
DMSP
18244 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18245 all available ciphers including rc5, which was forgotten until now.
18246 In order to let the testing shell script know which algorithms
18247 are available, a new (up to now undocumented) command
257e9d03 18248 `openssl list-cipher-commands` is used.
5f8e6c50
DMSP
18249
18250 *Bodo Moeller*
18251
18252 * Bugfix: s_client occasionally would sleep in select() when
18253 it should have checked SSL_pending() first.
18254
18255 *Bodo Moeller*
18256
18257 * New functions DSA_do_sign and DSA_do_verify to provide access to
18258 the raw DSA values prior to ASN.1 encoding.
18259
18260 *Ulf Möller*
18261
18262 * Tweaks to Configure
18263
18264 *Niels Poppe <niels@netbox.org>*
18265
18266 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18267 yet...
18268
18269 *Steve Henson*
18270
18271 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18272
18273 *Ulf Möller*
18274
18275 * New config option to avoid instructions that are illegal on the 80386.
18276 The default code is faster, but requires at least a 486.
18277
18278 *Ulf Möller*
18279
18280 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18281 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18282 same as SSL2_VERSION anyway.
18283
18284 *Bodo Moeller*
18285
18286 * New "-showcerts" option for s_client.
18287
18288 *Bodo Moeller*
18289
18290 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18291 application. Various cleanups and fixes.
18292
18293 *Steve Henson*
18294
18295 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18296 modify error routines to work internally. Add error codes and PBE init
18297 to library startup routines.
18298
18299 *Steve Henson*
18300
18301 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18302 packing functions to asn1 and evp. Changed function names and error
18303 codes along the way.
18304
18305 *Steve Henson*
18306
18307 * PKCS12 integration: and so it begins... First of several patches to
18308 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18309 objects to objects.h
18310
18311 *Steve Henson*
18312
18313 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18314 and display support for Thawte strong extranet extension.
18315
18316 *Steve Henson*
18317
18318 * Add LinuxPPC support.
18319
18320 *Jeff Dubrule <igor@pobox.org>*
18321
18322 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18323 bn_div_words in alpha.s.
18324
18325 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18326
18327 * Make sure the RSA OAEP test is skipped under -DRSAref because
18328 OAEP isn't supported when OpenSSL is built with RSAref.
18329
18330 *Ulf Moeller <ulf@fitug.de>*
18331
18332 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18333 so they no longer are missing under -DNOPROTO.
18334
18335 *Soren S. Jorvang <soren@t.dk>*
18336
257e9d03 18337### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
5f8e6c50
DMSP
18338
18339 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18340 doesn't work when the session is reused. Coming soon!
18341
18342 *Ben Laurie*
18343
18344 * Fix a security hole, that allows sessions to be reused in the wrong
18345 context thus bypassing client cert protection! All software that uses
18346 client certs and session caches in multiple contexts NEEDS PATCHING to
18347 allow session reuse! A fuller solution is in the works.
18348
18349 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18350
18351 * Some more source tree cleanups (removed obsolete files
18352 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18353 permission on "config" script to be executable) and a fix for the INSTALL
18354 document.
18355
18356 *Ulf Moeller <ulf@fitug.de>*
18357
18358 * Remove some legacy and erroneous uses of malloc, free instead of
18359 Malloc, Free.
18360
18361 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18362
18363 * Make rsa_oaep_test return non-zero on error.
18364
18365 *Ulf Moeller <ulf@fitug.de>*
18366
18367 * Add support for native Solaris shared libraries. Configure
18368 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18369 if someone would make that last step automatic.
18370
18371 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18372
18373 * ctx_size was not built with the right compiler during "make links". Fixed.
18374
18375 *Ben Laurie*
18376
18377 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18378 except NULL ciphers". This means the default cipher list will no longer
18379 enable NULL ciphers. They need to be specifically enabled e.g. with
18380 the string "DEFAULT:eNULL".
18381
18382 *Steve Henson*
18383
18384 * Fix to RSA private encryption routines: if p < q then it would
18385 occasionally produce an invalid result. This will only happen with
18386 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18387
18388 *Steve Henson*
18389
18390 * Be less restrictive and allow also `perl util/perlpath.pl
1dc1ea18
DDO
18391 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18392 because this way one can also use an interpreter named `perl5` (which is
5f8e6c50 18393 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
1dc1ea18 18394 installed as `perl`).
5f8e6c50
DMSP
18395
18396 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18397
18398 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18399
18400 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18401
18402 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18403 advapi32.lib to Win32 build and change the pem test comparison
18404 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18405 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18406 and crypto/des/ede_cbcm_enc.c.
18407
18408 *Steve Henson*
18409
18410 * DES quad checksum was broken on big-endian architectures. Fixed.
18411
18412 *Ben Laurie*
18413
18414 * Comment out two functions in bio.h that aren't implemented. Fix up the
18415 Win32 test batch file so it (might) work again. The Win32 test batch file
18416 is horrible: I feel ill....
18417
18418 *Steve Henson*
18419
18420 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18421 in e_os.h. Audit of header files to check ANSI and non ANSI
18422 sections: 10 functions were absent from non ANSI section and not exported
18423 from Windows DLLs. Fixed up libeay.num for new functions.
18424
18425 *Steve Henson*
18426
1dc1ea18 18427 * Make `openssl version` output lines consistent.
5f8e6c50
DMSP
18428
18429 *Ralf S. Engelschall*
18430
18431 * Fix Win32 symbol export lists for BIO functions: Added
18432 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18433 to ms/libeay{16,32}.def.
18434
18435 *Ralf S. Engelschall*
18436
18437 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18438 fine under Unix and passes some trivial tests I've now added. But the
18439 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18440 added to make sure no one expects that this stuff really works in the
18441 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18442 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18443 openssl_bio.xs.
18444
18445 *Ralf S. Engelschall*
18446
18447 * Fix the generation of two part addresses in perl.
18448
18449 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18450
18451 * Add config entry for Linux on MIPS.
18452
18453 *John Tobey <jtobey@channel1.com>*
18454
18455 * Make links whenever Configure is run, unless we are on Windoze.
18456
18457 *Ben Laurie*
18458
18459 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18460 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18461 in CRLs.
18462
18463 *Steve Henson*
18464
18465 * Add a useful kludge to allow package maintainers to specify compiler and
18466 other platforms details on the command line without having to patch the
257e9d03
RS
18467 Configure script every time: One now can use
18468 `perl Configure <id>:<details>`,
18469 i.e. platform ids are allowed to have details appended
5f8e6c50 18470 to them (separated by colons). This is treated as there would be a static
257e9d03
RS
18471 pre-configured entry in Configure's %table under key `<id>` with value
18472 `<details>` and `perl Configure <id>` is called. So, when you want to
5f8e6c50 18473 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
257e9d03 18474 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
5f8e6c50
DMSP
18475 now, which overrides the FreeBSD-elf entry on-the-fly.
18476
18477 *Ralf S. Engelschall*
18478
18479 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18480
18481 *Ben Laurie*
18482
18483 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
1dc1ea18 18484 on the `perl Configure ...` command line. This way one can compile
5f8e6c50
DMSP
18485 OpenSSL libraries with Position Independent Code (PIC) which is needed
18486 for linking it into DSOs.
18487
18488 *Ralf S. Engelschall*
18489
18490 * Remarkably, export ciphers were totally broken and no-one had noticed!
18491 Fixed.
18492
18493 *Ben Laurie*
18494
18495 * Cleaned up the LICENSE document: The official contact for any license
18496 questions now is the OpenSSL core team under openssl-core@openssl.org.
18497 And add a paragraph about the dual-license situation to make sure people
18498 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18499 to the OpenSSL toolkit.
18500
18501 *Ralf S. Engelschall*
18502
1dc1ea18
DDO
18503 * General source tree makefile cleanups: Made `making xxx in yyy...`
18504 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18505 Additionally cleaned up the `make links` target: Remove unnecessary
5f8e6c50
DMSP
18506 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18507 to speed processing and no longer clutter the display with confusing
18508 stuff. Instead only the actually done links are displayed.
18509
18510 *Ralf S. Engelschall*
18511
18512 * Permit null encryption ciphersuites, used for authentication only. It used
18513 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18514 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18515 encryption.
18516
18517 *Ben Laurie*
18518
18519 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18520 signed attributes when verifying signatures (this would break them),
18521 the detached data encoding was wrong and public keys obtained using
18522 X509_get_pubkey() weren't freed.
18523
18524 *Steve Henson*
18525
18526 * Add text documentation for the BUFFER functions. Also added a work around
18527 to a Win95 console bug. This was triggered by the password read stuff: the
18528 last character typed gets carried over to the next fread(). If you were
18529 generating a new cert request using 'req' for example then the last
18530 character of the passphrase would be CR which would then enter the first
18531 field as blank.
18532
18533 *Steve Henson*
18534
257e9d03 18535 * Added the new 'Includes OpenSSL Cryptography Software' button as
5f8e6c50
DMSP
18536 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18537 button and can be used by applications based on OpenSSL to show the
18538 relationship to the OpenSSL project.
18539
18540 *Ralf S. Engelschall*
18541
18542 * Remove confusing variables in function signatures in files
18543 ssl/ssl_lib.c and ssl/ssl.h.
18544
18545 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18546
18547 * Don't install bss_file.c under PREFIX/include/
18548
18549 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18550
18551 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18552 functions that return function pointers and has support for NT specific
18553 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18554 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18555 unsigned to signed types: this was killing the Win32 compile.
18556
18557 *Steve Henson*
18558
18559 * Add new certificate file to stack functions,
18560 SSL_add_dir_cert_subjects_to_stack() and
18561 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18562 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18563 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18564 This means that Apache-SSL and similar packages don't have to mess around
18565 to add as many CAs as they want to the preferred list.
18566
18567 *Ben Laurie*
18568
18569 * Experiment with doxygen documentation. Currently only partially applied to
18570 ssl/ssl_lib.c.
257e9d03 18571 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
5f8e6c50
DMSP
18572 openssl.doxy as the configuration file.
18573
18574 *Ben Laurie*
18575
18576 * Get rid of remaining C++-style comments which strict C compilers hate.
18577
18578 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18579
18580 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18581 compiled in by default: it has problems with large keys.
18582
18583 *Steve Henson*
18584
18585 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18586 DH private keys and/or callback functions which directly correspond to
18587 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18588 is needed for applications which have to configure certificates on a
18589 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18590 (e.g. s_server).
18591 For the RSA certificate situation is makes no difference, but
18592 for the DSA certificate situation this fixes the "no shared cipher"
18593 problem where the OpenSSL cipher selection procedure failed because the
18594 temporary keys were not overtaken from the context and the API provided
18595 no way to reconfigure them.
18596 The new functions now let applications reconfigure the stuff and they
18597 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18598 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18599 non-public-API function ssl_cert_instantiate() is used as a helper
18600 function and also to reduce code redundancy inside ssl_rsa.c.
18601
18602 *Ralf S. Engelschall*
18603
18604 * Move s_server -dcert and -dkey options out of the undocumented feature
18605 area because they are useful for the DSA situation and should be
18606 recognized by the users.
18607
18608 *Ralf S. Engelschall*
18609
18610 * Fix the cipher decision scheme for export ciphers: the export bits are
18611 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18612 SSL_EXP_MASK. So, the original variable has to be used instead of the
18613 already masked variable.
18614
18615 *Richard Levitte <levitte@stacken.kth.se>*
18616
257e9d03 18617 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
5f8e6c50
DMSP
18618
18619 *Richard Levitte <levitte@stacken.kth.se>*
18620
18621 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
257e9d03
RS
18622 from `int` to `unsigned int` because it is a length and initialized by
18623 EVP_DigestFinal() which expects an `unsigned int *`.
5f8e6c50
DMSP
18624
18625 *Richard Levitte <levitte@stacken.kth.se>*
18626
18627 * Don't hard-code path to Perl interpreter on shebang line of Configure
18628 script. Instead use the usual Shell->Perl transition trick.
18629
18630 *Ralf S. Engelschall*
18631
1dc1ea18 18632 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
5f8e6c50 18633 (in addition to RSA certificates) to match the behaviour of `openssl dsa
1dc1ea18
DDO
18634 -noout -modulus` as it's already the case for `openssl rsa -noout
18635 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
5f8e6c50 18636 currently the public key is printed (a decision which was already done by
1dc1ea18 18637 `openssl dsa -modulus` in the past) which serves a similar purpose.
5f8e6c50
DMSP
18638 Additionally the NO_RSA no longer completely removes the whole -modulus
18639 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18640 now, too.
18641
18642 *Ralf S. Engelschall*
18643
18644 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18645 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18646
18647 *Arne Ansper <arne@ats.cyber.ee>*
18648
18649 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18650 to be added. Now both 'req' and 'ca' can use new objects defined in the
18651 config file.
18652
18653 *Steve Henson*
18654
18655 * Add cool BIO that does syslog (or event log on NT).
18656
18657 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18658
18659 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18660 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18661 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18662 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18663
18664 *Ben Laurie*
18665
18666 * Add preliminary config info for new extension code.
18667
18668 *Steve Henson*
18669
18670 * Make RSA_NO_PADDING really use no padding.
18671
18672 *Ulf Moeller <ulf@fitug.de>*
18673
18674 * Generate errors when private/public key check is done.
18675
18676 *Ben Laurie*
18677
18678 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18679 for some CRL extensions and new objects added.
18680
18681 *Steve Henson*
18682
18683 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18684 key usage extension and fuller support for authority key id.
18685
18686 *Steve Henson*
18687
18688 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18689 padding method for RSA, which is recommended for new applications in PKCS
18690 #1 v2.0 (RFC 2437, October 1998).
18691 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18692 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18693 against Bleichbacher's attack on RSA.
18694 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
257e9d03 18695 Ben Laurie*
5f8e6c50
DMSP
18696
18697 * Updates to the new SSL compression code
18698
18699 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18700
18701 * Fix so that the version number in the master secret, when passed
18702 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18703 (because the server will not accept higher), that the version number
18704 is 0x03,0x01, not 0x03,0x00
18705
18706 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18707
ec2bfb7d
DDO
18708 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18709 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18710 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
5f8e6c50
DMSP
18711
18712 *Steve Henson*
18713
18714 * Support for RAW extensions where an arbitrary extension can be
ec2bfb7d 18715 created by including its DER encoding. See `apps/openssl.cnf` for
5f8e6c50
DMSP
18716 an example.
18717
18718 *Steve Henson*
18719
18720 * Make sure latest Perl versions don't interpret some generated C array
18721 code as Perl array code in the crypto/err/err_genc.pl script.
18722
18723 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18724
18725 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18726 not many people have the assembler. Various Win32 compilation fixes and
18727 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18728 build instructions.
18729
18730 *Steve Henson*
18731
18732 * Modify configure script 'Configure' to automatically create crypto/date.h
18733 file under Win32 and also build pem.h from pem.org. New script
18734 util/mkfiles.pl to create the MINFO file on environments that can't do a
18735 'make files': perl util/mkfiles.pl >MINFO should work.
18736
18737 *Steve Henson*
18738
18739 * Major rework of DES function declarations, in the pursuit of correctness
18740 and purity. As a result, many evil casts evaporated, and some weirdness,
18741 too. You may find this causes warnings in your code. Zapping your evil
18742 casts will probably fix them. Mostly.
18743
18744 *Ben Laurie*
18745
18746 * Fix for a typo in asn1.h. Bug fix to object creation script
18747 obj_dat.pl. It considered a zero in an object definition to mean
18748 "end of object": none of the objects in objects.h have any zeros
18749 so it wasn't spotted.
18750
18751 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18752
18753 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18754 Masking (CBCM). In the absence of test vectors, the best I have been able
18755 to do is check that the decrypt undoes the encrypt, so far. Send me test
18756 vectors if you have them.
18757
18758 *Ben Laurie*
18759
18760 * Correct calculation of key length for export ciphers (too much space was
18761 allocated for null ciphers). This has not been tested!
18762
18763 *Ben Laurie*
18764
18765 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18766 message is now correct (it understands "crypto" and "ssl" on its
18767 command line). There is also now an "update" option. This will update
18768 the util/ssleay.num and util/libeay.num files with any new functions.
18769 If you do a:
18770 perl util/mkdef.pl crypto ssl update
18771 it will update them.
18772
18773 *Steve Henson*
18774
257e9d03 18775 * Overhauled the Perl interface:
5f8e6c50
DMSP
18776 - ported BN stuff to OpenSSL's different BN library
18777 - made the perl/ source tree CVS-aware
18778 - renamed the package from SSLeay to OpenSSL (the files still contain
18779 their history because I've copied them in the repository)
18780 - removed obsolete files (the test scripts will be replaced
18781 by better Test::Harness variants in the future)
18782
18783 *Ralf S. Engelschall*
18784
18785 * First cut for a very conservative source tree cleanup:
18786 1. merge various obsolete readme texts into doc/ssleay.txt
18787 where we collect the old documents and readme texts.
18788 2. remove the first part of files where I'm already sure that we no
18789 longer need them because of three reasons: either they are just temporary
18790 files which were left by Eric or they are preserved original files where
18791 I've verified that the diff is also available in the CVS via "cvs diff
18792 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18793 the crypto/md/ stuff).
18794
18795 *Ralf S. Engelschall*
18796
18797 * More extension code. Incomplete support for subject and issuer alt
18798 name, issuer and authority key id. Change the i2v function parameters
18799 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18800 what that's for :-) Fix to ASN1 macro which messed up
18801 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18802
18803 *Steve Henson*
18804
18805 * Preliminary support for ENUMERATED type. This is largely copied from the
18806 INTEGER code.
18807
18808 *Steve Henson*
18809
18810 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18811
18812 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18813
257e9d03 18814 * Make sure `make rehash` target really finds the `openssl` program.
5f8e6c50
DMSP
18815
18816 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18817
18818 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18819 like to hear about it if this slows down other processors.
18820
18821 *Ben Laurie*
18822
18823 * Add CygWin32 platform information to Configure script.
18824
18825 *Alan Batie <batie@aahz.jf.intel.com>*
18826
257e9d03 18827 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
5f8e6c50
DMSP
18828
18829 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18830
18831 * New program nseq to manipulate netscape certificate sequences
18832
18833 *Steve Henson*
18834
18835 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18836 few typos.
18837
18838 *Steve Henson*
18839
18840 * Fixes to BN code. Previously the default was to define BN_RECURSION
18841 but the BN code had some problems that would cause failures when
18842 doing certificate verification and some other functions.
18843
18844 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18845
18846 * Add ASN1 and PEM code to support netscape certificate sequences.
18847
18848 *Steve Henson*
18849
18850 * Add ASN1 and PEM code to support netscape certificate sequences.
18851
18852 *Steve Henson*
18853
18854 * Add several PKIX and private extended key usage OIDs.
18855
18856 *Steve Henson*
18857
18858 * Modify the 'ca' program to handle the new extension code. Modify
18859 openssl.cnf for new extension format, add comments.
18860
18861 *Steve Henson*
18862
18863 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18864 and add a sample to openssl.cnf so req -x509 now adds appropriate
18865 CA extensions.
18866
18867 *Steve Henson*
18868
18869 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18870 error code, add initial support to X509_print() and x509 application.
18871
18872 *Steve Henson*
18873
18874 * Takes a deep breath and start adding X509 V3 extension support code. Add
18875 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18876 stuff is currently isolated and isn't even compiled yet.
18877
18878 *Steve Henson*
18879
18880 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18881 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18882 Removed the versions check from X509 routines when loading extensions:
18883 this allows certain broken certificates that don't set the version
18884 properly to be processed.
18885
18886 *Steve Henson*
18887
18888 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18889 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18890 can still be regenerated with "make depend".
18891
18892 *Ben Laurie*
18893
18894 * Spelling mistake in C version of CAST-128.
18895
18896 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18897
18898 * Changes to the error generation code. The perl script err-code.pl
18899 now reads in the old error codes and retains the old numbers, only
18900 adding new ones if necessary. It also only changes the .err files if new
18901 codes are added. The makefiles have been modified to only insert errors
18902 when needed (to avoid needlessly modifying header files). This is done
18903 by only inserting errors if the .err file is newer than the auto generated
18904 C file. To rebuild all the error codes from scratch (the old behaviour)
18905 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18906 or delete all the .err files.
18907
18908 *Steve Henson*
18909
18910 * CAST-128 was incorrectly implemented for short keys. The C version has
18911 been fixed, but is untested. The assembler versions are also fixed, but
18912 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18913 to regenerate it if needed.
18914 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18915 Hagino <itojun@kame.net>*
18916
18917 * File was opened incorrectly in randfile.c.
18918
18919 *Ulf Möller <ulf@fitug.de>*
18920
18921 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18922 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18923 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18924 al: it's just almost always a UTCTime. Note this patch adds new error
18925 codes so do a "make errors" if there are problems.
18926
18927 *Steve Henson*
18928
18929 * Correct Linux 1 recognition in config.
18930
18931 *Ulf Möller <ulf@fitug.de>*
18932
18933 * Remove pointless MD5 hash when using DSA keys in ca.
18934
18935 *Anonymous <nobody@replay.com>*
18936
18937 * Generate an error if given an empty string as a cert directory. Also
18938 generate an error if handed NULL (previously returned 0 to indicate an
18939 error, but didn't set one).
18940
18941 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18942
18943 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18944
18945 *Ben Laurie*
18946
18947 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18948 parameters. This was causing a warning which killed off the Win32 compile.
18949
18950 *Steve Henson*
18951
18952 * Remove C++ style comments from crypto/bn/bn_local.h.
18953
18954 *Neil Costigan <neil.costigan@celocom.com>*
18955
18956 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18957 based on a text string, looking up short and long names and finally
18958 "dot" format. The "dot" format stuff didn't work. Added new function
18959 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18960 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18961 OID is not part of the table.
18962
18963 *Steve Henson*
18964
18965 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18966 X509_LOOKUP_by_alias().
18967
18968 *Ben Laurie*
18969
18970 * Sort openssl functions by name.
18971
18972 *Ben Laurie*
18973
ec2bfb7d 18974 * Get the `gendsa` command working and add it to the `list` command. Remove
5f8e6c50
DMSP
18975 encryption from sample DSA keys (in case anyone is interested the password
18976 was "1234").
18977
18978 *Steve Henson*
18979
257e9d03 18980 * Make *all* `*_free` functions accept a NULL pointer.
5f8e6c50
DMSP
18981
18982 *Frans Heymans <fheymans@isaserver.be>*
18983
18984 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18985 NULL pointers.
18986
18987 *Anonymous <nobody@replay.com>*
18988
18989 * s_server should send the CAfile as acceptable CAs, not its own cert.
18990
18991 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18992
ec2bfb7d 18993 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
5f8e6c50
DMSP
18994
18995 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18996
18997 * Temp key "for export" tests were wrong in s3_srvr.c.
18998
18999 *Anonymous <nobody@replay.com>*
19000
19001 * Add prototype for temp key callback functions
19002 SSL_CTX_set_tmp_{rsa,dh}_callback().
19003
19004 *Ben Laurie*
19005
19006 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19007 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19008
19009 *Steve Henson*
19010
19011 * X509_name_add_entry() freed the wrong thing after an error.
19012
19013 *Arne Ansper <arne@ats.cyber.ee>*
19014
19015 * rsa_eay.c would attempt to free a NULL context.
19016
19017 *Arne Ansper <arne@ats.cyber.ee>*
19018
19019 * BIO_s_socket() had a broken should_retry() on Windoze.
19020
19021 *Arne Ansper <arne@ats.cyber.ee>*
19022
19023 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19024
19025 *Arne Ansper <arne@ats.cyber.ee>*
19026
19027 * Make sure the already existing X509_STORE->depth variable is initialized
19028 in X509_STORE_new(), but document the fact that this variable is still
19029 unused in the certificate verification process.
19030
19031 *Ralf S. Engelschall*
19032
ec2bfb7d 19033 * Fix the various library and `apps/` files to free up pkeys obtained from
5f8e6c50
DMSP
19034 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19035
19036 *Steve Henson*
19037
19038 * Fix reference counting in X509_PUBKEY_get(). This makes
19039 demos/maurice/example2.c work, amongst others, probably.
19040
19041 *Steve Henson and Ben Laurie*
19042
ec2bfb7d 19043 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
257e9d03 19044 `openssl` and second, the shortcut symlinks for the `openssl <command>`
5f8e6c50 19045 are no longer created. This way we have a single and consistent command
257e9d03 19046 line interface `openssl <command>`, similar to `cvs <command>`.
5f8e6c50
DMSP
19047
19048 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19049
19050 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19051 BIT STRING wrapper always have zero unused bits.
19052
19053 *Steve Henson*
19054
19055 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19056
19057 *Steve Henson*
19058
19059 * Make the top-level INSTALL documentation easier to understand.
19060
19061 *Paul Sutton*
19062
19063 * Makefiles updated to exit if an error occurs in a sub-directory
19064 make (including if user presses ^C) [Paul Sutton]
19065
19066 * Make Montgomery context stuff explicit in RSA data structure.
19067
19068 *Ben Laurie*
19069
19070 * Fix build order of pem and err to allow for generated pem.h.
19071
19072 *Ben Laurie*
19073
19074 * Fix renumbering bug in X509_NAME_delete_entry().
19075
19076 *Ben Laurie*
19077
19078 * Enhanced the err-ins.pl script so it makes the error library number
19079 global and can add a library name. This is needed for external ASN1 and
19080 other error libraries.
19081
19082 *Steve Henson*
19083
19084 * Fixed sk_insert which never worked properly.
19085
19086 *Steve Henson*
19087
19088 * Fix ASN1 macros so they can handle indefinite length constructed
19089 EXPLICIT tags. Some non standard certificates use these: they can now
19090 be read in.
19091
19092 *Steve Henson*
19093
19094 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19095 into a single doc/ssleay.txt bundle. This way the information is still
19096 preserved but no longer messes up this directory. Now it's new room for
19097 the new set of documentation files.
19098
19099 *Ralf S. Engelschall*
19100
19101 * SETs were incorrectly DER encoded. This was a major pain, because they
19102 shared code with SEQUENCEs, which aren't coded the same. This means that
19103 almost everything to do with SETs or SEQUENCEs has either changed name or
19104 number of arguments.
19105
19106 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19107
19108 * Fix test data to work with the above.
19109
19110 *Ben Laurie*
19111
19112 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19113 was already fixed by Eric for 0.9.1 it seems.
19114
19115 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19116
19117 * Autodetect FreeBSD3.
19118
19119 *Ben Laurie*
19120
19121 * Fix various bugs in Configure. This affects the following platforms:
19122 nextstep
19123 ncr-scde
19124 unixware-2.0
19125 unixware-2.0-pentium
19126 sco5-cc.
19127
19128 *Ben Laurie*
19129
19130 * Eliminate generated files from CVS. Reorder tests to regenerate files
19131 before they are needed.
19132
19133 *Ben Laurie*
19134
19135 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19136
19137 *Ben Laurie*
19138
257e9d03 19139### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
5f8e6c50
DMSP
19140
19141 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19142 changed SSLeay to OpenSSL in version strings.
19143
19144 *Ralf S. Engelschall*
19145
19146 * Some fixups to the top-level documents.
19147
19148 *Paul Sutton*
19149
19150 * Fixed the nasty bug where rsaref.h was not found under compile-time
19151 because the symlink to include/ was missing.
19152
19153 *Ralf S. Engelschall*
19154
19155 * Incorporated the popular no-RSA/DSA-only patches
19156 which allow to compile a RSA-free SSLeay.
19157
19158 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19159
257e9d03 19160 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
5f8e6c50
DMSP
19161 when "ssleay" is still not found.
19162
19163 *Ralf S. Engelschall*
19164
19165 * Added more platforms to Configure: Cray T3E, HPUX 11,
19166
19167 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19168
19169 * Updated the README file.
19170
19171 *Ralf S. Engelschall*
19172
19173 * Added various .cvsignore files in the CVS repository subdirs
19174 to make a "cvs update" really silent.
19175
19176 *Ralf S. Engelschall*
19177
19178 * Recompiled the error-definition header files and added
19179 missing symbols to the Win32 linker tables.
19180
19181 *Ralf S. Engelschall*
19182
19183 * Cleaned up the top-level documents;
19184 o new files: CHANGES and LICENSE
19185 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19186 o merged COPYRIGHT into LICENSE
19187 o removed obsolete TODO file
19188 o renamed MICROSOFT to INSTALL.W32
19189
19190 *Ralf S. Engelschall*
19191
19192 * Removed dummy files from the 0.9.1b source tree:
19193 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19194 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19195 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19196 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19197 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19198
19199 *Ralf S. Engelschall*
19200
19201 * Added various platform portability fixes.
19202
19203 *Mark J. Cox*
19204
19205 * The Genesis of the OpenSSL rpject:
19206 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19207 Young and Tim J. Hudson created while they were working for C2Net until
19208 summer 1998.
19209
19210 *The OpenSSL Project*
19211
257e9d03 19212### Changes between 0.9.0b and 0.9.1b [not released]
5f8e6c50
DMSP
19213
19214 * Updated a few CA certificates under certs/
19215
19216 *Eric A. Young*
19217
19218 * Changed some BIGNUM api stuff.
19219
19220 *Eric A. Young*
19221
19222 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19223 DGUX x86, Linux Alpha, etc.
19224
19225 *Eric A. Young*
19226
19227 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19228 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19229 available).
19230
19231 *Eric A. Young*
19232
19233 * Add -strparse option to asn1pars program which parses nested
19234 binary structures
19235
19236 *Dr Stephen Henson <shenson@bigfoot.com>*
19237
19238 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19239
19240 *Eric A. Young*
19241
19242 * DSA fix for "ca" program.
19243
19244 *Eric A. Young*
19245
19246 * Added "-genkey" option to "dsaparam" program.
19247
19248 *Eric A. Young*
19249
19250 * Added RIPE MD160 (rmd160) message digest.
19251
19252 *Eric A. Young*
19253
19254 * Added -a (all) option to "ssleay version" command.
19255
19256 *Eric A. Young*
19257
19258 * Added PLATFORM define which is the id given to Configure.
19259
19260 *Eric A. Young*
19261
19262 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19263
19264 *Eric A. Young*
19265
19266 * Extended the ASN.1 parser routines.
19267
19268 *Eric A. Young*
19269
19270 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19271
19272 *Eric A. Young*
19273
19274 * Added a BN_CTX to the BN library.
19275
19276 *Eric A. Young*
19277
19278 * Fixed the weak key values in DES library
19279
19280 *Eric A. Young*
19281
19282 * Changed API in EVP library for cipher aliases.
19283
19284 *Eric A. Young*
19285
19286 * Added support for RC2/64bit cipher.
19287
19288 *Eric A. Young*
19289
19290 * Converted the lhash library to the crypto/mem.c functions.
19291
19292 *Eric A. Young*
19293
19294 * Added more recognized ASN.1 object ids.
19295
19296 *Eric A. Young*
19297
19298 * Added more RSA padding checks for SSL/TLS.
19299
19300 *Eric A. Young*
19301
19302 * Added BIO proxy/filter functionality.
19303
19304 *Eric A. Young*
19305
19306 * Added extra_certs to SSL_CTX which can be used
19307 send extra CA certificates to the client in the CA cert chain sending
19308 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19309
19310 *Eric A. Young*
19311
19312 * Now Fortezza is denied in the authentication phase because
19313 this is key exchange mechanism is not supported by SSLeay at all.
19314
19315 *Eric A. Young*
19316
19317 * Additional PKCS1 checks.
19318
19319 *Eric A. Young*
19320
19321 * Support the string "TLSv1" for all TLS v1 ciphers.
19322
19323 *Eric A. Young*
19324
19325 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19326 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19327
19328 *Eric A. Young*
19329
19330 * Fixed a few memory leaks.
19331
19332 *Eric A. Young*
19333
19334 * Fixed various code and comment typos.
19335
19336 *Eric A. Young*
19337
19338 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19339 bytes sent in the client random.
19340
19341 *Edward Bishop <ebishop@spyglass.com>*
44652c16 19342
44652c16
DMSP
19343<!-- Links -->
19344
1e13198f 19345[CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
6ffc3127 19346[CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
44652c16
DMSP
19347[CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19348[CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19349[CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19350[CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19351[CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19352[CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19353[CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19354[CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19355[CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19356[CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19357[CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19358[CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19359[CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19360[CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19361[CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19362[CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19363[CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19364[CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19365[CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19366[CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19367[CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19368[CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19369[CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19370[CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19371[CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19372[CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19373[CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19374[CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19375[CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19376[CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19377[CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19378[CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19379[CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19380[CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19381[CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19382[CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19383[CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19384[CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19385[CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19386[CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19387[CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19388[CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19389[CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19390[CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19391[CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19392[CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19393[CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19394[CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19395[CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19396[CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19397[CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19398[CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19399[CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19400[CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19401[CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19402[CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19403[CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19404[CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19405[CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19406[CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19407[CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19408[CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19409[CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19410[CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19411[CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19412[CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19413[CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19414[CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19415[CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19416[CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19417[CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19418[CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19419[CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19420[CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19421[CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19422[CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19423[CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19424[CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19425[CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19426[CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19427[CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19428[CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19429[CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19430[CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19431[CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19432[CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19433[CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19434[CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19435[CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19436[CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19437[CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19438[CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19439[CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19440[CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19441[CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19442[CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19443[CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19444[CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19445[CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19446[CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19447[CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19448[CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19449[CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19450[CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19451[CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19452[CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19453[CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19454[CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19455[CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19456[CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19457[CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19458[CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19459[CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19460[CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19461[CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19462[CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19463[CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19464[CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19465[CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19466[CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19467[CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19468[CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19469[CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19470[CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19471[CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19472[CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19473[CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19474[CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19475[CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19476[CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19477[CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19478[CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19479[CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19480[CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19481[CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19482[CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19483[CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19484[CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19485[CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19486[CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19487[CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19488[CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19489[CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19490[CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19491[CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19492[CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19493[CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19494[CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19495[CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19496[CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19497[CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19498[CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19499[CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19500[CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19501[CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19502[CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19503[CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19504[CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19505[CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19506[CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655